article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
simulations of lattice qcd ( lqcd ) are in practice the only reliable approach to qcd thermodynamics which covers the broad region of strongly interacting matter properties from the hadron gas at low temperatures to a deconfined quark gluon plasma phase at high temperatures .
recently , finite temperature lqcd simulations have overcome the difficulties of reaching the low physical light quark masses and approaching the continuum limit which makes this theoretical laboratory now a benchmark for modeling qcd under extreme conditions .
we are going to present a combined effective model reproducing the equation of state of hadronic matter as obtained in recent lattice qcd simulations @xcite .
the model should reproduce basic physical characteristics of processes encountered in dense hadronic matter , from the hot qcd phase through the critical temperature region till the lower temperature hadron resonance gas phase . in medium properties of hadrons are different from those in the vacuum .
the very notion of the mass shell should be modified , as was postulated quite long ago @xcite .
the interaction becomes effectively nonlocal due to the mott effect and hadrons eventually gradually dissolve into quarks and gluons in the high temperature phase .
then , with the increasing temperature , quark masses are less and less important although the massless stefan - boltzmann limit would be eventually reached only at extremely high temperature
. it has been shown that the equation of state derived from that time qcd lattice calculation @xcite can be reproduced by a simple hadron gas resonance model . the rapid rise of the number of degrees of freedom in lattice qcd data around the critical temperature @xmath0 mev , can be explained quantitatively by a resonance gas below the critical temperature @xmath1 @xcite . for higher temperatures the model
is modified by introducing finite widths of heavy hadrons @xcite with a heuristic ansatz for the spectral function which reflects medium modifications of hadrons .
this fits nicely the lattice data , also above @xmath1 as is shown in fig .
[ fig.1 ] .
this mott - hagedorn type model @xcite has been constructed to fit nicely the lattice data , also above @xmath1 where it does so because it leaves light hadrons below a mass threshold of @xmath2 gev unaffected .
the description of the lattice data at high temperatures is accidental because the effective number of those degrees of freedom approximately coincides with that of quarks and gluons .
the qgp presence in this region is formally simulated here by the smart choice of the mass cut - off parameter such that cut - off defined stable light hadrons provide the same number of degrees of freedom as partonic components of the qgp .
the considered model is here gradually refined to take into account those physical processes present in the full qcd treatment .
the uniform treatment of all hadronic resonances , without artificial stability island , is reached by a state - dependent hadron resonance width @xmath3 given by the inverse collision time scale in a resonance gas @xcite . in order to remove this unphysical aspect of the otherwise appealing model one has to extend the spectral broadening also to the light hadrons and
thus describe their disappearance due to the mott effect while simultaneously the quark and gluon degrees of freedom appear at high temperatures due to chiral symmetry restoration and deconfinement . in the present contribution we will report results obtained by introducing a unified treatment of all hadronic resonances with a state - dependent width @xmath3 in accordance with the inverse hadronic collision time scale from a recent model for chemical freeze - out in a resonance gas @xcite .
the appearance of quark and gluon degrees of freedom is introduced by the polyakov - loop improved nambu
jona - lasinio ( pnjl ) model @xcite .
the model is further refined by adding perturbative corrections to @xmath4 for the high - momentum region above the three - momentum cutoff inherent in the pnjl model .
one obtains eventually a good agreement with lattice qcd data , comparable with all important physical characteristics taken into account .
we introduce the width @xmath5 of a resonance in the statistical model through the spectral function @xmath6 a breit - wigner distribution of virtual masses with a maximum at @xmath7 and the normalization factor @xmath8^{-1 } \nonumber\\ % & = & \frac{1 } { \frac{\pi}{2 } + \arctan \left ( \frac{m^2 - m^2_0 } { \gamma \cdot m } \right ) } \ , .
% \end{aligned}\ ] ] and the model ansatz for the resonance width @xmath5 is given by @xcite @xmath9 where @xmath10 mev , @xmath11 , @xmath12 and the hagedorn temperature @xmath13 mev .
the energy density of this model with zero resonance proper volume for given temperature @xmath14 and chemical potentials : @xmath15 for baryon number and @xmath16 for strangeness , can be cast in the form @xmath17 where @xmath18 gev and the energy density per degree of freedom with a mass @xmath19 is @xmath20 with the degeneracy @xmath21 and the chemical potential @xmath22 of hadron @xmath23 . for mesons , @xmath24 and for baryons @xmath25 . according to eq .
( [ one ] ) the energy density of hadrons consists of the contribution of light hadrons for @xmath26 and the contribution of heavier hadrons smeared with the spectral function for @xmath27 . for simplicity , we assume @xmath28 for the strangeness number density and @xmath29 for the baryon number density . then @xmath30 and @xmath31 always , so the temperature is the only significant statistical parameter here . in such a case and for a fixed volume
we have @xmath32 where @xmath33 and @xmath34 are the pressure and entropy density , respectively .
this is the first order ordinary differential equation for the pressure and the general solution reads @xmath35 where @xmath36 . to have well - defined solution for the initial temperature
@xmath37 one has to assume that @xmath38 .
then @xmath39 and the entropy density reads : @xmath40 where @xmath41 .
we put @xmath42 as suggested by the nernst postulate .
the sound velocity squared is given by @xmath43 in fig .
[ fig.1 ] we show the results for the thermodynamic quantities ( pressure , energy density and squared speed of sound ) of the mhrg model at this stage .
the nice correspondence with results from lattice qcd is not accidental for the temperature region @xmath44 mev , where has been shown in @xcite that a hadron resonance gas perfectly describes the lattice qcd data . for @xmath45
the broadening of the spectral function ( [ one ] ) which affects at this stage of the model only the hadronic resonances with @xmath46 leads to the vanishing of their contribution at about @xmath47 while the light hadrons with masses @xmath48 are not affected and gradually reach the stefan- boltzmann ( sb ) limit determined by their number of degrees of freedom . as has been noted in @xcite , this number ( @xmath49 ) accidentally ( or by duality arguments )
coincides with that of the quarks and gluons ( @xmath50 ) for @xmath51 .
therefore , imposing that all mesons lighter than @xmath2 gev are stable provides us with a sb limit at high temperatures which fakes that of quarks and gluons in the case for three flavors .
although providing us with an excellent fit of the lattice data , the high - temperature phase of this model is unphysical since it ignores the mott effect for light hadrons . due to the chiral phase transition at @xmath1 , the quarks loose their mass and therefore the threshold of the continuum of quark - antiquark scattering states is lowered . at the same time
the light meson masses , however , remain almost unaffected by the increase in the temperature of the system .
consequently , they merge the continuum and become unbound - their spectral function changes from a delta - function ( on - shell bound states ) to a breit - wigner - type ( off - shell , resonant scattering states ) .
this phenomenon is the hadronic analogue @xcite of the mott - anderson transition for electrons in solid state physics ( insulator - conductor transition ) .
( color online ) thermodynamic quantities for the old mott - hagedorn resonance gas model @xcite .
different line styles correspond to different values for the parameter @xmath52 in the range from @xmath53 ( dashed line ) to @xmath54 ( solid line ) .
lattice qcd data are from ref .
, scaledwidth=70.0% ] it has been first introduced for the hadronic - to - quark - matter transition in @xcite .
later , within the njl model , a microscopic approach to the thermodynamics of the mott dissociation of mesons in quark matter has been given in the form of a generalized beth - uhlenbeck equation of state @xcite , see also @xcite .
as a microscopic treatment of the mott effect for all resonances is presently out of reach , we introduce an ansatz for a state - dependent hadron resonance width @xmath3 given by the inverse collision time scale recently suggested within an approach to the chemical freeze - out and chiral condensate in a resonance gas @xcite @xmath55 which is based on a binary collision approximation and relaxation time ansatz using for the in - medium hadron - hadron cross sections the geometrical povh - hfner law @xcite . in eq .
( [ gamma ] ) the coefficient @xmath56 is a free parameter , @xmath57 is the partial density of the hadron @xmath58 and the mean squared radii of hadrons @xmath59 obtain in the medium a temperature dependence which is governed by the ( partial ) restoration of chiral symmetry . for the pion this was quantitatively studied within the njl model @xcite and it was shown that close to the mott transition the pion radius is well approximated by @xmath60 and a pion cloud contribution @xmath61 where from the vacuum values @xmath62 fm and @xmath63 fm one gets @xmath64 fm .
a key point of our approach is that the temperature dependent hadronic radii shall diverge when hadron dissociation ( mott effect ) sets in , driven basically by the restoration of chiral symmetry . as a consequence , in the vicinity of the chiral restoration temperature
all meson radii shall behave like that of the pion and all baryon radii like that of the nucleon .
the resulting energy density behaviour is shown in fig .
[ fig.1a ] .
( color online ) energy density ( red lines and symbols ) and pressure ( black lines and symbols ) for the state - dependent width model of eq .
( [ gamma ] ) and three values of the mass threshold @xmath65 : 1 gev ( solid lines ) , 980 mev ( dashed lines ) , 0 ( dash - dotted lines ) .
lattice qcd data are from ref .
@xcite.,scaledwidth=70.0% ] this part of the model we call mott - hagedorn - resonance - gas ( mhrg ) .
when all hadrons are gone at @xmath66 mev , we are clearly missing degrees of freedom !
we improve the pnjl model over its standard versions @xcite by adding perturbative corrections in @xmath4 for the high - momentum region above the three - momentum cutoff @xmath67 . in the second step ,
the mhrg part is replaced by its final form , using the state - dependent spectral function for the description of the mott dissociation of all hadron resonances above the chiral transition .
the total pressure obtains the form @xmath68 where @xmath69 stands for the pressure of the mhrg model , accounting for the dissociation of hadrons in hot dense , matter .
the @xmath4 corrections can be split in quark and gluon contributions @xmath70 where @xmath71 stands for the quark contribution and @xmath72 contains the ghost and gluon contributions .
the total perturbative qcd correction to @xmath4 is @xmath73 where @xmath74 .
the corresponding contribution to the energy density is given in standard way by eq .
( [ six ] ) .
we will now include an effective description of the dissociation of hadrons due to the mott effect into the hadron resonance gas model by including the state dependent hadron resonance width ( [ gamma ] ) into the definition of the hrg pressure @xmath75 from the pressure as a thermodynamic potential all relevant thermodynamical functions can be obtained . combining the @xmath76 corrected meanfield pnjl model for the quark - gluon subsystem with the mhrg description of the hadronic resonances we obtain the results shown in the right panel of fig . [ fig.1 ] where the resulting partial contributions in comparison with lattice qcd data from ref .
@xcite are shown .
we see that the lattice qcd thermodynamics is in full accordance with a hadron resonance gas up to a temperature of @xmath77 mev which corresponds to the pseudocritical temperature of the chiral phase transition .
the lattice data saturate below the stefan - boltzmann limit of an ideal quark - gluon gas at high temperatures .
the pnjl model , however , attains this limit by construction .
the deviation is to good accuracy described by perturbative corrections to @xmath4 which vanish at low temperatures due to an infrared cutoff procedure .
the transition region @xmath78\le 250 $ ] is described by the mhrg model , resulting in a decreasing hrg pressure which vanishes at @xmath79 mev .
( color online ) thermodynamic quantities for the new mott - hagedorn resonance gas where quark - gluon plasma contributions are described within the pnjl model including @xmath76 corrections ( dashed lines ) .
hadronic resonances are described within the resonance gas with finite width , as an implementation of the mott effect ( dash - dotted line ) .
the sum of both contributions ( solid lines ) is shown for the energy density ( thick lines ) and pressure ( thin lines ) in comparison with the lattice data from @xcite .
, scaledwidth=70.0% ] we have presented two stages of an effective model description of qcd thermodynamics at finite temperatures which properly accounts for the fact that in the qcd transition region it is dominated by a tower of hadronic resonances . to this end we have further developed a generalization of the hagedorn resonance gas thermodynamics which includes the finite lifetime of hadronic resonances in a hot and dense medium by a model ansatz for a temperature- and mass dependent spectral function .
( color online ) thermodynamic quantities as in fig . [ fig.1 ] for the mhrg - pnjl model compared to lattice data from ref . @xcite .
, scaledwidth=70.0% ] after having presented the mhrg - pnjl model with the state - dependent spectral function approach we show the summary of the thermodynamic quantities in fig .
[ fig.3 ] .
we have presented two stages of an effective model description of qcd thermodynamics at finite temperatures which properly accounts for the fact that in the qcd transition region it is dominated by a tower of hadronic resonances . in the first of the two stages of the developments we presented here , we have used the fact that the number of low - lying mesonic degrees of freedom with masses below @xmath80 gev approximately equals that of the thermodynamic degrees of freedom of a gas of quark and gluons . in the second one
we have further developed a generalization of the hagedorn resonance gas thermodynamics which includes the finite lifetime of heavy resonances in a hot and dense medium by a model ansatz for a temperature- and mass dependent spectral function which is motivated by a model for the collision time successfully applied in the kinetic description of chemical freeze - out from a hadron resonance gas .
a next step should take into account also the effects of continuum correlations in hadronic scattering channels in accordance with the levinson theorem @xcite as discussed recently for the example of pion dissociation within the pnjl model @xcite .
our account for @xmath4 corrections from quark and gluon scattering in the plasma may be seen from this perspective .
d.b . wants to thank k.a .
bugaev , k. redlich and a. wergieluk for discussions and collaboration .
this work was supported in part by the polish national science center ( ncn ) under contract no .
n n202 0523 40 and the `` maestro '' programme dec-2011/02/a / st2/00306 as well as by the russian foundation for basic research under grant no .
11 - 02 - 01538-a ( d.b . ) .
+ s. borsanyi _ et al .
_ , jhep * 1011 * , 077 ( 2010 ) .
a. bazavov _ et al .
_ , phys .
d * 80 * , 014504 ( 2009 ) .
f. karsch , k. redlich and l. turko , z. phys . c * 60 * , 519 ( 1993 ) .
f. karsch , lect .
notes phys .
* 583 * , 209 ( 2002 ) ; [ hep - lat/0106019 ] .
f. karsch , k. redlich and a. tawfik , eur .
j. c * 29 * , 549 ( 2003 ) . c. ratti
_ et al . _
[ wuppertal - budapest collaboration ] , nucl .
a * 855 * , 253 ( 2011 ) [ arxiv:1012.5215 [ hep - lat ] ] .
d. b. blaschke , k. a. bugaev , fizika b * 13 * , 491 ( 2004 ) ; + phys .
* 2 * , 305 ( 2005 ) .
d. b. blaschke , k. a. bugaev , phys .
* 2 * , 305 ( 2005 ) .
[ pisma fiz .
chast . atom .
yadra * 2 * , 69 ( 2005 ) ] .
l. turko , d. blaschke , d. prorok and j. berdermann , acta phys .
supp . * 5 * , 485 ( 2012 ) [ arxiv:1112.6408 [ nucl - th ] ] .
d. b. blaschke _ et al .
* 8 * , 811 ( 2011 ) .
k. fukushima , phys .
b * 591 * , 277 ( 2004 ) . c. ratti , m. a. thaler , w. weise , phys .
d * 73 * , 014019 ( 2006 ) .
g. e. brown , h. a. bethe , p. m. pizzochero , phys .
b * 263 * , 337 ( 1991 ) .
d. zablocki , d. blaschke and g. rpke , _ metal - to - nonmetal transitions _ , springer series in materials science , * 132 * , 161 ( 2010 ) . | a combined effective model reproducing the equation of state of hadronic matter as obtained in recent lattice qcd simulations is presented .
the model reproduces basic physical characteristics encountered in dense hadronic matter in the quark - gluon plasma ( qgp ) phase and the lower temperature hadron resonance gas phase .
the hadronic phase is described by means of an extended mott - hagedorn resonance gas while the qgp phase is described by the extended pnjl model .
the dissociation of hadrons is obtained by including the state dependent hadron resonance width .
pacs numbers : 12.38.gc , 12.38.mh , 12.40.ee , 24.85.+p |
single photons are one of the basic building blocks of quantum information technologies .
numerous experimental implementations of quantum computation schemes @xcite , quantum communication protocols @xcite as well as quantum metrology @xcite are based on single photons as the physical realization of the qubit . in order to code the state of such flying qubits , quantum interferences
are commonly exploited , which requires the photons to be in a pure state @xcite .
the creation of photon pairs is a crucial tool to many of the aforementioned applications , which are based on entangled 2-photon state ( e.g. entanglement swapping ) or require heralding the presence of a single photon by detecting its partner @xcite .
different approaches have been followed to produce photon pairs , e.g. quantum dot emission @xcite or atomic ensembles @xcite , but implementations are still quite cumbersome and in an early development stage .
frequency mixing processes in nonlinear optical media , however , are well understood , easier to realize and widely used @xcite .
still , a common drawback of all these techniques is the fact that photon pairs are in general created in a mixed state due to correlations in energy and momentum @xcite .
thus , in order to obtain a photon pair in a pure state , narrow spectral and spatial filtering is required .
this introduces additional losses which limits the overall efficiency @xmath1 of detecting a created photon . in experiments employing multiple photons ,
the probability @xmath2 of a successful detection of all n photons scales as @xmath3 . in practice
, this means that @xmath4 is usually limited to the order of 4 - 6 before prohibitively low n - fold coincidences render a successful experiment impossible within a reasonable period of time .
so far , high detection efficiency on the one hand and narrow filtering on the other hand have been mutually exclusive . in this paper , we present a source of photon pairs based on 4-wave mixing ( fwm ) in photonic crystal fibres ( pcf ) featuring both a high overall detection efficiency and narrowband photons . due to the careful engineering of the phase matching ,
the pairs are created in an intrinsically factorable state , i.e. free of correlations in any degree of freedom . in our case
this means , that signal and idler photon have decoupled spectral distributions without any frequency correlations between them .
hence , the detection of an idler photon heralds the presence of a signal photon in a pure state @xcite , which makes additional filters obsolete . as a consequence ,
losses are reduced and future multi - photon experiments with @xmath5 become possible .
in fwm , two photons from an intense pump field are converted into two daughter photons , signal and idler .
this process takes place according to conservation of energy and phase matching @xmath6 with @xmath7 the refractive index at pump , signal and idler frequency @xmath8 , respectively , @xmath9 the nonlinear coefficient of the fibre @xcite and @xmath10 the peak pump power .
fwm in pcf is a well understood process and has already been used for pair - photon generation .
both cases where the pump , signal and idler waves are all in the same polarization @xcite as well as in orthogonal polarization states @xcite have been studied . with an elliptical fibre core or an asymmetric cladding
, we obtain a slight birefringence and different effective indices @xmath11 along the fast ( @xmath12 ) and slow ( @xmath13 ) axis of the fibre . solving eq .
( [ eq : pm ] ) results in the phase matching curves as plotted in fig .
[ fig : phasematch](a ) , where _ _ ss__@xmath14_ss _ and _ _ ff__@xmath14_ff _ correspond to cases where all fields are polarized in the same direction ( @xmath15 ) and _ _ ss__@xmath14_ff _ and _ _ ff__@xmath14_ss _ correspond to cases where the pump field is orthogonal to the signal and idler fields ( @xmath16 ) .
if we ignore the wavelength dependence of the birefringence , @xmath17 , then to first order this appears in eq .
( [ eq : pm ] ) in the same way as the term @xmath18 , but with either a positive or negative sign , and usually much greater magnitude
. we fabricated several fibres with zero dispersion wavelengths close to 800 nm and different birefringence by controlling the hole sizes during the fibre draw @xcite .
the calculated phase matching curves for the fibre selected for these experiments are shown in fig .
[ fig : phasematch](a ) and an electron micrograph of the fibre is shown in fig .
[ fig : phasematch](b ) .
ff ( ss@xmath14ss ) represent the cases where pump and daughter photons are all polarized along the f ( s ) axis . note that the same curve is obtained for a non - birefringent fibre described by just one refractive index . in the case ff@xmath14ss ( ss@xmath14ff ) ,
orthogonally polarized pump and daughter photons see different modal indices which leads to a region in the phase matching curve featuring a horizontal tangent . at this point signal and idler photons
have no frequency correlation and are hence in a factorable state .
( b ) sem of the central region of the fibre used in these experiments .
the birefringence required was introduced by reducing the size of two holes adjacent to the core ( 11 oclock and 5 oclock).,title="fig:",width=264 ] ff ( ss@xmath14ss ) represent the cases where pump and daughter photons are all polarized along the f ( s ) axis .
note that the same curve is obtained for a non - birefringent fibre described by just one refractive index . in the case ff@xmath14ss ( ss@xmath14ff ) , orthogonally polarized pump and daughter photons see different modal indices which leads to a region in the phase matching curve featuring a horizontal tangent . at this point signal and idler photons
have no frequency correlation and are hence in a factorable state .
( b ) sem of the central region of the fibre used in these experiments .
the birefringence required was introduced by reducing the size of two holes adjacent to the core ( 11 oclock and 5 oclock).,title="fig:",height=128 ] for carefully controlled @xmath19 , @xmath20 and @xmath21 , we can achieve a situation where the group velocity of the pump photons ( e.g. in the @xmath13-axis ) equals the group velocity of one of the daughter photons , the idler in our case ( e.g. in the @xmath12-axis ) . in this configuration , due to matching group velocities of pump and idler light ,
the idler photon is maximally confined in time and hence minimally in energy , and vice versa for the signal . as a result ,
the entire energy uncertainty of the photon pair is concentrated on the idler photon and the temporal uncertainty on the signal photon .
furthermore , over a limited range the signal wavelength now becomes independent of the pump wavelength , which translates into a horizontal tangent on the phase matching curve , as can be seen in fig .
[ fig : phasematch](a ) for @xmath19=705 nm . this intrinsic lack of energy correlation between the two daughter photons means that their joint spectral amplitude ( jsa ) can be written in the factorable state .
see also @xcite for more details .
by thoroughly controlling the absolute and relative index of the two fibre axes via the core diameter and size of the air holes in the silica cladding , we are able to tailor the pcf to make it perfectly suited to our application .
this enables us to build a photon pair source with high single mode coupling efficiencies because no lossy filtering is required anymore and the photons , created and propagating in the fundamental spatial mode of the pcf , can efficiently be coupled into any other single mode fibre .
figure [ fig:3](c ) shows the jsa as a function of signal and idler wavelength , given by the product @xmath22 of the natural phase matching condition of the pcf @xmath23 ( fig .
[ fig:3](a ) ) and the pump envelope function @xmath24 ( fig .
[ fig:3](b ) ) with @xmath25 and @xmath26 the detuning from the perfect phase matching @xmath27 and @xmath28 , respectively , and @xmath29 the length of the fibre @xcite . in this picture
, the jsa of an ideal factorable state with a gaussian spectral distribution is represented by an ellipse with symmetry along the horizontal and vertical axes .
however , in the real case ( fig . [ fig:3](c ) and [ fig : ripples ] ) we can see that although the central peak may be well approximated by an elliptical gaussian there are significant outlying peaks which can make the state partly correlated .
in order to show the suitability of our source for quantum communication or quantum gates @xcite , we will now demonstrate non - classical 2-photon interference .
recently , such quantum interferences have been reported for photons created in a single source @xcite . here , we show non - classical interference between photons originating from two separate sources , as required for example for long distance quantum relays or quantum repeaters @xcite .
the experimental setup is shown in fig .
[ fig : setup ] .
two separate 40 cm long pcfs ( pcf1 , pcf2 ) are pumped by a mode - locked ti : sapphire laser in fs - mode at 705 nm .
this gives rise to pairs of signal and idler photons at 597 nm and 860 nm respectively .
the bandwidth @xmath30 of the pump laser is filtered to 0.9 nm fwhm , resulting in a coherence time of @xmath31=0.8ps .
the intrinsic bandwidth of the signal photon created is as narrow as 0.13 nm corresponding to a coherence time @xmath32=4ps@xmath33 and the idler bandwidth is 2 nm .
[ fig : signalidler ] shows the spectra of the created photons , taken with a spectrometer and a liquid nitrogen - cooled ccd camera , integrated over 1s . at the output of each pcf ,
a polarizer in parallel orientation with respect to that of the created photons blocks the remaining pump field and a dichroic mirror ( dm ) separates the daughter photons in frequency .
the idler photons are passed through broadband interference filters ( bf , fwhm 10nm@xmath34 ) to reject spontaneous raman emission and residual pump light , and are then directed onto single - photon si - avalanche photon detectors ( apds ) .
the successful detection of an idler photon heralds the presence of a signal photon in the corresponding source .
such a 2-photon coincidence measurement is shown in @xcite .
the heralded signal photons are then sent into the two input ports of a fibre 50/50 beam splitter ( bs ) , passed through broad band filters ( bf , fwhm 40nm@xmath35 ) and detected by apds , one at each output port .
the overall collection efficiency of our signal and idler photons are 0.21 and 0.18 , respectively .
this includes all losses due to absorption in the pcf , transmission of the blocking filters ( 81% and 65% , respectively ) and dichroic mirror , coupling to single - mode fibres , and detector efficiencies ( 59% and 40% , respectively ) .
to demonstrate the intrinsic purity of the heralded photons we show a hong - ou - mandel interference @xcite between two signal photons originating from different sources and sent into different input ports of a 50/50 bs . for photons ,
indistinguishable in spatial mode , polarization , wavelength and arrival time , photon bunching occurs and leads to detection of the two photons always in the same output port of the bs .
hence the coincidence detection rate between the two output ports will drop to zero . to ensure the modal indistinguishability of the two signal photons , we use a single mode fibre bs for perfect spatial overlap .
the signal polarization in input port 1 is manually controlled in order to match the polarization mode in port 2 .
spectral overlap of the two signal photons is obtained by temperature controlling the pcfs independently by a peltier element in order to tune their phase matching condition and hence the wavelength of the emitted signal photons .
the temperature dependence of the central signal wavelength is 11pm/@xmath36c , making it relatively easy to control as a stability of @xmath37c suffices . by varying the relative arrival time of two
heralded signal photons at the bs ( @xmath38 ) and recording the 4-fold coincidence probability of detecting each photon in a separate apd , a hong - ou - mandel dip can be demonstrated .
the experimental result is shown in fig .
[ fig : dip ] .
note that the coincidence count rate for two heralded signal photons , one in each output port , is constant for @xmath39t@xmath400 , but dropping for photons arriving simultaneously at the bs ( @xmath39t=0 ) . between the signal photons impinging on the beam splitter .
the error bars are determined for a poisson distribution , given by the square root of the obtained count rates .
the solid curves is a fit through the experimental data with a lorentzian function resulting in a visibility of 76.2%.,width=264,height=151 ] fitting the uncorrected experimental data with a lorentzian function , a raw visibility for the non - classical interference of @xmath41 is observed .
the limited visibility can be attributed to different reasons : the non - perfect splitting ratio of the bs ( 1% ) , noise due to creation of multiple pairs ( @xmath421% ) and raman emission ( @xmath421% ) , and difference in the spectra of the two signal photons leading to distinguishability .
however the main reason is frequency - correlation in the jsa of a signal - idler pair ( @xmath4320% ) .
looking at fig .
[ fig:3](c ) and [ fig : ripples ] we see that the sinc function nature of the phase matching function -given by the fourier transform of the temporal ( top hat ) creation window of a pair in the pcf- leaves significant ripples along the energy matching direction .
hence in a small way signal and idler photons are correlated .
detection of an idler photon gives partial information on the signal photon .
this reduces the indistinguishability , and hence the interference visibility , of the heralded signal photons . using the theory derived in @xcite we have simulated our pcf and find good agreement with our experimental results @xcite .
this slight distinguishability increases the schmidt number @xmath44 , a measure of the frequency correlations between signal and idler photons ( see @xcite ) . for a factorable state @xmath45 and for any mixed state @xmath46 .
it can be shown that the visibility will approach 100% ( i.e. @xmath47 ) for increasing fibre lengths .
however , this only holds true for a theoretical pcf , homogeneous over an infinite length @xmath29 .
due to fluctuations during the fabrication process of our pcf , we are restricted to @xmath481 m , before phase matching conditions vary too much . another way to reduce distinguishability , and hence increase visibility ,
is the use of a pcf featuring a higher birefringence @xmath17 .
phase matching conditions in such a fibre result in a wider spectral separation of signal and idler photons .
consequently , detection efficiencies of the idler photons will decreased due to low sensitivity of the si - apd beyond 900 nm , leading to a reduced overall coincidence count rate .
we have demonstrated the creation of photon pairs via four wave mixing in photonic crystal fibres .
the pairs are created in an intrinsically factorable state without any spectral filtering .
this allows us to herald the presence of a photon in a pure state and to show hong - ou - mandel interference between photons originating from separate sources with a raw visibility of 76.2% .
furthermore we achieve an overall detection efficiency of up to 21% , making future multi - photon experiments feasible due to high multi - fold coincidence count rates .
for instance in a 6-photon setup with 80mhz repetition rate , 20% overall detection efficiency and 0.1 photons per pulse we will achieve a 6-fold coincidence count rate of 5.1 per second , sufficient for an experiment with a reasonably short integration time .
the work was supported by the uk epsrc ( qip irc and ep / f002424/1 ) , the eu integrated project qubit applications ( ip qap ) and acdet .
w.j.w . is a royal society university research fellow , j.g.r .
is supported by a wolfson merit award and m.h . is supported by an axa fellowship . | in this paper , we demonstrate a source of photon pairs based on four - wave - mixing in photonic crystal fibres . careful engineering of the phase matching conditions in the fibres enables us to create photon pairs at 597 nm and 860 nm in an intrinsically factorable state showing no spectral correlations .
this allows for heralding one photon in a pure state and hence renders narrow band filtering obsolete .
the source is narrow band , bright and achieves an overall detection efficiency of up to 21% per photon .
for the first time , a hong - ou - mandel interference with unfiltered photons from separate fibre sources is presented .
10 url # 1`#1`urlprefix[2][]#2 e. knill , r. laflamme , and g. j. milburn , `` a scheme for efficient quantum computation with linear optics , '' nature * 409 * , 4652 ( 2001 ) .
http://dx.doi.org/10.1038/35051009 .
n. gisin and r. thew , `` quantum communication , '' nat photon * 1 * , 165171 ( 2007 ) . http://dx.doi.org/10.1038/nphoton.2007.22 .
t. nagata , r. okamoto , j. l. obrien , k. sasaki , and s. takeuchi , `` beating the standard quantum limit with four - entangled photons , '' science * 316 * , 726729 ( 2007 ) .
http://www.sciencemag.org/cgi/content/abstract/316/5825/726 .
a. b. uren , c. silberhorn , r. erdmann , k. banaszek , w. p. grice , i. a. walmsley , and m. g. l raymer , `` generation of pure - state single - photon wavepackets by conditional preparation based on spontaneous parametric downconversion , '' laser phys . * 15 * , 146161 ( 2005 ) .
a. soujaeff , s. takeuchi , k. sasaki , t. hasegawa , and m. matsui , `` heralded single photon source at 1550@xmath0 nm from pulsed parametric down conversion , '' j. mod . opt . * 54 * , 467471 ( 2007 ) .
http://www.informaworld.com/10.1080/09500340600742239 .
r. m. stevenson , r. j. young , p. atkinson , k. cooper , d. a. ritchie , and a. shields , `` a semiconductor source of triggered entangled photon pairs , '' nature * 439 * , 179182 ( 2006 ) . j. k. thompson , j. simon , h. loh , and v. vuletic , `` a high - brightness source of narrowband , identical - photon pairs , '' science * 313 * , 7477 ( 2006 ) . d. c. burnham and d. l. weinberg , `` observation of simultaneity in parametric production of optical photon pairs , '' phys .
rev .
lett . * 25 * , 8487 ( 1970 ) . http://link.aps.org/abstract/prl/v25/p84 .
e. giacobino , c. fabre , a. heidmann , r. horowicz , s. reynaud , d. grandclment , g. grynberg , and m. pinard , `` generation of non classical states of light by phase conjugation and parametric conversion , '' hyperfine interactions * 37 * , 109124 ( 1987 ) .
http://dx.doi.org/10.1007/bf02395706 .
k. garay - palmett , h. j. mcguinness , o. cohen , j. s. lundeen , r. rangel - rojo , a. b. uren , m. g. raymer , c. j. mckinstrie , s. radic , and i. a. walmsley , `` photon pair - state preparation with tailored spectral properties by spontaneous four - wave mixing in photonic - crystal fiber , '' opt . express * 15 * , 1487014886 ( 2007 ) .
http://www.opticsexpress.org/abstract.cfm?uri=oe-15-22-14870 .
w. p. grice , a. b. uren , and i. a. walmsley , `` eliminating frequency and space - time correlations in multiphoton states , '' phys .
rev .
a * 64 * , 063815 ( 2001 ) .
http://link.aps.org/abstract/pra/v64/e063815 .
g. agrawal , `` nonlinear fiber optics , '' ( 2000 ) .
http://www.springerlink.com / content / q022707lwv86154v/. l. j. wang , c. k. hong , and s. r. friberg , `` generation of correlated photons via four - wave mixing in optical fibres , '' j. opt .
b : quantum and semiclassical optics * 3 * , 346352 ( 2001 ) .
j. fulconis , o. alibart , j. l. obrien , w. j. wadsworth , and j. g. rarity , `` nonclassical interference and entanglement generation using a photonic crystal fiber pair photon source , '' phys .
rev .
lett . * 99 * , 120501 ( 2007 ) . http://link.aps.org/abstract/prl/v99/e120501 .
j. fan and a. migdall , `` generation of cross - polarized photon pairs in a microstructure fiber with frequency - conjugate laser pump pulses , '' opt .
express * 13 * , 57775782 ( 2005 ) .
http://www.opticsexpress.org/abstract.cfm?uri=oe-13-15-5777 .
c. xiong and w. j. wadsworth , `` polarized supercontinuum in birefringent photonic crystal fibre pumped at 1064 nm and application to tuneable visible / uv generation , '' opt .
express * 16 * , 24382445 ( 2008 ) .
http://www.opticsexpress.org/abstract.cfm?uri=oe-16-4-2438 .
o. alibart , j. fulconis , g. k. l. wong , s. g. murdoch , w. j. wadsworth , and j. g. rarity , `` photon pair generation using four - wave mixing in a microstructured fibre : theory versus experiment , '' new journal of physics * 8 * , 67 ( 2006 ) .
h. f. hofmann and s. takeuchi , `` quantum phase gate for photonic qubits using only beam splitters and postselection , '' phys .
rev .
a * 66 * , 024308 ( 2002 ) .
http://link.aps.org/abstract/pra/v66/e024308 .
o. cohen , j. lundeen , b. smith , g. puentes , p. mosley , and i. a. walmsley , `` tailored photon - pair generation in optical fibers , '' arxiv * quant - phys * , 0809.0071v1 ( 2008 ) . c. k. hong , z. y. ou , and l. mandel , `` measurement of subpicosecond time intervals between two photons by interference , '' phys .
rev .
lett .
* 59 * , 20442046 ( 1987 ) .
p. j. mosley , j. s. lundeen , b. j. smith , and i. a. walmsley , `` conditional preparation of single photons using parametric downconversion : a recipe for purity , '' new j. phys .
* 10 * , 093011 ( 2008 ) .
b. cemlyn , `` four - wave mixing in microstructured fibres , '' master s thesis , university of bristol ( 2008 ) . |
the equation of state ( eos ) in the vicinity of saturation density is pretty well understood .
but beyond saturation , the theories of dense matter present uncertainties .
neutron stars provide a fantastic astrophysical environment for testing theories of cold and dense matter . in the core of neutron stars
, densities could reach values of several times @xmath2 g @xmath3 . at such high densities
, the fermi energies of the constituent particles could exceed the rest masses of heavier particles , and hence favour the appearance of such particles in the core .
further , as the timescales associated with neutron stars are much greater than those associated with weak interactions , violation of strangeness conservation due to weak reactions in the core would result in the appearance of strangeness - containing particles such as hyperons . by producing new degrees of freedom , the appearance of strange particles
is expected to result in a softer eos of dense matter in the neutron star interior .
the highest neutron star mass that can be supported depends crucially on the eos .
observations of pulsars in neutron star binaries provide a precise measurement of neutron star masses through general relativistic effects .
the best determined pulsar mass ( @xmath4 ) is the hulse - taylor pulsar , and masses of most other pulsars are found to be clustered around this canonical value .
recently , several neutron stars with larger masses have been discovered .
radio timing observations of three post keplerian parameters led to the most precise measurement of the mass of a millisecond pulsar of @xmath5 @xcite .
shapiro delay measurements from radio timing observations of the binary millisecond pulsar psr j1614 - 2230 indicated a mass of [email protected]@xmath7 of the neutron star @xcite .
this is the largest rather precisely observed pulsar mass so far and thus poses the tightest reliable lower bound on the maximum mass of neutron stars .
any theory of ultradense matter requires that the eos produce a maximum mass at least as high as this measured value , i.e. , models with @xmath8 @xmath9 would be ruled out .
+ according to existing models of dense matter , the presence of hyperons leads to a considerable softening of the eos , resulting in a reduction of the maximum mass of the neutron star . with hyperons , including only the hyperon - nucleon interaction , brueckner - hartree - fock ( bhf ) calculations obtain maximum masses of the order 1.47@xmath7 @xcite .
the inclusion of the hyperon - hyperon interaction leads to a further softening of the eos and reduces the obtained masses to 1.34@xmath7 @xcite . on inclusion of three - body interactions ,
the maximum mass achieved in the bhf framework is 1.26@xmath7 @xcite . by employing a recently constructed hyperon - nucleon potential , the maximum masses of neutron stars with hyperons is computed to be well below 1.4@xmath7 @xcite .
another approach is to adopt a relativistic mean field ( rmf ) model @xcite , which we employ for the following investigation .
several attempts have been made using the rmf model to explain neutron star masses higher than 2@xmath7 , by artificially increasing the hyperon vector coupling away from their su(6 ) values @xcite .
dexheimer et al .
@xcite obtained large neutron star masses using a chirally motivated model including fourth - order self - interaction terms of the vector mesons @xmath10 and @xmath11 .
recently , bednarek et al .
@xcite achieved a stiffening of the eos by invoking quartic vector - meson terms proportional to @xmath12 and cross terms like @xmath13 and @xmath14 . when the stiffening on inclusion of hyperons is not sufficient , a transition to the quark phase can be considered to produce large maximum masses @xcite .
bonanno and sedrakian @xcite also succeeded in obtaining a large neutron star mass with a hyperon and quark core using a stiff eos and vector repulsion among quarks .
+ as the parameters in the rmf model are fitted to the saturation properties of infinite nuclear matter , extrapolation to higher densities and asymmetry involve uncertainties .
three of these properties - the saturation density , the binding energy and the asymmetry energy are more precisely known than the remaining ones - the effective nucleon mass and the compression modulus of nuclear matter . the uncertainty in the dense matter eos is basically related to the uncertainty in these two saturation properties .
in this paper , we determine which of the two influences most the high density part of the eos that determines the highest attained neutron star mass .
the parameters associated with attractive interaction among hyperons and nucleons are fitted to the potential depths of hyperons in nuclear matter , known from hypernuclear experiments .
we investigate in this paper how the uncertainty in hyperon potential depths influences the stiffness of the eos and hence the maximum mass .
we want to address the question : under which conditions can hyperons be present in massive neutron stars ?
we achieve this through a controlled parameter study within the rmf model . in our model
, we succeed in generating stiff eos by taking small values of effective nucleon mass @xmath15 , in the range of those obtained from fits to nuclei rather than bulk nuclear matter , and by including the strange vector meson @xmath11 . in a subsequent paper @xcite
, we will focus on the assumption of the underlying symmetries that govern the repulsive interactions among hyperons and nucleons .
+ this paper is organized in the following way : in sec .
2 , we describe the model to calculate the eos
. the parameters of the model are listed in sec .
the results of our calculations are discussed in sec . 4 , and the summary and conclusions are given in sec .
one of the possible approaches to describe neutron star matter is to adopt a rmf model subject to chemical equilibrium and charge neutrality . for our investigation of nucleons and hyperons in neutron star matter
we will choose the full standard ( @xmath16 ) baryon octet as well as electrons and muons . in this model ,
baryon - baryon interaction is mediated by the exchange of scalar ( @xmath17 ) , vector ( @xmath18 ) and isovector ( @xmath19 ) mesons .
the lagrangian density is given by @xcite @xmath20 the isospin multiplets for baryons b are represented by the dirac spinor @xmath21 with vacuum baryon mass @xmath22 , isospin operator @xmath23 , and @xmath24 and @xmath25 are field strength tensors . to reproduce the saturation properties of nuclear matter , the scalar self - interaction term @xmath26
is introduced @xcite .
we also included an additional self - interaction term @xmath27 for the vector field as proposed by bodmer @xcite . due to inclusion of this term ,
the vector field increases proportional to @xmath28 for high densities , where @xmath19 is the baryon density , instead of the linear dependence in absence of this term .
this results in a good agreement with brueckner - hartree - fock calculations .
we will denote the above model with the three exchange mesons as `` model @xmath29 '' .
the hyperon - hyperon interaction is usually incorporated through the exchange of additional strange scalar ( @xmath30 ) and strange vector ( @xmath11 ) mesons , again the scalar meson being responsible for attractive and the vector meson for repulsive interactions respectively : @xmath31 since it is our goal to obtain the stiffest possible eos within the model , we will only make use of the @xmath11 but omit the @xmath30 .
this choice is also in accordance with @xmath32-hypernuclear data , where now it is clear that the @xmath32 interaction is only weakly attractive ( see ref @xcite for a discussion ) .
the model obtained with this additional lagrangian we will call `` model @xmath33 '' .
the calculation is performed using the mean field approximation @xcite .
the effective baryon mass is given by @xmath34 while the chemical potential of baryon @xmath35 is @xmath36 , with @xmath37 denoting the corresponding fermi momentum .
charge neutrality is described by the condition @xmath38 where @xmath39 is the number density of baryon b , @xmath40 is the electric charge and @xmath41 and @xmath42 are charge densities of electrons and muons respectively .
we consider the hadronic matter to behave like an ideal fluid .
the energy - momentum tensor for such a fluid yields the eos , defined by the relationship between the total energy density @xmath43 and the pressure @xmath44
there are five input parameters for the rmf models : the nuclear saturation density @xmath45 ( or equivalently , the number density @xmath46 ) as well as the binding energy per baryon number b / a , the effective mass of the nucleon @xmath47 , the nuclear compression modulus k and the asymmetry coefficient @xmath48 , all taken at saturation density .
the parameters @xmath49 , @xmath50 , @xmath51 , @xmath52 and @xmath53 of `` model @xmath54 '' are determined from the saturation properties of nuclear matter in an analytic way ( see @xcite for explicit formulae ) .
the binding energy , the asymmetry coefficient and the saturation density are well determined .
we set them to @xmath55 , @xmath56 mev , @xmath57 mev @xcite .
the effective nucleon mass and the compression modulus are less well known . in principle
, the parameters of the model can be fitted to properties of nuclei , which result in low effective masses @xcite . however , here we want to explore the full parameter range of the model , as model parameters fitted to bulk nuclear matter and usually adopted in the community have large effective mass parameters . hence for this study , we use the parameter sets such as gm1 ( @xmath58=300 mev , @xmath59 ) and gm3 ( @xmath58=240 mev , @xmath60 ) which are fitted to the properties of bulk nuclear matter and have rather large effective nucleon masses in the range @xmath61 @xcite , and also the parameter sets fitted to the properties of nuclei , e.g. the whole nl parameter family and tm1 which yield very low values of @xmath62 @xcite . the value of compression modulus for symmetric nuclear matter can be extracted from the energies of the isoscalar giant monopole resonances ( isgmr ) in nuclei ( see e.g. @xcite and references therein ) .
systematic studies of isgmr energies in various nuclei suggest a value of @xmath63 mev for the incompressibility of symmetric nuclear matter @xcite . however , there are strong surface effects , as well as large uncertainties in the value of the compression modulus for higher densities and asymmetric nuclear matter .
another semiempirical parameter of nuclear matter crucial for determining the stiffness of the eos is the density dependence of symmetry energy , denoted by @xmath64 .
we computed the values of the density dependence of symmetry energy @xmath64 for the sets discussed above , and they were found to be fairly comparable with the values existing in the literature @xcite .
these however differ from the values of @xmath64 extracted from isospin diffusion data ( @xmath65 mev ) @xcite and isoscaling data ( @xmath66 mev ) @xcite .
we note here that when the effective mass @xmath67 is varied between 0.55 and 0.8 , @xmath64 changes from 108 to 94 mev ( @xmath6 1 mev depending on the compressibility ) .
the hyperon coupling constants @xmath68 , @xmath69 and @xmath70 are determined by using su(6 ) symmetry ( the quark model ) @xcite : @xmath71 the scalar meson - hyperon coupling constants @xmath72 are adjusted to the potential depths @xmath73 felt by a hyperon y in a bath of nucleons at saturation @xcite following the relation @xmath74 where the hypernuclear potential depths in nuclear matter @xmath75 are fixed in accordance with the available hypernuclear data .
the best known hyperon potential is that of the @xmath76 , having a value of about @xmath77 mev @xcite . in case of the @xmath78s and the @xmath79s
the potential depths are not as firmly known as for the @xmath76 .
the following values @xmath80 mev , @xmath81 mev are generally adopted from the hypernuclear experimental data ( see @xcite for a discussion ) .
we begin our investigations with `` model @xmath54 '' , namely with a @xmath82 model including non - linear scalar and vector self - interactions as well as @xmath19 mesons and the full baryon octet . in the previous section
, we already mentioned that there are uncertainties in the values for the potential depths of hyperons in nuclear matter , obtained from hypernuclear experiments .
we vary the potentials @xmath83 and @xmath84 systematically to study how they influence the stiffness of the hadronic eos and the limiting neutron star mass .
we start with a fixed @xmath79 potential depth of @xmath85 mev @xcite .
the potential depth of @xmath78 is varied in the range @xmath86 mev @xmath87 mev .
we plot in fig .
[ figuresigmafreeeos ] the eos for different values of @xmath83 ( lower branch ) for the gm1 parameter set and also for the same model with additional @xmath11 meson ( upper branch ) . for the sake of clarity
only the cases @xmath88 mev are displayed . in mev , indicated by the corresponding numbers in the key .
the lower branch is obtained for `` model @xmath54 '' for the gm1 parameter set while the upper branch is obtained including additionally the @xmath11 meson.,width=12 ] it can be inferred from the figure that for less deep potentials @xmath83 the eos becomes stiffer for each model .
however , this is only true for negative potentials , since the values @xmath89 mev and @xmath90 mev basically result in the same eos . for negative values of the potential @xmath83 , the @xmath78s
are bound in nuclear matter and the deeper the potential is , the more attractive must be the effective mesonic interaction and thus the softer the eos . for @xmath91 mev
, the @xmath78s are no longer bound to nuclear matter and the effective mesonic interaction becomes more and more repulsive for increasing @xmath83 .
this would in principle stiffen the eos .
however , up to neutron star densities , i.e. about @xmath92 , the @xmath78s are not present in hadronic matter if the potential is repulsive and thus the eos up to these densities becomes insensitive to the actual value of @xmath83 .
the composition of hadronic matter for positive and negative values of @xmath78 potential depth has already been investigated in @xcite . + .
the variation of @xmath93 in
model @xmath54 can not account for the observed neutron star mass limit ( lower branch ) , unless the @xmath11 meson is included in the model ( upper branch).,width=453 ] the repulsive potentials @xmath94 give the highest maximum star masses for each parameter set ( also for others which we did not plot , as e.g. tm1 and nl3 ) . in fig .
[ figuresigmafreemr ] we plot the parts of the mass - radius relations around the maximum mass star configurations for the eos in fig .
[ figuresigmafreeeos ] . the maximum mass for the lower branch of curves in fig .
[ figuresigmafreemr ] corresponding to `` model @xmath54 '' is @xmath95 .
we note , that the variation of @xmath83 in `` model @xmath54 '' for gm1 does not help to increase the maximum neutron star mass above the limiting value of [email protected]@xmath96 .
investigating other parameter sets like tm1 would not solve the problem since the maximum masses are comparable to those of gm1 . the various nl - models would be good candidates , since they are amongst the stiffest available eos @xcite .
however , the nl - models have ( like all models which are fitted to finite nuclei ) very small effective nucleon masses already at saturation density ( @xmath97 ) . at around five times
nuclear saturation density the @xmath98fields are so large that the effective nucleon mass becomes zero or in principle even negative .
this causes a well - known instability which makes it impossible to find a physical solution to the rmf equations @xcite .
however , the critical density for the onset of the instability is higher than the maximum central density in the neutron star interior , and hence this is not a relevant problem for the present investigation . + the `` model @xmath99 '' gives us the upper bundle of eos in fig .
[ figuresigmafreeeos ] . in fig .
[ figuresigmafreemr ] the upper branch of the mass - radius relations corresponds to these eos from `` model @xmath99 '' .
while the variation of @xmath78 potential only varies the maximum mass in the range @xmath100 0.03 @xmath96 , the impact of @xmath11 on the maximum mass is an order of magnitude larger : the maximum mass is increased by nearly 0.21 @xmath101 to @xmath102=1.96 @xmath101 . we should also mention that for purely nucleonic matter eos the obtained maximum masses are even higher ( about 2.4@xmath96 for gm1 and nearly 2.8@xmath101 for nl3 ) .
we have seen that the potential depth @xmath83 has but a little influence on the maximum mass of a hadronic neutron star . as it does not play an important role in our investigations ,
we shall fix it to the value @xmath83=+30 mev @xcite which means that the @xmath78s are not present and the eos is as stiff as possible .
we then vary the @xmath79 potential @xmath84 in the range @xmath103 + 40mev .
as before , we compute the eos for the gm1 parameter set for `` model @xmath54 '' and `` model @xmath99 '' for different @xmath79 potentials . we plot in fig .
[ figurexifreeeos ] the results for the values of @xmath84 from -40 mev to + 40 mev in steps of 20 mev .
. the lower bundle of curves corresponds to `` model @xmath54 '' and the upper to `` model @xmath99 '' .
all eos are obtained for gm1 parameter set.,width=453 ] '' for potential depths @xmath84 = + 28 mev ( a ) and -28 mev ( b ) are displayed . similarly , particle fractions for `` model @xmath99 '' are shown for @xmath79 potentials + 28 mev ( c ) and -28 mev ( d ) . the threshold for appearance of the @xmath79 hyperons is pushed to higher densities with increasing @xmath79 potential.,width=453 ] we see that for increasing @xmath79 potential , the eos stiffens . a difference to the influence of the @xmath78 potential depth
is found for positive values of @xmath84 : here the eos remains sensitive to the value of the potential due to the contributions from @xmath104 at neutron star densities while for even higher densities both @xmath79s contribute . in fig .
[ figureuxifrac ] , we display how the composition of the neutron star core varies for various values of @xmath79 potential in models `` @xmath54 '' and `` @xmath99 '' .
it is clear from the figure that for positive @xmath79 potential , @xmath76 hyperons dominate particle fractions at high densities , while for negative @xmath105 the particle fractions from all hyperons contribute . as @xmath79 potential increases from negative to positive values , the threshold density for appearance of @xmath104 and @xmath106 hyperons
is pushed to higher densities , and this effect further increases on inclusion of repulsion from the @xmath11 mesons . + . as before ,
the variation of @xmath107 in model @xmath54 can not explain the observed neutron star mass limit ( lower branch ) , unless the @xmath11 meson is included ( upper branch).,width=453 ] in fig .
[ figurexifreemr ] we plot the mass - radius relations of the neutron stars obtained from the various eos in fig . [ figurexifreeeos ] . within `` model @xmath54 ''
the maximum masses range from 1.69 @xmath96 for @xmath108 mev to 1.92 @xmath96 for @xmath84=+40 mev , i.e. over an interval of @xmath1090.22 @xmath96 . the maximum mass for the case @xmath84=+40 mev including @xmath11 is 2.04@xmath96 .
although the value of the @xmath79 potential depth helps to stiffen the eos far better than the @xmath83 it is still not enough to be consistent with the mass of psr j1614 - 2230 for `` model @xmath54 '' .
however , the mass constraint can be fulfilled within `` model @xmath110 '' for the gm1 parameter set .
we combine the variations of @xmath83 and @xmath84 and plot in fig .
[ figurepotentialcontour ] lines of constant maximum mass in the @xmath111 plane within the range @xmath86 to @xmath112 mev for both potentials . as before ,
we compare the results from `` model @xmath54 '' with those of `` model @xmath33 '' . ) in the @xmath111 plane .
upper plot for `` model @xmath54 '' in gm1 parameterization , lower one for the same model including also @xmath11 mesons.,title="fig:",width=453,height=453 ] - 3 cm ) in the @xmath111 plane .
upper plot for `` model @xmath54 '' in gm1 parameterization , lower one for the same model including also @xmath11 mesons.,title="fig:",width=453,height=453 ] - 1 cm in the light of our previous discussions we can quickly explain the shape of the lines in the plot : on each side of the @xmath113 line the maximum masses are dominated by that potential which is smaller and are rather insensitive to the value of the larger potential .
this means , that in the region where @xmath114 the lines of constant maximum mass are parallel to the @xmath115axis . in the other part of the plot , where @xmath116 , the behaviour is mirrored .
since the influence on the maximum masses is larger for smaller potentials , the mass lines ( which are plotted for equidistant mass steps of 0.01@xmath96 ) are densest towards the lower and the left parts of the plots but thin out with increasing potentials .
we conclude this section by noting that for the whole range @xmath86 mev @xmath117,@xmath118 mev the maximum masses of `` model @xmath54 '' for the gm1 parameter set are not compatible with the observed pulsar mass @xmath1190.04@xmath96 .
inclusion of the @xmath11 meson helps to sufficiently stiffen the eos and the whole range of @xmath120,@xmath121 + 40mev is now allowed .
however , in this case the new mass limit does not help to restrict the parameter range of the hyperon potentials . .
the upper curves refer to purely nucleonic stars while for the lower curves hyperons are included.,width=491 ] in the previous section , we found that the variation of hyperon potentials were not sufficient to explain the observed large neutron star mass .
we then proceed to find out how the maximum mass depends on the nucleon coupling constants .
we vary the least well known of the saturation parameters , namely the effective nucleon mass @xmath47 and the compression modulus k. we first consider purely nucleonic stars , and then repeat the exercise including hyperons in the `` model @xmath29 '' . in fig .
[ figuremefffree ] we plot the maximum masses as a function of the effective nucleon mass at saturation @xmath122 for several values of k. we see that the maximum masses for the nucleonic case ( upper bunch of curves ) show a strong dependence on the effective nucleon mass : they drop from 2.88@xmath96 at @xmath123 to 1.87 - 2.01@xmath96 at @xmath124 . from fig .
[ figuremefffree ] it is clear that for moderate effective masses , @xmath125 , the compression modulus gains a little influence on the maximum mass for pure nucleonic stars , and causes a mass difference of less than 0.05@xmath96 . above this value of the effective mass , the compression modulus causes a larger splitting of the maximum masses until at @xmath124 the mass difference is about 0.1@xmath96 .
for the hyperonic case ( lower bunch of curves ) , the maximum masses range from 2.0 - 2.1@xmath96 at @xmath123 to about 1.39 - 1.63@xmath96 at @xmath124 .
the mass splitting at @xmath126 for different k is about 0.2@xmath96 , increases to nearly 0.35@xmath96 at @xmath127 , and then again decreases to about 0.25@xmath96 at @xmath128 . for comparison we also mark in fig .
[ figuremefffree ] the maximum masses and the effective masses of several other rmf sets fitted to properties of nuclei , like tm1 , nl3 or nl - sh @xcite . for the `` model @xmath33 '' , the maximum masses range from 2.4 @xmath7 at @xmath129 to 1.6 @xmath7 at @xmath128 .
we also note that the results do not change substantially from the results obtained using `` model @xmath29 '' when we include the @xmath30 meson ( in `` model @xmath130 '' ) .
contrary to the statement made in @xcite , that k as well as @xmath47 strongly influence the high - density behaviour of the equation of state of symmetric nuclear matter this being an effect [ which ] shows up in neutron rich and neutron star matter , we find that the compression modulus has nearly no influence on the high - density behaviour of the nucleonic eos whereas the effective mass has a stark impact .
this finding is well known for the rmf model @xcite .
the effect of the compressibility on hyperonic stars is however remarkable , although this dependence also varies considerably with the effective nucleon mass . in @xcite ,
the investigated range of the effective nucleon mass was 0.7 - 0.8 for the hyperonic case , in which k has a greater influence on the maximum mass than @xmath122 , which we also confirm in our study .
however , we also extend our investigation for a broader range of @xmath122 , and find that this conclusion no longer holds for low effective nucleon masses .
also , at large @xmath122 where k is influential the maximum masses of hyperonic stars do not comply with the limiting pulsar mass of [email protected]@xmath96 . in `` model
@xmath29 '' with hyperons this mass is only reached for low values of the effective nucleon mass @xmath131 . within the `` model @xmath132 '' the mass constraint
can only be achieved for @xmath133 . in the case of nucleonic stars , for @xmath134
the pulsar mass constraint can not be reached if the compression modulus is below 225 mev .
+ plane.,width=453,height=453 ] to explore the whole range of k continuously , we plot lines of constant maximum mass of nucleonic stars in the @xmath135 plane in fig . [ figuremeffkcontour ] .
the results support our previous observation : the lines of constant maximum mass are nearly parallel to the x - axis , which means that the compression modulus does not influence the stiffness of the eos at neutron star densities . only for effective masses above @xmath136 do the lines become slightly slanted and show that a higher compression modulus then slightly stiffens the eos . we note , that in the investigated case of purely nucleonic neutron stars the observed mass of psr j1614 - 2230 requires an effective nucleon mass at saturation density of maximally @xmath137 at @xmath138 mev down to @xmath139 at @xmath140 mev .
the possibility of existence of hyperons in massive neutron stars has been investigated in this paper . within hadronic models
, we varied the coupling strengths of the scalar @xmath17 meson to the @xmath78 and @xmath79 hyperons by changing the potential depths of these hyperons in nuclear matter .
meanwhile , we assumed su(6 ) relations for the vector couplings and kept the @xmath76 potential fixed .
we found that for deeper potential depths the eos becomes softer and the maximum masses correspondingly smaller .
which of the two potentials has a larger impact on the stiffness of the eos depends mainly on their relative values : if @xmath141 the @xmath79s will appear earlier ( i.e. at lower densities ) than the @xmath78s and therefore have the greater influence on the eos . however , this can not exactly be mirrored , since for @xmath142 the @xmath143 might appear earliest of all @xmath78s and @xmath79s , but as soon as the @xmath104 appears , the number density of the @xmath104 overtakes that of the @xmath143 by far .
the overall effect of varying the two potential depths is rather small ( @xmath1090.2@xmath96 over the probed range ) .
+ since the impact of the potential depths on the eos is not enough to raise the maximum masses of neutron stars for `` model @xmath54 '' for gm1 above the limit of @xmath1190.04@xmath96 , we decided to increase the repulsive interaction amongst hyperons by including the vector @xmath11 meson in the rmf model .
the additional repulsion between hyperons results in a sufficiently stiff eos accompanied by an increase in the maximum masses of about @xmath1090.2@xmath96 . for the gm1 parameter set , the investigated area of the @xmath144 plane complies now to almost full extent with the mass of psr j1614 - 2230 .
+ in the case of purely nucleonic stars we varied the effective mass of the nucleon as well as the compression modulus at nuclear saturation density which are input parameters for the rmf model .
we found that the compression modulus has very little influence on the maximum mass of pure nucleonic neutron stars : in the range @xmath145 mev its influence is negligible for low effective masses @xmath146 , while its most prominent effect is for large effective masses @xmath147 where it causes a mass difference of just @xmath148 0.1@xmath96 .
the effective mass of the nucleon at saturation proved to be a very good parameter for obtaining large maximum masses . over the investigated range @xmath149 the maximum masses
change about @xmath100 1.1@xmath96 .
the most massive stars ( up to @xmath1502.9@xmath96 ) are obtained for low effective nucleon masses independently of the compression modulus , while for @xmath151 we can not reach the mass limit of [email protected]@xmath96 if the compression modulus is too low .
for the hyperonic case , maximum masses varied from 2.0 - 2.1@xmath96 at @xmath123 to about 1.39 - 1.63@xmath96 at @xmath124 in the `` model @xmath29 '' , whereas in the `` model @xmath33 '' the variation is from 2.4 to 1.6 @xmath7 respectively .
hyperonic neutron stars in the `` model @xmath29 '' are only compatible with the new pulsar mass constraint for low values of the effective nucleon mass @xmath131 mev , while within the `` model @xmath132 '' the mass constraint can only be achieved for @xmath133 mev .
we conclude that the nuclear compression modulus at saturation might be a good indicator for the stiffness of the nuclear matter eos at low densities ( around saturation ) , but the maximum mass is much more sensitive to the effective nucleon mass at saturation .
this result is in accordance with the fact that the high density behaviour of the eos in the rmf model is solely controlled by the effective nucleon mass at saturation density by virtue of the hugenholtz - van hove theorem @xcite . + as we have seen ,
a 2@xmath96 star like psr j1614 - 2230 helps to constrain the nuclear matter eos . then a possible 2.4@xmath96 star - as has been recently suggested to exist @xcite - would pose even tighter constraints on the neutron star composition .
should this measurement be confirmed , we could rule out many hadronic models ( e.g. tm1 or gm3 ) even in the limit of purely nucleonic stars .
furthermore , in most such models the appearance of hyperons together with the use of su(6 ) relations for the vector coupling constants would then become impossible . in an associated paper @xcite , we further investigate the possibility of existence of hyperons in massive neutron stars , by questioning the assumption of su(6 ) symmetry for the determination of vector meson - hyperon couplings .
+ @xmath152 + _ acknowledgements : _
b . is supported by the dfg through the heidelberg graduate school of fundamental physics . d.c .
acknowledges the support from the alexander von humboldt foundation .
is supported by the state of baden - wrttemberg through a lgfg stipend .
this work is supported by bmbf under grant fkz 06hd9127 , by the helmholtz alliance ha216/emmi and by compstar , a research networking program of the european science foundation .
[ 2]#2 p. c. c. freire , c. g. bassa , n. wex , i. h. stairs , d. j. champion , s. m. ransom , p. lazarus , v. m. kaspi , j. w. t. hessels , m. kramer , j. m. cordes , j. p. w. verbiest , p. podsiadlowski , d. j. nice , j. s. deneva , d. r. lorimer , b. w. stappers , m. a. mclaughlin , and f. camilo , http://dx.doi.org/10.1111/j.1365-2966.2010.18109.x [ _ mon . not .
r. astron .
* 412 * ( 2011 ) 2763 ] . | the constituents of cold dense matter are still far from being understood .
however , neutron star observations such as the recently observed pulsar psr j1614 - 2230 with a mass of @xmath0 m@xmath1 help to considerably constrain the hadronic equation of state ( eos ) .
we systematically investigate the influence of the hyperon potentials on the stiffness of the eos .
we find that they have but little influence on the maximum mass compared to the inclusion of an additional vector - meson mediating repulsive interaction amongst hyperons .
the new mass limit can only be reached with this additional meson regardless of the hyperon potentials .
further , we investigate the impact of the nuclear compression modulus and the effective mass of the nucleon at saturation density on the high density regime of the eos .
we show that the maximum mass of purely nucleonic stars is very sensitive to the effective nucleon mass but only very little to the compression modulus .
neutron stars , equation of state , hypernuclei , hadronic matter |
reed - solomon ( rs ) codes are among the most widely used codes .
their applications range from consumer electronics such as compact disc ( cd ) and digital versatile disc ( dvd ) players to optical communication and data storage systems .
most high - throughput rs codec architectures are based on reformulated versions @xcite of either the berlekamp - massey algorithm @xcite or the extended euclidean algorithm ( eea ) @xcite .
a variable number of iterations at most @xmath0 for correcting up to @xmath1 errors are used in most eea - based decoders .
one exception is the errors - only hypersystolic reed - solomon decoder proposed by berlekamp _
_ @xcite that uses exactly @xmath0 iterations regardless of the number of errors . of course , decoders based on the berlekamp - massey algorithm also use exactly @xmath0 iterations . in this paper ,
new modifications of the eea are proposed for both errors - only ( eo ) decoding and errors - and - erasures ( ee ) decoding .
first , a new modification of the eea for errors - only decoding is proposed based on the ideas in @xcite and @xcite
. the new algorithm also uses _ exactly _ @xmath0 iterations , thus eliminating the degree computation and comparison in most variants of the eea ( see , for example , @xcite ) .
one particular hardware implementation ( not described in this paper ) of this modified algorithm turns out to be the _ same _ circuit as that obtained by implementing the ribm algorithm of @xcite , with the difference that in one implementation the polynomials enter and leave the circuit in ascending order of coefficients while in the other implementation the polynomials enter and leave in descending order of coefficients !
the ribm algorithm is based on the berlekamp - massey algorithm , and this result gives yet _ another _ equivalence between the berlekamp - massey algorithm and the eea , different from those that have been described previously in @xcite and @xcite .
the modified eea proposed in this paper can be extended to errors - and - erasures decoding , and used to derive an errors - and - erasures decoding algorithm that also iterates _ exactly _ @xmath0 times .
this modified algorithm also combines the erasure - locator polynomial computation and the solution of the modified key equation in a seamless way .
hardware implementation of this algorithm eliminates not only the degree computation and comparison ( see , e.g. , @xcite ) but also the separate block used for computing the erasure - locator polynomial ( see , for example , @xcite ) , thus leading to considerable savings .
the codewords in a @xmath1-error - correcting cyclic reed - solomon code @xcite of block length @xmath2 over @xmath3 are the polynomials @xmath4 , with the property that the @xmath0 successive powers @xmath5 of @xmath6 , a primitive @xmath2-th root of unity in @xmath3 , are roots of @xmath7 . here
, @xmath8 can be any integer , but is often chosen to be @xmath9 or @xmath10 for ease of implementation .
the code has @xmath11 information symbols .
suppose the codeword polynomial @xmath7 is transmitted and the received word , corrupted by errors , is @xmath12 where @xmath13 .
the decoder computes the _ syndromes _ of the error polynomial @xmath14 : @xmath15 , @xmath16 the _ syndrome polynomial _ is defined as @xmath17 if @xmath18 errors have occurred , the error polynomial @xmath14 can be written as @xmath19 where @xmath20 , called the _ error values _ , are said to have occurred at the _ error locations _
@xmath21 respectively .
the _ error - locator polynomial _ @xmath22 of degree @xmath18 is defined to be @xmath23 while the _ error - evaluator polynomial _ @xmath24 of degree less than @xmath18 is defined as @xmath25 the error - locator and error - evaluator polynomials defined above are related to the syndrome polynomial by the _ key equation : _
@xmath26 note that @xmath27 is known to the decoder , while @xmath22 and @xmath24 are not .
as the name suggests , solving the key equation for both @xmath22 and @xmath24 is the most difficult part of the decoding process . in this paper , we focus on the eea algorithm @xcite for solving the key equation . after the key equation is solved , the errors can be corrected by finding the error locations and computing the error values .
the error locations can be found via the _ chien search _ : for each @xmath28 , @xmath29 , the decoder tests whether or not @xmath30 .
if @xmath30 , @xmath31 , that is , @xmath32 .
the value of the @xmath28-th transmitted symbol is computed via _
forney s formula _ : @xmath33 where @xmath34 is the formal derivative of @xmath22 .
it is worth noting that most implementations of rs decoders compute and use @xmath35 and @xmath36 where @xmath37 is a nonzero scalar whose value is immaterial : @xmath35 has the same roots as @xmath22 and so the decoder finds the same error locations , and @xmath37 cancels out in ( [ eq : forney ] ) and so the decoder finds the same error values .
henceforth , we ignore such scalar factors in @xmath22 and @xmath24 . in some cases , the received words enter the decoder with some symbols specially marked as being highly unreliable and hence more likely to be in error than other symbols .
these marked symbols are called erasures . for a code with minimum hamming distance @xmath38 , any pattern of @xmath39 erasures and @xmath18 errors
can be corrected @xcite as long as @xmath40 let @xmath41 denote the ( unknown ) error locations and @xmath42 denote the known _
erasure locations_. as before , the error - locator polynomial is defined as @xmath43 while the _ erasure - locator polynomial _ is defined to be @xmath44 note that @xmath45 can be computed from the known erasure locations whereas @xmath46 is unknown .
similarly , the error - evaluator polynomial is defined as @xmath47 and the _ erasure - evaluator polynomial _ is defined to be @xmath48 where @xmath49 and @xmath50 denote respectively the @xmath51-th error and _ erasure values_. also note that some of the erasure values might be zero . if we define the _ errata - locator polynomial _ @xmath52 of degree @xmath53 as @xmath54 and the _ errata - evaluator polynomial _ @xmath55 as @xmath56 , then the key equation ( [ eq : keyeq ] ) still holds for the errata - locator and errata - evaluator polynomials .
furthermore , the @xmath57 errata locations can be obtained from the errata - locator polynomial by the chien search and the correct values of the codeword symbols can be computed using forney s formula .
note that the errors - only decoding is simply the special case of the errors - and - erasures decoding where @xmath58 and @xmath59 .
as in errors - only decoders , typical implementations compute the same scalar multiple of all these polynomials , and the value of this scalar does not affect the results of any subsequent computations . as described above ,
the decoding of rs codes involve three successive stages syndrome computation ( sc ) , key equation solving ( kes ) , and errata correction ( ec ) .
the implementation of syndrome computation and errata correction , which is described in , for example , @xcite-@xcite is generally straightforward and will not be discussed further in this paper . here , we will focus on the implementation of key equation solving , which is the most difficult part of the decoding process .
sugiyama _ et al . _
@xcite first pointed out that the extended euclidean algorithm for computing the polynomial greatest common divisor ( gcd ) can be used to solve the key equation ( [ eq : keyeq ] ) .
the eea , tailored to solving the key equation , can be stated as follows : * eea for errors - only decoding : the eo algorithm * 1 .
* initialization : set * @xmath60 , @xmath61 , + @xmath62 , @xmath63 , and @xmath64 .
* iteration : while * @xmath65 \geq t$ ] , + @xmath66 by @xmath67 to obtain both + @xmath68 and the + @xmath69 .
. + @xmath71 .
* output : * @xmath72 , @xmath73 .
let @xmath74 denote the value of @xmath28 when the eo algorithm stops .
then , the outputs @xmath75 and @xmath76 are scalar multiples of @xmath52 and @xmath55 as defined in ( [ eq : errorlocator ] ) and ( [ eq : errorevaluator ] ) since @xmath77 is not necessarily @xmath10 . also , it can be shown that the polynomials @xmath78 computed by the eo algorithm have degrees @xmath79 that form a strictly decreasing sequence with @xmath80 , @xmath81 and @xmath82 .
the drawbacks to efficient implementation of the above algorithm are as follows .
* the degree @xmath83 of the quotient polynomial @xmath84 can vary with @xmath28 , and thus step 2 of the eo algorithm requires a variable number of computations .
this complicates the control mechanism .
furthermore , it is necessary to divide the coefficients of @xmath85 by the leading coefficient of @xmath67 in order to obtain the quotient polynomial @xmath84 . *
determining the stopping condition @xmath86 < t$ ] is difficult since data needs to be gathered from many different cells in the circuit .
these two drawbacks have motivated many improvements .
brent and kung @xcite proposed a systolic array implementation of the polynomial gcd algorithm in which each of the polynomial division operations involved is broken into a sequence of partial divisions , as humans often do in the `` long division '' method .
in fact , this idea had been pointed out even earlier ( see , for example , @xcite ) .
brent and kung also proposed using cross multiplications to avoid dividing one polynomial coefficient by another .
these notions can be explained as follows .
let @xmath87 and @xmath88 denote polynomials of degrees @xmath89 and @xmath90 respectively where @xmath91 .
then , in the long division " of @xmath87 by @xmath88 , the first step consists of subtracting @xmath92 from @xmath87 to cancel out the highest degree term in @xmath87 . if the remainder has degree at least @xmath90 , a different multiple of @xmath88 is subtracted to cancel out the highest degree term in the remainder , and so on .
but , @xmath93 where ( [ eq : crossmult ] ) follows from ( [ eq : longdiv ] ) because @xmath94 for any nonzero scalar @xmath37 .
thus , changing @xmath87 to @xmath95 instead of @xmath96 avoids a division while still zeroing out the highest degree term in @xmath87 and still having the same gcd .
since the computation of @xmath97 in the eo algorithm is of exactly the same form as the computation of @xmath98 , a similar calculation can be used to update these polynomials as well .
these two basic ideas have been used in different ways by many researchers to design different algorithms for gcd computation and rs decoding ( see , for example , @xcite ) .
all these algorithms actually compute scalar multiples @xmath99 and @xmath100 of the error - locator and error - evaluator polynomials defined in ( [ eq : errorlocator ] ) and ( [ eq : errorevaluator ] ) respectively .
our architectures also use the ideas of brent and kung , but compute @xmath101 and @xmath102 where @xmath103 .
since the nonzero roots of @xmath104 are the same as those of @xmath52 and the factors @xmath105 cancel out in forney s formula ( [ eq : forney ] ) , such factors are inconsequential and can be ignored .
as noted before @xcite , a polynomial division can be broken up into a sequence of partial divisions for ease of implementation , and the cross - multiplication technique can be used to avoid divisions of field elements @xcite .
the same ideas can be adapted to eliminate the comparison of @xmath106 $ ] with @xmath1 as well .
our modification of the eea solves the key equation in _ exactly _ @xmath0 steps ; rather than in _ at most _
@xmath0 steps as in previous work by others .
when @xmath107 errors have occurred , our algorithm computes @xmath108 and @xmath109 in @xmath110 steps instead of @xmath22 and @xmath24 .
our algorithm is also set up so that each of the additional @xmath111 steps simply multiplies the results by @xmath112 so that after a total of @xmath0 steps , our modification of the eea has computed @xmath113 and @xmath114
. these give the same error locations and error values as do @xmath22 and @xmath55 .
the advantages to our approach are that the degree checking is avoided completely , and the key equation solution is produced with a fixed latency , both of which properties simplify the control mechanism in an implementation .
we claim that the following modified version of the eea solves the key equation for rs decoding , producing polynomials @xmath115 and @xmath116 .
* algorithm i ( modified euclidean algorithm ) * 1 . * initialization : * @xmath117 , @xmath118 , @xmath119 , @xmath120 , and @xmath121 .
* iteration : repeat * @xmath0 times : 1 . *
set * @xmath122 , @xmath123 , @xmath124 .
if * @xmath125 and @xmath126 , + @xmath127 3 .
* set * @xmath128 3 .
* output : * @xmath129 , @xmath130 , and @xmath131 . if @xmath107 errors have occurred , then after @xmath132 iterations of step i.2 in algorithm i , @xmath133 and @xmath134 where scalar factors are ignored .
when step i.2 is iterated @xmath111 more times , @xmath135 and @xmath88 are multiplied by @xmath112 ( step 1.2a ) and the ignorable scalar factor @xmath136 ( step i.2c ) each time .
hence , when algorithm i ends , @xmath137 , @xmath138 , and @xmath139 . if @xmath140 , then algorithm i terminates with @xmath141 . in practice ,
steps i.2a - i.2c are not executed in succession but combined into a single calculation that computes a boolean control variable @xmath142 and then _ simultaneously _ sets @xmath143 note also that @xmath131 must be initialized to @xmath144 for this modified computation to work properly .
we refer to this variation of algorithm i as algorithm i*. the following theorem summarizes the results of algorithms i and i*. + _ theorem _ 1 : if @xmath107 errors have occurred , then when algorithm i or i * terminates , @xmath145 and @xmath146 where @xmath37 is nonzero . if algorithm i or i * terminates with @xmath147 , then more than @xmath1 errors have occurred and the error pattern @xmath14 is not correctable .
in errors - and - erasures decoding ( see , for example , @xcite ) , the key equation ( [ eq : keyeq ] ) relating the errata - locator polynomial @xmath148 and the errata - evaluator polynomial @xmath55 is usually solved via the following three steps executed in succession : 1 . using the known erasure locations @xmath149 , @xmath150 to compute the erasure - locator polynomial @xmath151 and the modified syndrome polynomial @xmath152 2 . solving the modified key equation @xmath153 for the _ error_-locator polynomial @xmath154 and the _ errata_-evaluator polynomial @xmath55 3 .
multiplying @xmath154 by @xmath151 to obtain the errata - locator polynomial @xmath52 .
the computations of @xmath151 and @xmath155 can be implemented as @xmath39-iteration procedures in which initial values @xmath156 and @xmath157 are multiplied successively by @xmath158 , @xmath159 , , @xmath160 .
alternatively , @xmath151 can be computed as described above and then the polynomial product @xmath161 computed in @xmath162 further iterations ( cf .
@xcite ) . of course
, if there are no erasures , then these calculations do not need to be carried out .
next , the ( modified ) key equation is solved in _ at most _
@xmath163 iterations via a slightly modified version of the extended euclidean algorithm for errors - only decoding .
a slightly modified version of the berlekamp - massey errors - only decoding algorithm also can be used for this purpose . finally , the last of the three steps above is not strictly necessary , but is usually implemented ( in fact , embedded into the second step ) because it is more convenient to use @xmath52 in computing errata values via forney s formula .
it was pointed out by blahut @xcite that if the registers used to compute @xmath46 are initialized to @xmath151 instead of @xmath10 , then the iterations during the solution of the modified key equation produce @xmath148 directly and thus the third step above is in effect embedded into the key equation solution .
blahut @xcite also noted that for the berlekamp - massey algorithm , it is unnecessary to compute the modified syndrome polynomial : if the registers used to compute @xmath46 are initialized to @xmath151 instead of @xmath10 , then the `` discrepancies '' calculated in the berlekamp - massey algorithm are exactly those needed for solving the modified key equation , and the algorithm produces @xmath22 directly instead of @xmath45 .
unfortunately , reformulated berlekamp - massey algorithms such as the ribm and ribm algorithms of @xcite as well as _ all _ key equation solvers that are based on the extended euclidean algorithm _ do _ need @xmath155 .
however , these algorithms are able to embed the third step above into the key equation solution .
finally , it has been noted by several researchers that the operations used for the solution of the key equation can be adapted to the computation of @xmath151 or @xmath155 or both .
thus , the same hardware can be used in these calculations , which reduces the number of finite - field multipliers required .
as pointed out in @xcite , the modified key equation @xmath164 can be solved by using the extended euclidean algorithm shown below : * eea for errors - and - erasures decoding * 1 . * initialization : set * @xmath60 , @xmath165 , @xmath62 , @xmath63 , and @xmath166 .
* iteration : while * @xmath86 \geq t+\mu /2 $ ] , + @xmath66 by @xmath67 to obtain both + @xmath68 and the + @xmath69 . + @xmath70
. + @xmath71 .
* output : * @xmath167 , @xmath73 .
this algorithm is clearly similar to the eo algorithm for errors - only decoding .
in fact , the only differences between the two algorithms are the initial values of @xmath168 and the stopping condition .
hence , direct implementation based on the above algorithm suffers the same problems we described in section iii.a . using the same reformulation steps as in section iii
, the above algorithm can be modified to a @xmath169-iteration algorithm that eliminates the degree checking and produces @xmath170 and @xmath171 instead of @xmath46 and @xmath24 respectively . as mentioned above ,
the operations used to compute @xmath172 in the third step can be embedded in the modified euclidean algorithm by initializing @xmath173 and @xmath174 to scaled values @xmath175 and @xmath176 respectively .
note that the updates of @xmath177 depend on @xmath84 , which are not at all affected by the change in the initial values .
thus , each @xmath177 is scaled by @xmath151 , leading to a final output @xmath178 . in order to combine the computation of @xmath151 and @xmath155 with the modified euclidean algorithm ( with the computation of @xmath52 embedded ) into a single algorithm with @xmath0 iterations
, we use a polynomial @xmath179 that can be formed easily during the syndrome computation stage by saving the marked erasure locations .
we allow for the possibility that more than @xmath0 erasures have occurred , even though such an errata pattern is not decodable .
our reformulated eea for errors - and - erasures decoding is as follows : * algorithm ii * 1 .
* initialization : * @xmath180 , @xmath118 , @xmath119 , @xmath121 , @xmath120 , and @xmath181 .
* iteration : repeat * @xmath0 times : + @xmath182 .
+ @xmath183 .
+ @xmath184 + @xmath185 @xmath186 3 .
* output : * @xmath187 , @xmath188 , @xmath131 , and @xmath189 .
this algorithm uses a boolean control variable first that has value 1 only when the erasure locations are being processed to compute @xmath151 and @xmath155 . during this time , swap is always @xmath9 and @xmath190 is set to the erasure location being processed currently . for each erasure location @xmath190 ,
algorithm ii sets @xmath88 to @xmath191 $ ] and @xmath135 to @xmath192 $ ] , thus obtaining @xmath155 and @xmath151 after all the @xmath39 erasure locations have been processed one by one .
note that the update @xmath193 discards the erasure location that was just processed and replaces it by the next erasure location to be processed , and thus first becomes zero after @xmath39 iterations . from this point
onwards , the updates of all the polynomials and of @xmath131 are the same as in algorithm i * , and thus solve the modified key equation in @xmath163 iterations .
we remark that the received words with no erasures can be decoded correctly by algorithm ii as well .
in fact , algorithm i * corresponds to the special case of algorithm ii where @xmath194 and first is always @xmath9 .
similar to theorem 1 , we have + _ theorem _ 2 : suppose that @xmath18 errors and @xmath39 erasures have occurred where @xmath195 .
let @xmath53 denote the total number of errata .
then when algorithm ii terminates , @xmath126 and + @xmath196 where @xmath37 is nonzero .
if algorithm ii terminates with @xmath147 or with @xmath197 , then the errata pattern is not correctable .
in this paper , modified euclidean algorithms that use _ fixed _ numbers of iterations are proposed for both errors - only and errors - and - erasures decoding of rs codes .
the salient feature of fixed numbers of iterations leads to simpler control mechanisms and hence hardware savings .
the new algorithm for errors - and - erasures decoding seamlessly combines the three steps typically used in previously proposed architectures into one procedure , leading to hardware savings .
j. h. baek , j. y. kang and m. h. sunwoo , `` design of a high - speed reed - solomon decoder , '' _ proceedings of the 2002 ieee international symposium on circuits and systems , iscas 2002 , _ may 2002 , vol . 5 , pp . 793796 .
e. berlekamp , g. seroussi , and p. tong , `` a hypersystolic reed - solomon decoder , '' in _ reed - solomon codes and their applications _ ,
s. b. wicker and v. k. bhargava , eds . , chapter 10 .
ieee press , piscataway nj , 1994 .
a. e. heydtmann and j. m. jensen , `` on the equivalence of the berlekamp - massey and the euclidean algorithms for decoding , '' _ ieee transactions on information theory _
46 , pp . 26142624 ,
november 2000 .
huang and c.w .
wu , `` vlsi design of a high speed pipelined reed - solomon codec , '' _ proceedings of the international symposium on multi - technology information processing ( ismip ) , _ hsinchu , december 1996 , pp .
517522 .
s. kwon and h. shin , `` an area - efficient vlsi architecture of a reed - solomon decoder / encoder for digital vcrs , '' _ ieee transactions on consumer electronics _ , vol .
43 , no . 4 pp .
10191027 , november 1997 .
t. k. truong , w. l. eastman , i. s. reed , and i. s. hsu , `` simplified procedure for correcting both errors and erasures of reed - solomon code using euclidean algorithm , '' in _ proc .
iee , part e _ ,
november 1988 , vol . | the extended euclidean algorithm ( eea ) for polynomial greatest common divisors is commonly used in solving the key equation in the decoding of reed - solomon ( rs ) codes , and more generally in bch decoding . for this particular application ,
the iterations in the eea are stopped when the degree of the remainder polynomial falls below a threshold .
while determining the degree of a polynomial is a simple task for human beings , hardware implementation of this stopping rule is more complicated .
this paper describes a modified version of the eea that is specifically adapted to the rs decoding problem .
this modified algorithm requires no degree computation or comparison to a threshold , and it uses a fixed number of iterations .
another advantage of this modified version is in its application to the errors - and - erasures decoding problem for rs codes where significant hardware savings can be achieved via seamless computation . |
circular and linear polarization are important diagnostics of nonthermal radio emission and its environment in extragalactic radio sources and galactic micro - quasars . in active galactic nuclei ( agn ) , linear polarization ( lp ) has been used to demonstrate that the emission mechanism is synchrotron radiation , that magnetic fields are present and that shocks propagate in jets , compressing the magnetic fields ( e.g. , * ? ? ?
lp is typically greater than 1% and is present in a very high fraction of powerful extragalactic radio sources @xcite .
rotation measures are used to study the plasma density and magnetic field in accretion and outflow regions of agn @xcite .
the significance of circular polarization ( cp ) is less clear .
a variety of mechanisms can produce the small levels of cp seen in these sources , including intrinsic and extrinsic origins .
synchrotron and cyclotron emission produce cp in varying degrees depending on the the electron population , magnetic field strength and orientation .
coherent emission processes may also lead to cp .
in addition , the presence of electrons with lorentz factors @xmath2 in or near a synchrotron source can significantly alter both cp and lp through faraday effects .
the combination of these effects can lead to complex polarized spectra .
finally , a magnetized scattering region can produce scintillating cp .
recently , there has been a sharp increase in interest in cp observations .
cp has been detected in sagittarius a * @xcite , a large number of powerful extragalactic radio sources @xcite including some idv sources @xcite , and galactic microquasars @xcite .
vlbi imaging of powerful extragalactic radio sources indicate that the cp is confined to the cores .
the spectral and variability properties in these sources vary widely indicating potentially diverse origins . in most sources
, lp dominates cp , as expected for simple synchrotron sources . a notable exception to this is sgr a * , which exhibits no lp .
the polarization properties of low luminosity agn ( llagn ) have not been previously investigated .
@xcite observed the lp of a sample of `` weak '' radio sources ( @xmath3 ) with flat spectra .
many of these sources were unpolarized at a level of 1% at 15 ghz , suggesting that they are not simply scaled down versions of powerful agn .
however , there exists a significant gap in luminosity between these sources and sgr a * , which has a radio luminosity @xmath4 .
the intrinsic nature of the llagn is still in doubt ( for recent reviews see * ? ? ?
* ; * ? ? ?
these sources display emission line luminosities that may be produced through a variety of mechanisms , including stellar photoionization , collisional ionization through shocks or in a starburst .
a variety of methods have demonstrated that some of the sources display agn - like characteristics .
however , as in the case of sgr a * , the luminosities are substantially sub - eddington and there is dispute over the roles of advection , convection and outflow in these sources .
the detection of high brightness temperature radio components has made a strong case for the role of jets in these sources .
lp and cp can act as probes of the radio - emitting region as well as of the accretion and outflow regions through which the emission propagates .
we present here a radio polarization survey of 11 nearby llagn with luminosities in the range @xmath5 to @xmath6 @xmath7 . in 2 , we describe the observations , our error analysis and the results .
cp is only detected in one source , ngc 3031 ( m81 * ) .
we have discussed the significance of this detection in another paper @xcite .
lp is detected in 4 sources .
we discuss our results in 3 and summarize in 4 .
a sample of 11 llagn was selected from the samples of @xcite and @xcite .
the selected sources ( table [ tab : results ] ) are the brightest radio sources in the catalogs for which we were able to obtain data .
we observed the sample with the very large array in the c configuration on 4 april 2000 for 12 hours at 8.4 ghz .
the correlator was configured for continuum observations with 50 mhz bandwidth each in two intermediate frequency bands in right ( rcp ) and left circular polarization ( lcp ) .
the angular resolution was about @xmath8 .
observations were carried out with the same techniques described in @xcite and @xcite .
absolute flux calibration was set with a short observation of 3c 48 .
scans of the target sources were bracketed with scans for pointing and calibration on a nearby background compact radio source .
a second nearby background compact radio source was observed in the same way as the target sources as a check for cp calibration errors .
we cycled through calibration and target sources in about 10 minutes .
calibrator , target and check sources were placed in five groups , labeled a through e , based on their proximity on the sky .
integration time for each target source was set to achieve rms noise errors of less than 0.1% except for the weak sources i 1024 and ngc 3516 .
observations were made over 12 hours , leading to significant changes in parallactic angle , azimuth angle and elevation angle for each source .
at least 4 observations were made of each source .
data were reduced in aips .
the calibration sources were self - calibrated assuming @xmath9 .
linear polarization calibration was also performed using the calibration sources .
amplitude gains , phase gains and polarization leakage terms were transferred to the target and check sources .
all sources were imaged in stokes i , q , u and v. flux densities were estimated by fitting gaussians at the source s location .
results are summarized in table [ tab : results ] .
some sources were resolved in total intensity . however , polarization images were unresolved for all sources except ngc 4374 and ngc 4486 .
we plot the linear and circular polarizations together in figure [ fig : lpcp ] .
systematic errors from beam squint , gain and leakage terms are estimated according to the model for the vla in @xcite to be between 0.03 and 0.05% in total , depending on the number of scans per source .
we see that that measured cp for the check sources range from 0.03 to 0.45% with a median of 0.14% .
the likely source of additional error is polarization of the calibrator and/or check source .
ngc 3031 is the only source where the measured cp exceeds that of the check source and is greater than 3 times the thermal noise .
the cp for ngc 3031 is , in fact , 10 times the thermal noise and repeatedly detectable @xcite .
we do not count ngc 6500 as a cp detection because of the large measured cp for the check source , j1745 + 173 .
the two large cp values with the same sign suggest that the calibrator , j1751 + 096 , may have cp at a level of @xmath10 .
@xcite report a tentative detection of @xmath11 for j1751 + 096 in their 15 ghz vlba observations , which is consistent in sign and in magnitude for a steep spectral index .
for all sources , the measured polarization fractions are lower limits to the polarization of the nuclear source . for a typical distance of 10 mpc , the linear resolution of these observations
is @xmath12 pc .
diffuse thermal radiation can contribute significantly to the total flux .
however , for the sources for which we know of vlbi observations , the compact radio flux is a substantial component of the total flux .
ngc 3031 and ngc 4486 ( m87 ) both have well - studied vlbi properties , which show that a compact core dominates the total flux @xcite . @xcite and @xcite
also show that the ratio of 5 ghz vlba to vla flux densities is on the order of unity for five of our sources : ngc 4278 , ngc 4374 , ngc 4552 , ngc 4579 and ngc 6500 .
for all 13 sources in the @xcite sample , the ratio is greater than 0.48 .
nevertheless , a factor of a few in the ratio of total to compact flux density would raise the polarization limits to the point where sgr a * and ngc 3031 would not be detected .
we detect lp in 3 of 11 sources : ngc 4278 , ngc 4486 , and ngc 4579 .
these results show that some llagn do exhibit lp , contrary to the conclusions of @xcite , which were based on less sensitive data .
this is strong evidence for a nonthermal contribution to the flux density .
however , the rate of detection of lp is lower than that of more powerful agn .
for example , @xcite detect lp in all but 3 of their 40 sources with a mean value of a few percent .
in addition , the fraction of lp detected in our sources is lower than in more powerful agn . excluding m87 , the mean fractional lp is @xmath13 .
of the 62 sources in the pearson - readhead flux - limited sample , 39 have polarization greater than 1% at 4.8 ghz and 57 have polarization greater than 1% at 15 ghz @xcite .
mean polarizations are on the order of a few percent .
these trends are true for quasars , bl lac objects and radio galaxies in the sample .
this difference can not be explained by confusion .
all three sources have been previously detected with vlbi and been shown to be dominated by their compact flux @xcite .
are the llagn intrinsically weakly polarized or are they depolarized by their surrounding medium ?
a number of scenarios may lead to intrinsic depolarization .
extreme magnetic field disorder in the source will leave the source weakly polarized .
this may be due to the absence of shocks in jets that order the field and produce regions of high polarization fraction . a high degree of symmetry in the magnetic field , i.e. , a face on azimuthal field or a radial field will also lead to a weakly polarized source .
such field configurations might be typical of disk accretion or quasi - spherical accretion .
finally , internal faraday depolarization may occur if the density of nonrelativistic particles in the emitting region is large enough .
on the other hand , depolarization in the accretion , outflow or surrounding regions is more than adequate to account for this result .
beam and bandwidth depolarization at these frequencies and in these bandwidths will occur for rotation measures of @xmath14 and @xmath15 , respectively .
@xcite has measured rms as large as @xmath16 towards the cores of powerful agn . @xcite
have measured a rm@xmath17 towards sagittarius a * , which may arise from the accretion and outflow regions or from an interstellar halo . even for the very low accretion rates of adaf and cdaf models of sgr a * , rms can easily reach @xmath18 and higher @xcite . the radio emission from our sources
almost certainly propagates through similar media . in the case of ngc 4258 ,
for example , the radio emission originates from within 0.02 pc of the black hole @xcite . on the other hand ,
the radio emission from much more powerful agn may begin parsecs away from the black hole .
the radio intermediate quasar iii zw 2 provides an interesting case for comparison .
this compact radio source ( @xmath19 pc ) shows superluminal expansion and spectral evolution consistent with shock excitation in a relativistic outflow @xcite .
however , the source exhibits no linear polarization in 20 years of monitoring ( m. aller , private communication ) . in order to look for the effects of bandwidth depolarization and large rms , we compared the difference between lower and upper sideband position angles for all sources .
these sidebands are each 50 mhz wide and separated from each other by 50 mhz .
we find a significant change only for ngc 4579 of @xmath20 degrees
. systematic error due to leakage term miscalibration is less than 10 degrees . by comparison ,
ngc 4278 , which is polarized at a level similar to ngc 4579 , shows a difference of 6 degrees in the position angle .
changes in the position angle for calibrator sources were also consistent with thermal noise .
the implied rm for ngc 4579 is @xmath21 .
this would lead to bandwidth depolarization @xmath22 , which does not significantly increase the intrinsic lp .
however , one can not rule out a much higher rm inducing more substantial depolarization as well as the angular rotation
. we can compute an upper limit for the rm @xmath23 with the requirement of a maximum of 70% intrinsic polarization . as discussed above
, this rm could arise from a variety of sources . without more detailed knowledge of the source geometry
, it is difficult to place limits on specific source models such as jet , adaf and cdaf models . unlike the case of millimeter wavelength polarization in sgr a * , the centimeter wavelength emission from ngc 4579 may originate far outside of the accretion region .
however , if the emission does pass through the accretion region , then the accretion rate must be very low @xcite .
spectropolarimetry , higher frequency polarimetry and vlbi polarimetry are potentially useful tests of these scenarios .
they could provide quantitative information about field strengths and particle densities in the inner few parsecs of these sources .
we compare the rates of detection of cp here with that in powerful extragalactic radio sources . at a level of @xmath0
, we detect 1 out of 9 sources .
@xcite detect cp greater than 0.1% in 12 out of 30 sources .
@xcite detect cp greater than 0.1% in 11 out of 40 sources .
due to the small number of sources in our sample , we can not conclude that the rate of cp in llagn is less than that in powerful agn .
however , the trend appears to be that way .
what physical condition leads to the presence of cp in some sources and not in others ?
ngc 3031 and sgr a * are both unique in that they have inverted spectra at the frequencies where cp is detected .
we summarize the spectral index information for sources where simultaneous spectra are available in figure [ fig : lpcp ] .
all spectral indices come from the measurements of @xcite . while the number of sources is small , the trend is apparent
. this effect may be the result of an inhomogeneous synchrotron source in the optically thick regime .
orientation effects may also be of importance since the magnitude of cp from a synchrotron source depends on the line of sight magnetic field strength .
finally , ngc 3031 is also the closest of our survey sources , suggesting that resolution of thermal sources in the nucleus and of extended optically , thin regions in the jet may be important .
we have observed linear and circular polarization in a sample of low luminosity agn .
the polarization properties of these sources differ from those of more powerful agn .
linear polarization is less frequently detected and less strong when detected .
circular polarization is less frequently detected .
the conclusion is not as simple as claiming that these sources differ intrinsically from the more powerful agn , however . while increased field disorder in the sources could explain the results , environmental depolarization effects
could also play a role .
these polarization measurements probe regions on a scale of parsecs and smaller in these sources .
detection of rotation measures in these sources is a direct measure of the environment of a supermassive black hole , which may be used to constrain models for accretion and outflow .
the national radio astronomy observatory is a facility of the national science foundation operated under cooperative agreement by associated universities , inc .
this research has made use of data from the university of michigan radio astronomy observatory which is supported by funds from the university of michigan . a & 1048 + 717 & calibrator & @xmath24 & @xmath25 & @xmath26 & @xmath27 & n + a & 1056 + 701 & check source & @xmath28 & @xmath29 & @xmath30 & @xmath31 & n + a & ngc 3031 & srbi - ii & @xmath32 & @xmath33 & @xmath34 & @xmath35 & n + a & ngc 3516 & rsbo@xmath36 & @xmath37 & @xmath38 & @xmath39 & @xmath40 & y + a & ngc 4589 & e2 & @xmath41 & @xmath42 & @xmath43 & @xmath44 & n + + b & 1400 + 621 & calibrator & @xmath45 & @xmath25 & @xmath46 & @xmath47 & n + b & 1335 + 587 & check source & @xmath48 & @xmath49 & @xmath50 & @xmath51 & n + b & ngc 5216 & e / s0 & @xmath52 & @xmath53 & @xmath54 & @xmath55 & n + + c & 1158 + 248 & calibrator & @xmath56 & @xmath57 & @xmath58 & @xmath59 & n + c & 1221 + 282 & check source & @xmath60 & @xmath61 & @xmath62 & @xmath63 & n + c & ngc 4278 & e1 & @xmath64 & @xmath65 & @xmath66 & @xmath67 & n + c & 1239 + 075 & check source & @xmath68 & @xmath69 & @xmath70 & @xmath71 & n + c & ngc 4374 & e1 & @xmath72 & @xmath73 & @xmath74 & @xmath75 & y + c & ngc 4486 & e0 & @xmath76 & @xmath77 & @xmath78 & @xmath79 & y + c & ngc 4552 & so@xmath80 & @xmath81 & @xmath82 & @xmath83 & @xmath84 & n + c & ngc 4579 & sabii & @xmath85 & @xmath86 & @xmath87 & @xmath88 & n + + d & 1430 + 107 & calibrator & @xmath89 & @xmath90 & @xmath91 & @xmath92 & n + d & 1445 + 099 & check source & @xmath93 & @xmath94 & @xmath95 & @xmath96 & n + d & i 1024 & s0 & @xmath97 & @xmath98 & @xmath99 & @xmath100 & y + + e & 1751 + 096 & calibrator & @xmath101 & @xmath25 & @xmath102 & @xmath103 & n + e & 1745 + 173 & check source & @xmath104 & @xmath105 & @xmath106 & @xmath107 & n + e & ngc 6500 & sbc & @xmath108 & @xmath109 & @xmath110 & @xmath111 & n + | we conducted a very large array survey of eleven low luminosity active galactic nuclei for linear and circular polarization at 8.4 ghz .
we detected circular polarization in one source ( m81 * ) and linear polarization in 3 sources .
sensitivity limits were @xmath0 for both modes of polarization in 9 of 11 sources .
the detections confirm the importance of nonthermal emission in llagn . however , detection rates for circular and linear polarization are lower for these sources than for more powerful agn .
fractional linear polarization in detected sources is also lower than in more powerful agn .
the weak linear polarization in the survey sources indicates their overall similarity to sgr a*. confusion with thermal sources , depolarization and weaker , less extended jets may contribute to these differences .
we detect a rotation measure @xmath1 for ngc 4579 .
this may arise from magnetized plasma in the accretion , outflow or interstellar regions .
inverted spectra are present in both m81 * and sagittarius a * and absent from all sources in which circular polarization is not detected .
this suggests that optical depth effects are important in the creation of circular polarization . |
the theoretical difficulty in understanding the smallness of the fermi scale in the standard model ( sm ) with respect to any new physics at very high energies is the main motivation for expecting new phenomena to show up already in the energy range that is probed by the cern lhc .
one of the most appealing solutions to this ` hierarchy problem ' is supersymmetry ( susy ) , to the point that it is sometimes called ` the standard way beyond the sm ' .
since superpartners have not been found at colliders , supersymmetry is apparently pushed more and more to higher energies and this in principle weakens its power in solving the hierarchy problem , or equivalently some amount of finetuning is apparently reintroduced in the determination of the electroweak scale . while this is more and more unavoidably true at least for its minimal implementation , the minimal supersimmetric standard model ( mssm ) , it is also clear that non - minimal implementations or non - standard configurations can still have the chance to be natural , i.e. not finely - tuned .
it is then crucial , from a conceptual point of view , to keep an eye on these natural configurations until they are excluded since , if it were not for the hierarchy problem , susy could manifest itself at much higher energies . ] .
the main recent experimental inputs that are relevent for our considerations are the following : * a sm - like higgs boson is now excluded from about 127 gev up to 600 gev @xcite@xcite . at the same time , there is a @xmath1 hint for a higgs - like scalar close to 125 gev . * direct searches of s - particles
@xcite-@xcite have already set very strong lower bounds on their masses .
these bounds are quite model dependent , however it can be said that typically the squarks of the first two generations have to be heavier than about 1 tev .
the case in which only the third generation is light , which is very motivated by naturalness as we shall see , is instead much less constrained , see @xcite@xcite for theoretical estimates . in such a situation
the gluino could still be as light as @xmath2 gev , with the third generation even down to @xmath3 gev , and the lhc phenomenology depends crucially on the mass difference between the gluino and the charginos / neutralinos ( see figure [ fig : gluinodecays ] ) . *
lhcb @xcite , and later cdf @xcite , have measured with a high precision the direct cp violation in the decays of a @xmath4 meson into @xmath5 and @xmath6 .
the world average now yields : @xmath7 which deviates by approximately 3.8@xmath8 from the no - cp violation point , and is quite larger than the value @xmath9 that one would expect in the sm .
we focus on naturalness considerations in section [ sect : nat ] , we consider other possible issues and points of view in section [ sect : other ] , and then conclude in section [ sect : concl ] .
the main purpose of this discussion is to conceptually clarify the various possibilities in a compact way , while providing a relatively rich list of references with particular attention to the most recent studies .
we hope that this summary may be useful at least as an orientation in this subject that is now a very hot topic @xcite@xcite@xcite .
as already said naturalness , or the problem of finetuning , is the main theoretical motivation for physics beyond the sm at the lhc energies .
it makes thus sense to discuss this point in some detail .
it is usually said that the problem stems from the ` quardatic divergences ' that are present in the radiative corrections to the mass of an elementary scalar , or from the fact that these corrections are quadratically sensitive to the ultraviolet cutoff of the theory .
this statement may sound strange , because it seems to depend on the way the theory is regularized : for example , where is the problem if one uses dimensional regularization instead of a sharp momentum cutoff to make the loop integrals finite ? given the importance of the issue ,
let us briefly see how it can be expressed in terms of renormalized quantities : this will give us more insight in the meaning of finetuning in the supersymmetric case .
consider a heavy fermion field @xmath10 coupled to the higgs boson @xmath11 through a yukawa interaction : @xmath12 and let us regularize the higgs self energy @xmath13 with a cutoff @xmath14 .
the relevant diagram is shown in figure [ figura : loops ] left , and we have : @xmath15\right|_{p^2=0}\ ] ] where @xmath16 , @xmath17 is the momentum of the @xmath11 line and ` @xmath18 ' are finite terms .
let us now consider a process involving energy scales @xmath19 much higher than the value @xmath20 of the mass of the higgs particle when it is produced on shell ( the ` pole mass ' , that corresponds to the pole of the propagator ) .
if we do renormalized perturbation theory using @xmath20 , then quantum effects will give in general significant corrections to the physical observables , and it will be necessary to compute the relevant amplitudes at many orders in perturbation theory .
it is instead convenient to treat @xmath21 as a parameter of perturbation theory , defined by a renormalization prescription at the scale @xmath19 , for example : @xmath22 where @xmath23 is the renormalized propagator of the scalar @xmath11 .
as a result @xmath21 becomes an ` effective mass ' and starts running with the energy scale @xmath19 according to the renormalization group ( rg ) flow , like the other parameters of perturbation theory , with initial condition : @xmath24 in this way the most important quantum corrections are resummed provided that one uses the value of this ` running mass ' @xmath25 to compute the mass effects in a process involving the typical momentum scale @xmath19 . in words , @xmath25 is ` the mass that minimizes the quantum corrections if we probe the theory at the energy scale @xmath19 ' . [
cols="^,^,^ " , ] to summarize , natural supersymmetry with stops and left - handed sbottom well below the tev and heavier first two generations stands now as the only ` conventional ' left out possibility for supersymmetry to solve the hierarchy problem with @xmath26 , or equivalently 10% finetuning ( other ` less conventional ' possibilities are discussed in section [ sect : escaping ] ) .
our considerations are briefly summarized in table [ table : naturalsusy ] .
the main point , as already said , is that the spectrum shown in figure [ fig : natspectrum ] has now the right to be called ` natural supersymmetry ' .
moreover , if the hint for a higgs boson at 125 gev is confirmed , @xmath27susy with singlet - doublet mixing stands as the most natural possibility .
notice also that , due to the large increase in the higgs quartic , in @xmath27susy all the squarks are naturally allowed to be heavier by a factor of 2 or 3 ( see also figure [ naturalness12gen ] ) .
thus , for example , if in the future no stops are found below 1 tev and/or the squarks of the first two generations are excluded up to about 5 tev , then only @xmath27susy - like configurations will have the right to be called ` natural supersymmetry ' ... in this last case , the easiest way to exclude natural supersymmetry might be to exclude the extended higgs sector
. we will come back on these issues in the conclusions . as a final remark we quote that , maybe also motivated by similar considerations , there has been recent interesting model building activity focussed on susy breaking with split families , or natural supersymmetry as we now say : see e.g. the recent idea of ` flavor mediation ' @xcite @xcite or other possibilities @xcite @xcite-@xcite , including even natural supersymmetry from string theory @xcite .
moreover , since a low @xmath28 can help in minimizing the amount of finetuning in ( [ eq : natbounds ] ) given the sparticle masses , we can also say that low - scale susy breaking is somehow favoured by data(+naturalness ) ; see e.g. @xcite@xcite for very recent studies .
besides taking heavier superpartners , another way to keep supersymmetry alive is escaping the experimental bounds through less - standard or peculiar configurations .
a first well - known possibility is r - parity violation .
recall that , from the superfield point of view , the down - type higgs doublet @xmath29 has exactly the same quantum numbers of the left - handed lepton doublet .
it is thus clear that the baryon number and the lepton number are not accidental symmetries of the theory at the renormalizable level , as instead is the case for the sm .
introducing all the couplings allowed by the gauge symmetry without any suppression mechanism , one ends up with large contributions to excluded processes like proton decay .
the standard way to avoid these terms is to impose by hand an additional discrete symmetry , namely the r - parity , whose quantum number is @xmath30 where @xmath31 is the baryon number , @xmath32 is the lepton number , and @xmath33 is the spin of the field @xmath34 and @xmath32 are separately violated by nonperturbative effects at high energies , while @xmath35 is conserved . ] .
notice that all the sm particles have r - parity @xmath36 , while all the superpartners have r - parity @xmath37 .
the most striking consequence of this fact is that the lightest supersymmetric particle ( lsp ) must be stable , and can thus be a very good dark matter candidate . from the point of view of collider signatures
, r - parity leads to the crucial prediction of abundance of events with missing energy , since the lsp necessarily escapes the detector .
most of the experimental susy - search analyses published so far assume in fact that the production of any supersymmetric particle is accompained by the missing energy carried away by the lsp at the end , eventually , of a decay chain .
for the case of r - parity violating ( rpv ) supersymmetry , on the contrary , very few experimental analyses have been published so far , see e.g. @xcite @xcite , and it is clear that the present stringent bounds on the ` conventional ' r - parity conserving case can be escaped , at least in part . from the conceptual point of view , for rpv to be ` natural ' one needs a rationale to make the effect small enough in order not to be in trouble with proton decay and other very constrained processes . for example a mechanism that one could use , and that is well known to be very efficient in suppressing the flavor violation that additional flavor structures can introduce in beyond - the - sm contexts , is minimal flavor violation ( mfv ) .
the possibility of rpv with mfv has been studied in @xcite@xcite and also recently in @xcite@xcite .
see also @xcite@xcite for recent discussions .
even with conserved r - parity , the signal of jets plus missing energy can be significantly reduced if the mass splittings between some superpartners are small .
for example , a small gluino - lsp splitting reduces the phase space for the gluino dacay : as a consequence the typical missing transverse energy and momentum coming from processes involving gluinos are reduced , leading to a reduced acceptance for given selection cuts . for recent studies of this configuration , called ` compressed spectrum ' , see e.g. @xcite@xcite . a similar way to reduce the missing energy , that in some cases is even more efficient , is ` stealth susy ' @xcite@xcite ( see also @xcite ) . in this case ,
r - parity is conserved but the lsp of the ordinary sector decays into particles belonging to a ` stealth sector ' in which the relative mass splittings within a supermultiplet are small . moreover , r - parity - even stealth particles can decay back into sm particles ; the phase space is thus again reduced for the decay chain that terminates with the true ` stealth lsp ' .
another example of possible non - standard signatures is given by dirac gauginos , which may distort the phenomenology in a significant way @xcite@xcite@xcite , see also @xcite@xcite@xcite for recent studies
. in summary , given the negative results of the susy searches so far , trying to see whether the detection may have been escaped for a reason that is not simply large mass is probably being to become a very hot topic in the near future , as an alternative way in which a natural supersymmetry could be realized .
known possibilities are suppressing the missing energy ( rpv ) , suppressing the ` visible energy ' that goes into jets ( compressed spectrum ) , suppressing both ( stealth susy ) , or suppressing the production cross section by giving the gauginos a large dirac mass ( dirac gauginos ) .
new phenomenological studies and ideas about the above issues would probably be very timely .
recall that the main motivations for supersymmetry are , broadly speaking : ( i ) it can solve the hierarchy problem , in a perturbative way ; ( ii ) it gives a good dark matter ( dm ) candidate ; ( iii ) gauge couplings unify at high energy ; ( iv ) it is very elegant and a necessary ingredient of string theory .
moreover all this is done in a way that is almost automatically compatible with the ewpt . retaining all the features ( i)-(iv )
provides a very attractive picture .
however , that nature must be without finetuning is not a necessity ( we will come back on this in the conclusions ) , and it is meaningful to study what happens retaining ( ii)-(iv ) while abandoning the requirement ( i ) .
it is clear that for dm one needs at least some neutralino at low energy , although its mass may be relatively heavier than the usual susy dm .
it can be seen that the only choice that is radiatively stable is to keep all the gauginos and higgsinos around the tev scale , while the squarks and one higgs scalar doublet can be at a very high mass scale @xmath38 .
remarkably , it turns out that this is also what is required for obtaining a precise gauge coupling unification .
this scanario is called ` split supersymmetry ' @xcite-@xcite , and the only finetuning that is needed is the ( large ) one that makes one of the two scalar doublets much lighter than its natural mass scale , which would be @xmath38 .
split supersymmetry has many interesting peculiar features , whose discussion go beyond our present scopes ; we refer for example to @xcite for a review .
to cite one of them , the gluino can be very non - standard and its dominant dacay width can be into a gluon and a gravitino , giving a single - jet signal with a distinctive energy distribution . alternatively , although it is not the lsp , the gluino can be very long lived if its dominant decay mode is mediated by the heavy squarks , with a variety of possible unusual signatures not only at the lhc @xcite-@xcite .
the experimental discovery of a slowly decaying gluino would be a strong indication for split susy , and would also provide a way to measure the high scale @xmath38 .
a more radical point of view is to retain only the motivation ( iv ) and to accept again a large finetuning in the value of the fermi scale .
actually , having accepted the finetuning , string theory is a very plausible framework since it provides a huge number of vacua ( the ` landscape ' ) , and one can speculate about environmental / anthropic selection among them @xcite . in any case , what one ends up with is ` high - scale supersymmetry ' @xcite , in which all the superpartners are relegated at a very high energy and dm is plausibly made of axions . what can lhc data to - date say about these scenarios
? an interesting feature of both models is that , assuming that there is no additional field content besides the one of the mssm , the value of the higgs mass is very precisely determined in terms of the parameters of the theory , in particular the scale @xmath38 , with some variability depending on the boundary condition at the scale @xmath38 . by carefully performing the running and the various matchings @xcite-@xcite@xcite , it can be seen that a higgs mass of 125 gev implies @xmath39 tev for the case of split susy , while no bound can be derived on @xmath38 in the case of high - scale susy .
finally , these models are practically designed so that the direct bounds on sparticle masses do not touch them , and in particular the second case would not be touched even by eventual strong exclusion bounds on higgsinos and gauginos . if , on the contrary , particles resembling charginos and neutralinos are found , then it will be crucial to measure their couplings with high precision and see whether they are compatible with supersymmetry or not . to do so
, a linear collider would be probably necessary .
so far we have assumed that the hint for a higgs boson at 125 gev stands indeed for a signal , and we discussed some of the consequences for supersymmetric models .
given the importance of the issue , however , it is clear that we can not content ourselves with a @xmath1 evidence , and more data is definitely needed ( as the experimental collaborations also say ) . moreover the fact that the channel in which most of the excess is seen , namely the diphoton one , prefers a rate that is _ larger _ than in the sm is precisely what one would expect in case of a background fluctuation .
let us than say something about the possibility ( that may be disproved quite soon ) that the present excess goes away and a sm - like higgs boson is excluded up to a mass of 500 - 600 gev .
the sm alone would then be excluded by the combination of higgs searches and electroweak precision tests . in the context of supersymmetry
, the most plausible explanation would be that the cp - even scalars of the higgs sector have a reduced production cross section times branching ratio with respect to the sm higgs boson .
namely , what counts is the quantity : @xmath40 where @xmath41 is a given final state . in particular , above @xmath42 the higgs boson decay into vector bosons is the dominant channel in the sm .
looking at the exclusion plots we see that a heavy ` not - so - much - sm - like higgs boson ' is still allowed by data provided that @xmath43 ( @xmath44 ) is smaller than @xmath45 .
this suppression is not difficult to achieve , generally speaking , in supersymmetric models @xcite@xcite , thanks to mixing effecs ( see e.g. @xcite ) , or top - loop effects that modify the gluon - fusion efficiency @xcite , or a depletion of @xmath46 due to a significant decay width into two neutralino lsp . in the case of @xmath27susy , in which such non - sm - like higgs particles above @xmath42
can be achieved very easily , it can be seen that it is not implausible to have @xmath47 for both the cp - even scalars , and it can also happen that the next to lightest state is more easily detectable than the lightest one @xcite-@xcite . in conclusion , strictly speaking at the moment of writing it is not excluded that the hint for a higgs boson at 125 gev is just a fluctuation of the background . in this case
there may be one or more not - so - much - sm - like higgs bosons with mass between 200 and 300 gev and @xmath48 .
waiting for more data to disprove these possibilities , we can say that if no sm - like higgs boson is found at 125 gev , then in reasonable models we typically expect at least one almost - sm - like such scalar below 500 gev with . ] @xmath49 .
if the lightest scalar , with reduced couplings , is found to be above 200 gev , then in the context of supersymmetry this would be a strong indication for models with largely increased quartic coupling , like @xmath27susy .
on the contrary , if a sm - like higgs boson is indeed found at 125 gev , then eventual additional scalars can have very low @xmath50-values . as a last issue ,
let us give attention to the possibility that the recently measured direct cp - violation in the @xmath4-meson decays , reported in equation ( [ eq : deltaacp ] ) , is a signal of physics beyond the sm . whether the experimental value can be explained within the sm or not is still under discussion .
what we can say is that it is _ possible _ that the sm accounts for the observed effect @xcite@xcite , but maybe not very easily or at least it is fair to say that _ new physics is not implausible _ ( see e.g. @xcite -@xcite ) . in the np interpretation ,
what one needs is a rationale to understand why the main effect comes out in a @xmath51 process without disturbing the very tightly - constrained @xmath52 ones . referring to @xcite for details , a way to understand
it in the context of supersymmetric models may be that flavor violation in the squark sector comes mainly from left - right mixings ( i.e. from the a - terms ) , while the left - left and right - right mixings are subleading .
the size of the experimental value can then be naturally explained in the configuration called ` disoriented a - terms ' , in which the size of the entries of the a - term matrices is the same of those of the corresponding yukawa matrices but without respecting exact proportionality . a concrete realization of this configuration that may actually be considered also independently of supersymmetry can be @xcite to invoke the paradigm of partial compositeness .
in fact , if the flavor structure is originated by this mechanism , than one expects an effective lagrangian of the type : @xmath53 while the yukawa matrices are given by : @xmath54 where the @xmath55 are ` suppression factors ' generated by the strong dynamics .
it is then clear that @xmath56 operators tend to be suppressed more than the @xmath57 ones , and thus we are going in the right direction .
moreover as discussed e.g. in @xcite and @xcite , the best candidates for producing a sizable effect in the charm sector while being consistent with the other flavor constraints are the @xmath58 chromomagnetic operators , such as : @xmath59 together with the other one with opposite chiralities whose effect is typically relatively suppressed by a factor @xmath60 . in the notation ( [ eq : finalparametrization ] ) we have , following @xcite@xcite : @xmath61 where im@xmath62 is the hadronic matrix element of @xmath63 , and we used the fact that @xmath64 is expected to be @xmath65 times the cabibbo angle . since a reasonable estimate can be @xmath66 , in order to be compatible with the observed value ( [ eq : deltaacp ] ) one needs either an enhanced sm contribution @xmath67 , or a np contribution that corresponds in our case to : @xmath68 in a supersymmetric context these flavor - violating operators are generated by the flavor - mixings in the squark mass matrices , that are of the form ( looking at the order of magnitudes , no proportionality between matrices ! ) : @xmath69 moreover the effects come at the loop level , so that ( [ eq : fixingthescale ] ) can translate into : @xmath70 where @xmath71 is the typical scale of superpartner masses . notice that until last year this would have been more than welcome , as a possible signal of supersymmetric particles , while now it starts being in tension with the bounds from direct detection ! for a detailed study along this direction
we refer to @xcite , in which this possibility is thoroughly discussed .
it can be seen that the other bounds on flavor - violation in the quark sector are easily satisfied , while the electric dipole moments ( edm ) turn out to be the strongest constraint .
this picture could also be extended to the lepton sector , and in this case the strongest bounds come from @xmath72 and the electon edm .
we discussed the status of a few motivated beyond - the - mssm models in the light of the recent lhc data .
our conclusions can be summarized as follows : 1 .
* insisting on naturalness : natural supersymmetry . *
the present hint of a higgs boson at 125 gev tells us that , in the mssm , the stop should be too heavy to be natural according to the criterion @xmath26 .
this instead is not a problem for non - minimal extensions , and thus we can at least say that beyond - the - mssm models with increased quartic coupling are more motivated than before . apart from that , the most solid input so far is that the first two generations must be above the tev , thus either one tolerates @xmath73 or typically one needs the third generation to be lighter than the first two . for this reason , ` more minimal ' or ` hierarchical ' scenarios
can now be renamed ` natural supersymmetry ' ( since it is the only left natural possibility ... unless point 4 , below ) .
actually in @xmath27susy the stop at about 1 tev , eventually degenerate with the other squarks , can be compatible with @xmath26 . in this case
however also the first two generations can naturally be much heavier .
low scale susy breaking is also favoured in this view , since broadly speaking it reduces the amount of tuning .
+ how to exclude natural supersymmetry ?
a relatively fair statement may be : _ not finding the higgs boson _
( which would actually exclude also most of the finely - tuned susy scenarios ) , _ and/or especially not finding superpartners at the lhc_. 2 .
* escaping the lhc bounds . * r - parity violation after all is not compulsory , and without it the present bounds are partly escaped .
other possibilities can be susy with a compressed spectrum , stealth mechanisms , or dirac gauginos .
it is important to be ready to these distortions of the ` usual ' lhc phenomenology .
+ this is clearly a very hot topic ! in the future ,
if the bounds get stronger , we might be forced to call it something like ` alternative natural supersymmetry ' ... 3 . * insisting only on dm and unification and/or strings : split and high - scale susy . * by construction
, these possibilities are not touched by direct bounds on squarks .
thus for the moment the only experimental input for these models is the value of the higgs boson mass , from which we can say that split susy has to be ` not - so - split ' ( @xmath39 tev ) , while nothing can be said so far in case of high - scale susy .
* higgs boson not at 125 gev . *
the importance of the issue is such that we can not content ourselves with a @xmath74 evidence : although there are a lot of good reasons for the higgs boson to be there , the significance of the present hint is such that it may well be a background fluctuation .
more importantly , heavier scalars with reduced couplings are present in susy as well as in other motivated extensions of the sm , and they must be looked for .
+ what can be said is that , if the present hint turns out to be a fluctuation , then it is obviously crucial to look for the most - sm - higgs - like particle ( with reduced couplings ) . in this case , it is fair to say that such a particle can not be considered to be reasonably excluded until the exclusion bound at 95% c.l .
is at least down to @xmath75 .
* direct cpv in @xmath4 decays . * in the context of supersymmetry , assuming that the sm contribution is not enough to explain the data , a very natural possibility is to have ` disoriented a - terms ' .
this may be explicitly realized in a context in which the flavor structure has something to do with the mechanism of partial compositeness .
if this is the case , new effects are predicted to be around the corner in the neutron edm and , if the picture is extended to leptons in the most straightforward way , in the electron edm and in @xmath72 . to conclude , in the near future a very fundamental question has a chance to be answered : _ was naturalness a good guiding principle ?
_ the lhc will tell us .
the very minimal requirement is a colored sector mainly coupled to the top , at or below the tev .
if supersymmetry is the way of nature , this will be the scalar partner of the top .
thus a very important point is whether the stop is there or not .
notice that if this sector is made of scalars the first option would be susy , while in case of fermions the most plausible possibility would be compositeness . in any case , also being open to the possibility that naturalness is implemented in an unexpected way , at least new particles / resonances within the lhc reach and/or a non - standard higgs boson are requred .
thus , in case the higgs boson is discovered , it will be crucial to measure its couplings and check whether it is sm - like or not , also looking at ww - scattering and looking for a second higgs doublet .
if the higgs particle is discovered to be exactly sm - like and no new particles are found , this will be a _ _ strong indication that there is fundamental finetuning in nature _ _ , we like it or not . on the contrary , a non - standard higgs boson and especially new particles would probably tell us that naturalness arguments were correct . confirming or discarding the naturalness of the fermi scale will be a very nontrivial conceptual input for the scientific though .
thus for the moment it makes sense to keep insisting on natural supersymmetry , until it is found or excluded .
a quite non - conventional alternative is that _ nothing _ is found at the lhc , not even the higgs particle , which is probably what most people expect least .
finally , we must not forget about flavor physics : even in presence of a totally - sm - like higgs boson and no extra particles , non - ckm flavor effects would at least imply that there is some new physics relatively at hand , although maybe beyond the lhc reach for direct discovery .
figure [ fig : outlook ] is a pictorial summary of these last considerations .
i thank riccardo barbieri , gino isidori , and riccardo rattazzi for useful discussions .
i also thank the organizers of the session `` electroweak and searches '' at the `` xx international workshop on deep inelastic scattering and related subjects '' for their kind invitation to talk about this subject . this work is supported by the swiss national science foundation under grant 200021 - 125237 .
g. aad _ et al . _ [ atlas collaboration ] , phys .
b * 710 * ( 2012 ) 49 http://arxiv.org/abs/1202.1408[_arxiv:1202.1408_ ] [ hep - ex ] .
updated results : http://cdsweb.cern.ch/record/1430033[atlas-conf-2012-019 ] .
g. aad _ et al . _ [ atlas collaboration ] , phys .
b * 710 * ( 2012 ) 67 http://arxiv.org/abs/1109.6572[_arxiv:1109.6572_ ] [ hep - ex ] .
updated results : http://cdsweb.cern.ch/record/1432199[atlas-conf-2012-033 ] .
r. aaij _ et al . _
[ lhcb collaboration ] , http://arxiv.org/abs/1112.0938[_arxiv:1112.0938_ ] [ hep - ex ] .
[ cdf collaboration ] , http://www-cdf.fnal.gov/physics/new/bottom/120216.blessed-cpvcharm10fb/cdf10784.pdf[cdf note 10784 ] r. barbieri , http://www.sns.it/didattica/scienze/menunews/personale/docenti/barbieri/materiale/supersymmetry/ download / supersymmetry.pdf[``where are we with the search for supersymmetry '' ] , talk given at the sns of pisa , 12 october 2011 ; http://download.sns.it/barbieri/barbieri_paris_nov_2011.pdf[``susy phenomenology , including susy higgs '' ] , talk given at the 2011hcp , paris , 14 november 2011 ; http://download.sns.it/barbieri/cms_school_2012.pdf[``physics at the electroweak scale after the first 5 inverse femtobarn at the lhc '' ] , talk given at the cms data analysis school , pisa , 23 january 2012 .
s. dimopoulos and g. f. giudice , phys .
b * 357 * , 573 ( 1995 ) , http://arxiv.org/pdf/hep-ph/9507282[_arxiv:hep-ph/9507282_ ] .
m. a. luty , http://arxiv.org/abs/hep-th/0509029[_arxiv:hep-th/0509029_ ] .
g. f. giudice , in * kane , gordon ( ed . ) , pierce , aaron ( ed . ) : perspectives on lhc physics * 155 - 178 http://arxiv.org/abs/0801.2562[_arxiv:0801.2562_ ] .
r. barbieri , `` effective field theories and physics beyond the standard model , '' proceedings of the international school of subnuclear physics _
effective theories and fundamental interactions _ , the subnuclear series * 34 * , world scientific ( 1996 ) .
s. weinberg , phys .
* 59 * ( 1987 ) 2607 .
n. arkani - hamed , s. dimopoulos and g. r. dvali , phys .
b * 429 * ( 1998 ) 263 http://arxiv.org/abs/hep-ph/9803315[_arxiv:hep-ph/9803315_ ] .
i. antoniadis , n. arkani - hamed , s. dimopoulos and g. r. dvali , phys . lett .
b * 436 * ( 1998 ) 257 http://arxiv.org/abs/hep-ph/9804398[_arxiv:hep-ph/9804398_ ] .
s. p. martin , `` a supersymmetry primer , '' http://arxiv.org/abs/hep-ph/9709356[_arxiv:hep-ph/9709356_ ] .
j. a. casas , j. r. espinosa and i. hidalgo , jhep * 0401 * ( 2004 ) 008 , http://arxiv.org/pdf/hep-ph/0310137[_arxiv:hep-ph/0310137_ ] .
m. dine , r. g. leigh and a. kagan , phys .
d * 48 * ( 1993 ) 4269 , http://arxiv.org/pdf/hep-ph/9304299[_arxiv:hep-ph/9304299_ ] .
p. pouliot and n. seiberg , phys .
b * 318 * ( 1993 ) 169 , http://arxiv.org/pdf/hep-ph/9308363[_arxiv:hep-ph/9308363_ ] .
a. pomarol and d. tommasini , nucl .
b * 466 * , 3 ( 1996 ) , http://arxiv.org/pdf/hep-ph/9507462[_arxiv:hep-ph/9507462_ ] .
r. barbieri , g. r. dvali and l. j. hall , phys .
b * 377 * , 76 ( 1996 ) , http://arxiv.org/pdf/hep-ph/9512388[_arxiv:hep-ph/9512388_ ] .
a. g. cohen , d. b. kaplan and a. e. nelson , phys .
b * 388 * , 588 ( 1996 ) , http://arxiv.org/pdf/hep-ph/9607394[_arxiv:hep-ph/9607394_ ] .
r. barbieri , l. j. hall and a. romanino , phys .
b * 401 * ( 1997 ) 47 , http://arxiv.org/pdf/hep-ph/9702315[_arxiv:hep-ph/9702315_ ] .
r. barbieri , e. bertuzzo , m. farina , p. lodone , d. zhuridov , jhep * 1012 * ( 2010 ) 070 .
http://arxiv.org/abs/1011.0730[_arxiv:1011.0730_ ] .
r. barbieri , p. lodone , d. m. straub , jhep * 1105 * ( 2011 ) 049 .
http://arxiv.org/abs/1102.0726[_arxiv:1102.0726_ ] .
r. barbieri , g. isidori , j. jones - perez , p. lodone , d. m. straub , eur .
j. * c71 * ( 2011 ) 1725
. http://arxiv.org/abs/1105.2296[_arxiv:1105.2296_ ] .
l. j. hall , d. pinner and j. t. ruderman , http://arxiv.org/abs/1112.2703[_arxiv:1112.2703_ ] .
a. arbey , m. battaglia , a. djouadi , f. mahmoudi and j. quevillon , phys .
b * 708 * ( 2012 ) 162 http://arxiv.org/abs/1112.3028[_arxiv:1112.3028_ ] .
p. draper , p. meade , m. reece and d. shih , http://arxiv.org/abs/1112.3068[_arxiv:1112.3068_ ] .
d. m. ghilencea , h. m. lee and m. park , http://arxiv.org/abs/1203.0569[_arxiv:1203.0569_ ] .
m. carena , s. gori , n. r. shah and c. e. m. wagner , http://arxiv.org/abs/1112.3336[_arxiv:1112.3336_ ] .
r. harnik , g. d. kribs , d. t. larson and h. murayama , phys .
d * 70 * , 015002 ( 2004 ) http://arxiv.org/abs/hep-ph/0311349[_arxiv:hep-ph/0311349_ ] .
s. chang , c. kilic and r. mahbubani , phys .
d * 71 * , 015003 ( 2005 ) http://arxiv.org/abs/hep-ph/0405267[_arxiv:hep-ph/0405267_ ] .
a. birkedal , z. chacko and y. nomura , phys .
d * 71 * , 015006 ( 2005 ) http://arxiv.org/abs/hep-ph/0408329[_arxiv:hep-ph/0408329_ ] .
a. delgado and t. m. p. tait , jhep * 0507 * ( 2005 ) 023 http://arxiv.org/abs/hep-ph/0504224[_arxiv:hep-ph/0504224_ ] .
p. batra , a. delgado , d. e. kaplan and t. m. p. tait , jhep * 0402 * , 043 ( 2004 ) http://arxiv.org/abs/hep-ph/0309149[_arxiv:hep-ph/0309149_ ] .
p. lodone , jhep * 1005 * ( 2010 ) 068 http://arxiv.org/abs/1004.1271[_arxiv:1004.1271_ ] .
j. r. espinosa and m. quiros , phys .
b * 279 * , 92 ( 1992 ) .
p. batra , a. delgado , d. e. kaplan and t. m. p. tait , jhep * 0406 * , 032 ( 2004 ) http://arxiv.org/abs/hep-ph/0404251[_arxiv:hep-ph/0404251_ ] .
a. delgado , http://arxiv.org/abs/hep-ph/0409073[_arxiv:hep-ph/0409073_ ] .
a. maloney , a. pierce and j. g. wacker , jhep * 0606 * , 034 ( 2006 ) http://arxiv.org/abs/hep-ph/0409127[_arxiv:hep-ph/0409127_ ] .
s. p. martin , phys .
d * 81 * , 035004 ( 2010 ) http://arxiv.org/abs/0910.2732[_arxiv:0910.2732_ ] .
a. brignole , j. a. casas , j. r. espinosa and i. navarro , nucl .
b * 666 * ( 2003 ) 105 , http://arxiv.org/pdf/hep-ph/0301121[_arxiv:hep-ph/0301121_ ] .
m. dine , n. seiberg and s. thomas , phys .
d * 76 * ( 2007 ) 095004 http://arxiv.org/abs/0707.0005[_arxiv:0707.0005_ ] .
i. antoniadis , e. dudas , d. m. ghilencea and p. tziveloglou , nucl .
b * 808 * ( 2009 ) 155 http://arxiv.org/abs/0806.3778[_arxiv:0806.3778_ ] .
m. carena , k. kong , e. ponton and j. zurita , phys .
d * 81 * ( 2010 ) 015001 http://arxiv.org/abs/0909.5434[_arxiv:0909.5434_ ] .
i. antoniadis , e. dudas , d. m. ghilencea and p. tziveloglou , nucl .
b * 831 * ( 2010 ) 133 http://arxiv.org/abs/0910.1100[_arxiv:0910.1100_ ] .
y. kats , p. meade , m. reece and d. shih , jhep * 1202 * ( 2012 ) 115 http://arxiv.org/abs/1110.6444[_arxiv:1110.6444_ ] .
r. essig , e. izaguirre , j. kaplan and j. g. wacker , jhep * 1201 * ( 2012 ) 074 http://arxiv.org/abs/1110.6443[_arxiv:1110.6443_ ] . c. brust , a. katz , s. lawrence and r. sundrum , http://arxiv.org/abs/1110.6670[_arxiv:1110.6670_ ] .
a. delgado and m. quiros , phys .
d * 85 * ( 2012 ) 015001 http://arxiv.org/abs/1111.0528[_arxiv:1111.0528_ ] .
z. kang , j. li and t. li , http://arxiv.org/abs/1201.5305[_arxiv:1201.5305_ ] .
j. cao , z. heng , j. m. yang , y. zhang and j. zhu , http://arxiv.org/abs/1202.5821[_arxiv:1202.5821_ ] .
u. ellwanger , http://arxiv.org/abs/1112.3548[_arxiv:1112.3548_ ] .
d. a. vasquez , g. belanger , c. boehm , j. da silva , p. richardson and c. wymant , http://arxiv.org/abs/1203.3446[_arxiv:1203.3446_ ] .
u. ellwanger and c. hugonie , http://arxiv.org/abs/1203.5048[_arxiv:1203.5048_ ] .
n. craig , m. mccullough and j. thaler , http://arxiv.org/abs/1201.2179[_arxiv:1201.2179_ ] .
n. craig , m. mccullough and j. thaler , http://arxiv.org/abs/1203.1622[_arxiv:1203.1622_ ] .
n. craig , d. green and a. katz , http://arxiv.org/abs/1103.3708[_arxiv:1103.3708_ ] .
g. larsen , y. nomura and h. l. l. roberts , http://arxiv.org/abs/1202.6339[_arxiv:1202.6339_ ] .
t. gherghetta and a. pomarol , phys .
d * 67 * ( 2003 ) 085018 http://arxiv.org/abs/hep-ph/0302001[_arxiv:hep-ph/0302001_ ] .
m. gabella , t. gherghetta and j. giedt , phys .
d * 76 * ( 2007 ) 055001 http://arxiv.org/abs/0704.3571[_arxiv:0704.3571_ ] .
r. sundrum , jhep * 1101 * ( 2011 ) 062 http://arxiv.org/abs/0909.5430[_arxiv:0909.5430_ ] .
l. j. hall and m. suzuki , nucl .
b * 231 * ( 1984 ) 419 .
j. r. ellis , g. gelmini , c. jarlskog , g. g. ross and j. w. f. valle , phys .
b * 150 * ( 1985 ) 142 .
g. g. ross and j. w. f. valle , phys .
b * 151 * ( 1985 ) 375 .
v. d. barger , g. f. giudice and t. han , phys .
d * 40 * ( 1989 ) 2987
. s. dawson , nucl .
b * 261 * ( 1985 ) 297 .
r. barbieri and a. masiero , nucl .
b * 267 * ( 1986 ) 679 .
jackson [ on behalf of the atlas collaboration ] , http://arxiv.org/abs/1112.0369[_arxiv:1112.0369_ ] [ hep - ex ] .
[ atlas collaboration ] , http://cdsweb.cern.ch/record/1432202/files/atlas-conf-2012-035.pdf[atlas-conf-2012-035 ] .
g. dambrosio , g. f. giudice , g. isidori and a. strumia , nucl .
b * 645 * ( 2002 ) 155 http://arxiv.org/abs/hep-ph/0207036[_arxiv:hep-ph/0207036_ ] .
g. isidori and d. m. straub , http://arxiv.org/abs/1202.0464[_arxiv:1202.0464_ ] .
e. nikolidakis and c. smith , phys .
d * 77 * ( 2008 ) 015021 http://arxiv.org/abs/0710.3129[_arxiv:0710.3129_ ] . c. smith , http://arxiv.org/abs/0809.3152[_arxiv:0809.3152_ ] . c. csaki , y. grossman and b. heidenreich , http://arxiv.org/abs/1111.1239[_arxiv:1111.1239_ ] .
g. arcadi , l. di luzio and m. nardecchia , http://arxiv.org/abs/1111.3941[_arxiv:1111.3941_ ] .
t. j. lecompte and s. p. martin , phys .
d * 84 * ( 2011 ) 015004 http://arxiv.org/abs/1105.4304[_arxiv:1105.4304_ ] .
t. j. lecompte and s. p. martin , phys .
d * 85 * ( 2012 ) 035023 http://arxiv.org/abs/1111.6897[_arxiv:1111.6897_ ] .
l. j. hall and l. randall , nucl .
b * 352 * ( 1991 ) 289 .
l. randall and n. rius , phys .
b * 286 * ( 1992 ) 299 .
p. j. fox , a. e. nelson and n. weiner , jhep * 0208 * ( 2002 ) 035 http://arxiv.org/abs/hep-ph/0206096[_arxiv:hep-ph/0206096_ ] .
n. arkani - hamed , s. dimopoulos , g. f. giudice and a. romanino , nucl .
b * 709 * ( 2005 ) 3 http://arxiv.org/abs/hep-ph/0409232[_arxiv:hep-ph/0409232_ ] .
n. bernal , a. djouadi , p. slavich , jhep * 0707 * ( 2007 ) 016 .
http://arxiv.org/abs/0705.1496[_arxiv:0705.1496_ ] .
w. kilian , t. plehn , p. richardson and e. schmidt , eur .
j. c * 39 * ( 2005 ) 229 http://arxiv.org/abs/hep-ph/0408088[_arxiv:hep-ph/0408088_ ] .
j. l. hewett , b. lillie , m. masip and t. g. rizzo , jhep * 0409 * ( 2004 ) 070 http://arxiv.org/abs/hep-ph/0408248[_arxiv:hep-ph/0408248_ ] .
k. cheung and w. -y .
keung , phys .
d * 71 * ( 2005 ) 015015 http://arxiv.org/abs/hep-ph/0408335[_arxiv:hep-ph/0408335_ ] .
l. anchordoqui , h. goldberg and c. nunez , phys .
d * 71 * ( 2005 ) 065014 http://arxiv.org/abs/hep-ph/0408284[_arxiv:hep-ph/0408284_ ] .
j. g. gonzalez , s. reucroft and j. swain , phys .
d * 74 * ( 2006 ) 027701 http://arxiv.org/abs/hep-ph/0504260[_arxiv:hep-ph/0504260_ ] .
d. s. m. alves , e. izaguirre and j. g. wacker , http://arxiv.org/abs/1108.3390[_arxiv:1108.3390_ ] .
m. e. cabrera , j. a. casas and a. delgado , phys .
* 108 * ( 2012 ) 021802 http://arxiv.org/abs/1108.3867[_arxiv:1108.3867_ ] .
g. f. giudice and a. strumia , http://arxiv.org/abs/1108.6077[_arxiv:1108.6077_ ] .
t. tsukamoto , k. fujii , h. murayama , m. yamaguchi and y. okada , phys .
d * 51 * ( 1995 ) 3153 .
j. l. feng , m. e. peskin , h. murayama and x. r. tata , phys .
d * 52 * ( 1995 ) 1418 http://arxiv.org/abs/hep-ph/9502260[_arxiv:hep-ph/9502260_ ] .
m. m. nojiri , k. fujii and t. tsukamoto , phys .
d * 54 * ( 1996 ) 6756 http://arxiv.org/abs/hep-ph/9606370[_arxiv:hep-ph/9606370_ ] .
k. -i .
hikasa and y. nakamura , z. phys .
c * 70 * ( 1996 ) 139 [ erratum - ibid .
c * 71 * ( 1996 ) 356 ] http://arxiv.org/abs/hep-ph/9501382[_arxiv:hep-ph/9501382_ ] .
d. m. pierce , j. a. bagger , k. t. matchev and r. -j .
zhang , nucl .
b * 491 * ( 1997 ) 3 http://arxiv.org/abs/hep-ph/9606211[_arxiv:hep-ph/9606211_ ] .
h. -c .
cheng , j. l. feng , n. polonsky , phys .
d56 * ( 1997 ) 6875 - 6884
. http://arxiv.org/abs/hep-ph/9706438[_arxiv:hep-ph/9706438_ ] .
h. -c .
cheng , j. l. feng and n. polonsky , phys .
d * 57 * ( 1998 ) 152 http://arxiv.org/abs/hep-ph/9706476[_arxiv:hep-ph/9706476_ ] .
e. katz , l. randall , s. -f .
su , nucl . phys .
* b536 * ( 1998 ) 3 - 28 .
http://arxiv.org/abs/hep-ph/9801416[_arxiv:hep-ph/9801416_ ] .
m. spira , a. djouadi , d. graudenz and p. m. zerwas , nucl .
b * 453 * ( 1995 ) 17 http://arxiv.org/abs/hep-ph/9504378[_arxiv:hep-ph/9504378_ ] .
a. djouadi , phys .
* 459 * ( 2008 ) 1 http://arxiv.org/abs/hep-ph/0503173[_arxiv:hep-ph/0503173_ ] .
m. carena , p. draper , t. liu and c. wagner , phys .
d * 84 * ( 2011 ) 095010 http://arxiv.org/abs/1107.4354[_arxiv:1107.4354_ ] .
i. low and s. shalgar , jhep * 0904 * ( 2009 ) 091 http://arxiv.org/abs/0901.0266[_arxiv:0901.0266_ ] .
p. lodone , int .
j. mod .
a * 26 * ( 2011 ) 4053 http://arxiv.org/abs/1105.5248[_arxiv:1105.5248_ ] .
e. bertuzzo , m. farina , http://arxiv.org/abs/1105.5389[_arxiv:1105.5389_ ] .
e. bertuzzo and m. farina , http://arxiv.org/abs/1112.2190[_arxiv:1112.2190_ ] .
y. grossman , a. l. kagan and y. nir , phys .
d * 75 * ( 2007 ) 036008 http://arxiv.org/abs/hep-ph/0609178[_arxiv:hep-ph/0609178_ ] .
g. isidori , j. f. kamenik , z. ligeti and g. perez , http://arxiv.org/abs/1111.4987[_arxiv:1111.4987_ ] .
h. -n .
li , c. -d .
lu and f. -s .
yu , http://arxiv.org/abs/1203.3120[_arxiv:1203.3120_ ] .
e. franco , s. mishima and l. silvestrini , http://arxiv.org/abs/1203.3131[_arxiv:1203.3131_ ] .
d. b. kaplan , nucl .
b * 365 * ( 1991 ) 259 .
t. gherghetta and a. pomarol , nucl .
b * 586 * ( 2000 ) 141 http://arxiv.org/abs/hep-ph/0003129[_arxiv:hep-ph/0003129_ ] .
k. agashe , g. perez and a. soni , phys .
d * 71 * ( 2005 ) 016002 http://arxiv.org/abs/hep-ph/0408134[_arxiv:hep-ph/0408134_ ] .
r. contino , t. kramer , m. son and r. sundrum , jhep * 0705 * ( 2007 ) 074 http://arxiv.org/abs/hep-ph/0612180[_arxiv:hep-ph/0612180_ ] .
r. contino , http://arxiv.org/abs/1005.4269[_arxiv:1005.4269_ ] .
m. redi and a. weiler , jhep * 1111 * ( 2011 ) 108 http://arxiv.org/abs/1106.6357[_arxiv:1106.6357_ ] .
r. barbieri , g. isidori and d. pappadopulo , jhep * 0902 * ( 2009 ) 029 http://arxiv.org/abs/0811.2888[_arxiv:0811.2888_ ] .
s. davidson , g. isidori and s. uhlig , phys .
b * 663 * , 73 ( 2008 ) , http://arxiv.org/abs/0711.3376[_arxiv:0711.3376_ ] . y. nomura , m. papucci and d. stolarski , phys . rev .
d * 77 * , 075006 ( 2008 ) , http://arxiv.org/abs/0712.2074[_arxiv:0712.2074_ ] . s. davidson , g. isidori and s. uhlig , phys .
b * 663 * , 73 ( 2008 ) , http://arxiv.org/abs/0711.3376[_arxiv:0711.3376_ ] .
y. nomura , m. papucci and d. stolarski , jhep * 0807 * , 055 ( 2008 ) , http://arxiv.org/abs/0802.2582[_arxiv:0802.2582_ ] . y. nomura and d. stolarski , phys .
d * 78 * , 095011 ( 2008 ) , http://arxiv.org/abs/0808.1380[_arxiv:0808.1380_ ] .
e. dudas , g. von gersdorff , j. parmentier and s. pokorski , jhep * 1012 * , 015 ( 2010 ) , http://arxiv.org/abs/1007.5208[_arxiv:1007.5208_ ] .
a. e. nelson and m. j. strassler , jhep * 0009 * ( 2000 ) 030 http://arxiv.org/abs/hep-ph/0006251[_arxiv:hep-ph/0006251_ ] .
a. e. nelson and m. j. strassler , jhep * 0207 * ( 2002 ) 021 http://arxiv.org/abs/hep-ph/0104051[_arxiv:hep-ph/0104051_ ] .
d. b. kaplan , h. georgi and s. dimopoulos , phys .
b * 136 * ( 1984 ) 187 .
d. b. kaplan and h. georgi , phys .
b * 136 * ( 1984 ) 183 .
h. georgi and d. b. kaplan , phys .
b * 145 * ( 1984 ) 216 .
m. j. dugan , h. georgi and d. b. kaplan , nucl .
b * 254 * ( 1985 ) 299 .
a. falkowski , s. rychkov and a. urbano , http://arxiv.org/abs/1202.1532[_arxiv:1202.1532_ ] .
a. azatov , r. contino and j. galloway , http://arxiv.org/abs/1202.3415[_arxiv:1202.3415_ ] .
j. r. espinosa , c. grojean , m. muhlleitner and m. trott , http://arxiv.org/abs/1202.3697[_arxiv:1202.3697_ ] .
a. djouadi , http://arxiv.org/abs/1203.4199[_arxiv:1203.4199_ ] .
giardino , k. kannike , m. raidal and a. strumia , http://arxiv.org/abs/1203.4254[_arxiv:1203.4254_ ] . c. englert , b. jager , m. worek and d. zeppenfeld , http://arxiv.org/abs/0904.2119[_arxiv:0904.2119_ ] .
p. borel , r. franceschini , r. rattazzi and a. wulzer , http://arxiv.org/abs/1202.1904[_arxiv:1202.1904_ ] .
a. ballestrero , d. b. franzosi and e. maina , http://arxiv.org/abs/1203.2771[_arxiv:1203.2771_ ] .
k. blum and r. t. dagnolo , http://arxiv.org/abs/1202.2364[_arxiv:1202.2364_ ] .
l. maiani , a. d. polosa and v. riquer , http://arxiv.org/abs/1202.5998[_arxiv:1202.5998_ ] .
b. bellazzini , c. csaki , a. falkowski and a. weiler , phys .
d * 80 * ( 2009 ) 075008 http://arxiv.org/abs/0906.3026[_arxiv:0906.3026_ ] .
b. bellazzini , c. csaki , j. hubisz and j. shao , phys .
d * 83 * ( 2011 ) 095018 http://arxiv.org/abs/1012.1316[_arxiv:1012.1316_ ] . | we briefly review the status of motivated beyond - the - mssm phenomenology in the light of the lhc searches to date .
in particular , we discuss the conceptual consequences of the exclusion bounds , of the hint for a higgs boson at about 125 gev , and of interpreting the excess of direct cp violation in the charm sector as a signal of new physics .
we try to go into the various topics in a compact way while providing a relatively rich list of references , with particular attention to the most recent developments .
given at http://indico.cern.ch/conferenceotherviews.py?view=standard&confid=153252[``xx international workshop on deep inelastic scattering and related subjects '' ] , 26 - 30 march 2012 , bonn , germany . ] @xmath0 1.0 cm * supersymmetry phenomenology beyond + the mssm after 5/fb of lhc data * 1.0 cm paolo lodone
+ institut de thorie des phnomnes physiques , epfl , lausanne , switzerland + 1.0 cm |
[ b ! ] [ cols="^ " , ] several gnss constellations and augmentation systems are presently operational or under development , including the pioneering us global positioning system ( gps ) and the forthcoming european system , galileo .
these all - weather , long - term , stable and precise l - band signals can be used for bistatic remote sensing of the ocean surface and beyond , an emerging concept known as gnss - r . among several applications , two
have been emphasized by the community : sea - surface altimetry ( see and references therein ) and sea - surface `` speculometry '' ( a term discussed below ) , related to the statistical properties of sea surface gravity wave slopes .
although this paper addresses the latter , we note the intrinsic capability of gnss - r for providing long term co - located measurements of both surface roughness and sea level with high spatial and temporal resolution . as recently demonstrated with scatterometer measurements ( ) , such a capability would help to better quantify the relationship of velocities in the upper ocean ( driven by wind stress forcing ) with surface height dynamics .
thanks to its passive multistatic character , gnss - r clearly holds the potential to provide an unprecedented spatio - temporal sampling of the ocean surface .
the expected high spatial and temporal measurements can certainly serve operational applications .
for instance , and to follow successful scatterometer measurements , gnss - r can complement ocean winds and wave models . being rain immune , such new data could help quantify atmosphere - ocean coupling , including momentum and energy fluxes under extreme conditions for hurricane modeling .
many other scientific applications can be cited , such as sea surface breaking / whitecapping and gas exchange global characterization , important ingredients for understanding the ocean s biogeochemical response to , and its influence on , climate change .
indeed , a very promising approach to quantify co@xmath0 flux is to better assess the surface fractional area , which is readily measurable from surface slope measurements ( ) . in that case , gnss - r can provide a more direct measurement to extract gravity surface slope statistical properties and to quantify the role of the ocean in taking up increases of co@xmath0 .
in addition , l - band sea surface roughness data can be used to support l - band radiometric missions , such as smos ( soil moisture and ocean salinity ) and aquarius , to both quantify and efficiently separate roughness and salinity contributions to l - band radiometric brightness measurements .
inferring sea roughness from gnss - r data requires ( i ) a parametric description of the sea surface , ( ii ) an electromagnetic and instrument model for sea - surface scattering at l - band and ( iii ) the choice of a gnss - r data product to be inverted .
there is quite an agreement on the two first aspects in the literature .
it has been recognized that the scattering of gnss signals can be approximated by an _
effective _ geometric optics model , where the fundamental physical process is the scattering from facet - like surface elements .
this is the reason for the use of the term `` speculometry '' here , which stems from the latin word for mirror , _ speculo _ : the detected gnss - r return is dominated by the statistics of facet slopes and their curvatures at scales larger than the electromagnetic wavelength ( @xmath1 ) . under a gaussian assumption ,
three parameters fully define the detected l - band sea surface slope probability distribution ( pdf ) .
these parameters are encapsulated by the directional mean square slope , dmss@xmath2 , a symmetric tensor which results from the integration of the ocean energy spectrum at wavelengths larger than @xmath1 , and which characterizes the ellipsoidal shape of the slope pdf .
it is important to note that dmss@xmath2 has rarely been emphasized as the geophysical parameter of interest in the literature .
instead , most authors link sea roughness to the near surface wind vector , which is thought to be more useful for oceanographic and meteorological users . unfortunately , this is somewhat misleading , as the relationship between surface wind and roughness is not one - to - one and requires an additional modeling layer .
the connection between dmss@xmath2 and wind is affected by other factors ( e.g. , swell , fetch and degree of maturity ) , as is well known in the altimeter community ( see ) .
moreover , the product traditionally used for inversion in gnss - r speculometry is the 1d delay waveform of the reflected signal amplitude , from which the wind speed is inferred assuming an isotropic slope pdf ( see , e.g. , , or ) .
attempts have also been made to estimate the wind direction by fixing the pdf isotropy to some theoretical value ( around 0.7 ) and using at least two satellites reflections with different azimuths ( see , e.g. , , and ) . here
, we will work with a product of higher information content : the 2d ddm of the reflected signal amplitude .
as proposed in , the provision of an extra dimension opens the possibility of performing a robust estimation of all the dmss@xmath2 parameters through the direct fitting / estimation of the entire ddm . in , a first order approximation for inversion was proposed .
the full ddm - inversion technique proposed in is used for the first time here to analyze gnss - r data collected during the eddy experiment . this campaign and the altimetric data analysis
is reported elsewhere ( and ) , and is only briefly described in section 2 .
the retrieval methodology relies on a least - squares fit of the speculometric model , as discussed in the third section . in the fourth section ,
results are compared to ancillary data ( jason-1 radar altimeter , ecmwf numerical weather model , optical data ) . finally ,
another important outcome of the exhaustive exploitation of the information contained in the ddm product is related to the expected mean sea surface motion .
we evidence that a small part of the doppler spread can be attributed to the mean `` scatterer velocity '' , i.e. , the rapid motion of the @xmath1 ( or larger ) sized sea - surface facets that contribute the most to the detected signals .
the data set ( i.e. , the recorded direct and reflected gps signals together with the aircraft kinematic data ) was gathered during an airborne campaign carried out in september 2002 .
the aircraft overflew the mediterranean sea , off the coast of catalonia ( spain ) , northwards from the city of barcelona for about 150 km at 1000 m altitude and 45 - 75 m / s speed .
the area is crossed by the ground track # 187 of the jason-1 radar altimeter , which the aircraft overflew during the satellite overpass for precise comparison .
the track was overflown twice : the first time during the ascending pass ( from south to north ) at low speed ( 45 - 60 m / s ) and the second during the descending pass ( from north to south ) at a faster speed ( 65 - 75 m / s ) due to wind .
the time shift between the two passes over a same point on the track ranged from 45 min to 2h 15 min . during the ascending pass , prns 08 , 10 and 24 were visible with elevations spanning 30@xmath3 to 85@xmath3 while prns
08 , 10 and 29 were visible during the descending track with elevations between 40@xmath3 and 75@xmath3 .
the configuration of this test flight was not optimized for speculometry : from such low altitude , the sea - surface reflective area is essentially limited by the prn c / a code , and the glistening zone is coarsely delay - doppler mapped .
the raw gps signals were acquired with a modified turborogue receiver , sampled at 20.456 mhz and pre - processed with a dedicated software composed of two sub - units fed with the direct and reflected signals .
correlations were computed at 81 delay lags while the doppler dimension spanned -200 to 200 hz with a step of 20 hz . the coherent / incoherent integration times were respectively set to 20 ms and 10 s , meaning that the averaged ddm were produced at the rate of 0.1 hz after summation of 500 incoherent looks ( see figure [ ddm_example ] for a sample ddm ) .
following , the link between the ddm mean power at delay - doppler @xmath4 and the effective l - band sea - surface slope pdf , @xmath5 , is given by @xmath6 , \label{radar_eq}\end{aligned}\ ] ] where @xmath7 is the receiver antenna pattern , @xmath8 and @xmath9 the distances from generic point on sea - surface to transmitter and receiver , @xmath10 the scattering vector , @xmath11 the woodward ambiguity function ( waf ) , @xmath12 and @xmath13 the delay - doppler coordinates on the sea - surface and @xmath14 the delay / doppler offset of the geometric specular - point with respect to the direct signal ( the ddm `` center '' ) .
accounting for the receiver mean thermal noise @xmath15 and including a scaling parameter @xmath16 , the mean amplitude of the ddm can be written as @xmath17 . as discussed above ,
@xmath18 is described by the dmss@xmath2 parameter set , which defines an elliptic quadratic form in the 2d space of facet slopes .
mean - square slopes along major and minor principal axes are often called mss up - wind ( @xmath19 ) and mss cross - wind ( @xmath20 ) respectively . in the following , we will refer to the total mss ( mss@xmath21=@xmath22 , proportional to the ellipse area and directly related to nadir @xmath23 ) , the slope pdf azimuth ( spa , the direction of semi - major axis with respect to north ) and the slope pdf isotropy ( spi , equal to @xmath24 ) .
the inversion was performed through minimization of the mean square difference between model and data ddms .
numerical optimization was carried out by a steepest - slope - descent algorithm with a levenberg - marquardt type adjustment .
the main difficulty stemmed from the presence of several nuisance parameters in the forward model ( mainly @xmath25 and @xmath26 but also @xmath16 ) .
the ddm centers were affected by the aircraft trajectory ( altitude and vertical velocity ) to first order but also by geophysical parameters ( such as sea level ) .
they needed to be accurately known in order to estimate dmss@xmath2 .
for this reason , the dmss@xmath2 and nuisance parameters were jointly estimated in an iterative manner .
the values of dmss@xmath2 estimated along the descending track of the flight are shown on figure [ dmss ] .
the top plot illustrates the variations of total mss .
the inter - prn consistency is reasonable in the southern part of the track but worsens slightly in the northern part . for comparison ,
the total mss in ku - band was derived from the jason-1 @xmath27 co - located measurements at 1 hz sampling ( 7 km ) and 20 km resolution .
the jason-1 mss was obtained through the simple relationship mss@xmath28 , @xmath29 being the effective ( empirical ) fresnel coefficient , here set to 0.45 . as expected
, we observed that the level and dynamic of mss decreased with longer wavelength ( from 2 cm in ku - band to 19 cm in l - band ) .
the lower dynamic of l - band mss impeded any clear trend comparison , although the measurements of prn-10 seem in good agreement with jason-1 .
the jason-1 wind speed , derived from both the ku - band @xmath27 and the significant wave height ( of about 2 m ) , ranged from 9 to 13 m / s along the track . translating this wind speed into l - band mss through the spectrum of yields values between 0.0220 and 0.0255 , in - line with gnss - r results .
however , we must emphasize that the assumption of a wind - driven spectrum was not really warranted during the campaign .
spa estimation results are presented on the middle plot .
the inter - prn consistency is here very satisfying , and the apparent discrepancy in the southern part of the track can be explained as a degenerate solution of the estimation problem .
indeed , the inversion of ddm for the spa is degenerate in at least two cases : when the transmitter is at zenith or when the receiver moves towards the transmitter . in these scenarios ,
the delay - doppler lines mapping the glistening zone are fully symmetric around the receiver direction and it becomes impossible to distinguish a slope pdf from its mirror image about the receiver direction .
this effect is clearly observed here where the two found spa ( -20@xmath3 and 80@xmath3 ) are indeed symmetric around the aircraft heading direction ( 30@xmath3 ) . in this part of the track ,
the azimuth of prn-08 is about 50@xmath3 , almost aligned with the aircraft heading direction . in the northern part of the track ,
the estimated spa matches very well with the wind direction provided by ecmwf . in the southern part , the mismatch reaches up to 50@xmath3 .
however , we underline that surface wind is not the only element driving spa and that swell is likely to have contributed .
finally , the bottom plot shows the spi variations along the track .
the reflected signals are strongly directional . the wind - driven spectrum of for a mature sea predicts a spi value around 0.65 , largely insensitive to wind speed .
the apparent significant departure from this reference value is a probable signature of an under - developed sea with the presence of swell .
however , the relatively poor consistency among prn remains an issue to clarify : further work is needed to validate the accuracy of these spi estimates and to better understand the potential information ( and possible applicability constraints ) of this product . as a second and new outcome of this analysis
, we now discuss the signature of `` scatterer velocity '' in the data , i.e. , the signature of fastly moving sea - surface facets with size ( curvature ) larger than @xmath3020 cm .
such a signature can be detected when comparing the total mss along the ascending and descending tracks , as estimated by the least - squares approach : a drastic discrepancy ( up to 33% ) was observed for two passes shifted by less than hour over the same track point .
[ b ! ]
multipath effects could conceivably lead to a doppler width modulation .
however , some azimuth dependence should have been observed and was not .
another possible cause could be a changing aircraft attitude between ascending and descending tracks , but the aircraft roll and pitch values were checked to be nominal along both tracks .
moreover , a change in yaw would slightly impact the antenna pattern ground projection but would translate only into a doppler bandwidth cut and never a broadening .
we concluded that the most likely explanation was of geophysical origin : the doppler spectral width had been modulated by the overall sea surface dynamics as detected from the mean specular facet motion .
while the inversion approach assumed a still surface , the relative velocity between receiver and scatterer should be taken into account for proper ddm inversion . at high receiver
velocities this assumption is fine because the scatterer velocity impact will not be significant . at low speeds , however , scatterer velocity becomes relevant .
this analysis is consistent with the fact that the mss estimated in the ascending track ( slower aircraft speed , from 45 to 60 m / s ) showed abnormal high values compared to the ones estimated during the ascending track ( faster speed , from 65 to 75 m / s ) . in order to test this idea in a simple manner
, we used the results in , where it was demonstrated that a first order relationship exists between the moments of the ddm and the full set of dmss@xmath2 if the impact of the bistatic waf and antenna gain are neglected : mss=@xmath31 , where @xmath32 is the transmitter elevation , @xmath33 the receiver speed and @xmath34 the ddm doppler bandwidth . assuming that the mss did not vary significantly between the times of the ascending and descending passes and considering the relative velocity between sea - surface scatterers and aircraft , i.e. , @xmath35 for the descending / ascending passes respectively
, we can roughly solve for the scatterer velocity @xmath36 ( for simplicity assumed here parallel to the ground track ) . applying this scheme to all possible pairs of measurements and averaging , the plot of figure [ mean_scatterer_velocity ] results .
the apparent mean scatterer velocity variations correlate with the jason-1 wind speed variations
. moreover , from the dispersion relation an average scatterer speed of 8 m / s corresponds to waves of about 45 m wavelength .
the detected motion thus seems to be associated with the longer and more coherent wave components , consistent with optical observations of the swell vector ( wavelength and direction , see ) , revealing the presence of a northerly ( almost aligned with the flight track ) generated 48 m wave system .
as hypothesized , the quasi - specular facets contributing the most to the bistatic reflected signals may , on average , be assumed to almost coherently travel with the dominating longer wave peak component .
we have reported the inversion of gnss - r signals using , for the first time , the full delay - doppler maps for the retrieval of the sea - surface directional mean square slope , dmss@xmath37 : the estimates show good inter - prn consistency ( except for the measured anisotropy spi ) and fair agreement with other sources of data . the use of the full ddm further has helped to reveal a geophysical signature in gnss - r associated with the mean sea surface scatterer velocity . under quasi - specular conditions ,
sea surface scatterers are mostly associated to small slopes corresponding to longer waves with velocities that can reach 5 - 10 m / s , impacting significantly the doppler bandwidth of slow - moving receivers ( e.g. , airborne or ground - based , ) .
the detection of such a geophysical signature opens new opportunities for gnss - r speculometry : to infer either dmss@xmath2 or a combination of dmss@xmath2 and scatterer velocity , depending on the aircraft speed .
further investigations will be carried out to take into account the correct deformation of the doppler lines on the surface for the search of a scatterer velocity vector .
finally , we emphasize that the flight was not optimized for speculometry : higher and faster flights are needed in the future to consolidate the ddm inversion technique and to test new higher resolution inversion concepts .
the data analysis and the experimental campaign were respectively carried out under the esa contracts 3 - 10120/01/nl / sf ( oppscat ) and trp - etp-137.a .
we thank the institut cartografic de catalunya for flawless flight operations and aircraft gps / ins kinematic processing .
all starlab authors have contributed significantly ; the starlab author list has been ordered randomly .
garrison , j.l .
, anisotropy in reflected gps measurements of ocean winds , in _ proceedings of the 2003 workshop on oceanography with gnss - r _ , starlab , july 2003 .
available at http://starlab.es / gnssr2003/proceedings/. komjathy , a. , v. zavorotny , p. axelrad , g.h
. born , and j.l .
garrison , gps signal scattering from sea surface : wind speed retrieval using experimental data and theoretical model .
_ , 73:162174 , 2000 . | we report on the retrieval of directional sea surface roughness , in terms of its full directional mean square slope ( including direction and isotropy ) , from global navigation satellite system reflections ( gnss - r ) delay - doppler - map ( ddm ) data collected during an experimental flight at 1 km altitude .
this study emphasizes the utilization of the entire ddm to more precisely infer ocean roughness directional parameters .
in particular , we argue that the ddm exhibits the impact of both roughness and scatterer velocity .
obtained estimates are analyzed and compared to co - located jason-1 measurements , ecmwf numerical weather model outputs and optical data . |
there are a lot of many - body systems of different physical nature where the clustering of matter from smooth or almost smooth distributions of particles can occur ( see , for example , @xcite - @xcite ) .
such a process typically proceeds in particle ensembles with long - range forces , such as gravity or electric interactions , when the clustering leads to the formation of stable spatially inhomogeneous structures @xcite - @xcite .
these processes define the system s dynamics and existence . for example
, the pancake and adhesion models predict the persistence of a density peak at the flow caustics which appears from the initial distributions having gaussian initial fluctuations and is due to the development of the gravitational instability @xcite . in this case
the singular behavior makes it impossible to reverse the process through the destruction of such structures .
one may say that singularities are the factor which stabilizes the space structures .
the formation of singularities is a necessary condition for the implementation of these clustering mechanism ( see , for example , refs .
@xcite - @xcite ) .
it should be noted that a similar behavior can be expected for ideal and/or dissipative plasmas .
an example is the generation of a fast ion peak in the plasma expansion problem which is caused by the density collapse of the ion fluid for negligible dissipation becoming a stabilized fast ion peak at the front edge of the expanding plasma , when dissipation is included @xcite .
considering the analogy between gravitational and electrostatic fields , it may seem that such nonlinear dynamics is realized in the case of the coulomb interaction too . however , in the coulomb systems or the force - free systems there exists another possibility for structure formation without a singular ( blow - up ) behavior .
the dynamics for some initial conditions can lead to the formation of dynamical space structures @xcite in which the initial conditions play the role of a driving parameter precluding the formation of singularities but providing the formation of transient dynamical structures .
the present paper considers some aspects of such a transient clustering mechanism in one - dimensional geometry which , of course , looks like somewhat artificial since all real systems , as a rule , are not one - dimensional ones , in higher spatial dimensions , a large variety of nonlinear dynamical structures can be expected ( see , for example , @xcite ) .
nonetheless , this model indicates what may come about in real systems . in the case worked out there are no stabilizing factors such as dissipative processes or formation of caustics .
so one can expect that the structure emerged may break down or simply disappear .
it means that some smooth initial density distributions will produce a wave - like structure which over time will emerge and subsequently disappear yielding a distribution becoming close to the initial distribution .
we shall consider such a possibility and we try to define the conditions for which such an evolution of the one - dimensional system may be realized .
here , we consider two types of one - dimensional , time - dependent patterns of infinite , compressible medium .
first type is the medium containing only movable , particles which interact with each other via self - consistent gravitational ( or coulomb ) fields .
another type of the medium is a cold plasma containing movable electrons and immobile ions of constant density @xmath0 . in both media ,
the structure formation comes about because of the self - consistent forces caused by the distribution of the movable particles within the system and the effect of initial conditions .
the dynamics of these systems is governed by the euler - poisson hydrodynamic equations written in dimensionless variables as @xmath1 @xmath2 @xmath3 where @xmath4 is the density of movable component and @xmath5 the velocity of the fluid .
the self - consistent potential @xmath6 corresponds for constant @xmath7 to an attractive force and for @xmath8 to a repulsive force in the system , whereas the case @xmath9 describes an ordinary flow without potential forces . in our model
the constant @xmath10 can take two values : @xmath11 for the case of zero background describing one component flow with pure gravitational or coulomb interaction and @xmath12 for a nonzero background describing the evolution of cold plasma medium .
it is also worth noting that though the euler - poisson system appears in numerous applications including plasma physics and astrophysical applications there is a sense to consider it as an inviscid limit of burges turbulence model @xcite in the systems with collective interaction . from such point of view
the influence of initial conditions on the regular and chaotic structures formation has been of great interest . in order to study this effect
we consider an initial - value problem for the equations ( [ 1_nature])([3_nature ] ) in the infinite interval @xmath13 .
( this allows us to avoid the influence of boundary conditions and finite size of the interval on the dynamics . )
presuming that the initial velocity profile of movable particles is specified by the dependence @xmath14 where @xmath15 and @xmath16 are the arbitrary positive constants .
@xmath17 will be smooth and monotonic and we assume that the sign of the velocity derivative , @xmath18 , is retained .
second term in ( [ 4_nature ] ) describes some regular perturbation in the initial conditions .
the initial density profile of the movable component is assumed to be an even function @xmath19 which preserves a finite mass in the interval @xmath13 , i.e. @xmath20 moreover , we must supplement the above - given relations by the following conditions , arising from the symmetry of the problem at @xmath21 : @xmath22 now we transform the original system ( [ 1_nature])([3_nature ] ) to a simpler form . using the definition @xmath23 and condition ( [ symmetry_nature ] ) ,
we can conveniently cast ( [ 1_nature])([3_nature ] ) into the form @xmath24
the study of the system ( [ 6_col])([7_col ] ) is conveniently carried out in the lagrangian variables ( see , for example , @xcite ) : @xmath26 where @xmath27 satisfies the initial condition @xmath28 and provides @xmath29 this transformation holds true when the jacobian ( see , for example , @xcite ) @xmath30 besides , the condition of sign conservation of @xmath31 eliminates singularities of the flow .
+ under the transformation ( [ 1_sh ] ) the eqs .
( [ 6_col ] ) and ( [ 7_col ] ) with ( [ 4_col ] ) in hand turn into @xmath32 @xmath33 eq.([9_col ] ) , first of all , proves that @xmath34 is an invariant of the flow , which also holds true for @xmath35 when @xmath36 . to simplify ( [ 9_col])([10_col ] ) further we derive some additional relations being connected with the basic parameters of the problem .
differentiating the relation ( [ 4_col ] ) with respect to @xmath37 , switching thereby to the mass variable @xmath38 as a lagrangian variable instead of @xmath37 ( see , for example , @xcite ) , we get @xmath39 where we define the quantity @xmath40 as the specific volume ( see , for example , @xcite ) .
finally , after differentiating ( [ 5_col ] ) with respect to @xmath41 and bearing in mind ( [ u_nature ] ) we find @xmath42 relations ( [ 10_col ] ) and ( [ 12_col ] ) can be used to obtain a wave equation for the specific volume @xmath43 instead of the system ( [ 1_nature])([3_nature ] ) . differentiating the relation ( [ 10_col ] ) with respect to @xmath38 and taking into account ( [ 12_col ] ) , we obtain @xmath44 which should be supplemented by the initial conditions : @xmath45 @xmath46^{-1}{\partial u \over
\partial \xi}\right|_{\tau=0 } = v_0{\partial u_0 \over \partial \xi}\/. \label{dv0}\ ] ]
depending on the parameter @xmath10 , the equation ( [ 13_col ] ) has got different solutions which describe different physical situations .
indeed , in the case @xmath11 when in the system there is only a gravitational or electrostatic interaction , the general solution of the problem ( [ 13_col])-([dv0 ] ) is then finally provided by @xmath47^{-1}\/ \equiv { n_0(\xi ) \over j(\xi,\tau)}. \label{cold_col}\ ] ] we mention that the expression for @xmath48 can also be directly obtained from the solutions of eqs .
( [ 9_col ] ) and ( [ 10_col ] ) , namely @xmath49 and @xmath50 , and by utilizing the definitions of @xmath38 and @xmath51 . the basic solution of our problem in the @xmath52- representation is hence given by ( [ cold_col ] ) and @xmath53 @xmath54 that is the well - known solutions which have been more than once used to study the some astrophysical problems ( see , for example , refs .
the non - neutral plasma systems such as different the charged particle traps @xcite , or the charged particle beams @xcite-@xcite , or the electron transport in planar diodes @xcite are natural media which are described by these relations for @xmath55 .
as mentioned , for @xmath7 such type of solutions geometry describes the formation of finite - time singularities owing to the development gravitational instability . in three - dimensional geometry
these singularities form cellular structures being concentrated along the cell boundaries @xcite .
for @xmath55 it possible to escape the development of singular behavior and here we focus on such issue . in the case @xmath12 and @xmath55 one can have @xmath56^{-1}\/. \label{d_dav}\ ] ] the relation ( [ 5_col ] ) can be rewritten as @xmath57^{-1 } = { 1 \over n_0}{\partial x \over \partial \xi}=v\/.\ ] ] then from this relation and ( [ d_dav ] ) we find @xmath58 + u_0(\xi)\sin(\tau)\/. \label{x_dav}\ ] ] from which and ( [ u_nature ] ) it follows @xmath59\sin(\tau)\/. \label{udav}\ ] ] the solutions ( [ d_dav])-([udav ] ) have been studied in detail by dawson @xcite , kalman @xcite , davidson and schram @xcite to describe the dynamical features of nonlinear langmuir oscillations in the different limits of a cold plasma .
these findings should be helpful in identifying or interpreting cold - plasma expansion phenomena found in space and laboratory experiments @xcite - @xcite . in works
@xcite , the special case of initial velocity @xmath60 and the electron density @xmath61 ( in our notation ) was considered , the similar issue was studied for a periodic , step - like initial density distribution in @xcite .
it has been shown that this distribution @xmath62 propels the periodic wave structure of density in extremum points @xmath63 , where @xmath64 .
in the framework of initial - value problem , the initial velocity perturbations of special kind ( @xmath65 ) in spatially homogeneous plasma ( @xmath66 ) have been discussed by kalman @xcite .
as was established , this process can lead to the formation of standing density waves whose form is determined by the form of the initial velocity profile .
now based on the relations ( [ cold_col])-([u_sh ] ) and ( [ d_dav])-([udav ] ) , we show that for @xmath67 and @xmath68 it is possible the formation of density structures differing from the above - described variants .
moreover , we shall show that the structure formation in the cases @xmath11 and @xmath12 occurs in different way . according to @xcite , the dependence of the density on the coordinate @xmath69 for any moment of time @xmath70 is qualitatively similar to the dependence @xmath71 in lagrangian coordinates on @xmath37 at any moment of time @xmath72 .
therefore , all the information on the characteristics of the density distribution is contained in ( [ cold_col ] ) and ( [ d_dav ] ) or in theirs derivatives @xmath73 .
expression ( [ cold_col ] ) and ( [ d_dav ] ) can then be used to obtain @xmath74\/ , \label{d_nature}\ ] ] where @xmath75 and functions @xmath76 and @xmath77 are defined as @xmath78 it is easy to see that @xmath79 and @xmath80 under @xmath81 .
as is seen from ( [ d_nature ] ) for @xmath82 and @xmath83 , the function @xmath84 reaches the local extremum with respect to @xmath37 at time @xmath85 in points where @xmath86 and points @xmath87 being the roots of the algebraic equation @xmath88 where @xmath41 must be treated as a parameter .
as is seen , the number of points of such type depends on the form of functions @xmath89 and @xmath77 and this number can be varied with respect to time .
this suggests that under some circumstances we can get a wavelet structure from the initially monotonic density profile and then the one can be destroyed when the eq .
( [ 7_nature ] ) has no solution .
we start from case @xmath11 which refers to the systems with a pure electrostatic or gravitational interactions .
the case of the force - free systems must be considered as a limit of these systems . in
order reveal the structure of @xmath89 we substitute ( [ 4_nature ] ) and ( [ 5_nature ] ) into ( [ f_nature ] ) and rewrite this result in the following form @xmath90 where @xmath91 and @xmath92\/. \label{nu_nature}\ ] ] it is easy to see from these relations that eq .
( [ 7_nature ] ) can have a lot of roots for monotonic initial distributions @xmath93 and @xmath94 when the second derivative of the velocity is non monotonic
. in this case the form of @xmath89 may be essentially defined by the perturbative part of the velocity via function @xmath95 having a wavelet structure , i.e. it consists of local minima and maxima . noting that @xmath95 reaches its local extrema in the points @xmath87 being the roots of equation @xmath96 , e.i .
@xmath97 however , the manifestation of @xmath95 in @xmath98 completely depends on the magnitude and sign of the function @xmath99 . thus , the possible way of evolution for the system is determined by the relation between the functions @xmath99 and @xmath95 .
in @xcite we applied such approach to study the formation of wavelet structures under the influence of the perturbative part of the initial velocity distribution when the function @xmath89 can have both negative and positive values . here
we consider the restricted case when the governing function @xmath100 , and discuss the physical consequences of this restriction .
this regime is of interest since it admits both clustering and dissolution of the pattern arising from smooth initial distributions .
so we proceed to find out the initial conditions when such regimes can exist .
first of all , we shall find out the conditions for the initial data under which the singular behavior has been eliminated . in the present case the jacobian takes the form : @xmath101 \tau + { \delta n_0\over 2}\tau^2\/. \label{1_jacob}\ ] ] from the condition @xmath102 we get the quadratic equation @xmath103 \tau + 1 = 0\/,\ ] ] whose positive solutions define the time @xmath104 when the first collapse may appear .
the roots of this quadratic equation are determined by @xmath105 where @xmath106 as seen from ( [ qua ] ) it is impossible to escape from the formation of singularities when we have a self - gravitating fluid , @xmath107 .
therefore , we focus our attention on the case @xmath108 where we have no positive roots of quadratic equation when @xmath109 . this condition is always fulfilled for @xmath110 in order to study the influence of the initial conditions on @xmath89 now we have to define the form of @xmath93 and @xmath111 which make it possible to avoid a singular dynamics and provide a wavelet structure for positive @xmath89 .
as is seen from ( [ nu_nature ] ) in the case @xmath112 where @xmath113 is some positive constant , it holds @xmath114 for any @xmath37 .
so , taking into account the definitions ( [ g_nature ] ) and ( [ nu_nature ] ) we conclude that the condition @xmath115 is satisfied provided that there exists such a function @xmath111 for which the inequality @xmath116 holds for any @xmath37 . here
we restrict ourselves to the simplest case when the inequality ( [ 3_n0a ] ) turns into equation @xmath117 which has solution @xmath118 where @xmath119 is an arbitrary constant .
this relation determines an admissible type of @xmath120 which is defined by initial profile @xmath62 . inserting ( [ 5_n0a ] ) into ( [ kb ] )
, we obtain the relation @xmath121 which together with eq.([3_n0a ] ) impose a restriction on the possible form of initial distribution @xmath62 . combining these expressions we get @xmath122 for @xmath123 having physical meaning the condition ( [ 7_n0a ] ) is reduced to @xmath124 integrating this relation we obtain @xmath125 \leq n_0(\xi ) \leq \exp\left[- { \xi \over m}\right]\/. \label{8_n0a}\ ] ] it is also worth emphasizing that this relation has a meaning only for @xmath126 thus , the relations ( [ 5_n0a ] ) , ( [ 8_n0a ] ) and ( [ 9_n0a ] ) define some non - empty set of the governing parameters @xmath16 , @xmath15 and @xmath113 when nonsingular dynamics occurs and the positive governing function @xmath89 has a wavelet form .
an explicit example in the 3d parameter space @xmath127 for which ( [ 9_n0a ] ) is valid is given by @xmath128 .
this leads us to the principal features in the dynamics in contrast to @xcite where @xmath89 can change its sign . for definiteness
, we display such dynamics with the initial profile of form @xmath129 where constant @xmath130 satisfies all necessary conditions , noting that for an initial gaussian density profile such a process of wave let creation and dissolution would be absent . fig .
[ fig_1 ] depicts the qualitative dependence of @xmath89 which may be typical for profile of type ( [ 10_n0a ] ) .
in this situation the formation of wavelet structure on the density profile in the course of time occurs similar to the case studied in @xcite for small times when the initial smooth and monotonic distribution ( [ 10_n0a ] ) is transformed into density profile with a large number of local extrema as sketched in fig .
[ fig_2 ] for @xmath131 .
such script scenario will be implemented until the eq .
( [ 7_nature ] ) will have a real root .
however , it is clear that owing to @xmath132 such behavior can not last indefinitely .
there is a moment of time , say @xmath133 when we lose one root of the eq .
( [ 7_nature ] ) , and consequently one hump in the dependence of density profile .
it means that the wavelet structure disappears in the course of time . for this @xmath98 there
always exists a moment of time , say @xmath134 when the eq .
( [ 7_nature ] ) has no real root anymore and the wavelet structure of density dependence is transformed into a distribution which is similar to the initial distribution .
thus , for the given initial conditions the wavelet structure can exist only for some finite time .
the physical reason for this unexpected behavior may be found in the temporal change of the strength of the nonlinearity , which gives rise to a steepening and structure formation in the earlier stage when it is strong , but later , when it becomes weakened by the expansion process , it is no longer in the position to sustain the structure , which then dissolves . in the case presented in figs .
[ fig_1 ] and [ fig_2 ] such a structure exists only in the interval @xmath135 . for times
@xmath136 the process may be again repeated in general . however , it is impossible to replicate this scenario identically since the new velocity distribution will differ from the initial distribution @xmath137 .
this property distinguishes the cases @xmath108 from the case @xmath107 in which the formation of singularities eliminates such behavior . in this connection , let us dwell on the time reversibility of the system ( [ 1_nature])([3_nature ] ) .
it is easy to see that the eqs .
( [ 1_nature])([3_nature ] ) remain invariant under the transformation @xmath138 where @xmath139 and @xmath140 are some parameters of transformation .
now let the time change the direction , i.e. we have to put @xmath141 in these relations . as is seen from ( [ sym_nature ] )
this leads only to @xmath142 .
it means that ( [ 1_nature])([3_nature ] ) is a reversible system and generally we can expect to get the initial space distribution of density and velocity ; at least such a possibility is not prohibited .
however , as was shown above , in the present case the destruction of space structure is caused by other reasons . though the same result follows directly from eq.([13_col ] ) , here we use the euler frame as physically more intuitive one .
now we pass to the plasma structures formation by considering the vlasov - poisson system with parameters @xmath55 and @xmath12 . following to @xcite we put @xmath143 then eq . ( [ d_nature ] )
is reduced to @xmath144 ^ 2}\/. \label{ds_n}\ ] ] as is seen from this relations , for @xmath145 and @xmath146 the number points of exremum coincides with ones of the initial profile @xmath62 as this was established in @xcite and no additional structure may appear in this case .
assume that @xmath147 as this was done in @xcite . from ( [ d_nature ] ) it easy to see that in this case there are no conditions for the development of the wave breaking .
then the relation ( [ d_nature ] ) is transformed into @xmath148 ^ 2 } u_0^{\prime \prime}\/ , \label{k_n}\ ] ] from which it follows that the formation of new extrema is determined by @xmath149 .
namely there is an accumulation of particles in the regions where @xmath150 and there is a diminishing density in the regions @xmath151 .
also , at the time moments @xmath152 we have @xmath153 this type of exteremum is perfectly caused by the electrostatic electron oscillations of a plasma medium .
we now get an oscillatory time behavior with repeated sign reversals of @xmath154 .
the temporal evolution is no longer an uni - directional expansion with increasing out - flow velocities , as in the previous cases , but it is now characterized by a time - periodic behavior , where out- and in - flows alternate each other .
this is physically expected since in the presence of an ion background the force influencing on the electrons changes periodically its direction .
this is due to a density variation in euler space as it holds @xmath155 the frequency of this oscillation is , of course , the plasma frequency , given by @xmath156 in our normalization .
thus , at moments @xmath152 the spatial structure is destroyed and the initial , uniform density distribution emerges and then the process repeats in some local region of phase space [ see eq .
( [ x_dav ] ) and ( [ udav ] ) ] . in general case
when @xmath157 and @xmath158 const .
the formation of density structure comes about by the similar way .
the dependence of @xmath159 is controlled by the behavior of the function @xmath89 via eq .
( [ d_nature ] ) with @xmath160 .
in contrast to the case @xmath11 here we have the periodic , unlimited function of time .
this leads to the formation and disappearance of density profile will be repeated with the period of the function @xmath161 . since function @xmath161 takes both positive and negative values therefore the eq .
( [ d_nature ] ) always has roots for any time @xmath41 for which it holds @xmath162 and the formation of density structure can occur only for these times .
outside the time intervals defined by relation ( [ time_nature ] ) there is a restoration of the initial profile . ) for @xmath132 [ fig_1 ] ] [ fig_2 ] ]
in this paper we have considered the influence of a initial velocity on the dynamics of one - dimensional euler - poisson system . for the coulomb systems ( @xmath7 ) and the force - free systems ( @xmath9 ) it was shown that the wavelet - like density structure can appear and disappear for some types of initial conditions . in the present case the initial conditions play the role of a driving parameter which determines the formation and disappearance of such structures .
our consideration was limited to the case of regular perturbations in the initial distribution of velocity [ see eq . ( [ 4_nature ] ) ] . for this class of initial conditions
we have shown how such space - time structures emerge and disappear during the evolution .
however , the assumption about regular perturbations is not necessary ; in the general case the perturbations may be stochastic , too .
the distribution of density will look like random distribution , although this structure has been born by a deterministic process and our conclusions about the finite time of pseudo - chaotic behavior remain valid . taking into account the high sensitivity of our pattern to perturbations in the initial conditions , one can come to the conclusion that there exists some interval in the system parameters [ see eqs .
( [ 5_n0a ] ) , ( [ 8_n0a ] ) and ( [ 9_n0a ] ) ] in which it is possible to see the effect described in reality . it should be noted that in the framework of the model the nonlinear effects ( such as a density collapse or the temporal emergence and disappearance of wavelets ) are , as a consequence , hence less pronounced since for its development there is not enough time available within one period of a plasma oscillation . nevertheless , we believe that our present approach can be generalized , for instance , to an inhomogeneous , many - component plasmas in fluid description @xcite . further generalizations of the model can be realized by some modifying the basis solution ( [ cold_col])-([x_sh ] ) .
such generalizations can also be used as a starting point for investigation of more complex and more realistic problems @xcite-@xcite . in conclusion
, we again would like to emphasize that the one - dimensional model , the transient nonlinear dynamics of which we have analyzed here rigorously on mathematical grounds , is still rather artificial because real systems are essentially higher dimensional and , in the absence of dissipation , kinetic .
however , from the presented results we may catch a glimpse of the formation of time - dependent structures in higher dimensions governed by the hydrodynamic nonlinearity as a reference to a more realistic dynamical evolution , which generally involves particle trapping processes @xcite .
99 bernstein j 2004 _ kinetic theory in the expanding universe _
( cambridge : cambridge university press ) biskamp d 1993 _ nonlinear magnetohydrodynamics _ ( cambridge : cambridge university press ) kadanoff l p 1999 _ from order to chaos ii , essays : critical chaotic and otherwise _ ( hong kong : world scientific ) shandarin s f and zeldovich ya b 1989 _ rev .
_ * 61 * 185 eliasson b and shukla p k 2006 _ phys .
* 422 * 225 sack ch and schamel h 1987 _ phys .
rep . _ * 156 * 311 dubin d h e and o@xmath163neil t m 1999 _ rev .
phys . _ * 71 * 20 arnold v i 1992 _ catastrophe theory _
( berlin : springer ) eggers j and smith a f 2010 _ j. fluid mech . _ * 647 * 187 fouxon i , meerson b , assaf m and livne e 2007 _ phys .
fluids _ * 19 * 093303 gibbon j d 2008 _ physica d _ * 237 * 1894 yinbin deng , jianlin xiang and tong yang 2003 _ j. math .
* 295 sack ch and schamel h 1985 _ plasma phys . contr . fusion _
* 27 * 717 karimov a r 1997 _ physica d _ * 102 * 328 karimov a r and schamel h 2001 _ phys . plasmas _ * 8 * 1180 stenflo l and yu m y 1996 _ nature _ * 384 * 224 schamel h 2004 _ phys . rep . _ * 392 * 279 burgers j m 1929 _ proc .
soc . _ * 32 * 414 davidson r c 1972 _ methods in nonlinear plasma theory _ ( new york : academic press ) jin d z and dubin d h e 2000 _ phys .
* 84 * 1443 anderegg f 2013 in _ physics with trapped charged particles : lectures from the les houches winter school _ ( london : imperial college press ) anderson o a 1987 _ part .
* 21 * 197 struckmeier j 1996 _ phys .
e _ * 54 * 830 struckmeier j 2000 _ phys . rev .
beams _ * 3 * 034202 reiser m 2008 _ theory and design of charged particle beams _
( weinheim : wiley - vch verlag gmbh & co. kgaa ) dawson j 1959 _ phys .
rev . _ * 113 * 383 kalman g 1960 _ ann .
* 1 davidson r c and schram p.p.j.m . 1968 _ nucl . fusion _ * 8 * 197 leroy m m 1983 _ phys
. fluids _ * 26 * 2742 fletcher r s , zhang x l , and rolston s l 2006 _ phys . rev .
lett . _ * 96 * 105003 osipyan d a , nersisyan h b and matevosyan h h 2003 _ astrophysics _ * 46 * 434 winske d and omidi n 2005 _ phys . plasmas _
, * 12 * 072514 karimov a r , schamel h and shcheglov v a 2000 _ phys .
lett a _ * 272 * 193 karimov a r 2002 _ phys .
_ * 65 * , 356 ( 2002 ) .
stenflo l and gradov o m 1998 _ phys .
e _ * 58 * 8044 infeld e and rowlands g 1987 _ phys
* 58 * 2063 rowlands g , brodin g and stenflo l 2008 _ j. plasma phys . _ * 74 * 569 verma p s , soni j k , segupta s and raw p k 2008 _ phys .
plasmas _ * 17 * 0445031 stenflo l 1996 _ phys .
_ * t63 * 59 schamel h 2012 _ phys .
plasmas _ * 19 * 020501 schamel h 2015 _ phys
. plasmas _ * 22 * 042301 | the lagrangian fluid description is employed to solve the initial value problem for one - dimensional , compressible fluid flows represented by the euler - poisson system .
exact nonlinear and time - dependent solutions are obtained , which exhibit a variety of transient phenomena such as a density collapse in finite space - time or the appearance and , for the first time , a successive dissolution of wavelet structures during the same event .
the latter are superimposed on the gross density pattern in the course of the uni - directional expansion of an initially localized density hump .
whereas self - gravitating fluids will always experience collapse , neutral fluids and fluids with repulsive forces , such as a non - neutral , pure electron fluid , can exhibit an evolution of the second type , being determined by the initial conditions . for an electron fluid ,
being embedded in a neutralizing ion background , these nonlinearities are , however , strongly diminished due to the omnipresent plasma oscillations , which weaken and alternatively change the sign of the collective force . + |
radiation transfer plays a critical role in star formation and affects the structure of accretion flow and the resulting adiabatic cores even in a low - mass regime
. however multi - dimensional radiation hydrodynamics ( rhd ) simulations have been rarely performed due to their high computational cost .
therefore , the barotropic approximation , which omits radiation transfer and simplifies the thermal evolution of the gas , is widely used in multi - dimensional simulations .
however , recent advancement in the computer technology and development of numerical techniques enable us to incorporate radiation transfer into a multi - dimensional magnetohydrodynamics ( mhd ) simulation within reasonable computational time , using a moment method with simplified closure relations .
we performed rmhd simulations of proto - stellar collapse with our newly developed numerical code and clarified the difference between rmhd and mhd with the barotropic approximation .
rhd and rmhd simulation codes are being actively developed and some of them are used to study star formation processes .
for example , @xcite solved radiation hydrodynamics by using the flux limited diffusion ( fld ) approximation on 2d nested - grids .
@xcite performed smoothed particle hydrodynamics simulations with fld radiation transfer and showed different thermal evolution between models with and without radiation transfer .
orion amr ( adaptive mesh refinement ) rhd code developed by @xcite is used to study the main accretion phase of low - mass star formation @xcite and high - mass star formation @xcite . recently
amr rmhd simulations based on ramses were reported by @xcite , who studied the influence of radiation on fragmentation .
magnetic fields and self - gravity are also the key physical processes in star formation because they assume leading roles in the transfer of angular momentum which dominates the global evolution of a cloud .
magnetic braking and outflow driven by magnetic fields efficiently carry the initial angular momentum away from the cloud @xcite .
if the cloud rotates very fast , it will fragment through self - gravitational instability @xcite . since the structure and fate of the core strongly depend on its rotation and the efficiency of angular momentum transport , 3d mhd simulations with self - gravity are required . molecular outflows from young stellar objects are universally observed with millimeter / submillimeter radio instruments ( see @xcite and references therein ) . in those observations some of outflows show complicated structure which can be interpreted as consisting of multiple components @xcite .
they are strong evidences that magnetic fields play a crucial role in angular momentum transport during protostellar collapse .
the argument about the driving mechanism of molecular outflow is not yet settled .
we propose that the molecular outflow is driven via magnetic fields in the first core @xcite ; however , some groups believe that the outflow is entrained by a high velocity jet from the central protostar @xcite .
next generation radio interferometric telescopes like the atacama large millimeter / submillimeter array ( alma ) have the potential to reveal the on - going star formation sites directly and contribute to our understanding about overall star formation processes and such important physical mechanisms . in this letter , we report the results of a low - mass star formation simulation obtained with our newly developed 3d nested - grid self - gravitational rmhd code .
we performed very deep nested - grid simulation of proto - stellar collapse just before the second collapse . in our simulation , a two - component outflow is driven from the first core . in this letter , we concentrate on disscussion about these core and outflow including comparison between the rmhd simulation and the barotropic approximation while another paper is under preparation .
the structure of this letter is as follows : in 2 we describe our numerical method and model setup .
we show numerical results in 3 , and 4 is devoted to discussions and conclusions .
we solve 3d self - gravitational rmhd equations on nested - grids . equations and quantities for the radiation
are defined in the comoving frame of the fluid .
we treat only the zeroth order moment equations of radiation transfer using the gray ( i.e. , equations are integrated over frequency assuming a blackbody spectrum of a local temperature ) fld approximation proposed by @xcite with @xcite closure relations .
we assume ideal mhd with no resistivity .
the equation of state ( eos ) of the gas is also assumed to be ideal , where the gas mainly consists of hydrogen molecules and the adiabatic index @xmath0 is set to @xmath1 throughout the simulation , which is better to trace the realistic thermal evolution than @xmath2 @xcite .
the gas and the dust are assumed to have the common local temperature .
here we do not describe the implementation of our code in detail but mention some points about the treatment of radiation transfer . for the mhd and self - gravity parts ,
see @xcite . our method for updating radiation and
gas energy is similar to zeus - mp code @xcite , although our operator - splitting and treatment of the mhd and radiation parts are different from theirs .
while our mhd and self - gravitaty parts are advanced explicitly with a second - order scheme , the radiation part is advanced implicitly with the first - order backward euler method for stable time - integration .
newton - raphson iteration is used to solve the resulting non - linear equations .
the sparse matrix appearing here is very large and not diagonally dominant because of the strong non - linearity of the system ; hence , a fast and robust iterative solver is required .
we use a stabilized biconjugate gradient ( bicgstab ) solver with a multi - color modified incomplete lu decomposition ( ilu(0 ) ) preconditioner , which is stable and efficient .
we determine the timestep by the courant - friedrich - levy ( cfl ) criterion derived for the mhd part .
all the grids have this common timestep and are advanced synchronously although individual ( or asynchronous ) timesteps are generally used in amr simulations . in usual mhd simulations using individual timesteps , boundary values
are constructed by time - interpolation of the values in the coarser level . however , if we use an implicit time - integrator and far larger timestep than that determined by the cfl condition for radiation transfer ( i.e. , @xmath3 ) , then this time - interpolation is not adequate , at least in principle .
we use a compiled table of opacity adopted from @xcite and @xcite .
we use @xcite for the rosseland mean opacity , and we smoothly combine the two for the planck mean .
that is , we use @xcite in high temperature ( @xmath4 ) and @xcite in low temperature because the planck mean opacity of @xcite is quite lower than other opacity tables in high temperature @xcite .
we adopt the surface formula with the limiter proposed by @xcite to evaluate radiation energy flux at the cell interface .
we take a bonnor - ebert sphere of 10k with central density @xmath5 for the initial condition .
initial rotation and magnetic fields are given uniformly along _
z_-axis , @xmath6 and @xmath7 . the outermost boundary values are fixed to keep their initial values and no geometric symmetry is assumed .
the number of grid points in each level of nested - grids is @xmath8 .
the size of the finer grid is half of the coarser grid , and the finer grid is placed around the center of the simulation box self - similarly .
the simulation starts with 5 levels of nested - grids , and finer grids are generated adaptively to resolve the local jeans length with 32 meshes not to induce artificial fragmentation @xcite .
this is also because insufficient resolution causes the gas entropy to be overestimated near the center of the core . at the end of the simulation
18 levels of nested - grids are generated and the finest resolution is @xmath9 .
we stop our simulation when the gas temperature of the central region reaches 2000k at which temperature hydrogen molecule starts to dissociate and the second collapse begins .
figure [ f1 ] is a typical 3d birds - eye view of the first adiabatic core and the outflow @xmath10 after the first core formation .
there coexist two components of outflow : well collimated fast outflow ( associated with red magnetic field lines in figure [ f1 ] ) and slow outflow with a large opening angle ( associated with yellow magnetic field lines ) .
the former is driven by magnetic pressure and the latter by magneto - centrifugal force @xcite .
@xcite shows that the magnetic pressure mode typically appears in case of weak magnetic fields , while the magneto - centrifugal mechanism appears in case of relatively strong magnetic fields .
vertical slices of the outflow scale ( level @xmath11 ; upper row ) and core scale ( level @xmath12 ; lower row ) are shown in figure [ f2 ] .
hereafter we discuss only vertical slices since the process occurs nearly axisymmetric in this case .
although the density distribution shows the complicated structure as a result of mhd processes , the temperature distribution is almost spherically symmetric .
the gas is heated up by the radiation before it enters the shock of the first core .
this means that the temperature distribution in the outer region is dominated by the radiation from the central hot region .
shock heating up to @xmath13 at the edge of outflow is observed .
this picture is considerably different from that of the barotropic approximation used in previous studies , in which the temperature is determined only by the local gas density .
the difference between the rmhd simulation and the barotropic approximation is clearly seen in the right column ( c ) in figure [ f2 ] where the ratio of the gas temperature obtained in the rmhd simulation to the barotropic temperature given by the local gas density is plotted . in the rmhd simulation
, the gas tends to attain the temperature typically 2 - 3 times higher than that in barotropic approximation around @xmath14 from the center , although the two models are very close in the innermost region of the core ( @xmath15 ) .
the difference is most striking just above the shock at the surface of the first core because of pre - shock heating by radiation from the core .
here we use the following polytropic relation to evaluate the barotropic temperature : @xmath16^{\gamma-1}\hspace{1em } { \rm k , } \ ] ] where @xmath17 is the critical density and @xmath18 is the adiabatic index of diatomic molecular gas . the parameters used here are chosen to trace the thermal evolution track of the central region in a spherical rhd simulation by @xcite .
we visualize the outgoing mass flux and angular momentum flux defined as @xmath19 and @xmath20 where @xmath21 respectively where @xmath22 is poloidal ( toroidal ) velocity at the early driving phase ( @xmath23 after the first core formation ) of outflow in figure [ f3 ] . in the figure we can recognize two components of outflow driven by different mechanisms : the outer magneto - centrifugal mode and the inner magnetic pressure mode .
the inner magnetic pressure mode has a relatively high velocity @xmath24 and is well collimated .
the front of this outflow gets hot through shock heating due to its high mach number .
however , the outer , slow outflow has the mass and angular momentum flux , one order of magnitude larger , since the gas in the inner region has a relatively small angular momentum .
this trend is enhanced as the outflow evolves .
accordingly we can say that the outer magneto - centrifugal mode is more important for the kinematic structure of the accretion flow as a carrier of angular momentum ; however , the fast , inner outflow will stand out more for observation because it travels a long distance and achieves a high temperature .
we can see a more detailed structure of the collapsing cloud from figure [ f4 ] where we plot the distribution of physical quantities at the end of the simulation , just before the second collapse starts ( @xmath25 after the first core formation ) .
panels ( a)-(c ) show the distribution of physical quantities in the disk midplane and along the rotational axis . at this time the fast outflow reaches @xmath26 .
the surface of the core is located at @xmath27 in the disk midplane and @xmath28 along the rotational axis .
this height of the first core is larger than that with the barotropic approximation , which is typically @xmath29 @xcite , because the outer region of the first core achieves a higher entropy in rmhd simulation than in the barotropic approximation .
the core with a higher entropy can support more mass even with the same central density ; therefore , the first core lives longer in rmhd under the same accretion rate , which is mostly determined by the initial condition . along the _
z_-axis there is a high entropy radiative precursor outside the core ( @xmath30 ) seen in ( b ) and ( c ) .
the central region within @xmath31 from the center seems nearly spherically symmetric due to the slow rotation because of efficient magnetic braking .
the structure of the disk and the behavior of angular momentum transport is visible in panel ( d ) of figure [ f4 ] and figure [ f2 ] .
the profiles of infalling velocity @xmath32 and rotational velocity @xmath33 are plotted in panel ( d ) of figure [ f4 ] .
the infalling gas initially decelerates at the weak shock near @xmath34 associated with the first core .
there exist two centrifugal barriers , and the outer one near @xmath35 corresponds to the driving region of the slow and wide outflow driven via the magneto - centrifugal force .
this outflow carries off angular momentum of the gas efficiently and the gas falls radially . nonetheless in this case
the initial magnetic fields are not so strong , the gas spins up again as it falls and hits the inner centrifugal barrier near @xmath36 . then , the magnetic field lines are wound up tightly ( also due to weak initial magnetic fields ) and magnetic pressure launches the fast , well collimated outflow .
the variety of the gas s thermal behavior is shown in the @xmath37 plane in figure [ f5 ] .
the distribution in this plane indicates a strong dependence on the geometry .
we can see again that the gas tend to be hotter in the rmhd simulation than in the barotropic approximation .
although this was already pointed out by @xcite , our simulation can resolve more detailed features .
the jump around @xmath38 is caused by the shock at the surface of the first core .
this density is very close to the critical density in the barotropic approximation .
this shock is nearly isothermal since radiation can transfer thermal energy produced at the shock to the upstream efficiently .
radiation transfer is not so efficient beyond this critical density and this forms the first adiabatic core .
when the temperature reaches @xmath39 the evolution track in the @xmath37 plane gets shallow . this is because all the components of the dust evaporate and the opacity drops sharply at this temperature , and then radiation transfer becomes effective again .
the structure of the first core is slightly modified by this effect but not drastically since the gas soon becomes optically thick and adiabatic again .
we emphasize that radiation transfer is required to handle the realistic thermal evolution described here .
we performed numerical simulations of the early phase of the low - mass star formation process just before the second collapse starts .
our simulation used a newly developed three - dimensional nested - grid self - gravitational fld rmhd simulation code without assuming artificial thermal evolution . in our case
, radiation transfer does not seem to change the qualitative scenario of low - mass star formation drastically .
however , the temperature distribution is significantly changed by introducing radiation transfer . realistic treatment of gas thermodynamics alters some properties and structure of the core quantitatively .
the mass and size of the first core at a certain central density are larger because of a higher entropy , and the lifetime of the first core becomes slightly longer .
furthermore the envelope and the outer region of the first core become hotter .
we suggest that the observational probability of such very young star forming sites rises when compared to previous predictions based on the simulations with the barotropic approximation . barotropic approximation fails to reproduce realistic thermal evolution ; needless to say radiation transfer plays a critical role there even in the low - mass regime . in the barotropic approximation
all the gas elements trace the thermal evolution track of the gas at the center of the cloud which experiences no shock and , therefore , has minimum entropy in the spherically symmetric simulation . in this sense barotropic approximation tends to underestimate the gas entropy and temperature . in 3d , the rmhd simulation has more striking and complex differences from the barotropic simulation . on the other hand
if there exist fast initial rotation and ineffective angular momentum transfer , then the entropy of the resulting thin disk - like first core may be lower than in the barotropic eos .
thus radiation transfer will affect the mechanical properties such as stability or fragmentation through the thermal property of the gas .
determining a realistic temperature distribution is essential for predicting or interpreting the properties acquired in both the molecular emission lines and thermal continuum observations , for strong temperature dependences of the emissivity .
chemical reactions are also sensitive to temperature , so our results can be applied to a study of chemistry in molecular clouds and star forming regions . in our case
, unlike previous mhd or rmhd simulations , there coexist two components of outflow from the first core .
this feature suggests that the angular momentum transfer may occur necessarily not in single but in multiple stages , at least under a certain condition of magnetic fields although the situation depends on the initial condition and will be changed when magnetic dissipation is incorporated .
nonetheless our results imply a rich diversity of the outflow and first core .
though on - going star formation like this simulation has never been observed directly , new telescopes are expected to reveal them , and alma seems to be the most promising in the near future .
we propose some vellos ( very low luminosity objects ; @xcite ) are candidates for such obscured star forming cores .
compact ( @xmath40 scale ) and warm ( @xmath41 ) molecular outflow can also be a good indicator of the first core .
our results can be compared with future observation .
+ we thank prof .
shu - ichiro inutsuka , dr . kazuyuki omukai and dr .
takashi hosokawa for fruitful discussions .
numerical computations were performed on nec sx-9 at center for computational astrophysics , cfca , of national astronomical observatory of japan .
this work is supported in part by the ministry of education , culture , sports , science and technology ( mext ) , grant - in - aid for scientific research , 21244021 , 16204012 ( tomisaka ) , 20540238 ( tm ) , 20740115 ( ko ) and 21740136 ( mm ) .
k. tomida is supported by the research fellowship from the japan society for the promotion of science ( jsps ) for young scientists . | we perform a three - dimensional nested - grid radiation magneto - hydrodynamics ( rmhd ) simulation with self - gravity to study the early phase of the low - mass star formation process from a rotating molecular cloud core to a first adiabatic core just before the second collapse begins .
radiation transfer is handled with the flux - limited diffusion approximation , operator - splitting and implicit time - integrator . in the rmhd simulation ,
the outer region of the first core attains a higher entropy and the size of first core is larger than that in the magnetohydrodynamics simulations with the barotropic approximation .
bipolar molecular outflow consisting of two components is driven by magnetic lorentz force via different mechanisms , and shock heating by the outflow is observed . using the rmhd simulation we can predict and interpret the observed properties of star - forming clouds , first cores and outflows with millimeter / submillimeter radio interferometers , especially the atacama large millimeter / submillimeter array ( alma ) . |
there were 20 experimental presentations at this conference , all by expert people in the field .
all these talks were plenary , so there is no point of me summarizing again the contents of each talk .
instead , i would like to present a personal view and touch upon few selected issues which were presented here .
in spite of the fact that diffractive processes have a long history , it is still not easy to give a precise and concise definition of what one calls a diffractive reaction . it is clearly a process where , in an exchange picture , no color is exchanged .
this however includes all the colorless particles .
a further requirement is that the exchanged messenger has the quantum number of the vacuum , which , in the regge picture is named the pomeron .
elastic scattering is usually presented as an example of a diffractive process . however , at low energies it can proceed also through an exchange with quantum numbers different from the vacuum , called a reggeon . at high energies , where the kinematics allows for a large rapidity range , it is more useful to talk about rapidity gaps . due to the fact that no color is exchanged ,
there is a suppression of gluon radiation and therefore a rapidity gap is produced between the two vertices of the interaction .
of course , both the pomeron and the reggeon are colorless , however the rapidity gap produced by a reggeon is exponentially suppressed , while that of the pomeron remains constant .
this prompted bjorken @xcite to define diffraction as processes in which the large rapidity gap is not exponentially suppressed .
how practical is this definition ?
this we shall see in the following sections .
i will try to define here all the variables needed in the following sections .
let us first look at a diffractive reaction @xmath0 at hera , depicted in figure [ fig : ddis ] .
a virtual photon @xmath1 ( @xmath2 ) interacts with the proton through a colorless exchange with vacuum quantum numbers .
a mass @xmath3 of the hadronic system recoils against the proton .
the square of the momentum transfer at the proton vertex is @xmath4 .
the fraction of the proton momentum carried by the exchange is denoted by @xmath5 .
the quark struck by the virtual photon carries a fraction @xmath6 of the momentum of the colorless exchange .
( note that sometimes @xmath7 is used instead of @xmath6 ) .
the last two variables are connected to bjorken @xmath8 as follows : @xmath9 . at the tevatron , where one studies the diffractive reaction @xmath10 , the equivalent variable to @xmath5 is denoted by @xmath11 .
there are three methods used at hera to select diffractive events .
one @xcite uses the leading proton spectrometer ( lps ) to detect the scattered proton and by choosing the kinematic region where the scattered proton looses very little of its initial longitudinal energy , it ensures that the event was diffractive .
a second method @xcite simply request a large rapidity gap ( lrg ) in the event and fits the data to contributions coming from pomeron and reggeon exchange . the third method @xcite uses the distribution of the mass of the hadronic system seen in the detector , @xmath3 , to isolate diffractive events
. we will refer to these three as lps , lrg and @xmath3 methods . at the fermilab tevatron
@xcite diffractive processes are being studied by tagging events with either a rapidity gap or a leading hadron .
the lps method has the advantage of detecting the scattered proton and thus excluding proton dissociative processes .
however , in order to ensure that the scattered proton resulted from a diffractive process , one requires @xmath12 0.01 , where @xmath5 is the amount of longitudinal momentum lost by the scattered proton .
this cut removes contributions coming from reggeon exchanges @xcite .
the lrg method selects events which also include some proton dissociative processes and reggeon contributions .
the latter can be removed by the same @xmath12 0.01 cut as above .
the proton dissociative processes can be removed provided their mass is large enough to produce signals in some forward tagging devices . the contribution of low mass proton dissociation can be estimated . in the analysis of the h1 collaboration ,
processes with proton dissociation into masses below 1.6 gev amount to about 10% @xcite . the @xmath3 method which subtracts the exponentially suppressed large rapidity gap events , in principle subtracts also the reggeon contribution and
is left only with the proton dissociative background .
these can not be removed for masses below 2.4 gev , which constitutes about 30% of the selected diffractive events @xcite .
at the tevatron , single diffractive events , @xmath13 , are selected by tagging the scattered @xmath14 . in figure
[ fig : allsf ] the diffractive structure function measurements with all three hera methods are presented .
the @xmath3 data have been multiplied by a factor of 0.69 to correct for the proton dissociation background .
no correction was done to the h1 data .
all methods seem in general to agree with each other in their overlapping kinematic region . measured by h1 and zeus , as a function of @xmath5 in overlapping bins of @xmath6 and @xmath15 . ] the @xmath8 behaviour of the inclusive structure function @xmath17 at a given @xmath15 is well described by an @xmath18 form .
the value of @xmath16 is connected to the pomeron intercept , @xmath19 .
the value of @xmath16 is approximately constant till @xmath20 1 gev@xmath21 , and then rises almost linearly with @xmath22 .
it is of interest to see whether the @xmath5 behaviour of @xmath23 shows a similar pattern .
to this end , a fit of the form @xmath24 was performed , for different @xmath15 intervals .
figure [ fig : lama ] shows the value of @xmath16 as function of @xmath15 from fits to the @xmath8 behaviour of @xmath17 and from the @xmath5 behaviour of @xmath23 .
the precision data of @xmath17 makes it possible to see a very significant rise with @xmath15 .
the @xmath23 data does not have the precision needed for a clear @xmath15 dependence .
there is a trend similar to that of @xmath17 , but given the large errors of the data , the behaviour is also consistent with no @xmath15 dependence .
dependence of @xmath16 obtained from fits to @xmath25 and @xmath24 .
the band indicates the value corresponding to the soft pomeron . ]
it has been proven @xcite that qcd factorization works for diffractive processes at hera .
this allows to use the dglap @xcite evolution equations to get diffractive parton distribution functions .
given the fact that for describing diffractive processes one needs more variables , @xmath4 , @xmath5 , @xmath6 , @xmath15 , one would actually like to evolve in @xmath6 and @xmath15 for fixed @xmath4 and @xmath5 .
@xmath4 is usually hard to measure and one integrates over it .
thus , ideally one would like to evolve for fixed @xmath5 values .
the statistics of the presently available data is not sufficient for carrying this out .
ingelman and schlein @xcite suggested to consider the exchanged pomeron as a particle having internal structure . under this assumption ,
the diffractive process is described as a multistep event : the proton radiates a pomeron having a fraction @xmath5 of the proton momentum .
the virtual photon interacts in a deep inelastic process off the pomeron , scattering of a parton in the pomeron carrying a fraction @xmath6 of the pomeron momentum .
there is thus a flux factor at the proton vertex , dependent only on @xmath5 ( @xmath4 is integrated out ) , and a pomeron structure function @xmath26 .
this picture assumes regge factorization , an assumption which has to be checked by the data .
using the proven qcd factorization together with the assumed regge factorization , one gets diffractive parton distributions .
the assumption of regge factorization clearly does not hold in the inclusive case .
the value of @xmath16 is clearly @xmath15 dependent .
however both the diffractive h1 data and the lps data can be described by an nlo qcd fit with one fixed value of @xmath27(0 ) , as shown by schtzel @xcite and by capua @xcite at this meeting . in case of the lps data ,
a cut on @xmath12 0.01 has to be made . for the h1 analysis
, one needs to add the reggeon contribution .
the statistics of the lps data were not sufficient to repeat the fit in different @xmath15 bins .
the h1 analysis , shows some indication of a rise of @xmath16 with @xmath15 ( see figure [ fig : lama ] ) , though with quite large error bars .
the uncertainty comes not only from the statistics but also from the way one needs to isolate the pomeron contribution from the pomeron + reggeon fit .
we can conclude that in inclusive diffractive processes , for @xmath12 0.01 , the data can be consistent with regge factorization .
there is a clear need for more precise data .
the h1 data , which have a wide kinematical coverage in @xmath15 and in @xmath6 , have been used to extract diffractive parton distributions ( dpdfs ) , shown in figure [ fig : dpdfs ] .
one sees a dominance of the gluon distribution , which is the outcome of the fact that the data show positive scaling violation up to quite high @xmath6 values .
this is quantified in figure [ fig : gluon - fract ] , which shows that for the region @xmath28 , the gluons carry 80% of the pomeron momentum . .
] the same conclusion is reached by the lps analysis .
note that the validity of the diffractive parton distribution functions ( dpdfs ) is in the following kinematic region : @xmath29 3 gev@xmath21 , @xmath30 2 gev and @xmath12 0.05 .
are these dpdfs portable just like in the inclusive pdfs case ?
in other words , does qcd factorization work ?
the diffractive parton distributions , obtained by h1 from the nlo qcd fit , were used to calculate expectations of other diffractive processes .
the agreement with data is very good for diffractive @xmath31 as well as for diffractive dijet production @xcite .
however , the expectations for the tevatron results are by one order of magnitude too high @xcite .
this is not surprising as qcd factorization should not hold for diffractive hadron - hadron reactions .
furthermore , the tevatron data is measured in the kinematic region @xmath32 , where the reggeon exchange dominates and thus would not be called diffraction . as mentioned above , the validity of the h1 fit is for @xmath12 0.05 ( @xmath5 at hera is @xmath11 at the tevatron ) . the fact that qcd factorization seems to fail for hadron - hadron data
is also explained by introducing the notion of survival probability of the rapidity gap @xcite .
taking it at face value , this would mean that for the tevatron processes , the survival probability is about 0.1 . since the photon has a hadronic part ( resolved photon ) , this notion can be tested in diffractive photoproduction of dijets @xcite .
indeed it seems that using a survival probability of 0.34 @xcite , one can describe the resolved photon data .
there seems to be some uncertainty about the conclusion concerning the direct part : while the h1 measurement is below the expectations , the zeus result is consistent with it .
an interesting attempt to fit the combined data of inclusive and inclusive diffraction data was carried out by martin , ryskin and watt @xcite , who included absorption correction to the qcd analysis .
while the quark distributions they get are not very different from those of h1 , their gluon distribution is significantly lower than the h1 one .
this results in expectations which are by almost factor of 3 lower than the h1 ones for the comparison with the tevatron data ( see figure [ fig : watt ] ) .
interactions as a function of @xmath6 , compared to expectations of different sets of diffractive parton distribution functions . ] in this case the survival probability would be closer to that of the resolved photon case .
diffractive processes were said to constitute about 10% of the total inclusive dis processes .
however , the ratio of @xmath35 is @xmath15 dependent .
it can be as high as 20% at @xmath36 3 gev@xmath21 , going down to about 10% at @xmath36 30 gev@xmath21 , as can be seen in fig [ fig : sigratio ] , where this ratio is shown as a function of @xmath15 , for the kinematic region @xmath37 gev , @xmath38 gev , and @xmath39 gev .
, for a selected kinematic region . ] one should however keep in mind that the @xmath15 dependence might be an outcome of the fact that in this presentation of the ratio , different regions of @xmath6 are covered for different values of @xmath15 .
note that this ratio at @xmath15 = 100 gev@xmath21 goes down to @xmath40 5% for @xmath12 0.03 @xcite .
this ratio has been measured for the first time for charged current diffractive processes @xcite .
it is in the range of 2 - 3 % for @xmath5 and @xmath41 0.05 .
one can calculate the ratio of diffractive to total cross section for specific processes and check whether the parton distributions obtained from the nlo qcd analysis fulfill the pumplin bound @xcite .
this was done in @xcite for diffractive to inclusive dijet production induced by gluons and is displayed in figure [ fig : unitarity ] . of the gluon for different scales of the hard scattering , for the recent h1 diffractive parton distribution functions .
also shown is the unitarity limit , called pumplin bound . ]
as seen , the bound is clearly violated for relatively low scales at low @xmath8 .
this might indicate that unitarity effects are already present in the gluon sector .
this subsection is more a presentation of some questions than a real summary .
is regge factorization broken ? also for @xmath12 0.01 ?
there is a need for more precise measurements to come to a clear conclusion .
ideally , so as not to be dependent on the regge factorization assumption in diffraction , one would like to do a qcd analysis for fixed values of @xmath5 .
this will need much higher statistics than presently available .
there is a breaking of qcd factorization when using the parton distribution densities to compare to hadron - hadron data .
this is interpreted by the introduction of the large rapidity gap survival probability .
the value of the survival probability seems to be in the range of 0.1 - 0.3 .
the presently obtained gluon momentum densities seem to give results which are violating the pumplin bound , in certain kinematical regions .
this could be the indication of the presence of unitarity effects .
there is a large ratio of diffractive to total cross section which decreases with @xmath15 .
this section describes exclusive processes like electroproduction of vector mesons or deeply virtual compton scattering ( dvcs ) .
the situation in this cases is much simpler as these processes are clearly diffractive processes at the high energies where they are measured .
there still exist the problem of isolating the elastic process from the proton dissociative one . by measuring the cross section for a limited low @xmath4 range ,
the contribution of the latter is minimized .
one of the nice features seen in these data is the transition from soft to hard processes as one increases the scale . .
the total @xmath42 cross section is plotted for comparison . ]
this transition is seen also for the photoproduction of vector mesons , where the mass serves as the scale .
figure [ fig : xsect ] shows the photoproduction cross section as a function of the @xmath42 center of mass energy , @xmath43 , for different vector mesons .
the light vector mesons , @xmath44 , @xmath45 and @xmath46 show an energy dependence which is characteristic of a soft process ( the total @xmath42 cross section is also shown for comparison ) . for the heavier vector mesons ,
the energy dependence becomes much steeper , as expected from hard processes .
note also that the real part of the amplitude increases with the hardness of the process and is a further reason for the sharp energy dependence .
the soft to hard transition can also be seen for a given vector meson , by changing the @xmath15 of the process .
the cross section is parameterized as @xmath47 and @xmath48 is seen to increase with @xmath15 . to compare all the vector mesons on one plot @xcite
one shows @xmath48 as function of @xmath49 , with @xmath50 being the mass of the vector meson .
as seen in figure [ fig : delvm ] , one gets an increase of @xmath48 from values of about 0.2 ( soft ) at the low scale end to a value of about 1 ( hard ) at high scales . from a fit of the form @xmath47 to the cross section data of exclusive production of vms , as a function of @xmath49 . ] using the energy dependence of vector meson electroproduction at fixed @xmath4 values , one can obtain the parameters of the effective trajectory exchanged in the process . this way the effective trajectory of the pomeron was determined for the @xmath44 , @xmath51 and @xmath52 electroproduction .
a summary plot of the intercepts and slopes for all three vms , as function of @xmath49 , is presented in figure [ fig : apom ] .
( 0 ) and @xmath53 values , extracted in exclusive vm production , as a function of @xmath49 .
] the intercept of the low mass vms are consistent with that of the soft pomeron .
this is not the case for the @xmath52 which has a significantly higher intercept . as for the slope , all values are lower that that of the soft pomeron , as expected from hard processes .
the @xmath4 distribution of the vms can be well described by an exponentially falling cross section , with a slope @xmath54 .
this slope is connected to the size of the vm . at low
@xmath15 the size of the light vms is large , decreasing with @xmath15 from a value of @xmath40 10 gev@xmath55 to about 5 gev@xmath55 . for the @xmath52
the size is small already at low @xmath15 .
this can be seen in figure [ fig : bvm ] , where the slope @xmath54 is plotted as function of @xmath15 .
distribution measured for exclusive vm production as a function of @xmath15 .
the lines are to guide the eye . ]
the dotted lines are just to guide the eye .
the data seem to converge at high @xmath15 on a value of @xmath56 5 gev@xmath55 .
the ratio @xmath58 of the longitudinal photon cross section , @xmath59 , to that of the transverse photon , @xmath60 , can be obtained by studying the decay distribution of the vm and assuming s - channel helicity conservation .
this has been done for @xmath44 @xcite , @xmath51 @xcite and @xmath52 @xcite .
dependence of @xmath58 for exclusive @xmath61 electroproduction . ]
figure [ fig : r - compass ] shows as an example the very impressive preliminary measurements of @xmath58 as a function of @xmath15 for the @xmath44 vm by the compass collaboration @xcite . for all vms
, @xmath58 is rising with @xmath15 . for @xmath61 , @xmath51 and @xmath52 , as a function of @xmath15.the lines
are a fit to the data of the form @xmath62 . ]
a compilation of @xmath58 for all three vms is shown in figure [ fig : rvm ] , as function of @xmath15 .
the lines are a fit to the data of the form @xmath62 , which indicate that @xmath58 scales with @xmath63 .
the photon is described as fluctuating into a @xmath64 pair . when the relative @xmath65 between the pair is small we speak of a large spatial configuration , while if the relative @xmath65 is large , the photon fluctuates into a small spatial configuration .
longitudinal photons have small configurations , while transverse photons consist of both large and small configuration .
a small spatial configuration leads to hard processes and thus to a steep energy dependence of the cross section .
a large configuration has a shallower energy dependence , as expected in soft processes .
it came therefore as a surprise that the ratio @xmath58 for the @xmath44 electroproduction process is @xmath43 independent for @xmath43 values up to 120 gev and @xmath15 up to 19 gev@xmath21 @xcite .
the ratio @xmath58 is @xmath43 independent also for @xmath52 @xcite .
this means that for some reason in case of vm electroproduction the large configurations of the transverse photon are suppressed .
another process which seems to send the same message is dvcs .
the energy dependence of this reaction indicates a hard process .
the @xmath47 dependence yields @xcite @[email protected] for the h1 measurement and @[email protected]@xmath67 for zeus .
such a steep energy dependence would be expected from a dominant longitudinal photon .
however , in dvcs one goes from a virtual photon to a real one , @xmath68 .
assuming s - channel helicity conservation , this means that , since the real photon is transverse , also the initial virtual photon has to be transverse .
thus , the steep energy dependence of the dvcs cross section means that the large configuration in the transverse photon is suppressed .
exclusive diffractive processes become hard once the scale gets large .
the properties of the effective pomeron exchange at the larger scales are consistent with that of a hard process .
the large configurations of the transverse photon seem to be suppressed in exclusive vm , including real photon , production .
diffractive processes are expected to be measured with improved machines and detectors at hera , the tevatron and at rhic . whether we want it or
not , a large portion of the interactions measured at lhc will be of diffractive nature @xcite .
in fact , the exclusive diffractive production of the higgs boson has been proposed as a potential background free method to search for the light higgs at lhc .
one can also dream about a future @xmath69 collider which will allow to reach kinematic region where phase transitions can be observed .
clearly diffraction is a subject which will occupy us for quite some time to come .
it is a pleasure to acknowledge and thank the organizers for the excellent organization of a very pleasant conference in a most beautiful location .
collins , phys . rev .
* d 557 * , 3051 ( 1998 ) ; erratum phys . rev . *
d 61 * , 2000 ( 1998 ) ; a. berera and d.e .
soper , phys . rev . * d 53 * , 6162 ( 1996 ) ; l. trentadue and g. veneziano , phys .
* b 323 * , 201 ( 1994 ) .
dokshitzer , sov .
jetp * 46 * , 641 ( 1977 ) [ zh .
. fiz . * 73 * , 1216 ( 1977 ) ] ; v.n .
gribov and l.n .
lipatov , yad . fiz . * 15 * , 1218 ( 1972 ) [ sov .
. phys . * 15 * , 675 ( 1972 ) ] ; v.n .
gribov and l.n .
lipatov , yad .
fiz . * 15 * , 781 ( 1972 ) [ sov .
. phys . * 15 * , 438 ( 1972 ) ] ; g. altarelli and g. parisi , nucl . phys .
* b 126 * , 298 ( 1977 ) . | some issues in inclusive and exclusive diffractive processes are discussed . |
the groundbreaking early results of nasa s _ kepler _ mission @xcite have spectacularly added to our understanding of the frequency and properties of planets in the galaxy @xcite .
eight planetary systems discovered by _
kepler _ have already been confirmed by radial velocity or transit timing : four hot jupiters @xcite ; two saturn - mass planets and a super - earth in the kepler-9 system @xcite ; a neptune - mass planet kepler-4b @xcite ; kepler-10b , the smallest exoplanet yet known @xcite ; and the closely - aligned six transiting planet system kepler-11 @xcite .
the properties of the exoplanet candidate ( ec ) population discovered by
_ kepler _ will provide strong constraints on models of planet formation , especially for low - mass stars and small planets .
it is well established that metal - rich solar - type stars are more likely to host giant planets than more metal - poor solar - type stars ( e.g. , * ? ? ?
* ; * ? ? ?
likewise , metal - rich low - mass stars also seem more likely to host giant planets ( e.g. , * ? ? ?
meanwhile , preliminary results indicate that near solar metallicity there is not much of a correlation between solar - type host star metallicity and the likelihood of hosting neptune - mass planets ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
there are also trends in host stellar mass , as higher mass stars are more likely to host giant planets than lower mass stars ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in @xcite
, we tentatively noted a hint that m dwarfs that host low - mass planets are more likely to be metal rich than expected based on random sampling from the field m dwarf population .
we explained this correlation in the context of the core - accretion model of planet formation ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in particular , we argued that a more metal - rich protoplanetary disk will almost certainly have a higher surface density of solids . that increased solid surface density may enable the rapid formation of the several earth - mass embryo necessary to form the core of an ice giant and accrete gas from the protoplanetary disk before the gaseous disk
is dissipated .
that correlation might also indicate a lower limit on the amount of solid material necessary to form planets . in this paper
, we determine whether _ kepler _ ec host stars are more metal rich than stars with no detected planets . to that end
, we use a metallicity - sensitive color - color plot to examine the optical and infrared properties of the sample of _ kepler _ ec host stars relative to a control sample of dwarf stars observed by _
kepler _ but that were not observed to host transiting planets .
we describe our analysis procedures in section 2 , we discuss the results and implications of our analysis in section 3 , and we summarize our findings in section 4 .
we are interested in using the available photometric data for stars in the _ kepler _ field to determine if there is an offset in metallicity between the _ kepler _ ec host population and the population of stars observed by _
kepler _ but with no detected planets .
in particular , we compare the mean @xmath1 color of _ kepler _ ec hosts with the mean @xmath1 color of stars observed by _
kepler _ but with no detected planets at constant @xmath0 color .
@xcite explored the potential of sdss photometry to produce accurate photometric metallicities for metal - poor stars in the halo of the milky way .
unfortunately , the accuracy of those methods depends crucially on the availability of precise sdss @xmath8-band photometry .
nevertheless , even when @xmath8-band data is lacking , @xmath9,@xmath10 , and @xmath11 photometry can still provide very useful photometric metallicities ( e.g. , * ? ? ?
our problem is made easier because we are only interested in the relative ( not absolute ) metallicities of the two populations . at the same time , unlike @xcite and @xcite , we have the added benefit of 2mass @xmath12 photometry in the _ kepler _ field and apparent magnitude range . as a result , we can use infrared colors ( e.g. , @xmath0 ) as a proxy for effective temperature @xmath13 .
the _ kepler _ input catalog ( kic - * ? ? ? * ; * ? ? ? * ) provides reasonably accurate estimates of @xmath14 , so we can ensure that giant stars do not contaminate our control sample of stars with no detected planets . stars with
@xmath15 have main sequence lifetimes much longer than 10 gyr , and because age has little effect on the colors of low - mass stars on the main sequence , age differences between the two populations ( if they exist ) are unlikely to affect our analysis . since we compare dwarf stars with similar @xmath13 and @xmath14 , the only stellar parameter that can vary between the _ kepler _ ec hosts and the control sample is [ fe / h ] .
as metallicity is the only stellar parameter that can vary , the best explanation for systematically red @xmath1 colors in the ec host population relative to the control sample is that the ec host population is enriched in metals . for each of the 997 stellar hosts of the 1,235
_ kepler _ exoplanet candidates announced in @xcite , we obtain all available broadband photometry ( including 2mass @xmath12 from * ? ? ?
* ) and reddening data from the kic .
simultaneously , we obtain equivalent photometry and reddening data from the same sources for a random sample of 10,000 dwarf stars from the @xmath2 stars observed during q1 and q2 of the _ kepler _ mission with no candidate planets and kic - based @xmath16 .
this control sample is subject to the same selection effects that were applied in the _ kepler _ field to produce a list of stars to search for transiting planets .
consequently , differences in the characteristics of the stars in the control sample and the stars that host ecs are not related to the selection effects applied in the _ kepler _ field to identify a sample of stars suitable for transit observations .
metal - rich stars typically have redder optical colors than solar metallicity stars , as the forest of iron lines in the atmospheres of metal - rich stars preferentially absorb blue photons . at the same time
, there are other factors that affect the color of a star , most notably reddening due to the interstellar medium , its age , and its evolutionary state ( e.g. , dwarf or giant ) . in figure [ fig01 ]
, we plot a @xmath0@xmath1 color - color plot for the control sample and the ec host sample .
we subdivide the ec host sample into two sub - samples : those stars that host at least one giant ec ( e.g. , @xmath17 ) and those stars that host no giant ecs .
we use padova isochrones @xcite for a 2 gyr stellar population at [ fe / h ] = -0.1 and 0.1 to determine the effect of metallicity on the stellar locus in a @xmath0@xmath1 color - color plot .
the yellow arrows show the effect of increasing [ fe / h ] by 0.2 dex on @xmath1 at constant @xmath0 .
note though that the yellow arrows do not connect stars of constant mass , as a 0.2 dex [ fe / h ] enriched star will be about 5% more massive than a solar metallicity star at constant @xmath0 color .
we do not use the padova isochrone redward of @xmath18 , as the theoretical stellar models and atmospheres become unreliable after that point ( e.g. , * ? ? ?
consequently , we use the m 67 fiducial sequence from @xcite to illustrate the morphology of a solar metallicity population at red @xmath0 color .
age can also affect the color of main sequence star , though near @xmath5 age has little effect on the main sequence .
giant stars have slightly bluer @xmath1 colors than dwarf stars at constant @xmath0 , so the presence of giant stars in the control sample could in principle produce a color offset between the _ kepler _ ec host population and the control sample .
fortunately , kic - based @xmath14 estimates are accurate enough to ensure that giant stars are at most a few percent of our control sample @xcite .
two trends are immediately apparent in figure [ fig01 ] .
first , at @xmath19 typical of late k dwarfs , giant ecs become very rare relative to smaller ecs .
we interpret this as evidence in support of the correlation between host stellar mass and likelihood of hosting a giant planet ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
second , at @xmath20 dominated by late k dwarfs , ecs are preferentially found around the reddest part of the stellar locus as defined by the control sample .
we quantify the significance of this latter trend in figure [ fig02 ] . to determine the significance of the observation that low - mass stars that host ecs are on average redder in @xmath1 than the stellar locus as defined by the control sample , we bin both the control sample and each sub - sample of ec hosts into 0.16 mag bins in @xmath0 color and compute the mean and median @xmath1 color in that bin .
we estimate the error in each mean and median by bootstrap resampling .
we find that at @xmath21 characteristic of solar - type stars , giant ec hosts have preferentially red @xmath1 colors at the 4-@xmath4 level relative to the control sample . on the other hand , at that same @xmath0 color ,
the hosts of small ecs are indistinguishable from the control sample .
in addition , at @xmath22 the stars that host small ecs are 4-@xmath4 redder in @xmath1 than the control sample .
we summarize the results of our calculations in table [ tbl-1 ] .
we argue that the significant color offsets apparent in figure [ fig02 ] and table [ tbl-1 ] are the result of the enriched metallicity of the ec host sample . indeed , it reassuring that we identify the hosts of giant ecs as significantly red in @xmath1 at constant @xmath0 , as it is well established that the stellar hosts of giant planets are preferentially metal rich ( e.g. , * ? ? ?
* ; * ? ? ?
other possible explanations for the observed offset are selection effects , differential reddening , systematic age differences between ec hosts and the control sample , and the presence of giant stars in the control sample but not the ec host sample .
the depth of a transit is proportional to @xmath23 , so it is easier to identify the signal of transiting planet around a small star . on the main sequence
, both age and metallicity can affect the radius of a star . according to the padova isochrones for a 2 gyr population @xcite , at @xmath24 a star with [ fe / h ] = 0.1
is about 5% larger than a star with [ fe / h ] @xmath25 and the same @xmath0 color . at @xmath5 , a star with [ fe / h ] = 0.1
is about 7% larger than a star with [ fe / h ] @xmath25 and the same @xmath0 color . according to the padova isochrones , at @xmath24 a 5 gyr
old star is about 6% larger than a 2 gyr old star at the same @xmath0 color . at @xmath5 ,
a 10 gyr old star is about 2% larger than a 2 gyr old star at the same @xmath0 color .
consequently , if there were no correlation between host metallicity and probability of hosting a planet or possible tidal destruction of aged exoplanet systems , transit surveys would be more likely to identify transiting planets around young , metal - poor stars . as we show in section 2.2.3 , age on the main sequence
has very little effect on the @xmath1 color of a star at constant @xmath0 , so these effects are not likely to produce the color offsets we observe .
the @xmath1 color offset at constant @xmath0 color that we observe between the _ kepler _ ec host sample and the control sample is only meaningful if the two samples were subject to the same selection effects .
as we argued in section 2.1 , both samples were selected according to the same _ kepler _ target selection algorithm . to verify that the two samples are similar
, we plot in figure [ fig03 ] six color - color plots and two color - magnitude diagrams ( similar to * ? ? ? * ) . in all cases ,
the control sample outlines the distribution of both the sample of giant ec hosts and the sample of small - radius ec hosts .
the ec hosts that are outliers in each panel may be unequal mass binary star systems or systems with poorly - estimated reddening .
the _ kepler _ ids ( koi numbers ) of a few of our most extreme color outliers are : 1161345 ( 984 ) , 2446113 ( 379 ) , 5356593 ( 644 ) , 8162789 ( 521 ) , 10470206 ( 335 ) , 10514430 ( 263 ) , and 11465813 ( 771 ) .
in short , the age or metallicity of the host star of an exoplanet does not significantly affect the probability of detecting a transiting planet , and our control sample is a reasonable control sample for our measurement of relative color offsets .
we therefore conclude that selection effects are unlikely to produce the color offsets we observe in figure [ fig02 ] and table [ tbl-1 ] . reddening due to
the interstellar medium can affect the observed color of a star . in our differential analysis , this will only affect our result if the reddening of the control sample is systematically different than the reddening of the _ kepler _ ec host sample . fortunately , figure [ fig04 ] indicates that control sample is subject to the same reddening distribution as the ec host sample .
like the ec host sample , the control sample is also spread more or less uniformly over the _
kepler _ field , so any angular dependence on reddening is unlikely to produce a significant color offset between the control sample and the ec host sample .
we therefore conclude that differential reddening between the control sample and the _ kepler _ ec host sample is unlikely to produce the color offsets we observe in figure [ fig02 ] and table [ tbl-1 ] . as stars age on the main sequence , their colors can evolve . according to the padova isochrones , at @xmath24
a 5 gyr old star is about 0.1% redder in @xmath1 than a 2 gyr old star at the same @xmath0 color . at @xmath26 , a 10 gyr old star is about 0.5% bluer in @xmath1 than a 2 gyr old star at the same @xmath0 color .
clearly , colors of stars do not evolve significantly on the main sequence .
tidal evolution in exoplanet systems can destroy close - in planets ( e.g. , * ? ? ?
* ; * ? ? ?
transit surveys are strongly biased towards close - in systems , so it is possible that transit surveys may be biased towards young stars , as tidal evolution may have already destroyed planets that once orbited older stars . assuming that planets arrive in the close proximity of their host star on nearly circular orbits and that the timescale for eccentricity damping is very short , then the timescale for tidal disruption is approximately ( e.g. , * ? ? ?
* ; * ? ? ?
* ) @xmath27 \mathrm{,}\end{aligned}\ ] ] where @xmath28 is the initial semimajor axis of the planet before tidal evolution , @xmath29 is the specific dissipation function of the host star , @xmath30 is newton s gravitational constant , @xmath31 is the mass of the planet , and @xmath32 is the radius of the host star . if tidal evolution efficiently destroys planets in the _ kepler _ sample , then the hosts of _ kepler _ ecs may be preferentially younger than stars in the control sample .
the median period of giant ecs is 13 days , while the median period of small - radius ecs is 10 days .
as we plot in figure [ fig05 ] , the timescale for tidal disruption of such systems is in excess of 100 gyr . at the same time , as we showed above , age has little effect on the @xmath1 color of a main sequence star at constant @xmath0 .
we therefore conclude that systematic age differences between the control sample and the _ kepler _ ec host sample or tidal evolution in ec systems are unlikely to produce the color offsets we observe in figure [ fig02 ] and table [ tbl-1 ] .
giant stars have stellar radii @xmath33 , and because transit depth is proportional to @xmath23 , the transit of a planet with radius @xmath34 in front of a giant star with @xmath35 is about 10,000 times harder to detect than the transit of the same planet in front of a dwarf star with radius @xmath36 . as a result , _ kepler _ is unlikely to identify transiting planets around giant stars , so the sample of _ kepler _ ec hosts is very likely free of giant stars .
while we only used stars with kic - based @xmath37 in our control sample , it is possible that a few percent of the stars in our control sample are giant stars @xcite .
giant stars have slightly bluer @xmath1 colors at constant @xmath0 color than dwarf stars , so the presence of a significant number of giant stars in our control sample could explain the 0.08 mag @xmath1 color offset we observe at @xmath5 between the sample of _ kepler _ ec hosts and our control sample . to investigate the affect of significant giant star contamination in our control sample , we obtain all available broadband photometry ( including 2mass @xmath12 from * ? ? ?
* ) and reddening data from the kic for a sample of stars observed by _ kepler _ during q1 and q2 with kic - based @xmath38 .
the stars in this sample are likely giant stars .
recall that the 10,000 stars in our control sample have kic - based @xmath37 , so we can add stars from the sample of likely giant stars to our original control sample of likely dwarf stars to create giant star contaminated versions of our our control sample . in particular , we create four versions of the giant star contaminated control sample with differing levels of contamination : 30% , 10% , 5% , and 3% contimination . to determine the @xmath1 color offsets expected at constant @xmath0 color between a sample of 997 dwarf stars ( the number of _ kepler _ ec hosts )
, we create a sample of 997 dwarf stars by randomly selecting 997 stars from the original , not - intentionally contaminated control sample .
as before , we compute the mean @xmath1 color in bins of constant @xmath0 color for each of the five samples described above .
we plot the result of this calculation in figure [ fig06 ] .
we find that between 10% and 30% of the stars in our control sample of stars with kic - based @xmath37 would need to be giants to reproduce the 0.08 mag offset in @xmath1 at @xmath5 we observe between the hosts of _ kepler _ ec and our original not - intentionally contaminated control sample .
the level of contamination is at least a factor of a few larger than the level of contamination found in the kic by @xcite and @xcite .
moreover , figure [ fig06 ] indicates that a giant star contaminated control sample would have redder @xmath1 color at @xmath39 than control sample of dwarf stars ( very much like the likely hosts of _ kepler _ ecs ) .
we do not observe this effect in figure [ fig02 ] . indeed , the hosts of _ kepler _ ecs are redder than the control sample at @xmath39 . at the same time
, the presence of giant stars has no effect on the @xmath1 color of our sample at @xmath40 .
for that reason , the presence of giant stars in our control sample can not explain the significant color offsets we observe between the hosts of giant _ kepler _ ecs and our original control sample .
we therefore conclude that contamination of our control sample by giant stars is unlikely to produce the color offsets we observe in figure [ fig02 ] and table [ tbl-1 ] .
we use two methods to transform the @xmath1 color offsets we observe in figure [ fig02 ] and table [ tbl-1 ] into approximate metallicity offsets .
first , we use dereddened fiducial sequences for the open clusters m 67 ( [ fe / h ] @xmath41 and age @xmath42 gyr ) and ngc 6791 ( [ fe / h ] = @xmath43 and age @xmath44 gyr ) in the sloan photometric system from @xcite plotted in figure [ fig07 ] . in this case , the fiducial sequences show that at constant @xmath45 , the metal - enriched ngc 6791 fiducial sequence is @xmath46 mag redder in @xmath1 than the solar metallicity m 67 fiducial sequence .
though these two clusters have different ages , the stars with @xmath47 have main - sequence lifetimes much longer than 10 gyr and therefore should still be on the main sequence in both clusters .
consequently , the @xmath46 mag offset in @xmath1 between the two fiducial sequences is likely a result of the @xmath48 dex offset in [ fe / h ] between the two clusters .
accordingly , at constant @xmath47 it seems that @xmath49 } = 0.1 $ ] results in @xmath50 . to transform the @xmath0 colors of the points in figure [ fig02 ] and table [ tbl-1 ] into @xmath51
, we use a 2 gyr solar metallicity padova isochrone . to more
precisely estimate the average @xmath1 offset between the m 67 and ngc 6791 fiducial sequences , we calculate the difference between the fiducial sequences at three points ( @xmath51 = 5.9 , 6.4 , and 6.8 corresponding to @xmath0 = 0.46 , 0.54 , and 0.62 ) and average the result . in this way , we find that in the interval @xmath52 , an offset of 0.1 dex in metallicity results in an offset of 0.035 mag in @xmath1 .
this method is only applicable to stars with @xmath53 , so in table [ tbl-1 ] we only give the metallicity offsets indicated by this method for the three reddest @xmath0 bins in our analysis . in order to transform a @xmath1 color offset into a metallicity offset for more massive stars , we use padova isochrones .
in particular , we use two gyr isochrones for two metallicities ( [ fe / h ] = -0.1 and 0.1 ) to determine the change in @xmath1 color at constant @xmath0 attributable to a 0.2 dex increase in metallicity .
we use the two isochrones to determine the offset in @xmath1 at constant @xmath0 that can be attributed to metallicity , then use that @xmath1 offset to convert the observed @xmath1 offsets in table [ tbl-1 ] into metallicity offsets ( extrapolating if necessary ) . interestingly , even in the presence of possible systematics in our analysis ( e.g. , differential reddening , age differences , or the possible presence of giant stars in the control sample ) , it is reassuring that we recover the fact that solar - type stars that host giant planets are metal enriched relative to the field population . in figure
[ fig08 ] , we plot @xmath13 and metallicity for both a volume - limited sample of stars ( @xmath54 20 pc ) from the geneva - copenhagen survey ( gcs - * ? ? ?
* ; * ? ? ?
* ) and confirmed giant planet hosts ( e.g. , * ? ? ?
the horizontal lines give the average metallicities of the two populations .
the vertical arrows are at the @xmath55 corresponding to @xmath0 = 0.22 and 0.30 .
the arrows are anchored at the mean metallicity of the gcs sample , and their length corresponds to the metallicity offset we need to explain the differences in @xmath1 color we observe in the giant planet bins at those @xmath0 colors reported in table [ tbl-1 ] .
reassuringly , the metallicity offsets we compute based on @xmath1 color offsets are in quantitative agreement with previous observations of metal - enhancement in giant planet hosts . in other words ,
our analysis is quantitatively correct in the regime where it is possible to compare with previous results .
that is suggestive that our result for the low - mass stars that host small - radius _ kepler _ ecs is also reliable .
we identified three properties of the _ kepler _ ecs in section 2 : 1 . the population of solar - type stars that host giant ecs is 4-@xmath4 redder by 0.04 mag in @xmath1 at constant @xmath24 than a control sample of stars observed by _ kepler _ but with no detected planets .
the population of solar - type stars that host small ecs is similar in @xmath1 at constant @xmath24 to a control sample of stars observed by _
kepler _ but with no detected planets . 3 .
the population of low - mass stars that host small ecs is 4-@xmath4 redder by 0.08 mag in @xmath1 at constant @xmath5 than a control sample of stars observed by _
kepler _ but with no detected planets .
the first observation is consistent with the known correlation between solar - type host star metallicity and likelihood of hosting a giant planet ( e.g. , * ? ? ?
* ; * ? ? ?
the second observation supports the tentative assertion that for solar - type stars in the solar neighborhood , metallicity is only weakly correlated with the likelihood of hosting a low - mass planet ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
if the @xmath1 color offset is the result of a 0.2 dex enhancement in [ fe / h ] of the low - mass hosts of small - radius _ kepler _ ecs , then the third observation supports the tentative assertion put forth in @xcite that m dwarfs that host low - mass planets are more likely to be metal rich than expected based on random sampling of the field m dwarf population .
the observation that there is a correlation between host star metallicity and the presence of low - mass planets around low - mass stars is expected in the core - accretion model of planet formation . in that scenario , the cores of neptune - mass planets ( and possibly super - earths ) form and
grow to several earth masses before their parent protoplanetary disk is fully dissipated . to first order , the mass of a protoplanetary disk scales as @xmath56 and the fraction of the disk in solid material scales as @xmath57 .
the total amount of solids in a disk able to form the core of a neptune - mass planet is then @xmath58 .
the protoplanetary disk in which the solar system formed likely had a solid mass @xmath59 ( e.g. , * ? ? ? * ) .
if the probability of forming a planet is proportional to the total amount of solids in its parent protoplanetary disk , a late k dwarf with @xmath60 would need to have a metallicity of [ fe / h ] @xmath61 to have the same chance of forming a planetary system as a solar metallicity solar - type star . as a result
, it may only be the metal - rich late k dwarfs that have protoplanetary disks with enough solid mass to form the several earth - mass core of an ice giant or super - earth .
though comparing stars at constant @xmath0 color forces us to compare stars of slightly different masses , the magnitude of this effect on the planet formation process is likely an order of magnitude smaller than the effect of a 0.2 dex enhancement in [ fe / h ] .
the correlation between the red optical color of late k dwarfs and probability of hosting a small ec indicates that _ kepler _ might boost its yield of planets by shading their late k dwarf sample to redder optical colors . in particular
, focusing on stars in the range @xmath62 above the line @xmath63 + 1\end{aligned}\ ] ] [ eq01 ] might produce a larger sample of small ecs than a search that does not use red @xmath1 color in its target selection .
we find that at @xmath64 , low - mass stellar hosts of small - radius _ kepler _ ecs are 4-@xmath4 redder in @xmath1 than a control sample of stars with the same @xmath0 color but with no detected planets .
this result is unlikely to be an artifact of reddening , age differences between the two populations , or the presence of giant stars in the control sample .
stellar models and fiducial sequences in the sloan photometric system of the galactic open clusters m 67 and ngc 6791 indicate that a 0.2 dex enrichment in [ fe / h ] of the ec host population can produce the offset .
we suggest that small planets are preferentially found around metal - rich low - mass stars ( confirming the tentative correlation advocated in * ? ? ?
we also confirm that giant _ kepler _ ecs are preferentially found around metal - rich solar - type stars and that the presence of small _ kepler _ ecs orbiting solar - type stars does not seem to be strongly correlated with host metallicity .
we thank tim brown , connie rockosi , and graeme smith for useful comments .
we also thank the anonymous referee for suggestions that improved this manuscript .
this research has made use of nasa s astrophysics data system bibliographic services , the exoplanet orbit database , and the exoplanet data explorer at exoplanets.org .
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 nnx09af08 g and by other grants and contracts .
this material is based upon work supported under a national science foundation graduate research fellowship .
an , d. , et al .
2008 , , 179 , 326 an , d. , et al . 2009a , , 700 , 523 an , d. , et al .
2009b , , 707 , l64 basri , g. , et al .
2011 , , 141 , 20 batalha , n. m. , et al . 2010a , , 713 , l103 batalha , n. m. , et al .
2010b , , 713 , l109 batalha , n. m. , et al .
2011 , , 729 , 27 borucki , w. j. , et al . 2010a , science , 327 , 977 borucki , w. j. , et al .
2010b , , 713 , l126 borucki , w. j. , et al . 2011a , , 728 , 117 borucki , w. j. , et al .
2011b , arxiv:1102.0541 bouchy , f. , et al .
2009 , , 496 , 527 brown , t. m. , latham , d. w. , everett , m. e. , & esquerdo , g. a. 2011 , arxiv:1102.0342 bryson , s. t. , et al .
2010 , , 713 , l97 caldwell , d. a. , et al .
2010 , , 713 , l92 covey , k. r. , et al .
2007 , , 134 , 2398 dunham , e. w. , et al .
2010 , , 713 , l136 fischer , d. a. , & valenti , j. 2005 , , 622 , 1102 ford , e. b. , et al . 2011 , arxiv:1102.0544 girardi , l. , et al .
2010 , , 724 , 1030 gu , p .- g . ,
lin , d. n. c. , & bodenheimer , p. h. 2003 , , 588 , 509 haas , m. r. , et al .
2010 , , 713 , l115 holman , m. j. , et al .
2010 , science , 330 , 51 holmberg , j. , nordstrm , b. , & andersen , j. 2007 , , 475 , 519 holmberg , j. , nordstrm , b. , & andersen , j. 2009 , , 501 , 941 howard , a. w. , et al . 2011 , arxiv:1103.2541 hubickyj , o. , bodenheimer , p. , & lissauer , j. j. 2005 , icarus , 179 , 415 ibgui , l. , & burrows , a. 2009 , , 700 , 1921 ida , s. , & lin , d. n. c. 2004 , , 616 , 567 ida , s. , & lin , d. n. c. 2005 , , 626 , 1045 ivezi , . ,
2008 , , 684 , 287 jenkins , j. m. , et al . 2010a , , 713 , l87 jenkins , j. m. , et al .
2010b , , 713 , l120 jenkins , j. m. , et al .
2010c , , 724 , 1108 johnson , j. a. , & apps , k. 2009 , , 699 , 933 johnson , j. a. , aller , k. m. , howard , a. w. , & crepp , j. r. 2010 , , 122 , 905 johnson , j. a. , butler , r. p. , marcy , g. w. , fischer , d. a. , vogt , s. s. , wright , j. t. , & peek , k. m. g.2007 , , 670 , 833 koch , d. g. , et al . 2010a , , 713 , l79 koch , d. g. , et al .
2010b , , 713 , l131 latham , d. w. , et al .
2010 , , 713 , l140 latham , d. w. , et al .
2011 , , 732 , l24 laughlin , g. , bodenheimer , p. , & adams , f. c. 2004 , , 612 , l73 lissauer , j. j. 1993 , , 31 , 129 lissauer , j. j. , et al . 2011a , , 470 , 53 lissauer , j. j. , et al . 2011b , arxiv:1102.0543 mardling , r. a. , & lin , d. n. c. 2004 , , 614 , 955 marigo , p. , girardi , l. , bressan , a. , groenewegen , m. a. t. , silva , l. , & granato , g. l. 2008 , , 482 , 883 moorhead , a. v. , et al .
2011 , arxiv:1102.0547 mordasini , c. , alibert , y. , & benz , w. 2009 , , 501 , 1139 mordasini , c. , alibert , y. , benz , w. , & naef , d. 2009 , , 501 , 1161 pollack , j. b. , hubickyj , o. , bodenheimer , p. , lissauer , j. j. , podolak , m. , & greenzweig , y. 1996 , icarus , 124 , 62 santos , n. c. , israelian , g. , & mayor , m. 2004 , , 415 , 1153 schlaufman , k. c. , & laughlin , g. 2010 , , 519 , a105 schlaufman , k. c. , lin , d. n. c. , & ida , s. 2010 , , 724 , l53 skrutskie , m. f. , et al .
2006 , , 131 , 1163 sousa , s. g. , et al . 2008 , , 487 , 373 udry , s. , et al .
2006 , , 447 , 361 wright , j. t. , et al .
2011 , , 123 , 412 cccccccccc 0.22 & 1.23 & 0.01 & 1.70 & @xmath65 & 0.11 & 0.03 & 2.06 & @xmath65 & 0.22 + 0.30 & 1.05 & 0.01 & 1.14 & @xmath65 & 0.05 & 0.04 & 4.44 & @xmath65 & 0.22 + 0.38 & 0.93 & 0.01 & 1.91 & @xmath65 & 0.07 & 0.02 & 2.59 & @xmath65 & 0.14 + 0.46 & 0.85 & 0.06 & 1.72 & 0.17 & 0.39 & 0.08 & 4.11 & 0.23 & 0.54 + 0.54 & 0.78 & 0.02 & 1.03 & 0.06 & 0.09 & 0.02 & 0.69 & 0.06 & 0.09 + 0.62 & 0.73 & 0.08 & 4.00 & 0.24 & @xmath65 & @xmath65 & @xmath65 & @xmath65 & @xmath65 | we find that _ kepler _ exoplanet candidate ( ec ) host stars are preferentially metal - rich , including the low - mass stellar hosts of small - radius ecs .
the last observation confirms a tentative hint that there is a correlation between the metallicity of low - mass stars and the presence of low - mass and small - radius exoplanets .
in particular , we compare the @xmath0@xmath1 color - color distribution of _ kepler _ ec host stars with a control sample of dwarf stars selected from the @xmath2 stars observed during q1 and q2 of the _ kepler _ mission but with no detected planets .
we find that at @xmath3 characteristic of solar - type stars , the average @xmath1 color of stars that host giant ecs is 4-@xmath4 redder than the average color of the stars in the control sample . at the same time , the average @xmath1 color of solar - type stars that host small - radius ecs is indistinguishable from the average color of the stars in the control sample .
in addition , we find that at @xmath5 indicative of late k dwarfs , the average @xmath1 color of stars that host small - radius ecs is 4-@xmath4 redder than the average color of the stars in the control sample .
these offsets are unlikely to be caused by differential reddening , age differences between the two populations , or the presence of giant stars in the control sample .
stellar models suggest that the first color offset is due to a 0.2 dex enhancement in [ fe / h ] of the giant ec host population at @xmath6 , while sloan photometry of m 67 and ngc 6791 suggests that the second color offset is due to a similar [ fe / h ] enhancement of the small - radius ec host population at @xmath7 .
these correlations are a natural consequence of the core - accretion model of planet formation . |
financial support from the dfg is gratefully acknowledged by jk and tp . financial support from the australian research council
is gratefully acknowledged by alo and ar .
alo also thanks the institut fr theoretische physik at the technische universitt clausthal . | a self - interacting polymer with one end attached to a sticky surface has been studied by means of a flat - histogram stochastic growth algorithm known as flatperm .
we examined the four - dimensional parameter space of the number of monomers up to 91 , self - attraction , surface attraction and force applied to an end of the polymer . using this powerful algorithm the _ complete _ parameter space of interactions and
force has been considered .
recently it has been conjectured that a hierarchy of states appears at low temperature / poor solvent conditions where a polymer exists in a finite number of layers close to a surface .
we find re - entrant behaviour from a stretched phase into these layering phases when an appropriate force is applied to the polymer .
we also find that , contrary to what may be expected , the polymer desorbs from the surface when a sufficiently strong critical force is applied and does _ not _ transcend through either a series of de - layering transitions or monomer - by - monomer transitions .
new experimental methods in the physics of macromolecules @xcite have been used to study and manipulate single molecules and their interactions .
these methods make a contribution to our understanding of such phenomena as protein folding or dna un - zipping ; one can push or pull a single molecule and watch how it responds .
it is possible to apply ( and measure ) forces large enough to induce structural deformation of single molecules .
one can monitor the mechanism of some force - driven phase transition occurring at the level of a single molecule .
theoretical understanding of this behaviour has attracted much attention @xcite .
the response of a single polymer to an external force under good solvent conditions was considered some time ago@xcite .
the response under poor solvent conditions ( below the @xmath0-point ) , where the self - attraction and an external force compete with each other , was examined later @xcite .
another phenomenon commonly studied in polymer physics is the adsorption of a polymer tethered to a `` sticky '' wall .
the response of such a polymer to a force perpendicular to the wall has also recently been considered @xcite .
however , when both the self - attraction ( ie .
monomer - monomer attraction that leads to polymer collapse ) and the surface attraction ( ie monomer - wall attraction that leads to adsorption ) compete the response to an external force has not yet been elucidated ( some interesting results can be found in @xcite ) .
certainly , the full phase diagram has not been considered .
making such a study now is all the more timely because of the very recent discovery @xcite of a new low temperature phenomenon of layering transitions ( without a force ) .
it is this layering phenomenon that raises the intriguing question about the response a low - temperature polymer may have to an external force . in the layering state a polymer is tightly confined within a fixed number of layers above the wall .
it may be therefore be especially interesting experimentally to examine such a situation .
we demonstrate for the _ first _ time how the full two - dimensional phase diagram of surface and self - attraction changes as the force is increased .
the desorbed extended regime , which changes its scaling behaviour as soon as the force is made non - zero , simply grows as the force is increased .
the second - order phase transitions of adsorption and collapse become first order .
the rest of the phase diagram remains relatively unaffected as long as the force is small .
after the force passes a critical value , which depends on the zero temperature force required to pull a polymer from a wall , a re - entrant behaviour occurs at low temperatures . for different values of the force
, this re - entrant behaviour occurs for both the adsorption and collapse of polymers , including the layering phases mentioned above .
we provide a full force - temperature diagram for all ratios of surface attraction to self - attraction .
all this is achieved with the use of a recently developed algorithm , flatperm @xcite , that is specifically designed to obtain information about the whole phase diagram in one simulation run : it is effectively a stochastic enumeration algorithm that estimates the complete density of states .
the model we have considered is a self - avoiding walk on a three - dimensional cubic lattice in a half - space .
the self - avoiding walk is attached at one end to the wall with surface energy per monomer of @xmath1 for _ visits _ to the wall .
the self - avoiding walk self - interacts via a nearest - neighbour energy of attraction @xmath2 per monomer - monomer _ contact_. ( note that the attractive energies @xmath2 and @xmath1 are taken to be positive . ) a force @xmath3 is applied in the direction perpendicular to the boundary of the half - space ( wall ) .
the total energy of a configuration @xmath4 of length ( number of monomers ) @xmath5 is given by @xmath6 and depends on the number of non - consecutive nearest - neighbour pairs ( contacts ) along the walk @xmath7 , the number of visits to the planar surface @xmath8 , and the height @xmath9 in the direction perpendicular to the boundary ( wall ) of the half - space .
figure 1 shows a diagram of the two - dimensional analogue . for convenience ,
we define @xmath10 , @xmath11 and @xmath12 where @xmath13 is the temperature and @xmath14 is the boltzmann constant .
the partition function is given by @xmath15 with @xmath16 being the density of states .
it is this density of states that is estimated directly by the flatperm simulation .
our algorithm grows a walk monomer - by - monomer starting on the surface .
we obtained data for each value of @xmath5 up to @xmath17 , and all permissible values of @xmath7 , @xmath8 , and @xmath9 . when @xmath18 the phase diagram of the model contains several phases and transitions between them @xcite .
[ cols="^,^ " , ] for small @xmath19 and @xmath20 there is a desorbed extended ( de ) phase with the polymer behaving as a free flexible polymer in solution ( ie . swollen or extended in three dimensions ) . for @xmath19
fixed and small , increasing @xmath20 leads to a second - order phase transition ( adsorption ) to a state in which the polymer is adsorbed onto the wall and behaves in a swollen two - dimensional fashion ( ae ) .
alternately , if @xmath19 is increased at small @xmath20 a second - order collapse transition occurs to a state resembling a dense liquid drop .
this phase is known as desorbed collapsed ( dc ) on the assumption that it has little contact with the wall @xcite .
however , it has been subsequently argued @xcite that for larger @xmath19 there is also a polymer - surface transition to a surface - attached globule ( sag ) phase , where the polymer behave as a liquid drop partially wetting the wall .
this transition will not be seen directly by studying thermodynamic polymer quantities as it occurs as a singularity in the surface free energy and not the bulk free energy of the polymer . when @xmath20 is large ,
so that the polymer is adsorbed onto the wall , increasing @xmath19 will result in a two - dimensional ( second - order ) transition to a adsorbed and collapsed phase ( ac ) . in recent work @xcite this ac phase
was also referred to as the _
1-layer _ phase because for very large @xmath19 and @xmath21 there exist meta - stable @xmath22-layer phases where the polymer is two - dimensionally collapsed and more - or - less restricted to @xmath22 layers parallel to the wall ( for small @xmath22 ) .
a series of first - order transitions between adjacent values of @xmath22 occur as @xmath20 is varied at fixed @xmath19 .
all these transition lines can be seen in the figure 2 ( a ) which shows a plot of the logarithm of the maximum eigenvalue of the ( @xmath23 ) matrix of second derivatives in the variables @xmath19 and @xmath20 of @xmath24 for fixed @xmath25 .
the local maxima indicate transitions .
the transition to the ac phase is expected to be first order in the thermodynamic limit .
using the evidence available in the literature @xcite , let us now consider what we can expect when @xmath26 . the first important feature to note is that the isotropic de phase is replaced by an anisotropic phase in which the height of the end point of the polymer scales linearly with @xmath5 ; we denote this phase as the _ stretched _ phase .
consequently the transition from stretched to adsorbed phases becomes first - order @xcite .
likewise , at least in three dimensions @xcite , the transition from the vertically stretched phase to the collapsed phase also becomes first - order .
this implies that the multicritical point ( where for @xmath18 the de , ae and dc phases meet ) is now a triple point : the meeting of three first order lines .
the transition from the ae to ac phases should not be effected by the application of a small force as it acts in a direction perpendicular to the driving phenomenon of planar collapse . to a polymer in
the 2-layer adsorbed collapsed phase._,width=302 ] finally , it is intriguing to ask what happens to the layering phases observed in @xcite .
one can imagine that the force simply extends a vertical ` tail ' from a layered block ( see figure 3 ) and that as the force is increased the monomers are peeled off one at a time with corresponding micro - transitions @xcite for each monomer pulled until a vertical rod is achieved .
instead we see at some point a sharp first order transition between the highly stretched vertical rod and a layered system with short tail . in figures 2(b ) and 2(c ) we show plots of the logarithm of the maximum eigenvalue of the matrix of second derivatives in the variables @xmath19 and @xmath20 of @xmath24 at fixed @xmath27 ( as in figure 2(a ) ) but at values of @xmath27 being @xmath28 and @xmath29 .
it is clear that as @xmath27 is increased the stretched phase that occurs for small @xmath19 and @xmath20 expands while the positions of the other phases and transitions move little .
we immediately note that these plots _ do not _ tell the whole story since physically one is usually interested in fixing the force @xmath3 rather than @xmath27 : fixing @xmath27 implies that the force applied goes to zero at low temperatures .
it is for this reason that the re - entrant behaviour for absorbing polymers @xcite is not seen directly in these plots .
however , re - entrant behaviour does occur and occurs for _ any _ ratio of surface to bulk energies .
let us now consider the more traditional force - temperature diagram and return to this point . in figure 4 a plot of the force @xmath30 needed to pull a polymer from the wall for a ranges of temperatures and a parameter @xmath31 which measures the relative strength of the surface ( wall ) interaction to the self - interaction .
needed to pull a polymer from the surface against temperature @xmath13 and a parameter @xmath31 .
the parameter @xmath31 controls the relative strength of wall attraction and self - attraction with @xmath32 and @xmath33 .
the limiting cases of surface desorption and of pulling a collapsed polymer are easily visible in the plot for @xmath34 and @xmath35 , respectively._,width=325 ] we have parameterised the energies of surface and self - attraction as @xmath32 and @xmath36 respectively .
using this parameterisation for @xmath37 gives the whole range of attractive activities : the ratio of surface to bulk activities is given as @xmath38 and so is constant for fixed @xmath31 .
for @xmath35 we have @xmath39 and @xmath40 which corresponds to pure self - attraction while the other boundary of the parameter space with @xmath34 gives @xmath41 and @xmath42 which is the pure surface adsorption case .
this extends the diagrams given in @xcite in which only adsorption is considered . if a force smaller than @xmath43 is applied the polymer is in the phase appropriate to the value of @xmath31 : either collapsed or adsorbed or both .
on the other hand for forces larger than @xmath43 the polymer is in the ` stretched ' phase .
we immediately note that the reentrant behaviour observed in the adsorption - only case @xcite persists for all @xmath31 . fixing the force to be at a value slightly larger than the zero temperature critical force and then increasing the temperature leads to transitions from the stretched state to a non - stretched phase and back again to the stretched state .
this arises due to the entropy of the zero temperature state ; one can easily extend the arguments in @xcite to demonstrate that re - entrant behaviour can occur even when the zero - temperature configuration of the non - stretched state is a hamiltonian ( fully compact ) walk in a cube rather than a totally adsorbed polymer .
let us discuss the re - entrant behaviour returning to the ( @xmath19,@xmath20,@xmath27)-parametrisation .
for fixed @xmath2 , @xmath1 , and @xmath3 , changing the temperature @xmath13 implies moving on a ray in the ( @xmath19,@xmath20,@xmath27)-space . at high temperatures the system is in the stretched phase near the origin . at low temperatures ,
the state of the system depends on the choice of @xmath2 , @xmath1 , and @xmath3 . for very large @xmath3
the system remains stretched at all temperatures , whereas for very small @xmath3 and low temperatures , the system is in a layered phase .
the re - entrant behaviour manifests itself in the following way : there is a range of @xmath3 for which the system , when moving along a ray in the ( @xmath19,@xmath20,@xmath27)-space ( ie . for fixed @xmath2 , @xmath1 , and @xmath3 ) , changes from a stretched state near the origin to a layered one at intermediate temperatures and then changes back to a stretched state as @xmath19 , @xmath20 , and @xmath27 become larger .
if the critical force is zero then the curve in the @xmath44 plane corresponds to the phase boundary of the de phase with the apex of the curve around @xmath45 being the location of the multicritical point . on the other hand for @xmath46
there is a kink in the function @xmath47 at exactly @xmath48 which is a consequence of the first order point coming from the transition from sag / layer phases from small @xmath31 to the ac phase for larger @xmath31 .
there is the appearance of a kink joining the multicritical point to the zero temperature transition which is presumably a finite temperature effect of the transition to the ac phase . in this paper
we have studied how the phase diagram of a self - attracting polymer that is also attracted and tethered to a flat wall changes as a vertical force is applied to the un - tethered end of the polymer .
we have accomplished this using a flat histogram monte carlo simulation that is capable of studying the whole range of microscopic energies , temperature and polymer length up to a maximum of @xmath49 monomers .
we demonstrate that re - entrant behaviour occurs at low temperature and for a range of forces for _ all _ relative strengths of self - attraction and surface attraction . for small forces
we have found that only the transition boundary of the `` stretched '' phase moves with increasing force while the rest of the phase diagram is relatively unchanged . in contradiction to what may be expected we have found that the novel layering meta - phases found for large but finite polymer length are unaffected by small forces . |
constructing a hyperkhler metric on the total space @xmath0 of the cotangent bundle to a khler manifold @xmath1 is an old problem , dating back to the very first examples of hyperkhler metrics given by e. calabi in @xcite .
since then , many people have obtained a lot of important results valid for manifolds @xmath1 in this or that particular class ( see , for example , the papers @xcite , @xcite , @xcite , @xcite , @xcite ) .
finally , the general problem has been more or less solved a couple of years ago , independently by b. feix @xcite and by the author @xcite .
the metrics constructed in @xcite and @xcite are the same .
in fact , this metric satisfies an additional condition which makes it essentially unique which justifies the use of the term `` canonical metric '' .
but the approaches in @xcite and @xcite are very different .
feix s method is very geometric in nature ; it is based on a direct description of the associated twistor space .
the approach in @xcite is much farther from geometric intuition .
however , it seems to be more likely to lead to explicit formulas .
unfortunately , the paper @xcite is 100 pages long , and it is not written very well .
some of the proofs are not at all easy to understand and to check .
the exposition is sometimes canonical to the point of obscurity .
recently the author has been invited to give a talk on the results of @xcite at the second quaternionic meeting in rome .
it was a good opportunity to revisit the subject and to streamline and simplify some of the proofs .
this paper , written for the proceedings volume of the rome conference , is an attempt to present the results of @xcite in a concrete and readable form .
the exposition is parallel to @xcite but the paper is mostly independent .
compared to @xcite , the emphasis in the present paper has been shifted from canonical but abstract constructions to things more explicit and down - to - earth .
the number of definitions is reduced to the necessary minimum .
i have also tried to give a concrete geometric interpretation to everything that admits such an interpretation . in a sense
, the exposition as compared to @xcite intentionally goes to the other extreme .
thus the present paper is not so much a replacement for @xcite but rather a companion paper the same story told in a different way .
i should note that the canonical hyperkhler metric on @xmath0 has one important defect
namely , it is defined only in an open neighborhood @xmath2 of the zero section @xmath3 .
this raises a very interesting and difficult `` convergence problem '' .
one would like to describe the maximal open subset @xmath4 where the canonical metric is defined , and to say when @xmath5 is the whole total space @xmath0 .
unfortunately , very little is known about this .
in fact , in the present paper we even restrict ourselves to giving the formal germ of the canonical metric near @xmath6 .
the fact that this formal germ converges to an actual metric at least on an open subset @xmath2 is proved in the last section of @xcite .
the proof is long and tedious but completely straightforward .
since i do not know how to improve it , i have decided to omit it altogether to save space .
this reader will find the precise statements of all the results in section [ deffs ] .
the last part of that section contains a brief description of the rest of the paper , and indicates the parallel places in @xcite , the differences in notation and terminology and so on .
the only thing in this paper which is completely new is the last section [ symmm ] .
it contains an explicit formula for the canonical metric ( or rather , for the canonical hypercomplex structure ) in the case when @xmath1 is an hermitian symmetric space .
the formula is similar to the general formula for symmetric @xmath1 obtained by o. biquard and p. gauduchon in @xcite .
* acknowledgments .
* i would like to thank the organizers of the rome quaternionic meeting for inviting me to this very interesting conference and for giving me an opportunity to present the results of @xcite .
part of the present work was done during my visit to ecole polytechnique in paris during the autumn of 1999 .
i have benefited a lot from the hospitality and the stimulating atmosphere of this institution .
i would like to thank p. gauduchon for inviting me to paris and for encouraging me to write up a streamlined version of @xcite .
the present paper owes a lot to discussions with o. biquard and p. gauduchon during my visit .
in particular , the last section is an attempt to compare @xcite with the results in their beautiful paper @xcite .
i am also grateful to m. verbitsky , who read the first draft of the manuscript and suggested several improvements .
to save space , we will assume some familiarity with hyperkhler and hypercomplex geometry . we only give a brief reminder .
the reader will find excellent expositions of the subject in @xcite , @xcite , @xcite , @xcite .
let @xmath7 be the algebra of quaternions .
a smooth manifold @xmath8 is called _ almost quaternionic _ if it is equipped with a smooth action of the algebra @xmath7 on the tangent bundle @xmath9 .
equivalently , one can consider a smooth action on the cotangent bundle @xmath10 . to fix terminology
, we will assume that @xmath7 acts on @xmath10 on the left .
an almost quaternionic manifold @xmath8 is called _ hypercomplex _ if it admits a torsion - free connection preserving the @xmath7-module structure on @xmath10 .
such a connection exists is unique .
it is called the _ obata connection _ of the hypercomplex manifold @xmath8 . a riemannian almost quaternionic manifold @xmath8 is called _ hyper - hermitian _ if the riemannian pairing satisfies @xmath11 where @xmath12 is the quaternion conjugate to @xmath13 .
a hyper - hermitian almost quaternionic manifold @xmath8 is called _ hyperkhler _ if the @xmath7-action is parallel with respect to the levi - civita connection @xmath14 . in other words
, @xmath8 must be hypercomplex , and the obata connection must be @xmath14 .
let @xmath8 be a almost quaternionic manifold .
every embedding @xmath15 from the field of complex numbers to the algebra of the quaternions induces an almost complex structure on @xmath8 . if @xmath8 is hypercomplex , then all these induced almost complex structures are integrable .
if @xmath8 is also hyperkhler , then all these complex structures are khler with respect to the metric
. throughout this paper it will be convenient to choose an embedding @xmath16 and an additional element @xmath17 such that @xmath18 with these choices , every left @xmath7-module @xmath19 defines a complex vector space @xmath20 and a map @xmath21 which satisfies @xmath22 we will call @xmath23 the _ main complex structure _ on the real vector space @xmath19 .
conversely , every pair @xmath24 of a complex vector space @xmath25 and a map @xmath21 which satisfies defines an @xmath7-module structure on the real vector space @xmath19 underlying @xmath25 . the map @xmath21 can be considered as an automorphism @xmath26 of the underlying real vector space .
this map induces a second complex structure on @xmath19
. we will call it the _ complementary _ complex structure and denote the resulting complex vector space by @xmath27 applying this to vector bundles , we see that an almost quaternionic manifold @xmath8 is the same as an almost complex manifold @xmath8 equipped with a smooth complex bundle map @xmath28 from the tangent bundle @xmath29 to its complex - conjugate bundle @xmath30 which satisfies
. it would be very convenient to have some way to know whether an almost quaternionic manifold @xmath8 is hypercomplex or hyperkhler without working explicitly with torsion - free connections .
for hypercomplex manifolds , the integrability condition is very simple .
an almost quaternionic manifold @xmath8 is hypercomplex if and only if both the main complex structure on @xmath8 and the complementary almost complex structure @xmath31 are integrable .
the simplest way to determine whether a hyper - hermitian almost quaternionic manifold @xmath8 is hyperkhler is to consider the complex - bilinear form @xmath32 on the tangent bundle @xmath29 given by @xmath33 where @xmath13 is the hermitian metric on @xmath8 . then insures that the form @xmath32 is skew - symmetric .
the manifold @xmath8 is hyperkhler if and only if both the @xmath34-form @xmath32 and the khler @xmath35 are closed .
alternatively , one can define a @xmath34-form @xmath36 for the complementary almost complex structure @xmath31 instead of the @xmath34-form @xmath32 for the main almost complex structure .
then @xmath8 is hyperkhler if and only if both @xmath32 and @xmath36 are closed .
moreover , it suffices to require that these @xmath34-forms are holomorphic , each in its respective almost complex structure on @xmath8 ( `` holomorphic '' here means that @xmath37 is a form of type @xmath38 ) . finally , if we already know that the manifold @xmath8 is hypercomplex , then it suffices to require that only one of the forms @xmath32 , @xmath36 is a holomorphic @xmath39-form .
if @xmath8 is a khler manifold equipped with a closed @xmath34-form @xmath32 , one can _ define _ a map @xmath40 by .
then @xmath8 is hyperkhler if and only if this map @xmath41 satisfies .
consider now the case when @xmath42 is the total space of the cotangent bundle to a khler manifold @xmath1 with the khler metric @xmath13 .
then @xmath8 carries a canonical holomorphic @xmath39-form @xmath32 .
moreover , the unitary group @xmath43 acts on @xmath8 by dilatations along the fibers of the projection @xmath44 , so that we have @xmath45 for every @xmath46 . using these data
, we can formulate the main result of @xcite as follows .
[ hk ] there exists a unique , up to fiber - wise automorphisms of @xmath47 , @xmath43-invariant khler metric @xmath13 on @xmath48 , defined in the formal neighborhood of the zero section @xmath49 , such that 1 .
@xmath13 restricts to the given khler metric on the zero section @xmath50 , and 2 .
the pair @xmath51 defines a hyperkhler structure on @xmath8 near @xmath50 .
if the khler metric @xmath13 on @xmath1 is real - analytic , then the formal canonical metric on @xmath8 comes from a real - analytic hyperkhler metric defined in an open neighborhood of @xmath52 .
the @xmath43-action on @xmath0 is very important for this theorem .
in fact , @xmath0 with this action belongs to a general class of hyperkhler manifolds equipped with a @xmath43-action , introduced in @xcite , @xcite .
[ comp ] we will say that a holomorphic @xmath43-action on a hyperkhler manifold @xmath8 is _ compatible with the hyperkhler structure _ if and only if 1 .
the metric @xmath13 on @xmath8 is @xmath43-invariant , 2 .
the holomorphic @xmath39-form @xmath32 satisfies , 3 .
the map @xmath40 satisfies @xmath53 it is easy to check using that every two of these conditions imply the third .
theorem [ hk ] can generalized to the following statement , somewhat analogous to the darboux - weinstein theorem in symplectic geometry .
[ darboux ] let @xmath8 be a hyperkhler manifold equipped with a regular compatible holomorphic @xmath43-action .
then there exists an open neighborhood @xmath54 of the @xmath43-fixed point subset @xmath55 and a canonical embedding @xmath56 such that the hyperkhler structure on @xmath5 is induced by means of the map @xmath57 from the canonical hyperkhler structure on @xmath0 . here _
regular _ is a certain condition on the @xmath43-action near the fixed points subset @xmath58 which is formulated precisely in definition [ reg ] ( roughly speaking , weights of the action on the tangent space @xmath59 at every point @xmath60 should be @xmath61 and @xmath62 ) . the map @xmath57 will be called the _ normalization map_. note that this theorem allows one to reformulate theorem [ hk ] so that the metric is indeed unique not just unique up to a fiber - wise automorphism of @xmath47 . to fix the metric , it suffices to require that the associated normalization map @xmath63 is identical .
metrics with this property will be called _
there is also a form of theorem [ hk ] for hypercomplex manifolds ( and it is this form which is the most important for @xcite all the other statements are obtained as its corollaries ) . to formulate it
, we note that out of the three condition of definition [ comp ] , the third one makes sense for almost quaternionic ( in particular , hypercomplex ) manifolds .
we will say that a holomorphic @xmath43-action on an almost quaternionic manifold @xmath8 is _ compatible with the quaternionic action _ if definition [ comp ] is satisfied .
let @xmath64 be the total space of the bundle @xmath65 complex - conjugate to the tangent bundle of the manifold @xmath1 .
then @xmath64 is a smooth manifold , and we have the canonical projection @xmath66 and the zero section @xmath67 .
the group @xmath43 acts on @xmath64 by dilatations along the fibers of the projection @xmath68 .
moreover , for any compatible hypercomplex structure on the @xmath43-manifold @xmath64 the corresponding obata connection @xmath69 induces a torsion - free connection @xmath70 on @xmath1 by the following rule @xmath71 the hypercomplex version of theorem [ hk ] is the following .
[ hc ] let @xmath1 be a complex manifold @xmath1 equipped with a holomorphic connection @xmath70 on the tangent bundle @xmath72 such that 1 .
@xmath70 has no torsion , and 2 .
the curvature of the connection @xmath70 is of type @xmath73 .
let @xmath74 be the total space of the complex - conjugate to the tangent bundle @xmath75 .
let the group @xmath43 act on @xmath8 by dilatations along the fibers of the projection @xmath76 .
then there exists a unique , up to a fiber - wise automorphism of @xmath47 , hypercomplex structure on @xmath8 , defined in the formal neighborhood of the zero section @xmath50 , such that the embedding @xmath77 and the projection @xmath76 are holomorphic and the obata connection on @xmath8 induces the given connection @xmath70 on @xmath1 .
moreover , if the connection @xmath70 is real - analytic , then the hypercomplex structure on @xmath8 is real - analytic in an open neighborhood @xmath54 of @xmath50 .
note that _ a priori _ there is no natural complex structure on the space @xmath74 ( we write @xmath78 instead of @xmath79 just to indicate the correct @xmath43-action ) .
therefore theorem [ hc ] in fact claims two things : firstly , there exists an integrable almost complex structure on @xmath8 , and secondly , there exists a map @xmath40 which extends it to a hypercomplex structure .
connections @xmath70 that satisfy the conditions of this theorem were called _ khlerian _ in @xcite ( see @xcite ) .
i would like to thank d. joyce for attracting my attention to his paper @xcite , where he uses the same class of connections to construct _ commuting _ almost complex structures on the total space @xmath75 of the tangent bundle to a complex manifold @xmath1 .
joyce calls these connection _ khler - flat_. theorem [ hc ] also admits a generalized darboux - like version in the spirit of theorem [ darboux ] ( see ( * ? ? ?
* proposition 4.1 ) ) ; we do not formulate it here to save space .
we will now give a brief outline of the remaining part of the paper .
in section [ nrm ] we consider an arbitrary @xmath43-equivariant hypercomplex manifold @xmath8 and construct the normalization map @xmath80 , thus proving theorem [ darboux ] .
this corresponds to ( * ? ? ?
* section 4 ) . what we call _ normalization _ here was called _ linearization _ in @xcite ; _ normalized _ corresponds to _ linear_. the terminology of @xcite has been changed because it was misleading : connections on the fibration @xmath81 that were called linear in @xcite are not linear in the usual sense of the word .
section [ hb ] introduces @xmath82-hodge structures and the so - called hodge bundles ( definition [ hb.def ] ) which are the basis of our approach to @xmath43-equivariant quaternionic manifolds @xmath8 .
this corresponds to ( * ? ? ?
* sections 2,3 ) .
proposition [ dol ] is a version of ( * ? ?
* proposition 3.1 ) .
in section [ hcon ] we turn to the case @xmath74 and introduce the so - called hodge connections ( definition [ hcon.def ] ) .
this corresponds to ( * ? ? ?
* section 5 ) . the proofs
have been considerably shortened .
there are also some new facts on the relation between our formalism and the objects one usually associates with hypercomplex manifolds .
in particular , lemma [ j.matrix ] is new .
section [ wa ] and section [ pf ] introduce the weil algebra @xmath83 of a complex manifold @xmath1 ( definition [ wa.def ] ) and then use it to prove theorem [ hc ] .
section [ wa ] contains the preliminaries ; section [ pf ] gives the proof itself .
this material corresponds to ( * ? ? ?
* sections 6 - 8 ) .
the approach has been changed in the following way .
keeping track of various gradings and bigradings and on the weil algebra presents considerable difficulties : when the proof of theorem [ hc ] is written down , the number of indices becomes overwhelming .
in @xcite we have tried to handle this by an auxiliary technical device called _ the total weil algebra _ ( @xcite ) .
it was quite a natural thing to do from the conceptual point of view .
unfortunately , the proof became more abstract than one would like .
here we have opted for the direct approach . to make things comprehensible ,
we rely on pictures ( figure [ fig.odd ] , figure [ fig.even ] ) which graphically represent the hodge diamonds of the relevant pieces of the weil algebra @xmath83 . finally , section [ metrics ] deals with things hyperkhler : we deduce theorem [ hk ] from theorem [ hc ] .
this corresponds to ( * ? ? ?
* section 9 ) .
we believe that the exposition has also been simplified , and the proofs are easier to check .
the last section [ symmm ] of this paper is new .
we try to illustrate our constructions by a concrete example of an hermitian symmetric space @xmath1 .
we obtain a formula similar to @xcite .
the last section of @xcite contains the proof of convergence of our formal series in the case when the khler manifold @xmath1 is real - analytic . in this paper , this proof is entirely omitted .
of all the statement formulated in the last section , the most straightforward one is the darboux - like theorem [ darboux ] and its hypercomplex version . in this section ,
we explain how to construct the normalization map @xmath57 .
most of the proofs are replaced with references to @xcite .
we begin with some generalities .
assume that the group @xmath43 acts smoothly on a smooth manifold @xmath8 .
for any point @xmath84 fixed by the action , we have an action of @xmath43 on the tangent space @xmath85 .
equivalently , we have the weight decomposition @xmath86 where @xmath46 acts on @xmath87 by multiplication by @xmath88 .
we will say that the fixed point @xmath89 is _ regular _ if the only non - trivial weight components @xmath87 correspond to weights @xmath90 . the subset @xmath58 of fixed points is a disjoint union of smooth submanifolds of different dimensions .
regular fixed points form a connected component of this set .
denote this component by @xmath50 .
let @xmath91 be the differential of the @xmath43-action that is , the vector field on @xmath8 which gives the infinitesimal action of the generator @xmath92 of the lie algebra of the group @xmath43 . assume further that @xmath8 is a complex manifold and that @xmath43 preserves the complex structure .
say that a point @xmath84 is _ stable _ if for any @xmath93 , @xmath94 there exists @xmath95 , and moreover , the limit @xmath96 also exist . when the limit point @xmath97 does exist , it is obviously fixed by @xmath43 .
say that a stable point @xmath84 is _ regular stable _ if the limit point is a regular fixed point , @xmath98 .
regular stable points form an open subset @xmath99 .
[ reg ] a complex @xmath43-manifold @xmath8 is called is _ regular _ if every point @xmath84 is regular stable , @xmath100 .
for example , the total space of an arbitrary complex vector bundle on an arbitrary complex manifold is regular ( if @xmath43 acts by dilatations along the fibers ) .
the submanifold of regular fixed points in this example is the zero section .
when the @xmath43-manifold @xmath8 is hypercomplex , we will say that it is regular if it is regular in the main complex structure . in this case
, the subset @xmath50 of regular fixed points is a complex submanifold .
setting @xmath101 defines a @xmath43-invariant projection @xmath76 .
the projection @xmath76 is a holomorphic submersion .
we can now define the normalization map .
consider the exact sequence @xmath102 of tangent bundles associated to the submersion @xmath76 .
the differential @xmath91 of the @xmath43-action is a vertical holomorphic vector field on @xmath8 , @xmath103 . applying the operator @xmath40 to @xmath91
gives a section of the bundle @xmath30 .
we can project this section to obtain a section @xmath104 of the pullback bundle @xmath105 .
but such a section tautologically defines a map @xmath80 from @xmath8 to the total space of the complex bundle @xmath78 on the manifold @xmath1 .
since @xmath91 is @xmath43-invariant , and @xmath41 is of weight @xmath62 with respect to the @xmath43-action , the section @xmath106 is also of weight one .
this means that the associated map @xmath107 is @xmath43-equivariant
. we will call it the _ normalization map _ for the regular hypercomplex @xmath43-manifold @xmath8 .
the normalization map @xmath80 is an open embedding .
this lemma essentially reduces theorem [ darboux ] to theorem [ hk ] .
a particular case occurs when @xmath74 is itself the total space of the complex - conjugate to the tangent bundle on a complex manifold @xmath1 or , more generally , an open @xmath43-invariant neighborhood @xmath108 of the zero section @xmath109 . as noted above , in this case the zero section @xmath110 coincides with the subset of fixed points .
therefore the normalization map @xmath111 is an open embedding from @xmath5 into @xmath64 , possibly different from the given one .
the hypercomplex structure on @xmath108 is called _ normalized _ if the normalization map @xmath112 coincides with the given embedding .
( in particular , when @xmath113 is the whole total space , the normalization map must be identical . ) to prove theorem [ hk ] and theorem [ hc ] , it is sufficient to be able to classify all normalized hypercomplex structures on the @xmath43-manifold @xmath114 and germs of such structures near the zero section @xmath109
. it will be convenient to slightly rewrite the normalization condition .
namely , the identity map @xmath115 defines a section on @xmath116 of the pullback bundle @xmath105 .
we will denote this section by @xmath117 and call it the _ tautological section_. then a hypercomplex structure on @xmath54 is normalized if and only if we have @xmath118
the first step in the proof of theorem [ hc ] is to give a workable description of hypercomplex structures on the total space @xmath74 . for this
we use the language of @xmath82-hodge structures .
recall that an _ @xmath82-hodge structure @xmath25 of weight @xmath119 _ is by definition a real vector space @xmath19 equipped with a grading @xmath120 such that @xmath121 equivalently , instead of the grading one can consider a @xmath43 action on @xmath25 defined by @xmath122 then becomes @xmath123 when the weight @xmath119 is equal to @xmath62 , this equation on the complex conjugation map becomes precisely .
the difference between the complex conjugation map and the map @xmath41 used to define quaternionic structures is that the first one is an involution , @xmath124 , while for the second one we have @xmath125 .
nevertheless , we will exploit the similarity between them to describe quaternionic actions via hodge structures .
to do this , we use the following trick .
let @xmath25 be an @xmath82-hodge structure of weight @xmath62 , and consider the map @xmath126 given by the action of @xmath127 in other words , let @xmath128 then the map @xmath129 still satisfies , and also holds .
this turns @xmath25 into a left @xmath7-module .
conversely , every left @xmath7 module @xmath24 equipped with a @xmath43-action on @xmath25 such that @xmath41 satisfies defines an @xmath82-hodge structure of weight @xmath62 . to use this for a description of hypercomplex structures on manifolds , we introduce the following .
[ hb.def ] let @xmath8 be a smooth manifold equipped with an action of the group @xmath43 , and let @xmath130 be the action of the element @xmath131 .
hodge bundle @xmath132 of weight @xmath119 _ on @xmath8 is by definition a @xmath43-equivariant complex vector bundle @xmath132 equipped with a @xmath43-equivariant isomorphism @xmath133 such that @xmath134 .
here @xmath135 is the bundle complex conjugate to @xmath132 , whose @xmath43-equivariant structure is twisted by tensoring with the @xmath62-dimensional representation @xmath136 of the group @xmath43 of weight @xmath119 , @xmath137 when the @xmath43-action on the manifold @xmath8 is trivial , a weight-@xmath119 hodge bundle @xmath132 on @xmath8 is just the bundle of @xmath82-hodge structures of weight @xmath119 in the obvious sense . in particular ,
if @xmath8 is an almost complex manifold , then the bundle @xmath138 of all complex - valued @xmath119-forms on @xmath8 is a hodge bundle of weight @xmath119 .
when the @xmath43-action on @xmath8 is no longer trivial , every bundle @xmath132 of @xmath82-hodge structures on @xmath8 still defines a hodge bundle . thus @xmath138 is still a weight-@xmath119 hodge bundle .
but this hodge bundle structure is not interesting , since it does not take into account the natural @xmath43-action on @xmath138 . assuming that the @xmath43-action preserves the almost complex structure on @xmath8 , we define instead a hodge bundle structure on @xmath139 by keeping the usual complex conjugation map and twisting the @xmath43-action so that @xmath140 as a @xmath43-equivariant vector bundle
it is easy to check that this indeed defines on @xmath139 a hodge bundle structure of weight @xmath62 .
assume now that the almost complex manifold @xmath8 is equipped with an almost quaternionic structure which is compatible with the @xmath43-action .
then the complex vector bundle @xmath141 of @xmath142-forms on @xmath8 already has a structure of a hodge bundle of weight @xmath62 .
the @xmath43-action for this structure is the natural one , and the complex conjugation map is induced by the map @xmath143 via .
this hodge bundle structure completely determines the quaternionic action .
more precisely , for every smooth @xmath43-manifold @xmath8 , every hodge bundle @xmath132 of weight @xmath62 whose underlying real vector bundle @xmath144 is identified with the cotangent bundle @xmath145 comes from a unique compatible almost quaternionic structure @xmath8 .
the natural embedding @xmath146 is not a map of hodge bundles
it is @xmath43-equivariant , but it obviously does not commute with the complex conjugation map . this can be corrected . to do this
, one has to look at the picture in a different way ( which will turn out to be very useful ) .
return for a moment to linear algebra .
let @xmath19 be a left @xmath7-module , and let @xmath25 , @xmath27 be the complex vector spaces obtained from @xmath19 by the main and the complementary complex structures .
consider the complex vector space @xmath147 .
this vector space does not depend on the @xmath7-action on @xmath19 .
given an @xmath7-action , we have the main and the complementary complex structure operators @xmath148 and @xmath26 and the associated eigenspace decompositions @xmath149 since the operators @xmath150 and @xmath151 anti - commute , these decompositions are distinct : we have @xmath152 in particular , the composition @xmath153 of the canonical embedding in and the canonical projection in is an isomorphism
. we will call it the _ canonical isomorphism between the main and the complementary complex structures_. on the level of the real vector space @xmath19 , the map @xmath154 is induced by a non - trivial automorphism @xmath155 ( in fact it is the action of the element @xmath156 ) .
conjugation with this map interchanges the operators @xmath150 and @xmath151 . return now to the case of an almost quaternionic manifold @xmath8 .
then we claim that the complementary almost complex structure operator @xmath157 is a map of hodge bundles .
indeed , it commutes with the complex conjugation map by definition .
therefore it suffices to show that it is @xmath43-equivariant on @xmath146 .
but for every @xmath158 we have @xmath159 , and the map @xmath41 is of weight @xmath62 .
thus the operator @xmath151 is indeed @xmath43-equivariant ( recall the twisting of the @xmath43-action in the definition of the hodge bundle structure on @xmath139 ) .
since the endomorphism @xmath157 is a map of hodge bundles , its eigenbundles @xmath160 are hodge subbundles .
therefore we obtain a canonical weight-@xmath62 hodge bundle structure on the bundle @xmath161 of @xmath142-forms for the _ complementary _ almost complex structure on @xmath8 .
this hodge bundle is not a new one .
indeed , the canonical isomorphism @xmath162 defined in is @xmath43-equivariant it is obtained as a composition of @xmath43-equivariant maps .
moreover , it is very easy to check that @xmath154 commutes with the complex conjugation
. thus @xmath163 as hodge bundles . but
the projections @xmath164 are different .
only the second one is a hodge bundle map .
all this linear algebra is somewhat tautological , but it becomes useful when we consider the integrability conditions on an almost quaternionic quaternionic structure .
the real advantage of @xmath82-hodge structures over @xmath7-modules is the presence of higher weights .
namely , the category of @xmath7-modules admits no natural tensor product . on the other hand ,
the category of @xmath82-hodge structures and the category of hodge bundles are obviously tensor categories .
thus , for example , the weight-@xmath62 hodge bundle structure on the cotangent bundle @xmath139 induces a weight-@xmath119 hodge bundle structure on the bundle @xmath165 of @xmath119-forms . to make use of these higher - weight hodge bundles , we need a convenient notion of maps between hodge bundles of different weights . a bundle map ( or , more generally , a differential operator ) @xmath166 between hodge bundles @xmath132 , @xmath167 of weights @xmath168 , @xmath169 is called _ weakly hodge _ if it commutes with the complex conjugation map and admits a decomposition @xmath170 where @xmath171 is of weight @xmath172 with respect to the @xmath43-action in other words , @xmath173 is @xmath43-equivariant when considered as a map @xmath174 we see that non - trivial weakly hodge maps between hodge bundles @xmath132 , @xmath167 exist only when their weights satisfy @xmath175 .
when the @xmath43-action on the manifold @xmath8 is trivial , the hodge bundles @xmath132 and @xmath167 come from bundles of @xmath82-hodge structures on @xmath8 , and the decomposition @xmath176 of a weakly hodge map @xmath177 is simply the hodge type decomposition , @xmath178 if the @xmath43-action is not trivial , but preserves an almost complex structure on @xmath8 , the de rham differential @xmath179 is weakly hodge .
if the almost complex structure is integrable , then the same is true for the de rham differential @xmath180 for every @xmath181 .
when the @xmath43-manifold is almost quaternionic , we have a hodge bundle structure of weight @xmath62 on the bundle @xmath161 .
then the dolbeault differential @xmath182 is weakly hodge .
indeed , it is the composition of the weakly hodge de rham differential and the projection which is a hodge bundle map .
let @xmath183 be the decomposition for the weakly hodge map @xmath184 .
looking at the definition of the canonical isomorphism @xmath185 , we see that the component @xmath186 in the decomposition coincides with the dolbeault differential @xmath187 for the main complex structure .
assume now that the almost quaternionic @xmath43-manifold @xmath8 is hypercomplex .
the bundle @xmath188 of @xmath189-forms on @xmath31 is a hodge bundle of weight @xmath119 for every @xmath181 , and we have the dolbeault differential @xmath190 since the projections @xmath191 are hodge bundle maps for every @xmath181 , this dolbeault differential is weakly hodge .
it turns out that this is a sufficient integrability condition for an almost quaternionic manifold equipped with a compatible @xmath43-action .
[ dol ] let @xmath8 be an almost quaternionic manifold equipped with a compatible @xmath43-action .
assume that the dolbeault differential @xmath192 extends to a weakly hodge derivation @xmath193 of the algebra @xmath194 satisfying @xmath195 . then the manifold @xmath8 is hypercomplex . _
_ it suffices to prove that both the main and the complementary almost complex structures on @xmath8 are integrable . for this , it is enough to prove that the dolbeault differentials @xmath196 extend to square - zero derivations of the exterior algebra @xmath194 .
the differential @xmath184 extends by assumption . to extend @xmath197 ,
take the component @xmath198 of the weakly hodge map @xmath193 .
then @xmath199 is a component in the decomposition of the weakly hodge map @xmath200 . since @xmath201
, we also have @xmath202 .
we will now say a couple of words about hyperkhler manifolds and hodge bundles .
let @xmath8 be an almost quaternionic @xmath43-manifold .
then every riemannian metric on @xmath8 defines a @xmath34-form @xmath203 .
it turns out that if the metric is hyper - hermitian and @xmath43-invariant , then in terminology of @xcite , the form @xmath36 _ of @xmath154-type @xmath73_. this means the following .
consider the form @xmath36 as a bundle map @xmath204 where @xmath205 is the trivial bundle on @xmath8 equipped with the so - called hodge - tate @xmath82-hodge structure of weight @xmath206 that is , the complex conjugation map on @xmath205 is minus the complex conjugation map on @xmath82 , and the @xmath43-equivariant structure is twisted by @xmath62 .
the form @xmath36 is said to be of @xmath154-type @xmath73 if the map is a hodge bundle map . conversely , every @xmath34-form @xmath203 of @xmath154-type @xmath73 on an almost quaternionic @xmath43-manifold @xmath8 which satisfies a positivity condition @xmath207 defines a @xmath43-invariant hyper - hermitian metric on @xmath8 .
( see @xcite , but the proof is almost trivial . ) if @xmath8 is hypercomplex , then , as indicated in section [ deffs ] , the metric corresponding to such a form @xmath36 is hyperkhler if and only if the form @xmath36 is holomorphic , @xmath208 . in fact , using the @xmath43-action on @xmath8 , one can even drop the integrability condition .
indeed , if a form @xmath36 of @xmath154-type @xmath73 on an almost quaternionic @xmath43-manifold @xmath8 satisfies @xmath209 , then it also must satisfy @xmath210 the canonical endomorphism @xmath211 , being the conjugation with a quaternion , preserves up to a coefficient the metric associated to @xmath36 and interchanges the almost complex structure operators @xmath150 and @xmath151 .
therefore it sends @xmath36 to a form proportional to @xmath32 .
then @xmath212 implies that not only @xmath36 is holomorphic , but that @xmath32 is holomorphic as well .
this proves that @xmath8 is hyperkhler ( _ a posteriori _ , also hypercomplex ) .
we will never need nor use this argument .
an interested reader will find details in @xcite .
we will now restrict our attention to the case when the @xmath43-manifold @xmath8 is the total space @xmath64 of the complex - conjugate to the tangent bundle of a complex manifold @xmath1 . in this case ,
proposition [ dol ] is really useful , because it turns out that the hodge bundle algebra @xmath213 does not depend on an almost quaternionic structure on @xmath8 . to see this ,
denote by @xmath214 the canonical projection , and let @xmath215 be the codifferential of the map @xmath68 .
then for every compatible hypercomplex structure on @xmath8 , we have canonical bundle maps @xmath216 assume that the manifold @xmath8 is a equipped with a hypercomplex structure satisfying the conditions of theorem [ hc ] namely , assume that the projection @xmath76 and the zero section @xmath217 are holomorphic maps .
then the codifferential @xmath218 is obviously a map of hodge bundles .
however , we also have the following .
[ isom ] the composition map @xmath219 is an isomorphism of hodge bundles in an open neighborhood of the zero section @xmath50 . from this point on , it will be convenient to only consider germs of hypercomplex structures defined near the zero section @xmath52 .
in other words , we replace @xmath74 with an unspecified and shrinkable @xmath43-invariant open neighborhood of the zero section .
since we are only interested in hypercomplex structures on @xmath8 that satisfy the conditions of theorem [ hc ] , lemma [ isom ] shows that no matter what the particular hypercomplex structure on @xmath8 is , we can _ a priori _ canonically identify the hodge bundle @xmath161 with the pullback bundle @xmath220 .
the only thing that depends on the hypercomplex structure is the derivation @xmath221 .
to formalize the situation , we introduce the following .
[ hcon.def ] 1 . a _ @xmath222-valued connection _
@xmath223 on @xmath47 is a bundle map @xmath224 which splits the codifferential @xmath225 of the projection @xmath76 .
2 . the _ derivation @xmath226 associated _ to a @xmath222-valued connection @xmath223 is the composition @xmath227 of @xmath223 with the de rham differential @xmath228 .
3 . a _ hodge connection _
@xmath223 on @xmath47 is a @xmath222-valued connection such that the associated derivation @xmath229 is weakly hodge .
a hodge connection @xmath223 on @xmath47 is called _ flat _ if the associated derivation @xmath184 extends to a weakly hodge derivation @xmath230 of the pullback of the de rham algebra @xmath231 , and the extended map @xmath184 satisfies @xmath195 . of course , a hodge connection @xmath223 is completely defined by the associated derivation @xmath226 .
conversely , an arbitrary weakly hodge derivation @xmath232 comes from a hodge connection if and only if we have @xmath233 for every smooth function @xmath234 .
say that a derivation @xmath232 is _ holonomic _ if the associated hodge connection @xmath223 induces an isomorphism @xmath235 between the real cotangent bundle @xmath145 and the real bundle underlying the complex vector bundle @xmath220 .
then proposition [ dol ] and lemma [ isom ] show that hypercomplex structures on @xmath8 satisfying the conditions of theorem [ hc ] are in one - to - one correspondence with hodge connections on @xmath47 whose associated derivations are holonomic .
indeed , the isomorphism induces a hodge bundle structure of weight @xmath62 on the cotangent bundle @xmath145 , hence an almost quaternionic structure on @xmath8 . applying proposition [ dol ]
, we see that flatness of the hodge connection is equivalent to the integrability of this almost quaternionic structure
. all derivations @xmath184 that we will work with will be automatically holonomic this will turn out to be a consequence of the normalization condition ( see lemma [ d.norm ] ) . the name `` hodge connection '' invokes the notion of a connection on a smooth fibration .
this is somewhat misleading .
the problem is that a hodge connection @xmath236 is only defined over @xmath222 .
so it has a real part @xmath237 and an imaginary part @xmath238 .
the real part @xmath239 is indeed a connection on the fibration @xmath76 in the usual sense that is , it defines a smooth splitting @xmath240 of the real cotangent bundle @xmath145 into a horizontal and a vertical part .
the vertical part @xmath241 is canonically isomorphic to the relative cotangent bundle @xmath242 .
the imaginary part @xmath238 , on the other hand , vanishes on the subbundle @xmath243 and defines therefore a certain map @xmath244 from the relative cotangent bundle @xmath242 to the pullback bundle @xmath245 .
since @xmath8 is an open subset in @xmath64 , we can canonically identify the bundle @xmath246 with the pullback bundle @xmath245 . under this identification ,
the map @xmath247 becomes an endomorphism of the bundle @xmath245 .
typically , when a hodge connection @xmath223 comes from a hypercomplex structure on @xmath8 , the associated real connection @xmath237 on @xmath47 is _ not _ flat .
it is only the sum @xmath248 which is flat but it is no longer a real connection .
this situation is somewhat similar to what happens in c. simpson s theory of higgs bundles and harmonic metrics ( @xcite ) .
the presence of a non - trivial imaginary part @xmath238 seems to imply a contradiction .
indeed , a hodge connection @xmath249 must by definition be compatible with the hodge bundle structures in particular , it must commute with the complex conjugation map .
but this is different from `` real '' .
the reason for this is the twist by the involution @xmath130 that we have introduced in the definition of a hodge bundle .
this can be seen clearly if instead of the connection @xmath223 one considers the associated derivation @xmath184 .
the derivation @xmath250 satisfies @xmath251 for any @xmath222-valued smooth function @xmath252 .
take the decomposition @xmath253 into the odd and the even part with respect to the involution @xmath254 , so that every function @xmath255 we have @xmath256 there is no reason for either one of these parts to vanish .
but shows that @xmath257 where @xmath258 is the de rham differential .
the imaginary part @xmath238 of a hodge connection @xmath223 on @xmath47 or rather , the associated map @xmath247 by itself has a very direct geometric meaning in terms of the hypercomplex structure on @xmath8 given by @xmath223 . to describe it ,
consider the splitting @xmath259 given by the real part @xmath237 and identify @xmath260 .
[ j.matrix ] the operator @xmath261 of the hypercomplex structure given by @xmath223 can be written with respect to the decomposition as the matrix @xmath262 where @xmath263 is the bundle endomorphism .
_ since @xmath264 , it suffices to prove that for every @xmath62-form @xmath265 which lies in the horizontal part of , the @xmath62-form @xmath266 is vertical , that is , @xmath267 and moreover , that we have @xmath268 let @xmath269 be such a form . by definition ,
the kernel @xmath270 of the projection @xmath271 is the subbundle @xmath272 of @xmath273-forms for the complementary complex structure on @xmath8 .
therefore we have @xmath274 which means that @xmath275 since @xmath276 , equations and are the real and the imaginary parts of this equality . in keeping with the general philosophy of this section
, we will use the formula for @xmath41 given by lemma [ rj ] to express the bundle endomorphism @xmath277 entirely in terms of operators on the algebra @xmath278 . to do this , consider the tautological section of the pullback tangent bundle @xmath279 , and let @xmath280 be the operator given by contraction with this tautological section .
thus @xmath117 vanishes on functions , and for every @xmath62-form
@xmath281 the function @xmath282 is just @xmath269 considered as a fiberwise - linear function on the total space @xmath114 .
[ rj ] for any @xmath62-form @xmath283 we have @xmath284 _ proof .
_ by , the right - hand side is equal to @xmath285 .
since the projection @xmath286 vanishes on the subbundle @xmath287 , this expression depends only the relative @xmath62-form @xmath288 obtained from the @xmath62-form @xmath289 by the projection @xmath290 . but
@xmath291 is precisely the image of the form @xmath292 under the canonical isomorphism @xmath293
. we will now rewrite in the same spirit the normalization condition on the hypercomplex structure on @xmath8 associated to @xmath184 .
for this we need to extends the canonical isomorphism @xmath294 to an algebra isomorphism @xmath295 .
then the map @xmath280 becomes the contraction with the relative euler vector field ( that is , the differential of the @xmath296-action by dilatations along the fibers of the projection @xmath76 ) .
the normalization condition involves a different vector field
namely , the differential @xmath91 of the standard action of the group @xmath43 .
it will be more convenient now to multiply it by @xmath297 ( or , equivalently , to change the generator of the lie algebra of the circle @xmath43 from @xmath92 to @xmath298 ) .
denote by @xmath299 the contraction with the vertical vector field @xmath300 .
the operators @xmath301 and @xmath117 are related by @xmath302 where @xmath303 is the complex structure operator in other words , @xmath304 [ d.norm ] let @xmath223 be a hodge connection on @xmath47 , and let @xmath305 be the even component of the associated derivation @xmath226 .
the hypercomplex structure on @xmath8 given by @xmath223 satisfies the normalization condition if and only for every @xmath62-form @xmath283 we have @xmath306 where @xmath307 .
moreover , if this is the case , then the hodge connection @xmath223 is holonomic .
_ it suffices to check by evaluating both sides on every @xmath62-form @xmath292 .
moreover , it is even enough to check it for forms of the type @xmath308 , where @xmath283 is a @xmath62-form on @xmath1 .
let @xmath269 be such a form .
we have to check that @xmath309 by lemma [ j.matrix ] this is equivalent to @xmath310 and by lemma [ rj ] this can be further rewritten as @xmath311 replacing @xmath282 with @xmath312 gives precisely the first claim of the lemma .
to prove the second claim , we have to show that the map @xmath313 is surjective . by definition , on the second term @xmath243 in the splitting we have @xmath314 .
therefore it suffices to prove that @xmath315 is surjective . but by and the first claim of the lemma , this is the inverse map to @xmath316
the final preliminary step in the proof of theorem [ hc ] is to reduce it from a question about the total space @xmath74 of the complex - conjugate to the tangent bundle on @xmath1 to a question about the manifold @xmath1 .
to do this .
we introduce the following . [ wa.def ] the _ weil algebra _ @xmath83 of a complex manifold @xmath1 is the algebra on @xmath1 defined by @xmath317 where @xmath66 is the canonical projection .
this requires an explanation indeed , for a vector bundle @xmath132 on @xmath64 , the direct image sheaf @xmath318 _ a priori _ is not a sheaf of sections of any vector bundle on @xmath1 .
we have to consider a smaller subsheaf . from now on and until theorem [ hc ] is proved , we will be interested not in hypercomplex structures on the total space @xmath64 but in their formal taylor decompositions in the neighborhood of the zero section @xmath319
. therefore it will be sufficient for our purposes to define the direct image @xmath318 as the sheaf of sections of the bundle @xmath132 on @xmath64 _ which are polynomial along the fibers of the projection @xmath66_. formal germs of bundle maps on @xmath132 will give formal series of maps between the corresponding direct image bundles .
having said this , we can explicitly describe the weil algebra @xmath83 .
our first remark is that @xmath320 is canonically a hodge bundle on @xmath1 of weight @xmath119 .
moreover , since the @xmath43-action on @xmath1 is trivial , hodge bundles on @xmath1 are just bundles of @xmath82-hodge structures in the usual sense .
thus we have a hodge type bigrading @xmath321 and a canonical real structure on every one of the complex vector bundles @xmath320 .
the projection formula show that for every @xmath119 we have a canonical isomorphism @xmath322 these isomorphisms are compatible with the hodge structures and with multiplication .
the degree-@xmath61 hodge bundle @xmath323 is a symmetric algebra freely generated by the bundle @xmath324 of functions on @xmath64 linear along the fibers of @xmath325 .
the complex vector bundle @xmath324 is canonically isomorphic to the bundle @xmath326 of @xmath62-forms on @xmath1 .
however , the hodge structures on these bundles are different .
the hodge type grading on @xmath324 is given by @xmath327 where @xmath328 and @xmath329 the grading is the same as on @xmath326 but graded pieces are assigned different weights .
moreover , the complex conjugation map on @xmath324 is _ minus _ the complex conjugation map on @xmath326 . this is the last vestige of the twist by the involution @xmath330 in definition [ hb.def ] . to simplify notation , denote by @xmath331 the @xmath119-th symmetric power of the hodge bundle @xmath324 .
then we have @xmath332 and the weil algebra @xmath333 is the free graded - commutative algebra generated by @xmath324 and @xmath326 ( where , contrary to notation , @xmath324 is placed in degree @xmath61 ) .
it will be convenient to introduce another grading on the weil algebra @xmath83 by assigning to both of the generator bundles @xmath324 , @xmath326 degree @xmath62 . we will call it _ augmentation grading _ and denote by lower indices , so that we have @xmath334 the augmentation grading corresponds to the taylor decomposition near the zero section @xmath109 .
namely , every formal germ near @xmath109 of a flat hodge connection on @xmath335 induces a formal series @xmath336 of algebra bundle derivations @xmath337 where each of the derivations @xmath338 is weakly hodge and has augmentation degree @xmath119 .
their ( formal ) sum satisfies @xmath339 conversely , every formal series induces a ( formal germ of a ) weakly hodge derivation @xmath340 on the total space @xmath64 .
this derivation comes from a flat hodge connection if and only if we have @xmath341 for every function @xmath342 .
since we have @xmath195 , this immediately implies that @xmath184 coincides with the de rham differential @xmath228 on the whole subalgebra @xmath343 more precisely , we must have @xmath344 on @xmath231 , and all the other components @xmath345 must vanish on this subalgebra . since @xmath184 is a derivation , this in turn implies that all the components @xmath346 for @xmath347 are not differential operators but bundle maps .
moreover , since the algebra @xmath83 is freely generated by @xmath324 and @xmath326 , and we know _ a priori _ the derivation @xmath184 on the generator subbundle @xmath326 , it always suffices to specify the restriction @xmath348 .
the decomposition @xmath253 of a hodge connection into an even and an odd part is quite transparent on the level of the weil algebra we simply have @xmath349 we will now rewrite the normalization condition in terms of the weil algebra . to do this , note that the map @xmath350 induces a bundle map @xmath351 .
this map is in fact a derivation of the weil algebra .
it vanishes on the generator bundle @xmath324 , while on the generator bundle @xmath326 it is given by @xmath352 then lemma [ d.norm ] immediately shows that the ( formal germ of the ) hypercomplex structure on @xmath64 induced by a derivation @xmath353 is normalized if and only if we have @xmath354 on the generator subbundle @xmath355 .
it is convenient to modify this in the following way .
let @xmath356 be the isomorphism inverse to @xmath357 .
set @xmath358 on the generator subbundle @xmath359 and extend it a derivation @xmath360 of the weil algebra . both derivations
@xmath361 and @xmath301 are real .
moreover , the derivation @xmath361 is weakly hodge ( the derivation @xmath301 is not simply because it decreases the weight ) .
then the normalization condition is equivalent to @xmath362 to sum up , formal germs near @xmath109 of normalized flat hodge connections on @xmath64 are in a natural one - to - one correspondence with derivations @xmath363 of the weil algebra @xmath83 which satisfy the following conditions .
1 . @xmath195 .
2 . @xmath364 is a weakly hodge algebra derivation of augmentation degree @xmath119 .
3 . @xmath365 .
4 . @xmath344 and @xmath366 , @xmath367 on the subalgebra @xmath368 .
5 . @xmath369 on @xmath370 for every @xmath371 .
[ cnd ] for every such derivation , the differential operator @xmath372 satisfies the leibnitz rule @xmath373 therefore it is a connection on the bundle @xmath324 .
we postpone the proof of the following lemma till the end of section [ metrics ] .
[ obata ] the connection @xmath374 on the bundle @xmath375 coincides with the connection on @xmath1 induced by the obata connection for the hypercomplex structure on @xmath64 defined by the derivation @xmath184 . with lemma [ obata ] in mind
, we see that theorem [ hc ] is reduced to the following statement .
[ hc.prop ] let @xmath70 be a torsion - free connection on the cotangent bundle @xmath326 of a complex manifold @xmath1 .
assume that the curvature of the connection @xmath70 is of type @xmath73
. then there exists a unique derivation @xmath376 of the weil algebra @xmath83 of the manifold @xmath1 such that @xmath184 satisfies the conditions - above and we have @xmath377 on @xmath355 .
this ends the preliminaries .
we now begin the proof of proposition [ hc.prop ] .
the proof proceeds by induction on the augmentation degree .
denote @xmath378 to base the induction , consider the derivation @xmath379 . by assumptions it is equal to @xmath380
let @xmath381 be the composition @xmath382 since the derivation @xmath379 of the graded - commutative algebra @xmath83 is of odd degree , up to a coefficient the composition @xmath383 coincides with the supercommutator @xmath384 .
in particular , it is also an algebra derivation . the derivation @xmath385 _ a priori _ has components @xmath386 , @xmath387 , @xmath388 of augmentation degrees @xmath61 , @xmath62 and @xmath39 .
however , @xmath389 moreover , @xmath390 is precisely the torsion of the connection @xmath391 .
thus it vanishes by assumption . since @xmath358 on @xmath392
, we also have @xmath393 on @xmath326 , which implies that @xmath393 everywhere .
what remains is @xmath388 . in general
, it does _ not _ vanish , and to kill it we have to add new terms @xmath338 .
we now turn to the induction step .
assume that for some @xmath394 we are already given the derivation @xmath395 which satisfies the conditions - on page , and assume that the composition @xmath396 has no non - trivial components of augmentation degrees @xmath397 .
denote by @xmath398 its component of augmentation degree @xmath119 .
we have to find a derivation @xmath399 which also satisfies - and such that @xmath400 in augmentation degree @xmath119 .
this is equivalent to @xmath401 the conditions and mean that @xmath364 must be a weakly hodge derivation which vanishes on @xmath402 .
the condition is only relevant for @xmath374 .
finally , the condition is relevant for all even @xmath119 and means that @xmath403 on @xmath355 . because of , it suffices to define @xmath338 on the generator subbundle @xmath324 .
since @xmath338 commutes with the complex conjugation map , it even suffices to consider only @xmath404 .
moreover , implies that it also suffices to check only on @xmath405 .
note that on this subbundle we have @xmath406
. there will be two slightly different cases .
the first is one when @xmath407 is odd , the second one is when @xmath408 is even . in both cases , the weakly hodge map @xmath409 only has non - trivial pieces of hodge bidegrees @xmath34 , @xmath73 and @xmath410 .
moreover , for any map @xmath223 of odd degree we tautologically have @xmath411 = 0 $ ] . applying this to @xmath412 and collecting terms of augmentation degree @xmath119
, we see that @xmath413 to track various components of the map @xmath409 it is convenient to refine the augmentation grading on the weil algebra @xmath83 to an _ augmentation bigrading _ by setting @xmath414 on the generator bundles @xmath415 and @xmath416 .
the augmentation bigrading will be denoted by lower indices , so that we have @xmath417 the relevant pieces of the augmentation bigrading on the bundle @xmath418 are shown on figure [ fig.odd ] for @xmath419 odd , and on figure [ fig.even ] for @xmath408 even .
the axes on the figures correspond to the grading by hodge type .
a hodge bidegree component @xmath420 can be non - trivial only when @xmath421 and @xmath422 .
thus the component @xmath423 is represented by an upward - looking angle with vertex @xmath424 : a graded piece @xmath420 can be non - trivial only if the point @xmath425 lies in the interior ( or on the boundary ) of this angle . [ fig.odd ] [ fig.even ] consider the hodge bidegree decompositions @xmath426 of the derivations @xmath361 , @xmath301 of the weil algebra @xmath83 .
then the augmentation bigrading is essentially the eigenvalue decomposition for the commutators @xmath427 and @xmath428 .
more precisely , we have @xmath429 indeed , since all these commutators are derivations of the weil algebra , it suffices to check this on the generator bundles @xmath324 , @xmath326 , which is elementary . in particular
, we see that both @xmath361 and @xmath301 preserve the augmentation bidegree .
the equalities also immediately imply that @xmath430 one further corollary of will be very important ( we leave the proof to the reader as an easy exercise ) .
[ bndry ] if @xmath431 , then the map @xmath432 is injective on every graded piece @xmath420 with @xmath433 , while @xmath434 is injective on @xmath420 with @xmath435 .
graphically , this means that @xmath432 is injective on @xmath420 when the point @xmath425 lies on the right - hand boundary of the angle representing @xmath423 , and @xmath434 is injective in this graded piece when the point @xmath425 lies on the left - hand boundary of the same angle .
we will call this _ the boundary rule_. we can now proceed with the proof of the induction step .
_ case when @xmath407 is odd .
_ looking at figure [ fig.odd ] , we see that the only non - trivial augmentation - bidegree components of the map @xmath409 are @xmath436 and @xmath437 , @xmath438 and the same is true for any weakly hodge map @xmath439 , @xmath440 moreover , on @xmath441 we have @xmath442 while @xmath443 further decomposes as @xmath444 .
we have to find @xmath338 which satisfies .
in particular , we must have @xmath445 but by the boundary rule the map @xmath432 is injective on @xmath446 , while the map @xmath434 is injective on @xmath447 .
therefore there exists at most one weakly hodge map @xmath448 which satisfies .
setting ( again on @xmath449 ) @xmath450 gives this unique solution to .
indeed , we have @xmath451 the second summand in the brackets vanishes since @xmath452 , while the first is equal to @xmath453 by .
this proves the first equation in .
the second one is proved in exactly the same way .
it remains to prove that this map @xmath338 satisfies not only but also the stronger condition . to do this , note that @xmath454 but from we see that @xmath455 is of hodge bidegree @xmath73 .
therefore this implies that @xmath456 the only possible non - trivial components of @xmath455 with respect to the augmentation bigrading have bidegrees @xmath457 and @xmath458 , and by the boundary rule @xmath432 is injective on @xmath459 , while @xmath434 is injective on @xmath460 .
thus @xmath461 .
_ case when @xmath462 is odd .
_ assume for the moment that @xmath463 , thus @xmath464 .
looking at figure [ fig.even ] , we see that _ a priori _ the map @xmath465 can have three non - trivial augmentation - bidegree components , namely , @xmath466 however , since @xmath452 , and the map @xmath432 is injective on @xmath467 by the boundary rule , we see that @xmath468 .
analogously , @xmath469 .
therefore @xmath465 is of pure augmentation bidegree @xmath470 .
since the map @xmath471 is weakly hodge , it must also be of augmentation bidegree @xmath470 .
conversely , looking at the angle representing @xmath472 , we see that _ every _ real map @xmath473 of pure augmentation bidegree @xmath470 is necessarily weakly hodge . in particular
, setting @xmath474 defines a weakly hodge map .
this map is a solution to : @xmath475 this solution is not unique
. however , since @xmath119 is even , we have the additional normalization condition @xmath476 .
this condition ( automatically satisfied by the solution ) ensures uniqueness .
indeed , the difference @xmath477 between two solutions @xmath338 , @xmath478 must satisfy @xmath479 , which implies @xmath480 finally , it remains to consider the case @xmath481 .
the general argument works in this case just as well , with a single exception .
since @xmath482 is no longer strictly positive , the boundary rule does not apply : it is not true that @xmath432 is injective on @xmath483 ( in fact , on this graded piece @xmath432 is equal to zero ) .
therefore the component @xmath484 does not vanish automatically .
however , this component @xmath485 is precisely the @xmath34-curvature of the connection @xmath70 on @xmath1 .
it vanishes by the second assumption on this connection .
the last section essentially finishes the proof of the hypercomplex theorem [ hc ] ( it remains to prove lemma [ obata ] ) .
we will now sketch a proof of the hyperkhler theorem [ hk ] .
as we have already noted , theorem [ hk ] will be a corollary of theorem [ hc ] .
namely , given a khler manifold @xmath1 we proceed in the following way .
first we note that the levi - civita connection @xmath14 on @xmath1 has no torsion and no @xmath34-curvature
. therefore theorem [ hc ] applies to @xmath14 and provides a hypercomplex structure on the total space @xmath74 .
then we show that every hermitian metric on @xmath1 which is preserved by @xmath14 ( in particular , the given khler metric ) extends uniquely from the zero section @xmath319 to a ( formal germ of a ) hyper - hermitian metric on the hypercomplex manifold @xmath74 which is compatible with the hypercomplex structure . after this
, we finish the proof by identifying the holomorphically symplectic manifolds @xmath64 and @xmath0 . we will go through these steps in reverse order , starting with the last one .
[ cotang ] assume given a hypercomplex structure on the total space @xmath486 which satisfies the conditions of theorem [ hc ] .
let @xmath13 be a @xmath43-invariant hyperkhler metric on @xmath8 compatible with this hypercomplex structure , and let @xmath487 be the associated holomorphic @xmath39-form .
let @xmath0 be the total space of the cotangent bundle to @xmath1 equipped with the standard holomorphic @xmath39-form @xmath32 .
then there exists a unique @xmath43-equivariant biholomorphic map @xmath488 such that @xmath489 . _ proof .
_ since the map @xmath490 must be @xmath43-equivariant , it must commute with the canonical projections @xmath491 and send the zero section @xmath50 to the zero section @xmath3 .
denote by @xmath91 the differential of the @xmath43-action .
then we also must have @xmath492 but the @xmath62-form @xmath493 is the tautological @xmath62-form @xmath494 .
therefore the @xmath62-form @xmath308 on @xmath8 completely defines the map @xmath490 .
conversely , the form @xmath495 satisfies @xmath496 for a unique map @xmath488 . since the metric @xmath13 is @xmath43-invariant , the forms @xmath497 and @xmath498 are of weight @xmath62 .
therefore the map @xmath488 is @xmath43-equivariant . by the cartan homotopy formula , we have @xmath499 we will now explain how to construct the metric @xmath13 or , equivalently , the associated holomorphic @xmath39-form @xmath500 .
keep the notation of last two sections .
let @xmath501 be the khler form ( more generally , any @xmath73-form preserved by the connection ) .
we have to prove that there exists a unique ( formal germ of a ) holomorphic @xmath34-form @xmath502 which is of @xmath154-type @xmath73 and whose restriction to the zero section @xmath50 coincides with @xmath35 ( since the positivity is an open condition , it is satisfied automatically in a neighborhood of the zero section @xmath50 ) . to reformulate this in terms of @xmath1 ,
consider the complex @xmath503 of hodge bundles on @xmath64 with the dolbeault differential @xmath504 .
denote by @xmath505 its direct image on @xmath1 ( the grading is shifted by @xmath39 to make it compatible with the hodge degrees ) .
we are given a section @xmath506 of hodge type @xmath73 .
we have to prove that there exists a section @xmath507 such that @xmath508 on the zero section @xmath50 and @xmath509 . since @xmath510 and @xmath511
, the complex @xmath512 is a free module @xmath513 over the weil algebra @xmath514 generated by some subbundle @xmath515 which is isomorphic to @xmath516 .
we introduce the augmentation grading on the @xmath83-module @xmath512 by setting @xmath517 . just as in proposition
[ hc.prop ] , the proof will proceed by induction on the augmentation degree
namely , we will construct the form @xmath518 as a sum @xmath519 with @xmath520 of augmentation degree @xmath521 .
we begin with the induction step .
it is completely parallel to proposition [ hc.prop ] , so we give only a sketch . _ induction step a sketch .
_ we can assume that we already have @xmath522 such that @xmath523 is of augmentation degree @xmath524 .
denote by @xmath525 the component of augmentation degree exactly @xmath521 .
we have to show that there exists a unique @xmath526 such that @xmath527 .
the derivations @xmath361 and @xmath301 of the weil algebra @xmath83 extend to endomorphisms of the free module @xmath528 by setting @xmath529 on @xmath530 . just as on @xmath83
, we have @xmath365 . for every @xmath181
, we have @xmath531 on @xmath532 .
this immediately implies that @xmath533 is injective on @xmath534 for @xmath371 , which proves the uniqueness of @xmath535 .
the space @xmath532 splits into the sum of parts of the form @xmath536 such a part can have a non - trivial piece of hodge bidegree @xmath537 only if @xmath538 and @xmath539 . having in mind the graphical representation as in figure [ fig.odd ] and
figure [ fig.even ] , we will say that the part @xmath540 is _ an angle based at @xmath541_. each angle is preserved by the maps @xmath361 and @xmath301 . in this terminology , @xmath542 can intersect non - trivially with various angles based at @xmath34 and @xmath73 , while @xmath543 can intersect non - trivially with angles based at @xmath73 and @xmath410 . but by induction we have @xmath544 , which implies that @xmath545 .
applying the boundary rule lemma [ bndry ] , we see that both @xmath546 and @xmath547 must lie entirely within angles based at @xmath73
. therefore @xmath548 must also lie within angles based in @xmath73 .
this means that @xmath549 is of hodge type @xmath73 , and we can set @xmath550 .
we now have to base the induction
namely , to find the section @xmath551 with correct restriction to the zero section @xmath50 , and to handle those angles @xmath552 to which the boundary rule does not apply which means the angles with @xmath553 or @xmath554 .
there are three such angles .
we denote the corresponding components of the derivation @xmath555 by @xmath556 we have to choose @xmath557 so that @xmath558 .
we note that @xmath559 implies @xmath560 by complex conjugation .
moreover , we note that we have one more degree of freedom .
so far nothing depended on the choice of the generator subbundle @xmath561 all we needed was to know that it exists .
we will now make this choice .
it will not be the most obvious one , but the one which will make computations as easy as possible .
we consider the splitting @xmath562 given by the hodge connection @xmath236 .
this splitting induces a bigrading on the de rham algebra @xmath563 .
the complex @xmath563 with this bigrading is a bicomplex which we will denote by @xmath564 .
more precisely , the de rham differential @xmath228 is a sum of two anti - commuting differentials @xmath565 ( this is essentially equivalent to the flatness of the hodge connection @xmath223 ) .
since the first term in the splitting is @xmath272 , the subcomplexes @xmath566 and @xmath567 of the de rham complex @xmath563 are the same for every @xmath119 . therefore the associated graded quotients @xmath568 and @xmath569 are also isomorphic for every @xmath119 .
in particular , we have @xmath570 on the other hand , since @xmath571 is the bundle of relative @xmath62-forms on @xmath47 , the quotient complex @xmath572 with the differential @xmath573 is canonically isomorphic @xmath574 to the relative de rham complex @xmath575 .
we use this identification and choose as @xmath576 the subbundle of @xmath119-forms which are constant along the fibers of the projection @xmath76 ( by a constant @xmath119-form on a vector space @xmath25 with a basis @xmath577 we mean a linear combination of forms @xmath578 with constant coefficients ) .
this choice guarantees that the relative de rham differential @xmath579 takes a very simple form .
namely , it vanishes on @xmath580 , and induces an isomorphism @xmath581 between the generator subbundles @xmath582 and @xmath583 .
this is important because by construction @xmath573 anti - commutes with @xmath184 .
moreover , since @xmath573 obviously preserves the augmentation degrees , it anti - commutes separately with each of the components @xmath338 .
since we already know the derivations @xmath338 on @xmath83 , the isomorphism will allow us to compute individual components @xmath584 .
the first result is the following : the map @xmath585 in is minus the connection @xmath586 on the bundle @xmath587 .
indeed , @xmath588 is a derivation of the exterior algebra @xmath589 .
thus it suffices to prove that @xmath590 on @xmath591 .
since @xmath592 , this follows from and the construction of the map @xmath353 given in section [ pf ] .
this shows that taking @xmath593 guarantees that @xmath594 . at this point
we will also choose the isomorphism @xmath595 namely , we take the composition of the embedding @xmath596 and the restriction @xmath597 to the zero section @xmath598 .
then the form @xmath557 defined by automatically restricts to @xmath35 .
it remains to prove that @xmath559 .
this is a corollary of the following claim .
the map @xmath599 defined in is the composition of the curvature @xmath600 of the connection @xmath601 and a certain bundle map @xmath602 _ proof . _ extend the map @xmath599 to an algebra derivation @xmath603 by setting @xmath604 on @xmath605 and taking as @xmath606 the corresponding component of the map @xmath607 . since @xmath599 is a derivation of the algebra @xmath589 which vanishes on @xmath605 , on @xmath608 it is equal to the composition @xmath609 (
here @xmath610 is the alternation map ) .
therefore it suffices to prove that on @xmath611 we have @xmath612 for a certain bundle map @xmath613 . since @xmath614
, this follows directly from and with @xmath481 .
this lemma implies that @xmath615 .
this finishes the proof of theorem [ hk ] .
the last application of the formalism that we have developed in this section is the proof of lemma [ obata ] .
_ proof of lemma [ obata ] . _ the obata connection @xmath69 on a hypercomplex manifold @xmath8 is defined as follows : for every @xmath142-form @xmath616 the @xmath273-part @xmath617 is equal to @xmath618
while the @xmath142-part satisfies @xmath619 the first condition in fact automatically follows from the absence of torsion .
consider a hypercomplex structure on @xmath74 corresponding to a torsion - free connection @xmath70 on @xmath1 , and let @xmath620 be a @xmath142-form on @xmath1 .
we have to prove that @xmath621 on the zero section @xmath598 .
it suffices to prove that @xmath622 as section of the bundle @xmath623 on the zero section @xmath50 .
the canonical isomorphism @xmath624 sends the dolbeault differential @xmath625 to the component @xmath626 of the dolbeault differential @xmath627 .
moreover , the composition of the map @xmath628 and the restriction to the zero section @xmath50 induces a bundle map @xmath629 it is easy to check that this map is proportional to the canonical embedding @xmath630 . therefore it commutes with the connection @xmath631 namely , we have @xmath632 . the equation becomes @xmath633 but we have already proved that on @xmath591 we have @xmath634 .
to illustrate our rather abstract methods by a concrete example , we would like now to derive a formula for the canonical hypercomplex structure on @xmath635 in the case when @xmath1 is a symmetric space . in this case
the classic equality @xmath636 bring many simplifications , so that the constructions of section [ pf ] can be seen through to a reasonably explicit final result .
the formula that we obtain is similar to the one obtained by o. biquard and p. gauduchon in @xcite .
let us introduce some notations .
let @xmath1 be a symmetric space with levi - civita connection @xmath70 and curvature @xmath383 .
consider the total space @xmath637 with the canonical projection @xmath66 .
let @xmath638 be the endomorphism of the pullback bundle @xmath245 given by @xmath639 here @xmath640 , @xmath641 , @xmath642 is a point in @xmath64 , @xmath643 is the curvature evaluated at the point @xmath644 , and @xmath645 is interpreted as a section of the bundle @xmath646 .
let also @xmath647 be the generating function for the recurrence relation @xmath648 with the initial condition @xmath649 .
in other words , @xmath650 is the solution of the ode @xmath651 with the initial condition @xmath652 . with these notations
, we can formulate our result .
[ symm ] let @xmath653 be the complex structure operator on @xmath1 .
then the map @xmath654 for the canonical normalized hypercomplex structure on @xmath8 is given by the matrix @xmath655 with respect to the decomposition @xmath656 associated to the levi - civita connection on @xmath1 .
we have already noted that this result is very similar to the formula obtained in @xcite .
however , it is not the same .
the reason is the following : biquard and gauduchon work with the cotangent bundle @xmath0 , and they use the normalization natural to the cotangent bundle . as the result ,
their analog of the map @xmath657 is slightly different , and the function @xmath650 is also different in particular , it is given by an explicit expression . to compare our results with those of @xcite
, one should either compute the normalization map @xmath658 for the biquard - gauduchon hypercomplex structure , or go the other way around and compute the map @xmath659 provided by lemma [ cotang ] .
unfortunately , i have not been able to do either .
the main simplification in the case of symmetric spaces is the well - known equality @xmath636 . for our construction
it immediately implies that the odd augmentation degree components @xmath660 of the weakly hodge derivation @xmath661 vanish for @xmath662 .
the only non - trivial component is @xmath374 . by
, this immediately shows that the connection @xmath237 on the fibration @xmath76 is simply the linear connection associated to the levi - civita connection @xmath70 .
applying lemma [ j.matrix ] , we see that all we have to prove is the equality @xmath663 . by lemma [ rj ]
this can be rewritten in terms of the weil algebra @xmath83 as @xmath664 moreover , by with @xmath481 the endomorphism @xmath657 of the @xmath323-module @xmath665 satisfies @xmath666 on @xmath326 .
in fact , this should be taken as the definition
i apologize to the reader for any possible mistakes in writing down the explicit formula .
next we note that since @xmath667 vanishes on @xmath326 , the endomorphism @xmath657 is in fact equal @xmath668 to the commutator @xmath669 . moreover , this formula holds not only on the generator subbundle @xmath359 , but on the whole @xmath323-module @xmath670 .
indeed , by the normalization condition this commutator vanishes on the generator subbundle @xmath582 .
since it is a derivation of the weil algebra , it vanishes on the whole @xmath323 .
therefore it restricts to a map of @xmath323-modules on @xmath671 .
we now trace one - by - one the induction steps in the proof of proposition [ hc.prop ] . since all the odd terms vanish
, we only need to consider the even terms @xmath672 .
we have to prove that @xmath673 since both sides are maps of @xmath323-modules , it suffices to prove this on @xmath359 , where we can replace @xmath674 with @xmath675 . by , on @xmath582 we have @xmath676 comparing this to , we see that it suffices to prove that @xmath677 for all even @xmath678 . writing out the commutators on the right - hand side , we see that the difference is equal to @xmath679 the first summand vanishes since by the normalization condition we have @xmath680 on @xmath681 .
the last two summands vanish since @xmath682 on @xmath326 .
o. biquard and p. gauduchon , _ hyperkhler metrics on cotangent bundles of hermitian symmetric spaces _ , in _ proc .
of the special session an geometry and physics held at aarhus university and of the summer school at odense university , denmark , 1995 _ , lecture notes in pure and applied mathematics , * 184 * , marcel dekker inc . 1996 , 768 pages .
a. dancer and a. swann , _ the structure of quaternionic khler quotients _ , in _ proc .
of the special session an geometry and physics held at aarhus university and of the summer school at odense university , denmark , 1995 _ , lecture notes in pure and applied mathematics , * 184 * , marcel dekker inc . 1996 , 768 pages . | a canonical hyperkhler metric on the total space @xmath0 of a cotangent bundle to a complex manifold @xmath1 has been constructed recently by the author in @xcite .
this paper presents the results of @xcite in a streamlined and simplified form .
the only new result is an explicit formula obtained for the case when @xmath1 is an hermitian symmetric space . |
more than three decades ago partridge and peebles ( 1967 ) predicted that galaxies in their first throes of star - formation should be strong emitters in the lyman-@xmath1 line .
their predictions were optimistic , based on converting roughly 2% of gas into stars in @xmath19 years in milky way sized galaxies , which translates into a luminosity of @xmath20 .
these objects are also expected to be common - if all the @xmath21 galaxies have undergone a phase of rapid star - formation one should see a surface density of about @xmath22 ( pritchet 1994 ) .
searches based on these expectations did not detect lyman-@xmath1 emitters ( laes ) .
( see review by pritchet 1994 ; koo & kron 1980 ; pritchet & hartwick 1987 , 1990 ; cowie 1988 ; rhea et al 1989 ; smith et al 1989 ; lowenthal et al 1990 ; wolfe et al 1992 ; de propris et al 1993 ; macchetto et al 1993 ; mller & warren 1993 ; djorgovski & thompson 1992 ; djorgovski , thompson , & smith 1993 ; thompson , djorgovski , & trauger 1992 ; thompson et al 1993 ; thompson , djorgovski , & beckwith 1994 ; thommes et al 1998 . ) only recently have lyman-@xmath1 emitters been observed , albeit at luminosity levels roughly a hundred times lower than the original prediction .
these lyman-@xmath1 emitters have been found from both deep narrow band imaging surveys ( cowie & hu 1998 ; hu , cowie & mcmahon 1998 ; hu , mcmahon , & cowie 1999 ; kudritzki et al 2000 ) , and from deep spectroscopic surveys ( dey et al 1998 ; manning et al 2000 ; but see stern et al 2000 ) .
weak lyman-@xmath1 emitters have also been found through targeted spectroscopy of lyman break objects ( e.g. , steidel et al 1996 , lowenthal et al 1997 ) .
the lower luminosity in the lyman-@xmath1 line may be because of attenuation by dust if chemical enrichment is prompt ; or because the star - forming phase is more protracted ; or because the star - formation happens in smaller units which later merge .
the first two scenarios will give a smaller equivalent width than early predictions , while the last scenario results in low luminosities but high equivalent width .
dust effects are expected to be severe even a small amount of dust can greatly attenuate this line , because it is resonantly scattered
. however , two factors can help the lyman-@xmath1 photons escape .
if lyman-@xmath1 photons are produced in diffuse regions of a clumpy interstellar medium , they can simply scatter off the dense clumps and escape ( neufeld 1991 ) , and some geometries can even lead to an increase in the equivalent width of the line . secondly , energetic winds are seen in low-@xmath23 lyman-@xmath1 emitters ( kunth et al 1998 ) .
these can displace the neutral gas and doppler - shift the peak wavelength of the resonant scattering , thereby reducing the amount of scattering and the path length for interaction with dust .
detailed predictions for luminosities and surface densities of laes using a press - schechter formalism and exploring a range of dust obscuration and star - formation time scales have been explored by haiman and spaans ( 1999 ) , who are able to reproduce the surface densities of laes reported by hu et al ( 1998 ) with a wide range of models . in order to narrow down the range of possibilities and characterize the high redshift lyman-@xmath1 population , better statistics over a wide range of flux and source density
are needed .
an efficient search for lyman-@xmath1 emitters ( and other emission line galaxies ) was started in 1998 using the ccd mosaic camera at the kitt peak national observatory s 4 m mayall telescope .
the mosaic camera has eight @xmath24 chips in a @xmath25 array comprising a @xmath26 field of view .
the final area covered by the lala survey is @xmath27 square - degrees in two mosaic fields centered at 14:25:57 + 35:32 ( 2000.0 ) and 02:05:20 -04:55 ( 2000.0 ) .
five overlapping narrow band filters of width fwhm@xmath28are used .
the central wavelengths are @xmath16 , @xmath29 , @xmath30 , @xmath31 , and @xmath32 , giving a total redshift coverage @xmath33 .
this translates into surveyed comoving volume of @xmath34 comoving @xmath35 per field for @xmath36 , @xmath37 , @xmath38 .
about 70% of the imaging at @xmath39 is complete , and an extension of the survey to @xmath40 is planned . in about 6 hours per filter per field
we are able to achieve line detections of about @xmath41 .
the survey sensitivity varies with seeing .
broadband images of these fields in a custom @xmath42 filter ( @xmath43 , @xmath44 ) and the johnson - cousins @xmath45 , @xmath46 , and @xmath47 bands are being taken as part of the noao deep widefield survey ( jannuzi & dey 1999 ) .
the images were reduced using the mscred package ( valdes & tody 1998 ; valdes 1998 ) in the iraf environment ( tody 1986 , 1993 ) , together with assorted custom iraf scripts .
in addition to standard ccd data reduction steps ( overscan subtraction , bias frame subtraction , and flatfielding ) , it was necessary to remove crosstalk between pairs of ccds sharing readout electronics and to remove a ghost image of the telescope pupil .
astrometry of usno - a catalog stars was used to interpolate all chips and exposures onto a single coordinate grid prior to combining final images .
cosmic ray rejection is of particular importance in a narrowband search for emission line objects .
we therefore identified cosmic rays in individual images using a digital filtering method ( rhoads 2000 ) and additionally applied a sigma clipping algorithm at the image stacking stage .
weights for each exposure were determined using the method of fischer & kochanski ( 1994 ) , which accounts for sky level , transparency , and seeing to optimize the signal to noise level for compact sources in the final image .
catalogs were generated using the sextractor package ( bertin & arnouts 1996 ) .
fluxes were measured in @xmath48 ( 9 pixel ) diameter apertures , and colors were obtained using matched @xmath48 apertures in registered images that had been convolved to yield matched point spread functions .
spectroscopic followup of a cross - section of emission line candidates was obtained with the lris instrument ( oke et al 1995 ) at the keck 10 m telescope on 1999 june 13 ( ut ) .
two dithered exposures of @xmath49 seconds each were obtained through a single multislit mask in good weather .
these data were reduced using a combination of standard iraf onedspec and twodspec tasks together with the homegrown slitmask reduction iraf task `` bogus '' ( stern , bunker , & stanford , private communication ) for reducing lris data .
our imaging data yield numbers of sources as a function of their fluxes in several filters , from which we can construct number densities as a function of magnitudes , colors , and equivalent widths . in order to gracefully handle sources that are not detected in all filters ,
we have chosen to use `` asinh magnitudes '' ( lupton , gunn , & szalay 1999 ) , which are a logarithmic function of flux for well detected sources but approach a linear function of flux for weak or undetected sources .
figure [ cmdfig ] shows the color - magnitude diagram for the 0 ( @xmath50 ) and @xmath45 filters .
the color scatter achieved for bright sources ( @xmath51 ) is @xmath52 magnitudes ( semi - interquartile range ) .
this includes the true scatter in object colors , and is therefore a firm upper limit on the scatter introduced by any residual systematic error sources , which we expect to be a few percent at worst . to sharpen our focus on the high redshift lyman-@xmath1 population
, we identify the range of parameter space occupied by known @xmath53 lyman-@xmath1emitters .
these sources have typical observed equivalent widths @xmath54 and line+continuum fluxes @xmath55 ( adjusted to @xmath7 and measured in an @xmath56 filter ) .
we further want to restrict attention to sources with sufficiently reliable detections that the number of false emission line candidates is a small fraction of the total candidate sample .
the data set includes many continuum sources ( @xmath57 in the flux range above ) .
it is expected to contain a few hundred emission line sources , based on earlier source counts from smaller samples ( hu et al 1998 ) and on our own findings ( below ) .
the fraction of continuum sources with measured equivalent widths above some threshold @xmath58 can be calculated as a function of signal to noise level @xmath59 and filter width @xmath60 : the measured equivalent width for a source with no line emission or absorption in the narrowband filter will be @xmath61 .
thus , a false positive becomes an @xmath62 event , where @xmath63 .
our sample would be reasonably safe from contamination with @xmath64 , which would correspond to about 20 false positives in a sample of 14000 sources ; and very safe for @xmath65 , corresponding to @xmath66 false positive in 14000 sources .
we have conservatively chosen detection thresholds @xmath67 and @xmath68 , which gives @xmath69 .
this keeps the number of false positives small even when the foregoing analysis is expanded to include errors in the continuum flux ( which increase photometric error in the color by a factor @xmath70 ) and a realistic distribution of equivalent widths for the low-@xmath23 galaxy populations . as a further check on our sample
, we use the photometric color error estimates from sextractor to demand that the source be an emission line source at the @xmath71 level . combining all of these requirements ,
the final criteria for good candidates in our survey become @xmath72 , @xmath73 , and @xmath74 , where @xmath75 .
there are @xmath76 such sources detected in the 0 filter in a solid angle of @xmath77 square degree and redshift range @xmath78 .
this corresponds to @xmath79 good candidate lyman-@xmath1
emitters per square degree per unit redshift .
the precise upper flux cutoff used is somewhat arbitrary , but including sources with larger fluxes in our list of good candidates has little effect on the total source counts .
the final 16 image had a broader psf than the 0 image , and consequently a reduced effective sensitivity .
in addition , it samples a slightly higher redshift , resulting in an @xmath80 magnitude increase in distance modulus ( @xmath81 , @xmath38 ) . accounting for both effects ,
the matched flux range is @xmath82 at 0 and @xmath83 at 16 .
the corresponding counts are @xmath84 candidates at 16 , and 104 candidates at 0 .
thus , the source density in directly comparable luminosity bins varies by a factor of @xmath85 , a difference that is significant at about the @xmath86 level .
we interpret this as a likely signature of large scale structure in the lyman-@xmath1 emitter distribution at @xmath0 .
the comoving distance between the centers of the two redshift slices is @xmath87 , while the thickness of each slice is @xmath88 and the transverse comoving size @xmath89 ( again assuming @xmath90 , @xmath37 , @xmath38 ) .
comparable variations have been observed in the density of lyman break galaxies at @xmath91 ( steidel et al 1998 ) .
our spectroscopic followup sampled a wide range of flux and equivalent width in order to characterize the different populations of sources with extreme narrowband colors , and to thereby tune candidate selection criteria for future spectroscopic observations . those sources with @xmath92 narrowband detections and photometrically measured @xmath93 were all confirmed as emission line objects at the narrowband wavelength . in total , we detected two emitters at @xmath94 , three emitters at @xmath95 and @xmath12 , one confirmed lyman-@xmath1 ( @xmath96 ) emitter at @xmath97 , and one source that could either be lyman-@xmath1 at @xmath11 or at @xmath12 .
we also found a @xmath98 galaxy with emission lines of , @xmath991640 , and @xmath991663 .
this source may have been included in the narrowband candidate list ( with @xmath100 ) because of a weak spectral break around @xmath101 .
finally , we found a few serendipitous emission line sources in the slit spectra .
one of these is a single - line source with large equivalent width , possibly lyman-@xmath1 at @xmath102 .
spectra of two interesting sources are shown in figure [ specfig ] .
by combining our imaging survey with these spectroscopic results , we can estimate the source density of lyman-@xmath1 emitters passing our selection cut .
our spectra included three sources fulfilling the criteria given above for good candidates .
of these , one was confirmed as a @xmath10 lyman-@xmath1 source .
a second remains a candidate @xmath11 source , but is more conservatively interpreted as a @xmath12 [ o ii ] emitter on the basis of a rather strong continuum on the blue side of the line .
the third is a clear @xmath94 [ o iii ] emitter .
we therefore estimate that roughly @xmath103 to @xmath104 of the good candidates will be confirmed as lyman-@xmath1 sources , yielding @xmath105 emitters per square degree per unit redshift .
this is compatible with earlier measurements from smaller volumes ( hu et al 1998 ) after accounting for differences in flux threshold .
our measurement is distinct from previous efforts in the field for its basis in a large number of candidate emitters .
poisson errors in our source counts are of order @xmath106 .
this is smaller than the variations observed in the comparison of two filters ( of order @xmath107 ) . by combining observations in multiple fields
, we will be able to average over local fluctuations in number densities effectively . when completed , the lala survey will yield comoving volume of @xmath108 ( [ obsdesc ] ) and a sample of several hundred laes , and will allow the luminosity function , equivalent width distribution , and correlation function of this population to be determined for the first time .
we thank andy bunker and steve dawson for help with the spectroscopic observations and frank valdes for writing and helping with the mscred package in iraf .
jer s research is supported by a kitt peak postdoctoral fellowship and by an stsci institute fellowship .
sm s research is supported by nasa through hubble fellowship grant # hf-01111.01 - 98a from the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555 .
bertin , e. and arnouts , s. , 1996 , 117,393 cowie , l. l. 1988 , in `` the post - recombination universe '' , ed . n.
kaiser & a. n. lasenby ( dordrecht , kluwer ) p. 1
cowie , l. l. , & hu , e. m. 1998 , 115 , 1319 de propris , r. pritchet , c. j. , hartwick , f. d. a. , & hickson , p. 1993 , 105
, 1243 dey , a. , spinrad , h. , stern , d. , graham , j. r. , & chaffee , f. h. 1998 , 498 , 93 djorgovski , s. , & thompson , d. j. 1992 , in `` the stellar populations of galaxies , '' iau symposium 149 , ed . b. barbuy & a. renzini ( dordrecht , kluwer ) , p. 337 djorgovski , s. , thompson , d. j. , & smith , j. d. 1993 , in `` first light in the universe , '' ed . b. rocca - volmerange , m. dennefeld , b. guiderdoni , & j. tran thanh van ( gif sur yvette , editions frontires ) , p. 67
fischer , p. , & kochanski , g. p. 1994
, 107 , 802 hu , e. m. , cowie , l. l. , & mcmahon , r. g. 1998 , 502 , l99 jannuzi , b. t. , & dey , a. , 1999 , in `` photometric redshifts and high redshift galaxies '' , asp conference series , vol .
191 , editors r. j. weymann , l. j. storrie - lombardi , m. sawicki , and r. j. brunner , p.111 koo , d. c. , & kron , r. g. 1980 , 92 , 537 kudritzki , r .- p . , et al 2000 , in preparation landolt , a. u. 1992 , 104 , 340 lowenthal , j. d. , hogan , c. j. , leach , r. w. , schmidt , g. d. , & foltz , c. b. 1990 , 357 , 3 lowenthal , j. d. , koo , d. c. , guzman , r. , gallego , j. , phillips , a. c. , faber , s. m. , vogt , n. p. , illingworth , g. d. , & gronwall , c. 1997 , 481 , 673 lupton , r. h. , gunn , j. e. , & szalay , a. s. 1999 , aj 118 , 406 . manning , xx , et al 2000 , astro - ph/0002239 kunth , d. , mas - hesse , j. m. , terlevich , e. , terlevich , r. , lequeux , j. , & fall , s. m. 1998 , 334 , 11 macchetto , f. , lipari , s. , giavalisco , m. , turnshek , d. a. , & sparks , w. b. 1993 , 404 , 511 mller , p. , & warren , s. j. 1993 , 270 , 43 oke , j.b .
et al 1995 , pasp , 107 , 375 pritchet , c. j. 1994 , 106 , 1052 pritchet , c. j. , & hartwick , f. d. a. 1987 , 320 , 464 pritchet , c. j. , & hartwick , f. d. a. 1990 , 355 , l11 rhee , g. f. r. n. , webb , j. k. , & katgert , p. 1989 , 217 , 1 rhoads , j. e. 2000 , in press ( may 2000 issue ) smith , h. e. , cohen , r. d. , burns , j. e. , moore , d. j. , & uchida , b. a. 1989 , 347 , 87 steidel , c.c . ,
giavalisco , m. , pettini , m. , dickinson , m. , & anderson , k. l. 1996 , 462 , l17 steidel , c. c. , adelberger , k. l. , dickinson , m. , giavalisco , m. , pettini , m. , & kellogg , m. 1998 , 492 , 428 . stern , d. , & spinrad , h. 1999 , 111 , 1475 stern , d. , et al 2000 , astro - ph/0002241 thompson , d. , djorgovski , s. , & trauger , j. 1992 , in `` cosmology and large - scale structure in the universe , '' asp conference series , vol .
24 ( san francisco , asp ) , p. 147
thompson , d. , djorgovski , s. , trauger , j. , & beckwith , s. 1993 , in `` sky surveys : protostars to protogalaxies , '' asp conference series , vol .
b. t. soifer ( san francisco , asp ) , p. 185 thommes , e. , meisenheimer , k. , fockenbrock , r. , hippelein , h. , rser , h .- j .
, & beckwith , s. 1998 , 293 , l6 thompson , d. , djorgovski , s. , & beckwith , s. 1994 , 107 , 1 tody , d. 1986 , proc .
spie instrumentation in astronomy vi , ed .
crawford , 627 , 733 tody , d. 1993 , adass ii , a.s.p .
conference ser . ,
vol 52 , eds .
hanisch , r.j.v .
brissenden , & j. barnes , 173 p valdes , f. , & tody , g. 1998 , proceedings of spie vol .
3355 valdes , f. 1998 , adass vii , asp conf .
series 145 , ed .
r. albrect wolfe , a. m. , turnshek , d. a. , lanzetta , k. m. , & oke , j. b. 1992 , 385 , 151 | we report on a new survey for @xmath0 lyman-@xmath1 sources , the large area lyman alpha ( lala ) survey .
our survey achieves an unprecedented combination of volume and sensitivity by using narrow - band filters on the new @xmath2 pixel ccd mosaic camera at the 4 meter mayall telescope of kitt peak national observatory .
well - detected sources with flux and equivalent width matching known high redshift lyman-@xmath1 galaxies ( i.e. , observed equivalent width @xmath3 , @xmath4 , and @xmath5 ) have an observed surface density corresponding to @xmath6 per square degree per unit redshift at @xmath7 .
spatial variation in this surface density is apparent on comparison between counts in @xmath8 and @xmath9 filters .
early spectroscopic followup results from the keck telescope included three sources meeting our criteria for good lyman-@xmath1 candidates .
of these , one is confirmed as a @xmath10 source , while another remains consistent with either @xmath11 or @xmath12 .
we infer that @xmath13 to @xmath14 of our good candidates are _ bona fide _
lyman-@xmath1
emitters , implying a net density of @xmath15 lyman-@xmath1 emitters per square degree per unit redshift .
0@xmath16 16@xmath17 17f_17 2s @xmath18 |
a theory of quantum gravity should not apply only to asymptotically locally anti - de sitter ( alads ) spacetimes . for this reason , the ads / cft correspondence @xcite , although immensely successful , has fallen short of a description of the quantum mechanics of spacetime .
the ads restriction is severe : maldacena s conjecture does not apply in an obvious way to even the cosmological spacetime we find ourselves in .
if a quantum theory applies to general spacetimes , it is desirable that it reduces to ads / cft in the appropriate cases .
this suggests a strategy for guessing properties of a complete theory : consider specific aspects of ads / cft and devise generalizations that are applicable to other spacetimes .
if one knew only of special relativity , she could guess aspects of general relativity by thinking to `` promote '' the flat metric to a dynamical one .
similar statements can be made about the relation between many other pairs of theories .
but retrospective examples obscure the challenge : one can not confidently know what to promote ( and how to promote it ) to enlarge the regime of validity of a given theory .
holographic entanglement entropy , proposed by ryu and takayanagi ( rt ) @xcite , proved by lewkowycz and maldacena @xcite , and made covariant by hubeny , rangamani , and takayanagi ( hrt ) @xcite , is a beautiful property ( or , in the covariant case , conjecture ) of ads / cft .
below we describe a promotion of holographic entanglement entropy beyond the scope of ads / cft that applies just as well to cosmological spacetimes as it does to asymptotically ads spacetimes . in the case of the latter , it reduces to the hrt proposal .
moreover , the promoted holographic entanglement entropy satisfies , for nontrivial reasons , expected properties of entanglement entropy like strong subadditivity .
the hrt prescription provides a way to compute entanglement entropy of a spatial region @xmath0 in a quantum state dual to an alads spacetime .
the procedure is to consider @xmath1 , the boundary of the spatial region , and to find the area of a codimension 2 extremal surface that is anchored to @xmath1 .
a nave extension of this idea to general spacetimes would be to take @xmath0 to be a region in the conformal boundary of an arbitrary spacetime .
this approach fails : what is the boundary of a closed frw universe with past and future singularities ? in our proposal , we anchor extremal surfaces to a _ holographic screen_. holographic screens are codimension 1 surfaces that appear to be the most natural place for quantum states dual to arbitrary geometries to live on .
in fact , they were proposed by bousso @xcite in an attempt to find the analogue of the ads boundary when extending holography to general spacetimes .
if one believes the covariant entropy bound @xcite , then there is essentially no other reasonable class of surfaces for this purpose .
+ _ outline_. in section [ section2 ] we first review the concept of holographic screens @xcite with an emphasis on the recent developments of bousso and engelhardt @xcite which identified a class of screens that satisfy an area monotonicity law .
we then give the definition of _ holographic screen entanglement entropy _ and list a number of its key properties .
we conclude the section by stating our _ screen entanglement conjecture_a proposal that holographic screen entanglement entropy actually measures von neumann entropy in a putative holographic description of general spacetimes .
section [ sec_ssa ] contains technical developments including proofs of the properties of screen entanglement entropy that are advertised in section [ section2 ] .
section [ sec_cosmology ] gives cosmological examples of holographic screens and their extremal surfaces .
we focus particularly on frw universes that approach de sitter space at late times .
section [ sec_conclusion ] concludes by reviewing the procedure for computing screen entanglement entropy and by suggesting extensions to our proposal such as possible methods for computing subleading contributions to holographic screen entanglement entropy .
we open this section with a brief review of holographic screens , especially past and future holographic screens .
readers that are already familiar with the content @xcite may still find it useful to read through these paragraphs to become familiar with our conventions and notation . throughout this paper
we will work in a globally hyperbolic spacetime @xmath2 of dimension @xmath3 that satisfies the null energy condition .
we assume that the spacetime satisfies the genericity conditions laid out in @xcite .
suppose that @xmath4 is an orientable spacelike codimension @xmath5 submanifold of @xmath2 .
it is possible to find an independent pair of future directed null vector fields on @xmath4 that are everywhere orthogonal to @xmath4 .
if one of these vector fields has vanishing null expansion on @xmath4 , we will say that @xmath4 is _
marginal_. if one vector field has zero expansion on @xmath4 while the other has negative ( positive ) expansion on @xmath4 , we say that @xmath4 is _ marginally trapped ( marginally anti - trapped)_. a _ past holographic screen _ is a codimension-1 submanifold @xmath6 of the spacetime that is foliated by marginally anti - trapped compact spacelike surfaces called _
leaves_. the foliation into leaves is unique : other splittings of @xmath6 can not satisfy the marginally anti - trapped condition .
future holographic screen _ is instead foliated by marginally trapped surfaces . in this paper
, we will always assume that leaves have the topology of @xmath7 .
holographic screens are generated by null foliations : if @xmath8 is a null foliation of a spacetime , it is possible to identify a family of leaves @xmath9 with @xmath10 by finding the codimension 2 surface of maximal area on each null surface . in general
, this will break the values of the parameter @xmath11 into open intervals , some of which correspond to past holographic screens and others corresponding to future screens .
, @xmath12 does not have a definite sign for @xmath13 .
we leave the investigation of this scenario to future work . ]
isolated values of @xmath11 that lie between past and future screens correspond to the case where @xmath12 is an isolated extremal sphere which can join a past and future screen .
such a sphere will not be considered to lie on a past or future holographic screen by convention .
this occurs in the case of a closed universe with a big crunch : see figure [ fig_big_crunch ] . .
one particular leaf @xmath14 is highlighted here along with its null orthogonal vector fields @xmath15 and @xmath16 satisfying @xmath17 and @xmath18 .
the causal region @xmath19 plays a critical role in our generalization of holographic entanglement entropy.,width=302 ] some of the simplest examples of holographic screens arise in the `` observer - centered '' case where we take @xmath8 to be the set of past light - cones of an observer s worldline in some spacetime . in the case of frw cosmology with the observer taken to be comoving , such holographic screens are just apparent horizons .
figure [ fig_screen_def ] shows an example of such a holographic screen .
see also figures [ ds_identifications ] ( top ) and [ fig_big_crunch ] . because past holographic screens are often generated in this way
, we will mostly focus on the case of past screens throughout this paper .
however , all results below apply equally well to future screens with appropriate modifications . because null foliations are highly non - unique , holographic screens are also non - unique . for example , in the observer - centered case , a past holographic screen can be obtained obtained by considering the surfaces of maximal area on the past light - cones of an observer s worldline if the maximal area surfaces are anti - trapped and compact which we assume . in this case
, performing a modification to the worldline will modify the holographic screen . from this point of view
, holographic screens appear to be `` pro - complementarity '' objects .
the potential importance of this aspect of screens is further discussed below .
suppose that @xmath6 is a past holographic screen .
let @xmath14 be a leaf of @xmath6 and let @xmath15 and @xmath16 denote , respectively , the ingoing and outgoing future - directed null surface - orthogonal vector fields on @xmath14 .
( it may be useful to refer to figure [ fig_screen_def ] . )
then , the condition that @xmath14 is marginally anti - trapped means that @xmath20 where @xmath21 and @xmath13 denote the expansion of congruences in the @xmath15 and @xmath16 directions at @xmath14 .
every holographic screen comes with a _
fibration_. a fibration is a family of curves generated by a nonvanishing vector field @xmath22 on and tangent to @xmath6 with the property that @xmath22 is orthogonal to every leaf .
if we extend the vector fields @xmath15 and @xmath16 to all of @xmath6 ( so that they are surface orthogonal to every leaf ) , then @xmath23 where @xmath24 and @xmath25 are scalar functions on @xmath6 .
@xmath22 is not required to be timelike , spacelike or null and , in fact , can switch between these three cases on one screen .
thus , holographic screens need not have definite signature . lacking a definite signature , normalization of @xmath22 is arbitrary . nonetheless , it is convenient to write the leaves of @xmath6 as @xmath12 where @xmath11 is some ( non - unique ) parameter and to then normalize @xmath22 by the condition @xmath26 .
bousso and engelhardt proved that @xmath27 at every point in @xmath6 and concluded that leaves have strictly increasing area @xcite .
more precisely , the area of @xmath28 is greater than the area of @xmath29 if @xmath30 .
in fact , if @xmath31 denotes the area functional , then @xmath32 which is positive by equation [ theta_cond ] and the fact that @xmath33 . here
, @xmath34 denotes the induced metric on @xmath12 .
note , in particular , that the area is _ strictly _ increasing for all intervals of @xmath11 . the inequality would not be strict if it were not for the genericity conditions of @xcite .
as before , let @xmath6 denote a past holographic screen .
everything below can be modified to the case of a future holographic screen without difficulty .
it is helpful to emphasize the following result which follows from the genericity conditions of @xcite : * _ strict focusing .
_ if @xmath4 is a codimension 2 spacelike surface , the four surface - orthogonal null congruences have strictly decreasing expansion as they move away from @xmath4 .
this means that there is always enough matter content everywhere in the spacetime to focus neighboring null geodesics .
if @xmath2 fails to satisfy this condition , it can be made to do so by sprinkling a very small amount of classical matter everywhere . as discussed above
, there is a unique foliation of @xmath6 into anti - trapped leaves .
let @xmath14 be a particular leaf in this foliation and let @xmath15 and @xmath16 denote the vector fields on @xmath14 that satisfy equation [ theta_cond ] . because @xmath2 is globally hyperbolic
, there exists a cauchy surface @xmath35
containing @xmath14 such that @xmath36 consists of a disconnected interior and exterior . the interior of @xmath35 is defined so that a vector on @xmath14 pointing toward the interior takes the form @xmath37 with @xmath38 .
let @xmath39 denote the union of the interior of @xmath35 with @xmath14 .
we will assume that @xmath39 is compact and that it has the topology of a solid ball .
now let @xmath19 be the domain of dependence of @xmath39 , @xmath40 , with the convention that @xmath19 includes orthogonal null surfaces generated by @xmath15 and @xmath41 . .
the black and red codimension 2 regions together form a single leaf @xmath14 .
the black segment represents a region @xmath0 and the extremal surface @xmath42 ( orange ) is anchored to its boundary .
the causal region @xmath19 is the green diamond ( both interior and boundary ) .
note that @xmath43.,width=302 ] suppose that @xmath0 is a @xmath44 dimensional submanifold of @xmath14 with a boundary .
consider the set of extremal codimension 2 surfaces that are anchored to and terminating at @xmath1 , and contained entirely in @xmath19 ( see figure [ fig_screen_3d ] ) . in section [ sec_ssa ] we will give conditions on @xmath19 that ensure that this set is not empty .
taking the existence of such a surface for granted , let the one of minimal area be denoted by @xmath42 and define the _ holographic screen entanglement entropy _ ( or _ screen entanglement entropy _ for brevity ) of @xmath0 as @xmath45 the quantity @xmath46 is the most natural generalization of the hrt proposal to general spacetimes .
we emphasize that we have defined screen entanglement entropy geometrically without reference to a quantum theory .
the term `` entanglement entropy '' is only meant suggestively . nonetheless , below we state a _ screen entanglement conjecture _
: that @xmath46 is in fact the von neumann entropy of a subsystem of a holographic quantum state for general spacetimes .
regardless of the validity of this conjecture , we are free to study @xmath46 as we have defined it . as we will see , the properties of holographic screens ensure that screen entanglement entropy possesses numerous properties reminiscent of von neumann entropy which we now discuss .
* _ existence and containment .
_ in section [ sec_ssa ] we provide conditions for @xmath47 to exist .
this is a nontrivial issue because of the `` containment condition '' that @xmath43 .
arguments that @xmath19 contains an extremal surface rely critically on the assumption that @xmath0 is in a leaf of a holographic screen .
moreover , the condition that @xmath48 gives rise to properties of holographic screen entanglement entropy like strong subadditivity ( see below ) and will allow us to reasonably define an entanglement wedge for @xmath0 .
for an example of the importance of the containment condition , see equation [ not_contained ] below and the paragraphs around it .
* _ ( strong ) subadditivity .
_ suppose that @xmath0 and @xmath4 are regions in @xmath14 .
then , @xmath49 where @xmath39 is the function defined in [ entropydef ] .
this result holds regardless of whether or not @xmath0 and @xmath4 intersect as long as we take the convention that @xmath50 .
as we will see in section [ sec_ssa ] , the proof of this is a modified version of wall s @xcite `` maximim '' proof for the hrt case .
this does not mean that strong subadditivity is an obvious result : most of the work in section [ sec_ssa ] is to show that the properties of leaves of holographic screens are sufficient to generalize wall s arguments to our context .
* _ page bounded .
_ define the extensive entropy of @xmath0 as @xmath51 .
then , the holographic screen entanglement entropy satisfies the following _ page bound _ : @xmath52 this is a simple consequence of the maximin construction we give in section [ sec_ssa ] .
note that the area law for holographic screens implies that this inequality becomes a weaker constraint if we transport @xmath0 along the fibration vector field defined above . in certain cases , the inequality saturates and @xmath46 approaches a `` random entanglement limit . ''
( see section [ sec_cosmology ] for examples of this in cosmology . )
* _ reduction to the hrt proposal . _
as explained in detail in @xcite , the ads boundary can be regarded as a holographic screen . in this case , surfaces of constant time in the dual field theory correspond to leaves , and our proposal becomes identical to the covariant holographic entanglement entropy conjecture of @xcite .
we are now in a position to state our conjecture about the role of @xmath46 in quantum gravity .
this conjecture is the primary concern of this paper .
nonetheless , we emphasize that the mathematical developments below ( e.g. the proof that @xmath46 satisfies standard properties of von neumann entropy ) do not rely on any conjectural statements .
our proposal can be regarded as an extension of a covariant holographic principle due to bousso which we now review very briefly . in @xcite , bousso integrated the ideas of @xcite with his covariant entropy conjecture @xcite and proposed that each marginal surface @xmath4 foliating a holographic screen is associated with a hilbert space @xmath53 of dimension @xmath54 and that states in @xmath53 holographically define the state on a null surface @xmath55 passing through @xmath4 in the marginal direction . for our purposes ,
this holographic principle takes the following form .
to each leaf @xmath14 of a past or future holographic screen we assign a density matrix @xmath56 .
the density matrix acts on a hilbert space of dimension @xmath57 which may be a subspace of a `` complete '' hilbert space .
the covariant entropy bound suggests that @xmath56 encodes the quantum information on the null slice generated by @xmath15 and @xmath58 where @xmath15 is the null vector field with @xmath17 on @xmath14 .
we now assume bousso s holographic principle and state our new conjecture .
we propose that every region @xmath0 of @xmath14 ( up to string scale resolution ) corresponds to a subsystem of the hilbert space that @xmath56 acts on .
we conjecture that the von neumann entropy of that subsystem in the density matrix @xmath56 is given , at leading order , by @xmath46 as we have defined it in equation [ entropydef ] .
we refer to this statement as the _ screen entanglement conjecture_. because a holographic quantum theory dual to arbitrary spacetimes is not known , the screen entanglement conjecture is not a mathematical statement about the relation between two known theories ( as in the case of hrt ) . instead , our conjecture suggests a way to compute properties of quantum states in an unknown theory .
it is our hope that this will , in fact , be a step toward developing a quantum theory for arbitrary spacetimes .
it was emphasized above that in a given spacetime , there is no unique preferred holographic screen . as a consequence ,
screen entanglement entropy can not even be defined before first deciding on a particular choice of a screen .
this might seem to put the screen entanglement conjecture on haphazard footing , but we explain here why this arbitrariness is , in fact , a necessary feature of any generalization of holographic entanglement entropy to general spacetimes . conventional holographic entanglement entropy in ads / cft is reference frame dependent in the following sense . consider an observer in an asymptotically ads spacetime @xmath2 with conformal boundary @xmath59 following a worldline @xmath60 . here
, @xmath61 is the proper time parameter of the observer . at a given value of @xmath61
, we can consider a spacelike _ cut _ of the boundary @xcite : @xmath62 here , @xmath63 denotes the causal past of a point @xmath64 .
a region @xmath65 on @xmath66 can be specified by considering a portion @xmath67 of a small sphere on the tip of the past light cone of the point @xmath60 and following points in @xmath67 down null geodesics until @xmath59 is reached .
thus , once the trajectory @xmath60 is decided upon , we can use the hrt formula to compute @xmath68 , the holographic entanglement entropy of the region @xmath65 on the cut @xmath66 .
if the trajectory is changed , @xmath68 correspondingly transforms . at the level of the dual cft
, this discussion corresponds to the fact that quantum states and their time - dependence have a gauge - redundancy that is fixed by making a choice of time - slicing on the boundary @xmath59 . in the case of the screen entanglement conjecture and a spacetime that is not asymptotically ads ,
a null foliation must be selected to fix a holographic screen . as discussed above , a simple way to do
this is to choose a curve @xmath60 , and , at any given @xmath61 , follow along the past light - cone of @xmath60 until a marginal surface @xmath69 is obtained .
the role of @xmath69 is analogous to that of the cut @xmath66 in asymptotically ads spacetimes .
the foliation dependence of screen entanglement entropy is closely tied to the frame - dependence of the hrt formula .
this is an example of `` fixing the gauge '' in quantum gravity , a concept developed in @xcite . in the case of asymptotically ads spacetimes ,
no matter what worldline @xmath60 is chosen , the union of all of its cuts will always be a subset of the boundary .. however , in some cases past directed null geodesics of @xmath60 may fail to reach @xmath59 . ] in general spacetimes however , the particular holographic screen obtained by taking the union over all @xmath61 of @xmath69 will depend on the choice of the worldline .
thus , the surface on which holographic quantum states are defined is no longer tethered to the spacetime .
this is a basic property of bousso s holographic principle , not one that arises only in the more extended framework of this paper .
we regard this aspect of holographic screens as being in the spirit of black hole complementarity , where quantum information is not attached to a fixed spacetime position ( e.g. a qubit is not inside or outside a black hole ) until an observer is selected to describe the system .
in this section we prove key technical results about holographic screen entanglement entropy including many of the properties advertised above .
the notation and conventions we will use are the same as those given in section [ section2 ] . in particular , @xmath6 is a past holographic screen in a globally hyperbolic spacetime of dimension @xmath3 that satisfies the genericity conditions of @xcite .
@xmath14 is a compact leaf of @xmath6 which we assume to have the topology of @xmath7 .
@xmath15 and @xmath16 are null orthogonal vector fields on @xmath14 satisfying equation [ theta_cond ] .
@xmath35 is a cauchy slice containing @xmath14 and @xmath39 is the portion of @xmath35 that is enclosed by @xmath14 including @xmath14 itself ( the enclosed side is defined in section [ section2 ] ) .
@xmath39 is assumed to have the topology of a compact @xmath70 ball .
@xmath19 is the domain of dependence of @xmath39 . as always , the case of a future holographic screen is omitted because it presents no additional subtlety . as discussed in section [ section2 ] , it is nontrivial and critical to show the existence of an extremal surface anchored to @xmath1 that lies entirely in @xmath19 .
we now prove that such a surface exists under very generic conditions .
our first step is to show that @xmath42 exists in the case that @xmath19 is compact .
this is a common situation terminate at caustics rather than singularities .
let @xmath71 and @xmath72 denote the set of the first caustics encountered ( local or nonlocal ) by null geodesics in the future and past light sheets respectively . then ,
if @xmath73 , we can conclude that @xmath19 is compact .
this follows from the fact that @xmath74 inherits the compactness of @xmath14 and from the fact that global hyperbolicity implies that @xmath75 is compact if @xmath76 and @xmath77 are compact . ]
although it is not the case if the ingoing light sheets of @xmath14 encounter a singularity .
involves a continuous family of surfaces @xmath78 along with their extremal surfaces ( dotted curves).,width=302 ] [ inside_lem ] if @xmath19 is compact , then there exists a codimenson 2 extremal surface anchored and terminating at @xmath1 that lies entirely in @xmath19 and that intersects @xmath79 only at @xmath1 .
let @xmath80 and @xmath81 denote the future and past ingoing light - sheets of @xmath14 .
we now extend @xmath81 to a slightly larger light - sheet , @xmath82 , by following the future directed null congruence of @xmath16 . because @xmath18 on @xmath14 , we can make this extension so that @xmath82 has @xmath18 everywhere and so that there exists an open set in @xmath82 containing @xmath14 . in the language of @xcite , both @xmath83 and @xmath82 are extremal surface barriers because they have negative expansion in the @xmath15 and @xmath41 directions respectively .
moreover , @xmath84 .
it follows that @xmath85 is itself an extremal surface barrier for extremal surfaces in the interior , i.e. @xmath86 , is analogous to exterior regions studied by wall and engelhardt .
] of @xmath19 . now consider the region @xmath0
. the spherical topology .
in particular , we are making simplifying topological assumptions about @xmath14 and @xmath39 .
we will leave it to future work to investigate the consequences of relaxing these assumptions .
] of @xmath14 ensures that it is possible to introduce a continuous one - parameter family of submanifolds of @xmath19 , @xmath78 , such that * @xmath87 consists of a single point in the interior of @xmath19 * @xmath88 * for @xmath89 , @xmath78 is a codimension 2 submanifold of the interior of @xmath19 that is diffeomorphic to @xmath0 .
this is shown in figure [ fig_lemma_1 ] .
note , in particular , that if @xmath90 , @xmath91 . if @xmath92 is sufficiently small , then the extremal surface of minimal area that is anchored to @xmath93 lies entirely in the interior of @xmath19 .
denote this extremal surface by @xmath94 .
consider increasing the value of the parameter @xmath95 from @xmath96 to @xmath97 .
for each value of @xmath95 , construct an extremal surface @xmath98 ( not necessarily the one of minimal area ) anchored to @xmath99 . the compactness of @xmath19 ( which ensures that it is bounded and has no singularities ) together with the fact that , as discussed above , @xmath85 is an extremal surface barrier , allows us to take @xmath98 to not jump discontinuously and to be contained in the interior of @xmath19 for all @xmath90 .
when we take the limit sending @xmath95 to 1 , the extremal surface anchored to @xmath1 must intersect @xmath85 at @xmath1 and nowhere else : if it did intersect @xmath85 outside of @xmath1 , the extremal surface would be locally tangent to an extremal surface barrier with strictly nonzero null extrinsic curvature .
is the shaded region along with its boundary , the blue and orange lines .
@xmath100 consists of two parts : an extremal surface barrier @xmath4 ( blue ) and a portion of @xmath85 ( orange ) . in this figure ,
the barrier @xmath4 protects extremal surfaces in @xmath101 from a singularity .
not shown are extremal surfaces in @xmath101 , none of which contact @xmath100 except at their anchor on the leaf @xmath14.,width=302 ] the unwanted assumption that @xmath19 is compact ( which fails in the event that @xmath80 or @xmath81 encounter a singularity ) can be dropped if there exists a codimension 0 submanifold ( with boundary ) of @xmath19 , @xmath101 , which `` restricts '' extremal surfaces ( see figure [ fig_compact_restriction ] ) . by this
we mean that 1 .
@xmath101 is compact , 2 .
there exists an open set @xmath102 containing @xmath39 with @xmath103 , and 3 .
@xmath104 where @xmath4 is an extremal surface barrier for codimension 2 extremal surfaces inside in @xmath101 .
these conditions for @xmath101 are designed to ensure that @xmath101 can be used in lemma [ inside_lem ] in place of @xmath19 without difficulty .
the existence of such regions @xmath101 relies on the existence of the barrier @xmath4 .
the arguments in theorem 11 of @xcite show that kasner singularities are always protected by such barriers .
hartman and maldacena @xcite encountered a barrier protecting black hole singularities from codimension 2 extremal surfaces .
constant time slices in frw spacetimes are another example of suitable barriers . in the proof of lemma [ inside_lem ]
. as a region @xmath78 is deformed from a point inside @xmath101 into @xmath105 , extremal surfaces anchored to @xmath99 can not smoothly pass @xmath4 or @xmath85 . ]
any region @xmath106 satisfying the conditions will be called a _ compact restriction _ of @xmath19 .
note that , in particular , if @xmath19 is compact then @xmath19 is a compact restriction of itself .
our findings can now be summarized by the following improvement upon lemma [ inside_lem ] : [ inside ] if @xmath19 possesses a compact restriction , then there exists a codimenson 2 extremal surface anchored and terminating at @xmath1 that lies entirely in @xmath19 and that intersects @xmath79 only at @xmath1 . to better appreciate this theorem ,
it is helpful show that the statement is false if @xmath14 is not a leaf of a holographic screen . consider @xmath107 dimensional minkowski space with inertial coordinates @xmath108 and let @xmath109 denote the large cylinder satisfying @xmath110 with @xmath111 .
consider the two line segments on @xmath109 that are approximately given by @xmath112 and construct any spacelike `` time slice '' on @xmath109 , @xmath14 , that includes @xmath113 .
it is easy to see that the extremal surface anchored to @xmath114 is a straight line that is timelike related to @xmath113 and thus fails to lie within the domain of dependence @xmath19 . to see how severe this problem is ,
note that the segments @xmath0 and @xmath4 fail to satisfy subadditivity of entanglement entropy .
that is , the inequality @xmath115 is false .
note that in this example @xmath14 fails to satisfy equation [ theta_cond ] because of the kink at @xmath116 .
theorem [ inside ] ensures that holographic screen entanglement entropy is a well - defined quantity in a broad set of cases .
we will now demonstrate that this quantity satisfies expected properties of entanglement entropy .
to do this , it is very useful to closely follow @xcite and introduce a maximin construction of @xmath117 .
our construction will be slightly modified from that used for hrt surfaces anchored to the ads boundary .
wall s maximin prescription involves considering a collection of cauchy slices that are anchored only to @xmath1 . because we already know
that @xmath117 lies inside of @xmath19 , we will introduce a stronger constraint requiring that we only consider achronal slices that are anchored to all of @xmath14 .
our setup remains unchanged .
fix a past ( or future ) holographic screen @xmath6 in a globally hyperbolic spacetime and let @xmath14 be a leaf .
we take a cauchy surface @xmath35 containing @xmath14 and define @xmath39 as the closure of the portion of @xmath35 inside of @xmath14 .
as before , we require that @xmath39 is compact and that it has the topology of a solid @xmath70 ball .
let @xmath40 .
we also fix a region @xmath0 in @xmath14 with a boundary . now define @xmath118 as the collection of codimension 1 compact achronal surfaces that are anchored to @xmath14 and that have domain of dependence @xmath19 .
note , in particular , that @xmath119 . moreover , note
that the global hyperbolicity of @xmath19 ensures that every element of @xmath118 has the same topology as @xmath39 : that of a compact @xmath70 ball .
take any @xmath120 .
let @xmath121 denote the codimension 2 surface of minimal area wall @xcite added the condition that @xmath121 be homologous to @xmath0 . while this condition ought to be included in our discussion as well , the assumption that @xmath39 ( and thus every element of @xmath118 ) has the topology of a compact @xmath70 ball makes a homology condition trivial .
we leave the task of investigating more general topologies to future work . ] on @xmath122 that is anchored to @xmath1 .
the existence of @xmath121 is guaranteed by the compactness of @xmath122 and theorem 9 of @xcite .
define a function @xmath123 $ ] by @xmath124 .
now assume that there exists a @xmath125 in @xmath118 that maximizes @xmath126 ( globally ) .
we now define @xmath127 as the maximin surface of @xmath0 , and we will denote it by @xmath128 .
if there are several maximin surfaces , @xmath129 can refer to any of them .
the existence of @xmath129 can be proven in many cases by appropriately importing the arguments of theorems 10 and 11 in @xcite which we only briefly describe here .
consider the cauchy surface @xmath35 which can be identified as a slice in a foliation of cauchy surfaces @xmath130
. using this definition of time , we can identify a surface @xmath131 with a function @xmath132 in a natural way : if @xmath133 denotes the integral curve of @xmath134 that passes through a point @xmath135 , then @xmath136 . from this viewpoint
, @xmath126 can be regarded as a real - valued functional on @xmath137 .
now if @xmath19 is compact , we can find the maximum and minimum values of @xmath138 for the set @xmath19 to obtain an upper and lower bound on @xmath139 that applies for all @xmath122 .
moreover , the condition that @xmath122 be compact and achronal ensures that @xmath137 is equicontinuous .
these facts imply that @xmath118 is compact ( with the uniform topology ) and that the extreme value theorem applies to the function @xmath126 . in the case where @xmath19 is not compact ( for instance
, due to a singularity terminating a light sheet of @xmath14 ) , we can still argue that @xmath126 has a maximum as long as @xmath19 satisfies a condition similar to but slightly stronger than the `` compact restriction '' idea discussed above .
suppose that @xmath140 is a surface in @xmath118 which is identical to @xmath80 in some neighborhood of @xmath39 .
for any @xmath131 , define another surface @xmath141 by @xmath142 .
if @xmath140 has the property that for any @xmath122 we have @xmath143 , then we will say that @xmath140 is a _ future maximin barrier_. a past maximin barrier is defined analogously as a surface @xmath144 , identical to @xmath81 in a neighborhood of @xmath39 , such that for any @xmath122 we have @xmath143 where @xmath141 is defined by @xmath145 .
now if @xmath19 possesses both a past and future maximin barrier , then we can restrict our attention to the subset of surfaces in @xmath118 that satisfy @xmath146 .
let @xmath147 denote this restricted set .
because @xmath148 and @xmath140 are compact , @xmath149 is compact and so the set @xmath147 is compact in the uniform topology and @xmath126 has a maximum @xmath150 .
the definition of past and future maximin barriers ensures us that if @xmath131 , then @xmath151 .
thus , @xmath125 is a global maximum for @xmath126 and we can safely define @xmath152 as the maximin surface of @xmath0 , @xmath128 . as in the case of the compact restriction of @xmath19 used in theorem [ inside ] , it is difficult to find examples where @xmath19 does not possess a past and future barrier .
wall @xcite argued that such barriers protect maximin surfaces from a wide range of singularities : approximately kasner singularities , bkl singularities , and frw big bangs all lead to past or future maximin barriers .
if @xmath153 simply terminate at caustics rather than singularities , then @xmath154 are barriers . in any event , if @xmath155 exist , then the region @xmath149 provides a compact restriction of @xmath19 in the sense of theorem [ inside ] .
thus , the existence of @xmath155 ensures both the existence of @xmath129 as well as the existence of @xmath47 . from here on , we will simply take for granted that a past and future maximin barrier exist .
below we will argue that @xmath156 .
however , it is first very useful to introduce two additional definitions first . 1 .
take @xmath131 and let @xmath157 be a codimension 2 surface anchored to @xmath1 that lies in @xmath19 .
consider the intersection between @xmath122 and the future and past - directed orthogonal null surfaces of @xmath157 that are directed toward @xmath0 .
this intersection is called the representative of @xmath157 on @xmath122 and will be denoted by @xmath158 .
the domain of dependence of codimension 1 achronal surfaces anchored to @xmath159 lying in @xmath19 will be called the entanglement wedge of @xmath0 . note
that @xmath158 is itself a codimension 2 surface anchored to @xmath1 that lies on @xmath122 .
moreover , if @xmath157 is extremal then , by the focusing theorem , @xmath160 .
we now demonstrate that our maximin procedure always finds @xmath117 , the extremal surface of minimal area that is anchored to @xmath1 and which lies in @xmath19 .
while much of the proof is similar to the arguments in @xcite , we will have to pay special attention to the possibility that the maximin surface could run into the boundary of @xmath19 .
[ mmext ] @xmath161 .
the argument of theorem 15 in @xcite immediately shows that if a point @xmath162 is also in the interior of @xmath19 ( i.e. @xmath86 ) , then @xmath129 is extremal at @xmath163 . in particular ,
if @xmath164 , then @xmath129 is an extremal surface everywhere .
we now argue that @xmath128 in fact can not ever intersect @xmath85 outside of @xmath1 .
suppose there exists @xmath165 .
there must be an open neighborhood of @xmath163 in @xmath129 ( open in the d-2 dimensional manifold @xmath129 ) that is entirely contained in @xmath85 .
if this were not the case , @xmath129 would be extremal at points arbitrarily close to @xmath163 and would thus be extremal at @xmath163 .
moreover , @xmath129 would be tangent to @xmath85 at @xmath163
. however , @xmath85 is an extremal surface barrier ( see lemma [ inside_lem ] ) so this is not possible .
there are now two cases to consider . .
note that the surface @xmath39 is shown here for reference and that it does not play a critical role in the proof .
the shaded region is @xmath40 and the green dot is ( a cross - section of ) the leaf @xmath14.,width=302 ] * _ case 1 _ : @xmath166 .
+ figure [ fig_mm_proof_case_1 ] illustrates a construction that we will use for this case .
take @xmath167 ( the case of @xmath168 is no different ) . by construction
, @xmath129 is minimal on a surface @xmath125 .
there exists a ( dimension d-1 ) open subset @xmath102 of @xmath125 containing @xmath163 such that @xmath169 .
moreover , we can require that @xmath102 is `` split '' by @xmath128 into two disconnected sets , @xmath55 and @xmath170 , such that @xmath55 is the side of @xmath102 closer to @xmath14 .
since @xmath125 is anchored to @xmath14 , we must have that @xmath171 and , in particular , @xmath55 is null . on the other hand , @xmath170 can not be a subset of @xmath80 .
if it were , then @xmath129 could decrease its area by being deformed up @xmath80 ( by the focusing theorem ) .
in particular , we can take @xmath102 small enough to ensure that @xmath170 is nowhere null in the direction of @xmath15 .
+ we now consider the process of slightly sliding @xmath125 down @xmath80 . more precisely ,
take a small parameter @xmath172 and a corresponding one - parameter family of slices @xmath173 that are slightly deformed from @xmath125 in a way we now describe ( an example of @xmath174 is depicted in figure [ fig_mm_proof_case_1 ] by an orange dashed line ) .
the surface @xmath175 is described by a function @xmath176 giving the affine distance from @xmath14 up to @xmath129 at a point @xmath177 . now put @xmath178 . here , @xmath179 $ ] is a smooth weighting function which equals 1 at the null generator @xmath180 that @xmath163 lies on .
we take @xmath181 to go to zero smoothly as @xmath182 moves away from @xmath180 , equaling zero exactly when @xmath182 corresponds to a point outside of @xmath183 . for @xmath184 , we require that the surface @xmath174 is identical to @xmath80 .
we extend @xmath174 beyond @xmath185 by parallel transporting tangent vectors on @xmath129 directed toward @xmath170 down to @xmath185 .
this prescription does not uniquely fix @xmath174 , but it is sufficient for our purposes . + consider the one - parameter family of codimension 2 curves @xmath186 .
for any @xmath92 , let @xmath187 denote the future - directed null congruence of @xmath186 that points toward the interior of @xmath19 ( see figure [ fig_mm_proof_case_1 ] ) .
the continuity of @xmath186 as @xmath96 varies and the fact that @xmath80 is a light sheet ensures that there exists and @xmath188 such that for @xmath189 , * * @xmath187 intersects @xmath125 to form a codimension 2 surface on @xmath125 anchored to @xmath1 and * * @xmath187 has negative future - directed expansion in the region between @xmath186 and its intersection with @xmath125 .
+ denote this intersection by @xmath190 and observe that @xmath191 .
but @xmath129 is minimal on @xmath125 so for sufficiently small @xmath96 , @xmath192 which contradicts the assumption that @xmath129 has area greater than or equal to the minimal area surface on any slice .
note that the last inequality above follows from the focusing theorem applied to @xmath187 .
* _ case 2 _ : @xmath193 .
+ assume that there exists a ( dimension d-2 ) open subset of @xmath128 that is contained in @xmath14 .
( if not , there must be such an open set in @xmath194 which just leads to case 1 above . )
now consider the null vector field @xmath15 on @xmath14 and the geodesics generated by it .
follow these geodesics from @xmath14 up along @xmath80 by a short affine distance @xmath92 to generate a new codimension 2 surface , @xmath195 , which limits to @xmath14 when @xmath196 .
the focusing theorem now gives rise to a modified version of equation [ theta_cond ] at @xmath195 : + @xmath197 + along with moving @xmath14 up the light - sheet , we also translate @xmath0 up the sheet to a one - parameter family of surfaces @xmath198 that limit to @xmath0 .
consider the maximin construction applied to codimension 2 surfaces anchored to @xmath93 that lie on codimension 1 surfaces anchored to @xmath195 .
we denote the result by @xmath199 .
we also define @xmath200 in the obvious way .
now this maximin procedure leads to the same two cases that we are now studying .
the first case , where @xmath199 intersects @xmath201 proceeds exactly as it did with @xmath202 .
now suppose that @xmath203 has an open set contained in @xmath195 .
@xmath203 must be minimal on some slice @xmath174 .
however , equation [ theta_cond_ep ] implies that @xmath195 has negative ( inward ) extrinsic curvature on @xmath174 .
it is thus impossible for @xmath199 to be minimal on @xmath174 since its area could be decreased by `` cutting corners . ''
+ we can thus conclude that @xmath204 .
this implies that @xmath199 is extremal .
taking the limit as @xmath196 , we conclude that @xmath129 is extremal .
but , given our assumption that part of @xmath129 lies on @xmath14 , equation [ theta_cond ] shows that @xmath129 can not be extremal since extremal surfaces have zero null expansion in all directions . at this point
it is proven that @xmath129 is extremal .
all that is left is to show that , of all the extremal surfaces in @xmath19 that are anchored to @xmath1 , @xmath129 is the smallest .
let @xmath205 be a slice on which @xmath129 is minimal .
if @xmath157 is another extremal surface anchored to @xmath1 then , as a result of the focusing theorem , we find that @xmath206 we are now in a position to prove a variety of properties of screen entanglement entropy .
we begin with the `` page bound '' advertised in section [ section2 ] .
[ thermal_bound_thm ] if @xmath0 is a region in the leaf @xmath14 , then @xmath207 where @xmath39 deonotes the holographic screen entanglement entropy of @xmath0 and @xmath208 denotes the area of a region @xmath209 divided by 4 .
@xmath210 but @xmath211 for some @xmath205 . both @xmath0 and @xmath212
are codimension @xmath44 dimensional surfaces on @xmath125 anchored to @xmath1 so the area of @xmath128 is less than or equal to the areas of both @xmath0 and @xmath212 .
next we turn to the proof of strong subadditivity for holographic screen entanglement entropy ( other properties of entanglement entropy that admit covariant geometrical bulk proofs can be imported here as well ) . unlike the case of theorems [ inside ] and [ mmext ] ,
the arguments below are essentially identical to those of @xcite with little additional subtlety .
we start with our version of theorem 17 in @xcite which states that if @xmath213 , then @xmath117 lies `` outside '' of @xmath214 .
[ same_surface ] suppose that @xmath0 and @xmath4 are regions in the leaf @xmath14 with @xmath213 .
then , 1 .
the entanglement wedge of @xmath0 contains the entanglement wedge of @xmath4 , 2 .
there exists a surface in @xmath118 on which both @xmath117 and @xmath214 are minimal .
the proof is the same as that of theorem 17 of @xcite so we only sketch it here . for any surface in @xmath131 , consider a pair of codimension 2 surfaces constrained to lie on @xmath122 , @xmath215 and @xmath216 , such that @xmath215 is anchored to @xmath1 and @xmath216 is anchored to @xmath217 .
then let @xmath218 .
we now minimize the value of @xmath219 by varying over all possible choices of @xmath215 and @xmath216 .
after that , we maximize the minimal values of @xmath219 by varying over all possible @xmath122 . this new maximin procedure gives a well - defined answer for the maximinimal value of @xmath219 .
moreover , a slice @xmath125 results on which both @xmath215 and @xmath216 are minimal . on this slice ,
it is impossible for @xmath215 to cross @xmath216 as this would necessarily give rise to a surface on @xmath125 anchored to @xmath1 with smaller area than @xmath215 .
a further observation is that if a connected component of @xmath0 is distinct from a component of @xmath4 , the corresponding connected components of @xmath215 and @xmath216 can not come into contact even tangentially .
the argument for this is that the component of @xmath216 would necessarily have a different trace of its spatial extrinsic curvature than @xmath215 at points close to the contact point .
this would mean that either @xmath215 or @xmath216 is not minimal on @xmath125 .
at this point it is known that components of @xmath215 or @xmath216 that are distinct have neighborhoods in @xmath125 that do not intersect the other surface . within such neighborhoods , small deviations @xmath125 and the minimal surfaces
can be made that prove that such surfaces are extremal .
the only remaining step is to show that , in fact , @xmath215 and @xmath216 are the extremal surfaces in @xmath19 of minimal area . if @xmath220 is an extremal surface in @xmath19 anchored to @xmath1 , then its representation on @xmath125 must have larger area than that of @xmath215 but smaller area than that of @xmath220 .
thus , @xmath221 . similarly , @xmath222 . by construction ,
both are minimal on the same surface @xmath205 . moreover , because @xmath125 is achronal , we must have that the entanglement wedge of @xmath0 contains that of @xmath4 .
[ ssa ] suppose that @xmath0 , @xmath4 , and @xmath223 are nonintersecting regions in @xmath14 .
then , @xmath224 where @xmath225 denotes @xmath226 and where the function @xmath39 is defined in equation [ entropydef ] . by theorem [ same_surface ]
, we can find a surface @xmath205 such that @xmath214 and @xmath227 are both minimal on @xmath125 .
let @xmath228 and @xmath229 denote the areas of the representations of @xmath230 and @xmath231 on @xmath125 .
then , @xmath232 where the first inequality follows from the focusing theorem and the second inequality follows from the standard geometric proof of strong subadditivity @xcite .
note that the inequality @xmath233 follows as a special case of this result .
the conventional holographic entanglement entropy prescription , with its limitation to asymptotically locally ads spacetimes , provides very little information about entanglement structure in cosmology .
one of the most intriguing applications of our proposal , therefore , is to calculate holographic screen entanglement entropy in frw universes . assuming the screen entanglement conjecture , the calculations below give the entanglement entropy of subsystems in quantum states that are dual to cosmological spacetimes .
first we review the holographic screen structure of frw spacetimes .
consider a homogeneous and isotropic spacetime with the metric @xmath234 where @xmath235 or @xmath236 in the open , flat , and closed cases respectively . before computing extremal surfaces
we must decide upon a null foliation for the spacetime and then identify the corresponding holographic screen .
null foliations ( and thus holographic screens ) are highly nonunique .
the foliation we will consider here is that of past light cones from a worldline at @xmath237 . to find the holographic screen for this foliation , it is convenient to introduce a conformal time coordinate @xmath238 such that @xmath239
then , the past light cone of the point @xmath240 satisfies @xmath241 .
spheres along the past light cone can be parameterized by the coordinate @xmath238 , and their area is given by @xmath242 assuming that @xmath243 is not merely a coordinate singularity , the condition that @xmath244 is maximized is equivalent to the condition that @xmath245 .
thus , equation [ area_eta ] gives the condition that fixes the holographic screen : @xmath246 the codimension 1 surface defined by this constraint may be timelike , spacelike , or null , depending on the particular choice of frw spacetime .
the foliating leaves of this holographic screen are spheres of constant @xmath61 and comoving radius @xmath247 satisfying equation [ screen_constraint ] .
the covariant entropy bound implies that each leaf has sufficient area to holographically encode the information on one past light cone from the worldline at @xmath237 @xcite .
let @xmath69 be the leaf of the holographic screen at time @xmath61 and let @xmath248 denote the energy density in the universe ( measured by comoving observers ) at time @xmath61 .
then , one can write a simple expression for the area of a leaf of the holographic screen at time @xmath61 , valid for any @xmath181 : @xmath249 in particular , this expression shows that holographic screens grow in area as the universe expands .
consider 3 + 1 dimensional de sitter space of radius @xmath24 .
this spacetime is @xmath250 with the metric @xmath251 where @xmath252 is the metric on a unit 3-sphere .
despite the fact that this spacetime has the form of equation [ frw_metric ] ( with @xmath253 ) , it is an awkward setting for the consideration of holographic screens : the null expansion on the past or future light cones of any point in de sitter space goes to zero only at infinite affine parameter .
this suggests that the appropriate `` boundary '' of de sitter space is past or future infinity .
even if we do attempt to anchor extremal surfaces to spheres at infinity , the analysis in section [ sec_ssa ] fails to apply because of the assumption made there that leaves are compact .
fortunately these difficulties can be averted completely by considering an frw spacetime that asymptotically approaches de sitter space at late times .
specifically , we will consider a spacetime of the form of equation [ frw_metric ] with vacuum energy density @xmath254 and , in addition , some matter content @xmath255 with the property the matter content gives rise to a big bang at @xmath256 and dilutes completely as @xmath257 .
equation [ screen_area ] immediately implies that @xmath258 where @xmath259 . because of the big bang singularity
, we must have that @xmath260 .
thus , by the area law for holographic screens @xcite , we can conclude that the leaves of our screen are spheres that monotonically increase in area , starting with @xmath261 area at the big bang , and expanding to approach the de sitter horizon of area @xmath262 at late @xmath61 .
now focus on a late time leaf @xmath69 . as discussed in section [ sec_ssa ] , given a region @xmath263 with a boundary , we can determine the holographic screen entanglement entropy of @xmath0 , @xmath46 , by considering an extremal surface anchored to and terminating at @xmath1 .
in the notation of section [ sec_ssa ] , @xmath264 is compact so theorem [ inside ] implies that an extremal surface anchored to @xmath1 exists and lies inside of @xmath264 .
for any time @xmath61 , define @xmath265 we allow this definition to extend to a function @xmath266 where @xmath0 is a region on a 2-sphere of radius @xmath24 .
this @xmath267 case is defined exactly as in equation [ thermal_entanglement ] if we take @xmath268 .
below we will present an argument that if @xmath263 , then @xmath269 ( note that in this limit , it is implied that @xmath0 is transported to later and later leaves . ) thus , we will find that as @xmath270 , @xmath46 approaches the random entanglement limit discussed in section [ section2 ] .
any interpretation of this result is necessarily speculative .
nevertheless , if one assumes the screen entanglement conjecture , then equation [ ds_entanglement ] implies that the the quantum state of an frw universe asymptotically approaching de sitter space has the property that its @xmath271 degrees of freedom are almost randomly entangled with one - another . at earlier times , the degrees of freedom are not randomly entangled because @xmath272 .
we now present a combination of rigorous arguments , numerical data , and analytic approximations suggesting that the approximate de sitter cosmological spacetimes discussed above saturate the random entanglement bound in the @xmath270 limit .
as before , @xmath69 denotes a leaf at time @xmath61 in an frw universe with vacuum energy as well as matter energy that dilutes at late time .
the entire region @xmath264 has a metric that can be made arbitrarily similar to that of a patch of empty de sitter space by making @xmath61 large . to see this , first note that points in @xmath264 have @xmath273 and @xmath274
can be made arbitrarily small by making @xmath61 large .
( this follows from equation [ ds_area_limit ] and the fact that @xmath275 .
) meanwhile , the conformal diagram for our spacetime immediately shows that the minimal value of @xmath61 in @xmath264 can be made arbitrarily large by making @xmath61 large .
thus @xmath264 can be made to only cover arbitrarily large @xmath61 and arbitrarily small @xmath247 , in which case our metric of equation [ frw_metric ] takes the form @xmath276 where @xmath277 is a constant and @xmath24 is the same constant as before .
here we have made use of the friedmann equations .
the right - hand side of this equation is precisely the metric of de sitter space in flat slicing .
de sitter space can also be described in static coordinates that make a time - translation killing vector field manifest : @xmath278 fortunately , @xmath264 lies in a region that is well - described by either the flat or static slicing of equations [ flat_ds ] and [ static_ds ] respectively
. we can now identify @xmath264 with a region @xmath279 where @xmath279 denotes a corresponding region in exact de sitter space obtained by finding a sphere @xmath280 in the static patch with area matching that of @xmath69 . while it may seem natural to put @xmath280 at large static time
, we can use the @xmath138 translational symmetry of de sitter space to place @xmath280 at @xmath281 for all @xmath61 .
the effect of increasing @xmath61 is simply to bring @xmath280 closer to the bifurcation sphere on the de sitter horizon .
this identification is illustrated in figure [ ds_identifications ] .
note that as @xmath270 , the geometry of @xmath282 and @xmath279 become arbitrarily similar .
consider the region @xmath263 which can be identified with a region @xmath283 . at large @xmath61 ,
@xmath284 approaches the equator of a 3-sphere of radius @xmath24 .
the equator itself is an extremal surface so with @xmath285 but still large , there must be an extremal surface that is close to @xmath286 but not exactly on it
. its area will be slightly less than that of @xmath286 .
note , moreover , that if the area of @xmath286 exceeds half the area of the equator , then a smaller extremal surface can be obtained by considering the complement of @xmath0 .
this suggests but does not yet prove that at large @xmath61 , the holographic screen entanglement entropy of @xmath0 is almost equal to a fourth of its own area in planck units if @xmath0 has less area than half of the de sitter horizon .
what we have proven so far is that an extremal surface exists with area almost equal to that of @xmath0 ( or @xmath287 ) .
what if there is another extremal surface with smaller area than the one we have found ?
it is easy to see that this is impossible . following the notation in section [ sec_ssa ] ,
consider the spacelike surface @xmath125 that , after mapping to @xmath288 , lies at static time @xmath281 , and that and terminates at @xmath289 .
( @xmath125 is most of a hemisphere of the 3-sphere . )
the riemannian geometry of @xmath290 shows that the surface of minimal area anchored to @xmath1 is the one we have already found .
if @xmath157 is another extremal surface ( not necessarily lying on @xmath125 ) , then its representation on @xmath125 , @xmath291 , necessarily has larger area than the extremal surface close to the horizon .
but @xmath292 so we conclude that @xmath157 does not have minimal area .
the arguments above show that the random entanglement limit is saturated at large @xmath61 . taking @xmath293 and @xmath294
, we now explain a way to obtain a more accurate estimate for @xmath46 than @xmath295 .
calculating @xmath46 without taking the large @xmath61 limit is more involved than what was done above . nonetheless , it is worthwhile to investigate this case to better understand how the page bound limit is approached .
in particular , it is of interest to understand how the discontinuity of the derivative of @xmath296 arises .
is half of a static slice in empty de sitter space and serves as a good approximation for @xmath264 at large @xmath61 .
the blue 2-sphere ( appearing as a circle here ) lies at constant @xmath297 ( equivalently , constant @xmath11 where @xmath11 is the radial coordinate in equation [ static_ds ] ) .
this 2-sphere is an approximation for the leaf @xmath69 .
green surfaces depict extremal spherical caps on @xmath290 that approximate @xmath298 for various values of @xmath299 .
the many samples of extremal surfaces shown here have evenly spaced values of @xmath299 .
figure [ fig_cc_dust ] provides evidence that this static sphere approximation is accurate at late @xmath61.,width=302 ] we begin by further discussing the role of the 3-sphere in de sitter space .
figure [ fig_3_sphere ] depicts a hemisphere of an @xmath290 of radius @xmath24 which is precisely half of a static slice of de sitter space ( which we can freely take to be @xmath281 ) .
define a parameter @xmath297 as @xmath300 where @xmath11 is the static radius appearing in equation [ static_ds ] .
note that a surface of constant @xmath297 ( and static time ) is an @xmath301 of area @xmath302 .
this suggests a way to obtain an approximation for @xmath46 if @xmath0 is a region in the leaf @xmath69 . rather than taking @xmath0 to be a region in @xmath69 , we take figure [ ds_identifications ] seriously and map @xmath0 to a region in the @xmath301 of constant @xmath303 which ensures that this @xmath301 has the same area as @xmath69 . after this mapping
is made , one computes @xmath46 by finding the extremal surface on the @xmath290 that is anchored to @xmath1 ( which we take to lie at constant @xmath297 ) . below we will refer to this procedure as the `` static sphere approximation . ''
consider regions in @xmath69 that are spherical caps .
such a cap can be fixed ( up to @xmath304 rotation ) by a zenith opening angle angle @xmath299 , so we will denote our region of @xmath69 by @xmath305 .
( with this notation , @xmath306 is a hemisphere and @xmath307 is the entire leaf . )
the static sphere approximation makes it is clear that for @xmath308 , @xmath298 is close to @xmath305 itself and that for @xmath309 , @xmath298 approaches @xmath310 .
as @xmath299 passes the transition angle @xmath311 , @xmath298 quickly passes over the top of the 3-sphere of radius @xmath24 .
the closer @xmath312 is to @xmath262 , the faster @xmath298 passes over the top of the sphere .
this explains how the discontinuity in the derivative of @xmath266 arises in the large @xmath61 limit .
, there is always another extremal surface on the 3-sphere which goes around the sphere the wrong way .
this surface always has area greater than @xmath298 and , in any case , fails to lie in @xmath264 .
however , if we consider the @xmath267 limit , then @xmath298 does not smoothly pass over the hemisphere of the 3-sphere , and in this case , the discontinuity in the derivative of @xmath266 is explained by the fact that the surface that wraps around the sphere the `` wrong way '' is now precisely the complement of @xmath305 in the equator . if @xmath299 exceeds @xmath311 in this case , then the complement of @xmath305 has smaller area than @xmath305 .
we see that a phase transition occurs only in the exact @xmath267 limit . ]
because the geometry of @xmath290 is simple , it is not difficult to obtain an explicit ( if cumbersome ) expression for @xmath313 in the static sphere approximation : @xmath314 \bigg)\ ] ] where @xmath297 is given by equation [ z_def ] and , as before , @xmath315 .
this expression can be thought of as giving a correction to the `` zeroth order '' expression @xmath316 .
taking @xmath285 will lead to corrections in @xmath317 that are not described by the static sphere method .
it is an open question as to whether or not such corrections can , in principle , be of the same ( or greater ) order in @xmath317 as the one we have studied here .
however , numerical data that suggests that the static sphere approximation is accurate at large @xmath61 as we will now see . as explained above , the cosmological spacetimes we have been discussing have vacuum energy @xmath254 as well as some matter content that dilutes at late time .
the simplest case of this is when the universe is flat ( @xmath318 ) and when the additional matter content consists of only one species with density @xmath319 and pressure @xmath320 .
the scale factor for this case is @xmath321^{2\over 3(1+w)}\ ] ] where the normalization factor @xmath223 is independent of @xmath61 .
this setting is very useful to test the theoretical apparatus developed in this section . in the case of @xmath322 ,
figure [ fig_cc_dust ] shows a variety of quantities we have discussed .
figure [ fig_cc_dust ] ( a ) and ( b ) depict the case of an earlier and later time leaf with @xmath323 and @xmath324 respectively .
the solid red curves show @xmath313 ( computed numerically ) while the green curves give the static sphere approximation of equation [ ds_entropy_static ] . the dotted horizontal line marks half of the de sitter entropy : @xmath325 . as expected , @xmath326 .
the orange curve with a discontinuity in its derivative is @xmath327 .
comparing figures [ fig_cc_dust ] ( a ) and ( b ) , one can see that @xmath313 is approaching @xmath327 as @xmath270 .
finally , the black curves shows extensive entropy : @xmath328 .
note that @xmath329 for all @xmath299 as required by corollary [ thermal_bound_thm ] .
the holographic screen entanglement entropy structure of a closed universe with a past and future singularity is similar to that of approximate de sitter space .
the spacetimes we consider have the metric of equation [ frw_metric ] with @xmath330 . in this case
the coordinate @xmath247 takes values from @xmath261 to @xmath331 .
we put one species of matter content in the spacetime that satisfies @xmath332 which gives rise to a big bang at @xmath256 as well as a big crunch .
as before , we introduce a conformal time coordinate @xmath238 in terms of which the scale factor is @xmath333 where @xmath334 and @xmath277 is constant .
this shows that the penrose diagram for this spacetime is a rectangle with a time - to - space aspect ratio of @xmath64 .
figure [ fig_big_crunch ] shows the holographic screen structure of this spacetime for two examples of null foliations .
we focus on the diagram to the left in which case the null foliation ( partially ) consists of past light cones of a comoving worldline at the @xmath237 .
as suggested by the figure , the holographic screen is given by @xmath335 however , a subtlety arises because the screen is a past holographic screen for @xmath336 and a future screen for @xmath337 . the sphere that connects
the past and future screen is extremal ( this was called an `` optimal '' surface in @xcite ) and has area @xmath338 .
let @xmath339 denote the leaf at conformal time @xmath238 .
we put @xmath340 . just as in the de sitter case
, this example leads to a saturation of the page bound of equation [ thermal_bound ] as leaves are maximized in area .
more precisely , if @xmath341 , then @xmath342 where in this case @xmath343 it appears that @xmath46 saturates the page bound in a great variety of cases where the areas of leaves are bounded above .
the proposal we have given above may open the door to a new research program : the study of the entanglement structure of general spacetimes . in light of this , and for the sake of clarity , we now summarize the recipe for computing von neumann entropy under the assumption of the screen entanglement conjecture discussed in section [ section2 ] : 1 . select a particular null foliation @xmath344 of a spacetime with dimension @xmath3 .
2 . find the codimension 2 surfaces @xmath345 with @xmath346 that have maximal area on each @xmath347 .
3 . take a @xmath44 dimensional subregion @xmath348 with a boundary @xmath1 .
4 . of all extremal surfaces anchored to @xmath1 and lying in the causal region @xmath19 ( see section [ section2 ] ) , select the one of minimal area . the conjectured entropy @xmath46 is then one fourth the area of the minimal extremal surface in planck units
. potential applications of our conjecture are numerous .
one example not considered above is case of a spacetime with a black hole .
black holes formed through the collapse of matter possess future holographic screens in their interiors that approach their horizons at late times .
it is of potential significance to investigate the entanglement structure of such spacetimes .
perhaps such an analysis will shed light on the firewall paradox @xcite .
if the screen entanglement conjecture is correct , it should still only be regarded as a leading order prescription for the computation of von neumann entropies .
a version of the analysis of @xcite may be extendible to the context of holographic screens .
it is not completely obvious how this should be done .
if @xmath0 is a region in a leaf @xmath14 lying on a cauchy slice @xmath35 , one may consider the region on @xmath35 bounded by @xmath0 and its extremal surface @xmath47 and compute the entanglement entropy of this region in a quantum field theory on the spacetime background . on the other hand
, it may be necessary to modify the spacetime position of the holographic screen itself as was done in @xcite .
we are very grateful to c. akers , r. bousso , z. fisher , d. harlow , j. koeller , m. moosa , m. van raamsdonk , and a. wall for for impactful discussions .
we thank r. bousso and a. wall for their comments on earlier versions of this work .
the work of fs is supported in part by the doe nnsa stewardship science graduate fellowship .
the work of sjw is supported in part by the bctp brantley - tuttle fellowship for which he would like to extend his gratitude to lynn brantley and douglas tuttle .
j. m. maldacena , int .
j. theor .
phys . * 38 * , 1113 ( 1999 ) [ adv .
* 2 * , 231 ( 1998 ) ] doi:10.1023/a:1026654312961 [ hep - th/9711200 ] .
e. witten , adv .
* 2 * , 253 ( 1998 ) [ hep - th/9802150 ] .
s. ryu and t. takayanagi , phys .
lett . * 96 * , 181602 ( 2006 ) doi:10.1103/physrevlett.96.181602 [ hep - th/0603001 ] .
a. lewkowycz and j. maldacena , jhep * 1308 * , 090 ( 2013 ) doi:10.1007/jhep08(2013)090 [ arxiv:1304.4926 [ hep - th ] ] .
v. e. hubeny , m. rangamani and t. takayanagi , jhep * 0707 * , 062 ( 2007 ) doi:10.1088/1126 - 6708/2007/07/062 [ arxiv:0705.0016 [ hep - th ] ] .
r. bousso , jhep * 9907 * , 004 ( 1999 ) doi:10.1088/1126 - 6708/1999/07/004 [ hep - th/9905177 ] .
r. bousso , jhep * 9906 * , 028 ( 1999 ) doi:10.1088/1126 - 6708/1999/06/028 [ hep - th/9906022 ] .
g. t hooft , salamfest 1993:0284 - 296 [ gr - qc/9310026 ] . l. susskind , j. math . phys . * 36 * , 6377 ( 1995 ) doi:10.1063/1.531249 [ hep - th/9409089 ] .
r. bousso and n. engelhardt , phys .
* 115 * , no . 8 ,
081301 ( 2015 ) doi:10.1103/physrevlett.115.081301 [ arxiv:1504.07627 [ hep - th ] ] .
r. bousso and n. engelhardt , phys .
d * 92 * , no . 4 , 044031 ( 2015 ) doi:10.1103/physrevd.92.044031 [ arxiv:1504.07660 [ gr - qc ] ] .
s. a. hayward , phys .
d * 49 * , 6467 ( 1994 ) .
doi:10.1103/physrevd.49.6467 s. a. hayward , class .
grav . *
15 * , 3147 ( 1998 ) doi:10.1088/0264 - 9381/15/10/017 [ gr - qc/9710089 ] .
a. ashtekar and b. krishnan , phys
. rev .
d * 68 * , 104030 ( 2003 ) doi:10.1103/physrevd.68.104030 [ gr - qc/0308033 ] .
a. ashtekar and g. j. galloway , adv .
* 9 * , no . 1 , 1 ( 2005 ) doi:10.4310/atmp.2005.v9.n1.a1 [ gr - qc/0503109 ] . m. headrick and t. takayanagi , phys . rev .
d * 76 * , 106013 ( 2007 ) doi:10.1103/physrevd.76.106013 [ arxiv:0704.3719 [ hep - th ] ] . a. c. wall , class .
grav . * 31 * , no .
22 , 225007 ( 2014 ) doi:10.1088/0264 - 9381/31/22/225007 [ arxiv:1211.3494 [ hep - th ] ] .
n. engelhardt and a. c. wall , jhep * 1403 * , 068 ( 2014 ) doi:10.1007/jhep03(2014)068 [ arxiv:1312.3699 [ hep - th ] ] .
t. hartman and j. maldacena , jhep * 1305 * , 014 ( 2013 ) doi:10.1007/jhep05(2013)014 [ arxiv:1303.1080 [ hep - th ] ] .
d. n. page , phys .
* 71 * , 1291 ( 1993 ) doi:10.1103/physrevlett.71.1291 [ gr - qc/9305007 ] .
y. nomura , jhep * 1111 * , 063 ( 2011 ) doi:10.1007/jhep11(2011)063 [ arxiv:1104.2324 [ hep - th ] ] . y. nomura , found .
phys . * 43 * , 978 ( 2013 ) doi:10.1007/s10701 - 013 - 9729 - 1 [ arxiv:1110.4630 [ hep - th ] ] .
y. nomura , j. varela and s. j. weinberg , phys .
b * 733 * , 126 ( 2014 ) doi:10.1016/j.physletb.2014.04.027 [ arxiv:1304.0448 [ hep - th ] ] . y. nomura , f. sanches and s. j. weinberg , jhep * 1504 * , 158 ( 2015 ) doi:10.1007/jhep04(2015)158 [ arxiv:1412.7538 [ hep - th ] ] . y. nomura , f. sanches and s. j. weinberg , phys . rev .
* 114 * , 201301 ( 2015 ) doi:10.1103/physrevlett.114.201301 [ arxiv:1412.7539 [ hep - th ] ] .
a. almheiri , d. marolf , j. polchinski and j. sully , jhep * 1302 * , 062 ( 2013 ) doi:10.1007/jhep02(2013)062 [ arxiv:1207.3123 [ hep - th ] ] .
e. t. newman , gen .
* 7 * , 107 ( 1976 ) .
doi:10.1007/bf00762018 n. engelhardt and g. t. horowitz , arxiv:1605.01070 [ hep - th ] .
t. faulkner , a. lewkowycz and j. maldacena , jhep * 1311 * , 074 ( 2013 ) doi:10.1007/jhep11(2013)074 [ arxiv:1307.2892 [ hep - th ] ] .
r. bousso and n. engelhardt , phys .
d * 93 * , no . 2 , 024025 ( 2016 ) doi:10.1103/physrevd.93.024025 [ arxiv:1510.02099 [ hep - th ] ] . | we present a natural generalization of holographic entanglement entropy proposals beyond the scope of ads / cft by anchoring extremal surfaces to holographic screens .
holographic screens are a natural extension of the ads boundary to arbitrary spacetimes and are preferred codimension 1 surfaces from the viewpoint of the covariant entropy bound .
a broad class of screens have a unique preferred foliation into codimension 2 surfaces called leaves .
our proposal is to find the areas of extremal surfaces achored to the boundaries of regions in leaves .
we show that the properties of holographic screens are sufficient to prove , under generic conditions , that extremal surfaces anchored in this way always lie within a causal region associated with a given leaf . within this causal region ,
a maximin construction similar to that of wall proves that our proposed quantity satisfies standard properties of entanglement entropy like strong subadditivity .
we conjecture that our prescription computes entanglement entropies in quantum states that holographically define arbitrary spacetimes , including those in a cosmological setting with no obvious boundary on which to anchor extremal surfaces . #
1#1 | # 1| # 1 # 1#2#1 | # 2 |
an increasing number of extragalactic objects is detected at energies greater than 100 gev .
the list of published sources comprises 20 objectsrwagner / sources ] : as expected ( e.g. , costamante & ghisellini 2002 ) , the bulk of them ( 17 ) belongs to the class of highly peaked bl lac objects . in fact , the position of the synchrotron peak , usually located in ( or close to ) the x ray band assures the existence , in these sources , of electrons with extremely large lorentz factors ( @xmath0@xmath1 ) , a key ingredient for the emission in the tev band via inverse compton ( ic ) scattering .
the remaining three sources are bl lac itself , which belongs to the lbl class , ( albert et al .
2007 ) , 3c279 , a fsrqs , ( teshima et al . 2007 ) and the nearby ( 16 mpc , tonry 1991 ) radiogalaxy m87 ( aharonian et al .
2003 , 2006 ) . already suggested as a possible source of high - energy radiation ( e.g. bai & lee 2001 ) based on its similarity with bl lac objects ( tsvetanov et al .
1998 ) , m87 was discovered as a tev source by the hegra array of cherenkov telescopes ( aharonian et al .
2003 ) and it has been recently confirmed by h.e.s.s .
( aharonian et al . 2006 ) .
due to the limited spatial resolution it is not possible to identify the emission region of this radiation .
however , the sensitivity of h.e.s.s . allowed the detection of variability on short timescales ( @xmath2 days ) , suggesting a compact emission region .
three main sites have been proposed as emission regions for the tev radiation : the resolved jet ( in particular the so
called knot hst1 ) ; the unresolved base of the jet ( in analogy with blazars ) ; the vicinity of the supermassive ( @xmath3 @xmath4 , marconi et al .
1997 ) black hole ( bh ) .
hst1 is a quite interesting jet feature located at 60 pc ( projected ) from the core of of m87 .
superluminal motions with apparent speeds up to @xmath5 were observed with _
hst _ ( biretta et al . 1999 ) .
continuous multifrequency coverage has shown spectacular activity from this knot in the past years ( perlman et al .
2003 ; harris et al . 2006 , 2007 ; cheung et al .
in particular , the x ray flux increases by a factor of @xmath6 from 2000 to 2005 and since then it is steadily decreasing .
these variations are accompanied by similar changes in the optical and in the radio bands and by the appearance of new superluminal components .
this extreme phenomenology is described by stawarz et al .
( 2006 ) on the assumption that hst-1 marks the recollimation shock of the jet . as such
, hst1 is thought to be a rather efficient particle accelerator and thus a possible source of intense tev radiation .
however , the small variability timescales observed by h.e.s.s . put strong limits to the size of the source ( @xmath7 cm , where @xmath8^{-1}$ ] is the relativistic doppler factor , @xmath9 is the viewing angle and @xmath10 days ) , which seems difficult to accomplish in this scenario ( e.g. neronov & aharonian 2007 ; levinson 2007 ) .
the rapid variability could be easily reproduced if the tev emission is produced close to the base of the jet , in the region associated to the emission from blazars
. however , standard one zone leptonic models face difficulties in reproducing the observed spectral energy distribution ( sed ) of the core including the tev data . a possible solution advocates that the emission comes from a decelerating jet ( georganopoulos et al .
alternatively , the emission can be produced by high energy protons ( reimer et al .
going further in towards the bh , tev photons could be emitted through ic by relativistic pairs produced by electromagnetic cascades in the bh magnetosphere ( neronov & aharonian 2007 ) .
the determination of the site producing the high - energy emission is rather important .
indeed , if the emission site will be eventually identified as knot hst-1 , this would have a broad impact on the current view of relativistic jets ( see the discussion in cheung et al .
it is thus extremely important to investigate whether the standard view ( or minimal variations of it ) for the inner jet is able reproduce the observed phenomenology . if not
, it would be mandatory to explore the other alternatives .
as discussed in ghisellini , tavecchio & chiaberge ( 2005 , hereafter paper i ) , there is compelling evidence supporting the view that jets of bl lac objects are structured at blazar scales ( at distances @xmath11 cm from the bh ) , with a fast core ( _ spine _ ) surrounded by a slower sheet ( _ layer _ ) . in paper i
we showed that the radiative coupling between the spine and the layer can naturally account for the deceleration inferred for the jet between the blazar and the vlbi scale .
furthermore , one component sees the radiation of the other boosted , since the relative velocity can be relativistic .
this enhances the ic emission of both the spine and the layer .
a direct consequence of such a structure is that while at small viewing angle ( as in the case of blazars ) the emission is dominated by the boosted spine emission , at large observing angles ( @xmath12 , typical for radio galaxies ) the emission from the spine would be suppressed , while the layer , characterised by a broader beaming cone , could substantially contribute to ( and sometimes dominate ) the overall emission . at intermediate angles both components
can significantly contribute .
following this line , in this letter we explore an alternative interpretation for the tev emission of m87 , assuming that the tev radiation is produced in the layer of the misaligned jet .
we use @xmath13 , @xmath14 , @xmath15 .
the sed of the core of m87 is reported in fig .
[ fig1 ] ( grey open points , see paper i for references ) together with the h.e.s.s .
spectra for 2004 ( open blue squares ) and 2005 ( open red triangles ) as reported by aharonian et al . ( 2006 ) . the x ray spectrum ( as observed in july 2000 )
is taken from balmavarde , capetti & grandi ( 2006 ) .
note that the slope [ @xmath16 ; @xmath17 is intermediate between that reported in marshall et al .
( 2002 ) ( @xmath18 , see also donato , sambruna & gliozzi 2004 ) and that of wilson & yang ( 2002 ) ( @xmath19 ) .
the x ray spectral shape was not determined in 2004 and 2005 , and for simplicity we assume that it is the same of 2000 , with a normalisation that is a factor 3 ( 2004 , blue ) and 4 ( 2005 , red ) greater .
this reproduces the increase of the x ray flux of the core recorded through the continuous monitoring of _ chandra _ ( harris et al .
2007 ) .
the data describe a bump peaking at @xmath20 hz and extending into the x ray band , while the tev emission clearly belongs to a second component .
the sed closely resembles that usually observed in blazars , in which the first component is synchrotron emission from relativistic electrons and the second peak is produced through ic emission ( e.g. , ghisellini et al .
1998 ) . for m87
there is no evidence of powerful sources of soft photons in the nucleus and hence the high energy emission should be dominated by the synchrotron self
compton emission ( ssc ; e.g. , maraschi et al .
1992 ) . however , a simple one zone synchrotron ssc model can not easily describe the observed sed .
the reason is that in order to have the synchrotron component peaking in the ir region ( @xmath21 hz ) and the ssc one close to the tev band ( @xmath22 hz , as required by the relatively flat tev spectra ) one is forced to assume a rather low magnetic field and an unreasonably large doppler factor .
following tavecchio , maraschi & ghisellini ( 1998 ) and defining the synchrotron and ssc peak luminosities , @xmath23 and @xmath24 and the variability timescale @xmath25 , one obtains : @xmath26 here @xmath27 , in cgs units , except for @xmath25 , that is in days .
the adopted numerical values are appropriate for m87 .
this is a general problem faced by any synchrotron ssc model using a single emission region . a direct way to overcome
this problem is to `` decouple '' synchrotron and the ic components , assuming that the two components are produced by two different regions .
such a `` two - zones '' model would allow us to reproduce the low frequency bump ( emitted by electrons with relatively low energy ) with one emission component and the tev emission with a separate source with high energy electrons .
this opportunity is directly offered by the spine layer scenario . in this case
the debeamed spine ( or , alternatively , the layer ) could be responsible for the low energy bump and the layer ( or the spine ) of the tev component .
we focus on the case of a sub pc scale emission region , due to the observed inter
day variability of the tev emission , difficult to explain if hst1 is responsible for this emission .
we are aware that we are almost doubling the number of the free parameters ( one set for the spine , one set for the layer ) , and therefore we pay the increased flexibility with an increased uncertainty concerning the value of the parameters themselves and a loss of predictive power . on the other hand ,
our main aim at this stage is not much to find whether / what specific physical conditions ( value of the magnetic field , beaming factor , number of particles , and so on ) are univocally determined by the model , but to show that one can attribute the radiation we see to the the sub pc jet ( with `` minor '' modification of the simplest model ) without radically changing our ideas about the emission models of blazars .
having said that , we will nevertheless do our best to find the most reasonable set of parameters .
this is why we will not only try to fit the sed of m87 , but we will worry about what kind of blazar would m87 be if seen at small angles . if the resulting sed would be too odd with respect to the known blazar sed , we will discard the found parameter set .
-0.7 true cm -0.8 true cm we refer to paper i for a complete description of the model .
briefly , we assume that the jet comprises two regions : i ) the spine , assumed to be a cylinder of radius @xmath28 , height @xmath29 ( as measured in the spine frame ) and in motion with bulk lorentz factor @xmath30 ; ii ) the layer , modelled as an hollow cylinder with internal radius @xmath28 , external radius @xmath31 , height @xmath32 ( as measured in the frame of the layer ) and bulk lorentz factor @xmath33
. each region is characterised by tangled magnetic field with intensity @xmath34 , @xmath35 and it is filled by relativistic electrons assumed to follow a smoothed broken power
law distribution extending from @xmath36 to @xmath37 and with indices @xmath38 , @xmath39 below and above the break at @xmath40 .
the normalisation of this distribution is calculated assuming that the system produces an assumed synchrotron luminosity @xmath41 ( as measured in the local frame ) , which is an input parameter of the model . as in paper
i we assume that @xmath42 .
the seed photons for the ic scattering are not only those produced locally in the spine ( layer ) , but we also consider the photons produced in the layer ( spine ) .
hereafter we will indicate this component as external compton ( ec ) . to properly calculate the observed emission ,
it is necessary to take into account the different beaming patterns associated to the two emission components .
this is because , in the comoving frame of the spine ( layer ) , the photons produced by the layer ( spine ) are not isotropic , but are seen aberrated .
most of them are coming from almost a single direction ( opposite to the relative velocity vector ) . in this case
the produced ec is anisotropic even in the comoving frame , with more power emitted along the opposite direction of the incoming photons ( i.e. for head on scatterings ) .
we therefore must take into account this anisotropy in the comoving frame when transforming in the observer frame .
consider also that this applies only to the ec radiation , while the synchrotron and the ssc emissions are isotropic in the comoving frame .
the ec radiation pattern of the spine will be more concentrated along the jet axis ( in the forward direction ) with respect to its synchrotron and ssc emission .
on the contrary , the ec radiation pattern of the layer will be _ less _ concentrated along the forward direction ( with respect to its synchrotron and ssc ) because , in the comoving frame , the layer emits more ec power in the backward direction ( e.g. back to the black hole ) .
the corresponding transformations are fully described in paper i , and we briefly summarise them in the following .
we here follow dermer ( 1995 ) , assuming that , due to the strong aberration , in the rest frame of the spine ( layer ) all the photons , assumed to be isotropic in the layer ( spine ) frame , can be considered as coming from a single direction , i.e. along ( opposite to ) the jet axis .
this approximation greatly simplifies the calculations , allowing us not to consider the detail of the geometry in the transformations .
the beaming of the synchrotron and ssc emission for both regions follows the usual @xmath43 ( spine ) and @xmath44 ( layer ) pattern , while the patterns of the ec components are described by more complex shapes .
in particular , the ec from the spine will have a rather narrow pattern given by @xmath45 , where @xmath46 is the doppler factor of the spine as measured in the layer frame . the first term ( @xmath47 )
describes the same pattern as the ec radiation produced when scattering the seed photons produced by the broad line regions of powerful blazars , as firstly pointed out by dermer ( 1995 ) .
it can be understood by going in the reference frame where the seed photons are isotropic ( in this case this means comoving with the layer ) . in this frame
we see the ec from the spine boosted by the dermer " pattern @xmath47 . finally , we go to the observer frame using the standard @xmath44 term .
analogously , the amplification of the ec from the layer is @xmath48 .
the net result is that the ec emission from the layer is less boosted in the observer frame than the corresponding ec emission from the spine .
[ fig2 ] illustrates this point , showing the amplification factors of the different components as a function of the viewing angle ( from 0@xmath49 to 45@xmath49 ) , calculated assuming @xmath50 and @xmath51 ( the values we used for m87 , see below ) .
solid lines refer to the spine , dashed lines to the layer .
black and red lines correspond to the synchrotron / ssc and ec amplification factors , respectively . for the spine ,
the ec amplification is larger than the synchrotron / ssc one for small angles , while at large angles the latter dominates . in this sense
one can say that the ec emission has a narrower pattern than the synchrotron / ssc emission .
the opposite occurs for the layer : at small angles the synchrotron / ssc amplification is larger then the ec one , which instead becomes more important at large angles .
note , moreover , that the ec / layer emission has its maximum at @xmath5210@xmath49 , while all the other curves peak at 0@xmath49 . to understand this , consider the emission pattern in the frame comoving with the layer . in this frame
the seed photons produced by the spine are coming along the black hole
layer direction : as a consequence the relativistic electrons will produce more ec radiation in the opposite ( layer black hole ) direction .
this is the radiation that will be boosted the least when we transform the intensity in the observer frame ( and the opposite for radiation along the outward radial direction : in the comoving frame it is the weakest , but it is boosted the most when going to the observer frame ) .
this compensation " between the intrinsic anisotropy as seen in the comoving frame and the beaming effect means that the maximum flux will be observed not at 0@xmath49 , but at a larger angle . at the extreme ,
if the layer is not moving , the maximum will be observed at 180@xmath49 .
this characteristic is particularly relevant for sources seen at large angles : for those , the ec produced by the layer becomes relatively more important .
-0.9 true cm -1.1 true cm we have two possibilities to reproduce the sed with this model : i ) either we assume that the low energy bump is produced by the layer and the tev emission originates in the spine or ii ) the layer produces the tev radiation and the spine mainly contributes at low frequencies .
the first choice is disfavoured in the present case .
in fact , to produce tev photons , the typical lorentz factor of the electrons in the spine must be large ( @xmath53 ) and , assuming typical magnetic fields of order 0.11 g , the corresponding synchrotron emission would peak in the hard x ray band . under this conditions
the ssc emission would be deeply in the klein
nishina regime and thus strongly depressed and the tev emission would be dominated by the ec component .
if the spine sed is dominated by the ec component at large ( 18@xmath49 ) angles , it will be even more so at small angles ( see fig .
2 ) , with @xmath54 . instead , in tev bl lacs , the synchrotron and the ic components show comparable luminosities [ with few exception , most notably 1es 1426 + 428 ( aharonian et al . 2002 ) , although the tev luminosity depends on the still unclear amount of intergalactic absorption ] . on the other hand ,
as we show below , the choice ii ) leads to a predicted sed for the beamed counterpart of m87 closely resembling those observed in lbl sources and thus we decided to adopt this view in the modelling .
[ cols= " < , < , < , < , < , < , < , < , < , < , < , < , < " , ] [ tab1 ] the seds obtained for the core of m87 with the model are shown in fig . [ fig1 ] for both 2004 and 2005 and the parameters
are reported in table 1 .
the model shown is intended as an example of the possible parameters giving a reasonable shape for the sed .
a whole family of acceptable solutions , each one with different values for the parameters , is allowed . however , the joint request to reproduce the data and to have a spine sed resembling that of a bl lac at small angles provides some constraints . since the layer has to produce tev radiation , the lorentz factor of the electrons at the peak
is constrained to be of the order of @xmath55 , independently on the component ( ssc or ec ) dominating the high energy emission .
this , in turn , fixes the peak of the synchrotron component close to or within the x ray band for the chosen magnetic field .
smaller @xmath56values are allowed , but not larger , since in this case the layer synchrotron emission would overproduce the observed data . for the spine , as already discussed ,
most of the constraints come from the request that the lorentz factor , the size and the magnetic field are similar to those typically inferred for ( lbl ) blazars .
the model slightly overproduces the emission in the optical band .
however , since the optical and the x ray emission are clearly correlated ( perlman et al . 2003 ) we expect that high x - ray fluxes correspond to high optical states
. note also that we predict quite bright gev emission in this period , exceeding the _ egret _ upper limit by a factor of 3 and easily detectable with _
glast_. the high - energy component of the layer is dominated by the ec emission from the beamed ( in the layer frame ) ir - optical photons from the spine . on the other hand , since the synchrotron component of the layer peaks in the x
rays , the ec emission of the spine is strongly suppressed by the effect of the klein
nishina cross section and hence the high energy component is mainly ssc .
the variability of the tev flux from 2004 to 2005 is reproduced by increasing the luminosity radiated by the layer ( and thus the corresponding density of the electrons ) and slightly decreasing the magnetic field .
note also that the small size of the emitting region allows the rapid changes such as those detected by h.e.s.s .
( @xmath10 days ) . a supplementary contribution to the ic emission from both the spine and
the layer could come from the scattering of the external ir radiation field , possibly associated to the putative low
efficiency accretion flow .
the luminosity of the core of m87 in the infrared band is @xmath57 erg s@xmath58 ( perlman et al .
this is an upper limit to the luminosity of the accretion flow , since the observed power law spectrum supports the idea that it is mainly associated to the emission from the jet ( perlman et al .
we can set an upper limit to the importance of this component by assuming @xmath59 erg s@xmath58 for the accretion luminosity , produced in a region whose size is of the order the distance of the active region of the jet ( around @xmath60 cm ) .
this maximizes the importance of this external component : for smaller radii the accretion ir radiation would be strongly deboosted in the jet frame , while for larger radii it will be more diluted .
even so , this external component has a radiation energy density smaller ( by more than an order of magnitude ) of the one already present in the spine and in the layer . inspection of fig .
[ fig1 ] reveals that we nicely reproduce the slope of the 2004 tev spectrum , while the 2005 spectrum appears to be harder than the model ( although the relatively large error bars prevent a firm conclusion ) .
the steep slope of the model continuum in this band is the result of the @xmath61 absorption of tev photons interacting with optical
ir photons of the spine and thus can not be simply avoided . in fig .
[ fig1 ] we also report ( black line ) the emission from the spine relative to 2004 as would appear to an observer almost aligned ( @xmath62 ) with the jet . for comparison we report ( orange , filled symbols ) the multifrequency sed for bl lac itself ( the prototype of bl lac objects ) , collected during different campaigns ( see ravasio et al .
2002 for references ) and reporting also the tev spectrum recently detected by magic ( albert et al .
we propose that the tev emission detected from m87 originates in the structured jet at blazar scale .
the most likely possibility is that the emission from the spine reproduces the low energy component , from radio to x
rays , while the layer contributes to the bulk of the tev radiation . with this choice ,
the beamed counterpart of m87 observed at small angle would have a sed closely resembling those of low peaked bl lac objects .
correspondingly the spine is characterised by physical parameters close to those usually inferred for those sources ( @xmath63 , @xmath64 g ) .
the layer , instead , would be characterised by a low magnetic field ( @xmath65 g ) and large peak lorentz factors ( @xmath66 ) .
it is possible that particles in the layer are energized by turbulent acceleration ( e.g. stawarz & ostrowski 2002 ) . in this case
the expected distribution would have a pile - up at the energy where acceleration and loss processes are in equilibrium ( e.g. katarzynski et al .
suggestively , in our conditions ( losses dominated by ic ) the expected equilibrium lorentz factor would be close to @xmath67 for @xmath65 g. note that our conclusions are rather different from that of georganopoulos et al .
( 2005 ) who discussed a similar model relying on the possibility that the jet experiences a strong deceleration at the blazar scale . in their model
the high - energy emission is produced in the innermost and faster portion of the jet , while the slow external portions mainly contributes at low frequency . under these conditions
the beamed counterpart of m87 would be a tev bl lac .
however , due to the different beaming patterns or the synchrotron and ic components ( analogously to our case ) , the resulting spectrum would be characterised by a strong dominance of the tev component with respect to the x ray one , contrary to what is usually observed in the known tev bl lacs . in our model the optical and the x
rays are produced by a region different from that responsible for the tev emission .
therefore a strict correlation between low energy and tev emission is not directly required ( even if it is possible ) .
analogously , the mev - gev emission should originate in the spine .
therefore we expect that the emission possibly detected by _ glast _ will not exactly follow the tev component . a potential weak point of our model concerns the slope of the tev spectrum .
the slope is mainly dictated by the absorption of tev photons by the dense optical radiation field and , in this sense , it is rather robust and does not strongly depend on the underlying slope of the intrinsic tev spectrum .
in particular , hard spectra such as that recorded in 2005 are quite difficult to achieve in this scheme .
a possibility to avoid important absorption of gamma - rays would be to enlarge the emission region , hence diluting the target photon density and decreasing the absorption optical depth .
however the increase of the source size is limited by the observed short variability timescales .
further observations with increased sensitivity could provide stringent constraints to our interpretation .
we thank m. chiaberge and b. balmaverde for useful discussions on the x - ray data of m87 and the referee , mark birkinshaw , for constructive comments that help to clarify the text . | simple one zone homogeneous synchrotron self
compton models have severe difficulties is explaining the tev emission observed in the radiogalaxy m87 .
also the site the tev emission region is uncertain : it could be the unresolved jet close to the nucleus , analogously to what proposed for blazars , or an active knot , called hst1 , tens of parsec away .
we explore the possibility that the tev emission of m87 is produced in the misaligned subpc scale jet .
we base our modelling on a structured jet , with a fast spine surrounded by a slower layer . in this context
the main site responsible for the emission of the tev radiation is the layer , while the ( debeamed ) spine accounts for the emission from the radio to the gev band : therefore we expect a more complex correlation with the tev component than that expected in one zone scenarios , in which both components are produced by the same region .
observed from small angles , the spine would dominate the emission , with an overall spectral energy distribution close to those of bl lac objects with a synchrotron peak located at low energy ( lbls ) .
[ firstpage ] galaxies : active galaxies : jets galaxies : individual : m87radiation mechanisms : non thermal . |
as is well - known , it is not possible to account for all three neutrino anomalies with just the three known neutrinos @xmath0 and @xmath1 .
if one or more of them can be explained in some other way , then no extra sterile neutrinos need be invoked .
this is one motivation for exotic scenarios . in any case
it is important to rule out all explanations other than oscillations in order to establish neutrino mixing and oscillations as the unique explanation for the three observed neutrino anomalies .
i should mention that , in general , some ( but not all ) non - oscillatory explanations of the neutrino anomalies will involve non - zero neutrino masses and mixings .
therefore , they tend to be neither elegant nor economical . but the main issue here is whether we can establish oscillations unequivocally and uniquely as the cause for the observed anomalies .
i first summarize some of the exotic scenarios and then consider each anomaly in turn .
there are three different ways that massless neutrinos may mix and even oscillate .
these are as follows .
\1 . if flavor states are mixtures of massless as well massive states , then when they are produced in reactions with q - values smaller than the massive state ; the flavor states are massless but not orthogonal@xcite .
for example , if @xmath3 , @xmath4 and @xmath5 for @xmath6 to 3 but @xmath7 = 50 gev ; then `` @xmath8 '' produced in @xmath9-decay and `` @xmath10 '' produced in @xmath11-decay are massless but not orthogonal and @xmath12 on the other hand @xmath8 and @xmath10 produced in @xmath13 decay will be not massless and will be more nearly orthogonal .
hence , the definition of flavor eigenstate is energy and reaction dependent and not fundamental .
current limits@xcite on orthogonality of @xmath14 and @xmath15 make it impossible for this to play any role in the current neutrino anomalies .
when flavor changing neutral currents as well as non - universal neutral current couplings of neutrinos exist , for example : @xmath16 such couplings arise in r - parity violating supersymmetric theories@xcite . in general
, in these theories , one expects neutrino masses at some level here it is assumed that the fcnc provides the dominant effect . in this case
the propagation of @xmath8 and @xmath15 in matter is described by the equation : @xmath17 where @xmath18 and @xmath19 .
there is a resonance at @xmath20 or @xmath21 and @xmath8 can convert to @xmath15 completely .
this matter effect which effectively mixes flavors in absence of masses was first pointed out by wolfenstein in the same paper@xcite where matter effects were first discussed .
the resonant conversion is just like in the conventional msw effect , except that there is no energy dependence and @xmath22 and @xmath23 are affected the same way .
this possibility has been discussed in connection with both solar and atmospheric anomalies .
( a ) flavor violating gravity wherein it is proposed that gravitational couplings of neutrinos are flavor non - diagonal@xcite and equivalence principle is violated . for example , @xmath24 and @xmath25 may couple to gravity with different strengths : @xmath26 where @xmath27 is the gravitational potential
. then if @xmath8 and @xmath10 are mixtures of @xmath24 and @xmath25 with a mixing angle @xmath28 , oscillations will occur with a flavor survival probability @xmath29 when @xmath27 is constant over the distance l and @xmath30 is the small deviation from universality of gravitational coupling . equivalence principle is also violated .
\(b ) another possibility is violation of lorentz invariance@xcite wherein all particles have their own maximum attainable velocities ( mav ) which are all different and in general also different from speed of light : then if @xmath24 and @xmath25 are mav eigenstates with mav s @xmath31 and @xmath32 and @xmath8 and @xmath10 are mixtures of @xmath24 and @xmath25 with mixing angle @xmath28 , the survival probability of a given flavor is @xmath33 where @xmath34 .
as far as neutrino oscillations are concerned , these two cases are identical in their dependence on le instead of l / e as in the conventional oscillations@xcite .
neutrino decay@xcite implies a non - zero mass difference between two neutrino states and thus , in general , mixing as well .
i will consider here only non - radiative decays .
we assume a component of @xmath35 i.e. , @xmath25 , to be the only unstable state , with a rest - frame lifetime @xmath36 , and we assume two flavor mixing , for simplicity : @xmath37 with @xmath38 . from eq .
( 2 ) with an unstable @xmath25 , the @xmath39 survival probability is @xmath40 where @xmath41 and @xmath42 . since we are attempting to explain neutrino data without oscillations there are two appropriate limits of interest .
one is when the @xmath43 is so large that the cosine term averages to 0 .
then the survival probability becomes @xmath44 let this be called decay scenario a. the other possibility is when @xmath45 is so small that the cosine term is 1 , leading to a survival probability of @xmath46 corresponding to decay scenario b. decay models for both kinds of scenarios can be constructed ; although they require fine tuning and are not particularly elegant .
in the lsnd experiment , what is observed is the following@xcite . in the decay at rest ( dar ) which is @xmath47
which should give a pure @xmath8 signal , there is a flux of @xmath48 at a level of about @xmath49 of the @xmath50 .
( there is a similar signal for @xmath51 accompanying @xmath52 in the decay in flight of @xmath53 . now
this could be accounted for without oscillations@xcite provided that the conventional decay mode @xmath54 is accompanied by the rare mode @xmath55 at a level of branching fraction of @xmath49 . assuming x to be a single particle , what can x be ?
it is straight forward to rule out x as being ( i ) @xmath10 ( too large a rate for muonium - antimuonium transition rate ) , ( ii ) @xmath8 ( too large a rate for fcnc decays of @xmath56 such as @xmath57 and ( iii ) @xmath15 ( too large a rate for fcnc decays of @xmath58 such as @xmath59 .
the remaining possibilities for x are @xmath60 or @xmath61 .
no simple models exist which lead to such decays .
rather baroque models can be constructed which involve a large number of new particles@xcite . experimental tests to distinguish this rare decay possibility from the conventional oscillation explanation are easy to state .
in the rare decay case , the rate is constant and shows no dependence on l or e. one can also ask whether the @xmath62 events seen in lsnd could have been caused by new physics at the detector ; for example a small rate for the `` forbidden '' reaction @xmath63 .
this is ruled out , since a fractional rate of @xmath49 for this reaction , leads ( via crossing ) to a rate for the decay mode @xmath64 in excess of known bounds .
oscillations of massless neutrinos via flavor changing neutral currents ( fcnc ) and non universal neutral currents ( nunc ) in matter have been considered@xcite as explanation for the solar neutrino observations , most recently by babu , grossman and krastev@xcite .
using the most recent data from homestake , sage , gallex and super - kamiokande , they find good fits with @xmath65 and @xmath66 or @xmath67 0.1 to 0.01 with @xmath68
. since the matter effect in this case is energy independent , the explanation for different suppression for @xmath69 , @xmath70 and pp neutrinos as inferred is interesting .
the differing suppression arises from the fact the production region for each of these neutrinos differ in electron and nuclear densities .
this solution resembles the large angle msw solution with the difference that the day - night effect is energy - independent .
the massless neutrino oscillations described in eq.(6 ) also offer a possible solution to solar neutrino observations .
there are three solutions .
two are characterized by @xmath71 and @xmath72 ; these are analogs of the small angle msw and large angle msw solutions@xcite
. these are now ruled out@xcite ( for both @xmath73 and @xmath74 by the recent nutev data@xcite .
the third one at @xmath75 and @xmath76 ( the analog of the vacuum solution ) is the only one allowed@xcite .
this possibility can be tested in long baseline experiments .
the possibility of solar neutrinos decaying to explain the discrepancy is a very old suggestion@xcite .
the most recent analysis of the current solar neutrino data finds that no good fit can be found : @xmath77 and @xmath78 ( e=10 mev ) @xmath79 to 27 sec .
come closest@xcite .
the fits become acceptable only if the suppression of the solar neutrinos is energy independent as proposed by several authors@xcite ( which is possible if the homestake data are excluded from the fit ) .
the above conclusions are valid for both the decay scenarios a as well as b.
the massless fcnc scenario has been recently considered for the atmospheric neutrinos by gonzales - garcia et al.@xcite with the matter effect supplied by the earth .
good fits were found for the partially contained and multi - gev events with @xmath80 as well @xmath81 .
the fit is poorer when higher energy events corresponding to up - coming muons are included@xcite .
the expectations for future lbl experiments are quite distinctive : for minos , one expects @xmath82 and @xmath83 the story for massless neutrinos oscillating via violation of equivalence principle or lorentz invariance is very similar . assuming large @xmath84 mixing , and @xmath85 a good fit to the contained events can be obtained@xcite ; but as soon as multi - gev and thrugoing muon events are included , the fit is quite poor@xcite .
turning to neutrino decay scenario a , it was found that it is possible to choose @xmath28 and @xmath86 to provide a good fit to the super - kamiokande l / e distributions of @xmath10 events and @xmath87 event ratio@xcite .
the best - fit values of the two parameters are @xmath88 and @xmath89 where @xmath90 km is the diameter of the earth .
this best - fit @xmath86 value corresponds to a rest - frame @xmath25 lifetime of @xmath91 however , it was then shown that the fit to the higher energy events in super - k ( especially the upcoming muons ) is quite poor@xcite . in all these three cases , the reason that the inclusion of high energy upcoming muon events makes the fits poorer is very simple .
the upcoming muons come from much higher energy @xmath92 and although there is some suppression , it is less than what is observed for lower energy events at the same l ( zenith angle ) .
this is in accordance with expectations from conventional oscillations .
the energy dependence in the above three scenarios is different and fails to account for the data . in the fcnc case
there is no energy dependence and so the high energy @xmath92 should have been equally depleted , in the fv gravity ( or lorentz invariance violation ) at high energies the oscillations should average out to give uniform 50% suppression and in the decay a scenario due to time dilation the decay is suppressed and there is hardly any depletion of @xmath92 .
turning to decay scenario b , consider the following possibility@xcite .
the three weak coupling states @xmath93 ( where @xmath94 is a sterile neutrino ) may be related to the mass eigenstates @xmath95 by the approximate mixing matrix . @xmath96 and the decay is @xmath97 .
the electron neutrino , which we identify with @xmath24 , can not mix very much with the other three because of the more stringent bounds on its couplings@xcite , and thus our preferred solution for solar neutrinos would be small angle matter oscillations .
then the @xmath98 in eq .
( 1 ) is not related to the @xmath99 in the decay , and can be very small , say @xmath100 ( to ensure that oscillations play no role in the atmospheric neutrinos ) . in that case , the oscillating term is 1 and @xmath101 becomes @xmath102 this is identical to eq .
( 13 ) in ref .
@xcite .
in order to compare the predictions of this model with the standard @xmath103 oscillation model , we have calculated with monte carlo methods the event rates for contained , semi - contained and upward - going ( passing and stopping ) muons in the super - k detector , in the absence of ` new physics ' , and modifying the muon neutrino flux according to the decay or oscillation probabilities discussed above .
we have then compared our predictions with the superk data @xcite , calculating a @xmath104 to quantify the agreement ( or disagreement ) between data and calculations . in performing our fit ( see ref .
@xcite for details ) we do not take into account any systematic uncertainty , but we allow the absolute flux normalization to vary as a free parameter @xmath9 .
the ` no new physics model ' gives a very poor fit to the data with @xmath105 for 34 d.o.f .
( 35 bins and one free parameter , @xmath9 ) . for the standard @xmath103 oscillation scenario the best fit has @xmath106 ( 32 d.o.f . ) and the values of the relevant parameters are @xmath107 ev@xmath108 , @xmath109 and @xmath110 .
this result is in good agreement with the detailed fit performed by the superk collaboration @xcite giving us confidence that our simplified treatment of detector acceptances and systematic uncertainties is reasonable .
the decay model of equations ( 3 ) and ( 4 ) above gives an equally good fit with a minimum @xmath111 ( 32 d.o.f . ) for the choice of parameters @xmath112 and normalization @xmath113 . in fig .
1 we compare the best fits of the two models considered ( oscillations and decay ) with the superk data . in the figure we show ( as data points with statistical error bars ) the ratios between the superk data and the monte carlo predictions calculated in the absence of oscillations or other form of ` new physics ' beyond the standard model . in the six panels we show separately the data on @xmath114-like and @xmath115-like events in the sub - gev and multi - gev samples , and on stopping and passing upward - going muon events .
the solid ( dashed ) histograms correspond to the best fits for the decay model ( @xmath103 oscillations ) .
one can see that the best fits of the two models are of comparable quality .
the reason for the similarity of the results obtained in the two models can be understood by looking at fig . 2 , where we show the survival probability @xmath116 of muon neutrinos as a function of @xmath117 for the two models using the best fit parameters .
in the case of the neutrino decay model ( thick curve ) the probability @xmath116 monotonically decreases from unity to an asymptotic value @xmath118 . in the case of oscillations
the probability has a sinusoidal behaviour in @xmath117 .
the two functional forms seem very different ; however , taking into account the resolution in @xmath117 , the two forms are hardly distinguishable .
in fact , in the large @xmath117 region , the oscillations are averaged out and the survival probability there can be well approximated with 0.5 ( for maximal mixing ) . in the region of small @xmath117 both probabilities approach unity .
in the region @xmath117 around 400 km / gev , where the probability for the neutrino oscillation model has the first minimum , the two curves are most easily distinguishable , at least in principle . decay model there are two decay possibilities that can be considered : ( a ) @xmath25 decays to @xmath119 which is dominantly @xmath94 with @xmath25 and @xmath120 mixtures of @xmath10 and @xmath15 , as in eq .
( [ eq : mixing ] ) , and ( b ) @xmath25 decays into @xmath119 which is dominantly @xmath121 and @xmath25 and @xmath120 are mixtures of @xmath10 and @xmath94 . in both cases the decay interaction has to be of the form @xmath122 where @xmath123 is a majoron field that is dominantly iso - singlet ( this avoids any conflict with the invisible width of the @xmath56 ) .
viable models for both the above cases can be constructed @xcite .
however , case ( b ) needs additional iso - triplet light scalars which cause potential problems with big bang nucleosynthesis ( bbn ) , and there is some preliminary evidence from superk against @xmath10@xmath94 mixing @xcite .
hence we only consider case ( a ) , i.e. @xmath124 with @xmath125 , as implicit in eq .
( [ eq : mixing ] ) . with this interaction ,
the @xmath25 rest - lifetime is given by @xmath126 where @xmath127 and @xmath128 . from the value of @xmath129 km / gev found in the fit and for @xmath130 ,
we have @xmath131 combining this with the bound on @xmath132 from @xmath133 decays of @xmath134 @xcite we have @xmath135 even with a generous interpretation of the uncertainties in the fit , this @xmath45 implies a minimum mass difference in the range of about 25 ev .
then @xmath25 and @xmath120 are nearly degenerate with masses @xmath136(25 ev ) and @xmath137 is relatively light .
we assume that a similar coupling of @xmath120 to @xmath137 and j is somewhat weaker leading to a significantly longer lifetime for @xmath120 , and the instability of @xmath120 is irrelevant for the analysis of the atmospheric neutrino data . for the atmospheric neutrinos in superk , two kinds of tests
have been proposed to distinguish between @xmath10@xmath15 oscillations and @xmath10@xmath94 oscillations .
one is based on the fact that matter effects are present for @xmath10@xmath94 oscillations@xcite but are nearly absent for @xmath10@xmath15 oscillations@xcite leading to differences in the zenith angle distributions due to matter effects on upgoing neutrinos @xcite .
the other is the fact that the neutral current rate will be affected in @xmath10@xmath94 oscillations but not for @xmath10@xmath15 oscillations as can be measured in events with single @xmath138 s @xcite . in these tests
our decay scenario will behave as a hybrid in that there is no matter effect but there is some effect in neutral current rates .
long - baseline experiments the survival probability of @xmath10 as a function of @xmath139 is given in eq .
the conversion probability into @xmath15 is given by @xmath140 this result differs from @xmath141 and hence is different from @xmath10@xmath15 oscillations .
furthermore , @xmath142 is not 1 but is given by @xmath143 and determines the amount by which the predicted neutral - current rates are affected compared to the no oscillations ( or the @xmath10@xmath15 oscillations ) case . in fig .
3 we give the results for @xmath144 , @xmath145 and @xmath146 for the decay model and compare them to the @xmath10@xmath15 oscillations , for both the k2k@xcite and minos@xcite ( or the corresponding european project@xcite ) long - baseline experiments , with the oscillation and decay parameters as determined in the fits above .
the k2k experiment , already underway , has a low energy beam @xmath147 gev and a baseline @xmath148 km .
the minos experiment will have 3 different beams , with average energies @xmath149 6 and 12 gev and a baseline @xmath150 km .
the approximate @xmath117 ranges are thus 125250 km / gev for k2k and 50250 km / gev for minos .
the comparisons in figure 3 show that the energy dependence of @xmath10 survival probability and the neutral current rate can both distinguish between the decay and the oscillation models .
minos and the european project may also have @xmath58 detection capabilities that would allow additional tests . big bang nucleosynthesis the decay of @xmath25 is sufficiently fast that all the neutrinos ( @xmath151 ) and the majoron may be expected to equilibrate in the early universe before the primordial neutrinos decouple .
when they achieve thermal equilibrium each majorana neutrino contributes @xmath152 and the majoron contributes @xmath153 @xcite , giving and effective number of light neutrinos @xmath154 at the time of big bang nucleosynthesis . from the observed primordial abundances of @xmath155he and @xmath156li , upper limits on @xmath157
are inferred , but these depend on which data are used@xcite .
conservatively , the upper limit to @xmath157 could extend up to 5.3 ( or even to 6 if @xmath158li is depleted in halo stars@xcite ) .
cosmic neutrino fluxes
since we expect both @xmath25 and @xmath120 to decay , neutrino beams from distant sources ( such as supernovae , active galactic nuclei and gamma - ray bursters ) should contain only @xmath8 and @xmath159 but no @xmath10 , @xmath160 , @xmath15 and @xmath121 .
this is a very strong prediction of our decay scenario .
we can compare the very different expectations for neutrino flavor mixes from very distant sources such as agn s or grb s .
let us suppose that at the source the flux ratios are typical of a beam dump , a reasonable assumption : @xmath161 .
then , for the conventional oscillation scenario , when all the @xmath45 s satisfy @xmath162 , it turns out curiously enough that for a wide variety of choices of neutrino mixing matrices , the final flavor mix is the same , namely : @xmath163 . in the case of the decay b scenario , as mentioned here , we have @xmath164 the two are quite distinct .
techniques for determining these flavor mixes in future km3 neutrino telescopes have been proposed@xcite .
reactor and accelerator limits the @xmath8 is essentially decoupled from the decay state @xmath25 so the null observations from the chooz reactor are satisfied@xcite .
the mixings of @xmath10 and @xmath15 with @xmath94 and @xmath8 are very small , so there is no conflict with stringent accelerator limits on flavor oscillations with large @xmath43 @xcite . in summary , neutrino decay remains a viable alternative to neutrino oscillations as an explanation of the atmospheric neutrino anomaly .
the model consists of two nearly degenerate mass eigenstates @xmath25 , @xmath120 with mass separation @xmath165 ev ) from another nearly degenerate pair @xmath24 , @xmath137 .
the @xmath10 and @xmath15 flavors are approximately composed of @xmath25 and @xmath120 , with a mixing angle @xmath166 .
the state @xmath25 is unstable , decaying to @xmath167 and a majoron with a lifetime @xmath168 sec .
the electron neutrino @xmath8 and a sterile neutrino @xmath94 have negligible mixing with @xmath169 and are approximate mass eigenstates ( @xmath170 , with a small mixing angle @xmath171 and a @xmath172 to explain the solar neutrino anomaly .
the states @xmath120 and @xmath137 are also unstable , but with @xmath120 lifetime somewhat longer and @xmath137 lifetime much longer than the @xmath25 lifetime .
this decay scenario is difficult to distinguish from oscillations because of the smearing in both l and @xmath173 in atmospheric neutrino events .
however , long - baseline experiments , where @xmath174 is fixed , should be able to establish whether the dependence of @xmath117 is exponential or sinusoidal . in our scenario
only @xmath24 is stable .
thus , neutrinos of supernovae or of extra galactic origin would be almost entirely @xmath8 .
the contribution of the electron neutrinos and the majorons to the cosmological energy density @xmath175 is negligible and not relevant for large scale structure formation .
another proposal for explaining the atmospheric neutrinos is based on decoherence of the @xmath92 in the flux@xcite .
the idea is that @xmath92 are interacting and getting tagged before their arrival at the detector .
the cause is unknown but could be a number of speculative possibilities such as a large neutrino background , new flavor sensitive interactions in an extra dimension , violation of quantum mechanics etc .
the @xmath10 survival probability goes as : @xmath176\ ] ] the super - kamiokande data can be fit by choosing @xmath177 and @xmath178 .
a detailed fit to all the data over the whole energy range has not been attempted yet .
as i mentioned at the beginning , the main motivation for this exercise is to try to establish neutrino oscillations ( due to mass - mixing ) as the unique explanation of the observed anomalies .
even if neutrinos have masses and do mix , the observed neutrino anomalies may not be due to oscillations but due to other exotic new physics .
these possibilities are testable and should be ruled out by experiments .
i have tried to show that we are beginning to carry this program out .
i thank professor klapdor - kleingrothaus for the invitation to this wonderful castle and for the hospitality here .
i thank andy acker , vernon barger , yuval grossman , anjan joshipura , plamen krastev , john learned , paolo lipari , eligio lisi , maurizio lusignoli and tom weiler for many enjoyable discussions and collaboration .
this work is supported in part by u.s.d.o.e . under grant de - fg-03 - 94er40833 .
some discussion of these scenarios can be found in s. bergmann and y. grossman , _ phys .
_ * d59 * ( 1999 ) 093005 ; l. m. johnson and d. mckay , _ phys .
lett . _ * b433 * ( 1998 ) 355 and p. herczeg , _ proceedings of international conference on particle physics beyond the standard model _ , jun 8 - 14 , 1997 ; castle ringberg , germany ; ed . by h. v. klapdor - kleingrothaus , ( 1998 ) 124 .
s. w. mansour and t - k .
kuo , hep - ph/9810510 ; see also a. halprin , c. n. leung and j. pantaleone , _ phys .
_ * d53 * ( 1996 ) 5365 ; j. n. bahcall , p. krastev and c. n. leung , _ phys .
* d52 * ( 1996 ) 1770 ; h. minakata and h. nunokawa , _ phys . rev . _
* d51 * ( 1995 ) 6625 .
s. pakvasa and k. tennakone , _ phys .
_ * 28 * ( 1972 ) 1415 ; j. n. bahcall , n. cabibbo and a. yahil , _ phys .
_ * 28 * ( 1972 ) 316 ; see also a. acker , a. joshipura and s. pakvasa , _ phys .
_ * b285 * ( 1992 ) 371 ; z. berezhiani , g. fiorentini , a. rossi and m. moretti , _ jetp lett . _
* 55 * ( 1992 ) 151
. p. f. harrison , d. h. perkins and w. g. scott , hep - ph/9904297 ; a. acker , j. g. learned , s. pakvasa and t. j. weiler , _ phys .
_ * b298 * ( 1993 ) 149 ; g. conforto , m. barone and c. grimani , _ phys .
_ * b447 * ( 1999 ) 122 ; a. strumia , _ jhep _ * 9904 * ( 1999 ) 026 .
y. fukuda et al.(the super - kamiokande collaboration ) , _ phys . rev
* 81 * ( 1998 ) 1562 ; _ ibid _ , * 82 * ( 1998 ) 2644 ; t. kajita hep - ex/9810001 .
the use of more recent data(k .
scholberg , hep - ex/9905016 ) does not change the conclusions .
j. valle , _ phys .
* 131b * ( 1983 ) 87 ; g. gelmini and j. valle ,
_ ibid _ * b142 * ( 1983 ) 181 ; k. choi and a. santamaria , _ phys .
* b267 * ( 1991 ) 504 ; a. joshipura and s. rindani , _ phys . rev . _ * d46 * ( 1992 ) 300 . | i review explanations for the three neutrino anomalies ( solar , atmospheric and lsnd ) which go beyond the `` conventional '' neutrino oscillations induced by mass - mixing .
several of these require non - zero neutrino masses as well . |
axially symmetric focusing of wave fields occurs in various areas of science , since physical systems often possess an intrinsic rotational symmetry .
in particular , the electromagnetic field enhancement by small spherical particles is important in many situations .
spheres have minimal surface energy for a given volume and thus are naturally formed as a result of phase separation , for example as aerosols or colloids .
applications of colloidal microspheres in photonic crystals and photonic crystal slabs led to an explosion of the experimental and theoretical studies of their optical properties.@xcite the majority of these investigations concentrate on their collective properties in a periodic arrangement .
single microspheres are used as high quality optical resonators and as agents that allow controlled and highly localized wavelength - dependent field enhancement for non - linear optical studies and in resonance spectroscopy.@xcite here , the emphasis is placed on the eigenmode analysis and the distribution of the field within the sphere or in the immediate vicinity of its surface . lately
, it was demonstrated that self - assembling arrays of transparent colloidal microspheres can be employed for high - throughput laser - assisted micro- and nano - structuring of materials.@xcite similar effects were observed in experiments on dry laser cleaning , where such particles are used as controlled contaminants.@xcite this necessitates better understanding of the focusing of light by microspheres with diameters of several wavelengths .
only few rigorous results are available for the intermediate range of sphere sizes and distances from the particle .
the majority of analytical approaches either deal with the properties of the eigenmodes , or refer to the integral characteristics and/or to the far field behavior.@xcite mie resonances were analyzed on the basis of advanced geometrical optics@xcite and detailed numerical calculations for transparent spheres of several wavelengths in size were performed in connection with the use of laser tweezers in biology@xcite and for needs of aerosol science.@xcite in this work we develop a theoretical description for an arbitrary non - paraxial strongly aberrated axially symmetric focusing and apply it to the case of dielectric microspheres .
our emphasis is on the fine structure of the field distribution in the exterior of the sphere up to the focal region , which can be used to control and improve the concentration of energy .
strong spherical aberration makes the focusing non - trivial .
usually , the exact solution is obtained using the mie theory,@xcite which does not give much of a physical insight as it requires the summation of a large number of terms in a multipole expansion even for moderate sphere sizes . at the same time , the main focusing properties of transparent dielectric microspheres originate rather from the picture of geometrical optics .
one might think that in the lowest approximation a small sphere acts as an ideal lens .
however , in the range of sizes we are interested in , this picture does not even provide a description which is qualitatively correct .
also classical formulas for weak spherical aberration @xcite do not yield useful results for the field behind a sphere : they predict that the maximum intensity is kept unchanged and its position does not depend on the wavelength .
our approach , following the method of uniform caustic asymptotics,@xcite is based on the canonical integral for the cuspoid ray topology of strong spherical aberration . though this bessoid integral
a member of the hierarchy of diffraction catastrophes@xcite appears naturally in the paraxial approximation , it can be used to describe arbitrary axially symmetric strong spherical aberration by appropriate coordinate and amplitude transformations . for angularly dependent vectorial amplitudes
the formalism uses higher - order bessoid integrals .
the bessoid integral is the axially symmetric generalization of the pearcey integral,@xcite which plays an important role in many short wavelength phenomena.@xcite therefore , the present approach can be applied in various areas of physics where axially symmetric focusing is of importance , e.g. , acoustics , semiclassical quantum mechanics , radio wave propagation and scattering theory .
we first consider the diffraction of a _
scalar _ spherically aberrated wave on a circular aperture with radius @xmath0 in the plane @xmath1 around the @xmath2-axis , where @xmath3 is the focal distance .
the origin of the coordinate system is put into the focus @xmath4 . in cylindrical coordinates @xmath5 , the paraxial fresnel
kirchhoff diffraction integral @xcite yields the field amplitude@xmath6 here @xmath7 is the amplitude of the incident wave in the center of the aperture , @xmath8 is the wavenumber ( @xmath9 , where @xmath10 is the wavelength ) and @xmath11 is the distance from the axis on the aperture .
the bessel function @xmath12 comes from the integration over the polar angle @xmath13 .
the parameter @xmath14 in the exponent determines the strength of the spherical aberration . for @xmath15
the diffraction focus shifts towards the aperture , while @xmath16 corresponds to ideal focusing.@xcite we introduce the dimensionless coordinates @xmath17% { 4\,k\,b}\,\widetilde{\rho}_{1}$ ] , @xmath18{k^{3}/4\,b}\,\rho / f$ ] and @xmath19 and consider an infinitely large aperture
. then the field ( [ eq fresnel - kirchhoff ] ) becomes proportional to the _ bessoid integral _
@xcite@xmath20 where@xmath21 its absolute square is shown in figure [ fig bessoid ] . in the cartesian representation @xmath22 and @xmath23
are dimensionless coordinates in the plane of integration .
expression ( [ eq bessoid ] ) is the axially symmetric generalization of the pearcey integral @xcite@xmath24 which is also shown in figure [ fig bessoid ] .
( top ) and the bessoid integral @xmath25 ( bottom ) .
the latter is proportional to the field of a spherically aberrated wave within small angles approximation . ]
both integrals correspond to so - called _
diffraction catastrophes_.@xcite their field distribution contains caustic zones where the intensity predicted by geometrical optics goes to infinity .
the pearcey integral corresponds to a cusp caustic , i.e. , a single one - dimensional curve in a two - dimensional space , and does not reveal a high intensity along the axis , while the bessoid integral corresponds to a _ cuspoid caustic _ , i.e. , to a surface of revolution of the cusp in three dimensions , as well as the caustic line up to the focus @xmath4 at @xmath26 . the equation of the cusp is given by the semicubic
parabola@xmath27 henceforth we will apply the term cusp also for the whole cuspoid .
a caustic is denoted as stable , if it does not change its topology under small perturbations .
this is the case for the pearcey integral .
the bessoid integral corresponds to a structurally unstable caustic , because an infinitely small perturbation will destroy the radial symmetry and the axis will not be a caustic zone any longer .
it is , however , stable on the class of axially symmetric wavefronts .
the cusp is the envelope of the family of rays .
the latter correspond to the _ points of stationary phase _ in the bessoid integral , i.e. , those points where the two first partial derivatives with respect to @xmath28 and @xmath29 of the phase @xmath30 in ( [ eq bessoid ] ) vanish . inside the cusp , for @xmath31 , three rays ( tangents to the cusp ) arrive at each point of observation @xmath32 , and outside , for @xmath33 , there is only one real ray ( figure [ fig rays ] ) .
thus , the cusp forms the border between the lit region and the ( partial ) geometrical shadow , where two rays merge .
-axis is represented by a dashed - dotted line . ] without loss of generality , we assume that all rays lie in the meridional plane @xmath34 ( @xmath35 ) and hence correspond to the roots @xmath36 ( @xmath37 ) of the cubic equation@xmath38 which are given by cardan s formulas.@xcite on the axis , @xmath39 , a cone formed by an infinite number of rays converges .
these rays originate from the circle @xmath40 on the aperture .
they are all in phase and produce a high intensity along the axis ( compare the two pictures in figure [ fig bessoid ] ) .
the oscillations occur due to interference with the ray propagating along the @xmath2-axis .
one can also directly observe that the bessoid integral has the topology of spherical aberration as the maximum of intensity does not lie in the geometrical focus @xmath41 but is spherically aberrated to a negative value of @xmath29 , i.e. , towards the aperture . off the caustic away from the cusp _ and the focal line _ the bessoid integral ( [ eq bessoid ] ) can be approximated by the method of stationary phase . as the integrand in ( [ eq bessoid ] ) is highly oscillatory , the only significant contributions to the integral come from those regions where the phase is stationary:@xcite@xmath42 where the summation runs over all real rays , i.e. , @xmath43 ( lit region ) or @xmath44 ( shadow ) .
the phase @xmath45 is obtained by inserting the @xmath46-th stationary point @xmath47 into the ( [ eq phase bessoid ] ) . the determinant and signature of the hessian are given by@xmath48 near the cusp the bessoid integral shows an airy - type behavior typical for caustics where two real rays disappear and become complex .
one can also derive a different approximation valid _ on and near _ the caustic axis , i.e. , for @xmath49 and small @xmath28 ( appendix [ appendix near axis bessoid]):@xmath50 here erfc is the complementary error function,@xcite which can also be written in terms of fresnel sine and cosine functions.@xcite expression ( [ eq near axis ] ) becomes exact at the axis @xmath39 , where @xmath51 .
it shows that near the axis the bessoid integral is virtually a bessel beam@xcite with a variable cross section . as the bessoid integrand is highly oscillatory ,
its evaluation for the whole range of coordinates @xmath28 and @xmath29 is non - trivial and of large practical importance .
direct numerical integration along the real axis and the method of steepest descent in the complex plane both have their disadvantages . by far
the fastest technique is based on the numerical solution of the _ ordinary differential equation _ ( derivation in appendix [ appendix differential])@xcite@xmath52 indices denote ( partial ) derivatives and @xmath53 is an abbreviation for the radial laplacian applied onto @xmath25 .
the three initial conditions at @xmath39 are@xmath54 @xmath55 was taken from ( [ eq near axis ] ) , @xmath56 vanishes due to symmetry , and the last condition arises from the fact that the bessoid integral satisfies the paraxial helmholtz equation @xmath57i@xmath58 , where @xmath59 is calculated from ( [ eq condition 1 ] ) . in the literature
the pearcey integral was calculated by solving differential equations,@xcite by a series representation @xcite and by the first terms of its asymptotic expansion.@xcite the bessoid integral was expressed in terms of parabolic cylinder functions @xcite and as a series.@xcite the latter work gives reference to an unpublished work of pearcey,@xcite stating that differential equations for the bessoid integral were employed there . in geometrical optics , the rays
carry the information of amplitude and phase . the total field in a point @xmath60 is given by the sum of all ray fields there .
a ray s field at @xmath60 is determined by@xcite@xmath61 where @xmath7 is the amplitude at some initial wavefront , @xmath62 is the eikonal , and @xmath63 is the generalized geometrical divergence , which can be calculated from flux conservation along the ray . for a homogeneous medium with constant refractive index@xcite@xmath64 @xmath65 ,
@xmath66 are the main radii of curvature at the point @xmath60 and @xmath67 , @xmath68 are the radii on the initial wavefront , where @xmath69
. when a ray touches a caustic , its radius of curvature ( the geometrical divergence in the general case ) changes the sign and the ray undergoes a _
caustic phase delay _
@xcite of @xmath70 , which is taken into account by the proper choice of the square root in ( [ eq ray ] ) .
when a ray touches several caustics , these delays must be added .
the total caustic phase shift , denoted as @xmath71 , can be explicitly written in the phase .
for the cuspoid topology and ray numbering ( @xmath37 ) according to figure [ fig rays ] , we obtain:@xmath72 with@xmath73{rl}% -\pi & \text { for } j=1,\\ 0 & \text { for } j=2,\\ -\pi/2 & \text { for } j=3 .
\end{array } \right.\ ] ] ray 1 touched the cuspoid and the focal line , ray 2 is not shifted , and ray 3 touched the cuspoid .
if we have found the phases @xmath74 and divergences @xmath63 of the rays , the ( scalar ) geometrical optics solution with an axially symmetric 3-ray cuspoid topology can be written as@xmath75 here @xmath76 are the real - space coordinates and we have allowed for different initial amplitudes @xmath77 of the rays .
this field shows singularities at the caustic , especially on the axis , which is the most interesting region for applications .
we want to describe _ arbitrary axially symmetric focusing _ by matching the solution of geometrical optics ( where it is correct ) with a wave field constructed from the bessoid integral ( [ eq bessoid ] ) , which naturally appears in the paraxial approximation and is finite everywhere , and its partial derivatives @xmath78 and @xmath59 ( _ method of uniform caustic asymptotics _ ) .
we make the _ ansatz _ @xcite@xmath79 the yet unknown arguments of the bessoid integral and its derivatives are @xmath80 .
@xmath81 , @xmath82 and @xmath83 are three amplitude factors and @xmath84 is a phase function . ( the indices @xmath28 and @xmath29 in the amplitudes do not indicate derivatives . ) now the geometrical optics solution ( [ eq geometrical solution ] ) is matched with the stationary phase approximation of ( [ eq ansatz ] ) by equating the amplitudes and phases:@xcite@xmath85 @xmath86 and @xmath87 are the partial derivatives of ( [ eq phase bessoid ] ) and@xmath88 where the determinant and signature of the hessian are written in ( [ eq det h ] ) and ( [ eq sign h ] ) , respectively . outside the cusp , the rays 2 and 3 are complex and
the general definition of @xmath89 is more subtle , namely @xmath90 with @xmath91/@xmath92 , @xmath93 ( the index @xmath46 denotes substitution of the @xmath46-th point of stationary phase as argument ) .
the three points of stationary phase were denoted as @xmath94 , where the @xmath95 are given by the ( correctly ordered ) cardan s solutions of ( [ eq cubic ] ) , i.e. , of@xmath96 note that they are functions of the bessoid coordinates , @xmath97 , and the latter depend on the real space coordinates : @xmath98 .
the partial derivatives with respect to @xmath28 and @xmath29 in ( [ eq matching amplitude ] ) must be evaluated in such a way as the @xmath95 were held constant , although they are functions of @xmath99 themselves .
the conditions ( [ eq matching amplitude ] ) and ( [ eq matching phase ] ) give 6 equations for the 6 unknowns @xmath28 , @xmath29 , @xmath100 , @xmath101 , @xmath102 , and @xmath103 .
it is convenient to solve ( [ eq matching phase ] ) , that is@xmath104 using quantities that are permutationally invariant with respect to the roots @xmath95.@xcite this yields@xmath105{\dfrac{2}{3}}\,\sqrt{-2\,\text{sgn}(b_{3})\,\sqrt{b_{2}% + q}+2\,\sqrt{d}}\,,\label{eq r , z , chi}\\ d & \equiv2\,b_{2}-q+2\sqrt{b_{2}^{2}-b_{2}\,q+q^{2}}\,,\nonumber\\ \chi & = b_{1}-\dfrac{1}{6}\,z^{2}\,,\nonumber\end{aligned}\ ] ] where sgn@xmath106sgn@xmath107 . the @xmath108 ( @xmath109 ) are given by @xmath110@xmath111 , @xmath112 and @xmath113 .
the quantity @xmath114 ( sometimes called discriminant ) can be expressed in different ways:@xmath115 hence , it vanishes exactly at the caustic where two phases are equal . at the cuspoid @xmath116 ( @xmath117 ) and on the axis @xmath118 ( @xmath39 ) . the solutions of ( [ eq matching amplitude ] ) , that is@xmath119 are@xcite@xmath120 where the cyclic terms permutate the numbering of rays : @xmath121 . the bessoid matching solution ( [ eq ansatz ] ) does not show the divergences of geometrical optics .
note that this method utilizes also the so - called complex rays which have less apparent physical meaning .
it turns out that both real and complex rays provide the _
geometrical skeleton for the wave flesh_.@xcite all formulas can be strongly simplified on and near the axis inside the cuspoid ( small @xmath122 , @xmath123 ) .
the bessoid coordinates have the simple form ( appendix [ appendix near axis coordinates])@xmath124% { ( \varphi_{1}+\varphi_{3})/2-\varphi_{2}}}\approx\dfrac{k\,\rho\,\sin\beta } { \sqrt{-z}}\,,\label{eq near axis r}\\ z & \approx-2\,\sqrt{\dfrac{\varphi_{1}+\varphi_{3}}{2}-\varphi_{2}}% \approx-2\,\sqrt{\varphi_{\text{np}}-\varphi_{\text{p}}}\ , , \label{eq near axis z}%\end{aligned}\ ] ] where @xmath125 is the local angle of the non - paraxial cone of rays with the axis and @xmath126 and @xmath127 denote the phases of the non - paraxial rays and the ( par)axial ray , respectively ( see figure [ fig nearaxis ] in appendix [ appendix near axis coordinates ] ) .
the simple natural combination@xmath128 also appears in the near axis approximation for the bessoid integral ( [ eq near axis ] ) . on the axis ( @xmath129 , @xmath118 ) we obtain @xmath39 and @xmath130 . the results ( [ eq near axis r])([eq bessel argument ] )
have transparent physical meaning .
indeed , near the axis the largest contribution to the field comes from the converging cone of non - paraxial rays ( similar to ray 1 ) that intersect the axis at an angle @xmath131 .
if the angle @xmath131 is constant and all rays have the same intensity , the result is the bessel beam.@xcite such beams have a propagation constant along the @xmath2-direction equal to the @xmath2-component of the wavevector of the plane waves which form them and correspondingly the argument of the bessel function ( cylindrical analog of a plane wave ) is equal to @xmath132 . as the angle @xmath131 gradually changes for the spherically aberrated wave , so does the argument of the bessel function .
additionally , there exists the axial ray 2 , which is not present in the canonical bessel beam ( though it often appears in real experimental situations ) .
the interference of this beam with the converging ray cone results in the intensity oscillations along the axis ( figure [ fig bessoid ] , bottom ) .
clearly , these oscillations are largely due to the phase difference @xmath133 . at large negative @xmath29 in ( [ eq near axis ] ) erfc@xmath134i@xmath135\rightarrow 2\,[1\!-\!\tfrac{1}{\sqrt{\pi}\,z}\exp($]i@xmath136 $ ] , and the oscillating behavior is governed by the phase of the exponent , which is equal to @xmath137 .
this clarifies the origin of expression ( [ eq near axis z ] ) , as it is @xmath138 entering the final formulas .
in particular , the global maximum is expected on axis at the first constructive interference of the axial and the non - paraxial rays . because @xmath139 in this region , the two terms of the erfc expansion are first in phase when the phase difference is @xmath140 .
the geometrical meaning of this result is that rays 1 and 3 are shifted by @xmath70 as they touch the cusp .
in addition , they acquire a further shift of @xmath70 when crossing the focal line .
but exactly _ on the axis _ only half of this delay has occurred yet , which yields the @xmath141 difference
. the numerical maximum of the bessoid intensity ( absolute square ) occurs at @xmath142 and hence this yields the condition@xmath143 which is close to @xmath144 .
the width of the focal line caustic , @xmath145 , is defined by the first zero @xmath146 of the bessel function in ( [ eq near axis ] ) .
hence , with ( [ eq bessel argument]),@xmath147 in the geometrical optics picture the first minimum occurs when rays 1 and 3 interfere destructively , i.e. , when their phase difference becomes @xmath148 .
this results in @xmath149 , where the term @xmath150 takes into account the caustic phase shift of ray 1 : @xmath151 .
note , that this is smaller than the airy spot for the same aperture angle@xcite and large angles @xmath131 are indeed realized , e.g. , in the case of the sphere studied below .
finally , we present an expression for the field ( [ eq ansatz ] ) on the axis . the equations for the amplitudes ( [ eq a , ar , az ] )
simplify tremendously ( appendix [ appendix on axis ] ) and result in@xmath152 \text{e}^{\,\text{i}\,\varphi_{2}}\ , .
\label{eq on axis field}%\ ] ] the structure of expression ( [ eq on axis field ] ) helps to understand its physical meaning .
it details the contribution of the cone of non - paraxial rays , represented by ray 1 ( first term ) , and the axial ray 2 ( second term ) to the overall structure of the field .
note that in the general case not only the angle @xmath131 , but also the amplitude of the converging cone may vary along @xmath2 ( @xmath29 ) , thus slowly modifying the properties of the bessel beam in the axial region .
this enters ( [ eq on axis field ] ) via amplitude transformations and is manifested by the presence of the initial ray amplitudes in both terms .
inside the cusp on the axis ( @xmath153 and @xmath154 ) both @xmath155 and the ratio @xmath156 remain finite as the divergence of the paraxial ray 2 is non - singular , while the sagittal divergence @xmath157 of the cone of non - paraxial rays 1 is proportional to @xmath122 . due to the bessoid matching procedure
the singularity of the converging cone is removed by the compensating factor @xmath158 . along the axis the last term in ( [ eq on axis field ] ) partly cancels with the second term in the parentheses of the first term . as a result ,
the on axis field behavior up to the focus is dominated by a single term proportional to the bessoid integral @xmath25 , which justifies the maximum condition ( [ eq phase difference axis ] ) discussed above .
often especially in _ vectorial _ problems there exists axial symmetry with respect to the wavefronts , ray phases and generalized divergences , but not with respect to the amplitudes . in this case ,
new functions are required to represent arbitrary angular dependence of the field .
the natural generalization of ( [ eq bessoid ] ) are the _ _ higher - order bessoid integrals__@xcite with the non - negative integer @xmath159:@xmath160 where @xmath161 and @xmath162 are higher - order bessel functions .
the higher - order bessoid integrals obey the recurrence relation@xmath163 the integral @xmath164 is canonical for angular dependent geometrical field components @xmath165 or @xmath166 . in matching similar to ( [ eq ansatz]),@xmath167 the angular dependence cancels . here
@xmath168 , @xmath169 and @xmath170 are the higher - order amplitude factors , whereas @xmath171 and @xmath172 are partial derivatives of the higher - order bessoid integrals @xmath164 .
since the latter can be written in terms of @xmath173 , it can be shown that the points of stationary phase , the matching of phases and thus the higher - order coordinates ( @xmath65 , @xmath174 ) and phases ( @xmath175 ) are identical with the original ones:@xmath176 from the physical point of view , this reflects the _ conservation of the wavefront _ and thus the ray phases and divergences .
the equations for the amplitudes have to be generalized .
the higher - order amplitudes @xmath168 , @xmath169 and @xmath170 have the same form as ( [ eq a , ar , az ] ) , but with an additional factor @xmath177i@xmath178 in each denominator , i.e.,@xmath179 a more detailed description of the higher - order bessoid integrals as well as the derivation of the recurrence relation and the amplitude equations can be found in appendix [ appendix higher - order ] .
consider a plane wave falling on a transparent sphere in vacuum .
figure [ fig sphere ] illustrates the refraction of a single ray in the meridional plane , containing the point of observation @xmath60 and the axis . within the frame of geometrical optics
the cuspoid is formed behind the sphere in analogy to figure [ fig rays ] .
to @xmath60 by a sphere with radius @xmath0 and refractive index @xmath180 .
the picture is drawn in the meridional plane and all indicated angles are positive . ]
let @xmath0 be the sphere radius and @xmath181 its refractive index .
in contrast to the previous sections , we choose the origin of the axially symmetric cylindrical coordinate system @xmath5 differently now , namely as the center @xmath182 of the sphere .
the incident plane wave propagates parallel to the @xmath2-axis .
the geometrical optics focus , formed by the paraxial rays , is located at @xmath183 with @xcite@xmath184 a ray passes the point @xmath185 , is first refracted at @xmath186 , a second time at @xmath187 and propagates to @xmath60 . the incident and transmitted angle , @xmath188 and @xmath189 ,
are related by snell s law , @xmath190 .
writing the position of @xmath32 in polar coordinates , @xmath191 and @xmath192 , one can find the following expression , determining the three rays that arrive at @xmath60:@xmath193 where one has to substitute @xmath194 $ ] .
this is a transcendental cubic - like equation which has three roots , either all real or one real and two complex conjugate .
( for @xmath195 this is true for @xmath196 ; if @xmath197 , the 3-ray region does not start until some distance behind the sphere . )
we denote them as @xmath198 and choose their order consistently with the previous notations .
therefore , @xmath199 is always real and negative , whereas @xmath200 and @xmath201 are either real and positive ( lit region ) with @xmath202 or complex conjugate ( geometrical shadow ) .
when the @xmath203 are known , we find the @xmath204 from snell s law and the @xmath205 and @xmath206 from@xmath207 omitting the index @xmath46 , the three ray coordinates can be written as @xmath208 the eikonal is the optical path accumulated from @xmath185 to @xmath60 ( on the dashed vertical line in figure [ fig sphere ] all rays are still in phase):@xmath209 the sphere radius @xmath0 was subtracted from the path contributions to make the eikonal zero in the center @xmath182 , if there were no sphere .
next we calculate the geometrical optics amplitudes by determining the meridional and sagittal radii of curvature , @xmath65 and @xmath66 , and their changes due to refraction .
formulas for the refraction on an arbitrary surface with arbitrary orientation of the main radii exist in the literature.@xcite a simple derivation for the sphere can be found in appendix [ appendix radii ] .
it yields the dependence of the actual radii of curvature @xmath65 and @xmath66 ( right after the refraction ) on the initial radii @xmath210 and @xmath68 ( just before the refraction):@xmath211 for a plane wave , @xmath212 , the radii of curvature in the points @xmath186 ( inside the sphere ) and @xmath187 ( outside the sphere ) have the compact form@xmath213 the overall geometrical generalized divergence after both refractions reads ( index @xmath46 omitted)@xmath214 where @xmath215 is the distance of propagation within the sphere .
note that ray 1 has a negative angle @xmath216 .
besides , a double caustic phase shift should be added ( manually ) to the phase of this ray ( minus sign ) as in ( [ eq caustic phase shift ] ) .
the caustic shifts of the rays 2 and 3 are taken into account automatically if the branch cut for the square roots in ( [ eq divergence sphere ] ) is along the negative real axis from @xmath217 to @xmath218 and the branch with @xmath219i is used . in this procedure
it is not allowed to multiply the radicands and write them under one common square root . also
the case of complex rays 2 and 3 is covered correctly by this convention .
finally , the geometrical optics solution for the sphere is given by ( [ eq geometrical solution ] ) , where the eikonal @xmath62 and divergence @xmath63 are given by ( [ eq eikonal sphere ] ) and ( [ eq divergence sphere ] ) .
the equation determining the three rays is ( [ eq cubic like sphere ] ) . in the geometrical shadow
the sum ( [ eq geometrical solution ] ) becomes only the term with @xmath220 .
to incorporate fresnel transmission coefficients , we assume that the incident light is linearly polarized in @xmath221-direction , i.e. , the incident electric field vector is@xmath222 with @xmath223 the unit vector in @xmath221-direction and @xmath224 . since axial symmetry is broken , we introduce the polar angle @xmath13 which is measured from @xmath221 to @xmath225
. the point of observation @xmath226 will be reached by three rays ( two may be complex ) and their angles @xmath203 are still determined by ( [ eq cubic like sphere ] ) , for all three rays lie in the meridional plane , containing @xmath60 and the @xmath2-axis ( figure [ fig projection]a ) . the initial @xmath148- and @xmath227-polarized components depend on @xmath13 ( figure [ fig projection]b ) : @xmath228 we define the overall transmission coefficients @xmath229 here the @xmath230 ( @xmath231 ) are the standard fresnel transmission ( reflection ) coefficients @xcite from the medium 1 , i.e. , vacuum , into the medium 2 , i.e. , the sphere .
to @xmath60 in the meridional plane ( plane of incidence ) .
( b ) decomposition of the initial electric field vector with length @xmath232 into its @xmath148- and @xmath233-component parallel and perpendicular to the meridional plane ] the ray field behind the sphere is found by the projection onto the original cartesian system @xmath234 .
we write the components of the transmission vector @xmath235 and show the ray index @xmath37 explicitly .
the @xmath13-dependence is indicated with the superscript @xmath236:@xmath237{lll}% t_{x , j}=t_{j}^{(0)}+t_{j}^{(2)}\cos2\,\varphi\ , , & & t_{j}^{(0)}\equiv \dfrac{t_{\pi , j}\cos\beta_{j}+t_{\sigma , j}}{2}\,,\smallskip\\ t_{y , j}=t_{j}^{(2)}\sin2\,\varphi\ , , & & t_{j}^{(1)}\equiv t_{\pi , j}\sin \beta_{j}\,,\smallskip\\ t_{z , j}=t_{j}^{(1)}\cos\varphi\ , , & & t_{j}^{(2)}\equiv\dfrac{t_{\pi , j}% \cos\beta_{j}-t_{\sigma , j}}{2}\ , . \end{array } \label{eq transmission components}%\ ] ] hence , the geometrical optics solution for the electric field @xmath238 including the eikonal @xmath62 ( [ eq eikonal sphere ] ) and divergence @xmath63 ( [ eq divergence sphere ] ) reads@xmath237{lll}% e_{x , j}=e_{j}^{(0)}+e_{j}^{(2)}\cos2\,\varphi\ , , & & \\ e_{y , j}=e_{j}^{(2)}\sin2\,\varphi\ , , & & e_{j}^{(m)}\equiv e_{0}%
\,\dfrac{t_{j}^{(m)}\,\text{e}^{\,\text{i}\,k\,\psi_{j}}}{\sqrt{j_{j}}}\,.\\ e_{z , j}=e_{j}^{(1)}\cos\varphi\ , , & & \end{array } \label{eq electric field components}%\ ] ] matching each term @xmath239 by the ansatz ( [ eq ansatz ] ) in its higher - order formulation ( [ eq ansatz higher - order ] ) with the corresponding integral @xmath164 , we obtain the vectorial electric field @xmath240 in the form@xmath241{c}% 1\\ 0\\ 0 \end{array } \right ) + e^{(1)}\!\left ( \begin{array } [ c]{c}% 0\\ 0\\ \cos\varphi \end{array } \right ) + e^{(2)}\!\left ( \begin{array } [ c]{c}% \cos2\,\varphi\\ \sin2\,\varphi\\ 0 \end{array } \right ) \!. \label{eq electric field matched}%\ ] ] figure [ fig contourxzyz ] illustrates the intensity , i.e. , the absolute square of the electric field @xmath242 , for @xmath34 ( @xmath221,@xmath2-plane ) and @xmath243 ( @xmath225,@xmath2-plane ) . in the normalized @xmath221,@xmath2-plane ( top ) and in the @xmath225,@xmath2-plane ( bottom ) .
contour shadings go from white ( zero ) to black ( @xmath244 ) .
parameters : refractive index @xmath245 , dimensionless wavenumber @xmath246 .
the initial electric field vector is @xmath247 .
the sphere with radius @xmath0 is situated in the origin .
the focus of geometrical optics is located at @xmath248 , whereas the diffraction focus ( the point of maximum intensity ) is significantly shifted towards the sphere : @xmath249 . in dimensional units , for a wavelength of @xmath250@xmath251
m the sphere radius is @xmath252@xmath251 m . ]
the magnetic field @xmath253 can be calculated similarly ( incident magnetic field @xmath254 , @xmath255 ) and the ( normalized ) poynting vector is given by @xmath256 .
on the axis the electric field is given by its @xmath221-component only ( direction of polarization ) due to averaging over @xmath13 in ( [ eq electric field matched ] ) .
for @xmath257 ( inside the cusp ) it is given by the analytical expression ( [ eq on axis field ] ) .
after several simplifications @xcite it can be written as@xmath258 \text{e}^{\,\text{i}\,\varphi_{2}}\ , , \label{eq electric field on axis}%\ ] ] where the transmission factors @xmath259 are given in ( [ eq transmission components ] ) and for dielectric spheres have the form:@xmath260 the phases in the coordinate @xmath130 are@xmath261 and @xmath262 is the first ray s _ compensated sagittal divergence_:@xmath263\,[(r_{s , q_{1}})_{1}+d_{1}]}}\nonumber\\ & \quad\quad\quad\times\dfrac{\sqrt{(r_{m , q_{2}})_{1}\,(r_{s , q_{2}})_{1}}% } { \sqrt{(r_{m , q_{2}})_{1}+s_{1}}}\,\sqrt{2\,k}\,\sin\beta\nonumber\\ & = -2\,\sqrt{2\,k\,a}\,\cos(\beta_{1}/2)\nonumber\\ & \quad\quad\quad\times\sqrt{\dfrac{\cot\theta_{i,1}\cos\theta_{t,1}% \sin(\beta_{1}/2)}{1 + 1/n^{2}-3\sin^{2}(\beta_{1}/2)/\sin^{2}\theta_{i,1}}}\ , , \label{eq compensated divergence}%\end{aligned}\ ] ] which manifestly has no singularity until the geometrical focus where @xmath264 .
the minus sign comes from the manually inserted phase shift of the first ray .
( all aforementioned quantities should be expressed in terms of the angles @xmath199 and @xmath265 as described in detail in section [ sect instruction ] below . )
the structure of the first two lines in ( [ eq compensated divergence ] ) is general and is valid for arbitrary axially symmetric systems .
@xmath266 is always finite on the axis , since both the sagittal radius of curvature and the phase difference @xmath267 are proportional to the distance @xmath122 .
equation ( [ eq electric field on axis ] ) is valid even near the focus , since the diverging terms @xmath268 and @xmath269 almost cancel . for @xmath270 , however , the divergence of @xmath266 itself becomes important , as the non - paraxial ray 1 becomes axial . in figure [ fig contouraxis ]
we show the position and the value of the maximum of @xmath271 as a function of the refractive index and the dimensionless product @xmath272 , calculated from ( [ eq electric field on axis ] ) .
the @xmath2-coordinate of this global maximum is denoted with @xmath273 ( _ diffraction focus _ ) and the intensity there is @xmath274 .
contrary to the square dependence for the case of an ideal lens,@xcite even for macroscopic spheres the maximum intensity turns out to be about proportional to @xmath272 , in agreement with the general theory.@xcite and @xmath272 ( contour lines from top to bottom go from 1.1 to 3.0 in steps of 0.1 ) .
right : intensity enhancement at @xmath273 ( contour lines from bottom left to top right are 20 , 50 , 100 , 200 , 500 , 1000 and 2000 ) ] the main contribution in ( [ eq electric field on axis ] ) stems from the bessoid integral , that is from the term @xmath275 .
thus , the position of the maximum can be estimated from condition ( [ eq phase difference axis ] ) , i.e. , @xmath276 .
if the phase difference @xmath277 from ( [ eq phases sphere 1 ] ) and ( [ eq phases sphere 2 ] ) is expressed as a function of @xmath199 , taylor expanded and equated to @xmath141 , then we get in the lowest non - trivial order of the inverse product @xmath272:@xmath278 hence , in the limit of small wavelengths or large spheres the relative difference between the diffraction and the geometrical focus decreases proportionally to the inverse square root of @xmath272 .
the factor @xmath279 can be replaced by the more exact bessoid value 2.327 from ( [ eq phase difference axis ] ) .
expression ( [ eq diffraction focus ] ) approximates the position of the maximum within an error of @xmath280 for @xmath281 and values of the refractive index in the range @xmath282 .
the transcendental phase difference condition ( [ eq phase difference axis ] ) , which holds for large angles , naturally has a wider range of applicability . with very good accuracy
the diffraction focus also provides the maximum for the absolute square of the magnetic field , @xmath283 , as well as for the @xmath2-component of the poynting vector @xmath284 .
note that on the axis @xmath285 and @xmath286 .
for convenience we summarize the sequence of steps that should be used for the calculation of the field behind a sphere irradiated with linearly polarized light on the basis of the formulas developed above : 1 . _ finding the rays _ : the origin of the coordinate system is in the center of the sphere .
choose a point @xmath287 behind the sphere and numerically calculate the 3 rays arriving at @xmath60 .
these rays are characterized by the 3 incident angles @xmath203 ( @xmath37 ) , found numerically from ( [ eq cubic like sphere ] ) , and numbered according to figure [ fig rays ] .
all other angles follow from snell s law and from ( [ eq alpha beta ] ) , respectively . outside the cuspoid the rays 2 and 3 are complex .
2 . _ the geometrical optics solution _ : compute the geometrical optics solution for the electric field ( [ eq electric field components ] ) , which is the sum of contributions from three rays .
the eikonals @xmath288 are calculated from ( [ eq eikonal sphere ] ) with ( [ eq path ] ) .
the geometrical optics amplitudes are given by the fresnel transmission components @xmath289 ( [ eq transmission components ] ) and the generalized divergence factors @xmath290 ( [ eq divergence sphere ] ) , which follow from the radii of curvature ( [ eq r1])([eq r4 ] ) and the distances @xmath291 from the sphere to @xmath60 ( [ eq path ] ) .
the conventions for the complex roots shall correctly add up all individual caustic phase shifts .
ignore the fact that the geometrical field diverges near caustic regions .
3 . _ bessoid matching _ : starting from the eikonals @xmath288 , determine the bessoid coordinates , i.e. , first @xmath29 , and then @xmath28 and @xmath292 ( [ eq r , z , chi ] ) . next , compute and correctly order the points of stationary phase @xmath95 ( [ eq stationary points ] ) , most conveniently using trigonometric formulas.@xcite with the generalized divergences ( [ eq divergence sphere ] ) and
the hessians ( [ eq hessian general ] ) the bessoid amplitudes @xmath168 , @xmath169 and @xmath170 can be computed from ( [ eq am ] ) for all orders @xmath293 .
the electric field component @xmath294 associated with the @xmath159-th order results from the ansatz ( [ eq ansatz higher - order ] ) .
the bessoid - matched field @xmath295 is finally given by ( [ eq electric field matched ] ) .
in the case of a scalar plane wave ( or exactly on the axis ) only the order @xmath296 contributes .
proceed accordingly for the magnetic field @xmath253 , employing different transmission coefficients in step 2 .
the bessoid integral _ : the final solution ( [ eq electric field matched ] ) contains the bessoid integral and higher - order bessoid integrals as well as their partial derivatives .
the bessoid integral @xmath297 can be efficiently computed numerically via the differential equation ( [ eq ordinary ] ) .
the higher - order bessoid integrals follow from a recursive relation ( [ eq bessoid recursive ] ) .
_ remarks _ : the speed limiting bottleneck of this procedure is the finding of the rays @xmath203 in step 1 .
the numerical evaluation of the bessoid integral is very efficient and in all other steps analytical expressions are applied . when approaching the caustic ( axis , cuspoid ) , individual quantities the geometrical optics amplitudes
diverge but their combinations remain finite . in our calculations we observed perfect numerical stability up to distances from the caustic of the order of @xmath298 times the sphere radius , which is more than enough for any practical purposes .
we presented a general way to match geometrical optics solutions with the bessoid integrals .
it can be applied to any axially symmetric system with the cuspoid topology of spherical aberration . for the sphere we can compare our approximate results with the theory of mie.@xcite a main quantity characterizing the sphere
is the dimensionless mie parameter @xmath299 .
figure [ fig comparisonaxis ] compares the intensity on the axis obtained from the mie theory with the bessoid approximation .
the parameters are as in figure [ fig contourxzyz ] and the mie parameter is @xmath300 , @xmath301 , @xmath302 and @xmath303 . on the axis .
dashed lines represent the mie theory , solid lines are the results of bessoid matching .
the parameters are as in figure [ fig contourxzyz ] and the cases ( a ) , ( b ) , ( c ) and ( d ) correspond to @xmath300 , @xmath301 , @xmath302 and @xmath303 , respectively . in dimensional units , for @xmath250@xmath251 m this corresponds to sphere radii of @xmath304@xmath251 m , @xmath305@xmath251 m , @xmath306 m and @xmath307 m . ] we see very good agreement down to @xmath308 ( @xmath309 ) . for @xmath310 ( @xmath311 ) the asymptotic behavior far from the sphere is still correct . however , for small @xmath114 the characteristic scale @xmath0 is no longer large compared to the wavelength @xmath10 and geometrical optics becomes invalid .
next , we compare the off - axis electric and magnetic field as well as the @xmath2-component of the poynting vector , @xmath312 ( figure [ fig comparisonmie ] ) .
right behind the sphere ( @xmath313 ) the agreement is not perfect ( see figure [ fig comparisonaxis ] ) , though all qualitative features are preserved .
sections at @xmath314 already show good agreement ( figure [ fig comparisonmie ] ) and for @xmath315 the pictures become visually almost indistinguishable . discussing the quality of these results , one has to differentiate between the accuracy of the method and the influence of those factors which can be taken into account , but were not included into the current consideration . , @xmath316 and @xmath312 in the normalized @xmath221,@xmath225-plane for @xmath314 calculated with bessoid matching ( left ) and with the mie theory ( right ) .
the parameters are the same as in figure [ fig contourxzyz ] . ]
the accuracy of the bessoid matching procedure itself was studied separately for the case of a spherically aberrated wave incident onto an aperture . the deviation
defined as the maximal relative error of the intensity @xmath317 between the bessoid matched geometrical optics solution and the corresponding ( exact ) rayleigh sommerfeld diffraction integral@xcite decreases as the aperture increases . if the aperture is large enough , the deviation is below @xmath318 for spherical abberation strengths and wavevectors approximately corresponding to the focusing by spheres studied in figures [ fig comparisonaxis ] and [ fig comparisonmie ] . for the sphere , for all investigated mie parameters @xmath319 , the deviations from the exact mie solution are about @xmath320 in those regions where the intensity is not very small ( including the caustic axis and cuspoid ) .
a detailed analysis indicates that this deviation originates from several factors : \(i ) influence of the finite size of the sphere .
there exist diffractive contributions from creeping rays@xcite propagating along the sphere surface .
they can in principle be accounted for at the expense of the simplicity of the procedure , essentially by considering the interference of the bessoid field with these additional rays .
this results in oscillations which can actually be seen in the mie curve on the right side of the bessoid tail in figures [ fig comparisonaxis]a and b. \(ii ) rays entering the sphere undergo multiple interior reflections .
some of them satisfy resonance conditions , accumulate significant energy inside the sphere and refract outside .
this produces an additional field behind the sphere , in particular the intensity becomes non - zero in the regions of geometrical shadow for the directly transmitted rays used in the bessoid matching . here
again , one can ( in principle ) study such multiply reflected rays separately and add them to the bessoid field , but their contribution to the focusing properties of the microspheres is of secondary importance .
\(iii ) finally , very close to the sphere surface at distances of the order of a fraction of @xmath10 , there exist evanescent contributions , which are taken into account in the mie theory , but are obviously absent in the bessoid matching procedure .
thus , the quality of the bessoid matching in the most interesting regions near caustic surfaces is quite satisfactory .
it rectifies the divergences of geometrical optics , which is asymptotically correct for large @xmath114 in non - singular regions of space . clearly , the procedure has to be extended whenever contributions from rays other than those 3 used for matching become significant .
the field distribution behind the sphere has a rich fine structure ( figure [ fig comparisonmie ] ) which our geometrical approach helps to clarify .
it is known that the ring - type field enhancement corresponds to the cuspoid caustic , having the approximate radial distance@xcite@xmath321 our approach explains the double - peak structure of @xmath271 along the direction of polarization .
it is related to the axial field component @xmath322 and can be understood in terms of geometrical optics . on the axis ,
the @xmath322 components from the rays 1 and 3 point into opposite directions and cancel , having an effective phase difference of @xmath148 . off the axis ,
ray 1 underwent a caustic phase shift of @xmath70 when crossing the axis , which makes the condition for constructive interference : @xmath323 .
then , according to ( [ eq bessel argument ] ) , the peak occurs at the radial distance@xmath324 more details on the derivation are given elsewhere@xcite together with the refined coefficient 0.293 ( instead of @xmath325 ) obtained from the bessoid asymptotic ( [ eq higher - order bessoid axis ] ) .
double - peak structures have been observed in nano - patterning experiments@xcite and were semi - quantitatively explained on the basis of the mie solution.@xcite in an actual experiment it may depend on the laser pulse parameters and the properties of the patterned material , whether the poynting vector or the electric field is responsible for the patterning process . for small spheres this double peak effect can be understood using the near field pattern for a scattering dipole.@xcite the present explanation ( for sphere diameters of a few wavelengths and larger ) results in the same orientation of the maxima and thus these two limiting cases cover almost all range of sphere sizes .
similar polarization dependence of the field distribution in focal regions can be used to improve the resolution.@xcite
we described theoretically arbitrary axially symmetric aberrated focusing and studied light focusing by microspheres as an example . following the method of uniform caustic asymptotics,@xcite we introduced a canonical integral describing the wave field for the given cuspoid ray topology .
this bessoid integral appears naturally in the paraxial approximation . in some regions ( off the caustic or exactly on the axis ) it reduces to simple analytical expressions . in other regions we efficiently computed this highly oscillatory integral via a single ordinary differential equation . for arbitrary axially symmetric focusing , coordinate and amplitude transformations match the bessoid wave field and the solution of geometrical optics .
the caustic divergences of the latter are removed thereby . for vectorial problems with angularly dependent field components ,
higher - order bessoid integrals are used for the matching procedure .
the formulas significantly simplify on and near the axis .
an approximate universal condition for the diffraction focus can be given in terms of phase differences . here
, the concept of caustic phase shifts is of main importance .
the central part of the bessoid integral is essentially a bessel beam@xcite with a variable cross section due to the variable angle of the non - paraxial rays .
its local diameter is always smaller than in the focus of an ideal lens with the same numerical aperture . besides , the largest possible apertures can be physically realized , which is hardly possible with lenses .
all this is achieved at the expense of longitudinal confinement . as an example
the focusing of a linearly polarized plane wave by a transparent sphere is studied in detail .
we calculate the geometrical optics eikonals and divergences , incorporate fresnel transmission coefficients and perform bessoid matching . using the general theory ,
simple expressions for the light field on the axis and for the diffraction focus are derived .
the two strong maxima in the intensity observed immediately behind the sphere can be explained as well .
finally , the results of the bessoid matching procedure are compared with the mie theory .
the agreement is good for mie parameters @xmath326 .
near the sphere the correspondence is worse due to unaccounted evanescent contributions .
the developed formulas can be directly applied in other areas of physics where non - paraxial axially symmetric focusing is of importance , e.g. , acoustics , semiclassical quantum mechanics,@xcite flat superlenses based on left - handed materials,@xcite radio wave propagation , scattering theory,@xcite chiral conical diffraction@xcite etc .
concluding , let us briefly enumerate several possibilities to extend and refine the developed formalism .
weak absorption can be incorporated easily , for it just changes the amplitudes along the rays and the transmission coefficients .
strong absorption additionally modifies snell s law of refraction , still preserving the axial symmetry .
one can consider incoming radially or azimuthally polarized beams , which are known to produce better resolution than linear polarization.@xcite the diffraction of light from regions beyond the sphere radius can be incorporated by considering the interference of the bessoid field with creeping rays.@xcite for other geometries , in particular finite apertures with sharp boundaries , edge rays or the rubinowitz representation,@xcite or an approach based on catastrophe theory@xcite have to be used .
such corrections become relevant , for example , for the ray structure and the field distribution immediately behind spheres with a refractive index @xmath197 .
finally , one can calculate the interference of the diffracted light with the original incident wave or the interference of the light refracted by several spheres or arrays of spheres .
the latter yields interesting secondary patterns @xcite related to the so called talbot effect.@xcite the authors thank d. buerle ( johannes kepler university , linz ) for many stimulating discussions on microsphere patterning experiments , which initiated this study , and for his continuous support of this work .
the authors also thank b. lukyanchuk and z. b. wang ( both at the data storage institute , singapore ) for their mie program and discussions on mie calculations .
j. k. appreciates helpful conversations with g. langer ( johannes kepler university , linz ) .
n. a. thanks v. palamodov ( tel aviv university ) for illuminating mathematical suggestions .
financial support was provided by the fwf ( austrian science fund ) under contract no .
n. a. also thanks the christian doppler laboratory of surface optics ( johannes kepler university , linz ) .
we make the substitution @xmath327 in ( [ eq bessoid polar]):@xmath328 near the axis ( small @xmath28 ) the bessel function is slowly varying compared with the exponent .
the integral will have significant contribution only from the region in which the exponent s phase is stationary , i.e. , regions near @xmath329 .
we consider the most interesting caustic part of the axis for which @xmath49 . in a lowest order approximation
the bessel function is considered as constant near the stationary point @xmath330 and can be pulled out of the integral .
the phase can be written as a complete quadratic form . with the full square of @xmath331:@xmath332
the remaining integral can be expressed in terms of the complementary error function@xcite erfc ( of complex argument ) and hence we arrive at ( [ eq near axis ] ) .
for @xmath333 the point @xmath334 should be taken as a stationary edge point of the integration.@xcite and the near axis approximation ( [ eq near axis ] ) remains valid as long as the bessel function is set to @xmath51 .
we derive the paraxial helmholtz equation@xmath335 as well as the following _ ordinary differential equation _ for the bessoid integral:@xmath336 indices denote partial derivatives .
both equations can be rewritten in the compact form@xmath337 where @xmath338 is the radial laplacian@xmath339 we begin with the proof of ( [ eq differential bessoid 1 ] ) and state that we may differentiate under the integral sign , since the partial derivatives of the integrand exist and are continuous functions . starting from the bessoid integral in the polar representation ( [ eq bessoid polar ] ) , its integrand can be written as@xmath340 with the abbreviation@xmath341 the ( multiple ) partial derivatives are@xmath342\,e\,,\label{eq
grr}\\ g_{z } & = -\text{i}\,\dfrac{\rho_{1}^{3}}{2}\,j_{0}(r\,\rho_{1})\,e\,.\end{aligned}\ ] ] here we used the derivative formula for bessel functions@xcite@xmath343 with @xmath296 to obtain ( [ eq gr ] ) and @xmath43 for ( [ eq grr ] ) . note that @xmath344 . applying the recurrence relation for bessel functions@xcite@xmath345 one can eliminate @xmath346 from ( [ eq grr ] ) .
and then it is enough to notice and verify that@xmath347 this proves equation ( [ eq differential bessoid 1 ] ) . for the proof of ( [ eq differential bessoid 2 ] ) , we need to note that its left hand side can be expressed as the integral of a partial derivative@xmath348 \!\text{d}\rho_{1}. \label{eq definite integral}%\ ] ] with the help of ( [ eq bessel derivative ] ) and ( [ eq bessel recurrence ] ) both the left hand side of ( [ eq differential bessoid 2 ] ) and @xmath349 become@xmath350\,e\,\text{d}\rho_{1}.\ ] ] thus , in order to prove ( [ eq differential bessoid 2 ] ) , it is enough to show that @xmath351 .
this follows from the newton - leibniz formula applied to the ( definite ) integral ( [ eq definite integral]):@xmath352 _ { 0}^{\infty}=0\,.\ ] ] the lower bound at @xmath218 vanishes for obvious reasons .
for the upper bound at @xmath353 one assumes an _ infinitely small _ imaginary part in front of the fourth order term in the exponent : @xmath354i@xmath355 with @xmath356 .
this completes the proof of the differential equations ( [ eq differential bessoid 1 ] ) and ( [ eq differential bessoid 2 ] ) for the bessoid integral .
near the axis the phases of the rays can be taylor expanded . from figure [ fig nearaxis ]
one infers that up to the first order in @xmath122 the phases can be written as@xmath357 here @xmath126 and @xmath127 denote the phases of the non - paraxial rays and the ( par)axial ray ( with @xmath129 ) and @xmath125 is the angle of ray 3 with the axis . ) by @xmath358 , whereas the phases of ray 2 and the ( par)axial ray are the same in first order . the non - paraxial ray and ray 3 cross the axis at an angle @xmath125 . ]
we insert these phases into the exact expressions for @xmath28 and @xmath29 in ( [ eq r , z , chi ] ) , taylor expand the result with respect to @xmath122 and resubstitute @xmath359 , @xmath360 and @xmath361 from ( [ eq phases taylor ] ) .
this yields ( [ eq near axis r ] ) and ( [ eq near axis z ] ) .
here we derive a simple on axis expression for the bessoid - matched field @xmath362 ( [ eq ansatz ] ) , namely equation ( [ eq on axis field ] ) . on the axis and inside the cusp @xmath129 ( @xmath39 ) and @xmath123 ( @xmath139 ) .
the stationary points , given by ( [ eq stationary points ] ) , are@xmath363 then , the amplitude @xmath101 in ( [ eq a , ar , az ] ) simplifies to@xmath364 because on the axis the ratios @xmath365 are both finite and the corresponding other terms disappear upon multiplication with @xmath366 . due to the restriction to the lit region ( @xmath139 ) ,
all rays are real and equation ( [ eq hessian lit ] ) holds . by virtue of ( [ eq det h ] ) @xmath367 , and due to ( [ eq sign h ] )
sign@xmath368 , one finds@xmath369 and thus:@xmath370 this approximation for the amplitude @xmath101 is valid up to the focus ( @xmath41 ) . as ray 2 converges like the inverse distance from the focus , @xmath371 is proportional to @xmath29 . the amplitude @xmath102 in ( [ eq a , ar , az ] ) vanishes due to @xmath372 and the fact that@xmath373 which means that rays 1 and 3 have equal amplitudes and that the caustic phase shifts are in accordance with the signature of the hessian .
consequently,@xmath374 with ( [ eq ts on axis ] ) and ( [ eq ray 1 and 3 ] ) the amplitude @xmath103 reads@xmath375 the first term is non - trivial . both @xmath376 and @xmath377
are zero on the axis , but their ratio is finite and well defined . indeed , the taylor expansion of cardan s solution @xmath378 in its trigonometric representation@xcite yields in the first order in @xmath28:@xmath379 therefore , again in first order in @xmath28 : @xmath380 .
due to sign@xmath381 we obtain@xmath382 and with ( [ eq sqrt h2]):@xmath383 this approximation for @xmath103 holds for small values of @xmath28 .
it is finite , since @xmath377 approaches zero as @xmath384 for @xmath385 . for the final representation of the field @xmath362 , we can substitute the near axis expression for @xmath386 ( [ eq bessel argument ] ) into @xmath103 . on the axis
the phase coordinate becomes @xmath387 , which results from substituting @xmath118 and @xmath388 into the corresponding expression in ( [ eq r , z , chi ] ) .
this leads to@xmath389 using the linear relationship between the bessoid integral and its @xmath29-derivative ( [ eq condition 3]),@xmath390 we end up with equation ( [ eq on axis field ] ) .
_ higher - order bessoid integrals _ ( [ eq higher - order bessoids ] ) appear naturally , if one expands an arbitrary initial field amplitude on the aperture in a fourier series:@xmath391\,.\ ] ] the form of the coefficients @xmath392 and @xmath393 can be seen from a two - dimensional taylor expansion in cartesian coordinates around the point @xmath394 , rewritten into polar coordinates:@xmath395 with@xmath396 thus , @xmath397 is the lowest possible power of @xmath398 which can be found in the term with @xmath399i@xmath400 .
an additional @xmath398 comes from the transformation from cartesian to polar coordinates . if we define the functions@xmath401 we find that they satisfy the paraxial helmholtz equation@xmath402 where @xmath403 .
due to ( [ eq bessel derivative ] ) and ( [ eq bessel recurrence ] ) , one can write the identity@xmath404\nonumber\\ & \quad\quad+\dfrac{m}{r}\,\rho_{1}^{m+1}\,j_{m}(r\,\rho_{1})\,.\end{aligned}\ ] ] hence , the recursive relation for the bessoid integrals ( [ eq bessoid recursive ] ) follows:@xmath405 i.e. , @xmath406 , @xmath407 , etc . using ( [ eq bessoid recursive1 ] ) and ( [ eq bessel recurrence ] ) as well as expression ( [ eq near axis ] ) for @xmath25 , one obtains@xmath408 which is the analytic near axis expression for the higher - order bessoid integrals . while the coordinates and phases ( @xmath28 , @xmath29 , @xmath292 ) remain unchanged , the derivation of the higher - order amplitudes ( @xmath168 , @xmath169 , @xmath170 ) requires some insight for @xmath409 . let us briefly consider the case @xmath410 . for the matching procedure we need the asymptotic behavior of @xmath411 far from the caustic regions where @xmath412 and where it is dominated by the term @xmath413 .
note that for the matching procedure we need exactly this asymptotic representation and in the non - caustic regions only .
thus although we need the second - order bessoid integral @xmath414 on and near the axis , where it vanishes we shall use its asymptotic stationary phase expressions far from the axis for the derivation of the amplitudes . in this region
it is equivalent to the asymptotic of @xmath413 .
in fact , we may generalize this statement to arbitrary order .
due to ( [ eq bessoid recursive1 ] ) the leading term in the stationary phase calculation is always@xmath415 therefore , the equations for the amplitudes ( [ eq matching amplitudes ] ) become@xmath416 which can be seen from the bessoid integral s cartesian representation with the phase ( [ eq phase bessoid]):@xmath417 the equations ( [ eq matching amplitudes higher - order ] ) have the same form as ( [ eq matching amplitudes ] ) except an additional factor @xmath177i@xmath178 on the left hand side , proving ( [ eq am ] ) .
let us consider a point source @xmath418 and start with the derivation of the meridional radius of curvature ( figure [ fig radius1 ] ) .
the initial radius of curvature is @xmath419 , the one after refraction is @xmath420 .
the infinitesimally neighbored beam ( @xmath421 which is refracted in @xmath422 ( the angles of incidence and transmission in @xmath423 are @xmath216 and @xmath189 , in @xmath422 they be denoted @xmath424 and @xmath425 ) also propagates to @xmath426 .
the normals onto @xmath210 through @xmath423 and onto @xmath65 through @xmath422 are @xmath427 and @xmath428 , respectively . in the necessary order
the length of the arc @xmath429 can be approximated by the distance @xmath430 . as all angles are small : @xmath431 , @xmath432 , and @xmath433 .
on the other hand , we find from the infinitesimal triangles : @xmath434 , @xmath435 .
this leads to@xmath436 where we have introduced a minus sign because the wave is converging after the refraction .
the remaining problem is the angle @xmath437 in the denominator . to find @xmath437 we write the relations between angles and primed angles:@xmath438 with snell s law@xmath439 a first order taylor expansion in @xmath440 yields@xmath441 we express @xmath437 from ( [ eq delta ] ) , substitute it into ( [ eq r1 first ] ) and finally obtain the meridional radius of curvature ( [ eq rm])@xmath442 a ray which emerged from @xmath418
is refracted in @xmath423 ( incident angle @xmath216 and transmitted angle @xmath189 ) .
the distance from @xmath423 to the intersection @xmath349 of the ray with the line passing through @xmath418 and the sphere center @xmath182 is the sagittal radius of curvature , as a neighbored ray , emerging from @xmath418 and hitting the sphere not in @xmath423 but infinitesimally shifted perpendicular to the meridional plane , will also propagate to @xmath349 due to symmetry around the line @xmath443 .
we have @xmath444 and @xmath445 .
the tangent theorem states ( all angles are in general large now)@xmath446 where @xmath447 is just the third angle in the triangle @xmath448 . due to @xmath449 we find@xmath450 in the triangle @xmath451 the sine theorem reads@xmath452 where we have again introduced a minus sign due to the convergence of the refracted wave .
trigonometric transformations finally give the sagittal radius of curvature ( [ eq rs])@xmath453 in the present work all plots of the bessoid integral ( and its derivatives ) contain @xmath458 mesh points , computed with the software package mathematica 5 ( wolfram research ) .
direct numerical integration of ( [ eq bessoid polar ] ) takes more than one hour on a modern personal computer .
integration along a line in the complex plane decreases the amount of time by approximately a factor of 3 . solving the ordinary differential equation ( [ eq ordinary ] ) ,
however , lasts only a few seconds .
j. n. connor and p. r. curtis , j. math .
phys . * 25 * , 2895 ( 1984 ) . in this work the term _ cuspoid _ stands for the whole family of canonical catastrophe integrals with corank 1 , beginning with fold ( codimension 1 ) , cusp ( 2 ) , swallowtail ( 3 ) and buttererfly ( 4 ) , but
does not mean the surface of revolution of a cusp with corank 2 and codimension 2 as we use it . | an analytical description of arbitrary strongly aberrated axially symmetric focusing is developed .
this is done by matching the solution of geometrical optics with a wave pattern which is universal for the underlying ray structure .
the corresponding canonical integral is the bessoid integral , which is a three - dimensional generalization of the pearcey integral that approximates the field near an arbitrary two - dimensional cusp .
we first develop the description for scalar fields and then generalize it to the vector case . as a practical example the formalism
is applied to the focusing of light by transparent dielectric spheres with a few wavelengths in diameter .
the results demonstrate good agreement with the mie theory down to mie parameters of about 30 .
compact analytical expressions are derived for the intensity on the axis and the position of the diffraction focus both for the general case and for the focusing by microspheres .
the high intensity region is narrower than for an ideal lens of the same aperture at the expense of longitudinal localization and has a polarization dependent fine structure , which can be explained quantitatively .
the results are relevant for aerosol and colloid science where natural light focusing occurs and can be used in laser micro- and nano - processing of materials . |
in many crystal - growing procedures of interest , a nano - scale faceted surface appears and proceeds to evolve , often exhibiting coarsening and even dynamic scaling , whereby characteristic statistics describing the surface remain constant even as the characteristic lengthscale increases through the vanishing of small facets . for many evolving faceted surfaces , a _ facet velocity law _
can be observed @xcite , assumed @xcite , or derived @xcite which specifies the normal velocity of each facet , often in configurational form which depends on the geometry of the facet . in this way , the dynamics of a continuous , two - dimensional surface can be concisely represented by a discrete collection of such velocities , and overall computational complexity reduced to that of a system of ode s ; the resulting system is known as a _ piecewise - affine dynamic surface _ , or pads . such theoretical simplification , in turn , enables the large - scale numerical simulations necessary for the statistical investigation of coarsening and dynamic scaling .
the numerics involved in the direct geometric simulation of an arbitrary pads is straightforward for one - dimensional surfaces , requiring nothing beyond traditional ode techniques except simple geometric translation between facet displacement and edge displacement , and a small surface correction associated with each coarsening event .
consequently , such simulations accompany many of the above facet treatments of facet dynamics , and have also been independently repeated elsewhere @xcite .
however , in two dimensions , the corrections due to coarsening events are much more involved , and any code must be able to deal with a family of non - coarsening _ topological events _ that alter the neighbor relations between nearby facets .
consequently , the fewer simulation attempts use either fast but poentially imprecise envelope methods @xcite , or more robust but slower phase - field @xcite or level - set @xcite methods to avoid explicitly performing topological changes . besides the speed / accuracy trade - off exhibited by these approaches , both methods obscure the natural geometric simplicity of the native surface , complicating the extraction of detailed surface statistics which
, after all , motivates large simulations in the first place .
additionally , as will be seen , the presence of non - unique topological events requires explicit intervention regardless of topological scheme , which negates much of the advantage of a `` hands - free '' treatment . in the previous chapter
, we introduced a direct - simulation method which explicitly performs topological events along the way , thus preserving both simulation speed and topological accuracy .
in addition , by representing the surface as a collection of facets , edges , and junctions , plus the neighbor relations between them , the method mirrors the natural geometry of the surface being modeled , which allows easy extraction of geometric statistics .
there , however , the restricted case of threefold symmetry was chosen for ease of topological implementation ; under this symmetry , a limited number of topological events were observed , and both vanishing facets and non - vanishing surface rearrangements could be handled explicitly using a priori knowledge of the before and after surface states . while many surfaces exhibit threefold symmetry , making the method useful even in this special case , it could not handle other common crystal symmetries , notably fourfold and sixfold . in this chapter , then , we generalize the previous model to allow the simulation of surfaces with arbitrary symmetry groups .
we begin in section [ sec : cellular - structure ] with a brief summary of the basic method , including surface representation , facet kinematics , and the application of a dynamics .
next , in section [ sec : topological events ] , we provide a careful enumeration of topological events which may occur on surfaces of arbitrary symmetry ; this includes discussion of the far - field reconnection algorithm , by which network holes left by vanishing facets may be consistently repaired without knowledge of the post - event state .
then , we provide in section [ sec : discretization and topology ] a careful consideration of the consequences of using ( necessarily discrete ) timesteps during the simulation of a surface whose evolution equations change qualitatively between steps ( at topological events ) ; the issues that arise are discussed in the context of three sample strategies .
the completed method is illustrated from three- , four- , and six - fold symmetric surfaces in section [ sec : demonstration ] ; these exhibit all of the topological events likely to be encountered on a real surface , and demonstrate that the method is robust enough to generically simulate faceted surfaces of any symmetry class for which a facet - velocity law is uniquely specified .
finally , in addition to detailing the ffr algorithm , the appendix includes a discussion of kinematically non - unique topological events , where two resolutions are possible , and highlights the need to refer to the dynamics or even first principles to decide how the surface should evolve in those cases .
we consider the evolution of a single - valued , fully - faceted surface @xmath1 ; this definition explicitly forbids overhangs and inclusions .
we assume that the surface bounds a single crystal which exists on exactly one lattice ; thus , we are not treating surfaces with multiple grains .
the surface is piecewise - affine , consisting of facets @xmath2 with fixed normals @xmath3 .
these are bounded by and meet at edges @xmath4 which are necessarily straight line segments ; edges in turn meet at triple - junctions @xmath5 .
this three - component structure is reminiscent of two - dimensional _ cellular networks _ @xcite and indeed , while we consider three dimensional surfaces , the projection of the edge set onto the plane @xmath6 is a 2d cellular network .
this structure and the neighbor relations inherent within it suggest a doubly - linked object - oriented data structure , consisting of : ( 1 ) a set of junctions , each having a location , pointing to three edges and three facets ; ( 2 ) a set of edges , each having a tangent , pointing to two junctions and two facets ; and ( 3 ) a set of facets , each having a normal , pointing to @xmath7 edges and @xmath7 junctions .
these objects and the associated neighbor relations are illustrated in figure [ fig : neighbor - relations ] ; this structure is the natural structure of the surface , and uniquely and exactly describes it .
we now consider each element in more detail .
[ fig : neighbor - relations ] a _ junction _ is a point in space formed where edges ( and hence , facets ) intersect .
the * order * @xmath8 of a junction is simply the number of edges which meet there . while junctions of any order @xmath9 are possible , we restrict ourselves here to the case of order @xmath10 junctions or `` triple junctions . ''
this greatly simplifies analysis and code , as triple junctions are uniquely positioned by the three facets meeting there .
the intrinsic geometric information carried by a junction is its location .
junctions are stored in a ` junction ` class , which contains this location , as well as pointers to the three edges and three facets which meet there .
an _ edge _ is a line segment formed by the intersection of exactly two facets , and bounded by exactly two junctions .
the intrinsic geometrical quantity of an edge is its orientation , which is fixed since facets have fixed normals .
edges are stored in the ` edge ` class , which records the tangent , as well as pointers to the two neighboring facets and two bounding junctions . at creation
, edges are `` directed '' : one junction is arbitrarily deemed the origin , and the other the terminus , establishing a _
tangent_. this has two important consequences .
first , if we imagine walking along the edge in the tangent direction , then one neighboring facet may be labeled `` left '' , and the other `` right . ''
this information allows us to distinguish between convex and concave edges , and also to determine the clockwise direction around a given facet , which is necessary for effective navigation of the network , as well as the proper calculation of boundary integrals on facets .
second , the tangent allows us to detect when an edge `` flips '' ( see @xcite ) ; this will be discussed in more detail in section [ sec : topology - class-1 ] . a facet is a simply - connected planar polygonal region in space , which is bounded by an equal number of edges and junctions . the intrinsic geometric information carried by a facet is its normal , which is fixed .
our surface definition @xmath11 requires that the normal of each facet is constrained to be on the hemispherical shell of unit - length vectors with positive @xmath12 component .
the imposition of a particular symmetry on the crystal may further restrict available normals , but no such restriction is here assumed .
facets are stored in a ` facet ` class , which contains the normal , as well as a list of bounding edges and junctions , sorted in counter - clockwise order .
the intrinsic geometric means of characterizing surface evolution is by specifying the normal velocity of each point on the surface .
a piecewise - affine surface is composed of a collection of planar , fixed - normal facets , whose motion is limited to displacement along the normal .
therefore , the kinematics @xmath13 of the entire surface may be expressed by a discrete set of individual facet velocities @xmath14 .
as edges and junctions are merely intersections between two and three facets , respectively , their motion is uniquely specified by the motion of the facets that neighbor them . in particular , if @xmath15 is the location in space of a triple junction , then the velocity of that a triple junction may be calculated through the expression @xmath16 where the rows of @xmath17 and entries of @xmath18 are the unit normals and normal velocity , respectively , of the three facets intersecting to form @xmath15 . in practice ,
facet velocities are only used indirectly to calculate junction velocities if junctions are moved correctly , edges ( connections between two junctions ) and thus faces ( collections of edges ) are necessarily moved correctly as well .
all that remains now is to select a particular dynamics ; that is , to specify an expression for the normal velocity @xmath14 of each facet . having chosen one , we follow @xcite and refer to the resulting evolving structure as a piecewise - affine dynamic surface ( @xmath19 ) .
example dynamics describing many different physical situations were listed in the introduction , and the exact dynamics is not of special concern here ( although we will select one for demonstration later ) .
it is worth noting here , however , that most of the dynamics proposed to date are _ configurational _ , depending on properties of the facet such as area , perimeter , number of junctions , or mean height .
thus , sudden changes in the geometric properties of a facet can lead to sudden changes in its velocity , an issue which will be explored in more detail in section [ sec : discretization and topology ] .
we have just discussed how elements of each class ( facet , edge , vertex ) neighbor members from each of the other classes . taken together
, the set of all of these neighbor relations comprises the * topological state * of the surface .
it is a complete record of every neighbor relationship on the surface , and is unique for a given surface .
as the system evolves , these neighbor - relations may change as facets exchange neighbors , join together , split apart , or vanish .
each of these cases is an example of a * topological event * , and represents a change to the topological state of the surface ( topological events are a defining feature of evolving cellular networks again see @xcite ) . to maintain an accurate representation of the surface , a direct geometric method like that described here must manually perform topological events as necessary . because actual surface evolution is fairly trivial , this is the main difficulty of our method .
a natural first question to ask at this point is `` how many topological events are possible ? '' to begin answering this question , we point out that on a physical surface , topological events occur automatically , and by geometric necessity . if a detected event signals the need to change neighbor relationships at some location on the surface , we may therefore infer that failing to change them would produce a cellular network with `` wrong '' relationships , that do not correspond to a physical surface .
we call such erroneous configurations * geometrically inconsistent * ; examples include primarily edge networks that intersect when viewed from above , since these correspond to overhangs and inclusions , which are prohibited .
since topological events serve to avoid possible geometric inconsistencies , we may discover what events are possible by considering how inconsistencies may occur .
this is most easily accomplished by considering each surface element in turn .
we first consider junctions , which are simply a location in space .
a junction can , in the course of surface evolution , leave the periodic domain , in which case it is wrapped to the other side .
however , this is only a bookkeeping operation , and does not represent a real topological event . turning to edges , we note that edges possess a directed length . as already hinted in section [ sec : edges ] , this length could become negative if the edge were to `` flip '' @xcite a flipped edge has no geometrical meaning on a single - valued surface , and so we introduce a class of * vanishing edge * events which occur when edges reach zero length .
finally , we consider facets .
since a facet has fixed orientation , its changing properties are loosely its shape and size .
specifically , a facet is a _ simply connected _ planar region with _
positive area_. these two defining properties of facets lead , through consideration of their potential violation , to two additional classes of topological event : * facet constriction * events which prevent the formation of self - intersecting facets , and * vanishing facet * events which remove facets from the network when they reach zero area . an adjacent point - point event occurs when an edge shrinks to zero length , and its junctions meet . to consider what might happen to the faceted surface
when this occurs , we first label the immediate surroundings of an edge .
each edge is composed of two faces of which it is the intersection , its * composite faces * , and stretches between two faces at which it terminates , its * terminal faces*. in addition , we will also use the term * emanating edges * to refer to those edges immediately neighboring the shrinking edge .
now , consider the hemispherical shell of available facet normals ( section [ sec : faces ] ) .
the ( necessarily distinct ) normals of the composite faces specify a great circle about this hemisphere , which divides it into two parts .
the normals of the terminal faces can not lie on this boundary , and unless they are identical ( a special case ) , they form a second great circle around the hemisphere .
while terminal normals may not lie on the composite great circle , the reverse is not true , and this fact effectively divides vanishing edge events into three sub - classes . [
fig : shrinking edge classes ] figure [ fig : shrinking edge classes ] illustrates this idea , and gives an example of each of the three possible cases . if the terminal great circle touches neither composite point , then the well - studied * neighbor switch * occurs .
if the terminal great circle touches one composite point , then an * irregular neighbor switch * results .
finally , if the terminal normals occupy the same point , then no great circle is defined the terminal facets have he same normal , and when the edge between them shrinks to zero , they join into a single facet : a * facet join*. on a general surface , the most common vanishing edge event is the neighbor switch , which is frequently encountered in other evolving cellular networks . in this event , neither composite normal touches the terminal great circle , so any three of the normals involved form a linearly independent set this property is the defining feature of the neighbor switch .
when an edge with this configuration shrinks to zero length , the surrounding facets simply exchange neighbors . figure
[ fig : t1-illustration ] gives an example of this event .
+ [ fig : t1-illustration ] _ resolution .
_ the neighbor switch is performed by the ` ns_repairman ` class . to resolve this event
, it simply deletes the old edge , and creates a new edge .
the composite faces which formed the old edge become terminal faces of the new edge , and cease to neighbor each other .
conversely , the terminal faces of the old edge become the composite faces of the new edge , and thus become neighbors . this symmetric exchange in neighbor relations
is the cause of the name neighbor switch , which comes from the grain - growth literature the less - descriptive name `` t1 process '' in often used in the soap froth literature .
in addition to replacing the vanishing edge , the junctions on either side of this edge are replaced .
each new junction is formed by the intersection of the deleted edge s ( formerly non - adjacent ) terminal faces with one of its composite faces .
_ comments .
_ readers familiar with other cellular - network literature will note that the example neighbor switch in figure [ fig : t1-illustration ] lacks the typical `` x '' shape .
this is due to the constrained nature of facet normals , and hence , edge orientations .
additionally , we note that the neighbor switch is a reversible event ; in fact it is its own reversal .
finally , a certain sub - class of neighbor switches posessing `` saddle '' structure are non - unique , as was observed by thijssen @xcite . for a discussion of this non - uniqueness and its consequences , see appendix [ sec : non - uniqueness ] .
when the normal of one of the composite faces lies on the great circle formed by the terminal normals , the neighbor switch can not occur . here
, the terminal faces can not form a new junction with the offending composite face because the three normals involved are not independent . instead ,
when an edge with this configuration shrinks to zero , two closely related events are possible , depending on the configuration of the nearby edges .
these events are collectively called irregular neighbor switches , with two varieties called a `` gap opener '' and `` gap closer '' that are exact opposites .
these are illustrated in figure [ fig : gap opener ] .
[ fig : gap opener ] _ resolution . _
the irregular neighbor switch is performed by the ` ins_repairman ` class . because one composite normal lies on the terminal great circle , exactly two of the emanating edges are parallel in @xmath20 .
the gap opener occurs when these edges emanate from the shrinking edge in _ opposite _ directions , while the gap closer occurs when the edges emanate in the _ same _ direction . to resolve the gap opener , we select one of the parallel emanating edges to be split apart ( see below ) .
the gap will go here , filled by the terminal face that touches the other parallel edge , and will extend all the way to the far end of the split edge , where a new edge is introduced to link the two edges resulting from the split edge .
this is all illustrated in figure [ fig : gap opener ] . to resolve the gap closer , simply reverse the steps .
_ comments .
_ several comments on this pair of events are in order .
first , while the gap closer is uniquely resolved , the gap - opener is an inherently non - unique event , as either of the parallel edges could be the one split ( we will discuss this further in section [ sec : non - uniqueness ] ) .
second , both resolution options have the potentially dissatisfying property of being non - local in effect , because the collision of two junctions causes an entire edge to split apart . what is perhaps more likely is the nucleation of a new , tiny facet at the moment the junctions collide ; however , we have excluded that possibility from consideration here .
finally , while common experimentally - encountered surfaces usually have either high symmetry ( only a few facet orientations ) or no symmetry ( as many orientations as facets ) , the irregular neighbor switch with its three coplanar orientations requires what may be called `` intermediate symmetry , '' where orientations are limited , but many are available . because it poses resolution difficulties , and
because it is not encountered in any surfaces we wish to study , we have not yet actually implemented this event .
finally , we consider the special case where the terminal normals are identical .
when such an edge shrinks to zero length , the terminal faces meet exactly .
having the same orientation , they then join to form a larger face .
figure [ fig : face - join - illustration ] depicts a representative facet join event .
[ fig : face - join - illustration ] _ resolution .
_ facet joins are performed by the ` fjoin_repairman ` class . to perform a facet join ,
a new face is created to replace the joining faces , and all edges and junctions that neighbored the old faces are re - assigned to this new face .
next , the vanishing edge and its two junctions are deleted , leaving the four emanating edges to be considered .
these are most logically grouped into the ( necessarily parallel ) pairs of edges bordering , respectively , the left and right composite faces of the vanished edge . in the example event shown in figure
[ fig : face - join - illustration ] , these two pairs look different : one pair meets side - to - side , while the other pair meets end - to - end .
computationally , however , this makes no difference ; each pair is replaced by a single edge connecting their remaining non - deleted junctions .
this behavior is generic for all face joins .
_ comments .
_ we note that the face join is , strictly speaking , non - reversible ( though see section [ sec : face - split ] ) .
the exact opposite of the face join would be a facet which spontaneously `` shatters , '' as described in @xcite ; this behavior is certainly worth studying , but is not currently implemented .
second , although this is a `` special case '' in general , for high - symmetry crystal surfaces it may be very common indeed , for the case of a cubic crystal with only three available facet orientations considered in chapter 2 , facet joins are the only vanishing edge event exhibited .
finally , we note that this event is the only vanishing edge event which does not conserve the number of facets .
it is , in fact , one mechanism by which coarsening may occur , and may be the dominant mechanism for high - symmetry surfaces .
the second class of topological event occurs whenever a facet ceases to be simply - connected , and results in that facet being split into two new facets . remembering that the edges of a facet trace out a polygon in the plane , we observe that the non - simply connected polygon , if allowed to continue evolving , would become self - intersecting , which clearly has no geometrical interpretation .
so , how may an evolving polygon become self - intersecting ?
since the boundary consists of edges and junctions , there are three possible modes : ( a ) two non - adjacent junctions meet , ( b ) a junction meets an edge , or ( c ) two edges meet .
each case has a distinct `` signature , '' illustrated in figure [ fig : self - intersection - signatures ] , which can be used to tell them apart .
[ fig : self - intersection - signatures ] the junction - junction collision shown in figure [ fig : self - intersection - signatures]a represents the formation of a perfect @xmath21 junction . while theoretically interesting , such events are not considered here
; we hypothesize that , given random initial data , two junctions not connected by an edge will never exactly meet .
furthermore , by considering figure [ fig : self - intersection - signatures ] , it can be seen that all junction - junction collisions , if perturbed as we hypothesize , result in either junction - edge or edge - edge collisions , and can therefore be resolved accordingly .
junction - edge collisions occur when a facet is pinched into two pieces by three of its neighbors , depicted in figure [ fig : self - intersection - signatures]b .
there , two adjacent neighbors of the facet , forming a wedge , meet a third neighbor and pierce it .
two separate events are possible in this class . in most cases ,
the wedge simply splits the central facet into two parts , in an event called a * facet pierce*. however , if the normals of the wedge facets and the normal of the central facet lie on the same great circle , then , as the central facet is split , the opposing facet opens up a gap in the wedge : an * irregular facet pierce*. edge - edge collisions occur when a facet is pinched by four neighbors , shown in figure [ fig : self - intersection - signatures]c . in these events , two non - adjacent , exactly parallel edges meet , which requires that the normals of the impinging facets be coplanar with the normal of the pinched facet .
again , two variations are possible .
if the impinging faces have different normals , the event is called a * facet pinch*. however , if they have the same normal , they join even as they pinch the facet in question , in a process called a * joining facet pinch*. in addition , each event may occur in either symmetrical or asymmetrical flavors , which are shown in figure [ fig : self - intersection - signatures]c1,c2 respectively .
the meeting of two edges requires the involvement of two junctions ; these lie on the same edge for the symmetrical case , and on different edges for the asymmetrical case , as seen in the figure .
the first self - intersection we will study is the simplest ; the facet pierce .
it is a point - line event as described above ; that is , a facet is split when a triangular wedge formed by two adjacent neighboring facets intersects the edge formed with a third , opposing neighbor .
the facet pierce is functionally the opposite of a facet join , and is illustrated in figure [ fig : face - split - illustration ] .
[ fig : face - split - illustration ] _ resolution .
_ each facet pierce is performed by the ` fjoin_repairman ` class .
given the constricted facet , as well as the junction and edge which meet , it can label all of the surrounding facet elements and deterministically reconnect them correctly .
first , two new facets are created to replace the constricted facet .
the junctions and edges that bordered the old facet can be reassigned to these based on the labels created initially .
the colliding junction and edge are deleted , to be replaced by three new junctions and two new edges .
the locations of the former and neighbor relationships of each can be determined by considering figure [ fig : face - split - illustration ] and using the labels .
_ comments . _ first , technically , at the moment of the event , an @xmath22 junction forms , which as shown in figure [ fig : face - split - illustration ] may proceed to break in one of three ways .
this does not , however , constitute a non - uniqueness ; rather , the dynamics governing the surface evolution at the moment of topological change specify which exit pathway is chosen .
second , while thijssen @xcite rightly objected to this resolution for the case of separate grains , we find it satisfactory for the case of a single crystal considered here .
a special modification of the facet pierce just described occurs when the normal of the opposing facet shares a great circle with the normals of the facets forming the wedge .
this event is called an irregular facet pierce .
recall that three new junctions were created during the facet pierce .
however here , since the two newly created facets have identical normals , and the remaining three have normals which are not independent , those junctions can not be created . instead
, as the wedge facets meet the opposing facet , one of two things happen either the center edge of the wedge is split apart by the opposing facet ( a `` wedge split '' ) , or the opposing facet is split apart by the wedge ( a `` wedge extension '' ) . we see an illustration of each possibility in figure [ fig : table - split - illustration ] .
[ fig : table - split - illustration ] _ resolution . _
the irregular facet pierce is performed by the ` ifp_repairman ` class , which at instantiation is given the constricted facet , as well as the junction and edge which meet .
this event is repaired quite similarly to the regular facet pierce , with modifications .
as is done there , two new facets are created to replace the constricted facet , and junctions and edges bordering the old facet are reassigned to the new ones .
the resolution differs in how to replace the colliding junction and edge .
if the `` wedge split''resolution is chosen , then the middle edge of the wedge and its far junction are also deleted
these are replaced by two parallel edges and junctions .
finally , an edge is formed which links them and borders the facet on the far side of the deleted edge . if the `` wedge extension '' resolution is chosen , not only is the constricted facet split apart , but so is the one opposite the edge split by the wedge .
one must first determine which edge of this second split facet the extended wedge will intersect .
having done so , that facet is deleted , to be replaced by two new facets .
the extension is formed by adding two edges parallel to the middle edge of the wedge , and the edge it intersects is split in two .
two new edges and three junctions must be created to link the extension with the edge it intersects .
finally , all edges and junctions bordering the deleted facet , plus those created to form the extension , are re - assigned appropriately to the new facets .
figure [ fig : table - split - illustration ] is especially helpful here .
_ comments . _
the event clearly recalls the `` gap opener '' described above .
it shares with that event three coplanar surface normals , and as a result , two possible resolutions .
additionally , while the two options here are qualitatively different compared to the symmetric options of the gap opener , they are additionally both non - local effects due to a local cause .
again , perhaps the best resolution is to nucleate a new facet , which we do not yet consider .
finally , both events require `` intermediate symmetry , '' and for the same reasons discussed above , we have not implemented this event .
we now turn to consider the case of edge - edge events , the first of which is called a face pinch . here
the normals of the pinching facets are not identical , and so junctions can be created as needed an illustration of this event is shown in figure [ fig : face - pinch - illustration ] .
this event is philosophically similar to the face split described above . in each case , a facet is split into two by non - joining neighbors ; the difference is just whether the procedure is `` sharp '' or `` blunt '' ; i.e. , caused by parallel edges or a junction and an edge .
[ fig : face - pinch - illustration ] _ resolution .
_ each facet pinch is performed by the ` fpinch_repairman ` class .
because of the similarities between the facet pierce and facet pinch , the associated ` repairman ` classes behave similarly as well . here , the ` repairman ` class constructor takes the constricted facet and the two colliding edges . with this information
, it can label all of the surrounding facet elements and deterministically achieve the change shown in figure [ fig : face - pinch - illustration ] . as with the facet pierce ,
two new facets are created to replace the constricted facet , and the junctions and edges that bordered the old facet are reassigned as required .
the colliding edges are deleted , as are the associated junctions discussed above .
five edges and four junctions are created to complete the reconnection , as shown in figure [ fig : face - pinch - illustration ] .
_ comments .
_ this event , like the irregular neighbor switch and irregular facet pierce , requires a surface with `` intermediate symmetry . ''
while it is uniquely resolved and poses no great difficulty of implementation , we have not yet implemented it for this reason .
finally , a special modification of the face - pinch occurs when the impinging facets have identical normals .
the constricted is split in exactly the same way as in a face pinch ; however , since the two facets doing the `` pinching '' are identically oriented , they join together to form a larger facet .
we see an illustration of this situation in figure [ fig : face - swap - illustration ] .
[ fig : face - swap - illustration ] _ resolution .
_ each joining facet pinch is performed by the ` jfpinch_repairman ` class , which operates similarly to the ` repairman ` classes associated with the facet pierce and facet pinch .
this class is again instantiated with the constricted facet and the two meeting edges , which allows the necessary labeling .
again , two new facets are created to replace the constricted facet , but in this case the two facets which meet must join , and so another new facet must be created to replace them
necessary junction and edge reassignments are again easily carried out . finally , rather than deleting the edges which meet and the associated junctions involved , the meeting edges are simply re - connected as shown in figure [ fig : face - swap - illustration ] .
_ comments . _
note that the final configuration is similar to the original configuration ; in fact , with suitable facet motion , the surface could return to its original configuration via another face swap ; the event is thus self - reversible in a sense .
also , since both a facet pinch and a facet join occur simultaneously , the total numbers of each surface element remain unchanged during this event .
the final class of topological event occurs when a facet shrinks to zero area and is removed .
however , as has been noted numerous times previously in the context of cellular networks , very small facets can result in stiff dynamics that are difficult to numerically simulate accurately .
for this reason , we follow previous authors by pre - emptively removing facets with areas below some small threshold ( but see section [ sec : pure - predictive - method ] ) .
this process is summarized in figure [ fig : facet - removal - illustration ] . there , we see a single small flat facet vanishing into a pentagonal well ( [ fig : facet - removal - illustration]a ) . being smaller than the allowed threshold , it is removed , leaving a `` hole '' in the network ( [ fig : facet - removal - illustration]b ) . the facets and edges bordering this hole we call the * far field * , and
they need to be reconnected correctly to patch the hole .
the correct reconnection for this particular well is shown in figure [ fig : facet - removal - illustration]c .
[ fig : facet - removal - illustration ] the principal difficulty in this process occurs during the reconnection step ( figure [ fig : facet - removal - illustration]c ) . here
, we are assigning new neighbor relationships to the far - field facets , which also involves the creation of new edges and junctions to form boundaries between them .
in other cellular - network problems , these neighbor relationships ( and hence the reconnection ) is usually chosen randomly , under the reasoning that any error introduced is small enough to neglect and quickly corrected . however , because the faceted - surface network represents a piecewise - planar geometrical surface , we are not free to choose randomly . since each facet in the far field has a normal and a local height , neighbor relationships determine junction locations and thus edge placement .
however , the final reconnection must be geometrically consistent all facets must be simply connected , and thus no edges may intersect . if we were to randomly choose our neighbor relationships , the resulting reconnection would likely fail this test , and would thus represent a non - physical `` surface . '' to guarantee a geometrically consistent reconnection , we must search through all _ virtual _ reconnections until we find one that does not result in any self - intersecting facets . several questions immediately arise : 1 : : how can we effectively characterize a `` reconnection '' ? 2 : : how many virtual reconnections are there to search ? 3 : : how can we efficiently list all these choices ? 4 : : can we be sure a good reconnection exists ? 5 : : is this reconnection unique ? for our method to be effective , all but the last of these questions must be answered satisfactorily .
the detailed answers to ( 1 - 3 ) are found in the appendix , but we will summarize them here . the edges and junctions created during an @xmath23 reconnection may be effectively characterized as a binary tree with @xmath24 nodes .
the number of @xmath25noded binary trees is given by the _
catalan number _ @xmath26 .
finally , these trees may be efficiently listed using a greedy recursive algorithm in @xmath27 time . for the fourth question regarding existence ,
we argue heuristically that a facet reaching zero area proves the existence of its own reconnection , since a surface with a zero - area facet is functionally the same as the surface with that facet removed .
we then assume the existence of that same reconnection for some window of time before the facet reaches zero .
a fuller proof would appeal to manifold theory .
finally , the fifth question regarding uniqueness is addressed in section [ sec : non - uniqueness ] .
having established these facts , we have a robust method for reconnecting an arbitrary far field of facets . before considering some special cases of this method ,
let us summarize the general process so far : whenever facets smaller than a threshold area are detected , we : a : : remove them , leaving a hole in the mesh .
b : : list all virtual reconnections ( vr s ) as n - node binary trees .
c : : use associated neighbor relationships to find edge locations .
d : : test each vr until one with no intersecting edges is found .
we note that this approach represents a comprehensive reconnection method for _ any _ cellular network problem .
though it is necessary for the faceted surface problem , it may be useful in any situation where a verifiably optimal reconnection is sought .
it is possible for _ groups _ of facets to shrink together , in such a way that they can not be removed sequentially . for an example , consider the configurations in figure [ fig : volume - removal ] .
in such a case , it is necessary to identify and collect a contiguous group of small facets for simultaneous deletion we call this a * near field*. any facet neighboring the near field is assigned to the far field , which may be reconnected as described previously after the near field is deleted .
[ fig : volume - removal ] to gather the near - field facets , we maintain a second , more liberal threshold . whenever a face shrinks below the first threshold , as described above
, its neighbors are recursively examined to collect those smaller than the second threshold .
this method is rather simplistic , and , in cases of oddly - shaped pyramids , may not return the entire near field .
this , in turn , will result in an incorrect far field , which will most likely be non - reconnectable .
however , a group of facets vanishing together eventually all head to zero area , and for some window of time before they would physically vanish , all are small enough to be detected in this way .
thus , we allow the code to `` skip over '' small facet combinations that it can not remove successfully , and try again during a future timestep .
it is also possible , on high - symmetry crystal surfaces , that the small facet or group of facets forms a `` step '' between two much larger facets of identical orientation , but different height .
figure [ fig : step - removal ] illustrates this situation , in which the near field is bounded by exactly four facets , two of which have identical orientations . in such a case , the final fate of the surface is that the small facets vanish _ as the large facets join together_. the method described above contains no provision for joining far - field facets during reconnection , and so there is no way to reconnect the far field produced in this case .
[ fig : step - removal ] having identified a near field as forming a step , one solution is to delete the small facets , then move the two large faces to the same height and join them .
this results in two pairs of unconnected edges , which are each deleted and replaced with an appropriate single edge .
since facet groups forming steps are , in fact , bordered by four facets generically , a separate ` repairman ` class could be written to handle this case .
however , the small adjustment to the positions of the large facets can lead to subtle problems , as will be seen in section [ sec : discretization and topology ] .
therefore , a more robust if less elegant approach is to simply add one of the large parallel facets to the ( step - forming ) near field ; a good choice is the one with fewer edges .
since the far field surrounding this modified near field requires no joins , it can be repaired using the ffr method .
again , failures are possible as described in the above section , but resolution is always possible near enough to the time the event would physically occur .
we have now discussed the general kinematics of a @xmath19 , and surveyed all topological events which may occur as the surface evolves . before our treatment is complete , however , we must consider with care the application of a time - stepping scheme .
the accurate performance of topological events under such a scheme is problematic because , while events on a continuously evolving surface happen at precise times ( @xmath28 at @xmath29 ) , any time - stepping method invariably skips over these times .
this has three consequences , concerning detection , consistency , and accuracy . after discussing them briefly
, we will present three possible timestepping methods which illustrate them in more detail .
* detection . * because timestepping will always skip over moments of topological change , we must abandon hope of simply finding topological events ready to perform . instead
, we must either look ahead before each timestep and anticipate when events will occur ( a _ predictive _ method ) ; or step before looking , and then by examining the network infer where events should have occurred ( a _ corrective _ method ) .
class a events can be easily be detected either way , while class b events are easier to correct , and class c are easier to predict .
* consistency .
* once the occurrence of an event has been detected by either means , it must be performed in a way that preserves geometric consistency
i.e. , the network always corresponds to a physical surface @xmath30 .
for example , two joining facets can only be mechanically fused if they exhibit the same local height .
if , in addition to the occurrence of an event , a detection scheme can determine the exact time at which it occurred , then one strategy is to move the network to the precise event time , at which resolution is trivial .
however , one may wish to attempt resolutions at other times , and the geometric consequences of doing so must be weighed .
* accuracy * finally , we must consider the possibility of error that is produced during topological change .
this error is most easily understood if we view the evolving surface in its abstract form as a highly nonlinear system of ode s .
the ( usually configurational ) evolution function is moderated by the topological state ; thus , topological events can represent sudden , qualitative changes in the evolution function .
a naive time - stepping scheme which steps over these without appropriate measures will produce large localized errors at moments of topological change .
assume that , at all times , we accurately predict the time and location of the next topological event .
then , a straightforward timestepping strategy which avoids consistency and accuracy concerns is to continually calculate the time of the next topological event ( accurate to the the order of the time - stepping method ) , and then step from event to event . under this approach ,
time is divided into slices with constant equations of motion , guaranteeing that that the system always evolves under the correct equations , and accurately representing the continuously evolving surface .
in addition , high - order single - step methods such as runga - kutta methods may be used to obtain high accuracy . though neatly eliminating consistency and accuracy concerns
, this method has a serious disadvantage .
the frequency of topological events scales with the system size , and since we can never step farther than the next event , we effectively make the timestep dependent on system size @xmath31 .
since moving the system through a single timestep is itself an @xmath32 operation , then advancing the system through any @xmath33 period of time takes @xmath34 time . while acceptable for the detailed study of a small surface , it is obviously undesirable for the statistical study of large surfaces
this is chiefly because , consistency concerns aside , it makes little sense to halt the entire surface at every single topological event , when each of these involves only a few facets .
thus , our next method has as its chief objective the use of timesteps which are independent of system size .
a second strategy is to take fixed timesteps , use a corrective method of topological detection , and attempt to perform topological corrections late .
since timestep is independent of system size , many events will now occur per timestep , the size of which is chosen to produce a fixed small _ percentage _ of facets undergoing topological change each step .
while this approach theoretically eliminates the @xmath34 contribution to running time , it introduces hurdles to event detection , as well as geometrically consistent and accurate resolution . * detection .
* we just stated that , in this corrective detection scheme , more than one event occurs per timestep .
whether or not this is a problem depends on the * domain of influence * of each event , defined to be the set of network elements that event affects .
if these sets contain no common elements , then the associated events occur too far apart in space to affect each other they are independent .
consequently , a detection routine can hand them in arbitrary order to the repair routines , there to be confidently performed in isolation .
however , occasionally two or more domains of influence overlap . in this situation , called a * discrete compound event * , the associated topological events are no longer independent , and a detection routine can no longer ensure _ a priori _ their correct , consistent resolution when handed off .
even worse , the very signatures used to identify separate events may be obscured in the resulting `` tangle , '' such that the routine does not even recognize what has happened .
given the variety of event signatures described in section [ sec : topological events ] , and the many combinations in which they might occur , creating a complete list of all dce s would be prohibitive if not impossible .
instead , we reason that , on a random surface , no two events will ever occur at exactly the same moment ( it is possible to artificially construct faceted surfaces such two or more events must occur simultaneously
we do not consider this case ) .
thus , if we simply refine our timestep when necessary , formerly overlapping events can be sorted out , and detected in sequence . a robust strategy for handling compound events
is thus to ( a ) _ retrace _ the problematic timestep , ( b ) _ refine _ it into smaller slices , and ( c ) _ repeat _ steps ( a ) and ( b ) recursively , until only single events are detected
. * consistency . *
since the surface is allowed to evolve unrepaired past numerous topological events per timestep , surface regions near these events will be geometrically inconsistent after the step .
to say the same thing , facets involved in the bypassed events will have incorrect neighbor relationships .
however , we have already classified all possible events , so having identified which event occurred , and which facets were involved , we know _ a priori _ what the correct neighbor relationships should be after the event .
this knowledge , along with knowledge of the position of each facet involved , allows us to reconstruct the consistent surface that should have emerged during the event .
unfortunately , not all events can be consistently corrected at a late time in this way .
in particular , facet joins and joining facet pinches involve the joining of two facets that meet each other at a single local height .
since this condition exists for only a single instant , such events can not be performed in a geometrically consistent way at any time other than the `` correct '' one . to accommodate this requirement
while preserving a topology - independent timestep , we are forced to manually adjust the height of the joining facets before the event is performed .
besides the error induced by this strategy ( discussed next ) , this need illustrates a second problem that can arise . in a * repair - induced inconsistency * , the very act of performing one event , because it is done late , triggers a second event that was not detected originally .
an example is when the just - described height adjustment required for the delayed repair of a facet join triggers , say , a neighbor - switching event .
since this newly - triggered event was not originally detected , the system is left in an inconsistent state after all repairs are made .
an ad - hoc strategy to find such rii s is infeasible for the same reason as is a complete listing of all possible discrete compound events ( indeed , an rii may produce a dce , which rules out a simple multiple - rechecking strategy ) .
thus , a similar retrace / refine / repeat strategy is required , with the added requirement that all events performed prior to detecting the rii must first be undone . * accuracy . * finally ,
as alluded above , repairing topological events _ after _ they occur can introduce large isolated errors .
this can be due to the `` fudging '' required for the delayed repair of facet joins and swaps , but more generally is caused by facets involved in ( uncorrected ) topological events having been evolved under the wrong equations of motion for part of the relevant timestep .
consider an event @xmath35 , with domain of influence @xmath36 .
since topological events likely correspond to a change in the surface s evolution equation , the facets in @xmath36 are moved using the wrong equations for the time interval @xmath37 $ ] .
since the equations guiding @xmath36 are wrong by as much as @xmath33 for a time of order @xmath38 , facets in @xmath36 may accumulate @xmath39 location errors during the timestep in which the event occurs .
since the quantity of topological events does not depend on @xmath40 , the method retains first order accuracy globally .
however , this error introduces a barrier to achieving higher - order accuracy later on . the previous method , alas , contains one subtle problem that keeps it from being a true @xmath32 method .
this problem is that the frequency of dces and riis , though small , still scales with the system size , and these necessitate timestep refinement .
so although the late method does not have to explicitly step according to the @xmath41 time between topological events , yet to accurately detect and resolve those events it is still implicitly driven by the refinement strategy to step along a time associated with dces and riis .
while this characteristic time is longer than that between individual topological events , and does not greatly slow the simulation of tens of thousands of facets , it still results in a method that is formally @xmath34 , which becomes prohibitive when considering systems of millions of facets .
thus , we now sketch a third method , not yet implemented , which eliminates this effect all together .
in addition , the method allows us to perform topological events in a way which confers all the accuracy benefits of the first , predictive method .
we first re - state that , on any given timestep , _ most _ facets are not involved in any topological changes .
while it was therefore obviously wasteful to move the entire surface from event to event in the first method , it is also conceptually wasteful to perform a global retrace / refine / repeat step to dces and riis in the second method . instead , after every timestep , we should identify for each dce / rii the * topological subdomain * containing all facets involved in the event .
the few facets within these subdomains would be retraced / refined / repeated as required , while the rest of the ( unaffected ) facets would be left undisturbed in their post - timestep state .
since operating on a given , constant number of facets takes @xmath33 time , and since the number of events per timestep scales only like @xmath32 , we see that a single timestep and all associated corrections including dces and riis can now be performed in @xmath32 time , with a final state that is guaranteed to be consistent .
this produces a true @xmath32 method .
in addition , this `` localized replay '' strategy has an accuracy benefit .
regular , recognized topological events also have easily identifiable topological subdomains .
if the facets within these domains are retraced , then the predictive detection mechanism of the first method can be applied within the domain to eliminate consistency and accuracy problems associated with late removal .
one difficulty remains , however .
facets involved in topological events may , under configurational facet - velocity laws , exhibit abrupt changes in velocity a result of the event . during the remaining segment of timestep
, these facets may `` break out '' of the subdomain initially created to contain them , and begin interacting with facets outside of it .
thus , we would need a mechanism to detect this , and start over with a larger subdomain if it occurs .
finally , if subdomains can change size , then there is the possibility that two nearby subdomains will come to overlap as the algorithm progresses .
therefore , we must include the ability to merge them if necessary , start over with the new , larger sub - domain , and repeat adaptively until everything can be sorted .
this adaptivity ensures the robustness of the method , as highlighted by the method s formal name of * adaptive localized replay*. the reader may note that the pattern of adaptive repetition is similar to that used to resolve dces and riis above , and worry that another , even smaller @xmath34 effect lurks in the shadows .
however , in both of the previous methods , such effects were due to the _ global _ response to a _ local _ problem .
since this latter method is designed to be localized , there is no longer any mechanism to generate such effects .
we demonstrate our method using the sample dynamics derived in chapter 1 , associated with the directional solidification of a strongly anisotropic dilute binary alloy .
when a sample is solidified at a pulling velocity which is greater than some critical value , solute gradients caused by solute rejection at the interface create a solute gradient which opposes and overcomes the thermal gradient , resulting in a negative effective thermal gradient . in this environment , facets move away from the freezing isotherm at a rate proportional to their mean distance from the isotherm , as given by the dynamics @xmath42 in figures [ fig : threefold ] , [ fig : fourfold ] , and [ fig : sixfold ] , this dynamics is applied to surfaces with common three- , four- , and six - fold symmetries to illustrate the flexibility of our topology - handling approach .
a series of snapshots from the coarsening surface are presented , in which surface configurations near to many of the topological events described above may be observed .
( however , neither the irregular neighbor switch , irregular facet pierce , nor facet pinch occur because no three facet normals are coplanar in these symmetries ; indeed , these events are not expected to occur on most physical surfaces , and were included for theoretical completeness . )
[ fig : threefold ] about half of computational time is spent looking for topological changes , which is significant but not prohibitive . with appropriate timestep choice , using even the still - inefficient timestepping method 2 above , a surface of @xmath43 facets
may be simulated to a @xmath44 percent coarsened state in about an hour on currently available workstations . with the implementation of method 3 above
, this time should be cut in half , and since method 3 is truly @xmath32 , a single million - facet simulation should take about a day .
looking further ahead , since facet velocity calculations and topology checks require only local information , the method should be easily parallelizable , making possible even larger speed gains .
we have presented a complete method for the simulation of fully - faceted interfaces of a single bulk crystal , with arbitary symmetry , where an effective facet velocity law is known .
the surface , which is reminiscent of two - dimensional cellular networks , is encoded numerically in a geometric three - component structure consisting of facets , edges , and junctions , and the neighbor relationships between them .
consistent surface evolution specified by the facet velocity law is accomplished via a simple relationship between facet motion and junction motion .
although requiring the explicit handling of topological events , the method is efficient , using the natural structure of the surface , and accessible , allowing easy extraction of geometrical data .
this combination makes it ideal for the statistical study of extremely large surfaces necessary for the investigation of dynamic scaling phenomena .
a comprehensive listing of all topological events has been presented .
these allow single - crystal surface with arbitrary symmetry ( or no symmetry at all ) to be simulated .
events are classified into three categories , representing three ways that surface elements can become geometrically inconsistent .
these are : edges which approach zero length , facets which become constricted , and facets which approach zero area .
resolution strategies for the former two classes can be determined a priori , while repairing surface `` holes '' left by vanishing facets requires a novel far - field reconnection algorithm , which iteratively searches through all virtual reconnections to find one which produces a consistent surface .
finally , intrinsic non - uniqueness of several events is discussed ; since ours is a purely kinematic method , decisions regarding resolution of these events must be made ahead of time through consideration of the dynamics or other physics .
in addition , a detailed discussion of the issues associated with a discrete time - stepping scheme has been presented .
the core issue is that topological events , which occur at discrete times throughout surface evolution , invariably fall between timesteps , with consequences for the detection of events , as well as their geometrically consistent and numerically accurate resolution . since topological change corresponds
( under configurational facet velocity laws at least ) to qualitative changes in the local evolution function , some way to reach these in - between times must be introduced , while recognizing that only a few facets are involved in topological change during each timestep .
a comparison of three approaches showed that the optimal solution is one of localized adaptive replay , where large timesteps are taken to improve speed , but local surface subdomains associated with topological change are reverted , and then replayed in a way that re - visits events with the necessary precision as necessary . while further work remains to implement this approach , the method as presented is capable of comparing million - facet datasets via averaged runs .
our method of removing facets and facet groups requires patching a `` hole '' in the network left by the deleted facets .
this requires selecting a geometrically consistent reconnection from a list of potential , or virtual reconnections . as outlined in the text , this involves searching through a complete list of virtual reconnections and testing each for geometric consistency . in this appendix
, we address in more detail questions ( 1 - 3 ) posed in section [ sec : vanishing - facets ] regarding the details of this method . for convenience , we repeat them here : 1 : : how can we effectively characterize a `` reconnection '' ? 2 : : how many potential reconnections are there to search ? 3 : : how can we efficiently list all potential choices ? we show here that an effective means of answer these questions is to think of reconnections as extended binary trees .
this characterization enables us to easily count potential reconnections , distinguish them through naming , and suggests an algorithm for efficiently listing them for testing .
an exhaustive illustration of the process is given for the case of an @xmath22 far field in figure [ fig : binary tree listing ] .
it will be useful to refer to that diagram during the following discussion .
[ fig : binary tree listing ] patching network holes always involves finding unknown neighbor relations between a given number of adjacent facets that is , no facets are ever created , only edges and junctions .
these are always connected into a single graph .
in fact , the edges and junctions created during reconnection ( the `` reconnection set '' ) form a binary tree . in figure
[ fig : binary tree listing ] , the trees associated with each possible reconnection are shown in thick blue lines . the far - field edges touching
the reconnection set shown in gray represent the _ completion _ of this tree .
that is , they take the interior tree , and add leaves to it so that every node of the interior tree is a triple - node . each virtual reconnection corresponds to a unique interior tree and completed tree in this manner .
the counting of binary trees is , fortunately , a solved problem of graph theory . given @xmath8 nodes ,
they may be arranged in @xmath45 distinct binary trees , where @xmath45 is the @xmath46 _ catalan number _ ; @xmath47 now , re - connecting an @xmath23 far field requires the creation of @xmath48 edges and @xmath24 nodes ; this may be visually confirmed for the case @xmath49 in figure [ fig : binary tree listing ] .
this creates an @xmath24 noded binary tree , and so an @xmath23 far field has @xmath50 virtual reconnections to search .
if take the completed binary tree and arbitrarily select a _
root node _ , then from each virtual interior tree we can generate a unique sequence of letters which identify it and encode its construction . this can be formed in one of two ways .
the first way is to specify the tree by a set of recursive function calls .
each branch point has left and right branches , each of which may terminate in either a leaf , or another branch point .
the middle column of figure [ fig : binary tree listing ] gives such a function for each tree shown in the left column .
the second way is to walk around the tree in a counter - clockwise manner , recording each branch point or leaf as it is encountered .
either method produces a series of letters that uniquely identify the tree .
since the beginning lb and terminal ll are guaranteed , we may use only an abbreviated version consisting of @xmath48 of each letter .
the problem is now reduced to generating all possible letter combinations .
we can recursively build these combinations letter by letter using a greedy algorithm which chooses b over l if possible .
this approach is subject to three restrictions which must be true of a `` legal '' word . at each step ,
( a ) @xmath51 , ( b ) @xmath52 , ( c ) @xmath53 .
the function we use is sufficiently short that we simply reproduce it here : .... list_trees(n ) { rec_list_trees(n , 0 , 0 , '' ) ; } rec_list_trees(n , leaves , branches , word ) { max = n-3 ; if ( branches < max ) rec_list_trees(n , leaves , branches+1 , word+'b ' ) ; if ( leaves < max and leaves < = branches ) rec_list_trees(n , leaves+1 , branches , word+'l ' ) ; if ( leaves = = max and branches = = max ) print(word ) ; } .... thus , using this algorithm to generate a complete list of reconnection labels , we perform the reconnection associated with each one , and test it for geometric consistency .
this allows us to efficiently find the correct reconnection .
we have already mentioned that certain topological events are ambiguous in their resolution . here
we review the instances of non - uniqueness , discuss their cause and implications , and suggest a method of treating them in a numerical scheme .
we begin our list with an important ambiguity not discussed in the main text .
as noted by thijssen @xcite , the lowly neighbor switch can be non - unique if the four facets involved have a `` saddle '' configuration where the edges neighboring the vanishing edge ( the emanating edges ) form an alternating sequence of two valleys and two ridges .
this is illustrated in figure [ fig : non - uniqueness]a .
there , from either starting position on the left , two topological resolutions are possible .
one involves changing neighbor - relations , while the other does not .
this latter resolution , which we term a `` neutral pass , '' represents a vanishing edge event which needs no resolution .
however , it still results in a ( prohibited ) flipped edge , so if it is to actually occur , a bookkeeping operation must take place to correct this . moving on
, we recall the non - unique `` gap opener '' flavor of irregular neighbor switch , which is not reproduced here . however , as a vanishing edge event , it may also come in the saddle variety , and admits a neutral - pass resolution option which is displayed in figure [ fig : non - uniqueness]b .
finally , the irregular facet pierce event is also not reproduced .
none of these cases of ambiguity of resolution can be resolved at the kinematic level . since the particular evolution pathway leading to such ambiguities is the product of a long chain of mathematical reasoning , we therefore look backward in this chain to resolve them .
the kinematics are most directly provided by the dynamics , which is a good first place to start .
one option is to see which resolution is more strongly self - reinforcing ; if one choice under the given dynamics immediately works to reverse itself , for example , the other option should probably be chosen . moving farther back the chain of reasoning , one may consider the argument used to derive the given dynamics .
for example , the dynamics ( [ eqn : solidification - dynamics ] ) comes from the constant reduction of an undercooling energy by moving away from the maximally - expensive @xmath6 isotherm .
therefore , an energy - informed choice is to choose the resolution which minimizes this energy , i.e. , maximizes the total integrated distance from @xmath6 . taking another step back , perhaps such arguments come , as this one does , from a partial differential equation describing surface evolution .
such equations can be simulated directly , and the resulting resolution choices studied .
ultimately , the original physical model may have to be considered , perhaps at the atomic level .
for each dynamics we wish to study , one must apply this chain of reasoning to find the correct ambiguity - resolution strategy .
ideally , considerations at all levels should produce identical results .
the non - uniqueness of particular topological events reflects a deeper , underlying problem . for each non - unique event , facet heights of competing resolutions
are identical ; only neighbor relationships between these facets differ .
thus , non - uniqueness in topological events indicates the presence of , and indeed is caused by , multiple possible ffr - style reconnections of a given group of far - field facets .
what are the implications for the ffr algorithm of early facet removal ?
while it may seem helpful at first to list prohibited resolution configurations and feed these to the ffr algorithm , this approach grows increasingly brittle as the number of prohibited configurations grows .
instead , a more robust approach is to cause an ffr yielding multiple valid results to fail .
having discovered resolution strategies for each individual event as described above , we simply let these rules apply until the far field is small enough to be unique ( similar to @xcite ) .
a theory predicting when far fields will have multiple resolutions would aid in this process .
finally , we note that the benefits of phase - field and level - set methods , which automatically capture topological change , are only clear if topological changes are unique .
these methods are ultimately only kinematic , and offer no clear criteria for resolving kinematic non - uniqueness .
thus , a dynamic selection criteria would seem to be required no matter the method used , and indeed , the implicit handling of topology offered by these methods may actually hinder the dynamic selection of kinematically ambiguous events . | we fully generalize a previously - developed computational geometry tool @xcite to perform large - scale simulations of arbitrary two - dimensional faceted surfaces @xmath0 .
our method uses a three - component facet / edge / junction storage model , which by naturally mirroring the intrinsic surface structure allows both rapid simulation and easy extraction of geometrical statistics .
the bulk of this paper is a comprehensive treatment of topological events , which are detected and performed explicitly .
in addition , we also give a careful analysis of the subtle pitfalls associated with time - stepping schemes for systems with topological changes .
the method is demonstrated using a simple facet dynamics on surfaces with three different symmetries .
appendices detail the reconnection of `` holes '' left by facet removal and a strategy for dealing with the inherent kinematic non - uniqueness displayed by several topological events . |
the hard state of the black hole ( bh ) x - ray binaries is characterized by a power - law x - ray spectrum with photon index @xmath0 , and relatively strong radio emission ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
during the hard states , there exists an interesting nonlinear correlation between the radio and x - ray luminosity , i.e. @xmath1 with @xmath2 ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) .
although this relation is established from a few bh binaries like gx 339 - 4 and v404 cyg , it even holds for active galactic nuclei @xcite .
moreover , a similar correlation between optical / infrared and x - ray flux has been found @xcite .
it seems that this correlation is universal and can be taken as `` standard '' .
however , radio observations of some new bh transients in more recent outbursts in the following years reveal that their behavior deviates significantly from the aforementioned correlation ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
based on the first precise measurement of an outlier h1743 - 322 , ( * ? ? ?
* * c11 * ) discovered a new tight correlation in its bright hard state , with the slope being @xmath3 . at low luminosity h1743 - 322
rejoins the standard correlation with @xmath4 . at the intermediate luminosity
, there exists a transition region .
the reasons for the existence of the two correlation and their transition remain to be clarified . to date three classes of models
have been proposed to explain the hard states : ( i ) the hot accretion flow , which includes adaf ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and lhaf ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , ( ii ) the disk corona ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and ( iii ) the jets @xcite .
all these models can explain the `` standard '' correlation between radio and x - ray emission . in the models of hot accretion flow and corona ,
x - rays come from the accretion flow while radio emission from the jet .
since the jet originates from the accretion flow , it is natural to expect the radio / x - ray correlation ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in the jet model , both radio and x - ray emission
come from the same electrons in the jet , and thus strong correlation is expected ( e.g. * ? ? ?
* ; * ? ? ?
as for the steep correlation with index of @xmath5 , @xcite proposed * that the efficient accretion flow like lhaf or disk corona can explain it*. to understand the physics behind the correlation , more sample of sources are needed .
very recently , @xcite ( * p11 * ) reported the observation of a newly discovered x - ray transient , igr j17177 - 3656 .
they analyzed the quasi - simultaneous data by , integral and atca , and argued that the source is probably a low - mass bh x - ray binary in the hard state , although the neutron star nature can not be excluded yet . by assuming the distance to be 8 kpc
, they found that the source follows the behavior of the bhc h1743 - 322 at intermediate x - ray luminosity . in this paper , by using an accretion - jet model , we obtain the source distance by modeling the observed multi - band spectrum , and re - investigate its position in the radio / x - ray correlation diagram .
it is found that igr j17177 - 3656 probably follows the steep radio / x - ray correlation of h1743 - 322 and not the transition one .
after the discovery by integral , igr j17177 - 3656 became a focus of radio , nir , and x - ray telescopes ( * p11 * and references therein ) . conducted an observation of 20 ks at march 22nd , 2011 .
the 2 - 8kev observation shows a power - law spectrum , @xmath6 , with unabsorbed 3 - 9 kev flux being @xmath7 . a simultaneous integral observation in 20 - 200 kev shows a power law with @xmath8 and flux being @xmath9 .
this result is similar to the observation on march 15th , 2011 .
the radio observation on the same day is conducted by australia telescope compact array ( acta ) at two frequency bands , which are centered at 5.5 ghz and 9 ghz .
the corresponding flux density are @xmath10 mjy and @xmath11 mjy , respectively .
figure 1 shows the results of our modeling and the data points of * p11*. the accretion - jet model is briefly described here .
the reader can refer to @xcite for additional details .
there are four components in the model , the inner hot accretion flow ( adaf / lhaf ) , the outer standard accretion disk , the outflow and the jet . in the adaf regime , the accretion rate @xmath12 , where @xmath13 and @xmath14 is the viscous parameter , and the viscously dissipated energy can be advected inwards with the accreting gas . while in the lhaf regime , @xmath15 , and the radiative
cooling is a little more efficient than the viscous heating . although the temperature of lhaf is lower than that of adaf , it is also high because during the falling the compression work plus viscous heating together balance the radiative cooling .
the jet is shown in the flat / inverted radio spectra and the direct images ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the existence of outflow in adaf / lhaf is from theoretical work or numerical simulations ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the outer standard accretion disk is needed when modeling the uv spectra of some well - known bhs ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the radiation processes in the hot accretion flow include synchrotron , bremsstrahlung , and inverse compton scattering of soft photons .
the emission from the outer cool disk is modeled as a multi - color blackbody spectrum . and in the jet ,
synchrotron emission and self - absorption of the relativistic electrons are considered .
the accretion disk is described by a set of parameters .
the inner hot accretion flow truncates the outer standard accretion disk at the transition radius @xmath16 .
since the physics of the transition is highly uncertain , the range of the value of @xmath16 is large , from @xmath17 to @xmath18 , where @xmath19 ( e.g. * ? ? ? * ) . due to the lack of uv spectra
, we set it a normal value , @xmath20 .
the influence of outflow on the accretion is mimicked with the parameter @xmath21 by @xmath22 , where @xmath23 is the accretion rate at @xmath16 .
the value of @xmath21 describes the strength of the outflow and varies in the range @xmath24 . for the case
@xmath25 , there is no outflow , while for the case @xmath26 , 90% of the accreted gas will flow away and only 10% finally fall into the bh given @xmath20 , which roughly agrees with numerical simulations ( e.g. * ? ? ?
a recent numerical simulation shows that the fraction of outflow is almost the same for adaf and lhaf @xcite .
so in this paper , the value of @xmath21 is assumed to be the same for different accretion rates , instead of a function related to accretion rate as assumed in @xcite . to determine the dynamics of the hot accretion flow , we need three more parameters , which are associated with the mhd turbulence .
these parameters are the viscous parameter @xmath14 , the fraction of gas pressure to total pressure @xmath27 , the fraction of the viscous dissipation directly heating electrons @xmath28 .
the value of @xmath27 is fixed at a typical value of 0.9 , the value of @xmath14 changes in the range @xmath29 , and @xmath30 @xcite .
the jet is modeled phenomenologically with quite a few parameters .
we assume a fraction of the accretion flow , @xmath31 , is transferred into the vertical direction , and the jet has a conical geometry with a small half - opening angle @xmath32 . for simplicity
we assume each ejection moves at a constant velocity along the jet .
different ejections may have different velocities , however , their averaged bulk lorentz factor , @xmath33 , is assumed to be constant .
when different ejections collide , internal shocks occur , which accelerate a fraction of the electrons into power - law energy distribution with index @xmath34 .
the energy density of accelerated electrons and amplified magnetic field are specified with @xmath35 and @xmath36 . @xcite
have shown that the contribution of the jets to the x - ray emission in bh sources is negligible when the luminosity is @xmath37 . therefore in the transient source
igr j17177 - 3656 , the jet contributes only to the radio emission .
it should be noted that the main results of the paper are from the modeling of the x - ray observation , and we model the radio spectra just to show the ability of the model to explain the multi - band spectra .
so we fix most of the jet parameters .
considering that there are only two radio data points , we leave @xmath31 and @xmath34 free and the other parameters fixed at the value that has been used for other bh binaries , i.e. @xmath38 , @xmath39 , @xmath40 and @xmath41 @xcite .
the bh mass is set it to be of normal value @xmath42 .
the binary inclination angle @xmath43 is estimated to be high @xmath44 @xcite .
detailed calculations show the influence of @xmath43 is not significant , so we set it to be a middle value @xmath45 .
there are 7 free parameters and quantities in the model , @xmath46 , @xmath21 , @xmath14 , @xmath28 , @xmath31 , @xmath34 , and the distance @xmath47 .
the modeling steps are as follows .
first , we try with different value of @xmath46 , @xmath21 , @xmath14 , and @xmath28 to model the slopes of the x - ray spectra , and obtain the possible ranges of these parameters .
second , the range of @xmath47 is obtained by comparing the observed x - ray flux and theoretically calculated disk luminosity .
at last , @xmath31 and @xmath34 are adjusted to check whether the radio spectrum can be modeled simultaneously . , @xmath48 , and @xmath49 kpc .
for case i , the _ short dash - long dashed _ , _ long dashed _ , and _ dot - dashed _ lines correspond to radiation from the jet , the outer thin disk , and the inner hot accretion flow , respectively .
thick dashed _ line is the total spectrum and corresponds to case ii with @xmath50 , @xmath51 , and @xmath52 kpc .
, width=245 ] [ fsp ] [ fadv ] [ frx ] it is useful to investigate the effects of parameters on the output spectra in order to find the possible ranges of the parameters .
note that the x - ray spectral index decreases with the increasing @xmath53 compton parameter , which is the product of electron temperature and optical depth .
given other parameters fixed , the surface density of the accretion flow of higher @xmath46 is higher , therefore the optical depth is higher and the spectral index is smaller . the value of @xmath14 determines the maximum accretion rates of the adaf and lhaf , and therefore the maximum optical depth and hardest spectra . for larger @xmath28 , more viscously dissipated energy is transferred to thermal electrons .
consequently , the electron temperature is a little higher and the x - ray spectrum is harder .
when @xmath21 increases , the final x - ray spectrum is softer even if the electron temperature increases due to less cooling .
this is because , as shown by detailed calculations , the decrease of the surface density and optical depth are more significant than the increase of the electron temperature .
it should be mentioned that if the x - ray luminosity rather than @xmath46 is given , the x - ray spectral hardness increases with @xmath21 .
this is because more photons come from the region of larger radius , where the optical depth is higher @xcite .
our detailed calculations show that in order to model the observed spectrum , the parameters are limited in the following ranges , @xmath54 , @xmath55 , @xmath56 , and @xmath57 .
when @xmath58 , @xmath59 or @xmath60 , the output photon index are greater than 1.54 .
when @xmath61 , the obtained luminosity is as large as @xmath62 , and the obtained distance is greater than 30 kpc , which seems unreasonable since the diameter of the galaxy is about 30kpc . when @xmath63 , the highest photon index obtained with @xmath64 is 1.65 , which is not hard enough to explain the observed x - ray spectra
the modeling is not unique . as an example
two cases of the modeled spectra are shown in figure 1 . in case i ,
the set of parameters are @xmath65 , @xmath66 , @xmath67 , @xmath48 , @xmath68 , @xmath69 , and @xmath49 kpc . while in case
ii , @xmath50 , @xmath51 , and @xmath52 kpc , with others being the same as case i. the thick solid and dashed lines correspond to case i and ii , respectively . from this figure it can be seen that the radio and x - ray emission are dominated by the jet and the hot accretion flow , respectively .
the infrared and optical spectra are the sum of the emissions from the jet , hot accretion flow and the truncated thin disk .
it can also be seen that the model parameters can be further constrained by optical / uv observation .
the accretion rate is in the range @xmath70 , therefore the accretion flow is in the regime of lhaf .
this regime is characterized by the negative advection factor , which is defined as @xmath71 , with @xmath72 , @xmath73 , @xmath74 being the viscous heating rate , advection cooling rate , and the cooling rate of ions by ion - electron coupling , respectively .
it is easy to find in figure 2 that @xmath75 except in the innermost region @xmath76 .
the key reason why lhaf can model the spectra but adaf can not is because of the observed hard x - ray spectra with * photon index being @xmath77 * , which demands optically thick hot gas that is difficult for adaf to form .
the distance is 21 - 25 kpc , rather large compared to most of the observed bh x - ray binaries . however , this is still smaller than that of the farthest dynamically confirmed bh gs 1354 - 64 , which is estimated to lie at 25 - 60 kpc @xcite .
therefore it is still well acceptable .
considering the diameter of the galaxy is 20 - 30 kpc , we think that the source may be located on the side of the galaxy opposite to the sun . moreover , if the mass of the bh is adjustable , the distance could be as close as @xmath7813 - 15 kpc for @xmath79 .
from the spectral modeling results , the distance of the source should be 13 - 30 kpc by considering all possible bh mass and the radius of the galaxy .
the possible position of igr j17177 - 3656 is shown in figure 3 .
it can be seen that except for low mass bh the possible location of the source is close to the branch with @xmath80 .
this indicates that igr j17177 - 3656 is a new source like h1743 - 322 , which follows the steep radio / x - ray luminosity correlation .
moreover , our results show that lhaf can not only explain the relation between the radio and x - ray luminosity , but also model the multi - band spectrum .
this supports the result of * c11 * that lhaf is a candidate to explain the correlation index with @xmath80 .
the observation of h1743 - 322 shows that the correlation index is @xmath81 at low luminosity and @xmath5 at high luminosity .
the change of the correlation index can be naturally explained by virtue of the accretion - jet model .
when @xmath82 , the hot accretion flow is in the regime of adaf , the luminosity is low and the index is @xmath81 @xcite .
when @xmath46 increases within a few times , the hot accretion flow is in the regime of lhaf , the luminosity is higher and the index is @xmath5 .
lhaf is more luminous than adaf , therefore it is used to explain luminous hard state with @xmath83 .
a luminous hard state occurs during the rising part of an outburst , corresponding to the upper - right corner in the hardness - intensity diagrams ( hids ) . considering the above results it is natural to expect prevalent steep radio / x - ray correlation in the luminous hard state .
some observations seem to support this result ( yuan , private communication ) . it should be noted that in addition to the change of correlation index at high luminosity , the accretion - jet model predicts another change of the correlation index at low luminosity .
@xcite predicted that when @xmath84 , the x - ray emission should be dominated by the jet rather than the accretion flow .
consequently the correlation index changes from b@xmath81 to @xmath85 .
this prediction was well confirmed by later direct observations of agns @xcite and was compatible with theoretical works based on observational data @xcite .
@xcite , however , found that the radio and x - ray data of the quiescent state of a 0620 - 00 are consistent with an extrapolation of the v 404 cyg and gx 339 - 4 correlations .
since the exact track of a 0620 - 00 in outburst is still unknown @xcite , it is doubtful to say this result conflicts with the prediction of @xcite .
@xcite reported the radio and x - ray observations to the quiescent state of v404 cyg and found that the correlation of this source remain the original @xmath86 one .
this result may also not contradict with the 1.23 correlation , considering the quiescent luminosity of this source is very close to the critical luminosity of @xmath87 @xcite .
it should be noted that the results of this paper depend on the observed slope of the x - ray spectrum . in the above calculation
we take the upper limit of the x - ray photon index to be 1.52 , which is of 68% confidence .
if the upper limit is extended to 1.6 , which is of 90 - 99% confidence as shown in figure 4 of * p11 * , the above results are less robust . in this case
the adaf with weak outflow , @xmath88 , can marginally model the x - ray spectrum .
moreover , the minimal possible distance of igr j17177 - 3656 reduces to 18 kpc , which means the source locates around the point where the correlation begins to deviate from the @xmath89 branch .
the physics of the transition from the outer standard accretion disk to the inner hot accretion flow remains unclear .
although different scenarios may explain the transition , such as evaporation , turbulent diffusive heat transport and secular instability in the cold disk ( * ? ? ?
* and references therein ) , the position and gas temperature can not be estimated self - consistently . in this paper , the transition radius
is simply assumed to be 100@xmath90 and the ion temperature to be 0.5 - 0.8 times of the virial temperature .
if the variation of igr j17177 - 3656 is similar to h1743 - 322 , there will be a transition region between the radiatively inefficient and efficient branches .
this means that a radio - quenching or x - ray - increasing process occurs during the transition from adaf to lhaf .
however , a physical explanation for such processes has not been well understood . in this paper ,
the outflow only affects the density of the accretion flow . however , its contribution to the loss of angular momentum and energy can produce difference by a factor of few @xcite .
moreover , with the existence of ordered magnetic field , the loss of angular momentum can be even more significant @xcite .
these effects of outflow on the critical accretion rates and output spectrum will be studied in our future work .
our modeling is based on the argument of * p11 * that igr j17177 - 3656 is regarded as a bh x - ray transient in hard state .
however , the neutron star nature of the source can not be excluded yet .
more observations on the source are needed to test these results .
in this paper we model the quasi - simultaneous multi - band spectrum of a newly discovered hard x - ray transient , igr j17177 - 3656 , using an accretion - jet model .
the model is almost the same as that in @xcite , but with a constant outflow parameter @xmath21 considering the recent numerical simulation @xcite .
the required accretion rate is probably in the regime of lhaf , and the distance is about 21 - 25 kpc for @xmath91 .
all possible locations of igr j17177 - 3656 in the radio / x - ray luminosity diagram are shown in figure 3 .
this source is close to the steep correlation branch with @xmath80 .
the point of @xcite is supported , that lhaf can explain the steep radio / x - ray correlation .
the author thanks feng yuan and ding - xiong wang for valuable comments and discussions .
the author is grateful to the anonymous referees for the constructive and very useful comments .
this work is supported by the national natural science foundation of china under grants 11143003 and 10833002 , the natural science foundation of fujian province of china ( no.2011j01023 ) , and in part by the national basic research program of china under grant 2009cb824800 . | the correlation between radio and x - ray luminosity in the hard state of black hole x - ray binaries is important for unveiling the relation between the accretion flow and the jets . in this paper , we have modeled the quasi - simultaneous multi - band observations of a recently discovered transient x - ray source , igr j17177 - 3656 .
it is found that the source is probably an outlier following the steep radio / x - ray correlation rather than an outlier in the transition region as suggested by paizis et al .
( 2011 ) .
it is also found that the multi - band spectrum can be successfully modeled by the luminous hot accretion flow ( lhaf ) but less likely by the advection dominated accretion flow ( adaf ) .
our results support the point that lhaf can explain the steep radio / x - ray correlation .
[ firstpage ] x - rays : binaries - x - rays : individual : igr j17177 - 3656 - stars : winds , outflows - black hole physics - accretion , accretion disks |
to show whether the predictions of the fractal markets hypothesis hold , we analyze six developed ( liquid ) world indices .
the usa is represented by nasdaq composite index ( nasdaq ) .
the uk is covered by ftse 100 index ( ftse ) .
continental europe is represented by stock indices of two most developed eurozone countries
german dax ( dax ) and french cac 40 ( cac ) .
asia is represented by two stock indices hang seng index ( hsi ) of hong kong and japanese nikkei 225 ( nikkei ) .
the analyzed period ranges between 3.1.2000 to 30.5.2013 , which attributes around 3400 observations for each index .
evolution of index values is illustrated in figs .
[ fig1]-[fig6 ] .
we can see that the indices have evolved similarly but there are several distinct features for each market .
nasdaq as primarily a technological index was most severely hit by the dotcom bubble at the break of the millennium and it plummeted from its highs around 5000 points down to 1000 points at the end of 2002 when a slow but steady increasing trend began and persisted up to the point of the current global financial crisis .
regardless the after - dotcom growth , the heights reached before the current crisis were only around a half of the 2000 peak .
ftse index underwent a similar path loosing around 50% during the period after the dotcom bubble . compared to nasdaq ,
the decrease was more gradual . in a similar manner ,
the dotcom deflation of the index was followed by a continuous growth which was much more profound compared to nasdaq as the 2000 heights were reached in the half of 2007 . after the lehman brothers collapse , the index plunged to the levels experienced after the dotcom bubble .
however , the index has recovered quite rapidly and as of half of 2013 , the peaks of 2000 and 2007 have been reached again . in comparison to the nasdaq index ,
ftse experienced an additional correction in the second half of 2011 .
this correction can be seen for all european indices and it is connected to the debt crisis of the eu ( the greek crisis , the piigs portugal , italy , ireland , greece , spain crisis , and the eurozone crisis to mention some other names ) .
this dip in the index value is evident also for cac and dax but not so much for the non - european nasdaq , hsi and nikkei .
cac and dax indices can be described in the same words as the ftse index which shows the interconnection between the european markets .
hsi index shows a very different picture .
even though the global downturn of the beginning of the millennium is reflected here as well , the surge of the index after 2003 up to late 2007 is much more profound .
nonetheless , the global financial crisis breakdown is evident here in the same proportions as for the other indices . even until the half of 2013 , the index has not recovered and the market has stabilized at around 65% of its 2007 peak value .
nikkei follows the dynamics of the european indices in the beginning of 2000s and of the nasdaq index during the gfc .
the market was not able to recover from the gfc downturn for practically four years and only recently , it has started an explosive growth since the end of 2012 most likely connected to the changing direction of the economic policies in japan . figs .
[ fig1]-[fig6 ] depict the wavelet power spectra for the logarithmic returns of the analyzed indices . in the charts ,
the significance of the wavelet powers is tested against the null hypothesis of a red noise ( ar(1 ) process ) @xcite , which is considered sufficient for the financial series areas of the significant power are bordered by a bold black curve . for finite series as the ones analyzed , we arrive at the limitation of wavelets at high scales where the transforms can not be efficiently calculated .
the so - called cone of influence marked with paler colors separates the time - frequency space into two the lower one where the inference is reliable and the upper one where it is not . for the interpretation of the powers , the hotter the color the higher the power ( variance ) for the particular scale and time .
for all the analyzed indices , we observe at least two common features .
firstly , most of the time , there is no scale - characteristic power which can be distinguished from a pure noise so that no investment horizon dominates .
this is true mainly for the period between 2003 and the first half of 2008 .
note that this period is connected to a very stable growth after the long - lasting decline following the dotcom bubble .
the period is thus connected to a uniformly distributed activity at separate investment horizons ( or at least , there is no dominating horizon ) , which is well in hand with fmh . the whole period between 2003 and 2008
is characterized by low total variance at all frequencies and the colors of the power spectra remain cold .
secondly , the turning point between the post - dotcom decreases and the upcoming upward trends is connected with an increased power at high frequencies .
this is true for the european and the us indices , the asian indices do not experience such behavior .
note that for all the indices , the frequent dips in the market are connected with a short - term increased power at the highest frequencies , which shows that these short - lived corrections of the markets are mainly dominated by the short - term investors .
note also that the wavelet power of nasdaq displays a rather erratic behavior during the dotcom bubble which only highlights the fact that nasdaq has been the market which was most severely hit by the dotcom bubble bursting .
turning now to the current global financial crisis , recall that the biggest movements occurred after the bankruptcy of lehman brothers on 15.9.2008 . in figs .
[ fig1]-[fig6 ] , we can see that mainly the period between september and december 2008 is strongly dominated by an increased energy at the highest frequencies .
even though the variance increases at more scales , the dominance of the very low scales is apparent ( very hot colors even bordering with black for october and november 2008 ) and is again in hand with the predictions of fmh the turbulent times are connected with the dominance of specific investment horizons so that the efficient market clearing is not possible .
the situation calmed around may 2009 where the total variance decreased markedly . apart from the most severe periods of the gfc
, we also observe a changing structure of the active participants at the european stock markets during the second half of 2012 .
note that during this period , the eurozone and the eu in general underwent probably the most severe financial and economic turmoil of its existence .
the situation is known by various names as the symptoms were many the eu debt crisis , the piigs ( for portugal , ireland , italy , greece and spain as the most troublesome countries in that times ) crisis , the greek crisis and several more .
the fact that this crisis was geographically specific is well illustrated in its effects being seen only for cac , dax and ftse .
again the dominance of the shortest investment horizons is evident .
in summary , we have analyzed whether the prediction of the fractal markets hypothesis about the dominance of specific investment horizons during the turbulent times holds .
to do so , we have utilized the continuous wavelet transform analysis and the wavelet power spectra which give crucial information about the variance distribution across scales and its evolution in time .
we have showed that the most turbulent times of the current global financial crisis can be very well characterized by the dominance of short investment horizons which is well in hand with the fractal markets hypothesis .
misbalance between short and long investment horizons thus created a tension between supply and demand , leading to decreased liquidity which has been repeatedly shown to lead to occurrence of extreme events @xcite .
fractal markets hypothesis is thus able to describe events of the global financial crisis in a more satisfying way than the mainstream efficient markets hypothesis .
daily logarithmic returns @xmath0 , where @xmath1 and @xmath2 are closing and opening prices , respectively , are analyzed between 3.1.2000 and 31.5.2013 .
the us stock markets are represented by nasdaq composite index ( ticker symbol @xmath3 ) .
the european stock markets are covered by cac 40 ( @xmath4 ) , dax ( @xmath5 ) and ftse 100 ( @xmath6 ) and the asian ones by hang seng index ( @xmath7 ) and nikkei 225 ( @xmath8 ) .
all series were obtained from http://finance.yahoo.com .
a wavelet @xmath9 is a real - valued square integrable function defined as @xmath10 with scale @xmath11 and location @xmath12 at time @xmath13 .
any time series can be reconstructed back from its wavelet transform if the admissibility condition @xmath14 holds , where @xmath15 is the fourier transform of a wavelet .
wavelet has a zero mean so that @xmath16 and is usually normalized so that @xmath17 . to obtain the continuous wavelet
transform @xmath18 , a wavelet @xmath19 is projected onto the examined series @xmath20 so that @xmath21 where @xmath22 is a complex conjugate of @xmath19 .
importantly , the continuous wavelet transform decomposes the series into frequencies and can then reconstruct the original series so that there is no information loss , and energy of the examined series is maintained as well , i.e. @xmath23 @xmath24 where @xmath25 is the wavelet power at scale @xmath26 .
there is a large number of specific wavelets @xcite out of which we choose the morlet wavelet standardly used in the economic and financial applications @xcite .
the morlet wavelet with a central frequency @xmath27 is defined as @xmath28 and for @xmath29 , it provides a good balance between the time and frequency localization @xcite .
10 url # 1`#1`urlprefix[2]#2 [ 2][]#2 .
_ _ * * , ( ) . .
_ _ * * , ( ) .
_ _ * * , ( ) .
_ _ ( , , ) . .
_ _ * * , ( ) . & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . & .
_ _ * * , ( ) . , , & .
_ * * , ( ) . & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . , , , & .
_ _ * * , ( ) . .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . .
_ _ * * , ( ) . & .
_ _ * * , ( ) . & .
_ _ * * , ( ) . , , , & _ _ * * , ( ) . & _ _
* * , ( ) . & .
_ _ * * , ( ) . .
_ _ * * , ( ) . .
_ _ * * , ( ) . .
_ _ * * , ( ) . .
_ _ * * , ( ) . & .
_ _ * * , ( ) . & _ _ ( , ) . & .
_ _ * * , ( ) . & .
_ _ * * , ( ) . , , & .
_ _ * * , ( ) .
the support from the grant agency of the czech republic ( gacr ) under projects p402/11/0948 and 402/09/0965 , grant agency of charles university ( gauk ) under project @xmath30 are gratefully acknowledged .
we also thank grinsted , moore & jevrejeva @xcite for the matlab package for continuous wavelet analysis .
l.k . solely wrote the main manuscript text , prepared the figures and reviewed the manuscript .
* competing financial interests : * the author declares no competing financial interests .
+ * data retrieval : * data were obtained from http://finance.yahoo.com on 8.6.2013 . + | we analyze whether the prediction of the fractal markets hypothesis about a dominance of specific investment horizons during turbulent times holds .
to do so , we utilize the continuous wavelet transform analysis and obtained wavelet power spectra which give the crucial information about the variance distribution across scales and its evolution in time .
we show that the most turbulent times of the global financial crisis can be very well characterized by the dominance of short investment horizons which is in hand with the assertions of the fractal markets hypothesis .
critical events and turbulences on the financial markets have always attracted attention of financial researchers as these are the biggest challenges for prevailing theories and potentially good starting points for new paradigms .
the efficient markets hypothesis @xcite has been a cornerstone of mainstream financial economics for decades even though it provides no testable predictions about extreme events and even considers them highly improbable ( or even non - existent ) @xcite . in this paper
, we try to resurrect the practically forgotten fractal markets hypothesis @xcite and test whether its predictions about dominance of specific investment horizons during the turbulent times hold in practice . to do so , we assume that an increased volatility is a reflection of an increased trading activity @xcite and utilize the continuous wavelet transform analysis to get the wavelet power spectra , which can be interpreted as a scale- and time - characteristic variance .
wavelets have been widely used across various disciplines
cardiology @xcite , genetics @xcite , geophysics @xcite , neurology @xcite , traffic modeling @xcite , economics @xcite and others . using wavelets methodology ,
we are able to comment on evolution of trading activity in time but importantly also across scales and we show that predictions of the fractal markets hypothesis are indeed observed during the latest turbulences on the financial markets .
the current global financial crisis ( 2007 onwards ) has once again showed that an ongoing growth of financial asset prices is not limitless
. commenced in the usa , the uncertainty about the future economic prospects has spread globally . during the most turbulent times ,
the major stock indices decreased considerably and lost usually around 50% of their 2007 peak values during the 1.5 year long decline . according to the mainstream efficient markets hypothesis ( emh ) , such a crisis should have never happened because investors are assumed to price the assets rationally and with respect to all available information @xcite .
moreover , investors are assumed to be homogeneous so that the arriving information has the same effect on all investors .
however , it is well known that market participants are greatly heterogenous with investment horizons ranging from a fraction of seconds and minutes ( market makers , noise - traders ) up to several years ( pension funds ) .
investors with short investment horizons focus on technical information and crowd behavior of other market participants .
on the other hand , investors with long investment horizons base their decisions on fundamental information and care little about the crowd behavior . in turn , the financial markets become highly complex systems , which are rather hard to be analyzed with standard linear methods . the fractal markets hypothesis ( fmh ) , on contrary , is based on empirically observed characteristics of the financial markets and thus considers the market to consist of heterogenous agents who react to the inflowing information with respect to their investment horizon @xcite . what is considered a negative information and thus a selling signal for an investor with short horizon might be a buying opportunity for an investor with long horizon , and vice versa .
if a sufficient number of buyers and sellers trade and are efficiently cleared in the market mechanism , a smooth functioning of this market is guaranteed .
this brings us to the crucial notion of fmh liquidity .
if investment horizons are uniformly represented in the market ( i.e. there is a sufficient number of investors at a wide range of investment horizons ) , then supply and demand for assets are met , market works efficiently and remains stable .
however , if an investment horizon ( or a group of horizons ) becomes dominant , buying and selling orders are not efficiently cleared and extreme events are likely to occur @xcite .
therefore , fmh directly predicts that critical events are connected to dominating investment horizons . it
is then straightforward that short investment horizons should dominate during financial turmoils and that is for two possible reasons
( i ) the long - term investors panic and start selling as a reaction to the negative news and crowd behavior , and in turn activity at short horizons increases compared to the long ones , and/or ( ii ) the long - term investors stay out of the market until the situation calms making the short horizons dominant . to uncover whether this prediction of fmh is correct for the global financial crisis ( gfc ) of the late 2000s , we apply the continuous wavelet transform procedure on developed and liquid world stock indices the usa
is represented by nasdaq , the uk by ftse , continental europe by french cac and german dax , and asia by hsi of hong kong and japanese nikkei between 3.1.2000 and 31.5.2013 .
the basic idea behind application of wavelets on the financial processes is their ability to analyze the underlying process both in the time and frequency domain . in a similar manner as the standard fourier analysis ,
the series is decomposed into frequencies and a scale - specific power is obtained .
however , the wavelets add a significant improvement over the fourier analysis as they provide a way to study how this scale - specific power evolves in time ( for more technical details , see the methods section ) . in terms of financial economics ,
the wavelet power can be understood as a scale - specific variance . compared to the
standardly used techniques such as various versions of ( generalized ) autoregressive conditional heteroskedasticity models @xcite , we obtain information about time evolution of variance as well as its distribution across frequencies ( scales ) . according to the fmh arguments , we should observe increased power at low scales ( high frequencies ) during the critical periods .
moreover , we might observe a changing structure of variance across frequencies before the turbulences due to the changing structure of investors activity as reported in our previous study @xcite . |
spiral arms are one of the the most striking , beautiful and scientifically challenging structures of disc galaxies .
they have fascinated and intrigued astronomers for centuries . as a rough initial approximation to treat this complex problem , the spiral arms had to be assumed as almost massless and/or with extremely modest pitch angles , in order to obtain a solution as a linear perturbation to the axisymmetric background potential .
the proposed solution to the permanence of spiral arms in galaxies was based on the spiral density wave linear theory ( @xcite , inspired in the 60 s work of b. lindblad and p.o .
lindblad ) .
the analytical solution at first order of the theory , known as the tight - winding approximation ( twa ) , that represents a weak potential model , extremely idealized as a smooth , even negligible perturbation to the background potential , was modeled as a simplified cosine function to represent the spiral arms gravitational potential .
probably followed by this initial attempt of solving the nature of spiral arms , a lot of work has tackled repeatedly the problem of modeling them , as a simplified periodic function that disregards the importance of their dynamical effects . however ,
spiral arms have proven to be an influential feature on galactic modern models ( such as perlas and n - body models , based on three dimensional density structures ) , far beyond of a simple smooth perturbation .
if we compare the spiral arms with the galactic bar , the latter has a mass between 10 and 20% of the disc mass @xcite , unlike the spiral arms that have a mass smaller in general than 5% of the disc mass @xcite .
consequently , the largest influence on the disc of strongly barred galaxies is , of course , due to the central bar .
therefore one would be tempted to oversimplify the scenario , by assuming that spiral arms are not influential at all on the dynamics of the disc . however , there are determinations that suggest that the majority of spiral galaxies , or at least those with clearly delineated spiral arms , are rather far from being linear @xcite . based on all those rigorous studies ,
spiral arms seem to merit an extra effort to model them beyond a simple perturbing term .
this fact is even more significant in the case of the interstellar gas , which is considerably more sensitive to the details of the potential , and in general , responds strongly even if the mass of the spiral arms is only a small fraction of the axisymmetric background @xcite . * for almost a century , formal studies of spiral arms have been carried on . yet , the spiral arms morphology , origin and nature are still poorly understood . with respect to their morphology
, there is plentiful theoretical and observational literature on structures related to spiral arms , such as spurs , branches , feathers and beads , and their plausible explanations @xcite .
regarding spurs and feathering , interesting scenarios to explain these features include plain hydrodynamic simulations , where the kelvin - helmholtz instability creates spurs produced by shocks with the spiral arms ( @xcite ; an extension to self - gravitating three dimensional models of this work is presented in @xcite ) .
other scenarios include the use of global mhd simulations , adding gas self - gravity and magnetic fields that produce differential compression of gas flowing through the arms resulting in the formation of sheared structures in the interarm regions that resemble spurs and feathers @xcite , or through gravitational or magneto - jeans instabilities @xcite .
* on the other hand , features such as branches are significantly longer than spurs that emerge almost parallel to the main spiral arms but with different pitch angles ( usually smaller ) , and are generally associated with resonances .
it is worth to note that , in the literature , the names spurs , feathers and branches are assigned to slightly different structures . in this work
we will adopt the definition given by @xcite and @xcite , where the `` spurs '' and `` feathers '' are small structures that arise almost perpendicularly from the main spiral arm , and `` branches '' are narrow structures that arise almost parallel to the main spiral arm , with smaller pitch angles and extension longer than spurs ( examples of branches can be seen in galaxies like ngc 309 , ngc 1637 , ngc 2997 , ngc 6946 , among others ) .
regarding such structures , theory and gas simulations have already shown a bifurcation of spiral arms produced by ultraharmonic resonances induced by the main spiral arm on the background gas flow @xcite . in this case , the bifurcation of the arms is expected due to the topology of the stable periodic orbits at the 4/1 resonance .
other interpretations also include the possibility that branches are the response of the gas produced by strong shocks against the main massive spiral arms , i.e. the ones made mainly of older and smaller stars * @xcite . *
specifically , regarding the milky way galaxy , its general structure has been extensively studied and the latest determinations seem to agree with two grand design symmetric spiral arms , seen on ir and optical , and some other weaker arms , detectable mainly in optical observations @xcite .
the weaker arms ( * branches * ) have been explained through hydrodynamic simulations as the response to the two - armed stellar pattern .
for example , @xcite and @xcite found that the gas response to a barred potential can produce a four - armed spiral structure .
likewise , from hydrodynamic and magnetohydrodynamic ( mhd ) simulations , the gas responds to an imposed two - armed spiral potential with four spiral arms @xcite . in this context , the gas component shocks at the position of the spiral arms , producing a density enhancement and star formation @xcite that , depending on the relative velocity between the gas and the spiral arm , dwell upstream or downstream or in the spiral arm .
in particular , @xcite showed that the nonlinear response of the gas to a stationary stellar spiral arm potential may produce two shocks , which then would be associated to a doubling of the spiral arm number as seen in the gas component .
@xcite demonstrated that shocks in galaxies arise necessarily if the spiral arm strength exceeds a certain critical value .
additionally , they found a range of values for the wave frequency that generates an ultraharmonic resonance which can provoke a secondary compression of the interstellar gas .
this effect has been related to the origin of the carina arm in the milky way @xcite , for example . however , since branches have been traditionally explained as shocks induced by the spiral arms , it is important to mention other observational work that shows there is little difference ( or non at all ) in the star formation efficiency ( i.e. no shocks ) between the spiral arms regions and the rest of the disk @xcite . finally , regarding their nature , a problem of great interest is whether spiral arms are a transient or a long - lived feature . on this matter ,
recent numerical simulations show that spiral arms are transient and recurrent structures @xcite .
however , in a study by @xcite on a sample of 27 galaxies , the authors concluded that the break found in the radial metallicity distribution near to corotation resonance ( cr ) , implies that the spiral structure is a rather long lived feature .
other authors , employing different techniques , seem to find observational proofs of long - lasting spiral arms @xcite . in this work
we explore the gas response to a bisymmetric spiral arm potential . for this purpose we employed two different models , a cosine potential ( generally used in literature ) and the three - dimensional , density distribution based potential perlas , applied to the particular case of a milky way - like galaxy . with this study
we find an alternative explanation to the formation of branches in disk galaxies , and their relation to transient or long - lived spiral arms .
this paper is organized as follows . in section [
sec : method ] , the galactic potential used to compute the stellar orbits and the mhd initial simulation set - up are described . in section
[ results ] , we present first a comparison between the models of the spiral arms : perlas and the cosine potential ; second , we present the gas response to spiral arms perlas model changing the structural parameters such as the pitch angle and the mass of the spiral arms , and their connection with the presence of galactic branches as a signature of transient spiral arms . finally , we present a discussion and our conclusions in section [ sec : conclusions ] .
motivated by the fact that the gas in galaxies is dynamically colder than the stellar disk , in addition to being collisional , we can expect it to be much more sensitive to details of the given potential than stars .
therefore , any differences between the cosine and perlas potentials might be magnified in gas with respect to the stellar response . * and so , * for purposes of comparison
, we produce the whole study employing both potentials . in this section
we introduce briefly the cosine and perlas potential models and the mhd setup applied to a milky way like galaxy . in all cases ,
the spiral arm models are superimposed to the axisymmetric background potential of @xcite , which includes a miyamoto - nagai bulge and disk , and a supermassive spherical halo .
table [ tab : param ] presents the basic parameters of the axisymmetric background potential . [ cols="<,^,^",options="header " , ] [ tab : param ] + references (1 ) @xcite ; ( 2 ) @xcite ; ( 3 ) @xcite ; ( 4)@xcite ; ( 5 ) @xcite . as mentioned before , a large majority of the investigations on the gas associated to spiral arms model them as a linear perturbation of the axisymmetric background , represented by @xmath0,\ ] ] where @xmath1 are cylindrical coordinates , @xmath2 is the amplitude function of the perturbation , given by @xcite as @xmath3 , where @xmath4 is the amplitude and @xmath5 is the inverse of the scale - length .
finally , @xmath6 describes the geometry of the spiral pattern ( _ locus _ ) , given by @xcite , as @xmath7,\ ] ] where @xmath8 is the pitch angle , @xmath9 is the start position for the spiral arms , and @xmath10 is a constant that shapes the starting point of the spiral arms , * so that * @xmath11 represents a @xmath12 tip transition from the bar to the spiral arms , while @xmath13 means a @xmath14 tip transition from the bar to the spiral arms . in the present work , we set it equal to 100 ( for further details , see @xcite ) .
in contrast , the perlas spiral arms model @xcite is a bisymmetric , three - dimensional , self - gravitating stationary model based on an adjustable mass distribution , rather than a local arm approximation , as the cosine potential .
several studies have shown that there are differences in the stellar orbital dynamics when the spiral arms are modeled with the cosine potential or with the perlas model @xcite . furthermore , chaotic orbital studies have demonstrated that a more detailed spiral arm potential ( the perlas model ) induces an important fraction of chaos , enough to destroy the spiral structure @xcite , when all chaotic behavior was originally attributed to effects produced by the bar , such as overlapping of resonances @xcite .
perlas is constructed by a superposition of individual oblate inhomogeneous spheroids along a given locus ( the same than in the cosine potential , equation [ eq : locus ] ) and superimposed to the axisymmetric background . in this model ,
the spiral arms have a well defined mass , unlike the cosine potential , where the spiral arms are treated as a periodic function of the potential not straightforward translatable to mass .
the mass assigned to build the spiral arms in the perlas model is subtracted from the disk mass .
thus , the inclusion of perlas to the galactic model does not modify the total mass of the original axisymmetric background . in table
[ tab : param ] we present the parameters of spiral arms that we used in our simulations . for further details about the perlas model ,
we refer the reader to @xcite . in order to compare perlas with the cosine potential
, we need to fit the spiral arm strength to make them as similar as possible .
we achieve this by adjusting the amplitude of the cosine potential ( the factor @xmath4 in eq .
[ eq : cosine ] ) so that the resulting radial and azimuthal forces approximate the ones obtained using the perlas model .
figures [ fig : frad ] and [ fig : fazi ] show the resulting fit for the case of a milky way - like galaxy , where the pitch angle is @xmath15 and spiral arm mass ( in perlas ) is @xmath16 of the disk mass . in this case
, the cosine amplitude @xmath4 is @xmath17 . in figure
[ fig : frad ] , each panel represents different radial lines , starting in @xmath18 up to @xmath19 . in the figure [ fig : fazi
] , each panel shows the computed azimuthal force at different radii , from @xmath20 to @xmath21 .
continuous lines represent perlas , dotted lines represent our fit to the cosine potential .
the spiral strength is related to the pitch angle and the mass of the arm . in the perlas model
, the spiral arm mass is a small fraction of the disk mass . to quantify the strength of the spiral arms , we calculated the @xmath22 parameter @xcite , frequently used to quantify the strength of bars and spiral arms @xcite .
the value of @xmath22 is given by @xmath23 where @xmath24 represents the maximum amplitude of the tangential force at radius @xmath25 , and @xmath26 , is the average axisymmetric radial force .
figure [ fig : qt ] shows @xmath22 for the perlas ( solid line ) and cosine ( dotted line ) potential models , for the case where the spiral arms mass of the perlas model has a 3% of the disk mass .
parameter for the spiral arms of a milky way - like galaxy for the perlas ( _ solid line _ ) and cosine ( _ dotted line _ ) potential models.,scaledwidth=50.0% ] as opposed to figures [ fig : frad ] and [ fig : fazi ] , where the force amplitude is almost the same , figure [ fig : qt ] shows that there is a difference between both potentials . while the maximum value of @xmath27 for perlas is @xmath28 , for cosine is only @xmath29 .
therefore , if the arm strength is measured using equation ( [ eq : q_t ] ) instead of the arm force , it is necessary to increase the cosine model amplitude . figure [ fig : qt_am20 ] shows the @xmath27 parameter corresponding to * an * increased cosine amplitude , in this case , @xmath30 . in this case , since @xmath27 for the cosine model is larger than perlas along all the radial range , this case should be considered as an example to test the gaseous disk response to an extreme cosine model , in order to verify if at this forcing the cosine can reproduce what perlas does . in this way we are bracketing the cosine potential within the values of the force for the perlas model .
, with the amplitude of the cosine model increased .
, scaledwidth=50.0% ] the initial set - up of the simulations consisted on a gaseous disk , with density profile given by @xmath31 $ ] , where @xmath32 , @xmath33 and @xmath34 .
the gas follows an isothermal equation of state with temperature @xmath35 .
additionally , the gas is permeated by a magnetic field , initially in the azimuthal direction , with an intensity given by @xmath36 $ ] , where @xmath37 and @xmath38 .
the disk is set up in rotational equilibrium between the centrifugal force , the thermal and magnetic pressures , magnetic tension and the background axisymmetric potential @xcite .
this equilibrium is perturbed by the spiral arm potentials under study , both rotating with a pattern speed @xmath39 .
we employed the zeus code @xcite to solve the mhd equations , which is a finite difference , time explicit , operator split , eulerian code for ideal mhd .
we used a 2d grid in cylindrical geometry , with @xmath40\kpc$ ] and a full circle in the azimuthal coordinate , @xmath41 , using @xmath42 grid points .
both boundary conditions in the radial direction were outflowing .
all calculations are performed in the reference frame of the spiral arms .
no self - gravity of the gas was considered .
we present in this section a comparison between the gas response to the simplified cosine potential for the spiral arms , and the density distribution based model perlas . in the limit for weak and/or small pitch angles ( approximately linear regime ) , both models behave very similar as expected , however , for stronger arms ( more massive or larger pitch angles ) , from this comparison we find severe differences in the gas behavior and formation of spiral arms and branch - like structures .
the deviation of the response between both models is induced by the basic differences of these potentials .
we then present an interpretation on the presence of galactic branches prior to the ultraharmonic 4:1 resonance and connect them to a possible signature of transient spiral arms . in order to compare both potentials
, we fit the cosine potential with perlas to make it as similar as possible in the force amplitude ( [ sec : fit ] ) . in all cases , the axisymmetric gaseous disk , initially in rotational equilibrium , was perturbed by the spiral potential either perlas or the cosine .
as mentioned already , the cosine potential represents a simple solution from the density wave linear theory , self - consistent for tightly wound spiral arms ( twa , i.e. where the perturbation is very small , which means small pitch angles or with very reduced masses ) . the milky way galaxy and the most of spiral
galaxies are actually far from this regime .
thus , it should not be surprising that the gaseous disk @xcite and the stellar orbits @xcite show a different structure when subjected to a self - gravitating , more realistic model instead of a local approximation . with this in mind
, the gaseous disk response to both models should be similar if we focus on a region of the parameter space where both potentials are valid , i.e. , if we set perlas and the cosine potential so that it is approximately in the linear regime , with a very small pitch angle and spiral arm mass . in figure
[ fig : gas_cos_perlas ] we show the linear regime for both potentials .
the density distribution with the cosine potential is presented in the left panel and perlas model in the right panel , the pitch angle is @xmath43 , the spiral arms mass is 1% of the disk mass , @xmath44 km@xmath45 s@xmath46 kpc@xmath47 . indeed ,
the gas response to the potentials is similar as expected , forming two spiral arms in both cases . for the specific case of a milky way - like galaxy
, we constructed a model that reproduces some of the observational parameters for the background and spiral arms potential to compare them with the cosine potential for the spiral arms ( force fitted with perlas ) .
the spiral arms pattern angular speed is @xmath48 , on a logarithmic locus with a pitch angle of @xmath49 .
the mass of spiral arm for perlas model is 3% of the disk mass , which is equivalent to a cosine amplitude of @xmath50 , when the non - axisymmetric force is employed for the fitting ( see [ sec : fit ] ) , and @xmath30 , when the arm strength ( i.e. the @xmath27 parameter ) is considered for the fitting instead ( see [ sec : strength ] ) .
we * follow the evolution of the system * for @xmath51 .
figure [ fig : gas_coseno ] shows the resulting density distribution of the gaseous disk when is perturbed by the cosine potential with a small amplitude ( @xmath50 ) .
after the simulation starts , the gas very rapidly settles into a spiral pattern , forming two spiral arms at @xmath52 ( top left panel ) .
@xmath53 into the simulation ( top right panel ) , the two spiral arms are better defined , and the gas is forming a secondary structure . as simulation progresses ( @xmath54 and @xmath55 , bottom panels ) , the gas continues responding to the spiral arms and even more substructure forms , but the large scale density response consists of two arms only .
the pitch angle of the formed arms is @xmath56 , equal to the imposed potential at the beginning of the simulation . in the last panel , a gas instability at the corotation radius is seen , as reported previously @xcite .
+ + in contrast , figure [ fig : gas_perlas ] shows the density distribution of the gaseous disk when is perturbed by the perlas spiral arm potential .
the gas very rapidly settles into a spiral pattern , as in the cosine potential case , forming two spiral arms at 30 myr . at @xmath53 and
later , the gas forms four spiral arms instead of the two arms in the cosine potential simulation .
even though the simulation develops four spiral arms , these are associated in two * pairs , each composed of * a strong arm and a weak arm .
the strong arms have a pitch angle of @xmath57 , and the weak arms have a pitch angle of @xmath58 .
this doubling of the spiral arms in the gas response has been seen in other * mhd * simulations using the perlas model ( @xcite ; see also [ sec : response ] ) .
+ in [ sec : strength ] , we noted that the spiral arm strength , as measured by eq .
( [ eq : q_t ] ) , was very different for both potentials , even if the force amplitudes were almost the same .
therefore , we decided to perform a second experiment with a larger amplitude @xmath59 for the cosine potential , that would be equivalent to a factor of @xmath60 times the force of perlas .
this much larger amplitude for the cosine potential is stronger than perlas at any radii , as shown in figure [ fig : qt_am20 ] .
the purpose of this overestimated experiment was to test if the four spiral arms formed with perlas were due to the strength of spiral arms only .
figure [ fig : gas_cos_aum ] shows the gas density distribution resulting from perturbing the disk with the larger force amplitude of the cosine potential .
it is readily seen that the effect of the spiral potential on the gas disk is larger and generates more substructure , but the gas still responds forming two spiral arms , unlike the density based potential perlas that forms four arms for larger pitch angles , i.e. for stronger arms .
consequently , this difference does not come from the force amplitude , but it seems rather originated from the self - gravitating nature and specific details of the potential that a local approximation for the forces given by the cosine potential is unable to reproduce .
we further explain this in the next section .
in [ sec : comparison ] we showed that the intrinsic differences in the nature of the cosine and perlas potentials induce a distinct gaseous arm structure as compared to a disk perturbed by a cosine arm potential .
while employing the spiral arm cosine potential , a bisymmetric gaseous structure seems to be an invariable outcome , for the perlas potential a two or four armed structure could be obtained .
this result points to the necessity of understanding how the gas responds to the perlas model when we vary the structural parameters of the spiral arms , such as its mass and pitch angle , considering the uncertainties in the determination of these parameters for the milky way s spiral arms .
therefore , we developed a set of simulations varying spiral arm pitch angles ( @xmath61 ) and spiral arms masses .
we present here some of them ranging from @xmath62 to @xmath63 , and the mass of the spiral arms ( @xmath64 ) , from 1.75% to 5% of the total disk mass ( @xmath65 ) .
figure [ fig : periodic_gas ] shows a mosaic of simulations .
the panels in the left column are the stable periodic stellar orbits and the panels in the right column are the gas density distribution after @xmath66 . * the maximum density response ( where the orbits crowd producing a density enhancement ) quantifies the orbital support to a given spiral perturbation through periodic orbits .
we computed the stellar periodic orbits in order to explore the orbital support to the imposed perlas spiral arm potential ( open squares in the left column - panels of figure [ fig : periodic_gas ] ) . to estimate the density response ,
we employ the method of @xcite .
this method assumes that stars in circular orbits in an axisymmetric potential , with the same sense of rotation of the spiral arms , are trapped around the corresponding periodic orbit in the presence of the spiral arms . for this purpose
, we calculated a set of central periodic orbits ( between 50 and 60 ) and found the density response along them using the conservation of the mass flux between any two successive periodic orbits . with this information we seek the position of the maximum density response along each periodic orbit ( filled squares in the left column - panels of fig .
[ fig : periodic_gas ] ) . these positions
are compared with the imposed spiral arms ( perlas model ) .
the method implicitly considers a small dispersion ( with respect to the central periodic orbit ) since it studies a region where the flux is conserved . on the other hand , this dispersion is based on parameters for the galaxies where dynamics is quite ordered , orbits follow their periodic orbit closely , in such a way that we consider this study a good approximation .
this method to estimate the density response has been widely used in literature @xcite .
we refer the reader to the work of @xcite for more details . * * with this in mind , a model in figure [ fig : periodic_gas ] * where the open and filled squares coincide would represent a stable , approximately orbitally self - consistent system , while a lack of coincidence would mean that the spiral is unlikely * to be * long - lasting @xcite . for @xmath67 and @xmath68 ( upper left panel of figure [ fig : periodic_gas ] )
, the stellar density response follows approximately the imposed spiral potential prior to the 4/1 resonance .
after that , the stellar response forms a slightly smaller pitch angle than the imposed spiral . in the mhd simulation
, the gaseous disk responds to the two - arm potential with the now familiar four spiral armed structure , where the pitch angle of the stronger pair of arms corresponds to that of the imposed pitch angle potential , while the other pair of gaseous arms has a systematically smaller pitch angle , corresponding closely to the regions of periodic orbits crowding .
now , with the same spiral arm mass but a pitch angle of @xmath69 ( second row of fig .
[ fig : periodic_gas ] ) , the stellar density response ( i.e. periodic orbits ) forms again a smaller pitch angle compared with the imposed , while in the gas , the four spiral arms seem stronger and the difference in the pitch angle between the gaseous and imposed arms is larger than the previous case . for @xmath70 and @xmath68 ( third row ) , the stellar density response closely follows the imposed spiral arm potential prior to the 4/1 resonance .
after that , the stellar response forms a slightly smaller pitch angle than the imposed spiral , while the gas responds with four spiral arms , but the second pair is very weak but with a significantly smaller pitch angle than the imposed spiral arm potential .
finally , with the same spiral arm mass ( 5% of the disk mass ) but a pitch angle of @xmath69 ( fourth row ) , we found no periodic orbits beyond 4/1 resonance and the stellar density response forms a pitch angle smaller than @xmath63 .
the gaseous disk responds with well defined four spiral arms that extend up to the corotation radius .
notice that in this last simulation , there is not much that can be said about the ( stellar or gaseous ) orbital support since periodic orbits tend to disappear due to the strong forcing of the imposed spiral arms , meaning that spiral arms would rather be transient by construction in this case and the mhd gaseous disk behavior is difficult to predict from periodic orbit computations .
however , such as the stellar arms constructed in this case , the gaseous spiral would be transient in likely even shorter timescales than in the case where periodic orbits exist but settle down systematically in smaller pitch angles than the original imposed spiral arms in the region where periodic orbits do exist .
notice that , in general , the arms that should eventually disappear in this scenario , are the stronger stellar imposed spiral arms ( see second row of fig . [
fig : periodic_gas ] ; @xcite ) .
therefore , the stellar response density maxima represent the regions of the arms where stars would crowd for long timescales , this is , where the existence of stable , long - lasting spiral arms would be more likely . on the other hand , if the stellar density response forms a spiral arm with a different pitch than the imposed angle , then the imposed spiral arms triggered on the disk ( by a bar , an interaction , etc . ) , would rather be structures of transient nature since those are not orbitally supported @xcite .
likewise , for the case of the spiral arms in the milky way , the values frequently seen in literature for the pitch angle , range from @xmath71 to @xmath63 .
these values and the knowledge of the galactic type could provide some information about their nature , i.e. whether they are long - lasting or transient structures . * following the pitch angle restrictions found by @xcite , the larger pitch angle values reported in literature for the milky way galaxy would imply that its spiral arms are a transient feature . *
the formation of four spiral arms in the gas response is , in this scenario , another piece of evidence of a transient nature of the spiral arms in the milky way galaxy , as we claim it is the secondary pair of arms in figure [ fig : periodic_gas ] , that coincides * more * closely with the stellar density response * from * periodic orbits , as expected from @xcite .
summarizing , the first ( imposed ) pair of massive spiral arms formed in the disk , with a larger pitch angle , triggers a second pair of arms ( traced approximately by the periodic orbits ) , with smaller pitch angles @xcite . in this outline
, the gas responds forming a second pair of arms aligned with the locus of the orbital crowding .
this lighter structures would likely be preferentially formed by young stars and gas than by evolved stars because of their transient nature , i.e. similar to what we call `` branches '' .
finally , in this framework , the presence of clear and strong branches in spiral galaxies , with smaller pitch angles than the corresponding couple of massive spiral arms on a galaxy , would be a signature of the transient nature of the spiral arms in a given galaxy .
on the other hand , spiral galaxies without evidence of branches could indicate the presence of a longer lasting spiral arm structure .
+ + +
with the use of a detailed three dimensional , density - distribution based potential for the spiral arms , combined with mhd simulations on a milky way - like galactic disk , we have studied the stellar orbital and gaseous response to the galactic potential . as a first experiment
, we constructed a simple cosine potential ( as the ones commonly employed in literature ) that reproduced approximately what the density based potential perlas exerts on the stellar and gaseous dynamics . the first set of simulations compare the gas response when the disk is perturbed by both spiral arm potential models .
we also varied the structural parameters of the spiral arms in the perlas model , such as the pitch angle ( @xmath62 to @xmath63 ) and the mass of spiral arms ( 1.75% to 5% of the stellar disk mass ) in order to understand how these parameters affect the gaseous disk dynamics .
additionally , we constructed stellar periodic orbits and calculate the stellar response density maxima . with these exercises we found that only in the case of the perlas model the gas and stellar density response ( based on the existence of periodic orbits ) is a consistent four - armed spiral structure : a couple of strong gaseous arms located at the position of the imposed stellar arms , and a second pair of weaker gaseous arms located at the position of the stellar orbit crowding .
* since the potential and the stellar density response do not coincide , the spiral arms are * prone to destruction .
* so , the presence of the second ( gaseous ) pair is interpreted as a sign of this transiency . * our main conclusions can be summarized as follows : * * we performed a * study of the gas response to * two * galactic potentials : the density distribution based model perlas and the widely employed in literature cosine potential .
we verified that the gas response ( a two armed structure ) is the same for both models close to the linear regime only , i.e. * for low - mass spiral arms and pitch angles smaller than @xmath72 . * in the general case however ( i.e. large pitch angle or arm mass ) , even when the cosine potential has been fitted as close as possible to the perlas model , they produce quite distinct outcomes on the gas response . in the case of the cosine potential ,
the gas responds invariably forming two spiral arms while , with the perlas model , the gas responds with four gaseous spiral arms ( in the general strong arm case ) .
* we increased the strength of the spiral arms represented by the cosine potential up to a point the arms were equivalent and even beyond the mass of a strong bar as an experiment to try to reproduce the gas response provided by the perlas model .
however , the answer was always a bisymmetric structure .
we conclude that the spiral arm strength is not responsible for the four - arm gas response , but rather it is the product of the forcing generated by the whole density distribution better represented by the perlas model that , in turn , forces the periodic orbit response to shift its crowding regions inside the imposed locus of the massive spiral arms . *
using the perlas model , we changed the structural parameters of the spiral arms according to the observational and theoretical uncertainties in the determination of the milky way s spiral arms in literature . within these values , for the general case , the stellar response density maxima systematically forms spiral arms with a smaller pitch angle than the imposed spiral , meaning that the spiral arms might be a transient feature in the milky way galaxy @xcite .
the presence of a second pair of lighter spiral arms ( `` branches '' ) , with smaller pitch angles induced by the first more massive stellar spiral arms , might be evidence of the lack of support to the stronger arms on a galactic disk , and therefore evidence of the transient nature of spiral arms in a given galaxy . applying this scenario to the milky way ,
for the stronger spiral arm values reported in literature ( i.e. pitch angles larger than @xmath72 , and masses larger than @xmath73 of the disk ) , the spiral arms in the milky way would be of transient nature . *
although in this work we applied the models to milky way - like galactic discs , it is worth noticing that the results are general .
this means that the presence of branches with smaller pitch angles than the main arms might be the `` smoking gun '' that proves transiency of spiral arms in any given galaxy .
we thank edmundo moreno for enlightening discussions that helped to improve this work .
apv acknowledges the support of the postdoctoral fellowship of dgapa - unam , mxico . this work has received financial support from dgapa papiit grants in111313 and in114114 . | with the use of a background milky - way - like potential model , we performed stellar orbital and magnetohydrodynamic ( mhd ) simulations .
as a first experiment , we studied the gaseous response to a bisymmetric spiral arm potential : the widely employed cosine potential model and a self - gravitating tridimensional density distribution based model called perlas .
important differences are noticeable in these simulations , while the simplified cosine potential produces two spiral arms for all cases , the more realistic density based model produces a response of four spiral arms on the gaseous disk , except for weak arms -i.e .
close to the linear regime- where a two - armed structure is formed . in order to compare the stellar and gas response to the spiral arms ,
we have also included a detailed periodic orbit study and explored different structural parameters within observational uncertainties .
the four armed response has been explained as the result of ultra harmonic resonances , or as shocks with the massive bisymmetric spiral structure , among other . from the results of this work , and
comparing the stellar and gaseous responses , we tracked down an alternative explanation to the formation of branches , based only on the orbital response to a self - gravitating spiral arms model .
the presence of features such as branches , might be an indication of transiency of the arms .
[ firstpage ] mhd galaxy : disc galaxy : strcuture galaxy : kinematics and dynamics galaxies : spiral galaxies : structure |
a diamod doped with boron can become superconducting at temperatures of few kelvins @xcite .
such critical temperatures are surprisingly high for impurity band conductivity .
moreover , one can expect that @xmath4 will be soon further increased
. it would be useful to predict an optimal sample composition from a microscopic theory .
to this end we evaluate the density of states at the fermi level within the dynamical cluster approximation and show that it can be approximated by a simple power - law . using this density of states
the @xmath4 can be predicted from a mcmillan type formula .
diamond is an insulator with a band gap of @xmath5 .
boron atoms create shallow acceptor levels close to the valence band with activation energy of holes @xmath6 . at a doping concentration above @xmath7
the system presents a metal - insulator transition @xcite and might become superconducting at @xmath8 k. with increasing concentration @xmath4 increases .
a favoured theory of superconductivity driving mechanism is , that electrons mostly interact with localized vibrational modes on the boron atoms in an optic spectra .
however other possible theories are discussed as well @xcite .
experimental data on boron doped diamond are controversial . majority of samples were prepared in thin layers and their properties strongly depend on crystallographic orientation of the surface .
there are two widely studied types with @xmath9 and @xmath10 orientation , see references @xcite and @xcite , respectively .
both groups have used the microwave plasma assisted chemical vapor deposition method though under different growth conditions .
their samples differ in many characteristics including the @xmath4 .
the @xmath10 samples have higher @xmath4 in general . in this paper
we focus on the @xmath9 samples which are also supported by bulk samples prepared with the high temperature high pressure method @xcite revealing a comparable @xmath4 .
the theory of boron - doped diamond can benefit from an extensive experience with superconductivity in disordered materials .
@xcite unlike in metallic alloys , where the disorder mainly modifies the coherence length and can be handled within simple approximations , in the case of doping induced superconductivity a sophisticated treatment of the disorder is necessary to describe a formation of the impurity band .
shirakawa @xcite have shown that a concentration dependence of @xmath4 measured in boron doped diamond can not be reproduced by the bcs theory in which the disorder is described on the level of the coherent potential approximation ( cpa ) .
since this theory yields too low values of @xmath4 , they proposed that disorder effects beyond the cpa are responsible for this disagreement .
we will show that corrections beyond the cpa do increase @xmath4 , however in the extent which is not sufficent to cover discrepancies in question .
in the present paper we compute the density of states at the fermi energy level for boron doped diamond within the dynamical cluster approximation ( dca ) .
while the cpa assumes a single atom embeded in the effective medium , the dca generalises this idea by embedding a cluster .
the cluster allows us to describe splitting of closely located impurity bound states which contributes to the shape of the impurity band . with respect to the superconductivity it is essential that the density of states at the fermi level computed within the dca is higher than the cpa value . apparently , the density of states at fermi level is much lower in the impurity band than in ordinary metals . since values of @xmath4 are comparable , the boron doped diamond belongs to the family of the strong coupling materials .
one can expect large dicrepancies from the bcs theory and we will show , that this is indeed the case . in this paper
we use the belitz theory @xcite which generalises the mcmillan formula to disordered superconductors .
the structure of the paper is as follows . in sec .
ii we simplify the belitz theory so that it depends on the doping exclusively via the disorder dependent density of states . in sec .
iii we evaluate the density of states within the dynamical cluster approximation .
the coherent potential approximation is evaluated for comparision . in sec .
iv we discuss the concentration dependence of the critical temperature comparing predictions of the belitz theory , the mcmillan formula and the bcs theory . to this end
we first establish density independent material parameters with the help of which one can construct the density dependent coupling strength @xmath11 , the coulomb pseudopotential @xmath12 and the belitz correction @xmath13 .
section v brings conclusions .
let us outline the belitz theory , first . in the absence of currents , the eliashberg selfenergy has three terms @xmath14 where @xmath15 is the kinetic energy and @xmath16 is the matsubara frequency . for pure superconductors
@xmath17 renormalised by @xmath18 gives a gap function @xmath19 while @xmath20 is a scalar that only shifts the chemical potential having no effect on the critical temperature .
in disordered superconductors , parameter @xmath20 becomes energy and frequency dependent .
belitz has shown that its energy derivative at fermi energy @xmath21 modifies the mcmillan formula as @xmath22 here @xmath11 describes an electron - phonon coupling and @xmath23 is the screened pseudopotential , in which @xmath12 characterises a strength of the coulomb interaction , @xmath24 is its effective range and @xmath25 is the debye frequency . setting @xmath26
one recovers the mcmillan formula .
in general , belitz formula includes four disorder dependent material parameters @xmath25 , @xmath11 , @xmath12 and @xmath13 . based on ab - initio computations @xcite which show that @xmath25 changes negligibly with doping we ignore the disorder dependence of @xmath25 and use a pure diamond value @xmath27 .
the strong disorder dependence of remaining parameters stems from a rapid variation of the density of states at the fermi energy level @xmath28 in vicinity of the metal - insulator transition .
we will use a simplified version of the belitz theory neglecting disorder corrections to interaction vertices in perturbation series .
this brings a significant reduction of numerical demands , because @xmath11 , @xmath12 and @xmath13 reduce to functions of @xmath28 only .
moreover , with the vertex correction neglected , @xmath29 so that we are left with @xmath11 and @xmath12 similarly as in the mcmillan theory . we take @xmath11 from ab - initio computations .
since the published results cover only few concentrations , we evaluate @xmath11 in the spirit of morel and anderson @xcite formula @xmath30 here @xmath31 represents a phonon - electron coupling strength and @xmath32 describes a screening .
the pseudopotential @xmath12 reads @xcite @xmath33 where @xmath34 is the strength of the coulomb interaction .
since we have found no measurement of @xmath34 in the literature , we have to treat it as a parameter which is set from experimental data .
because in the boron - doped diamond the fermi energy level is localised very close to the top of the valence band we associate @xmath35 with the effective range of the coulombic interaction and estimate both as the distance of the fermi level from the top edge of the impurity band .
the equation then has a simple form @xmath36 using formulas , and in and , we obtain @xmath4 as a function of @xmath2 .
to proceed we have to establish @xmath28 as a function of the boron concentration @xmath2 .
this single particle property is independent of the electron - electron interaction , would it be the coulomb repulsion or the phonon - mediated pairing potential .
the density of states is thus given by a hamiltonian of the valence band in diamond @xmath37 and a random potential of boron impurities @xmath38 in the random potential @xmath39 at impurity sites and zero elsewhere , and @xmath40 is the potential amplitude .
we note that this single - site s - type potential does not cover the triple degeneracy of impurity state of boron .
we adopt this hamiltonian already studied by shirakawa @xcite to make a link between their and present results .
we have evaluated @xmath1 within the dca for clusters of @xmath41 atoms on a cubic lattice . compared to the cpa density of states employed by shirakawa ,
our cluster density of states includes nontrivial corrections given by bonding and anti - bonding states at neighbour impurity sites .
difference between the cpa and dca density of states in the impurity band is ilustrated in figure [ fig0_dos ] for 5% of boron doping . within the cpa ( dashed line ) the impurity band yields rather featureless density of states having a slightly skewed semieliptic shape .
the valence band starts approximately at an energy of @xmath42 ev and ends at @xmath43 ev .
we focus on the impurity band because the fermi energy , @xmath44 , lies there . since boron is an acceptor , its bound states have higher energy than extended states in a valence band . in the dca ( solid line )
one can distinguish additional contributions of two and three boron clusters .
clusters of four to eight borons are also included , but their contributions are invisible for the given concentration . the main part of the impurity band density of states is formed by a bound state on a single boron . a shoulder on the right side of the impurity band results from a symmetric bound state of two neighbouring borons . the nonsymmetric state is not bounded but forms a resonant state visible in the valence band .
the tree boron bound state is splited off at an energy @xmath45 ev .
higher states of the three boron clusters make a negligible contribution . with respect to superconductivity
the fermi energy region is the most important .
as one can see the density of states at @xmath44 is dominated by the single boron states .
this is a reason why the cpa and dca give comparable @xmath1 .
the resulting density of states at the fermi level @xmath1 is presented in figure [ fig1_n ] .
for comparison we also show the cpa result .
one can see that the dca density of states is higher than the cpa value at all impurity concentrations .
both resulting densities with a good accuracy obey power - law , @xmath46 and @xmath47 .
a very good approximation by power - law also holds for the distance of the fermi level from the top of the impurity band , @xmath48 for the dca . in the rest of this section we provide more details of our computation .
the reader not interested in these technical details can skip to the next section .
we only briefly introduce the dca here , the reader can find all details in reference @xcite .
a basic idea of the dynamical cluster approximation is to divide the brillouin zone into subzones in our case @xmath41 subzones . within the subzone
the selfenergy is momentum - independent , i.e. , the selfenergy is represented by eight complex functions of frequency . by symmetry arguments
this number can be reduced to four functions .
all subzones contribute to the density of states @xmath49 where @xmath50 is a subzone index .
in general , the subzone contribution @xmath51 is obtained integrating over the subzone with the electron dispersion of the valence band . for simplicity
we approximate these functions by a semielliptical distribution with a proper width , shift and normalisation @xmath52 where @xmath53 is number of subzones ( 8 in our case ) , and @xmath54 are width and center of subzone energy bands with @xmath55 and @xmath56 being the maximum and minimum energy in the subzone @xmath50 , respectively .
the density of states by definition maintains the width of the valence band .
since one of the subzones is centered around the @xmath57 point of the brillouine zone , this approximation also yields a correct curvature at the edge , i.e. , it correctly reproduces an effective mass of holes near the band edge .
this feature is vital for a realistic description of the relatively shallow impurity state .
we employ the model valence band in the cpa as well .
the parameters of the hamiltonian are fitted according to the facts that the valence band is @xmath58 wide and the single impurity bound state appears at energy @xmath59 above the valence one . using the local green function corresponding the density of states , we come to @xmath60 ev . we determine the fermi energy level from a condition for a local averaged density of electrons at zero temperature @xmath61 .
we are aware that a measured number of charge carriers can differ significantly from a number deduced from the concentration of boron atoms in the sample .
as has recently been pointed out by mukuda @xcite quite a large fraction of boron atoms can form the neutral b - h complexes , reducing the concentration of charge carriers . on
the other hand klein @xcite found that the effective number of carriers deduced from hall - effect measurements was much larger than the number of boron atoms in samples .
since the charge carrier concentrations for all samples are not accessible we assume for simplicity the films to be doped ideally .
in this section we discuss the concentration dependence of the critical temperature @xmath0 . to be able to implement the belitz theory
we need the coupling strength @xmath11 and the pseudopotential @xmath12 as functions of @xmath1 .
the coupling strength @xmath11 we deduce from ab - initio computations . in literature one
finds the studies within the virtual crystal approximation @xcite and the supercell method @xcite . in figure [ fig3_l ]
we show fit of ab - initio results by formula .
we found that parameters @xmath62 and @xmath63 yield a reasonable fit of rather scattered computed values .
it remains to establish the pseudopotential . according to formula we have to find a single parameter @xmath34 , which determines @xmath12 for all concentrations .
dots in figure [ fig4_mu ] show values of @xmath12 deduced from the belitz theory and experimental values of @xmath0 .
one can see that all dots stay close to a line given by formula with @xmath64 .
the parameter @xmath34 holds for all concentrations and it is the only material parameter fitted to experimental values of @xmath0 in the present theory .
now all relations and parameters are ready for predictions of the critical temperature .
the concentration dependence of @xmath0 given by the belitz formula and power - law approximation of the dca value of @xmath1 is shown in figure [ fig5_t ] .
as one can see , the theory describes a steep increase of @xmath0 with doping at the region of small concentrations . at higher concentrations the critical tempeture satures at about 3 k. this saturation reminds trends described by the theory of osofsky @xcite based on a heuristic rescaling of the bcs parameters . for comparison
we also show the value predicted by the belitz theory for the cpa density of states .
as one expects from the lower density of states , the cpa leads to lower values of @xmath4 , namely in the region of small concentrations .
our results thus confirm a trend predicted by shirakawa that corrections beyond the cpa will lead to higher @xmath4 .
we would like to emphasis that the disorder corrections by belitz are necessary for a good agreement between experimental data and theory .
it can be seen from attempt to the fit @xmath4 with the mcmillan formula as shown in figure [ fig5_t ] . in the comparison
we have not used the pseudopotential @xmath12 fitted within the belitz theory , but made a seprate fit directly from the mcmillan formula , see figure [ fig4_mu ] .
finally we show that the boron - doped diamond has to be treated as a strong coupling superconductor . in figure [ fig2_bcs ]
we show @xmath9 experimental @xmath4 data and its description with the bcs formula @xmath65 , where @xmath66 is the bcs interaction .
apparently , the bcs theory yields an incorrect concentration dependence .
it is noteworthy how much the corrections beyond the cpa increase @xmath0 within the bcs theory . in the strong coupling theory
these corrections are smaller by an order of magnitude .
to conclude we have discussed corrections beyond the cpa to the critical temperature in diamond doped with boron . from numerical studies of the dynamical cluster approximation with cluster of @xmath41 atoms it follows that these corrections increase the density of states at fermi level .
as expected this causes an increase of the critical temperature . comparing different approximations
we have shown that the strong coupling theory is necessary to predict realistic critical temperature .
it also turns out that disorder corrections of belitz improve agreement with experimental data .
the present theory offers a simple analytic formula for the critical temperature as a function of boron density .
indeed , in a simplified belitz theory all material parameters depend exclusively on the density of states and can be expressed as its linear or rational functions .
numerical studies of @xmath67 show that the density of states at the fermi level follows a power - law @xmath68 , what allows one to predict the critical temperature without any numerical effort .
author would like to thank pavel lipavsk ' y for many useful discussions and also to j. h. samson for a critical reading of the manuscript .
the access to the metacentrum computing facilities provided under the research intent msm6383917201 is also highly appreciated .
this work was supported by grants no .
gar 202/07/0597 , no .
av0z10100521 , no .
gauk 135909 , and no .
gaav 100100712 , the daad and european esf program aqdjj .
30 ekimov e a , sidorov v a , bauer e d , melonik n n , curro n j , thompson j d and stishov s m 2004 _ nature _ ( london ) * 428 * 542 klein t 2007 _ phys .
_ b * 75 * 165313 mare j j , hub ' ik p , nesldek m , kindl d , kritofik j 2007 _ diamond relat .
_ 16 takano y , nagao m , sakaguchi i , tachiki m , hatano t , kobayashi k , umezawa h and kawarada h _ appl .
lett . _ * 85 * 2851 weinkauf a and zittartz j 1975 _ j. low temp .
* 18 * ; 1974 _ solid state commun . _
* 14 * wysoki ' nski k f and kuzemsky a l 1983 _ j. low templ phys . _ * 52 * 81 wysoki ' nski k f 1987 _ solid state commun . _ * 63 * 3 belitz d 1986 _ phys
_ b * 35 * 4 ; 1987 _ phys rev .
_ b * 35 * 1651 shirakawa t , horiuchi s , ohta y and fukuyama h 2007 _ j. phys .
jpn . _ * 76 * 014711 giustino f , yates j r , souza i , cohen m l and louie g 2007 _ phys . rev
. lett . _ * 98 *
047005 boeri l , kortus j and andersen o k 2004 _ phys .
lett . _ * 93 * 237002 morel p and anderson p w 1962 _ phys .
rev . _ * 125 * 1263 vonsovsky s v , izyumov yu a and kurmaev e z 1982 _ superconductivity of transition metals _
( springer - verlag berlin heidelberg new york ) jarrell m and krishnamurthy h r 2001 _ phys .
_ b * 63 * 125102 mukuda h 2007 _ phys . rev .
_ b * 75 * 033301 lee k w and pickett w e 2004 _ phys .
lett . _ * 93 * 237003 ma y , tse j s , cui t , klug d d , zhang l , xie yu , niu y and zou g 2005 _ phys .
_ b * 72 * 014306 blase x , adessi c and connetable d 2004 _ phys . rev
93 * 237004 xiang h j , li z , yang j , hou j g and zhu q 2004 _ phys . rev .
_ b * 70 * 212504 bustarret e 2004 _ phys .
lett . _ * 93 * 237005 takano y , takenouchi t , ishii s , ueda s , okutsu t , sakaguchi i , umezawa h , kawarada h and tachiki m 2007 _ diamond relat .
_ * 16 * 911 osofsky m s , soulen r j , jr .
, claassen j h , trotter g , kim h and horwitz j s 2001 _ phys .
lett _ * 87 * 197004 ; 2002 _ phys .
_ b * 66 * r020502 | we study the concentration dependence of the superconducting critical temperature @xmath0 in a boron - doped diamond .
we evaluate the density of states at fermi level @xmath1 within the dynamical cluster approximation obtaining higher values than from the coherent potential approximation .
we discuss the @xmath0 as a function of @xmath1 within the bcs , the mcmillan , and the belitz theory .
the simplified belitz theory gives the best agreement with experimental data . since the density of states follows a simple power - law for accessible doping concetrations @xmath2 , the present theory offers an analytical formula for @xmath3 . |
the diffuse stellar halo component of galaxies represents a tiny fraction of the mass ( about 1% in the case of the milky way ; morrison 1993 ) and the low surface brightness of the population makes study in extragalactic settings difficult . however , galactic halos are unique laboratories for investigating fundamental galaxy properties
. age and/or metallicity distributions of halo stellar populations and their kinematics provide fossilized glimpses of the earliest conditions of galaxy formation .
halo populations may give us answers to key questions about the chronology of galaxy formation : how the halo formation was related to the assembly of galactic mass and the formation of galaxies .
different views have been advanced toward understanding the formation of galactic stellar halos .
the first scenario was proposed by eggen , lynden - bell , & sandage ( 1962 ) , where the authors proposed that the metal - poor stars in the galactic halo were formed during a rapid collapse of a relatively uniform , isolated protogalactic cloud .
this picture is generally viewed as advocating a rapid dissipative and monolithic collapse of protogalaxies .
an increasing number of observational findings , both at high redshift and in nearby galaxies , together with cosmological models of structure formation , challenge this general view , and suggest an alternative picture of how halos of galaxies may have formed . this alternative paradigm for halo formation
is based on the idea that the galactic halo was formed via the accretion , over an extended period , of small metal - poor satellites which underwent independent chemical evolution before being accreted ( searle & zinn 1978 ; see also freeman 1987 ) .
the disruption of globular clusters , revealed by tidal tails ( grillmair et al .
1995 ) , may also contribute to the stellar halo ( aguilar et al .
this is supported by the observation that halo field stars and globular clusters in the milky way have similar mean metallicities ( carney 1993 ) .
note that this is not the case for m 31 stellar halo ( durrell et al .
2001 , perrett et al .
an interesting consequence of this hypothesis is that a fraction of globular clusters may be stripped relics of highly nucleated dwarf satellites ( see freeman & bland - hawthorn 2002 for a detailed discussion ) .
the discovery of the sgr dwarf galaxy ( ibata et al .
1994 ) , streams and moving groups in halos of both the galaxy ( yanny et al . 2000 ; ivezic et al .
2000 ; dohm - palmer et al . 2001 ) and m31 ( ibata et al . 2001 ) encourage renewed interest in halo formation through accretion of smaller protogalactic structures .
although great effort has been dedicated in recent years to investigate stellar populations and/or kinematic of local group spiral halos , it is unlikely that the problem of galactic halo formation may be understood entirely by studies of these galaxies alone .
questions about the universality of halo formation mechanism(s ) , the similarity of halo global properties and their sensitivity to disk and/or bulge properties , among others , will require a larger sample of galaxies .
one means of addressing these questions is to extend the study of galaxy halos to other spiral galaxies .
such studies are now within reach of the hst .
our purpose is to study systematically the _ global properties of population ii halos _ of nearby normal spiral galaxies .
such a program may reveal whether there was a universal formation mechanism , or whether there are mixture of different stellar halos ; if the halos are dominated by formation of the nuclear bulge and perhaps enriched by gas outflow , or composed of debris of disintegrated dwarf galaxies revealed by the presence of a large fraction of intermediate age stellar populations .
also of great interest is the investigation of whether halo and disk / bulge properties are correlated . in order to interpret the color - magnitude diagrams ( cmds ) of resolved stars in nearby galaxies
it is important to have reliable distance estimates .
fortunately , sometimes the cmds themselves provide such an estimate .
stars on the first - ascent red giant branch ( rgb ) climb this phase with an expanding convective envelope and an hydrogen burning shell .
while evolving through the rgb , low - mass stars develop an electron - degenerate core , which causes an explosive start of the core - he burning phase , the so - called he - flash , almost independently from the initial stellar parameters , such as initial mass or abundance ( chiosi et al .
the he - flash is followed by a sudden decrease of the stellar luminosity due to the extinction of the h - burning shell .
hence low - mass stars will accumulate along the rgb , reaching their maximum luminosity , during this phase , at the tip of rgb ( trgb ) .
this behavior translates to an abrupt discontinuity in the luminosity function ( renzini 1992 ) .
as the evolution of the rgb stars depends essentially on the core - mass , the number of stars per luminosity interval will be related directly to the he - core growth rate ( iben & renzini 1983 ) .
theoretical and observational investigations find the luminosity function of rgb stars to follow a simple power - law ( see zoccali & piotto 2000 and references therein ) .
the trgb magnitude is the location where the rgb luminosity function truncates .
below metallicities @xmath2 \approx -0.7}$ ] , the location of the trgb in the @xmath0 band is expected to be quite insensitive to age and/or metallicity ( e.g. , salaris & cassisi 1997 ) . at higher metallicities ,
bolometric corrections in the @xmath0 band become important and the trgb magnitude becomes increasingly sensitive to metallicity ( e.g. , salaris & girardi 2005 ) .
the trgb method is also well supported observationally .
the location of the absolute i - band trgb luminosity for a sample of galactic globular clusters spanning a large range of abundances , @xmath3 \leq -0.7}$ ] , over ages spanning 2 - 15gyr ( da costa & armandroff 1990 ) , is quite stable and insensitive to age and metallicity ( for @xmath4 \leq -0.7 $ ] ) , changing by less than @xmath5mag ( lee et al .
1993 ) .
good agreement is generally found between the distances obtained using the trgb method and classical distance indicators such as cepheid and rr lyrae variables ( e.g. , sakai et al .
1996 , ferrarese et al .
being a population ii distance indicator , unlike the cepheid method ( population i distance indicator ) , the trgb method can be applied to any class of galaxies , making this an attractive and easy technique to use to estimate galaxy distances . because the outer regions of galaxies are devoid of significant concentrations of gas and dust and the stellar density is low , studies of standard candles in the halo provide a means of estimating distances free of assumptions concerning internal reddening in galaxies or crowding . in [ data ] , the details of observations using the hst wide field planetary camera 2 ( wfpc2 ) and data reduction are reported . in [ methods ] , we present the detection of rgb stars and the methods for measuring the trgb using the @xmath0-band luminosity function that we have employed . finally , in [ dist ] , we present the color magnitude diagrams , luminosity functions , and derived distances for the galaxies studied .
until recently , no spiral galaxy halo outside of the local group had been imaged to a depth sufficient to permit study of the halo metallicity distribution function or other stellar population studies , such as the presence of intermediate - age stars .
hst observations have begun to change this . we have observed a sample of spiral galaxies with the aim of unambiguously resolving their stellar halos down to one or two magnitudes below the trgb .
our selection criteria of the sample galaxies are as follows .
we seek spiral galaxies with ( i ) morphological types between sa to sc , ( ii ) distance moduli @xmath6 , ( iii ) high inclinations , i.e. , @xmath7 degrees , to reduce any contamination from the outer disk stars , and ( iv ) absolute magnitudes @xmath8 .
the resulting sample of eight galaxies represents most of the nearby luminous inclined spiral galaxies . in this paper
, we estimate the distances to four galaxies in our sample , i.e. , ngc 253 , ngc 4244 , ngc 4258 , and ngc 4945 , using the tip of the red giant branch technique .
[ field_location ] shows the locations of the observed halo fields , superimposed on the digitized sky survey images of these galaxies .
madore & freedman ( 1995 ) argue that in order to minimize potential biases and to securely identify the trgb , there should be at least 50100 stars in the first magnitude interval of the rgb star luminosity function . with sparser
data sets the distance modulus tends to be overestimated .
four galaxies in our sample have luminosity functions that are too poorly populated to yield reliable trgb distances : ngc 55 , ngc 247 , ngc 300 , and ngc 3031 .
hence these galaxies are not discussed in the present paper ; the color - magnitude diagrams of the observed halo fields for these galaxies are shown and discussed in paper iii of this series .
tikhonov , galazutdinova & drozdovsky ( 2005 ) estimate trgb distances for three of these galaxies using larger hst data sets . the galaxy sample is summarized in table 1 .
the galaxies were observed through the f814w and f606w filters .
exposure times through the f814w filter were set to reach s / n = 5 in the wf camera for an absolute magnitude @xmath9 ( for galaxies with @xmath10 ) or @xmath11 ( for ngc4258 and ngc4945 , which are more distant ) .
the f606w exposure times were set to reach the same s / n at the same absolute magnitude for metal - poor rgb stars .
the exposures were typically dithered over 3 separate pointings ( the range being from 2 to 6 ) with a pattern extending over approximately 0.25 arcseconds to allow rejection of hot pixels and detector artifacts .
the images were reduced through the standard hst pipeline , using the latest flatfield observations and using contemporaneous super - dark reference frames .
the dithered frames were combined with iterative cosmic ray rejection using software in the stsdas.dither package based on the drizzle algorithm of fruchter and hook ( 2002 ) .
briefly , the images were drizzled ( shifted , geometrically corrected , and resampled ) to a common frame to construct a median image reasonably free of cosmic rays .
this image was geometrically transformed back to the pixel grids of the original image and used as the truth image for identifying cosmic rays .
the images were then re - drizzled onto a pixel scale of 0.1@xmath12 , masking out the cosmic rays , hot pixels , and detector artifacts , using optimal weights based on the counts at the sky level in each image .
a value pixfrac=0.8 was used to minimize the degradation of the image quality due to the resampling .
the stellar magnitudes were measured through circular apertures with a radius of 0.15 .
these aperture magnitudes were corrected to total magnitudes using tinytim model ( krist 2004 ) point - spread functions .
aperture - corrections and zeropoints are applied separately for each ccd chip .
magnitudes are in the vega system ( i.e. , vega has magnitude=0 through both filters , when measured through a 3aperture ) .
diffuse sources have been excluded from the catalogs based on the magnitude difference between measurements through 0.09and 0.3apertures .
objects where this difference is greater than 1.3 magnitudes are generally galaxies or blends of stars . this demarcation was determined by starting with a tiny - tim prediction of the ratio of fluxes for a point source measured in those apertures .
we mark objects in the images whose flux growth exceeds that prediction and iterate to arrive at a value which preserves point sources while excluding those with extended radial profiles .
the stellar photometry was corrected for charge - transfer efficiency ( cte hereafter ) effects using the 31 may 2002 version of the dolphin ( 2000 ) equations .
the cte correction amounts to a maximum of @xmath13 mag for faint stars at the top of the chip in fields with the highest background , such as ngc3031 , while reaching @xmath14 mag in fields with low sky background , such as ngc55 .
typically the maximum correction for the data is @xmath15 mag .
transformation of the instrumental magnitudes to standard v and i magnitudes followed the prescriptions of holtzman et al .
first , we have corrected from the foreground extinction using @xmath16 and @xmath17 , then given the instrumental magnitudes , we have determined the standard johnson - cousins magnitudes by solving iteratively the second degree equations relating the wfpc2-to - vi magnitudes using the coefficients tabulated by holtzman et al .
fig.[phot_error ] shows the evolution of the resulting photometric errors for i - band as a function of the apparent magnitude .
the typical trend of increasing photometric errors with apparent magnitude can be seen , where most of the stars observed in the fields are affected by photometric errors of a few tenths of a magnitude .
to avoid stellar populations that may bias the analysis of the luminosity function , we have restricted our analysis to stars that match the location of the rgb sequence of metal - poor stellar populations .
to do so , we have retained only stars that are bluer than a certain cut .
for all the galaxies discussed in this paper , the remaining number of red giant stars after the color cut is sufficient for a reliable trgb estimate by the criterion of madore & freedman ( 1995 ) ; i.e. , there are more than 100 stars in the one - magnitude interval fainter than the trgb .
it is worth mentioning that as long as the aim is measurement of the trgb , there is no need to include _ all _ the observed stars to construct the luminosity function .
when the red giant branch is not well populated , metal - rich stars , i.e. , @xmath18 \ga -0.5}$ ] , that occupy the red part of the cmd where no calibration is available , might make the tip of the red giant branch less well identifiable .
our simple color cut accomplishes this goal , but will lead to different overall luminosity functions for different metallicity distributions .
this is not a problem , provided that the trgb feature itself is clearly visible .
the mean metallicities of halo stellar populations of all our sample galaxies are at or below @xmath18\la\,-0.6}$ ] ( see the second paper of this series for more details ) , so any contamination from metal - rich stars is expected to be small .
we used two different methods to detect and to measure the trgb i - band magnitude from the luminosity function of a stellar population . in the following subsections we present briefly our methodology .
the most widely used procedure to estimate the trgb i - band magnitude uses the abrupt termination of the rgb luminosity function at the trgb magnitude .
the trgb discontinuity causes a peak in the first derivative of the observed stellar magnitude distribution ( lee et al .
1993 ; madore & freedman 1995 ) .
sakai et al . ( 1996 ) update this method by gaussian - smoothing the luminosity function to avoid the arbitrary choice of the binning size , and use a continuous edge detection function . due to the power - law form of the rgb luminosity function
, we use a kernel - smoothed logarithmic luminosity distribution function filtered by a sobel kernel to determine the luminosity of the trgb ; the edge detection function is : @xmath19 where @xmath20 is the photometric error , estimated as the mean of photometric errors shown in fig.[phot_error ] , within a bin of @xmath21 magnitude about a magnitude @xmath22 ( sakai et al .
the edge detection filter essentially measures the slope of the luminosity function , and hence it is very sensitive to the noise ; any sudden change in the signal is extremely amplified by the filtering scheme . to suppress any detection of statistically non - significant peaks in the first derivative of the luminosity function
, we apply a weighting scheme to the edge detection filtering output ( sakai et al . 1997 ; mendez et al . 2002 ) .
each @xmath23 is weighted by the poisson noise ( i.e.,@xmath24 ) .
cioni et al .
( 2000 ) have found the edge detection technique to be a biased estimator of the trgb magnitude . however , the bias depends on the photometric errors and the amount of smoothing in the luminosity function at the location of the trgb . for our data
set , the bias is less than @xmath25 magnitude ( cioni et al .
2000 ) , much smaller than the random errors ; therefore the distance modulus to the galaxy sample we derive using the edge detection technique would not be significantly affected in a comparison with the maximum likelihood ( see below ) . to estimate the statistical uncertainty in measuring the @xmath0-band trgb magnitude
, we use bootstrap resampling of the data to generate a large number of samples drawn from the original data set to simulate the act of observing multiple times . using this procedure to quantify the formal errors in measuring the trgb
has the advantage of using the empirical distribution function derived from the data itself , rather than a particular functional form ( e.g. , gaussian ) for the photometric errors . for each galaxy in our sample , we perform the bootstrap resampling and calculate the statistic of interest @xmath26 ^ 2 $ ] times , where @xmath27 is the number of observed stars within the magnitude range where the luminosity function model is valid in the case of maximum likelihood analysis , and the number of stars used to construct the luminosity function in the case of sobel edge detection analysis .
to this extent the bootstrap version of a statistic s sampling distribution matches the asymptotic sampling distribution ( see babu & feigelson 1996 and efron & tibshirani 1986 for reviews of the theory and application of this technique ) . in these bootstrap samples , the standard deviation of the i - band trgb distribution is taken as the random error in the measurement of the trgb magnitude .
this procedure allows a more robust estimate of the statistical errors than using the full - width at half maximum of the observed peak in the first luminosity function derivative , as usually done ( sakai et al .
1996 ; see also mendez et al .
2002 and cioni et al .
2001 ) .
an alternative to filtering is to fit models to the data using the technique of maximum likelihood ; this avoids concerns over binning and kernel smoothing .
while theory does not provide strong guidance on the exact functional form of the luminosity function near the trgb , a simple two - power - law model appears to provide a good empirical fit over a restricted range of magnitudes near the trgb itself .
the trgb itself is identified as the break between the two power laws . in this technique , the underlying luminosity function is treated as a probability distribution function ( pdf ) and we compare relative likelihoods of drawing the observed data set from the model pdfs as we vary the model parameters .
an advantage of this procedure is that no data binning is needed . in practice
we have binned the both the probability distributions and the data in intervals of 0.002 mag , as it simplifies the fitting procedure . given a model of the stellar distribution on the rgb , @xmath28k , where @xmath29 is the set of free parameters characterizing the model , the likelihood that the observed data set is drawn from this model
is given simply by the product of individual probabilities that star @xmath30 has an observed magnitude @xmath31 within a range @xmath32 : @xmath33 where @xmath34 is the individual probability defined as the ratio of the number of stars actually _ observed _ at a given magnitude by the total number of stars expected given a luminosity function model @xmath28 .
photometric errors will affect the magnitude distribution of the observed stars . to account for this , we convolve the intrinsic luminosity function models with an appropriate _ broadening _ function , @xmath35 , to obtain the distribution of _ observed _ magnitudes .
this function describes the probability that a star with an intrinsic magnitude @xmath36 is observed to have the magnitude @xmath37 .
thus the distribution of the observed magnitudes is : @xmath38 where @xmath39 is the intrinsic luminosity function model .
we represent the measurement error by a gaussian function with a total dispersion @xmath20 .
that is , we have @xmath40\ ] ] where @xmath41 is the photometric error at the magnitude @xmath42 .
the probability of observing a star of a magnitude @xmath31 in a magnitude range @xmath32 is simply @xmath43 .
the photometric errors smear the probability to the shape of the error function , centered at @xmath31 ; then the probability becomes @xmath44 .
the photometric errors were determined from artificial star tests .
we take the output of these simulations in 0.5 mag intervals for input stars of a fixed color v - i = 1.0 . in each 0.5 mag interval
, we determine the mean and standard deviation of the residuals of recovered minus input i - band magnitudes . for smearing the model luminosity functions , we use a set of gaussian kernels with the mean and standard deviation that were determined from the simulations and with an overall normalization set by the fractional completeness in that 0.5 mag interval .
the model luminosity functions are convolved with these kernels prior to estimating the logarithm of the likelihood .
the maximum likelihood fitting is restricted to the magnitude range where the effects of both the photometric errors and the incompleteness are not large . to avoid contamination from metal - rich stars that may bias the estimate of the trgb location
, we consider only blue stars , i.e. , stars within the color range @xmath45 ( in practice there are few stars bluer than @xmath46 ) .
ideally , it would be best to fit a curved trgb to a locus of rgb models of all metallicities .
however , in practice the resolution of existing isochrones near the trgb is insufficient for this purpose , and also the isochrones themselves have been constructed via a series of interpolations that become increasingly uncertain near the trgb . for this reason ,
the trgb distance indicator remains an empirical test : a search for a clear jump or inflection in the luminosity function . to approximate the intrinsic luminosity function @xmath39
, we use a broken power - law model ( cioni et al .
2000 ; mendez et al . 2002 ; gregg et al .
the free parameters characterizing the intrinsic model are the logarithm of the amplitude of the discontinuity at the trgb , the apparent magnitude of the trgb break , and the slopes of luminosity function on either side of the break .
we used both the nelder - mead ( 1965 ) simplex algorithm and `` simulated annealing '' ( e.g. , metropolis , 1953 ; press et al . 1992 ) to determine the best - fit model .
results of the two methods were similar , and well within the uncertainties derived from the subsequent error analysis . to determine the uncertainties in the trgb magnitude , we have stochastically sampled the full model parameter space by building a monte - carlo markov chain ( mcmc ) using the metropolis - hastings update mechanism .
there is a growing body of literature on this technique ( e.g. , knox , christensen & skordkis 2001 ; verde et al .
2003 ) , which provides an efficient way to characterize the likelihood manifolds in a multi - dimensional parameter space .
the metropolis - hastings mcmc algorithm is a restricted random walk through parameter space . at each step in the random
walk , magnitude and direction of the next step is drawn from a normalized `` proposal density '' @xmath47 , where the new state @xmath48 depends only on the current state @xmath49 . the candidate @xmath48 is accepted or rejected with a probability @xmath50 where @xmath51 is the likelihood ratio for state @xmath49 relative to state @xmath48 . after a period of `` burn - in , '' the distribution of locations in parameter space sampled by the chain converges to the `` posterior distribution '' @xmath52 .
thus likelihood contours can be constructed simply by drawing contours of the density of points sampled by the chain . for our proposal distribution
, we have used a gaussian with @xmath53 in all of the fit parameters .
we have also restricted the chains to operate within a `` reasonable '' range of parameter space : break amplitudes within a factor of 100 of the best fit , trgb magnitudes within 1.5 - 2 mag of the best fit , and lf slopes between @xmath54 and @xmath55 .
the chains were burned in for 1000 steps , and then 10000 steps were used to sample the posterior distributions .
the probability distributions thus derived are shown in the lower - right panels of figures 3 - 6 .
we have verified that the results are insensitive to small changes in the range of magnitudes being fit and the parameters used to define the mcmc .
the confidence intervals ( and indeed the best fit values ) are nevertheless subject to a number of implicit prior assumptions .
these include the assumption that the distribution can be adequately described by two power laws with a break , over the full magnitude range being fit . in practice
, we do see small but significant differences in the confidence intervals as we change the magnitude limits and color limits
. however shifts of 0.2 mag in these limits in all of our tests shift the peak of the likelihood distribution by less than the @xmath56 widths show in figs .
as a first step to investigate the properties of field stellar populations in spiral galaxy halos , we have determined distances to a sample of galaxies . a summary of trgb distances and uncertainties is presented in table [ gal_dist ] .
color - magnitude diagrams , luminosity functions , and other properties for our galaxy sample are presented in the next subsections . the best fitting luminosity function model , @xmath57 , where @xmath58 is the free parameter set that maximize the likelihood , is overplotted on the observed luminosity function for each galaxy .
note that the smoothing by the photometric errors , accounted for in the maximum likelihood analysis , converts the abrupt break at the trgb magnitude into a continuous rgb star distribution near the trgb .
once the i - band magnitude of the trgb is measured , we use the semi - empirical calibration given by lee et al .
( 1993 ) to determine the distance modulus , starting from the relation : @xmath59 the bolometric magnitude , m@xmath60 , and the bolometric correction , bc@xmath61 , are function of the metallicity and the color of rgb stars ( da costa & armandroff 1990 ) : @xmath62 - 3.81\ ] ] @xmath63 where @xmath64 is the color of stars at the trgb magnitude .
the metallicity [ fe / h ] is function of the rgb star colors ( lee et al .
1993 ) : @xmath65=-12.65 + 12.6\times(v - i)_{-3.5}-3.3\times(v - i)^{2}_{-3.5}\ ] ] where @xmath66 is measured at an absolute i - band magnitude of @xmath67 mag . to measure @xmath64 ,
we calculate the color distribution of stars within @xmath68 , and fit a gaussian to the distribution .
a similar procedure was repeated for the simulated bootstrap stars to measure the uncertainties , and a similar procedure was adopted to measure @xmath66 .
the absolute i - band magnitudes are given for each galaxy in table [ gal_dist ] .
we assume that internal extinction is negligible as we are dealing with halo regions that are most likely dust free . to correct for the foreground extinction
, we use the 100@xmath69 m dirbe / iras all - sky map of schlegel et al .
( 1998 ) .
ngc253 is a well studied nearly edge - on sc - type l@xmath70 galaxy in the sculptor group ; it is an archetypal starburst galaxy , and is the closest edge - on infrared luminous galaxy ( radovich et al . 2001 ) . extended soft x - ray emission from a hot gaseous component was detected in the halo of ngc253 .
this x - ray emission seems to be powered by a large - scale outflow , driven by the nuclear starburst activity ( pietsch et al .
2000 ; strickland et al .
the left panel of fig.[ngc253 ] displays the color - magnitude diagram of the observed halo field .
the stars in the halo are predominantly red ; the main feature of the color - magnitude diagram is the rgb structure , which finishes in a well - defined trgb .
the red giant branch is prominent and wide , indicating a large spread in the metallicity distribution of the halo stellar population .
the stars above the trgb and extending up to @xmath71 may be asymptotic giant branch ( agb ) stars covering a range of ages and metallicities .
the color - magnitude diagram shows no indication of the presence of blue early - type main sequence stars in the observed halo field of ngc253 , as was reported previously by comern et al .
the reason for this may be that the small number of blue stars detected toward the halo of ngc 253 are distributed over a large field .
wide field and deep imaging are needed to investigate in detail the star formation history in the halo of ngc253 , and its potential link with the observed galactic outflow . in the panel
next to the color magnitude diagram , fig .
[ ngc253 ] shows the results of the edge detection filtering of the i - band luminosity function .
the right panel of fig.[ngc253 ] shows the logarithmic luminosity function .
the power - law distribution of the upper rgb stars is obvious , and has a clear break at @xmath72mag .
overplotted are the maximum - likelihood best fit model before and after convolution with the photometric errors and application of incompleteness , shown as dashed and solid lines respectively .
the bottom panel shows the posterior probability distribution of the trgb magnitude derived from the monte - carlo markov chain analysis .
also shown are the @xmath73 and @xmath74 confidence intervals .
both the sobel edge detector and the maximum likelihood model agree on the location of the i - band trgb magnitude . for the maximum likelihood analysis , with a foreground reddening of e(b - v)=0.02mag and fitting data in the range @xmath75 , we derive that the trgb is located at @xmath76 mag .
the monte - carlo markov chain analysis gives that the @xmath73 confidence interval spans the range @xmath77 and the @xmath74 confidence interval is @xmath78 .
the faint extention of the @xmath74 confidence level is quite large , when the trgb is well defined , as there are a variety of models with different amplitudes for the jump and slopes above and below the trgb that are consistent with the data to within the @xmath74 confidence interval .
there is a family of solutions with a steep powerlaw at the bright end and a much smaller jump than the best - fit shown in the plot .
the edge detection method locates the trgb i - band magnitude at @xmath79 mag with a bootstrap uncertainty of @xmath80mag . following the procedure described in
[ dist ] , we find a population ii distance modulus @xmath81(random)@xmath82(systematic ) . this result is consistent with ngc 253 being a member of the nearby sculptor group , located at the far side of the group ; however , it is significantly larger than previously reported values derived using the globular cluster luminosity function ( blecha 1986 ) , and the brightest halo ( davidge & pritchet 1990 ) and disk ( davidge , le fvre , & clark 1991 ) stars . the edge - on galaxy ngc 4244 is a late - type ( scd ) spiral ; its total b - band magnitude is almost a magnitude below the peak of the field sc luminosity function , making this galaxy less massive than a typical sc .
the radial light distribution is well represented by an exponential profile ( olling 1996 ) , with almost no dark lanes ( i.e. , internal extinction is not important ) .
the color magnitude diagram for the observed halo field , shown in the left panel of fig.[ngc4244 ] , reveals a simple stellar population ; the red giant branch is prominent and clearly dominates the halo light , and there seems to be no blue early - type stellar population in the halo of this galaxy .
the red giant branch is relatively tight , indicating a tight metallicity distribution in the halo stars , with a clear tip at @xmath83 mag .
the stars above the tip are likely agb stars .
also shown is the output of the edge detection filtering of the i - band luminosity function .
the upper right panel of fig.[ngc4244 ] shows the logarithmic luminosity function for all stars detected in the field . for ngc 4244
, the trgb appears as an inflection in slope , rather than a clear discontinuity .
the power - law luminosity distribution of stars on the red giant branch is clear down to the magnitude where the effect of incompleteness starts to be important ; the break of the luminosity function due to the trgb , at @xmath83 , is obvious . again both the edge detection method and the maximum likelihood model agree on the location of the trgb i - band magnitude . using the maximum likelihood analysis , and fitting data in the range @xmath84 , we find that the trgb is located at @xmath85 mag .
the monte - carlo markov chain analysis gives a @xmath73 confidence interval of @xmath86 and a @xmath74 confidence interval of @xmath87 .
the edge detection method locates the trgb at @xmath88 mag .
resampling simulations give uncertainties of @xmath89 mag for the edge detection method . with a foreground extinction of @xmath90mag
, we find the distance modulus to ngc 4244 to be @xmath91(random)@xmath82(systematic ) , corresponding to a heliocentric metric distance of @xmath92mpc .
our distance modulus is in agreement with the previously published distance modulus for ngc 4244 of @xmath93 using the tully - fisher relation ( aaronson et al .
1986 ) , but significantly different from the distance derived by karachentsev & drozdovsky ( 1998 ) , who used the mean b - band magnitude of blue supergiant candidates to derive a distance modulus of @xmath94 .
this method is affected by large uncertainties ( rozanski & rowan - robinson 1994 and references therein ) . a preprint by tikhonov & galazutdinova ( 2005 ) analyzing a larger data set on this galaxy , including the field analysed in this paper , appeared after our paper was submitted .
their derived distance modulus using the tip of the red giant branch technique is 28.16 , which is consistent within our 2@xmath95 confidence interval but not our 1@xmath95 confidence interval .
the magnitude of the tip of the red giant branch derived by tikhonov & galazutdinova ( 2005 ) is fainter than what is estimated here ( see fig . 3 of paper iii of this series , and fig . 4 of tikhonov & galazutdinova 2005 ) . on the other hand ,
their estimated mean stellar metallicity is lower than our estimate ( see paper ii of this series ) .
tikhonov & galazutdinova ( 2005 ) have not well documented their data reduction and the procedure they used to measure the distance modulus to ngc 4244 to trace back accurately the source of the observed discrepancy .
ngc4945 is an edge - on sb galaxy , hosting a powerful nuclear starburst region , with a ring morphology ( moorwood et al .
1996 ) , powering a galactic superwind ( heckman , armus , & miley 1990 ) .
similar to ngc 5128 , the prominent galaxy of the group to which ngc 4945 is believed to belong , its optical image is marked by dust extinction in the nuclear regions ( marconi et al .
2000 ; lipari et al .
1997 ) .
the left panel of fig.[ngc4945 ] shows the color - magnitude diagram of the observed halo field stellar population .
the red giant branch is wide and well populated , indicating a large spread of the halo stellar metallicities .
also shown is the output of the edge detection filtering of the i - band luminosity function .
the right panel of fig.[ngc4945 ] shows the logarithmic i - band luminosity function , compared to the maximum - likelihood best model derived by fitting the data in the range @xmath96 .
again the upper red giant branch power law distribution is evident with a break around @xmath97 .
the edge detection method and the maximum likelihood model disagree slightly on the location of the i - band trgb . using the maximum likelihood technique to fit the stellar luminosity function in the range @xmath96
, we locate the trgb magnitude at @xmath98 mag , with a @xmath73 confidence interval of @xmath99 and a @xmath74 confidence interval of @xmath100 .
the edge detection method gives a brighter location of the trgb at @xmath101 mag , with a bootstrap uncertainty of @xmath102mag .
the average location of the trgb is @xmath103 mag . using this value with a foreground reddening of e(b - v)=0.18mag
, we find the distance modulus for this galaxy to be @xmath104(random)@xmath82(systematic ) .
this result is consistent with ngc 4945 being member of centaurus a group , which has a centroid at a distance modulus of @xmath105 ( @xmath106mpc ; karachentsev et al .
2002 ) .
ngc 4258 is a nearby barred spiral galaxy hosting an obscured active nucleus and a massive nuclear black hole inside a highly inclined thin gaseous disk ( miyoshi et al .
1995 ; maoz 1995 ) .
nuclear water maser sources were discovered moving in a keplerian orbit around the central black hole of ngc 4258 ( watson & wallin 1994 ) . by resolving these masers , measuring their acceleration ( greenhill et al .
1995 ) , and monitoring their proper motions , herrnstein et al .
( 1999 ) derived a distance modulus to the galaxy of @xmath107 .
ngc 4258 is one of the only two galaxies with a geometrical distance estimate , the other galaxy being the large magellanic cloud , with a geometrical distance measurement based on the light echo of sn 1987a ( panagia et al .
thus ngc 4258 is a unique object for a cross - check of extragalactic distance techniques .
recently , newman et al . ( 2001 ) used a sample of cepheids in ngc 4258 to report the distance modulus to this galaxy is @xmath108 mag , with a distance modulus to the lmc of @xmath109 mag .
to do so , they used revised calibrations and methods for the hst key project on the extragalactic distance scale .
this distance is @xmath110 larger the the maser determination . to reconcile the geometrical and the cepheid distances to ngc 4258 , they argue that the true distance modulus to the lmc should be @xmath111(random)@xmath112(systematic ) , and should be used to calibrate the cepheid brightness . using the planetary nebulae luminosity function technique , ciardullo et al . (
2002 ) derived a distance modulus to ngc 4258 of @xmath113 .
similar to newman et al ( 2001 ) , they argue that the short distance scale to the lmc should be the correct one . on other hand ,
caputo et al . (
2002 ) argue that the theoretical metallicity correction , as suggested by pulsation models , is @xmath114 mag lower than the empirical correction adopted by newman et al .
then if this theoretical correction is applied to account for the chemical abundance differences between the lmc and ngc 4258 , it may reconcile the cepheid and the maser distances to ngc 4258 , with the _ long distance scale _ to the lmc , a scale supported by different empirical non - cepheid methods ( panagia et al .
1991 ; sakai et al .
2000 ; cioni et al .
2000 ; groenewegen & salaris 2001 ) .
this debate raises again the well - known controversy between the short and long distance scales to the lmc .
we take advantage of our halo field observations to estimate the distance modulus to ngc 4258 using the trgb technique .
provided we restrict the ourselves to @xmath115 , this method is relatively insensitive to metallicity .
the trgb luminosity calibration does not depend on the assumed distance to the lmc , but rather is computed from theory and tested and calibrated via observations of globular clusters in our galaxy . because ngc 4258 is metal rich and the lmc is relatively metal poor , the trgb distance to ngc 4258 is an important cross - check of the metallicity effect on the cepheid - based extragalactic distance scale .
the left panel of fig.[ngc4258 ] shows the color - magnitude diagram of the observed halo field stellar population . again
the halo stellar population is dominated by red giant branch stars .
the red giant branch is wide , indicating a large spread of the stellar metallicity in the halo .
also plotted is the output of the edge detection filtering of the i - band luminosity function .
the right panel of fig.[ngc4258 ] shows the logarithmic luminosity function for all stars observed in the field , compared to the maximum likelihood best fit model .
the bottom panel shows the posterior probability distribution of the trgb magnitude derived from the monte - carlo markov chain analysis .
again both techniques agree on the location of the i - band trgb magnitude .
for the maximum likelihood analysis , with a foreground foreground of e(b - v)=0.02mag and fitting data in the range @xmath116 , we locate the trgb at @xmath117 mag .
the monte - carlo markov chain analysis gives that the @xmath73 and @xmath74 confidence intervals are @xmath118 and @xmath119 , respectively .
the edge detection technique locates the trgb at @xmath120 mag with a bootstrap uncertainty of @xmath121 mag . following the procedure described in
[ dist ] , we find a population ii distance modulus for this galaxy to be @xmath122(random)@xmath123(systematic ) , or a metric distance of d=@xmath124 mpc .
this is in superb agreement with both the maser and the metallicity - corrected cepheid distances to the galaxy .
our distance modulus argue for the _ long distance scale _ to the lmc , where the true distance modulus to the lmc is close to @xmath125 mag .
the results provide support for hubble constant @xmath126 km s@xmath127 mpc@xmath127 , inferred by freedman et al .
( 2001 ) and based on the long distance scale to the lmc , in contrast to other recent suggestions ( maoz et al . 1999 ; ciardullo et al .
we report distance measurements using data obtained from a new program of deep hst imaging of the halos of nearby bright , edge - on spiral galaxies .
we have used the trgb as a distance indicator for four nearby galaxies , ngc 253 , ngc 4244 , ngc 4945 , and ngc 4258 , observed with hst / wfpc2 .
we employ two methods , edge detection and maximum likelihood fitting , to measure the trgb i - band magnitude from the luminosity function of rgb stars .
our results are in good agreement with prior distance measurements based on a variety of methods , including cepheid variables .
however , we now have a common method of distance measurement for these galaxies .
the uncertainty in the estimate of the luminosity function break is calculated using the bootstrap resampling technique , which makes no assumption about the underlying distributions .
using our trgb distance to ngc 4258 , in combination with the galaxy s geometric and new metallicity - corrected distances , we argue that the long distance scale to the large magellanic cloud is correct , and that the freedman et al .
( 2001 ) hubble constant should not be revised .
finally , increasing the number of galaxies beyond the local group with accurate trgb measurements will certainly refine our understanding of the extragalactic distance scale , and will help making the systematic comparison between population i and population ii distance indicators more meaningful . m.m would like to thank bryan mndez for useful discussions .
we acknowledge grants under hst - go-9086 awarded by the space telescope science institute , which is operated by the association of the universities for research in astronomy , inc .
, for nasa under contract nas 5 - 26555 aaronson , m. , bothun , g. , mould , j. , huchra , j. , schommer , r.a . , cornell , m.e .
, 1986 , , 302 , 536 aguilar , l. , hut , p. , & ostriker , j.p .
, 1988 , , 335 , 720 babu , g.j . , &
feigelson , e.d . , ed .
1996 , astrostatistics interdisciplinary statistics blecha , a. , 1986 , , 154 , 321 comern , f. , torra , j. , mndez , r.a . ,
gmez , a.e . , 2001 , , 366 , 796 caputo , f. , marconi , m. , musella , i. , 2002 , , 566 , 833 carney , b.w .
, 1993 , in the globular cluster - galaxy connection , eds . smith , g.h .
, & brodie j.p . , asp conf .
234 chiosi , c. , bertelli , g. , & bressan , a. , 1992 , , 30 , 235 ciardullo , r. , et al . , 2002 , , 577 , 31 cioni , m .- r.l , van den marel , r.p . , loup , c. , & habing , h.j . , 2000 , , 359 , 601 da costa , g.s . , & armandroff , t.e . , 1990 , , 100 , 162 davidge , t.j . ,
& pritchet , c.j . , 1990 , , 100 , 102 davidge , t.j .
, le fvre , o. , & clark , c.c . , 1991 , , 370 , 559 dohm - palmer , r.c . , helmi , a. , marrison , h. , et al . , 2001 , , 555 , l37 dolphin , a.e
. , 2000 , , 112 , 1397 durrell , p.r . , harris , w.e . , & pritchet , c.j . , 2004 , , 128 , 260 efron , b. , & tibshirani , r. , 1986 , stat
sci , 1 , 54 eggen , o. , lynden - bell , d. , & sandage a. , 1962 , , 136 , 748 ferrarese , l. , mould , j.r .
, kennicutt , r.c . , jr . , et al . ,
2000 , , 128 , 431 freedman w.l .
2001 , , 553 , 47 freeman , k.c . , 1987 , , 25 , 603 freeman , k.c . , & bland - hawthorn , j. , 2002 , , 40 , 487 fruchter , a.s . , & hook , r.n . , 2002 , , 114 , 144 greenhill , l.j . , jiang , d.r . ,
moran , j.m . ,
reid , m.j . , lo , k.y . , claussen , m.j . , 1995 , , 440 ,
619 gregg , m.d . , ferguson , h.c .
, minniti , d. , tanvir , a. , catchpole , r. , 2004 , , 127 , 1441 grillmair , c.j . ,
freeman , k.c . , irwin , m. , quinn , p.j . , 1995 , , 109 , 2553 groenewegen , m.a.t . ,
salaris , m. , 2001 , . , 366 , 752 heckman t.m . , armus l. , & miley g.k . , 1990 , , 74 , 833 herrnstein j.r . ,
et al . , 1999 , nature , 400 , 539 ho , l.c . , filippenko a.v .
, sargent w.w , 1996 , , 462 , 183 holtzman j.a .
et al . , 1995 ,
pasp , 107 , 1065 ibata , r.a . ,
gilmore , g. , & irwin m.j .
, 1994 , nature , 370 , 194 ibata , r.a .
, irwin m.j . , lewis , g. , ferguson , a.m.n . , & tanvir , n. , 2001 , nature , 412 , 49 ivezic , z. , goldston , j. , finlator , k. , et al . ,
2000 , , 120 , 963 karachentsev , i.d .
, & drozdovsky , i.o . , 1998 , , 131 , 1 karachentsev , i.d . ,
sharina , m.e . , dolphin , a.e . , et al . , 2002 , , 385 , 21 knox , l. , christensen , n. & skordis , c. , 2001 , apj , 563 , l95 krist , j. , 2004 , tinytim manual , http://www.stsci.edu/software/tinytim/tinytim.pdf lee , m.g . ,
freedman , w.l .
, & madore , b.f .
1993 , , 417 , 553 lipari , s. , tsvetanov , z. , macchetto , f. , 1997 , , 111 , 369 madore , b.f . , &
freedman , w.l .
1995 , , 109 , 1645 marconi , a. , oliva , e. , van der werf , p.p .
, maiolino , r. , schreier , e.j .
, macchetto , f. , moorwood , a.f.m .
, 2000 , , 357 , 24 mendez , b. , davis , m. , moustakas , j. , newman , j. , madore , b.f . , &
freedman w.l . , 2002 , , 124 , 213 miyoshi , m. , moran , j. , herrnstein , j. , greenhill , l. , nakai , n. , diamond , p. , & inoue , m. , 1995 , nature , 373 , 127 maoz , e. , 1995 , , 455 , 131 maoz d. , et al . , 1999 ,
nature , 401 , 351 metropolis , n. , rosenbluth , a. w. , rosenbluth , m. n. , tellar , a. h. , & tellar , e. , 1953 , journal of chemical physics , 21 , 6 .
moorwood , a.f.m .
1996 , , 308 , l1 morrison , h.l . , 1993 , , 106 , 578 nelder , j. a. & mead , r. , 1965 , computer journal , 7 , 308 newman j.a . , et al . , 2001 , , 553 , 562 olling , r.p . , 1996 , , 112 , 481 panagia , n. , gilmozzi , r. , macchetto , f. , adorf , h .-
m . , & kirshner , r.p .
, 1991 , , 380 , l23 perrett , k.m . , bridges , t.j . ,
hanes , d.a . , et al . , 2002 , , 123 , 2490 pietsch , w. , vogler , a. , klein , u. , & zinnecker , h. , 2000 , , 360 , 24 press , w.h . ,
teukolsky , s.a . , vetterling , w.t . & flannery , b.p .
, 1992 , numerical recipes in c , cambridge university press , cambridge .
puche , a. , & carignan , c. , 1988 , , 95 , 1025 radovich , m. , kahanpaa , j. , lemke , d. , 2001 , , 377 , 73 renzin , a. , 1992 , in iau symp .
149 , the stellar populations in galaxies , ed .
b. barbuy & a. renzini ( dordrecht : kluwer ) 235 rozanski r. , & rowan - robinson , m. , 1994 , , 271 , 530 sakai , s. , madore , b.f .
, & freedman , w.l .
1996 , , 461 , 713 sakai , s. , madore , b.f . , freedmen , w.l . ,
lauer , t.r .
, ajhar , e.a . , & baum , w.a .
, 1997 , , 478 , 49 sakai , s. , zaritsky , d. , kennicutt , r.c . , jr . , 2000 , , 119 , 1197 salaris , m. , & cassisi , s. , 1997 , , 289 , 406 salaris , m. , & girardi , l. , 2005 , , 357 , 669 searle , l. , & zinn r. , 1978 , , 225 , 357 schlegel , d.j .
, finkbeiner , d.p . , &
davis , m. , 1998 , , 500 , 525 strickland , d.k . , heckman , t.m . , weaver , k.a . , hoopes , c.g .
, dahlem , m. , 2002 , , 568 , 689 tikhonov , n.a . , galazutdinova , o.a . , & drozdovsky , i.o . , 2005 , , 431 , 127 .
tikhonov , n.a . , & galazutdinova , o.a . , 2005 ,
astrofizica , in press , astro - ph/0503235 .
watson , w.d .
, & wallin , b.k .
, 1994 , , 437 , l35 verde , l. , et al . , 2003 ,
apjs , 148 , 195 .
yanny , b. , newberg , h.j . ,
kent , s. , et al . , 2000 , , 540 , 825 zoccali , m. , & piotto , g. , 2000 , , 358 , 943 ngc 4258 & @xmath128 & maser @xmath129 & @xmath130 & @xmath131 & + & @xmath132 & cepheids + z - correction@xmath133 & & & + & @xmath134 & cepheids@xmath135 & & & + & @xmath136 & pnlf@xmath137 & & & + ngc 253 & 27.30 & globulars lf@xmath138 & @xmath139 & @xmath140 & + & @xmath141 & brightest halo and disk stars@xmath142 & & & + ngc 4244 & 28.28 & brightest supergiants@xmath143 & @xmath144 & @xmath145 & + & 27.78 & tully - fisher relation@xmath146 & & & + ngc 4945 & 27.82 & cen - a group centroid distance@xmath147 & @xmath148 & @xmath149 & + [ gal_dist ] | we have imaged the halo populations of a sample of nearby spiral galaxies using the wide field planetary camera 2 on broad the hubble space telescope with the aim of studying the stellar population properties and relating them to those of the host galaxies . in four galaxies ,
the red - giant branch is sufficiently well populated to measure the magnitude of the tip of the red - giant branch ( trgb ) , a well - known distance indicator .
using both the sobel edge - detection technique and maximum - likelihood analysis to measure the @xmath0-band magnitude of the red giant branch tip , we determine distances to four nearby galaxies : ngc 253 , ngc 4244 , and ngc 4945 , ngc 4258 .
for the first three galaxies , the trgb distance determined here is more direct , and likely to be more accurate , than previous distance estimates . in the case of ngc 4258
, our trgb distance is in good agreement with the the geometrical maser distance , supporting the the large magellanic cloud distance modulus @xmath1 that is generally adopted in recent estimates of the hubble constant . |
since the discovery of the giant dipole resonance ( gdr ) as a collective thermal excitation in highly excited ( hot ) nuclei @xcite , many experiments were carried out in the last three decades to extract the gdr width and its shape ( linearized strength function ) as functions of nuclear temperature @xmath1 and angular momentum @xmath2 . a recent compilation of gdr built on excited states is given in ref .
@xcite . at present
, the well - established systematics accumulated by measuring the @xmath3 decays of various hot compound nuclei formed in heavy - ion fusion reactions and inelastic scattering of light particles on heavy targets has shown that the gdr width increases with temperature @xmath1 within the temperature region 1 mev @xmath4 3 4 mev .
it has also shown that the gdr width s increase with angular momentum @xmath2 becomes noticeable only at @xmath5 27 30 @xmath6 in heavy nuclei , whereas its location ( peak energy ) remains mostly unchanged as @xmath1 and @xmath2 vary .
experimental studies often refer to the thermal shape fluctuation model ( tsfm ) @xcite as one of theoretical descriptions of the width s increase in this temperature region .
the tsfm takes the thermal average of the gdr photoabsorption cross section over the shape - dependent cross sections caused by all fluctuating quadrupole shapes , which are assumed to be coupled to the gdr vibration .
the width s increase as a function of @xmath1 arises as a results of such thermal average .
the tsfm , however , fails to describe the temperature dependence of the gdr width beyond the temperature region 1.5 @xmath7 3 mev . at @xmath8 3
4 mev several experimental evidences have shown that the gdr width seems to saturate at high @xmath1 @xcite , whereas the tsfm predicts a continuously increasing width .
in the low temperature region , at @xmath9 1 mev , a measurement of @xmath3 decays in coincidence with @xmath10o particles scattered inelastically from @xmath11sn @xcite has obtained a gdr width in @xmath11sn of around 4 mev at @xmath12 1 mev , that is smaller than its value of 4.9 mev at @xmath12 0 .
this result and the existing systematics for the gdr width in @xmath11sn up to @xmath13 1.5 mev are significantly lower than the prediction by the tsfm .
meanwhile , the gdr width as a function of temperature @xmath1 is well described by the phonon damping model ( pdm ) @xcite in the entire region 0@xmath4 5 - 6 mev , including the increase in the width at @xmath9 3 mev as well as the width saturation at high @xmath1 . within the pdm ,
the damping of gdr at @xmath14 0 is caused by coupling of the gdr to noncollective particle - hole ( @xmath15 ) and particle - particle ( @xmath16 ) [ hole - hole ( @xmath17 ) ] configurations .
the coupling to @xmath15 configurations exists even at @xmath18 0 , and leads to the quantal width @xmath19 , whereas the thermal width @xmath20 arises owing to coupling to @xmath16 and @xmath17 configurations , which appear only at @xmath14 0 because of the distortion of the fermi surface at @xmath14 0 . in the low temperature region
@xmath9 1 mev , it has been shown within the pdm that thermal pairing plays a crucial role in reducing the gdr width in @xmath11sn @xcite . as a matter of fact , in finite systems such as nuclei , the pairing gap does not collapse at the critical temperature @xmath21 0.57@xmath22 [ @xmath22 being the pairing gap at @xmath18 0 ] as in the case of the superfluid - normal phase transition in infinite systems , but decreases monotonically as @xmath1 increases .
this decrease of pairing tends to restore the fermi surface , which is diffused in the presence of pairing , back to the sharp step - function distribution .
this competes with the thermal smoothing of the fermi surface , which increases with @xmath1 . as a result of such competition
, a compensation takes place , which leaves the gdr width almost unchanged or even decreases slightly at @xmath9 1 mev . at @xmath23 2 mev ,
pairing becomes significantly smaller than its value at @xmath18 0 , so the thermal distortion of the fermi surface becomes dominant and the width starts to increase .
very recently , @xmath24 induced fusion reactions were used to measure the gdr width at low temperature @xcite .
these reactions can describe temperature more precisely whereas the associated angular momentum in the mass region @xmath25 115 121 is rather small ( @xmath26 24@xmath6 ) .
the data extracted from these latest experiments for the gdr width in @xmath27sb at 0.98 @xmath4 1.23 mev @xcite are similar to those obtained previously for @xmath11sn , including the data point at @xmath18 1 mev mentioned above , in good agreement with the prediction by the pdm for the gdr width in @xmath11sn when thermal pairing is included . however , these experiments also reported the data for the gdr width in @xmath0tl @xcite , which were extracted within the temperature interval 0.8 @xmath28 1.2 mev .
these values are significantly smaller than the prediction by the tsfm for the gdr width in @xmath29pb even after including the shell effect .
the authors of ref .
@xcite also made a comparison with the prediction by the pdm using the results for the gdr width in @xmath29pb .
however @xmath29pb is a doubly closed shell nucleus , that is the neutron and proton pairing gaps are both zero , and , naturally , no pairing was taken into account in the pdm prediction for the gdr width in this nucleus , whereas @xmath0tl is an open - shell nucleus for both neutrons ( n=120 ) and protons ( z=81 ) . therefore
, an adequate comparison should be made with the prediction within the pdm for the gdr width of the same @xmath0tl nucleus , including the effects owing to thermal pairing of neutrons as well as protons .
the aim of the present paper is to make such prediction .
as has been mentioned above , the conventional finite - temperature ( ft ) bcs theory , which exhibits a collapse of the pairing gap at a critical temperature @xmath30 , should be modified to include thermal fluctuations when it is applied to finite nuclei . among such modifications
are the modified bcs ( mbcs ) @xcite , the finite - temperature bcs1 ( ftbcs1 ) as well as the lipkin - nogami projected ftbcs1 ( the so - called ftln1 ) @xcite .
however , these approaches fail for a close - to - magic nucleus such as @xmath0tl ( z=81 ) .
therefore , in the present paper , we employ the exact treatment of thermal pairing within the canonical ensemble ( ce ) , which has been elaborated in ref . @xcite . within this approach ,
the thermal pairing gaps are calculated from the exact pairing energy , which is obtained by averaging the exact eigenvalues of the pairing problem in the ce at temperature @xmath1 .
this approach also allows us to calculate the exact single - particle occupation numbers , chemical potentials , as well as the exact quasiparticle energies . by using the latter
, one can determine the quantities that approximate the coefficients @xmath31 and @xmath32 of the bogolyubov s transformation from particles to quasiparticles as well as the corresponding quasiparticle occupation numbers and use them as ingredients to calculate the gdr width as a function of @xmath1 in @xmath0tl within the quasiparticle representation of the pdm .
because the recent data for the gdr width in @xmath0tl at low @xmath1 were obtained at low angular momentum ( below 25@xmath6 ) , the effect of angular momentum on the gdr width is negligible .
therefore , for simplicity , angular momentum is not included in the present calculations ( see ref .
@xcite for the recent extension of pdm to finite angular momentum . ) .
the paper is organized as follows .
the formalism is derived in sec .
[ formalism ] .
the results of numerical calculations are analyzed and compared with the experimental data in sec .
[ results ] .
the paper is summarized in the last section , where conclusions are drawn .
because the pdm has been discussed in great details in a series of papers @xcite , we summarize below only the main results necessary for the numerical calculations in the present paper .
the pdm considers a model hamiltonian , which consists of three terms .
the first term describes the independent single - particle ( quasiparticle ) field with single - particle ( quasiparticle ) energies @xmath33 ( @xmath34 ) , the second term stands for the phonon field with phonon energies @xmath35 , whereas last term treats the coupling between these two fields [ see eq .
( 1 ) in ref .
@xcite for example . ] .
this coupling causes the damping of the phonon vibrations , for example , the gdr phonon ( @xmath36 ) . as a result ,
the gdr acquires a width .
the expression of the gdr full width at half maximum ( fhwm ) @xmath37 is obtained within the pdm as @xmath38 where @xmath39 is the phonon damping at the gdr peak energy @xmath40 . in the presence of superfluid pairing , the phonon damping
@xmath41 has the explicit form as @xcite @xmath42^{2}(1-n_{p}-n_{h } ) \delta(\omega - e_{p}-e_{h})+{\cal f}_2 ^ 2\sum_{ss'}[v_{ss'}^{(-)}]^{2}(n_{s'}-n_{s } ) \delta(\omega - e_{s}+e_{s'})\bigg\}~. \label{gamma}\ ] ] in this expression , @xmath43 stands for @xmath44 and @xmath45 with @xmath46 and @xmath47 denoting the particle ( @xmath46 ) levels , that is those situated above the fermi level at @xmath18 0 and @xmath48 0 , and hole ( @xmath47 ) levels , that is those situated below it . @xmath49 and @xmath50 are the model parameters for @xmath15 and @xmath16 ( @xmath17 ) couplings , respectively .
functions @xmath51 and @xmath52 are combinations of the bogolyubov s coefficients @xmath53 , @xmath54 , namely @xmath55 , and @xmath56 .
the first sum at the right - hand side of eq .
( [ gamma ] ) with the factors @xmath57 represents quantal damping caused by coupling of the gdr phonon to noncollective @xmath15 configurations , whereas the second sum with the factors @xmath58 stands for thermal damping , which arises from coupling of the gdr phonon to @xmath16 and @xmath17 configurations . because @xmath59 0 , the second sum over @xmath43 is finite ( and positive ) only at @xmath60 .
the quasiparticle occupation number @xmath61 ( @xmath62 ) is given in the form of a fermi - dirac distribution @xmath63^{-1 } , \label{nk}\ ] ] smoothed with a breit - wigner - like kernel , whose width is equal to the quasiparticle damping with the quasiparticle energy @xmath64^{2}+\delta(t)^{2}}~ , \label{ek}\ ] ] where @xmath65 and @xmath66 denote the temperature - dependent chemical potential and pairing gap , respectively [ see eq .
( 2 ) of ref .
] . for the gdr in medium and heavy nuclei ,
the quasiparticle damping is usually small , therefore the breit - wigner - like kernel can be safely replaced with the @xmath67-function . as a result , the quasiparticle occupation number @xmath61 has the form of the fermi - dirac distribution ( [ nk ] ) .
this also means that thermal damping of the gdr appears only at @xmath14 0 because all @xmath68 vanish at @xmath18 0 , whereas quantal damping exists already at @xmath18 0 , when the factors @xmath69 1 , as well as at @xmath14 0 . in closed - shell nuclei ,
the pairing gap @xmath66 is zero , so we have @xmath70 , @xmath71 , and @xmath72 , where @xmath73 are the single - particle occupation numbers . correspondingly , the sums and differences of quasiparticle energies become @xmath74 , @xmath75 , and @xmath76 . as a result ,
( [ gamma ] ) reduces to the expression for the phonon damping for the case without pairing @xcite , that is @xmath77^{2}(f_k - f_{k'})\delta(\omega-\epsilon_{k'}+\epsilon_k)~ , \label{gamma_nopair}\ ] ] with @xmath78 and @xmath79 , where @xmath80 denote the corresponding values of parameters @xmath81 in the zero - pairing case ( @xmath82 ) . as has been demonstrated in ref .
@xcite ( see the discussion below fig . 3 therein ) , the factors @xmath83 ^ 2(1-n_p - n_h)$ ] cause a slight decrease in the quantal width as @xmath1 increases , whereas the factors @xmath84 ^ 2(n_{s'}-n_s)$ ] are responsible for the strong increase in the thermal width with @xmath1 at low and moderate @xmath1 , and its saturation at high @xmath1 .
the combined effect makes the total width increase with @xmath1 at low and moderate @xmath1 and saturate at high @xmath1 . at very low @xmath1
( @xmath9 1 mev ) pairing effect leaves the total width almost unchanged with @xmath1 or , in some cases , makes it even smaller than the value at @xmath18 0 , as has been discussed in the introduction .
the escape width @xmath85 , which arises because of coupling to the continuum and is related to the direct decay by particle emission , is usually small ( few hundred kev ) , and does not seem to the change with @xmath1 in medium and heavy nuclei . in the numerical calculations within the pdm
, its effect is taken into account via the smoothing parameter @xmath86 , which replaces the @xmath67-functions in eqs .
( [ gamma ] ) and ( [ gamma_nopair ] ) with @xmath87 $ ] , where @xmath88 and @xmath89 in eq .
( [ gamma ] ) , and @xmath90 in eq .
( [ gamma_nopair ] ) .
a value of @xmath91 0.5 mev is adopted in the present calculations .
the results do not change significantly with @xmath86 varying between 0.5 and 1 mev .
the pdm does not take into account the evaporation width @xmath92 of the compound nuclear states , which comes from the quantum mechanical uncertainty principle @xcite because its effect on the gdr width is expected to be significant only at high @xmath1 ( @xmath93 3.3 mev ) and @xmath2 ( @xmath93 30@xmath6 ) .
the gdr strength function @xmath94 is calculated as @xmath95 ^ 2+\gamma_q^2(\omega)}~. \label{s}\ ] ] the gdr energy @xmath40 is found as the solution of the equation [ see eqs .
( 3 ) and ( 4 ) of ref .
@xcite . ] : @xmath96^{2}(1-n_{p}-n_{h } ) } { \omega - e_{p}-e_{h}}+{\cal f}_2 ^ 2\sum_{ss'}\frac{[v_{ss'}^{(-)}]^{2}(n_{s'}-n_{s } ) } { \omega - e_{s}+e_{s'}}\bigg\}= 0~ , \label{egdr}\ ] ] where the expression within the figure brackets is the real part of the polarization operator @xmath97 at @xmath98 , which causes the energy shift of the phonon energy @xmath35 under the effect of quasiparticle - phonon coupling , whereas the imaginary part of @xmath99 is the phonon damping in eq .
( [ gamma ] ) .
as for the pairing problem , the present work considers the exact treatment of thermal pairing within the canonical ensemble ( ce ) with the pairing hamiltonian given as @xmath100 where @xmath101 are the creation ( annihilation ) operators of a particle ( neutron or proton ) having the single - particle energy @xmath33 , angular momentum @xmath102 with projections @xmath103 , denoted with @xmath104 , and with projection @xmath105 , denoted with @xmath106 .
the exact eigenvalues @xmath107 of the eigenstates @xmath108 with degeneracies @xmath109 , obtained by diagonalizing this hamiltonian , are used to construct the partition function within the ce at temperature @xmath1 @xcite : @xmath110 by using this ce partition function , the total exact ce energy @xmath111 , entropy @xmath112 , and occupation number @xmath113 on the @xmath114th level are obtained in terms of the ensemble averages of the exact eigenvalues @xmath107 and state - dependent occupation numbers @xmath115 , respectively , as @xmath116 knowing the total energy @xmath117 and entropy @xmath118 of a system with @xmath119 particles ( n = n neutrons or z protons ) , we can calculate its helmholtz free energy @xmath120 and the exact ce chemical potential @xmath121 as @xcite @xmath122~. \label{lambda}\ ] ] this approach does not produce a pairing gap , which is a mean field concept . instead , an exact ce pairing gap is introduced to mimic the mean - field pairing gap as [ see eq .
( 18 ) of ref . @xcite and the discussion therein . ] , namely : @xmath123 it is worth mentioning that , although the ce approach to thermal pairing was employed by several authors @xcite , the gap ( [ gap ] ) is very similar to that defined in eq .
( 52 ) of ref .
@xcite , but it is different from the canonical gap defined in refs .
@xcite because the term @xmath124 is taken at @xmath125 0 in the latter , whereas no thermal pairing gap was calculated in ref .
@xcite .
given the chemical potential @xmath121 in eq .
( [ lambda ] ) and the pairing gap @xmath66 in eq .
( [ gap ] ) , the quasiparticle energies @xmath34 can be calculated by using eq .
( [ ek ] ) .
in the same way the exact ce pairing gap is introduced to mimic the mean - field pairing gap , the quantities that mimic the bogolyubov s coefficients @xmath31 and @xmath32 can be obtained separately for neutrons and protons by using the standard expressions @xmath126}~,\hspace{5 mm } v_k = \sqrt{\frac{1}{2}\bigg[1-\frac{\epsilon_k-\lambda(t)}{e_k}\bigg]}~ , \label{uv}\ ] ] which , strictly speaking , are valid only within the bcs - based theories .
a justification for such approximation is that , in the cases where both the exact ce gaps and the ftbcs1 as well as the ftln1 are possible ( including the corrections owing to coupling to the self - consistent quasiparticle random - phase approximation ) , these pairing gaps are rather close to each other , especially at @xmath9 2 mev ( see fig . 1 of ref .
@xcite . ) .
the 323 neutron and proton doubly folded single - particle energy levels for @xmath0tl employed in the present calculations are obtained within the axially deformed woods - saxon potentials including the spin - orbit and coulomb interactions @xcite .
the neutron spectrum spans a space of 157 doubly folded levels ( degenerated into 41 spherical orbitals ) , starting from the bottom at -38.36 mev up to 12.57 mev .
the proton spectrum has 166 doubly folded levels ( degenerated into 44 spherical orbitals ) , starting from the bottom at around -34 mev , up to around 22.3 mev .
they cover the energy intervals similar to those used in the calculations of gdr width in @xmath29pb @xcite .
in the construction of the ce partition function ( [ zce ] ) one needs to include all the eigenvalues of the ground state as well as excited states .
meanwhile , the fortran imsl subroutine for matrix diagonalization at the riken integrated cluster of clusters computing system implies that the numbers of levels @xmath127 and particles @xmath119 should satisfy the condition @xmath128 < 10 ^ 4 $ ] , which means it is impossible to carry out the exact diagonalization of the pairing hamiltonian ( [ hpair ] ) with the entire single - particle spectra @xcite .
therefore , knowing that pairing has a significant effect around the fermi surface , we calculate the exact ce thermal gap ( [ gap ] ) only for 14 doubly folded neutron ( proton ) levels situated around the fermi one , @xmath129 , with 7 levels below and 7 levels above it .
the selected levels belong to the group of 18 neutron ( proton ) levels within five @xmath130-folded spherical orbitals @xmath131 for the neutrons , and @xmath132 , @xmath133 for protons , as listed in table [ table1 ] .
the pairing parameters @xmath134 0.182 mev for neutrons and @xmath135 = 0.37 mev for protons are chosen to reproduce the empirical values of the neutron and proton gaps @xmath136 , which are both equal to 1 mev for @xmath0tl according to fig . 2 - 5 of ref .
this value agrees well with the three - point and five - point gaps calculated in ref .
@xcite and shown in fig . 1 therein .
the remaining 4 neutron levels ( two on @xmath137 and two on @xmath138 orbitals ) and 4 proton levels ( two on @xmath139 and two on @xmath133 orbitals ) of this group are assumed to have the same thermal neutron and proton pairing gaps , respectively .
for the levels beyond this group , pairing is assumed to be negligible so that @xmath140 , @xmath141 , @xmath142 , and @xmath143 .
.single - particle spherical orbitals obtained within woods - saxon potentials and used in the calculations of neutron and proton exact ce pairing gaps for @xmath0tl .
[ table1 ] [ cols="^,^,^,^,^ " , ] the pdm assumes that the matrix elements of gdr coupling to @xmath15 configurations , causing the quantal width , are all equal to @xmath49 , whereas those of coupling to @xmath16 ( @xmath17 ) configurations , causing the thermal width at @xmath14 0 , are all equal to @xmath50 [ see sec .
ii b of ref .
@xcite for the detail discussion on the justification of this assumption . ] .
the third parameter , @xmath35 , in the case of gdr ( @xmath144 ) , is chosen to be close to @xmath145 . because the mechanism of the spreading width @xmath146 at @xmath18 0 is known ,
which is owing to coupling to more complicate configurations @xmath147 configurations , the pdm has no ambition to calculate it microscopically but is interested only in its temperature dependence incorporated in the quantal width .
therefore the parameter @xmath49 is selected to reproduce the gdr experimental width at @xmath18 0 , which is essentially the sum of the spreading width , @xmath146 , and the escape width , @xmath85 .
the parameter @xmath50 is usually adjusted so that the gdr energy @xmath148 , found as the solution of eq .
( [ egdr ] ) , does not change appreciably with @xmath1 . because the gdr energy in @xmath0tl does not depend on @xmath1 , namely @xmath149 mev @xcite , for simplicity
, we adopt in the present paper @xmath150 13.8 mev , and select @xmath50 so that the calculated width at @xmath13 2 mev matches the corresponding experimental value for the gdr width in @xmath29pb @xcite .
and protons [ ( blue ) dashed line ] for @xmath0tl .
[ gaptl201],width=510 ] shown in fig .
[ gaptl201 ] are the exact ce neutron and proton pairing gaps for @xmath0tl as functions of @xmath1 . in different with the bcs gaps ,
the exact ce gaps do not collapse at @xmath151 0.57 mev for @xmath152 1 mev .
instead , they both decrease as @xmath1 increases . in particular ,
the proton gap remains almost unchanged or even slightly increases in the region @xmath153 1 mev , and it is significantly larger than the neutron gap at the same value of @xmath1 already starting from @xmath13 0.5 mev . at @xmath18 1 mev
the proton gap is still equal to around 1 mev whereas the neutron gap drops to 0.34 mev . at @xmath18 5
mev the proton gap remains as large as around 0.4 mev , whereas the neutron gap is depleted to 0.13 mev .
versus experimental data .
the ( red ) thick solid line is the gdr width for @xmath0tl obtained with the effect of exact ce pairing .
the ( green ) thin solid line denotes the same width but obtained without pairing .
the ( purple ) dashed line stands for the same quantity , but obtained by using the exact single - particle occupation numbers @xmath73 and effective @xmath154 discussed in the text .
the dotted line represents the gdr width for @xmath29pb obtained previously in refs .
the full circles and open boxes are experimental gdr widths for @xmath0tl @xcite and @xmath29pb @xcite , respectively .
[ widthtl201],width=415 ] the gdr fwhm @xmath155 in @xmath0tl obtained within the pdm in three different approximations are plotted as functions of @xmath1 against the experimental data in fig .
[ widthtl201 ] .
in the first approximation , the zero - pairing one , the width @xmath155 is calculated by using eq .
( [ fwhm ] ) and the phonon damping given in eq .
( [ gamma_nopair ] ) .
the single - particle occupation numbers @xmath73 are approximated with the fermi - dirac distribution @xmath156 + 1\}~ , \label{ffd}\ ] ] where the chemical neutron and proton potentials change with @xmath1 to conserve the particle numbers , according to the equation @xmath157 with @xmath158 .
the calculations adopted the following values of the parameters for the @xmath15 and @xmath16 ( @xmath17 ) couplings : @xmath159 4.0@xmath16010@xmath161 mev and @xmath162 13.27@xmath16010@xmath161 mev .
the gdr width predicted by this approximation ( thin solid line ) clearly overestimates three lowest experimental data points at @xmath18 0.82 , 0.84 and 0.97 mev .
the overall temperature dependence of this width is similar to that obtained previously for the doubly closed - shell @xmath29pb ( dotted line ) except at @xmath8 3 mev , where it becomes larger than the gdr width in @xmath29pb , to saturate at @xmath8 4 mev at a value of 12.7 mev compared to that of around 10.5 mev for @xmath29pb .
the small bump at @xmath163 0.3 mev occurs because of the neutron single - particle energies of the orbitals @xmath164 and @xmath165 , which are equal to 6.903 mev and 6.587 mev , respectively , and located just below the fermi level at @xmath18 0 .
they cause a local minimum in the temperature dependence of the neutron chemical potential at @xmath18 0.3 mev as shown by the solid line in fig .
[ widthtest ] ( a ) .
a test by changing the energies of these levels to 7.903 and 7.587 mev , respectively , removes this local minimum and , consequently , flattens the gdr width at @xmath163 1 mev , as shown by the dash - dotted lines in figs .
[ widthtest ] ( a ) and [ widthtest ] ( b ) , respectively . .
in ( a ) the ( green ) solid line denotes the neutron chemical potentials as a function of @xmath1 obtained for @xmath0tl before modifying the energies of the orbitals @xmath164 and @xmath165 , whereas the ( blue ) dash - dotted line stands for the same quantity obtained after modifying the energies of these orbitals as specified in the text . in ( b )
the ( green ) solid and ( blue ) dash - dotted lines show the gdr widths for @xmath0tl obtained within the no - pairing approximation before and after this modification , respectively .
[ widthtest],width=415 ] in the second approximation , we consider an effective way of taking thermal pairing into account , namely , in eq .
( [ gamma_nopair ] ) , the single - particle occupation numbers @xmath73 for the 18 selected levels around the fermi surface , which were discussed previously , are replaced with their exact ce values @xmath166 calculated by using eq .
( [ ef ] ) .
the effective values of @xmath167 for these 18 levels are found by inverting the fermi - dirac distribution ( [ ffd ] ) to obtain @xmath168 $ ] .
the calculations used @xmath169 2.45@xmath16010@xmath161 mev and @xmath162 14.83@xmath16010@xmath161 mev .
starting at the same value equal to around 4 mev at @xmath18 0 , the width predicted by this approximation ( dashed line ) decreases as @xmath1 increases up to @xmath18 0.7 mev to perfectly match the three lowest data points at @xmath18 0.82 , 0.84 and 0.97 mev . at @xmath8 0.7 mev this
width increases with @xmath1 but remains smaller than that obtained without pairing up to @xmath13 3.6 mev when they cross and reach the same value at @xmath18 5 mev .
finally , thermal pairing is fully taken into account within the quasiparticle representation of the pdm by using eqs .
( [ fwhm ] ) and ( [ gamma ] ) , which include the exact ce thermal pairing gap for 18 levels around the fermi surface for neutrons and protons , separately , as has been discussed in sec .
the gdr width predicted by this approach ( thick solid line ) nicely describes all the data points for @xmath0tl .
its values at @xmath12 0.8 , 0.9 , 1.0 , 1.1 and 1.2 mev are found equal to 3.58 , 3.75 , 3.96 , 4.22 and 4.5 mev , respectively , to be compared with the corresponding experimental values of 3.4@xmath170 0.45 , 3.7@xmath170 0.45 , 3.8@xmath170 0.45 , 4.6@xmath170 0.45 and 4.5@xmath170 0.45 mev at @xmath18 0.82 , 0.94 , 0.97 , 1.09 and 1.12 mev , respectively .
it also matches fairly well the experimental width for @xmath29pb at 1.3@xmath171 2 mev , although , up to @xmath13 3 mev , it remains smaller than the width predicted by the pdm for @xmath29pb ( dotted line ) .
however , to obtain such description , the value for @xmath49 has to be smaller than that used in the case without pairing so that @xmath172 3 mev instead of 4 mev for the gdr width in @xmath29pb at @xmath173 , and , consequently , the value for @xmath50 has to be larger to reproduce the same data point at @xmath18 2 mev .
eventually , this brings @xmath174 closer to the recent parametrization of the gdr width for quasipherical nuclei @xcite , which implies that the gdr fwhm for @xmath0tl at @xmath18 0 amounts to around 3.33 mev , rather than 4 mev as that of @xmath29pb .
the slope of the width s increase as a function of @xmath1 also becomes slightly steeper , improving the agreement between theory and experiment within the entire region of @xmath1 . a value for @xmath49 that reproduces @xmath175 4 mev leads to an overestimation of the width in the entire temperature region of @xmath14 0 .
the values for the coupling parameters used in this approximation are @xmath169 2.81@xmath16010@xmath161 mev and @xmath162 16.73@xmath16010@xmath161 mev .
it is important to point out that , although one could fit the data points for @xmath0tl within the zero - pairing approximation by reducing the value of parameter @xmath49 from 4.0@xmath16010@xmath161 mev to 3.54@xmath16010@xmath161 mev , it is not possible by readjusting only parameters @xmath81 to achieve the overall agreement with both sets of data for the gdr widths in @xmath0tl and @xmath29pb as does the last approach that fully includes exact ce pairing gaps within the quasiparticle representation of the pdm . at @xmath8 2.5 mev the gdr width obtained by using such reduced parameter @xmath49 in the zero - pairing approximation becomes even smaller than that predicted by the second approximation ( of taking effectively thermal pairing into account ) .
this indicates that the effect owing to thermal pairing on the gdr width has its microscopic origin , which can not be accounted for by simply adjusting model s parameters
. tl obtained within the quasiparticle representation of the pdm ( red solid lines ) at @xmath12 0.8 1.2 mev as shown in the panels in comparison with the experimental linearized strength functions ( green data points ) @xcite at @xmath18 0.82 , 0.94 , 0.97 , 1.09 , and 1.12 mev from the bottom panel , ( e ) , to the top one , ( a ) , respectively .
[ stre],width=377 ] the gdr strength functions obtained at @xmath18 0.8 1.2 mev by using eq .
( [ s ] ) within the quasiparticle representation of the pdm that includes exact ce pairing are displayed in fig .
[ stre ] in comparison with the corresponding experimental data adapted from fig .
1(a ) of ref . @xcite .
the theoretical strength functions have been renormalized so that the value at @xmath176 5 mev and the maximum at @xmath12 1.2 mev match the corresponding experimental values .
this figure and fig . [ widthtl201 ] show that the pdm describes reasonably well not only the temperature dependence of the gdr width but also that of the gdr linearized shape .
in the present paper , we calculated the width and strength function of the gdr in @xmath0tl at finite temperature within the framework of the quasiparticle representation of the pdm .
thermal pairing is taken into account by using the exact treatment of pairing within the canonical ensemble .
this treatment allows us to calculate the exact equivalences to the pairing gaps for protons and neutrons in a nucleus neighboring a proton closed - shell one .
because of thermal fluctuations owing to the finiteness of the system , which are inherent in the ce , the exact ce thermal pairing gaps do not collapse at the critical temperature @xmath30 of the superfluid - normal phase transition as in the case of infinite systems , but decrease monotonically as @xmath1 increases , and remain finite up to @xmath1 as high as 5 mev .
the theoretical predictions within the pdm are compared with the data , which were recently obtained for the gdr width and strength functions in @xmath0tl at 0.8@xmath177 1.2 mev at low angular momentum below 25@xmath6 .
the good agreement between the pdm predictions including thermal pairing and the experimental data is a clear demonstration of the manifestation of the effect owing to thermal pairing , which plays a vital role in reducing the gdr width at low @xmath1 in open - shell nuclei . under the influence of thermal pairing
, the gdr width in @xmath0tl becomes as low as around 3.7 mev at @xmath18 0.8 mev , and the width @xmath174 of the gdr built on the ground state ( @xmath18 0 ) can be as small as 3 mev , that is smaller than the gdr width in @xmath29pb ( 4 mev ) at @xmath18 0 .
the results obtained in the present work as well as the previous predictions for the gdr width in @xmath11sn , where the important role of neutron thermal pairing has been shown to reduce the gdr width at @xmath9 1 mev @xcite , confirm that , in order to have an adequate description of gdr damping at low @xmath1 , a microscopic model needs to take into account thermal pairing at least up to @xmath178 1.5 mev .
the numerical calculations were carried out using the fortran imsl library by visual numerics on the riken integrated cluster of clusters ( ricc ) system .
the authors are grateful to s.r .
banerjee for useful discussion and help with the experimental data of gdr linearized strength functions for @xmath0tl .
nqh acknowledges the support by the national foundation for science and technology development(nafosted ) of vietnam through grant no .
103.04 - 2010.02 . 99 j.o .
et al . _ ,
lett . * 46 * , 1380 ( 1981 ) .
a. schiller and m. thoennessen , atomic data and nuclear data tables * 93 * , 548 ( 2007 ) .
j. j. gaardhje _
et al . _ ,
59 * , 1409 ( 1987 ) ; a. bracco _
et al . _ ,
lett . * 62 * , 2080 ( 1989 ) ; g. enders _ et al .
lett . * 69 * , 249 ( 1992 ) ; j. kasagi _
et al . _ ,
a * 538 * , 585c ( 1992 ) ; h. j. hofmann _ et al .
a * 571 * , 301 ( 1994 ) ; t. suomijrvi _
c * 53 * , 2258 ( 1996 ) .
ormand , p.f .
bortignon , r.a .
broglia , and a. bracco , nucl . phys .
a * 614 * , 217 ( 1997 ) ; d. kusnezov , y. alhassid , and k.a .
snover , phys .
* 81 * , 542 ( 1998 ) .
p. heckman _
et al . _ ,
b * 555 * , 43 ( 2003 ) .
n. dinh dang and a. arima , phys .
lett * 80 * , 4145 ( 1998 ) .
n. dinh dang and a. arima , nucl .
a , * 636 * 427 , ( 1998 ) . n. dinh dang and a. arima , phys .
c * 68 * , 044303 ( 2003 ) .
s. mukhopadhyay _
et al . _ ,
b * 709 * , 9 ( 2012 ) .
d. pandit , s. pal , a. de , and s.r .
banerjee , phys .
b * 713 * , 434 ( 2012 ) .
n.dinh dang and a. arima , phys . rev .
c * 68 * , 014318 ( 2003 ) .
n. dinh dang and n. quang hung , phys .
c * 77 * , 064315 ( 2008 ) .
n. quang hung and n. dinh dang , phys .
c * 79 * , 054328 ( 2009 ) .
n. dinh dang , phys .
c * 85 * , 064323 ( 2012 ) .
n. dinh dang , k. tanabe , a. arima , phys .
c * 58 * , 3374 ( 1998 ) ; nucl .
a * 645 * , 536 ( 1999 ) ; ibid . * 675 * , 531 ( 2000 ) . n. dinh dang , phys . rev .
c * 84 * , 034309 ( 2011 ) .
n. dinh dang , k. eisenman , j. seitz , and m. thoennessen , phys .
c * 61 * , 027302 ( 2000 ) ; n. dinh dang , t. suzuki , and a. arima , phys .
c * 61 * , 064304 ( 2000 ) ; n. dinh dang , v. kim au , and a. arima , phys .
lett . * 85 * , 1827 ( 2000 ) ; n. dinh dang , t. suzuki , and a. arima , phys .
c * 64 * , 027303 ( 2001 ) .
chomaz , nucl .
a * 569 * , 203c ( 1994 ) ; phys .
b * 347 * , 1 ( 1995 ) .
n. dinh dang and n. quang hung , phys .
c * 81 * , 034301 ( 2010 ) .
j. von delft and d. c. ralph , phys
345 * , 61 ( 2001 ) .
r. rossignoli , n. canoza , and p. ring , ann . phys . * 275 * , 1 ( 1999 ) .
s. frauendorf , n. k. kuzmenko , v. m. mikhajlov , and j. a. sheikh , phys .
b * 68 * , 024518 ( 2003 ) . t. sumaryada and a. volya , phys . rev .
c * 76 * , 024319 ( 2007 ) .
et al . _ ,
phys . commun . *
46 * , 379 ( 1987 ) .
a. bohr and b.r .
mottelson , _ nuclear structure _ , vol .
1 ( benjamin , ny , 1969 ) .
m. bender , k. rutz , p .-
reinhard , and j. a. maruhn , eur .
j. a * 8 * , 59 ( 2000 ) .
t. baumann _
et al . _ ,
a * 635 * , 428 ( 1998 ) ; e. ramakrishnan _ et al . _ , phys .
* 76 * , 2025 ( 1996 ) .
et al . _ ,
b * 670 * , 200 ( 2008 ) . | the thermal pairing gap obtained by embedding the exact solutions of the pairing problem into the canonical ensemble is employed to calculate the width and strength function of the giant dipole resonance ( gdr ) within the phonon damping model .
the results of calculations describe reasonably well the data for the gdr width as well as the gdr linearized strength function , recently obtained for @xmath0tl in the temperature region between 0.8 and 1.2 mev , which other approaches that neglect the effect of non - vanishing thermal pairing fail to describe . |
complementarity has become a catch - all concept to cover the weirdness of quantum mechanics ( qm ) .
mostly , this is because its originator , bohr , actually proposed this notion in order to capture exactly these weird features and legitimize them .
it is arguably the case , however , that having been unable to remove what were at first recognized as antinomies , they were just redubbed as deep , albeit preternatural , holistic insights into atomic scale ontology . following the popularization of the term in physics , bohr extended the concept to arenas outside the customary boundaries of science even , until it became for him a universal precept , the foundation of a total _ weltanschauung .
_ history seems to show that he was only marginally successful at promulgating philosophy .
none of his students , nor readers of his papers , agree completely on just whatever it was he tried to convey under the term `` complementarity , '' even when restricted to within physics.@xcite nevertheless , through the decades , authors of textbooks and other secondary literature have focused the notion of complementarity on the issue of particle - wave identity . out of this literature over time a consensus has crystallized : complementarity refers to that feature of quantum theory to the effect that ensembles of all entities , depending on the scale of the interaction or measurement scheme , exhibit alternately either the properties of waves or particles .
this has led some to suggest that there exists a single , fundamental category , instead of particles and waves , namely `` wavicles . ''
such abstract unification , however , does not overcome the reality of the matter , namely that this category has two obvious components : those entities which classically are particles and those that are waves .
this division is obvious in terms of the characteristic of mass ; classical waves can be ascribed only a mass equivalent of energy . through the years
, various proposals have been made to render this `` schizoid '' phenomena less mysterious .
for each of the two subcatagories different conceptions have been proposed ; the wave - like behavior of particle beams , for example , has been attributed to various forms of ` pilot waves . ' as is well known , de broglie , who first suggested that particles should exhibit wave phenomena , more or less explicitly envisioned that particles were a sort of material kernel embedded in such a pilot wave.@xcite the wave portion was credited with navigating for the kernel without itself alone interacting with measuring devices , a behavior that eventually led to coining the term `` empty wave . ''
proponents of this notion seem to have abandoned it , perhaps because of the internally contradictory conception that a pilot wave can react to all material boundaries except those used in connection with measurements .
how can any wave know the difference in purpose of any particular material object it encounters ? when is it supposed to communicate with objects so that it can navigate , and when is it to ignore them as the intervention of an experimenter ? this writer has proposed an extention of the pilot wave idea to the effect that such pilot waves are classical electromagnetic background waves attached to particles by virtue of their motion through the random electromagnetic background radiation that is hypothesized as the basis of a theory known as ` stochastic electrodynamics ' ( sed ) , a theory which seeks to rationalize quantum phenomena.@xcite whether this effort is successful , the reader is invited to pursue independently ; it is not the central theme of this paper , which concerns the other subcategory , the particle - like behavior of electromagnetic waves .
particle - like behavior of waves is nowadays credited mostly to einstein in connection with the photoelectric effect .
actually , however , einstein proposed only that energy transfer occurs in ` packets , ' it was later that the idea , that the incoming radiation itself is somehow subdivided , captured the common imagination .
still , even today , the conception is vague . sometimes considered a ` mode ' of the vacuum , and thereby present throughout the universe on the one hand , elsewhere such packets
are envisioned as `` needle radiation '' with small lateral cross - section , because absorption seems to be point - like . on other occasions
they are thought to be shockwave - like , because absorption can be rapid , seemingly instantaneous . in combination ,
such solutions are difficult to associate with the inhomogeneous wave equation ( with sources ) , as derived from maxwell s equations .
additional support for packaged radiation was found in the scattering of electromagnetic radiation from electrons , i.e. , the compton effect .
this effect is , in a certain elementary manner , just a particular application of the photoelectric effect . in any case , the notion of packages , as a model for waves , particularly electromagnetic waves , has become virtually universal nowadays as the ` photon ' paradigm . no modern commentator fails to note that , fundamentally , light is a stream of photons without branding himself an ignoramus .
this , in spite of the spectacular applicability of non quantum maxwell theory .
it is interesting to speculate on just why this is so ; this writer likes to suggest that the cause is that electromagnetic waves , as such , are never detected .
in fact the mechanism of the interaction of charged particles is completely beyond the reach of experimental science , which is restricted to ` photocurrents , ' for which there are two variants : those acting as sources in antennas , and those driven by the consequences of the former on the other end in detectors . in spite of the vividness of the images from paradigms of what transpires in between , either as electromagnetic waves or photon streams , all effective experience for mortals
is actually limited to these currents , information on which is then used mentally to _ infer _ just how these two currents have interacted in detail across space and through time .
this understanding makes photo detection the crux of the matter , and arguably , the essential reason for the utility of the `` photon '' concept .
here we come to the point of this paper .
the digitization , or ` quantization ' if one prefers , of charges , is _ not _ a result of quantum theory , but an independent empirical fact that serves as an hypothetical input into both classical and quantum theories of physical phenomena ; thus , the `` quantum '' behavior of electromagnetic waves as ensembles of particles , may not be at all `` quantum , '' but simply a consequence of the discreteness of the charges making up the material of detectors .
if we take this observation as a scientific proposition to be tested for internal ( theoretical ) consistency and external ( empirical ) verification , then it can not be regarded as anything other than just another task for work - a - day science . in this spirit
, it shall be addressed herein .
electromagnetic wave detectors come in a variety of constructions , each intended for a particular frequency range . in those lower portions of the electromagnetic spectrum where current technology enables following time development of electromagnetic fields ,
there is little of special interest for fundamental quantum physics analysis .
it is only in that range in which this is no longer possible that quantum phenomena seem to arise.@xcite detectors in this range do not follow the undulatory time development of incoming radiation .
instead , they absorb an undetermined number of cycles , and respond then with some bulk transition .
photographic plates are a good example .
incoming radiation at some point triggers a chemical reaction in molecules embedded in an emulsion that , as differentiated compounds , then serve during development as seeds for some macroscopically visible effect .
likewise , for the photoelectric effect .
incoming radiation eventually results in the transition of a valence electron to the conduction band where it is then drawn off into circuitry to be amplified and registered as the ` detection event . '
the historical photoelectric effect considered by einstein has an additional feature that is widely taken as symptomatic of an essential quantum character .
it is that the stimulated photocurrent appears virtually instantaneously after the start of illumination .
the argument that this is to be explained only by assuming that the incoming energy is bundled in compact packages , nowadays called ` photons , ' assumes that the electrons , before illumination , were all resting at an energy so low that it would require a measurable duration of exposure to the incoming radiation before they could be energized sufficiently to enable them to overcome binding forces.@xcite such an assumption is not , however , beyond reproach .
it could just as well be assumed that the electrons in the detector mass had a distribution of energies attributable to prevalent or incoming noise so that of the whole population , a certain portion at any given instant has noise energy bringing them up to nearly the escape level .
then when a coherent stimulation signal arrives , those electrons at this otherwise temporary ` almost ' escaped energy level are boosted over the threshold very quickly , virtually instantly .
in fact , lamb and scully long ago have shown with such ideas that photoelectric phenomena do not entail quantum theory ; semiclassical ideas suffice.@xcite the vital point made here is : the digitization of the detector reaction to incoming radiation need not necessarily be attributed to particle - like packaging of the incoming stimulus , i.e. , photons .
all the behavior involved can be explained fully also in terms of continuous , wave - like radiation of the classical sort known to radio , radar and communication engineers , where the discrete aspects of low intensity measurements are due to discreteness of detector charges . in other words , because charges are the ineluctable intermediaries between mortals and electromagnetic waves , their contribution to the nature of an observation can not be evaded .
it is the purpose here to parse the logic of some modern optics experiments plumbing the mysteries of complementarity of electromagnetic interaction .
these are , first , two quantum eraser experiments , and then ` afshar s ' experiment ; all three are intended to focus specifically on the issue of delimiting complementarity in light .
there are several versions of experiments demonstrating this conception proposed first by scully and drhl in 1982.@xcite the basic idea is to elaborate on a standard two - slit experiment so as to be able to mark the signal passing through at least one of the slits , thereby permitting the determination of which slit the purported ` photon ' passes through .
an extra feature in this proposal was the notion that , with clever design , it could be possible to `` erase information '' so as to flip the case between wave and particle , most spectacularly after - the - fact even .
as is well known , textbook orthodoxy nowadays would have it that if the slit of passage can be determined ( a particle - like property ) , then no interference pattern ( a wave - like property ) will result . the difficult trick in conceiving of and performing such
an experiment is to find a method to mark a signal on passage through a slit .
the original proposal was to employ multilevel atoms in place of slits .
the atoms were chosen to have , at least in principle , the same emitted decay radiation but different final states readable by some other measurement .
this scheme should allow the experimenter to do , even well after - the - fact , or not do the second measurement ; as a consequence , the interference pattern should appear or not appear .
( the sense in which the pattern appears or not results from the alternate data reduction calculations feasible with the existence of different elements in the data sets , after - the - fact .
it should not be understood that some visual image of an object is made to come and go . )
we gather that exactly this setup was never realized , however .
a particularly clean and purely optical ` quantum eraser ' experiment involving two slits , inspired by a conceptually similar proposal of scully et al .
@xcite , was carried out by walborn et al . and reported in 2002.@xcite the essential feature of this setup is that the beam sent towards the slits is the signal output of a parametric down conversion crystal ( pdc ) ; the idler is directed separately through another polarizer and then to a detector . see fig 1 . here
the marking in the slits is considered achieved by placing polarizers before each slit .
if the two polarizers before the slits are set to orthogonal positions , then the signals impinging on the registration screen are independent and do not form an interference pattern .
this feature is thought to conform with the complementarity principle , i.e. , that which - way information provided by a polarization tag put on the signals as they pass through the slits , is said to destroy the wave - information revealed by the interference pattern .
now , however , if coincident counts between data points taken behind the slits with points seen from the idler beam are analyzed , then the interference effect can be made to reappear in the coincidence counts .
that is , if the polarizer in the idler beam ( let us assume from a type - i pdc ) is set to horizontal , then the signal in the vertical channel behind the slits will yield correlated `` hits , '' ( fig .
2 ) by effectively filtering out those hits in the horizontal channel for lack of a mate in the idler channel .
if the idler polarizer is changed to vertical , then the correlated hits are in the horizontal channel behind the slits , as now the vertical component is filtered out ( fig .
3 ) . if the polarizer is removed from the idler beam , then no correlations with the slits can be made ; the only hit pattern available shows no interference , and it is said that `` which - way information has been erased '' ( fig .
4 ) . this behavior is said to illustrate quantum complementary .
but does it ?
first note that the various signals used are just different states of polarization .
this is very significant because polarization is a phenomenon fully explained by stokes circa 50 years before the need for quantum theory was known.@xcite it is a fully classical phenomenon .
how then , can this phenomenon be used to plumb a principle of quantum mechanics ?
the only even remotely quantum aspect to this story is the implicit assumption that the signals are made up of ` quantum ' objects , i.e. , photons , so that it is presumed their behavior must be regulated by quantum mechanics .
but , as was argued above , the possibility of fully non quantum character has not been rigorously excluded .
moreover , since there are two slits and two different interference patterns , with arguments that are brief and loose ( i.e. , `` it is clear that @xmath0 '' ) in the customary interpretation , an association is made to ` which - way ' information .
this is done in spite of the fact that each interference pattern is indisputably a two slit pattern .
thereafter , even while observing that the _ sum _ of these two patterns is what is observed when the idler polarizer is removed , is is said that this results from _ erasing _ some information .
one ought be excused for finding this apologia at least lexicographically pathological .
thus , those who find such imagery based on quantum mechanics at all obscure and contrived may be well disposed to consider the following classical rendition .
let us take it that the signal directed towards the slits is a vertically polarized classical electromagnetic beam .
if there are no polarizer filters in front of the slits , after passing through both slits the signal shows interference on the registration screen .
next let us place polarizers with axes set at @xmath1 to the vertical before one slit , and at @xmath2 before the other slit .
now , each polarizer will split the incoming vertically polarized electric field into two components one vertical and one horizontal , each with amplitude @xmath3 times the ` mother ' signal as determined by malus s law , once as a projection onto the polarizer axis and then again to find the component in either the vertical or horizontal direction . in turn , each orthogonal component or ` daughter ' signal will result in an interference pattern on the registration screen .
but , one pattern will be fringes ( fig .
2 ) , while the other is antifringes ( fig .
3 ) . because the horizontal diffraction pattern is the sum of opposing components , a phase term of @xmath4 is inserted and the pattern comprises antifringes ; in contrast the vertical components are codirectional and in phase .
these two patterns together add up to a total pattern without interference ( fig 4 ) .
that is , if the daughter horizontal interference pattern can be written @xmath5 and the daughter vertical pattern in these expressions for intensity equals @xmath4 in @xmath6-field amplitudes . ]
@xmath7 then the sum @xmath8 shows no interference . in the terminology of this explanation
, it can not be said that information has been erased , although it has been concealed , using , as it were , a kind of secret writing .
it can be refound ( rendered legible ) as follows .
as the signal was one output from a parametric down conversion crystal ( pdc ) , the conjugate idler then , depending on whether the crystal was of type i or ii , will be horizontal or vertical polarization respectively .
if type i , then the vertical contribution to the total pattern behind the slits will have nearly perfect correlated partner pulses in the horizontal mode in the idler branch . by counting only those hits in the signal pattern behind the slits
for which there is a companion hit in the horizontal idler mode , effectively the horizontal contribution to the signal pattern is filtered out , revealing the vertical interference pattern .
this procedure can be called ` coincidence filtering . '
when such coincidences are the result of polarization , such filtering is a purely non quantum procedure . on the basis of this analysis
it can be said that such a `` quantum eraser '' is neither quantum nor an eraser ; as such , it has little to contribute to a debate regarding the significance of quantum complementarity . furthermore , the only feature resembling quantum theory is the fact that the intensity of the signals used is so low that detailed structure , i.e. , its composition as a stream of point charges , that is electrons , is visible as individual `` hits . '' another particularly interesting ` quantum eraser ' experiment put the decision to measure wave- or particle - like characteristics to the internals of the setup , thereby taking it away from the experimenter .
the experimental setup is depicted in fig 5 .
the salient feature in this experiment , and distinguishing it from the one just considered , is the use of beam splitters . at the so - called single photon level
they are known to be virtually absolutist in splitting a beam by whole photons ; either a whole photon goes through , or a whole photon is reflected , not both by dividing incoming energy . in this experiment
their role is to take the decision regarding which correlation shall be registered out of the hands of a mortal , i.e. , the experimenter , and put it in the hands of zeus .
then , after - the - fact , each subensemble marked by which detector combination fired , is separately analyzed for intensity as a function of lateral detector displacement in the idler branch .
ensembles that correspond , again , to those made by coincidences triggered by just one of the components in one or the other dimension , corresponding to eqs .
( [ 1 ] ) or ( [ 2 ] ) , show interference .
beam splitters operating on a beam of the intensity yielding single hits in the detectors , show strict anticorrelated hits , which is interpreted nowadays as reflecting the essential , ineluctable , distinct identity of photons . according to current understanding ,
the random choice of transmission or reflection is a reflection of the fundamental uncertain , statistical , i.e. , `` quantum '' character of the universe at an atomic scale .
this feature might seem , therefore , to vest this experiment with an essential quantum element .
for this experiment , however , it is immaterial how the choice is made ; the phenomena under study does not depend on the mechanism ( or lack thereof ) of choice .
the only relevant feature is that the separate subensembles be compiled ; because only within each does the phenomenon of interest takes place .
the quantum character of beam splitters , although logical on its surface , is actually not inevitable . for it to be necessary , would require that alternative theories are logically and rigorously excluded .
insofar as non quantum alternatives exist , e.g.:@xcite , the viability of the photon paradigm is just provisional .
this experiment also aims to study complementarity.@xcite the basic idea is to set up a double - slit diffraction pattern at a middle stage of an optics train which is continued through a lens to end in a sharp image of the slits ; see : fig .
, its conceptual novelty consists in showing that thin opaque obstacles placed in the beam exactly at the locations of destructive interference in the fringe pattern midway through the train have no effect on the intensity of the slit s image at the end of the train .
the image of the slits can be sharp only if those photons , which it would seem have to have been intercepted by the obstacles , are in fact not removed from the beam . indeed ,
if one slit is blocked , the obstacles do substantially degrade the image of the open slit .
thus , one is led to question how to reconcile these contradictory facts .
the very precise final image of the slits gives very precise which - way , i.e. , particle - type information ( in the sense of geometrical optics ) on which slit the light for each slit image has passed through . at the same time , knowing just exactly where to place the opaque objects requires very precise knowledge of the wave - optical character of the beam .
since both types of knowledge are in play and are known to the experimenter at the same time , this knowledge is , arguably , in violent contradiction to the complementarity principle .
once again , however , a central question is just what is ` quantum mechanical ' in the involved phenomena ?
there is nothing in the device chain for which otherwise quantum theory is required .
indeed , in afshar s paper analyzing this experiment , one does not find a single occasion for which a factor of planck s constant is required .
the only quantum - appearing structure that is used is the born association of the probability of presence ( here of a ` photon ' ) being proportional to the square of the field . in this case
the field is actually the electromagnetic field , so that the born hypothesis completely overlaps the theory of photocurrent generation , i.e. , that it is proportional to the square of the electric field , a fact that might be seen even as the underpinning for born s interpretation . moreover , afshar s analysis calls on the notion that properties of a light beam are considered continously variable from wave to particle , a concept that wreaks havoc with the concept of ` particle . '
once again , there is little to nothing here addressing any feature of the complimentary principle from bohr . on this basis
it can be taken that this experiment actually addresses the issue of the tenability of the ` photon ' paradigm .
it renders the ray - optical viewpoint of photon propagation problematical while endorsing wave propagation .
in other words , it is conceptually a variation of the extensive experimental program of roychoudhuri@xcite , who identified a number of internal inconsistencies in photon imagery .
when is a phenomenon a quantum phenomenon ?
this question can be deeper than it appears at first sight .
a complication arises in that some non quantum structure from classical physics fits perfectly well within the vocabulary , notation , algorithms and interpretation of quantum theory without , however , actually depending on any hypothesis unique to quantum mechanics .
vectors of absolutely any sort , for example , can be symbolized by ` bras ' and ` kets . '
insofar as the bra - ket notation encompasses abstract vector space structure , it can be self - consistently used whenever this structure obtains .
to this writer s mind , this leaves only one option open for defining the category of quantum phenomena , namely : _ a quantum phenomenon is one that absolutely can not have its patterns encoded mathematically without calling on a unique fundamental hypothetical input of quantum mechanics .
_ in the final analysis , this means that the noncommutivity of phase ( or quadrature ) space must be essential . therefore , any phenomenon that can be explained without reference to this noncommutivity , no matter how unappealing the non quantum argument , is not a genuine `` quantum phenomenon , '' even when it can also be explained with quantum mechanics .
just here there arises a premier example of non quantum structure embedded in the quantum formalism .
the noncommutivity found in ` qubit ' space of polarization space or of spin space is widely thought to be of a quantum nature .
however , the group in play here is @xmath9 which is homeomorphic to @xmath10 .
the latter encodes the structure of rotations on a sphere which is a totally geometric structure having nothing to do with quantum mechanics . by cause of homomorphism ,
if @xmath10 is geometric , then @xmath9 must be also .
the noncommutivity uniquely evident in quantum mechanics is in phase space and is a dynamical feature , not just geometry . in the experiments considered above ,
noncommutivity on phase space plays no role in describing the observed phenomenon .
thus , it is concluded that ` quantum ' is not an appropriate adjective .
the term ` eraser ' is also inappropriate . to erase , etymologically ,
is derived from latin words meaning ` to scrape out . ' clearly this meaning is derived from times when writing was often carved in stone .
carving out words is essentially different than filling in the letters rendering them illegible ( with due attention to color and texture ) .
competent restoration can recover a concealed message ; destroyed informations is lost forevermore .
none of this has anything to do with the depths of a complementarity principle for electromagnetic radiation .
the key or essential feature of what heretofore goes under the rubric of ( optical ) `` quantum eraser '' is the phase shift between the vertical and horizontal components .
this shift allows the two components to fully compliment each other as seen in eq .
( [ 3 ] ) .
because this is a simple geometric effect , it must arise in a great variety of situations .
the euphoric bedazzlement at its implications for the nature of ` time ' in quantum mechanics as reported for some observers in @xcite , from this point of view , is a reaction to the logical jujitsu entailed in trying to understand nature on the basis of inappropriate assumptions , in this case , the photon paradigm .
this is not to claim that a complementarity principle has nothing to do with quantum mechanics .
although any such principle seems to be an expression of bandwidth considerations from classical wave theory , in fact the quantum one is intimately connected to the heisenberg uncertainty relation , which , unlike electromagnetic bandwidth , is scaled by planck s constant , and must , therefore , refer essentially to a quantum phenomenon .
the pertinent question then may be : how and why for de broglie waves ( sometimes called ` matter waves ' ) is bandwidth regulated in a fundamentally distinct manner from that for electromagnetic waves ?
but , this issue pertains to entities that are indisputably particulate in a classical limit , and not to continuous , undulatory radiation .
10 beller , m. , ` quantum dialogue , ' ( u. chicago press , chicago , 1999 ) .
de broglie , l. , ` recherches sur la thorie des quanta , ' _ ann . de phys .
, 10@xmath11 srie , _ * iii * ( 1925 ) .
selleri , f. , ` quantum paradoxes and physical reality , ' ( kluwer , dordrecht , 1990 ) , chapter iv .
kracklauer , a. f. , _ found .
_ * 12 * ( 5 ) , 441 ( 1999 ) .
kracklauer , a. f. and kracklauer , p. t , ` electrodynamics : action - at-_no_-distance , ' in _ has the last word been said on classical electrodynamics ? , _ chubykalo , a. ; espinoza , a. ; onoochin , , v and smirnov - rueda , r. ( eds . ) ( rinton press , princeton , 2004 ) pp .
118 - 132 .
pearsall , t. p. , ` photonics essentials , ' ( mc - graw - hill , new york , 2003 ) .
greenstein , g. and zajonc , a. g. , ` the quantum challenge , ' ( jones and bartlett , sudbury , 1997 ) .
lamb , w. and scully , m. o. , ` the photoelectric effect without photons , ' in _ polarization : matiere et rayonement , _
( presses univ .
fr . , paris , 1969 ) 363 - 369 .
scully , m. o. and drhl , k. , _ phys . rev .
a _ , * 25 * ( 4 ) , 2208 ( 1982 ) .
scully , m. o. , englert , b. g. and walter , h. , _ nature ( london ) _ , * 351 * , 111 ( 1991 ) .
walborn , s. p. , terra cunha , m. o. padua , s. and monken , c. h. , _ phys .
a _ , * 65 * , 033818 .
stokes , g. g. , _ trans .
soc . _ * 9 * , 339 ( 1852 ) .
kim , yoon - ho ; rong , yu . ; kulik , s. p. ; shih , yanhua and scully , m. o. , _ phys .
* 84 * ( 1 ) 1 ( 2000 ) .
santos , e. , _ proc .
spie , _ * 5866 * , 36 ( 2005 ) .
afshar , s. , _ proc .
spie , _ * 5866 * , 229 ( 2005 ) .
roychoudhuri , c. , _ proc .
spie , _ * 5866 * , 26 ( 2005 ) .
aharonov , y. and zubairy , m. s. _ science _ * * 307 * * , 875 ( 2005 ) . | optical experiments designed to explore quantum complementarity are reanalyzed .
it is argued that , for each , a classical explanation is not only possible , but more coherent and less contrived .
the final conclusion is that these experiments actually constitute support for criticism of the photon paradigm of electric charged particle interaction .
they offer little or nothing to say about quantum complementarity once the photon concept is not imposed by mandate . |
a framing for a smooth @xmath0-manifold @xmath1 is a smooth choice of ordered basis @xmath2 for the tangent space at each point of @xmath1 .
a _ projective _ framing is a smooth choice of a _ pair _ of bases of the form @xmath3 .
it has long been established that any orientable 3-manifold admits a framing ( s ) and any non - orientable 3-manifold admits a projective framing ( hh ) .
the original proofs rely on the notion of characteristic classes .
we present a proof for compact @xmath1 , which will only use basic facts about immersions of surfaces in 3-space .
we now present the facts on immersions that we will need .
denote @xmath4 , which is a cyclic group of order 4 .
let @xmath5 denote an annulus or mobius band .
there are two regular homotopy classes of immersions of @xmath5 into @xmath6 .
to each such class we attach a value in @xmath7 as follows : for @xmath5 an annulus , the regular homotopy class of immersions which includes an embedding whose image is @xmath8 { \subseteq}{{\bbb r}}^2 \times { { \bbb r}}= { { { \bbb r}^3}}$ ] , will have value @xmath9 .
the other class , ( containing an embedding differing from the previous embedding by one full twist ) will have value @xmath10 . for @xmath5 a mobius band we attach once and for all the value @xmath11 to one of the classes and the value @xmath12 to the other ( which again differ from the previous by one full twist ) .
now let @xmath13 be a closed surface and @xmath14 an immersion .
we define a map @xmath15 as follows : given @xmath16 let @xmath17 be an embedded circle which represents @xmath18 .
let @xmath5 be a thin neighborhood of @xmath19 in @xmath13 , then @xmath5 is an annulus or mobius band .
we define @xmath20 to be the value in @xmath7 attached above to the immersion @xmath21 .
it has been shown in p that @xmath22 is indeed well defined on @xmath23 and satisfies the following property : @xmath24 where @xmath25 denotes the intersection form on @xmath23 .
note @xmath26 .
we remark that the notation in p differs from ours in that @xmath7 is taken there to be @xmath27 rather than @xmath28 .
and so the numerical value of @xmath22 appearing there is twice the value here , and our property ( fp ) appearing above is replaced there by @xmath29 .
we point out that the above facts about immersions are indeed basic , in the sense that they may be obtained without knowledge of the smale - hirsch theorem .
e.g. for the classification of regular homotopy classes of immersions of @xmath5 into @xmath6 where @xmath5 is an annulus or mobius band , one needs to show that the easily constructed map into @xmath30 is bijective . in general
one would use the smale - hirsch theorem , but in this case surjectivity follows by direct construction of the two immersions , and injectivity follows by familiarity with the `` belt trick '' .
we will prove the statement for @xmath1 a closed 3-manifold .
the statement for @xmath1 with boundary will follow by restricting a framing from the double of @xmath1 .
so let @xmath1 be a closed 3-manifold , and let @xmath31 be a heegard surface , i.e. @xmath13 splits @xmath1 into two handlebodies @xmath32 ( orientable or non - orientable ) with common boundary @xmath13 .
let @xmath33 be a system of disjoint circles which may be compressed in @xmath34 , reducing @xmath34 to a ball , and let @xmath35 be such system compressible in @xmath36 .
then a thin neighborhood @xmath5 of each of the @xmath37s and @xmath38s is an annulus .
if @xmath1 is orientable then embed @xmath34 into @xmath6 as in figure f1a .
this induces an embedding @xmath14 as its boundary .
if @xmath1 is non - orientable , then take an immersion @xmath39 into @xmath6 as in figure f1b . in both cases we have @xmath40 for all @xmath41 .
( we allow @xmath37 to denote both the circle in @xmath13 and its homology class in @xmath23 . ) for any diffeomorphism @xmath42 , define @xmath44 by @xmath45 .
note that indeed @xmath46 and by ( fp ) @xmath47 is a linear functional .
let @xmath48 denote the dehn twist along @xmath37 , then the induced homomorphism @xmath49 is given by @xmath50 .
we get ( by ( fp ) ) for all @xmath16 : @xmath51 let @xmath52 be the subspaces spanned by the @xmath37s and @xmath38s respectively , then the intersection form vanishes on @xmath53 and @xmath54 .
it follows by ( fp ) that @xmath55 is identically 0 and @xmath56 is a @xmath57 valued linear functional . by ( e2 ) @xmath53 is contained in the kernel of each @xmath58 and since the intersection form is non - degenerate on @xmath23 , @xmath59 span all linear functionals on @xmath23 which vanish on @xmath53 . in particular , they span a linear functional @xmath60 vanishing on @xmath53 and satisfying @xmath61 .
now let @xmath62 be the composition of dehn twists along those @xmath37s participating in the presentation of @xmath60 as linear combination of @xmath59 , then @xmath63 and so for all @xmath41 : @xmath64 .
we now change the gluing between @xmath34 and @xmath36 using the @xmath42 given by lemma dhn .
this does not change the diffeomorphism type of @xmath1 since dehn twists along the @xmath37s may be extended to @xmath34 . after this change of gluing we have that @xmath65 for all @xmath41 , and still @xmath40 for all @xmath41 .
we will first construct a ( projective ) framing on @xmath66 .
let @xmath67 denote the tangent spaces of @xmath13 and @xmath1 at @xmath68 .
let @xmath69 be the unit vector normal to @xmath70 and pointing away from @xmath34 .
( assume we have chosen a riemannian metric on @xmath1 . )
if @xmath1 is orientable then there is well defined a unit normal @xmath71 at @xmath72 pointing away from the embedding of @xmath34 . and
so @xmath73 together with the choice of normals @xmath74 and @xmath71 induce an isomorphism of @xmath75 with @xmath76 , with which we pull back the standard framing of @xmath76 to @xmath75 .
if @xmath1 is non - orientable then a normal to @xmath77 may not be well chosen .
let @xmath71 be one of the two unit normals .
we define the isomorphism from @xmath75 to @xmath76 using this @xmath71 , and pull back the frame in @xmath76 which is obtained from the standard frame by rotating it by @xmath78 around @xmath71 in the positive sense defined by the direction of @xmath71 .
when choosing the opposite @xmath71 , then the isomorphism will differ by the reflection defined by @xmath71 , and the frame we pull back from @xmath6 will differ by a @xmath79 rotation around @xmath71 , and so altogether the frame induced in @xmath75 will differ by @xmath80 .
so such pull back defines a projective framing on @xmath66 . the next step is to extend the framing to a thin neighborhood of the compressing discs of the @xmath37s and @xmath38s in @xmath32 .
let @xmath19 be one of the circles @xmath37 or @xmath38 and let @xmath5 be a thin neighborhood of @xmath19 in @xmath13 , then @xmath5 is an annulus , and since @xmath81 , there is a regular homotopy @xmath82 , @xmath83 , with @xmath84 and @xmath85 an embedding with image @xmath86 where @xmath87 denotes @xmath88 $ ] . let @xmath89 be a neighborhood of the compressing disc of @xmath19 in @xmath34 or @xmath36 , so @xmath90 where @xmath5 corresponds to @xmath91
. for @xmath92 let @xmath93 denote @xmath94 where @xmath95 are respectively the disc and circle of radius @xmath96 in @xmath97 .
for any @xmath98 take @xmath74 to be the unit normal to @xmath99 pointing inward in @xmath89 . now
if @xmath1 is orientable then we follow the increasing time @xmath83 of our regular homotopy @xmath100 with decreasing radius from @xmath101 to @xmath102 in @xmath97 , and pull back the standard framing of @xmath6 to any point @xmath103 as before , using the normal @xmath74 , and normal @xmath71 to the image of the immersion in @xmath6 , where the side of the normal is chosen at time 0 to be pointing outward from the embedding of @xmath34 .
this extends the framing we have constructed in @xmath13 . at @xmath104 we have an embedding of @xmath105 as @xmath106 , and by additional regular homotopy if necessary we may assume the normals @xmath71 are pointing inward .
this extends to an embedding of @xmath107 onto @xmath108 .
now if the metric on @xmath1 is chosen so that this embedding of @xmath107 into @xmath6 is isometric , then the pull back of the standard framing of @xmath6 to @xmath105 that we have , coincides with the one induced by this embedding of @xmath107 , and so the framing can be extended to the whole of @xmath109 .
this completes the construction of the framing in @xmath89 .
we now only need to extend the framing to the remaining 3-balls in @xmath34 and @xmath36 , but this may always be done ( e.g. since @xmath110 ) .
this completes the proof for the case @xmath1 is orientable . for non - orientable @xmath1
we argue as follows : since @xmath5 is an annulus , it is orientable and so we may choose a unit normal @xmath71 in @xmath6 along @xmath111 . from the projective framing we have defined along @xmath13 select the proper ( i.e. non - projective ) framing determined above by this choice of @xmath71 .
we will extend this to a proper framing on @xmath89 as in the orientable case , only that now the framing in @xmath89 will be defined in three layers , @xmath112 , @xmath113 , @xmath114 . in the outer layer we will realize a rotation of the framing around @xmath71 in the negative sense , until we have the actual pull back of the framing in @xmath6 .
then the middle and inner layers will be defined as in the orientable case , using a regular homotopy of @xmath5 to a standard annulus in @xmath6 .
again we may complete the framing to the remaining balls in @xmath32 since once we have a projective framing on the boundary of such ball , we may select a proper framing from it ( since the ball is orientable ) , and then extend the framing to the ball as before . | we give an elementary proof of the fact that any orientable 3-manifold admits a framing ( i.e. is parallelizable ) and any non - orientable 3-manifold admits a projective framing .
the proof uses only basic facts about immersions of surfaces in 3-space .
[ section ] [ thm]lemma [ thm]proposition [ thm]corollary [ thm]definition [ thm]remark |
after the discovery of half - metallic ferromagnetism in half - heusler alloy nimnsb by de groot et al.@xcite many of the compounds were found to be half - metallic in different experimental @xcite and theoretical studies@xcite .
the half - metallic ferromagnets ( hmfs ) have band gap at the fermi level ( _ e@xmath3 _ ) in one spin channel while the other spin channel is strongly metallic .
these materials show a complete spin polarization of the conduction electrons at the
_ _ e@xmath3__. because of their exceptional band structures at the _ e@xmath3 _ , these materials are of great interest from theoretical and applications point of view .
co containing full - heusler alloys were firstly proposed by ishida et al.@xcite and were synthesised by webster@xcite .
fe@xmath0mnz type heusler alloys have also been proposed to show half - metallic ferromagnetism by fujii et al.@xcite . in the family of fe
based heusler alloys only fe@xmath0mnsi@xcite and fe@xmath0crz ( z = si , ge , sn ) @xcite were predicted to be half - metallic ferromagnts theoretically .
fe@xmath0mnsi is a ferromagnetic material with a curie temperature of 214 k and transforms at 69 k to the low temperature phase with smaller spontaneous magnetization @xcite .
the total magnetic moment of fe@xmath0mnsi obtained by magnetization measurements is 2.1 @xmath4/f.u .
( f.u.@xmath5 formula unit ) at 4.2 k @xcite which is smaller than the calculated moment of 3.0 @xmath4/f.u .
fe@xmath0mnal on the other hand is not a half - metallic as there exist slight density of states at the _ e@xmath3 _ for minority spin channel @xcite .
it is well known that the local spin density approximation ( lsda ) and the generalized gradient approximation ( gga ) schemes for the exchange - correlation potential are not sufficient to describe the electronic structure and magnetism of some heusler alloys like co@xmath0fesi@xcite .
such strongly correlated systems which contain atoms with open _ d _ or _ f _ shells , can be treated by adding on - site coulomb interaction ( _ u _ ) term as modification to lsda i.e. by using lsda + _ u _ approach @xcite .
the lsda + _ u _ method accounts an orbital - dependent on - site electron - electron coulomb interaction which is not included in the pure lda or gga .
rai et al.@xcite have studied some co based heusler alloys and reported the increase in band gap , hybridization of _
d_-_d _ orbitals as well as _ d_-_p _
orbitals when on - site coulomb interaction is added to lsda approach .
they have also concluded that some co - based heusler alloys show half - metallic character when are treated with lsda + _ u_. it is well known that the systems in which there exist the density of states ( dos ) at the fermi energy for one spin , the on - site coulomb interaction ( _ u _ ) may bring a drastic change into their electronic and magnetic properties @xcite . as the electron - electron correlation plays an important role in the heusler compounds so it can be expected that it will also affect the fe - based heusler alloys . in the present work
we have employed the full - potential linearized augmented - plane wave methods to compute the electronic and magnetic properties of three full - heusler alloys viz .
fe@xmath0mnsi , fe@xmath0mnal , and co@xmath0mnge .
the electronic structures in the fm solutions without _ u _ show that fe@xmath0mnsi is half - metallic compound whereas fe@xmath0mnal is metallic .
the large effect of _ u _ is observed on the electronic structures and magnetic moments of fe - based compounds whereas on co@xmath0mnge compound its effect is negligibly small .
the magnetic moment of fe atom increases with _ u _ and in the mn atom it decreases with _ u_. the magnetic moments of fe and mn atoms are found to be coupled ferromagnetically in fe@xmath0mnsi for all values of _ u _ , whereas in fe@xmath0mnal they coupled antiferromagnetically below _
u _ = 2 ev and ferromagnetically above it .
the ground state of fe - based compounds remains half - metallic for all values of _ u _ , whereas it changes from metallic to semiconductor in fe@xmath0mnal after _
u _ = 2 ev .
also the ferromagnetic coupling between fe and mn atom is found to be responsible for the presence of semiconducting ground state in this compound .
the electronic and magnetic properties of fe@xmath1mnsi , fe@xmath1mnal and co@xmath1mnge were calculated by using the full - potential linearized augmented plane - wave ( fp - lapw ) method within the density functional theory ( dft ) implemented in wien2k code@xcite . the local spin density approximation ( lsda ) of perdew and
wang @xcite was employed for exchange - correlation energy of electrons .
the effect of on - site coulomb interaction ( _ u _ ) under lsda + _ _ u _ _
@xcite formulation of the dft was also considered in the calculations . the around - the - mean - field ( amf ) version of the lsda+__u _ _ @xcite method was employed to account for the `` double - counting '' correction terms in the energy functional .
the effective coulomb - exchange interaction _ _ u__@xmath6 is given by ( _ u _ - _ j _ ) , where _ u _ and _ j _ are the coulomb and exchange parameter . the value of _ u _ was varied from 1 to 5 ev and _ j _ was kept fixed to 0 ev , therefor _
_ u__@xmath6 was equal to _ u _ in our calculations .
the values of muffin - tin radii used in the calculations were 2.2 bohr for fe , mn , si and al atoms .
the maximum _
l _ value ( _ l@xmath7 _ ) for the expansion of wave function in spherical harmonics inside the atomic spheres was equal to 10 . for convergence of energy eigenvalues
the wave function in the interstitial regions were expanded in plane waves with cutoff @xmath8 = 8 , where @xmath9 is the smallest atomic sphere radius and @xmath10 is largest k vector in the plane wave expansion .
the electronic and magnetic properties of these compounds were calculated by using the experimental lattice parameters .
the self consistent iteration was repeated until calculated total energy / cell and charge / cell of the systems converge to less than 10@xmath11 ry and 0.001e , respectively .
all these compounds belong to the family of full heusler alloys and crystallizes in @xmath12 crystal structure with space group @xmath13-@xmath14 .
these compounds have composition x@xmath0yz , where x and y are transition metals and z is the main groups element .
x atoms ( fe and co ) are placed at wyckoff position 8c ( 1/4 , 1/4 , 1/4 ) .
y atoms ( mn ) and z atoms ( si , al and ge ) are located at wyckoff position 4a ( 0 , 0 , 0 ) and 4b ( 1/2 , 1/2 , 1/2 ) , respectively @xcite .
the dispersion curves of fe@xmath1mnsi and fe@xmath1mnal along the high symmetry directions of the first brillouin zone are presented in fig .
1(a and b ) . in the dispersion curve of fe@xmath1mnsi , the bands labeled by 1 , 2 and 3 are lying above the _ e@xmath3 _ , bands 4 , 5 , 6 and 7 are crossing it at 14 different k - points and band 8 is lying below it . around w - point ,
bands 4 - 7 are concentrated in the energy range of about -0.2 to 0.2 ev .
the total energy of the system can be minimized if there will be shifting of these bands and this shifting may lead to the fm ground state for this compound . from fig .
1(b ) it is clear that in fe@xmath1mnal the bands labeled by 1 - 5 are lying above the _ e@xmath3 _ , bands 6 and 7 which are crossing the _ e@xmath3 _ at 10 different k - points and band 8 is lying just below it .
bands 4 - 7 are concentrated around the w - point in the energy range of about -0.2 to 0.2 ev .
therefore here also one can expect that shifting of these bands will minimize the total energy of the system which may lead to the ferromagnetic ground state in the compounds .
these results are consistent with our earlier reported results for the paramagnetic phase of co@xmath1mnge@xcite . the total and partial density of states plots of fe@xmath1mnsi and fe@xmath1mnal are presented in fig . 2 and fig .
3 , respectively . from total density of states plots ( tdos ) ( fig . 2(a ) and 3(a ) ) it is clear that there is very large density of states of about 6 states / ev / f.u .
( f.u.@xmath5 formula unit ) at _ e@xmath3 _ for both the spins . according to the stoner theory
the large value of tdos may be considered as the indication of the ferromagnetic ground states in the compounds@xcite .
the antibonding bands are extended upto 0.4 ev and 0.3 ev below the _
e@xmath3 _ for fe@xmath1mnsi and fe@xmath1mnal , respectively . as per stoner
theory the total energy of the systems will be minimized if there is a shifting in spin - up and spin - down bands by @xmath20.4 ev and 0.3 ev below and above the _
e@xmath3 _ for fe@xmath1mnsi and fe@xmath1mnal , respectively
. this may be responsible for the half - metallic fm ground state in the compounds as observed in the case of co@xmath1mnge@xcite .
also the total energies of fm phase of fe@xmath1mnsi and fe@xmath1mnal are about 0.77 ev and 0.54 ev less than pm phase , which further confirm that both compounds should have fm ground states .
the partial density of states ( pdos ) plots for fe , mn and si atoms of fe@xmath1mnsi are presented in fig .
2(b - d ) . from fig .
2(b ) it is evident that the pdos of fe atom at _
e@xmath3 _ is mainly contributed by _ _
t@xmath15 _ _ and _ e@xmath16 _ states with contribution of about 0.3 states / ev / atom and about 2.0 states / ev / atom , respectively for both spin channels .
the pdos of mn atom is shown in fig .
2(c ) and in both spin channels the contribution from _
t@xmath15 _ and _ e@xmath16 _ is @xmath21 states / ev / atom and @xmath21.6 states / ev / atom , respectively .
the pdos of si atom ( fig .
2(d ) ) show that the occupancy of 3__s _ _ and 3__p _ _ orbitals at _ e@xmath3 _ is very small , which can be neglected .
the pdos plots for fe , mn and al atoms of fe@xmath1mnal are shown in fig .
3(b - d ) .
it is clear from fig .
3(b ) that the contribution of _
t@xmath15 _ is @xmath20.3 states / ev / atom and _ e@xmath16 _ is @xmath21 states / ev / atom at _ e@xmath3 _ for both spin channels . in the pdos of mn ( fig .
3(c ) ) the occupancy of _
t@xmath15 _ at _ e@xmath3 _ is @xmath20.7 states / ev / atom and _ _ e@xmath16 _ _ at _ e@xmath3 _ is @xmath22.0 states / ev / atom . from pdos plot of al atom ( fig .
3(d ) ) it is evident that there is negligibly small contribution from 3__s _ _ and 3__p _ _ orbitals at _ e@xmath3_. it is also clear from these figures that _
t@xmath15 _ and _ e@xmath16 _ have main contribution to the total dos at _ e@xmath3 _ for fe@xmath1mnsi and fe@xmath1mnal . the _ e@xmath16 _ state has the largest contribution to the tdos of both compounds at _
e@xmath3_. the spin resolved dispersion curves of fe@xmath1mnsi and fe@xmath1mnal along the high symmetry directions of the first brillouin zone are shown in fig .
4 . from the dispersion curve shown in fig .
4(a ) it is clear that fe@xmath1mnsi is metallic for spin - up channel .
bands labeled by 1 and 2 are lying just above the _ e@xmath3 _
, bands 3 and 4 are crossing the _ e@xmath3 _ at 7 different k - points and bands 5 - 7 are lying just below it .
4(b ) shows that this compound behave as semiconductor for down spin channel and there exist an indirect gap of about 0.44 ev from @xmath17 to x - direction .
the computed value of indirect band gap using lsda is less than the reported value @xcite .
this compound is found to be half - metallic similar to the previous studies @xcite .
bands labeled by 1 - 6 have shifted into the conduction band while bands 7 and 8 have shifted into the valence band and thus there is presence of ferromagnetic ground state in this compound as stated earlier . from the dispersion curves shown in fig .
4(c and d ) , it is very clear that fe@xmath1mnal is metallic in nature . in
the spin - up channels bands 1 - 4 are lying above the _
e@xmath3 _ and bands 5 - 8 are crossing it at 8 different k - points . in down spin channel bands 1 - 7
are lying above _
e@xmath3 _ and only one band , labelled by 8 is crossing it at two different k - points . thus there
exist a slight density of states at _ e@xmath3 _ in the spin - dn channel and making this compound metallic in nature .
there is almost flat conduction band along @xmath17 to x direction in the down spin channel of both compounds and this can be responsible for large value of effective masses of these compounds .
total and partial dos plots of fe@xmath1mnsi for fm solution are shown in fig .
5 . from fig .
5(a ) , it is clear that spin - up channel is occupied at _
e@xmath3 _ with occupancy of about 5 states / ev / f.u . whereas spin - down channel is unoccupied .
thus this compound behave as metal for majority spin states and semiconductor for minority spin states . after comparing fig .
2(a ) and 5(a ) , one can conclude that tdos shifts towards lower energy in spin - up channel whereas , in spin - down channel it shifts towards higher energy .
because of this shift there is creation of band gap in the minority spin channel as is observed in co@xmath1mnge .
this band shift appears to be responsible for the presence of fm ground state in this compound .
however , in this compound the shift in tdos is very small in comparison to rigid shift observed in spin - up channel of co@xmath1mnge@xcite .
also the value of tdos at _ e@xmath3 _ is very large in comparison to tdos of co@xmath1mnge studied earlier and one can expect very large effect of on - site coulomb interaction ( _ u _ ) on such systems . from the pdos of fe atom ( fig .
5(b ) ) it is clear that _ t@xmath15 _
( @xmath20.5 states / ev / atom ) and _ e@xmath16 _ ( @xmath21.4 states / ev / atom ) states have main contribution at _ e@xmath3 _ for spin - up channel while the minority spin channel is empty . near the _ e@xmath3 _ in spin - dn channel , the valence band maximum has main contribution from _ t@xmath15 _ and conduction band minimum is contributed by _
e@xmath16 _ states .
the pdos of mn atom is presented in fig .
5(c ) and it is clear from this plot that _
t@xmath15 _ bands have occupation of @xmath21.0 states / ev / atom with negligibly small contribution from _
e@xmath16 _ band at _ e@xmath3 _ for spin - up channel .
the _ e@xmath16 _ band of mn atom has shifted towards lower energy with no contribution at the _
e@xmath3_. the pdos of si atom ( fig .
5(d ) ) shows negligibly small contribution from 3__s _ _ and 3__p _ _ orbitals .
the total and partial density of states plots of spin - up and spin - down channels for fe@xmath1mnal are shown in fig .
the spin - up channel of fig .
6(a ) shows that total density of states at _ e@xmath3 _ is @xmath21 states / ev / f.u . and
spin - down channel show very small ( @xmath20.1 states / ev / f.u . ) density of states at _
e@xmath3_. on comparing fig .
3(a ) and 6(a ) one can find that there is no rigid shifting of bands rather there is splitting of bands at _ e@xmath3 _ , which we have not observed in co@xmath1mnge and fe@xmath1mnsi .
the pdos of fe , mn and al atoms for both spin channels are shown in fig .
6(b - d ) . from figs .
3(b ) and 6(b ) it is clear that there is no rigid shift in band as per stoner theory as is observed in co@xmath1mnge@xcite .
e@xmath16 _ states split in such a way that there exists a minimum at _
e@xmath3 _ with contribution of @xmath20.3 states / ev / atom and 0.1 states / ev / atom , from _ t@xmath15 _ and _ e@xmath16 _ bands , respectively .
this minimum is responsible for the existence of the pseudo gap . in the minority spin channel
the valence band maximum has mainly _
t@xmath15 _ character and conduction band minimum has _
e@xmath16 _ character .
this compound show metallic nature for down spin also because of very small contribution from _
t@xmath15 _ bands at _ e@xmath3_. in fig .
6(c ) the pdos for mn _
t@xmath15 _ and _ e@xmath16 _ bands are shown .
the spin - up channel of pdos of mn atom is contributed by _
t@xmath15 _ state ( @xmath20.2 states / ev / atom ) with negligibly small contribution from _
e@xmath16 _ state at _
e@xmath3_. it is evident from fig .
6(d ) that in pdos of al atom there is negligibly small contribution of 3__s _ _ and 3__p _ _ orbitals .
the total magnetic moment per formula unit for fe@xmath1mnsi is 3.0 @xmath4 with contribution from fe , mn , si and interstitial region is 0.22 , 2.52 , -0.01 and 0.042 @xmath4 , respectively .
the similar value of total magnetic moment is also predicted theoretically in @xcite , whereas the experimental results @xcite gave a saturation magnetic moment less than this value .
this may be due to the reason that it is difficult to obtain the pure phases experimentally .
the mn atom coupled ferromagnetically with fe atom as is found by galanakis et al.@xcite .
the magnetic moment of si is very small and it is coupled antiferromagnetically with fe and mn atoms .
the total magnetic moment per formula unit for fe@xmath1mnal is 2.0 @xmath4 and contribution from fe , mn , al and interstitial region is -0.23 , 2.44 , -0.008 and 0.05 @xmath4 , respectively . the calculated value of the total magnetic moment matches with earlier reported value @xcite . the mn atom coupled antiferromagnetically with fe atom in this compound
. by using the full - potential screened korringa - kohn - rostoker ( fskkr ) green s function method in conjunction with the local spin density approximation galanakis et al.@xcite
have also found similar results . but fujii et al .
@xcite have observed ferromagnetic coupling between fe and mn atoms which is contradictory to our results .
the magnetic moment of al is very small but it is coupled ferromagnetically with fe atom and antiferromagnetically with mn atom .
the mn atom carries the largest magnetic moment in all these compounds , while the fe atom of fe@xmath1mnsi and fe@xmath1mnal carry the modest magnetic moment which is in agreement with the earlier reported results @xcite .
from the study of electronic structure of fe@xmath1mnsi it is clear that for spin - up channel there is very large density of states at _ e@xmath3 _ and spin - down channel is empty .
the electronic structure of fe@xmath1mnal show comparatively small value of tdos from 3__d _ _ bands at _ e@xmath3 _ for both spin channels .
it is well known that 3__d _ _
bands are less dispersive therefore on - site coulomb interaction can have very large effect on the electronic and magnetic properties of such compounds .
the lsda+__u _ _ method may bring a drastic change on the magnetic properties and electronic properties of these compounds .
so we have study the effect of _ u _ varying from 1 to 5 ev on the magnetic moment and electronic structure of these compounds along with previously studied co@xmath0mnge compound @xcite .
firstly , we discuss the effect of _ u _ on the magnetic moments of all these compounds . the total number of 3__d _ _ electrons in both spin channels and local magnetic moments in the presence of _ u _ are presented in table 1 for fe@xmath1mnsi , fe@xmath1mnal and co@xmath1mnge .
it is clear from table that in fe@xmath1mnsi , the number of 3__d _ _ electrons of fe atom in up spin channel increases while that of down spin channel decreases with _
u_. however the total number of electrons remains fixed to the value @xmath26.1 .
the local magnetic moment of fe atom is found to increase with increasing the value of _ u_. the magnetic moment increases from 0.29 @xmath4 at _ u _ = 1 ev to 0.87 @xmath4 at _ u _ = 5 ev . in mn atom , the total number of 3__d _ _ electrons in spin - up channel decreases and spin - dn channel increases with _ u _ in such a way that total number of 3__d _ _ electrons remains fixed to the value @xmath24.9 .
also the magnetic moment of mn atom decreases from 2.41 @xmath4 at _
u _ = 1 ev to 1.29 @xmath4 at _ u _ = 5 ev .
the above results suggest that the hund s like exchange interactions between fe 3__d _ _ electrons are increasing and that of mn 3__d _ _ electrons decreasing with increase in _
u_. in fe@xmath1mnal , also the number of electrons of fe ( mn ) atom in up spin channel increases ( decreases ) while that of down spin channel decrease ( increases ) with _
u_. the value of local magnetic moment of fe atom also goes on increasing with increasing the value of _ u_. the value of magnetic moment increases from -0.14 @xmath4 at _ u _ = 1 ev to 0.55 @xmath4 at _ u _ = 5 ev . there is also a very anomalous effect of _ u _ on the magnetic moment of fe atom , as the magnetic moment changes its sign after _
u _ = 2 ev and coupled ferromagnetically with mn atom . on the other hand the magnetic moment of mn decreases from 2.24 @xmath4 at _ u _ = 1 ev to 0.77 @xmath4 at _ u _ = 5 ev . at _ u _ = 2 ev , the magnetic moment of fe atom become almost zero and the total magnetic moment is contributed only by the mn atom . in the fe atom the hund s like exchange interactions
are also found to increase with _ u _ while in mn atom decrease with _
u_. in comparison to fe@xmath1mnsi the effect of _ u _ on the magnetic properties of fe@xmath1mnal compound is found to be more .
the on - site coulomb interactions are affecting above two compounds drastically but we have observed no drastic effect on the magnetic moment of co@xmath1mnge . in this compounds the number of electrons in spin - up and spin - dn channels are no changing significantly with _
u_. the magnetic moment of co atom increases slightly from 1.07 @xmath4 at _ u _ = 1 ev to 1.15 @xmath4 at _
u _ = 5 ev .
the magnetic moment of mn atom decreases from 2.80 @xmath4 at _ u _ = 1 ev to 2.61 @xmath4 at _ u _ = 5 ev .
the value of _ u _ is affecting the magnetic properties of these compounds although it has very less effect on co@xmath1mnge .
therefore we have also studied the electronic structures of all these compounds in the presence of _ u_. the total and partial dos of fe@xmath1mnsi , fe@xmath1mnal and co@xmath1mnge are presented in fig . 7 - 9 only for two selected values of _ u _ i.e. _ u _ = 2 and 4 ev .
the spin - up channel of fig .
7(a ) shows that the tdos at _ e@xmath3 _ decrease very slowly with _
u _ and there exists a gap at _ e@xmath3 _ in down spin channel .
the value of band gap in spin - up channel increases from @xmath20.8 ev at _ u _ = 2 ev to @xmath20.9 ev at _ u _ = 4 ev .
tdos in the spin - up and spin - dn channels shifts towards the lower energy as the value of _ u _ is increased from 2 ev to 4 ev .
the pdos of fe atom is presented in fig .
7(b ) and it is evident from figure that for these two values _
u _ , _ e@xmath16
_ and _ t@xmath15 _ states are contributing to the tdos at _ e@xmath3 _ with more contribution from _
e@xmath16 _ states .
also the contribution from _
t@xmath15 _ and _ e@xmath16 _ states at _ e@xmath3 _ decreases from @xmath20.45 states / ev / atom at _ u _ = 2 ev to @xmath20.27 states / ev / atom at _ u _ = 4 ev and @xmath21.7 states / ev / atom at _ u _ = 2 ev to @xmath21.0 states / ev / atom at _ u _ = 4 ev , respectively . the contribution from both these states decrease with _ u _ and pdos appears to shift towards lower energy in the both spin channels .
the pdos of mn ( fig .
7(c ) ) at _ e@xmath3 _ is occupied by _
t@xmath15 _ states with negligibly small contribution from _
e@xmath16 _ states . also the occupancy of _
t@xmath15 _ states decreases very slowly with _
u _ and _ t@xmath15 _ and _ e@xmath16 _ states shift towards lower energy in spin - up channel . in the down spin channel
the _ t@xmath15 _ states also shift towards lower energy while , _
e@xmath16 _ states shift towards higher energy .
the pdos of si atom has negligibly small contribution from 3__s _ _ and 3__p _ _
orbitals for these values of _
u_. the tdos of fe@xmath1mnal for _ u _ = 2 ev and 4 ev are shown in fig .
8(a ) . for _
u _ = 2 ev , the soft gap is appearing in the spin - up channel and when value of _ u _ is increased this compound become semiconductor with band gap of @xmath20.7 ev . in the spin - dn channel
the band gap does not change significantly with _
u_. this is also observed in magnetic moment calculations , where at _ u _ = 2 ev , the magnetic moment of fe atom become almost zero and for _ u _ = 4 ev the value of magnetic moment is positive .
the tdos for both the spin channels shift towards higher energy at _ u _ = 4 ev . the pdos of fe , mn and al atoms shown in fig .
8(b - d ) for _ u _ = 2 ev and _ u _ = 4 ev . from pdos of fe atom it is clear that _
t@xmath15 _ and _ e@xmath16 _ states shifts towards higher energy for both the spin channels .
the pdos of mn atom show that in spin - up channel shifting of _
t@xmath15 _ states is dominating and in spin - dn channel shifting of _ e@xmath16 _
states is more dominating .
also shifting is taking place at higher rate in the conduction band .
the pdos of al atom has also negligibly small contribution from 3__s _ _ and 3__p _ _
orbitals for these values of _
u_. the tdos of co@xmath1mnge in the presence of _ u _ are shown in fig .
this figure shows that the spin - up channel of co@xmath1mnge is metallic for all values of _ u _ and down spin channel is semiconducting . as the value of _ u _
is increased from 2 to 4 ev , there is broadening of the band gap in the spin - dn channel .
the band gap is found to increase from @xmath21.0 ev at _ u _ = 2 ev to @xmath21.4 ev at _ u _ = 4 ev .
we have not observed any significant shifting in the tdos for both spin channels .
thus the on - site coulomb interactions are not playing a significant effect on this compound , which is also evident from the calculations of magnetic moments .
it is clear from pdos of co atom shown in fig .
9(b ) that at _ e@xmath3 _ , the _ t@xmath15 _ and _ e@xmath16 _ states contribute same for both values of _ u_. increasing value of _ u _ is not going to affect the pdos of co atom .
the pdos of mn atom ( fig .
9(c ) ) also shows no difference in the dos at _ e@xmath3 _ for these two values of _
9(d ) shows that contribution from 3__s _ _ and 3__p _ _ states to the tdos is negligibly small and also not affected by the on - site coulomb interactions . from the above results it is clear that the metallic ground state of fe@xmath1mnal compound changes directly to the semiconducting ground state when the value of _ u _ is increased , whereas ground state of other two heusler alloys remain half - metallic ferromagnetic .
very few experimental work related to fe@xmath1mnal compound is found in the literature . in order to verify our predicted result
it is necessary to perform electrical conductivity and neutron diffraction experiments , which directly probe the electronic transport behaviour and the nature of magnetic coupling between the fe and mn moments . to the best of our knowledge
there are no experimental data on electrical conductivity in the @xmath12 phase
. however liu et al .
@xcite have reported resistivity data of the compound when it is in the b@xmath1 phase , which show insulating behaviour at the low temperature .
similarly neutron diffraction experiments on fe@xmath1mnsi compound are also desirable to know the magnetic moments of the fe and mn atoms . these experiments will help in understanding the role of on - site coulomb interaction among the 3__d _ _ electrons of fe and mn by studying the magnitude and directions of the magnetic moments of fe and mn atoms in the fe@xmath1mnsi and fe@xmath1mnal compounds .
the full - potential linearized augmented - plane wave methods have been employed to study the electronic and magnetic properties of fe@xmath0mnsi , fe@xmath0mnal and co@xmath0mnge .
the ferromagnetic ( fm ) solutions without using on - site coulomb interaction _ u _ show the presence of half - metallic fm ground state in fe@xmath0mnsi however , in fe@xmath0mnal the ground state is found to be metallic .
the total magnetic moment is contributed by mn atom with small contribution from fe atom in both cases .
the electronic and magnetic properties of fe@xmath0mnsi and fe@xmath0mnal are affected significantly by _ u _ , whereas the almost negligible effect of _ u _ is found in co@xmath1mnge .
the magnetic moment of fe atom in fe - based compounds is found to increase with _ u _ and for mn atom its value decreases . in fe@xmath1mnsi
the fe and mn moments are coupled ferromagnetically for all values of u , whereas in fe@xmath1mnal they coupled antiferromagnetically below u = 2 ev and ferromagnetically above it .
the study of electronic structures show that in fe@xmath0mnsi and co@xmath1mnge the ground state remains half - metallic ferromagnetic for all values of _ u _ , whereas in fe@xmath0mnal compound the ground state become semiconducting after _ u _ = 2 ev .
the ferromagnetic coupling between fe and mn moments appears to be responsible for this in fe@xmath0mnal compound .
r. a. degroot , f. m. mueller p. g. van engen and k. h. j. buschow , phys .
lett . * 50 * , 2024 ( 1983 ) .
k. e. h. m. hanseen and p. e. mijnarends , phys .
b * 34 * , 5009 ( 1986 ) . m. m. kirillova , a. a. makhnev , e. i. shreder , v. p. dyakina and n. b. gorina , phys .
status solidi b * 187 * , 231 ( 1995 ) .
i. galanakis , p. h. dederichs and n. papanikolaou , phys .
b * 66 * , 134428 ( 2002 ) .
j. k@xmath18bler , a. r. williams and c. b. sommers , phys .
b * 28 * , 1745 ( 1983 ) .
s. ishida , s. fujii , s. kashiwagi and s. asano , j. phys .
soc . jpn . *
64 * , 2152 ( 1995 ) .
s. fujii , s. ishida and s. asano , j. phys .
* 64 * , 185 ( 1995 ) .
s. picozzi , a. continenza and a. j. freeman , phys .
b * 66 * , 094421 ( 2002 ) .
j. barth , g. h. fecher , b. balke , s. ouardi , t. graf , c. felser , a. shkabko , a. weidenkaff , p. klaer , h. j. elmers , h. yoshikawa , s. ueda and k. kobayashi , phys .
b * 81 * , 064404 ( 2010 ) .
a. candan , g. u@xmath19ur , z. charifi , h. baaziz and m. r. ellialt@xmath20o@xmath19lu , j. alloys compd . *
560 * , 215 ( 2013 ) .
p. j. webster , j. phys .
solids * 32 * , 1221 ( 1971 ) .
k. r. a. ziebeck and p. j. webster , philos . mag . * 34 * , 973 ( 1976 ) .
m. kawakami , physica b * 186 - 188 * , 1037 ( 1993 ) .
s. fujii , s. ishida and s. asano , j. phys .
jpn . * 63 * , 1881 ( 1994 ) .
b. hamad and q. m. hu , phys .
status solidi b * 248 * , 2893 ( 2011 ) .
h. c. kandpal , g. h. fecher and c. felser , phys .
b * 73 * , 094422 ( 2006 ) .
s. wurmehl , g. h. fecher , h. c. kandpal , v. ksenofontov , c. felser , h. -j .
lin and j. morais phys .
b * 72 * , 184434 ( 2005 ) .
m.t . czy@xmath21yk and g. a. sawatzky phys .
b * 49 * , 14211 ( 1994 ) .
d. p. rai , sandeep , m. p. ghimire and r .k .
thapa , bull.mat.sc . * 34 * , 1219 ( 2011 ) .
d. p. rai and r .k .
thapa , j. alloys and comp . *
542 * , 257 ( 2012 ) .
s. ishida , d. nagatomo , s. fujii and s. asano , mater .
49 * , 114 ( 2008 ) .
p. blaha , k. schwarz , g. k. h. madsen , d. kvasnicka , and j. luitz , wien2k , an augmented plane wave plus local orbitals program for calculating crystal properties ( vienna university of technology , vienna , 2001 ) .
j. p. perdew and y. wang , phys .
b * 45 * , 13244 ( 1992 ) .
j. brown , k. u. neumann , p. j. webster and k. r. a. ziebeck , j. phys .
: condens .
matter * 12 * , 1827 ( 2000 ) .
h. c. kandpal , g. h. fecher and c. felser j. phys .
* 40 * , 1507 ( 2007 ) .
b. hamad , j. khalifeh , i. a. aljarayesh , c. demangeat , h. -b .
luo and q. -m .
hu , j. appl
. phys . * 107 * , 093911 ( 2010 ) .
i. galanakis , p. h. dederichs and n. papanikolaou , phys .
b * 66 * , 174429 ( 2002 ) .
k. ueda , k. hamaya , k. yamamoto , y. ando , t. sadoh , y. maeda and m. miyao , appl .
. lett . * 93 * , 112108 ( 2008 ) .
l. hongzhi , z. zhiyong , m. li , x. shifeng , l. heyan , q. jingping , l. yangxian and w. guangheng , j. phys .
d * 40 * , 7121 ( 2007 ) .
s. plogmann , t. schlath@xmath22lter , j. braun , m. neumann , yu .
m. yarmoshenko , m. v. yablonskikh , e. i. shreder , e. z. kurmaev , a. wrona and a. @xmath23 lebarski , phys .
b * 60 * , 6428 ( 1990 ) .
z. liu , x. ma , f. meng and g. wu , j. alloys compd . *
509 * , 3219 ( 2011 ) .
+ & & & + [ 0.5ex ] compound&_u _ ( ev ) & 3__d__@xmath24 & 3__d__@xmath25 & m(@xmath26 ) & 3__d__@xmath24 & 3__d__@xmath25 & m(@xmath26 ) + + & 1&3.21&2.92&0.29&3.67&1.26&2.41 + & 2&3.25&2.89&0.36&3.62&1.32&2.30 + fe@xmath1mnsi&3&3.32&2.82&0.50&3.49&1.44&2.05 + & 4&3.39&2.74&0.65&3.34&1.59&1.75 + & 5&3.50&2.63&0.87&3.11&1.82&1.29 + [ 1 ex ] & 1&2.97&3.10&-0.14&3.55&1.31&2.24 + & 2&3.03&3.03&0.00&3.41&1.44&1.97 + fe@xmath1mnal&3&3.11&2.94&0.17&3.24&1.60&1.64 + & 4&3.20&2.84&0.36&3.03&1.81&1.22 + & 5&3.29&2.74&0.55&2.80&2.03&0.77 + [ 1 ex ] & 1&4.16&3.09&1.07&3.91&1.11&2.80 + & 2&4.16&3.08&1.08&3.89&1.13&2.76 + co@xmath1mnge&3&4.17&3.06&1.11&3.88&1.15&2.73 + & 4&4.17&3.05&1.12&3.85&1.17&2.68 + & 5&4.18&3.03&1.15&3.82&1.21&2.61 + [ 1 ex ] | the electronic band structures , density of states plots and magnetic moments of fe@xmath0mnsi , fe@xmath0mnal , and co@xmath0mnge are studied by using the first principles calculations .
the fm solutions using lsda without _ u _ show the presence of half - metallic ferromagnetic ( hfm ) ground state in fe@xmath0mnsi , whereas the ground state of fe@xmath0mnal is found to be metallic . in both compounds
the maximum contribution to the total magnetic moment is from the mn atom , while the fe atom contributes very less .
the electronic structures and magnetic moments of fe - based compounds affected significantly by _ u _ , whereas its effect is very less on co@xmath1mnge .
the magnetic moment of fe atom in fe@xmath0mnsi ( fe@xmath0mnal ) , increased by @xmath2 70 % ( @xmath2 75 % ) and in mn atom it decreases by @xmath2 50 % ( @xmath2 70 % ) when the value of _ u _ is increased from 1 to 5 ev .
the hund s like exchange interactions are increasing in fe atom while decreasing in mn atom with increase in _
u_. the fe and mn moments are ferromagnetically coupled in fe@xmath0mnsi for all values of _ u _ , whereas in fe@xmath0mnal they coupled antiferromagnetically below _
u _ = 2 ev and ferromagnetically above it . above _ u _ = 2 ev the metallic ground state of fe@xmath0mnal changes to semiconducting ground state and the ferromagnetic coupling between fe and mn atoms appears to be responsible for this . |
the formation and evolution of structure in the universe is a fundamental field of research in cosmology .
clusters of galaxies represent the most extreme deviation from initial conditions , and are therefore good evolutionary probes for studying the formation of the large - scale structure .
while clusters of galaxies have been studied extensively in the relatively nearby universe , their evolutionary history becomes obscure beyond roughly half the hubble time ( e.g. blakeslee et al .
2006 , mullis et al . 2005 , stanford et al .
their progenitors are difficult to identify when the density contrast between the forming cluster and the field becomes small , and mass condensations on the scales of clusters are extremely rare at any epoch ( kaiser 1984 ) .
while deep pencil beam surveys have allowed us to study many different kinds of galaxies and their relative contributions to the cosmic star formation rate out to @xmath3 , the study of the evolution of galaxy clusters has not progressed much beyond @xmath4 .
however , the relatively advanced evolutionary state of some clusters even at @xmath5 suggests a formation epoch at @xmath2 , consistent with predictions for the ( early ) formation of large - scale structure in cosmological simulations ( springel et al .
2005 ) . finding and studying the progenitors of these clusters will give new clues to how the most massive structures in the universe came about .
distant radio galaxies are important laboratories for studying the formation and evolution of massive galaxies in the early universe .
they are among the most luminous and largest known galaxies at @xmath6 , having stellar masses in excess of @xmath7 @xmath8 ( e.g. de breuck et al .
2002 , villar - martin et al .
there is evidence that protoclusters in the early universe ( @xmath9 ) can be found around ( but are not limited to ) radio galaxies ( e.g. kurk 2000 , 2004 ; miley et al .
2004 ; overzier et al . 2006a,2006b ; pentericci et al .
2000,2002 ; venemans et al .
2002 , 2004 , 2005 ) .
these protoclusters have galaxy and mass overdensities that are sufficiently large for such structures to break away from the expanding universe , and form a virialized cluster before @xmath10 .
note that although powerful 3c - type radio sources are very rare objects , their number density is comparable to that of abell clusters ( @xmath11@xmath12 mpc@xmath13 ) if estimated radio source life - times ( @xmath14 yr ) are taken into account ( bahcall & soneira 1983 , dunlop & peacock 1990 , west 1994 ) .
it is believed that the fueling of a supermassive black hole due to cooling flows , mergers and interactions are responsible for triggering a powerful ( radio ) quasar ( e.g. heckman et al .
1986 , kauffmann & haehnelt 2000 , canalizo & stockton 2001 , croton et al .
if this indeed provides the trigger for radio activity , then the cores of protoclusters are likely locations to find radio galaxies at high redshift . taken together with the highly special intrinsic properties of the host galaxies of distant luminous radio sources ,
it is probable that radio galaxies could be progenitors of the brightest cluster galaxies ( bcgs ) .
the characteristic luminosity of radio sources increases with redshift in a flux limited sample , meaning that one could be observing an increasing fraction of proto - bcgs with increasing redshift , while at low redshift ( @xmath15 ) 3c sources span a factor of @xmath16 in radio luminosity and most of these will therefore not correspond to bcgs ( west et al .
we obtained deep images of the radio galaxy mrc 1138262 with the _ hubble space telescope _
advanced camera for surveys ( acs , ford et al .
1998 ) , as part of an acs gto programme to map the central regions of radio galaxy protoclusters in order to study star formation and galaxy morphologies in overdense environments at @xmath18 . by coadding the images taken through the filters @xmath19 and @xmath20 ( both in the rest - frame uv ) we created a 44 ksec image of a @xmath21 kpc@xmath22 region surrounding the radio galaxy ( fig .
1 ) . initially studied in detail by pentericci et al .
( 1997,1998,2001 ) , the increased depth of the acs image reveals more fine details and complexity in this remarkable object ( miley et al .
2006 ) : image of the radio galaxy mrc 1138262 at @xmath0 .
the system consists of a large ( @xmath23 kpc ) conglomeration of sub - galactic clumps embedded in a region of diffuse emission .
the total star formation rate in the clumps and the diffuse component ( several hundred @xmath8 yr@xmath24 ) are about equal .
comparison with the rest - frame optical magnitude ( pentericci et al .
1997 ) indicates that the mass of the star - forming component seen in this image is only a small fraction of the total mass of the system , suggesting that the mrc 1138262 system has elements of both early and late formation ( see miley et al .
2006 for details ) .
contours show the radio emission at 4.5 ghz .
the postage stamps to the right of the image show close - ups of objects having interesting morphologies ( marked with the letters a e in the image ) . ]
the optical continuum emission of the galaxy consists of at least 10 distinct clumps , presumably satellite galaxies that are still merging with mrc 1138262 .
several of the faint satellites have interesting morphologies .
object ` a ' in fig .
1 shows four bright knots of emission in a line like a string of beads .
these longitudinal galaxies that are barely resolved in their transverse dimension are known as ` chain ' galaxies , and could be distant , ancestral forms of local hubble type galaxies ( cowie et al .
they are believed to form from star formation progressing along filaments of gas previously blown out from massive forming galaxies ( taniguchi & shioya 2001 ) . because the chain galaxies are believed to be inherently unstable and therefore short - lived
, the chance of finding one could have been enhanced by the vigorous merging and gaseous outflows observed in the mrc 1138262 system .
the objects marked ` c ' and ` d ' in fig .
1 can be classified as ` tadpole ' galaxies , which consist of a knot at one end and an extended tail at the other .
these objects are believed to be early - stage mergers ( straughn et al .
other objects also show evidence for disruption , multiple nuclei , and tidal tails .
interestingly , faint diffuse emission is visible throughout the @xmath25 kpc inter - clump region .
the extended emission is unlikely to be caused by scattered light of the obscured quasar nucleus , because its roughly spherical morphology and radial profile do not resemble that of a scattering cone .
the mean colour of the diffuse emission is comparable to that of the star - forming satellites , consistent with the occurrence of ongoing star formation over a very extended region .
the small - scale variations in the surface brightness of the extended emission seen to the south - west of the nucleus could be caused by gradients in the amount of dust , or by the subclustering of star - forming clumps too faint to be detected individually .
the total extended luminosity ( comprising 45% of the total emission in @xmath19 ) corresponds to a star formation rate of @xmath26100 @xmath8 yr@xmath24 .
this indicates that star formation over extended regions such as observed in mrc 1138262 could be important for building the most massive galaxies , besides merging and gas accretion .
contours ( blue ) of the giant emission line nebula surrounding mrc 1138262 , and vla 8ghz radio contours ( red ) of the non - thermal radio structure superimposed on the composite ( @xmath19 + @xmath20 ) acs image .
the gaseous nebula extends for @xmath27200 kpc . ]
the galaxy is surrounded by a giant ( @xmath28 kpc ) emission line nebula ( fig .
2 ) , similar to the large low surface brightness halos of ionized gas seen around other high redshift radio galaxies . these halos could be remnant reservoirs of gas from which the massive galaxies are forming , or they could be due to gas swept up by previous episodes of starburst superwinds or large - scale agn outflows ( e.g. haiman et al .
2000 , taniguchi & shioya 2000 , haiman & rees 2001 , villar - martn et al .
when a powerful source of ionizing radiation becomes active ( e.g. , agn or starburst ) the halo will begin to emit luminous line emission , and high velocities of gas can be observed in regions where it is perturbed by interaction with the radio structure .
the size of the halos is comparable to that of the diffuse envelopes of cd galaxies in the local universe , but it is not clear whether these are related phenomena .
we interpret the new hst image as showing hierarchical merging in a proto - bcg .
the morphological complexity of the mrc 1138262 system agrees , at least qualitatively , with predictions of massive galaxy formation ( e.g. dubinsky 1998 , gao et al .
the extensive merging provides a plausible mechanism for fueling the radio source , which may subsequently interact with its surrounding medium to regulate its own growth by preventing gas accretion from the larger halo when the source is ` on ' ( croton et al .
evidence for this kind of coupling between the radio structure and the surrounding medium has been revealed through recent integral field spectroscopy of the gaseous halo surrounding mrc 1138262 ( fig .
2 ) . nesvadba et al .
( 2006 ) found very high gas velocities and velocity dispersions associated with the radio structure .
the energy of the outflow is sufficiently high to expel a significant fraction of the mass in gas during the lifetime of the radio source .
after the star formation has switched off , the galaxy will continue to grow through merging ( de lucia & blaizot 2006 ) .
however , the total stellar mass contained in the population of satellites is believed to be small compared to the mass of the central galaxy ( @xmath29 @xmath8 ) , estimated from its @xmath30-band magnitude ( pentericci et al .
this may indicate that mrc 1138262 is an exceptionally rare object that accumulated a large amount of its mass within a relatively short time at @xmath31 .
the observations presented here provide a unique testbed for simulations of forming massive galaxies at the centers of clusters and protoclusters .
bahcall , n. a. , & soneira , r. m. 1983 , apj , 270 , 20 blakeslee , j. p. , et al .
2006 , apj , 644 , 30 canalizo , g. , & stockton , a. 2001 , apj , 555 , 719 cowie , l. l. , hu , e. m. , & songaila , a. 1995 , aj , 110 , 1576 croton , d. j. , et al .
2006 , mnras , 365 , 11 de breuck , c. , van breugel , w. , stanford , s. a. , rttgering , h. , miley , g. , & stern , d. 2002 , aj , 123 , 637 de lucia , g. , & blaizot , j. , 2006 , mnras , submitted dubinski , j. 1998 , apj , 502 , 141 dunlop , j. s. , & peacock , j. a. 1990 , mnras , 247 , 19 ford , h. c. , et al .
1998 , proc .
spie , 3356 , 234 gao , l. , loeb , a. , peebles , p. j. e. , white , s. d. m. , & jenkins , a. 2004 , apj , 614 , 17 haiman , z. , spaans , m. , & quataert , e. 2000 , apj , 537 , l5 haiman , z. , & rees , m. j. 2001 , apj , 556 , 87 heckman , t. m. , smith , e. p. , baum , s. a. , van breugel , w. j. m. , miley , g. k. , illingworth , g. d. , bothun , g. d. , & balick , b. 1986 , apj , 311 , 526 kaiser , n. 1984 , apj , 284 , l9 kauffmann , g. , & haehnelt , m. 2000 , mnras , 311 , 576 kurk , j. d. , et al .
2000 , a&a , 358 , l1 kurk , j. d. , pentericci , l. , overzier , r. a. , rttgering , h. j. a. , & miley , g. k. 2004 , a&a , 428 , 817 miley , g. k. , et al .
2004 , nature , 426 , 47 miley , g. k. , et al .
2006 , apj , 650 , l29 mullis , c. r. , rosati , p. , lamer , g. , bhringer , h. , schwope , a. , schuecker , p. , & fassbender , r. 2005 , apj , 623 , l85 pentericci , l. , rttgering , h. j. a. , miley , g. k. , carilli , c. l. , & mccarthy , p. 1997 ,
a&a , 326 , 580 nesvadba , n. , et al .
2006 , apj , in press overzier , r. a. , et al . 2006a , apj , submitted ( astro - ph/0601223 ) overzier , r. a. , et al .
2006b , apj , 637 , 58 pentericci , l. , rttgering , h. j. a. , miley , g. k. , spinrad , h. , mccarthy , p. j. , van breugel , w. j. m. , & macchetto , f. 1998 , apj , 504 , 139 pentericci , l. , et al .
2000 , a&a , 361 , l25 pentericci , l. , mccarthy , p. j. , rttgering , h. j. a. , miley , g. k. , van breugel , w. j. m. , & fosbury , r. 2001 , apj , 135 , 63 springel , v. , et al .
2005 , nature , 435 , 629 stanford , s. a. , et al .
2006 , apj , 646 , l13 straughn , a. n. , cohen , s. h. , ryan , r. e. , hathi , n. p. , windhorst , r. a. , & jansen , r. a. 2006 , apj , 639 , 724 taniguchi , y. , & shioya , y. 2000 , apj , 532 , l13 taniguchi , y. , & shioya , y. 2001 , apj , 547 , 146 venemans , b. p. , et al .
2002 , apj , 569 , l11 venemans , b. p. , et al .
2004 , a&a , 424 , l17 venemans , b. p. , et al .
2005 , a&a , 431 , 793 villar - martn , m. , et al .
2006 , mnras , 366 , l1 west , m. j. 1994 , mnras , 268 , 79 | we present deep observations taken with the hst advanced camera for surveys of the central massive galaxy in a forming cluster at @xmath0 . the galaxy hosting the powerful radio source mrc 1138262
is associated with one of the most extensive merger systems known in the early universe .
our hst / acs image shows many star - forming galaxies merging within a @xmath1200 kpc region that emits both diffuse line emission and continuum in the rest - frame uv . because this galaxy lives in an overdense environment , it represents a rare view of a brightest cluster galaxy in formation at @xmath2 which may serve as a testbed for predictions of massive cluster galaxy formation . early universe , galaxies : clusters : general |
let @xmath0 denote the heights of @xmath2 simple random walks on @xmath3 , conditioned on satisfying @xmath4 , \left| z^{(i+1)}_n - z^{(i)}_n \right| = 1.\ ] ] more precisely , the random walk is a markov chain on the state space of @xmath5-step walks in @xmath6 @xmath7 , |z^{(i+1 ) } - z^{(i)}| = 1 \}\ ] ] where the next step from @xmath8 is selected uniformly among the neighbours of @xmath9 in the usual lattice @xmath10 that belong to @xmath11 .
in other words , we consider @xmath2 simple random walks on the lattice @xmath3 coupled under a shape condition . as in the case of a simple random walk ,
the rescaled trajectory of a walker , say @xmath12 , will converge in law to a brownian motion .
however , it is interesting to note that the constraint between each coordinate only slow down the walk by decreasing its variance .
since it is classical to illustrate for our students the simple random walk as the motion of a drunk man , we can illustrate the previous mathematical fact by considering the random walk as the motion of a chain of prisoners .
it should convince even non mathematicians that the motion of the walk is slowed by the constraint .
however it seems very hard to guess the variance from this comparison !
more precisely , denote @xmath13 where @xmath14 is the integer part of the real number @xmath15 [ t : var ] the rescaled random walk @xmath16 converges in law , as @xmath17 goes to infinity , to a brownian motion with variance @xmath18 convergence to the brownian motion is the usual invariance principle : the noteworthy statement here is that it is possible to give an explicit expression for the limit diffusivity of the process , and that its expression is particularly simple .
our object of interest , the motion of @xmath19 is a non - markovian process that falls into the class of random walks with internal structure .
related questions of limit diffusivity for random walks conditioned to respect some geometric shape have been studied in the literature , under the name of `` spider random walks '' or `` molecular spiders '' , see @xcite .
the computation of the limit diffusion coefficient is also a central aim there , although the model and methods are different .
our initial motivation was however more remote .
actually we first addressed this question starting from combinatorial problems related to @xmath20-vertex model in relation with the razumov stroganov conjecture ( see @xcite for instance ) .
the problem can also be related to random graph - homomorphisms ( see @xcite ) or the square ice model ( see @xcite where @xmath12 evolves on a torus . ) roughly speaking we can say that , in the literature we read , the authors consider questions related to the uniform distribution on a sequence of finite graphs @xmath21 and wonder about various asymptotics when @xmath22 later in the article the evolution of @xmath23 will be described as the simple random walk on a graph @xmath24 hence on the one hand our problem is a very simplified version of the problems stated above , on the other hand we were surprised to have such a simple formula for @xmath25 which is true for all @xmath26 and not only for @xmath22 we thought in the beginning that the proof of this fact should be simple but it turns out that , although elementary , the tools used to obtain the result are more sophisticated than expected .
it is the aim of this note to show these tools . to prove the theorem
, we will look for a decomposition @xmath27 where @xmath28 is a martingale , and where @xmath29 is a bounded function .
we will then show that the following limit exists : @xmath30.\ ] ] and that it is indeed the desired diffusivity .
the path to this conclusion is akin to classical results for central limit theorems for markov chains ( e.g. @xcite ) .
we will use another equivalent , albeit more geometric , point of view on this decomposition .
we split the chain in two parts : on the one hand , the motion of one of the walkers , and on the other hand , the relative positions of the walkers ( which we call the `` shape '' of the chain at a given time ) .
the latter part is a markov chain over the state space @xmath31 and our quantity of interest is ( almost ) an observable of this chain .
computing the martingale decomposition that we wish for amounts to decomposing a discrete vector field over this new state space into a divergence - free part ( corresponding to the martingale part ) and a gradient part ( corresponding to the function @xmath32 ) . owing to a particular geometric property of this vector field ,
for which we coin the term `` stationarity '' , it is indeed possible to perform this calculation explicitly .
let us denote @xmath33 , y_n^{(i ) } = z_n^{(i+1)}- z_n^{(i)},\ ] ] and @xmath34 here @xmath35 describes the shape of @xmath0 , i.e. the position of each @xmath36 relatively to the previous one , and belongs to @xmath37 , whereas @xmath38 can be seen as the height of the first walker .
obviously , the evolution of the chain of walkers may be described by the variables @xmath39 . for a convenient analysis
, we will represent our process as the simple random walk on a ( multi)-graph @xmath40 , which we define below . set @xmath41 : the multi - graph @xmath40 is given as a triplet @xmath42 where @xmath43 are two edge sets , called respectively the set of positive `` and ' ' negative edges .
a couple @xmath44 , @xmath45 , belongs to @xmath46 if the vector @xmath47 has nonzero entries of alternating signs , with the first one negative .
moreover , @xmath48 also contains a loop from each @xmath49 to itself , noted @xmath50 .
likewise , @xmath51 contains those couples @xmath44 , @xmath45 , such that @xmath47 has nonzero entries of alternating sign , with the first one positive , and self - loops noted @xmath52 for each @xmath49 .
set @xmath53 .
finally , we consider the following function on @xmath54 [ def : a ] let @xmath55 be the function that takes the value @xmath56 on @xmath48 and @xmath57 on @xmath51
. note that we have , for any @xmath49 , @xmath58 .
let @xmath59 be the simple random walk on @xmath40 .
the processes @xmath60 and @xmath61 have the same distribution .
it is sufficient to prove that the two markov chains ( taking values in @xmath62 ) @xmath60 and @xmath61 have the same transition matrix .
we claim that @xmath63 recall that @xmath64 , \left| z^{(i + 1 ) } - z^{(i ) } \right| = 1 \right\}$ ] .
denote @xmath65 , \left| z_1^{(i ) } - z_2^{(i ) } \right| = 1 \right\}.\ ] ] since @xmath66 and @xmath67 are both simple random walks , the corresponding transition matrices are given respectively by @xmath68 denote now @xmath69 so that @xmath70 . to prove the proposition , it is sufficient to prove that @xmath71 where @xmath72 ( and @xmath73 is the sign function ) .
first note that , if @xmath74 , and if @xmath75 , z_1^{(i ) } = z_2^{(i ) } - \epsilon$ ] , then @xmath76 .
moreover , we have in this case @xmath77 , and @xmath78 as claimed . on the other hand , if @xmath77 and @xmath79 , then @xmath75 , z_1^{(i ) } = z_2^{(i ) } + \epsilon$ ] .
for the chain , it means that , if @xmath80 then @xmath81 is @xmath82 with equal probability @xmath83 independently of @xmath84 .
assume now that @xmath85 .
we will prove that @xmath86 , where @xmath87 , i.e. @xmath88 , with nonzero entries of alternating signs , and a first one of the sign of @xmath89 .
indeed , there is a first index @xmath90 , \ ; \mbox{such that } \quad \delta(z_{1})^{(i ) } \neq \delta(z_2)^{(i ) } \}.\ ] ] there are two possible cases @xmath91 @xmath92 in the two cases we have @xmath93 furthermore , we have @xmath94 then if @xmath95 let us define @xmath96 , \ ; \mbox{such that } \quad \delta(z_2)^{(i ) } \neq \delta(z_1)^{(i ) } \},\ ] ] where we set by convention @xmath97 if the condition defining the infimum is never satisfied . using the same arguments
, we get @xmath98 by induction one can define @xmath99 , \ ; \mbox{such that } \quad
\delta(z_1)^{(i)}\neq \delta(z_2)^{(i ) } \},\ ] ] until @xmath100 we have @xmath101 by definition , we get @xmath102 where @xmath103 . on the other hand , if @xmath104 , where @xmath105 , then one can recover explicitly @xmath106 from the definition of @xmath107
. moreover , the condition @xmath108 , \left| z_2^{(i ) } - z_1^{(i ) } \right|=1 $ ] is implied by the previous arguments ( following the definitions of the @xmath109 ) .
we may denote , @xmath110 when @xmath111 ( @xmath112 ) and @xmath113 when @xmath114 ( @xmath115 ) . for a general @xmath116
the previous enumeration of its neighbors is surprisingly complicated but we can provide some simple examples .
for instance if @xmath117 has only @xmath118 neighbors : @xmath119 , a & \rightarrow ( 1,\cdots , -1 , \cdots,1 ) , \end{aligned}\ ] ] where the @xmath57 is in the @xmath120-th position . note
now that the graph @xmath40 can also be described inductively : there are only six following possibilities for @xmath121 , described below : in the figure , we have used the concatenation notation : given a string @xmath122 , the string @xmath123 , resp .
@xmath124 , is obtained by adding a @xmath56 , resp . @xmath57 in front of @xmath122 . looking only at the @xmath125 cases such that @xmath126
, we can deduce the construction of @xmath127 from @xmath128 : @xmath129 figure 1 shows the first two graphs @xmath130 and @xmath131 . note that each edge of @xmath132 gives @xmath125 edges for @xmath133 , one on each facet @xmath134 , and @xmath135 and one crossing from the facet @xmath136 to the facet @xmath135 .
[ f : constr ] and @xmath131 . ]
we obtain the cardinality @xmath137 of @xmath138 ( as a multigraph ) by induction : @xmath139 we will also make use of the number @xmath140 of edges of the form @xmath141 which can be computed by induction : @xmath142 let us now describe vector fields on this graph .
in the previous section a function @xmath143 has been defined on edges of @xmath144 we will consider here @xmath143 as a vector field on @xmath145 let us first recall some classical definitions .
a vector field on @xmath146 is a function @xmath147 such that @xmath148 and such that , for any @xmath49 , @xmath149 .
we say that the vector field @xmath150 on @xmath40 is a gradient vector field if there exists a function @xmath32 on the vertices of @xmath40 such that for each edge @xmath151 , @xmath152 .
the gradient vector field associated with @xmath32 is denoted by @xmath153 .
the divergence of a vector field @xmath150 at point @xmath154 is defined by @xmath155 we say that a vector field @xmath150 is divergence - free if its divergence vanishes at all points
. we can endow the set of vector fields with a scalar product @xmath156 please note that the sum runs over all edges , including loops @xmath157 .
denote also , for any subsets @xmath158 of @xmath159 , @xmath160 the flux of @xmath150 going from @xmath161 to @xmath162 . note that a divergence free field @xmath163 verifies @xmath164 in analogy with the case of vector fields in euclidean spaces
, we can decompose any vector field on @xmath40 into the sum of a gradient vector field and a divergence - free field .
the following proposition is well - known .
[ prop_dec_usual ] let @xmath150 be a vector field on @xmath40 .
there exist a unique gradient vector field @xmath153 and a unique divergence - free field @xmath163 such that @xmath165 moreover @xmath166 the last identity simply means that gradient fields and divergence - free fields are orthogonal complements of each other in the vector space of vector fields over @xmath40 .
in our case we are interested in stationary vector fields .
[ def : stat ] a subgraph @xmath167 of the complete graph on @xmath168 is _ stationary _ , if the following holds . for @xmath169 and @xmath170 such that @xmath171 if @xmath172 is an edge of @xmath167 , then @xmath173 is an edge of @xmath167 .
a vector field @xmath150 defined on a subgraph of the complete graph on @xmath168 with a stationary domain is _ stationary _ if for all @xmath174 edges of @xmath167 , @xmath175 only depends on @xmath176 ( where @xmath168 is embedded in @xmath177 in an obvious way ) . note that , thanks to the construction of @xmath48 , it is stationary . moreover if @xmath178 and @xmath179 , then @xmath180 and thus the vector field @xmath181 taking values @xmath56 , resp .
@xmath57 , on @xmath182 , resp .
@xmath51 , is stationary .
so we may expect the gradient vector field @xmath183 in the hodge decomposition of @xmath143 to be stationary .
unfortunately if @xmath150 is a stationary vector field on @xmath40 and its decomposition is @xmath184 as per proposition [ prop_dec_usual ] , then @xmath153 is not always stationary .
nevertheless it turns out that the gradient vector field @xmath183 in the hodge decomposition of @xmath143 is actually stationary as it will be shown in the next section .
let us recall the definition [ def : a ] the vector field @xmath181 on @xmath40 is such that @xmath185 in this section our aim is to compute a function @xmath186 such that @xmath187 one can first remark that @xmath188 @xmath189 in the previous equation we used the notation @xmath190 for cardinality of sets .
we will introduce various notations related to other cardinalities @xmath191 then we need also to define for @xmath192 @xmath193 the number of the vertices @xmath194 such that @xmath195 with @xmath196 digits @xmath197 such that @xmath198 similarly @xmath199 is the number of the vertices @xmath194 such that @xmath200 and @xmath196 digits @xmath197 such that @xmath201 then we consider @xmath202 and @xmath203 similarly @xmath204 and @xmath205 let us consider the function @xmath206 on @xmath207 such that @xmath208 the last equation is trivial since @xmath209 is the number of the @xmath210s equal to @xmath211 and @xmath212 the number of @xmath210s equal to @xmath213 obviously we also get @xmath214 for any vertex @xmath122 in @xmath215 let us remark that for any function @xmath186 on @xmath159 @xmath216 since @xmath217 yields the sum of the digits of any vertex , we first observe that if @xmath218 and the number of digits @xmath197 such that @xmath219 is even then @xmath220 if the number of digits @xmath197 such that @xmath219 is odd and @xmath221 then @xmath222 one can then deduce that @xmath223 it turns out that if we consider the function @xmath224 on @xmath207 such that @xmath225 then @xmath226 we will prove by induction on @xmath227 to do that we split @xmath228 into the sum of two functions @xmath229 to proceed the induction argument we remark that for any vertex @xmath122 in @xmath159 @xmath230 we will then compute @xmath231 the easiest computation is @xmath232 since @xmath233 then @xmath234 to evaluate @xmath235 we use that @xmath212 is the number of digits equal to @xmath236 in @xmath237 if @xmath238 and if the number of @xmath239 is even then @xmath240 if this number is odd then @xmath241 therefore @xmath242 hence @xmath243 one also get in the same way @xmath244 the induction is a bit more involved for @xmath245 because of @xmath246 then @xmath247 hence @xmath248 similarly we get @xmath249 we can now evaluate the functions @xmath250 [ lem : phi_barphi ]
@xmath251 @xmath252 @xmath253 we will only sketch the proof performed by an easy induction on @xmath26 for @xmath254 computations are similar for @xmath255 let us assume that , hold for @xmath256 we have to compute @xmath257 and @xmath258 and check that they fulfill , for @xmath259 because of @xmath260 one can check that @xmath261 using for @xmath256 we get for @xmath262 and @xmath263 the computations for @xmath258 are left to the reader . by summing and
we get , and we deduce that if we take @xmath264 @xmath265 is divergence free
. please note that obviously the additive constant in is arbitrary but it yields the following convenient expression of @xmath266 in terms of the digits of @xmath267 [ lem : form - grad a ] @xmath268 where @xmath269 moreover @xmath183 is a stationary gradient vector field . since we already know
that @xmath270 it is enough to show by induction on @xmath26 that @xmath271 obviously the formula is true for @xmath272 because of and , @xmath273 hence is proved for @xmath274 owing to @xmath270 and , it is obvious that @xmath275 are stationary and consequently @xmath183 is a stationary gradient vector field .
we denote by @xmath276 the decomposition of @xmath143 as per proposition [ prop_dec_usual ] . back to the original problem , we recall that @xmath277 let us denote @xmath278 .
let @xmath279 , we have @xmath280 = ( \nabla \cdot b)(y_n ) = 0,\ ] ] and @xmath281 is a martingale .
we now sketch out how to apply the central limit theorem for markov chains .
let @xmath282 @xmath283 is a markov chain on @xmath54 then our quantity of interest @xmath284 is an additive observable of the process @xmath283 , as @xmath285 the central limit theorem for markov chains ( see e.g. @xcite ) shows that @xmath286 converges as @xmath287 to a brownian motion , with variance given by @xmath288 \\ & = \lim_{n \rightarrow + \infty } \frac{1}{n } \mathbb{e } \left[m_n^2\right]+ \lim_{n \rightarrow + \infty } \frac{1}{n } \mathbb{e } \left [ f(y_n)^2\right ] + \lim_{n \rightarrow + \infty } \frac{2}{n } \mathbb{e } \left[\left(f(y_n ) m_n \right)\right].\end{aligned}\ ] ] let us first compute @xmath289.$ ] we can remark that @xmath290 is a @xmath291 martingale , hence @xmath292=\lim_{n \rightarrow + \infty } \frac{1}{n } \mathbb{e } \left [ \sum_{i = 0}^{n-1 } \big(b(y_i , y_{i+1 } ) \big)^2\right].$ ] if @xmath293 denotes the invariant measure for the random walk @xmath294 , by ergodicity , we get @xmath295=\mathbb{e}_\mu \left[\left(b(y_0 , y_1)\right)^2\right].\ ] ] under @xmath296 the distribution of @xmath297 is uniform on @xmath298 because @xmath299 is uniformly chosen among all neighbours of @xmath300 hence @xmath301 using that @xmath32 does not depend on @xmath302 we obtain @xmath303 = 0,\ ] ] and by cauchy schwarz inequality and @xmath304=0.\ ] ] so we get @xmath305,\ ] ] and by ergodicity , @xmath306.\ ] ] since , under @xmath307 the distribution of @xmath297 is uniform on @xmath298 @xmath308 by orthogonality of @xmath163 and @xmath153 , @xmath309 thus , it remains to compute @xmath310 ( since @xmath311 , by definition ) . at this point
we use the fact that @xmath153 is a stationary field , in the sense of definition [ def : stat ] .
then if we denote by @xmath312 ( where @xmath57 is in @xmath120-th position ) and , for @xmath313 , we have by @xmath314 now we compute @xmath310 as a function of @xmath315 , the value @xmath153 on the edge @xmath316 . by @xmath317 if @xmath318 and @xmath319 , then using the relation between @xmath320 and @xmath321 @xmath322 by stationarity of @xmath323 @xmath324 then , because of the definition of @xmath325 and @xmath326 @xmath327 then by stationarity of @xmath153 @xmath328 in the proof , the way we guessed is a bit mysterious . assuming that @xmath183 is a stationary gradient vector field , the family @xmath329}$ ] can be computed considering the system of equations given by @xmath330 , j_{m_i , n_i } = 0,\ ] ] where @xmath331 } \in v_k , a_i = 1\right\}$ ] and @xmath332 } \in v_k , a_i = -1\right\}$ ] .
this leads to the following system : @xmath333 f = \left[\begin{matrix}3^{k-1 } \\ 3^{k-2 } \\
\vdots \\ 3 \\
1\end{matrix } \right].\ ] ] the unique solution is given by : @xmath334 even if the guess is correct , we did not find another way as the techniques used in the section [ sec : hodge - a ] to show that @xmath183 is a stationary gradient vector field .
in this final section we briefly outline some related problems .
* it is possible to make sense of the process when @xmath5 is infinite .
several questions arise : what happens to the process of one marked walker ?
is there a scaling limit under equilibrium for the `` shape''process @xmath335 ?
+ another natural step would be to let @xmath5 grow with @xmath336 in a suitable way , so as to get a scaling limit for the two - parameter process @xmath337 .
* one may also ask about different quantities , such as the diameter of the set of walkers under the invariant measure for the entire walk .
* one may also consider random walkers conditioned on satisfying different shape constraints , and on graphs more general than @xmath3 . as a starting example , what happens if we work on a torus , i.e. if we force also @xmath338 ? the shape chain changes in this case and it is no longer irreducible over @xmath339 ( one may check that the number of @xmath57 symbols is fixed , and that this enumerates the recurrence classes ) .
it is interesting to point out that this setup is the one chosen by e. lieb for the computation of the six - vertex constant in @xcite .
we warmly thank charles bordenave for fruitful conversations and useful comments . | we consider a random walk @xmath0 with the constraint that each coordinate of the walk is at distance one from the following one . in this paper , we show that this random walk is slowed down by a variance factor @xmath1 with respect to the case of the classical simple random walk without constraint .
_ keywords _ : random walk , graph , central limit theorem _ ams classification ( 2000 ) _ : 05c81 , 60f05 . |
in type iib string theory there exists an @xmath2 multiplet of strings labeled by their nsns and rr charges @xmath3 @xcite .
these strings are allowed to connect at a point or a `` junction '' as long as charge is conserved at the junction @xcite . of particular interest are bps saturated junctions @xcite , which preserve a fraction ( 1/4 ) of the supersymmetry , and are therefore guaranteed to be stable .
recent interest in string webs has focused on their application to quantum field theory @xcite .
one of the successes of string web technology was the first explicit construction of @xmath4 bps states in @xmath5 @xmath6 super - yang - mills ( sym ) theory @xcite .
in this example the dilaton and axion of type iib string theory are constant , and it is relatively straightforward to analyze the bps properties of string webs
. a richer set of examples involve a dilaton and axion which are position dependent due to the presence of 7-branes in the background .
however , the bps nature of string webs has only been analyzed in the 7-brane background corresponding to pure @xmath0 @xmath1 sym so far @xcite .
these examples display a number of interesting phenomena including `` marginal stability '' , whereby space ( moduli space in the world - volume field theory ) may contain regions in which string webs fail to be bps . in this paper
we study another interesting phenomenon , in which the stringy description of a bps state may undergo a transition , as one moves in moduli space , from either a single open string or a string web , to another string web . in a recent paper we showed how to represent bps states in @xmath0 @xmath1 sym theory using string webs @xcite .
the field theory is constructed as the world volume theory of a d3-brane in a 7-brane background .
bps states correspond to string webs with one ( or two in the case of vector multiplets ) string ending on the d3-brane and the others on the 7-branes . here
we show how in different regions of moduli space the string configuration representing a bps state changes from a string web to either a different string web or to a single string .
these transitions are driven by the string creation phenomenon @xcite , which we generalize here to multiple string creation between @xmath3 7-branes and @xmath7 strings .
this also allows us to generalize to some extent the `` s - rule '' @xcite , which limits the number of strings that can link a 7-brane and another string .
transitions of the type addressed here were first discussed in @xcite for the case where all strings end on 7-branes .
a closely related paper @xcite argues that for any given configuration of 7-branes a bps state has a unique stringy representative .
the results presented here , aside from their application to @xmath0 theories , are of independent interest in string theory .
in fact , the 3-brane plays little role in what follows except to give strings a place to end .
the rest of the paper is organized as follows . in section 2
we review and generalize the string creation effect and the s - rule . in section 3
we express the supersymmetry conditions on strings and string webs in terms of @xmath3-geodesics and geodesic webs , and relate the latter to constant phase curves of analytic functions . in section 4 we show that the complex plane ( moduli space ) is divided into four regions , and exhibit the transitions that occur as one moves from one region to another .
we include an appendix , in which we give relevant information about the analytic functions in question , including a few numerically obtained constant phase curves .
when a d@xmath8-brane crosses a mutually transverse d@xmath9-brane a fundamental string , whose ends terminate on the two d - branes , is created .
this phenomenon was originally discovered in @xcite , and is related by a chain of dualities to the creation of d3-branes between crossing 5-branes @xcite .
it has since been exhibited in a number of different ways @xcite . for @xmath10
the effect involves a d1-brane , _ i.e. _ @xmath11 string , and a d7-brane , _ i.e. _ @xmath12 7-brane 7-branes are objects on which @xmath3 strings can end .
] , in type iib string theory .
when the branes cross each other , a @xmath12 string is created between them . by @xmath13 invariance
one can generalize this result to the creation of a single @xmath3 string between a @xmath3 7-brane and an @xmath7 string , where @xmath14 . if @xmath15 the configuration can not be related by duality to any of the above , and the arguments in @xcite are not applicable . nevertheless , it is straightforward to generalize the string creation phenomenon to an arbitrary @xmath3 7-brane and an arbitrary ( transverse ) @xmath7 string .
we associate a branch cut to the @xmath3 7-brane in the transverse plane , which one can think of as the place where the following monodromy occurs : m_(p , q ) = ( cc 1-pq & p^2 + -q^2 & 1+pq ) .
[ monodromy ] when an @xmath7 string living in the transverse plane crosses this cut in a counter - clockwise manner it undergoes a monodromy transformation , ( c r + s ) m_(p , q ) ( c r + s ) = ( c r + s ) + ( ps - qr ) ( c p + q ) . the asymptotic charges of the string are therefore @xmath7 at one end , and @xmath16 at the other ( fig .
these can not change as we pull the string through the 7-brane .
however , since now the string does not cross the cut , something else must supply the additional charge .
we conclude then that @xmath17 @xmath3 strings are created between the string and the 7-brane , oriented in such a way as to conserve the charges ( fig .
= 4 in in the special case of @xmath14 we obtain the known result of single string creation . for @xmath15
one can use @xmath13 to relate the system to a @xmath12 7-brane and an @xmath18 string .
one can then think of the @xmath18 string as a bound state of @xmath19 d - strings with @xmath20 units of electric flux .
each of the d - strings will create a @xmath12 string when it crosses the 7-brane , giving a total of @xmath21 strings .
the orientation of the created strings is determined by charge conservation , but is also restricted by supersymmetry .
consider the case of d - branes .
the initial configuration , consisting of a d7-brane and a transverse d - string , is supersymmetric , and therefore the final configuration must also be supersymmetric . on the other hand
it can be shown @xcite that such a configuration will only be supersymmetric if the fundamental string is oriented in the right way .
it is a matter of convention which of the two orientations preserves supersymmetry .
the important point is that only one of them does .
charge conservation in the string creation process fixes this convention , and gives us the unique supersymmetric orientation .
this leads to a non - trivial selection rule , known as the `` s - rule '' @xcite , for configurations of 7-branes and transverse strings , and configurations related to them by dualities . consider a configuration of @xmath22 fundamental strings between a d7-brane and a transverse d - string ( fig .
supersymmetry restricts the orientation of the fundamental strings .
as the d - string crosses the d7-brane two things happen : the @xmath22 fundamental strings reverse orientation , and a single fundamental string oriented supersymmetrically is created .
the latter will annihilate with one of the former , leaving @xmath23 strings which are oriented so as to break supersymmetry ( fig .
the final configuration is therefore _ not _ supersymmetric , and therefore neither is the initial configuration , even though it seemed to be .
we conclude that supersymmetry restricts not only the orientation of strings between the branes , but also their number . in the above case
there can be at most _ one _ fundamental string between a d7-brane and a d - string .
more generally , the maximum number of @xmath3 strings between a @xmath3 7-brane and an @xmath7 string allowed by supersymmetry is @xmath17 .
= 4 in in the case of @xmath14 there are other ways of deriving this rule @xcite . in particular , in the dual configuration of a d0-brane and a d8-brane the `` s - rule '' can be understood as a pauli exclusion principle , since the ground state of a 0 - 8 string corresponds to a single fermion in the d0-brane world - volume .
alternatively , in the d2@xmath24d6 configuration one can derive the `` s - rule '' by lifting to m - theory , and finding all the holomorphic curves corresponding to a supersymmetric embedding of the m2-brane in a kaluza - klein monopole background . as an application of this selection rule
let us consider a configuration of a @xmath12 7-brane , a @xmath25 7-brane , and a @xmath11 7-brane , connected by a `` 3-string junction '' consisting of two @xmath12 strings , a @xmath25 string , and a @xmath11 string ( fig . 3b ) . since the @xmath25 and @xmath11 7-branes can be thought of as the quantum resolution of an orientifold 7-plane @xcite , this configuration corresponds to the quantum resolution of a string which begins on the d7-brane , reflects off the orientifold plane , and ends on the d7-brane ( fig .
( in the covering space this becomes a string between the d7-brane and its image ) . in this _ classical _ description , the usual massless ( vector ) ground state of the string
is removed by the orientifold projection , leaving a massive _ non - bps _ ground state .
this state is extremal in the sense that it is the lowest mass charged state . in the quantum picture
this is understood as follows .
the configuration in ( fig .
3b ) violates the `` s - rule '' , since it contains _ two _ @xmath12 strings between the d7-brane and the @xmath11 string .
it therefore breaks supersymmetry , and the corresponding state is non - bps .
one can compare this to the situation in which the d7-brane is replaced by a d3-brane .
in this case the massless vector survives the orientifold projection , and leads to an enhanced gauge symmetry @xmath26 when the d3-brane coincides with the orientifold 7-plane . in the quantum picture the `` 3-string junction '' with the d3-brane
satisfies the s - rule , and is therefore supersymmetric .
a @xmath3-geodesic between points @xmath27 and @xmath28 is defined as the path which minimizes the mass of a @xmath3 string stretched between @xmath27 and @xmath28 , @xmath29 where @xmath30 .
since the geodesic is independent of the overall sign of @xmath3 we shall assume , without loss of generality , that @xmath31 .
in a generic 7-brane background @xmath32 is position - dependent , and in fact undergoes monodromies as branch cuts are crossed .
the metric on the transverse ( complex ) plane is given by @xcite ds^2 = |f(z)dz|^2 , and is smooth everywhere .
the function @xmath33 is determined by the 7-brane background .
one can define a flat `` @xmath3-metric '' ds_(p , q ) = t_(p , q)ds = @xmath34 given by dw_(p , q ) = ( p+q)f(z)dz , [ pqcoord ] in which @xmath3-geodesics are straight lines , w_(p , q)(t ) = ct + d . a necessary condition for @xmath3 strings to preserve supersymmetry is that they lie along @xmath3 geodesics @xcite .
if in addition the 7-brane background is supersymmetric , this is a sufficient condition .
multi - pronged strings , and more generally string webs ( which can include several junctions and internal strings ) , arise quite naturally in type iib string theory .
the simplest example corresponds to the configuration of a fundamental string ending on a d - string . by charge conservation , this turns the d - string into a @xmath35 string , and one ends up with a @xmath36 three - pronged string . due to the @xmath13 symmetry of type iib string theory there is actually an entire @xmath13 multiplet of such objects , each of which satisfies the charge conservation condition @xmath37 .
by gluing several of these objects together one can form a rich variety of string webs .
a necessary condition for a string web to preserve supersymmetry is that it lies on a _
geodesic web_. geodesic webs consist of several connected @xmath3-geodesics , w_(p_i , q_i)(t_i ) = c_it_i + d_i , such that the phases of @xmath38 are the same for all @xmath39 , _ i.e. _ c_i = e^i |c_i| . [ slopes
] this reduces to the usual angle conditions in flat space @xcite .
this is not a sufficient condition however , as the string web may violate the `` s - rule '' ( or a generalization of it given in @xcite ) , as in the example of the previous section .
the background of interest consists of a @xmath11 7-brane at @xmath40 and a @xmath25 7-brane at @xmath41 .
the metric on the complex plane is given by ds^2 = 2| ^2()(z-1)^-1/12(z+1)^-1/12dz|^2 , where we have set @xmath42 as compared to @xcite .
we shall choose the branch cuts associated to the @xmath11 and @xmath25 7-branes to run along the real axis from @xmath40 to @xmath43 , and from @xmath41 to @xmath43 , respectively .
the monodromies associated with the cuts along @xmath44 $ ] and @xmath45 $ ] are respectively given by m_= ( rr -1 & 4 + 0 & -1 ) , m_(0,1 ) = ( rr 1 & 0 + -1 & 1 ) . as a result ,
the charge of the 7-brane at @xmath41 is @xmath25 when viewed from above the cut , and @xmath46 when viewed from below the cut .
the above metric is related to the solution of @xmath0 @xmath1 sym @xcite as follows @xcite : t_(p , q ) ds = |pda + qda_d| , [ swconnection ] where @xmath47 and @xmath48 are given by integrals of the meromorphic seiberg - witten differential over the two cycles of the auxiliary riemann surface .
these can be expressed in terms of hypergeometric functions ( eq . ( [ hypergeo ] ) , appendix ) . comparing with ( [ pqmetric ] ) and ( [ pqcoord ] ) we see that the coordinates in which a @xmath3 geodesic is straight are given by w_(p , q ) = pa(z ) + qa_d(z ) .
the connection with @xmath0 sym is not a coincidence , and in fact follows by considering a d3-brane probe in the above background @xcite . in particular ,
the bps states are realized as strings ending on the d3-brane @xcite . by supersymmetry
, these strings should lie on geodesics given by pa(z(t ) ) + qa_d(z(t ) ) = ct + d , [ gensolution ] where @xmath49 is the position of the d3-brane .
note however that @xmath3 and @xmath50 undergo monodromies as one crosses branch cuts .
we therefore define smooth multi - valued functions @xmath51 and @xmath52 by analytic continuation . in terms of these , the geodesic equation ( [ gensolution ] ) holds across cuts as well .
the bps spectrum of @xmath0 @xmath1 sym in the weakly coupled regime consists of @xmath53-boson vector multiplets carrying charges @xmath54 , and dyon hypermultiplets carrying charges @xmath55 .
the corresponding geodesics would have to be given by : 1 . a @xmath56-geodesic that begins on the d3-brane , goes around the two 7-branes , and ends on the d3-brane , 2(z(t ) ) = c(t-1/2 ) .
[ w_geodesic ] 2 .
@xmath55-geodesics that begin on the d3-brane and end on one of the 7-branes , 2n(z(t ) ) _
d(z(t ) ) = c(t-1 ) .
[ hyper_geodesic ] in both cases there is a single complex parameter @xmath57 .
these geodesics are therefore given by _
constant phase curves _ ( cpc ) of the appropriate analytic function appearing on the left hand side of ( [ w_geodesic ] ) and ( [ hyper_geodesic ] ) . in order to end on either 7-brane
, a geodesic of the second kind with @xmath58 would have to cross the cut along @xmath44 $ ] ( and thereby undergo the monodromy @xmath59 ) @xmath60 times if @xmath22 is even , and @xmath61 times if @xmath22 is odd , without ever crossing the cut along @xmath45 $ ] ( where it would undergo the monodromy @xmath62 ) . alternatively , the same states ( with the exception of @xmath63 in the second case ) could also be described by string webs connecting all three branes , and lying on geodesic webs consisting of three geodesics @xcite .
if we denote the positions of the d3-brane and junction by @xmath64 and @xmath65 respectively , the appropriate webs are given by @xmath66(z(t_2 ) ) & = & [ 2\tilde{a}+\tilde{a}_d](z_0)t_2 \\
\tilde{a}(z(t_3 ) ) & = & \big(\tilde{a}(z_0 ) - a(z_3)\big)t_3 + a(z_3)\nonumber\end{aligned}\ ] ] for the @xmath53-boson , and @xmath67(z(t_2 ) ) & = & [ 2\tilde{a}+\tilde{a}_d](z_0)t_2 \\
\mbox{}[2n\tilde{a}\pm\tilde{a}_{d}](z(t_3 ) ) & = & \big([2n\tilde{a}\pm \tilde{a}_d](z_0 ) - [ 2na\pm a_d](z_3)\big)t_3 + [ 2na\pm a_{d}](z_3)\nonumber\end{aligned}\ ] ] for the dyons . in the above
we have assumed that the geodesics begin on the branes and end at the junction , so @xmath68 corresponds to the position of the @xmath11 7-brane , @xmath25 7-brane , and d3-brane for @xmath69 respectively , and @xmath70 to the position of the junction .
consequently , we can drop the tilde from the functions evaluated at the positions of the branes .
note that the equations are independent of the positions of the 7-branes , since the corresponding functions @xmath71 vanish there .
the condition on the slopes ( [ slopes ] ) is satisfied provided that @xcite = 0 , > -2 , [ condition1 ] for both the @xmath53-boson and the dyons , and = 0 , > 1 [ wcondition2 ] for the @xmath53-boson , and = 0 , > 1 , [ dyoncondition2 ] for the dyons .
the conditions in ( [ condition1 ] ) imply that the junction ( @xmath65 ) must lie on the curve of marginal stability @xmath72 ( see appendix for definition of @xmath72 ) .
the first condition in ( [ wcondition2 ] ) , implies that the third geodesic in ( [ wweb ] ) _ i.e. _ the one beginning at the d3-brane ( @xmath73 ) and ending at the junction ( @xmath65 ) , is a cpc of the function @xmath51 .
similarly , the first condition in ( [ dyoncondition2 ] ) implies that the third geodesic in ( [ dyonweb ] ) is a cpc of the function @xmath74 . the second condition in both ( [ wcondition2 ] ) and ( [ dyoncondition2 ] )
implies that the d3-brane must lie _ outside _
@xmath72 ( see appendix b of @xcite ) . since the the other two geodesics begin on 7-branes , they too are cpc s , and we conclude that the entire geodesic web consists of three cpc s of equal phase , which begin at the location of the branes and end at a common point on @xmath72 .
uniqueness of the bps state demands that there be only one string configuration which describes it at a given point in moduli space .
this can either be a single string on a smooth geodesic , or a string web on a geodesic web . as the above geodesics and geodesic webs are associated with cpc s of analytic functions
, one can determine their existence by studying these curves .
specifically , one is interested in cpc s of the functions @xmath51 and @xmath75 .
these can be determined numerically from the explicit expressions ( [ hypergeo ] ) for @xmath76 .
we have plotted cpc s of the single - valued functions @xmath47 and @xmath77 for a few values of @xmath22 using mathematica ( fig .
6 , appendix ) .
the cpc s of the corresponding smooth functions @xmath51 and @xmath75 are obtained by gluing those of the appropriate single - valued functions at the cuts . from figure 6
it is apparent that all the cpc s of @xmath47 intersect @xmath72 , and then proceed to intersect the cut along @xmath45 $ ] .
a closed geodesic which goes around both 7-branes therefore does not exist .
this has also been shown in a different way in @xcite and in @xcite .
on the other hand , since all the cpc s cross @xmath72 before crossing any cuts , there exists a geodesic web of the form ( [ wweb ] ) everywhere _ outside _
@xmath72 @xcite .
this is consistent with the observation that the @xmath53-boson is stable everywhere outside @xmath72 .
the cpc s of @xmath48 and @xmath78 , _
i.e. _ @xmath63 , all end on the corresponding 7-brane .
this means that there exists a geodesic starting at any point in the complex plane and ending on the appropriate 7-brane .
this is consistent with the observation that the corresponding bps states exist everywhere in moduli space .
for @xmath79 the moduli space is divided according to the behavior of the cpc s into four distinct regions , which are separated by @xmath72 , the cut along @xmath44 $ ] , and the two cpc s corresponding to @xmath80 and @xmath81 ( fig . 4 ) . * in region i ( inside @xmath72 ) cpc s intersect the cut along @xmath45 $ ] .
as this would imply a monodromy @xmath62 , @xmath55-geodesics with @xmath79 do not exist in this region ( this has also been shown in @xcite ) .
in addition , geodesic webs can only exist when the d3-brane is _ outside _ @xmath72 , so they too do not exist in region i. this is consistent with the observation that there are no bps dyons with @xmath79 inside @xmath72 . * in regions ii and iii cpc s intersect the lower half and upper half of @xmath72 , respectively .
they then proceed to intersect the cut along @xmath45 $ ] .
smooth geodesics therefore do not exist .
in these regions the bps states are represented by string webs with a single junction point located on @xmath72 .
since the charge of the 7-brane at @xmath41 is different in the upper and lower half planes , the appropriate string webs will be different in regions ii and iii as well .
* in region iv cpc s intersect the cut along @xmath44 $ ] .
the string therefore undergoes the monodromy @xmath59 .
for @xmath82 it becomes a @xmath11 string which then follows a @xmath11 geodesic ending on the @xmath11 7-brane at @xmath40 .
for @xmath83 the @xmath55 geodesic in region iv becomes a @xmath84 geodesic in region ii , which in turn intersects the lower half of @xmath72 .
these states are therefore represented by string webs .
as the d3-brane moves from one region to another one expects transitions between different string webs which represent the same bps state . in the particular case of the @xmath85 dyon
we expect a transition from a single string to a @xmath86 string web as the d3-brane crosses the @xmath81 contour , and a transition from a @xmath86 web to a @xmath87 web as it crosses the @xmath80 contour ( figure 5 ) . from the discussion in section 2
we see that this is precisely what happens . when the @xmath85 string crosses the @xmath25 7-brane at @xmath41 _ two _
@xmath25 strings are created , leading to the first transition . as the d3-brane moves towards the @xmath80 contour the junction point slides along @xmath72 , until it coincides with the @xmath11 7-brane at @xmath40 when the d3-brane is on the @xmath80 contour . at this point
the @xmath85 string crosses the @xmath11 7-brane , thereby creating _ four _ @xmath11 strings .
one of these annihilates with the original @xmath11 string which has reversed orientation , leaving the required _
three_. this is the second transition . = 2.4 in = 2.4 in = 2.4 in one can easily generalize this discussion for all the @xmath55 dyons with @xmath79 . in each case
the string description undergoes two transitions at @xmath88 , which are associated with crossing the two 7-branes .
this work is supported in part by the nsf under grant phy-92 - 18167 .
the functions @xmath47 and @xmath48 can be expressed in terms of hypergeometric functions as @xmath89 we choose the branch cuts ( dashed lines in the figures ) to run along the real axis from @xmath41 to @xmath43 for @xmath47 , and from @xmath40 to @xmath43 for @xmath48 .
the values of the functions at the branch points @xmath90 are given by a(1)=2/ , a(-1)=2i/ , a_d(1)=0 , a_d(-1 ) = -4i/ , [ values ] where the sign in @xmath91 depends on whether one is just above the cut , or just below it .
these points lie on the _ curve of marginal stability _ , defined by _
this curve is diffeomorphic to a circle centered at the origin .
denote the upper and lower halves by @xmath92 and @xmath93 , respectively .
it then follows from ( [ values ] ) that constant phase curves ( cpc ) are defined by ( f(z ) ) = .
if @xmath96 is in the above ranges the curve intersects @xmath72 and ends on the cut along @xmath45 $ ] .
otherwise it ends on the cut along @xmath44 $ ] without intersecting @xmath72 . when we consider the smooth continuation of @xmath33 using @xmath97 and @xmath98 , the cpc s continue across the cuts .
in terms of @xmath99 , a cpc of the function @xmath100 connects smoothly at the cut to a cpc of a function @xmath101 , where @xmath102 are related to @xmath3 by the appropriate monodromy .
the shapes of the cpc s can be qualitatively deduced from the asymptotics of @xmath103 and @xmath104 , @xmath105 \ ; . \label{zinf}\end{aligned}\ ] ] for @xmath106 @xmath33 is dominated by @xmath104 , whose phase at infinity varies over the range @xmath107 $ ] .
results of a quantitative analysis using mathematica are shown in figure 6 . | the string theory description of bps states in d - brane world - volume field theories may undergo transitions from open strings to string webs , as well as between different string webs , as one moves in the field theory moduli space .
these transitions are driven by the string creation phenomenon .
we demonstrate such transitions in the d3-brane realization of @xmath0 @xmath1 super - yang - mills theory . by -1.5 cm by -1.5
cm = 16.5 cm = 20.0 cm # 10= -.025em0 - 0 .05em0 - 0 -.025em.0433em0 h *s * *a * hep - th/9806011 + hutp-98/a027 0.5 cm 1.0 cm oren bergman and ansar fayyazuddin 0.5 cm _
lyman laboratory of physics + harvard university + cambridge , ma 02138 _ 1.5 cm june 1998 1.5 cm 1.5 cm pacs codes : 11.25.-w , 11.15.-q , 11.30.pb |
the vision of molecular electronics has stimulated great interest in both the experimental @xcite and the theoretical @xcite understanding of electronic transport through single molecule devices .
the coupling of electronic degrees of freedom to few , well - defined molecular vibrations , a property that discriminates transport through single molecules from that through other nanostructures such as quantum dots , sets the stage for observing novel quantum - transport phenomena .
@xcite one of the simplest and most intensively studied models of such a device is a molecule with a single electronic level and a one - dimensional potential energy surface for the nuclear displacements .
this model already gives rise to a large variety of interesting transport phenomena such as vibrational sidebands , @xcite electron shuttles , @xcite franck - condon blockade , @xcite avalanche - like transport , @xcite pair tunneling , @xcite and dynamical symmetry breaking .
@xcite generalizations of this model to two or more electronic levels have been suggested in order to account for non - degenerate but competing molecular states @xcite and for degenerate electronic states on a molecular dimer .
@xcite extensions of the simple molecular model to higher - dimensional potential energy surfaces may seem to be straightforward without yielding qualitatively new physics .
this is , however , _ not _ the case for molecules with symmetry - induced degeneracies of both electronic and vibrational states . for such molecules , the jahn - teller effect@xcite leads to nontrivial vibrational dynamics , originating in the conical intersections of adiabatic potentials and their associated geometric ( berry ) phases .
we expect the transport properties to be particularly sensitive to this intriguing dynamics when the tunneling of electrons between molecule and leads switches the molecule between a symmetric and a jahn - teller distorted charge state . in this paper , we investigate the consequences of such a berry phase for transport through single jahn - teller molecules .
we focus on the @xmath0 jahn - teller effect , for it is the simplest example with a conical intersection of the potential energy surfaces .
this type of jahn - teller effect occurs for a large variety of triangular ( x@xmath1 ) , tetrahedral ( ml@xmath2a transition metal ion m surrounded by a tetrahedron of ligands l ) , and octahedral molecules or complexes ml@xmath3 .
we show in section [ ssec : selectionrule ] that the berry phase induces a nontrivial selection rule for transitions between vibronic excitations of different molecular charge states .
this selection rule , in combination with generic properties of the energy spectrum of the jahn - teller molecule , leads to the formation of trapping states in the sequential tunneling limit ( section [ sec : theory ] ) .
the consequences are a current - blockade outside the coulomb blockade window , asymmetry of the current - voltage characteristics , and strong negative differential conductance ( section [ sec : rateequations ] ) .
the appearance of trapping states also strengthens the role of higher order processes , which introduce a second time scale that is much larger than the scale on which sequential tunneling takes place .
the interplay of these two scales renders the details of the tunneling dynamics irrelevant and has a non - negligible effect on the line shape of the differential conductance peaks .
we conclude the paper with a discussion of some experimental issues ( section [ sec : perturbations ] ) and an outlook to interesting problems related to electronic transport through jahn - teller molecules ( section [ sec : conclusion ] ) . jahn - teller distorted ( charged ) molecule ( left ) and the undistorted ( neutral ) molecule ( right ) as a function of @xmath4 . * bottom : * displacement patterns of the degenerate @xmath5 and @xmath6 vibrations for octahedral ml@xmath3 molecules .
[ fig : mexican_hat ] ] jahn - teller distorted ( charged ) molecule ( left ) and the undistorted ( neutral ) molecule ( right ) as a function of @xmath4 . *
bottom : * displacement patterns of the degenerate @xmath5 and @xmath6 vibrations for octahedral ml@xmath3 molecules .
[ fig : mexican_hat ] ] + jahn - teller distorted ( charged ) molecule ( left ) and the undistorted ( neutral ) molecule ( right ) as a function of @xmath4 . *
bottom : * displacement patterns of the degenerate @xmath5 and @xmath6 vibrations for octahedral ml@xmath3 molecules .
[ fig : mexican_hat ] ] jahn - teller distorted ( charged ) molecule ( left ) and the undistorted ( neutral ) molecule ( right ) as a function of @xmath4 . *
bottom : * displacement patterns of the degenerate @xmath5 and @xmath6 vibrations for octahedral ml@xmath3 molecules .
[ fig : mexican_hat ] ]
the @xmath0 jahn - teller effect generally occurs in molecules where two degenerate electronic states are coupled to two degenerate vibrational modes @xmath5 and @xmath6 . in the case of an octahedral ml@xmath3 molecule ,
the relevant electronic states are of @xmath7- and @xmath8-type , and the degenerate vibrations correspond to the symmetric oscillations of the ligands as they are pictured in fig [ fig : mexican_hat ] . for definiteness , we assume that without an excess charge on the molecule the degenerate electronic @xmath9 states are empty . with this choice ,
the molecular hamiltonian is @xcite @xmath10 we use a pseudospin notation with pauli matrices @xmath11 for the electronic states @xmath12 , @xmath13 .
the dimensionless strength of the linear coupling of electronic and vibrational degrees of freedom is @xmath14 , and @xmath15 is the frequency of the vibrational mode with effective mass @xmath16 and oscillator length @xmath17 .
the dimensionless coordinates @xmath18 and momenta @xmath19 of the normal modes @xmath5 and @xmath20 are measured in units of @xmath21 . the gate voltage @xmath22 tunes the energy of the degenerate electronic levels .
@xmath23 means that for vanishing bias , the molecular levels are degenerate with the fermi levels of the electrodes .
electronic occupations different from @xmath24 ( neutral molecule ) and @xmath25 ( charged molecule ) are effectively projected out by assuming a large charging energy @xmath26 .
the potential surfaces of both charge states are @xmath27 symmetric in the @xmath28 plane and thus have a conserved angular momentum quantum number
. we shall refer to this angular momentum as a _ pseudo _ angular momentum to make clear that it does _ not _ describe rotations in real space but rather a continuous change of the shape of the molecule s distortion .
a vibronic state of the neutral molecule is labeled by its _ integer _ pseudo angular momentum @xmath29 and radial excitation @xmath30 .
as it is well - known from the occupation number representation of the two - dimensional harmonic oscillator , its energy depends on the total number of vibrational quanta no matter in which direction of the oscillator they are excited .
the energy manifold @xmath31 is @xmath32-fold degenerate . in the pseudo angular momentum representation ,
the degeneracy is reflected in the relation @xmath33 , compare fig .
[ fig : spectra_a ] ( a ) . in the charged configuration , on the contrary , with an explicit coupling of vibrations to the pseudospin , it is the _ total _ pseudo angular momentum @xmath34 that is conserved . that the conserved pseudo angular momentum must be a _
half - integer _ quantity can also be seen by inspection of the adiabatic potentials , which are computed by diagonalizing the potential energy of @xmath35 for fixed nuclear displacements @xmath36 .
the two sheets of the @xmath25 adiabatic potential have , in contrast to the single sheet of the neutral state , a conical intersection at the origin . along any loop enclosing the conical intersection ,
the electronic pseudospin of the adiabatic eigenstates rotates by @xmath37 .
@xcite the adiabatic eigenstates consequently acquire a berry phase of @xmath38 along such closed loops , which must be offset by the vibrational wavefunction in order to make the full , electronic plus vibrational , wavefunction single valued .
this implies _
anti_periodic boundary conditions for the vibrational wavefunction and thus _ half - integer _ pseudo angular momentum @xmath39 . the spectrum of the charged configuration is plotted in fig . [ fig : spectra_a ] ( b ) for an intermediate value of the electron - phonon coupling , @xmath40
in contrast to the regularity of the oscillator spectrum , the jahn - teller effect lifts the large degeneracy ; each state is only doubly degenerate , @xmath41 .
the jahn - teller effect also lowers the ground state energies of each pseudo angular momentum column with respect to the harmonic oscillator energies , which is one of the keys to the current - blockade mechanism discussed in this paper .
.[fig : spectra_a ] ] the molecular system is coupled to two non - interacting fermi - gas electrodes labeled by @xmath42 .
we define electron annihilation operators @xmath43 for a particle in lead @xmath44 with wavevector @xmath45 and spin @xmath46 , and @xmath47 for a particle on the molecule .
the tunneling between the molecule and the electrodes is described by a conventional tunnel hamiltonian , @xmath48 jahn - teller distortions usually occur on a sub - ngstrm scale,@xcite which is small compared with the scale on which electronic wavefunctions fall off .
for this reason , we can neglect the dependence of @xmath49 on the vibrational coordinates .
the full hamiltonian of the single molecule device is @xmath50 where @xmath51 is the hamiltonian of the electrodes .
the states of the molecule will be labeled @xmath52 , @xmath53 is the charge on the molecule , @xmath39 is the pseudo angular momentum ( @xmath29 in the neutral state ) , and @xmath30 enumerates the radial excitations .
our discussion is restricted to the weak tunneling case , where the tunneling induced broadening of the molecular states @xmath54 is small compared with the temperature @xmath55 , and the predominant transport process is sequential tunneling of electrons .
we will only consider a regime of intermediate electron phonon coupling , where the system is not in the franck - condon blockade.@xcite the franck - condon blockade effect would not alter the validity of our conclusions but merely complicate the discussion unnecessarily .
any computation of transport properties via an expansion of the system s reduced density matrix in powers of the tunnel hamiltonian requires the computation of the matrix elements of @xmath56 .
an expansion of the jahn - teller states @xmath57 in terms of harmonic oscillator eigenstates @xmath58 times the two - dimensional electronic manifold @xmath59 shows that the electron - phonon interaction due to the @xmath0 jahn - teller effect only couples the states @xmath60 with the states @xmath61 of the non - interacting system.@xcite the associated half - integer pseudo angular momentum number of the jahn - teller state is @xmath62 .
the eigenstates of the jahn - teller active charge state thus have the general form @xmath63\ ] ] for suitable expansion coefficients @xmath64 and @xmath65 . the matrix elements
@xmath66 will therefore vanish unless the pseudo angular momenta fulfil @xmath67 this condition causes many zeroes in the tunnel matrix and has a great influence on the transport properties of the device as will be shown in the following sections .
the main result of our paper is a current - blockade effect _
outside _ the coulomb blockade diamonds . in certain areas of the @xmath68 plane , the system exhibits `` trapping states '' that , if populated , lock an excess charge on the molecule and , due to the large charging energy , obstruct any stationary current to flow through the device .
these states are vibrational eigenstates of the jahn - teller active molecule with pseudo angular momentum @xmath69 .
the effect becomes more pronounced for larger pseudo angular momenta . for certain voltage parameters that lie well outside the coulomb - blockaded region ,
an electron tunneling from the source electrode can excite the molecule into such states , but it can not tunnel out towards the drain electrode . the trapping effect
itself comes about , because these states become isolated in the tunneling dynamics : either a transition to a neutral target state is forbidden by the selection rule or it is forbidden by conservation of energy .
the implications of this effect for the differential conductance are the formation of structures like those in fig . [
fig : results ] : a suppression of the stationary current outside the coulomb blockaded region .
the suppression , however , is not complete ; a small stationary current is admitted due to the presence of higher order processes such as co - tunneling , relaxation to thermal equilibrium , or mixing of states by external perturbations , which effectively break the strict selection rule , allow to release the system to a vibrational state that can then be left by tunneling .
since these processes are of higher order in @xmath49 , they define a second time scale @xmath70 .
this time scale , which describes the time that the system will effectively spend in a trapping state , is much larger than the time scale of sequential tunneling , given by @xmath56 , @xmath71 , where @xmath72 is the density of states in the electrodes . and source - drain voltage @xmath73 for @xmath74 , @xmath75 . ( a ) @xmath76 , ( b ) @xmath77
white correspond to positive , black to negative values .
neutral gray is zero differential conductance.[fig : results ] ] the technical and formal discussion of the trapping effect is subtle , because due to the high level of degeneracy , one has to employ full master equation dynamics instead of intuitive rate equations for sequential tunneling . the qualitative picture
can nonetheless be grasped by rate equations in the pseudo angular momentum basis .
we confine ourselves to the sequential tunneling regime , which means that transitions will only take place between different charge states .
we use an expansion of the reduced density matrix @xmath78 in the tunneling hamiltonian,@xcite employ both the markov and the born approximation , and eventually arrive at a liouville - von neumann equation , @xmath79 + \mathcal{l } \rho$ ] . the operator @xmath80 has the well - known lindblad form for dissipative linear operators and describes the tunneling in second order perturbation theory.@xcite since we are only interested in the stationary transport properties , we can safely neglect those off - diagonal matrix elements of @xmath78 that belong to states of different energies and thus effectively produce a rate equation dynamics between different energy manifolds .
the stationary von neuman equation reduces with this approximation to @xmath81 .
but since the system under consideration has degenerate states , the equations can , at this point , _ not _ be further simplified to rate equations for the states @xmath57 and @xmath82 .
we do , however , know that the density matrix is a self - adjoint operator and hence admits an eigenbasis in the degenerate subspaces for _ each _ set of voltage parameters @xmath83 . in this basis
, the von neumann equation assumes the form of a rate equation , where the rate for a transition between two eigenstates @xmath84 and @xmath85 is given by a golden - rule expression , @xmath86 which is of order @xmath87 , multiplied by a fermi function that accounts for the availability of an electron or a hole in the leads .
if we want to compute the rates between different energy manifolds in the tunneling problem , we shall consider the , unknown , diagonal basis of the degenerate subspaces of @xmath78 at fixed voltages .
let @xmath88 be one of these states in the neutral configuration and @xmath89 one in the charged configuration .
the franck - condon factor is @xmath90 the selection rule , which has been shown to be strict in the pseudo angular momentum basis , has thus implications on the transition matrix elements whatever the eigenbasis of @xmath78 is .
if the angular momentum of a charged state @xmath84 is too large compared with those present in a neutral state @xmath85 , a transition may not be allowed .
we will show in the next section that such forbidden transitions between energy manifolds are the cause of the current - blockade that we observe analytically and numerically . considering a sufficiently large bias and a gate voltage that places the ground state of the charged molecule slightly above the fermi level of the drain electrode , the fermi functions for relevant excitations are @xmath91 , and we only need to discuss the rates for an electron to leave the molecule towards the drain electrode . in this configuration ,
the system is not in the coulomb blockaded region and ought to be conducting .
this is , however , _ not _ the case due to the selection rule eq .
( [ eq : selection_rule ] ) and the spectral structure of the molecule .
consider fig .
[ fig : spectra_b ] , in which the vibronic spectra of both charge states have been overlaid and shifted by @xmath92 and @xmath93 along the energy axis to account for the electronic energies of the charged and the neutral state respectively .
conservation of energy then implies that transitions are energetically possible if the charged target state lies below the neutral initial state in fig .
[ fig : spectra_b ] ( a ) and vice versa in fig . [
fig : spectra_b ] ( b ) , see the figure caption for more details .
take for example the state @xmath94 , the charged state with pseudo angular momentum @xmath95 and radial ground state .
once the system occupies this state , it will stay there forever ; the electron can not tunnel to the drain electrode , see figs .
[ fig : spectra_b ] and [ fig : spectra_d ] .
conservation of energy requires that any target state has to have less energy than @xmath84 , so only @xmath96 and suitable @xmath97 are possible .
the selection rule eq .
( [ eq : selection_rule ] ) only allows the dynamics to connect neighboring columns of fig .
[ fig : spectra_b ] .
but then all transition rates are zero .
the only states that have non - vanishing franck - condon factors belong to energy manifolds that due to the specific choice of voltages are out of reach energetically .
the state @xmath84 is called absorbing or `` trapping state '' , that is to say there is no escape rate but yet a finite rate for populating the state .
the latter of the two properties of @xmath84 is easily verified with fig . [
fig : spectra_b ] ( a ) , which shows the shift of the two spectra for tunneling processes from the source electrode . in this configuration
there are enough states from which @xmath84 can be populated , for instance the @xmath98 manifold .
the energy threshold above which a trapping state can be reached from the vibronic ground - state of the neutral system marks the onset of the current - blockade in the @xmath99 diagram .
it can thus cause negative differential conductance peaks , like those in fig .
[ fig : results ] , if the system were conducting below the threshold .
positive differential conductance peaks , on the contrary , can be observed for the energy threshold above which the trapping state acquires a finite escape rate and is no longer trapping .
the current - blockade effect , as it has been discussed in the previous paragraph also produces a significant asymmetry of the @xmath99 diagram with respect to @xmath22 . in fig .
[ fig : spectra_c ] , we plot the energies of the charged and uncharged states for voltage parameters that place the molecule closely below the fermi level of the source electrode , that is for the same @xmath22 as before but with opposite sign . in both processes , tunneling onto and off the molecule , there is no state of the charged system that is isolated due to the selection rule or the spectral distortion .
the property of the jahn - teller effect to _ lower _ the energy of the distorted states is the actual reason for this asymmetry . if our jahn - teller molecule were such that it was jahn - teller active _ without _ an excess charge and symmetric _ with _ an additional electron , the situation would be inverted , current - blockade for positive , conducting at negative gate voltage . ) ( diamonds ) and the neutral molecule ( circles ) .
radial excitations are grouped for each pseudo angular momentum .
panel ( a ) shows the situation for tunneling at the source electrode , panel ( b ) the situation at the drain electrode .
the chosen parameters are @xmath100 , @xmath101 .
consider panel ( a ) .
an electron can only tunnel from the source electrode and induce the vibronic transition @xmath102 if the difference in the electronic energy , @xmath103 , is larger than the difference in the vibronic energy , @xmath104 .
more precisely , the fermi function @xmath105 has to be positive . by shifting the neutral spectrum by @xmath106 and the charged spectrum by @xmath92
, this relation translates into the intuitive rule that in the diagram the target state of a tunnel process has to lie below the initial state .
energetically allowed transitions are therefore directed downwards in the figure .
a horizontal transition corresponds to the borderline case of an electron sitting exactly at the fermi level of the electrode . that conservation of energy has the form of an inequality is due to the filled continuum of states in the fermi - gas electrode .
panel ( b ) shows the corresponding energetic situation for sequential tunneling processes at the drain electrode .
the presence of trapping states can be seen clearly .
[ fig : spectra_b ] ] . in this case
the jahn - teller molecule is placed slightly below the chemical potential of the source electrode .
the chosen parameters are @xmath100 , @xmath107 .
contrary to the situation shown in fig .
[ fig : spectra_b ] , there are no trapping states.[fig : spectra_c ] ] , @xmath101 .
the trapping state @xmath108 is marked in red .
the three nearest energy manifolds of the uncharged system are shaded in gray .
the dashed arrows show transitions that are forbidden , because the target manifold s energy is too large .
the dotted arrow shows a transition that is forbidden due to the selection rule.[fig : spectra_d ] ]
in the previous paragraphs , we have argued that due to the interplay of the spectral distortion and the selection rule , there are two time scales present in the problem .
one describes the tunnel dynamics and is of order @xmath87 .
the other is due to higher - order processes and is an estimate of how long the system spends in the trapping states .
it is thus evident that the precise implementation of the internal dynamics is irrelevant as long as the key ingredients to the separation of time scales , namely selection rule and spectral distortion , are included . in this paragraph
, we present such a numerical implementation that accounts for the selection rule by employing rate equations in the pseudo angular momentum basis . by doing so
, we neglect all coherences between _ any _ two states in the density matrix , and therefore also those between degenerate states .
this will , as we have pointed out before , _ not _ affect the appearance and structure of the current - blockade , because the effect is independent of the specific form of the eigenbasis of the stationary density matrix .
the only error that is introduced by the restriction to rate equations is a quantitative one .
the rate equations , on the contrary , allow for a very intuitive and generic treatment of transport problems for systems with spectral distortions and selection rules .
the occupation probability @xmath109 of the molecular eigenstates @xmath110 ( @xmath111 denotes the charge of the molecule , and @xmath112 is a multiindex for the internal degrees of freedom .
@xmath54 is the broadening of the molecular levels due to the coupling to the leads . )
is governed by the rate equation@xcite
@xmath113 + \gamma_{\rm rel}.\ ] ] the tunnel rates @xmath114 are a product of the golden - rule expressions eq .
( [ eq : goldenrule ] ) and the fermi functions of the leads .
@xcite the last term , @xmath115 $ ] , describes vibrational relaxation towards the thermal equilibrium distribution @xmath116 by higher order processes not included in the hamiltonian . by doing so
, we effectively implement the aforementioned processes , which release the system from the trapping states .
our results , however , are independent of the details of this implementation .
the maximal tunneling rate , @xmath117 , defines the proper time scale on which we measure all times . the stationary solution of eq .
( [ eq : rate ] ) permits us to compute the stationary current through , say , the left junction , @xmath118,\ ] ] as a function of the source - drain voltage @xmath73 and the gate voltage @xmath22 .
the results for the differential conductance @xmath119 shown in fig .
[ fig : results ] for ( a ) fast and ( b ) slow vibrational relaxation differ strongly from those obtained for single - mode models .
@xcite when the molecular orbitals are close to the fermi energy of the drain electrode ( negative @xmath22 in fig .
[ fig : results ] ) , we find , in accordance with the discussion in section [ sec : theory ] , an extended region of current suppression outside the coulomb - blockaded region with large negative differential conductance ( ndc ) .
no such suppression occurs when the molecular orbitals are close to the chemical potential of the source electrode ( positive @xmath22 in fig .
[ fig : results ] ) , thus rendering @xmath99 asymmetric in the gate voltage .
the positions of the @xmath99-lines that mark the area of the suppressed current in the diagrams are defined by those values of @xmath83 for which the trapping state is connected to an energy manifold that participates in the transport process . if an increase of the bias voltage generates a finite escape rate for population from the trapping state , that state is no longer absorbing , and a stationary current can flow , hence a line of _ positive _ differential conductance is observed . in the opposite case , that is trapping behaviour
, a line of _ negative _ differential conductance appears as soon as the principal current carrying states most often it is sufficient to consider the vibronic ground state of the system have tunnel rates into the trapping state .
because trapping states are excited vibronic states and intermediate excited states are necessary to access these , there is a certain energy threshold below which a finite current may flow but the trapping state is still out of reach
. this threshold will thus be visible as a negative differential conductance peaks in the @xmath99 diagram .
since the trapping states with larger pseudo angular momentum have larger energy thresholds before they can be populated in steady state , the current - blockade mechanism will yield a step - like structure of negative differential conductance peaks and plateaus of suppressed current .
these are clearly visible in our numerical results in fig .
[ fig : results ] .
we further support the irrelevance of the details of the fast , order @xmath87 dynamics by reproducing the positions and line shapes of the peaks in the numerically obtained differential conductance , see fig .
[ fig : results ] , within a simple model .
this model only retains the most basic features of the dynamics , namely the presence of trapping states and the two relevant time scales .
consider a two - level system with a trapping state @xmath84 and a state @xmath120 from which the electron can tunnel to the drain electrode , say the absolute ground state of the charged molecule .
the current is given by the inverse of the time @xmath121 the electron spends on the molecule . in our model
, this time is the sum of the trapping time @xmath70 and the time the electron needs to leave the second , non - trapping state .
the rate to leave this latter state is the product of the bare tunneling rate @xmath122 and the probability @xmath123 to find an empty state of appropriate energy in the drain electrode , where @xmath124 is this electrode s fermi function .
the typical time for tunneling from @xmath120 out to the leads is thus @xmath125 and the stationary current can be estimated as @xmath126 this yields the relation @xmath127 .
its consequences for the line - shape are twofold . as a purely quantitative effect
, the height of the peak in @xmath99 decreases with increasing @xmath70 . as a qualitative effect , however , the position of the peak is shifted . in a model with just a single time scale , be it a quantum dot or a single - mode molecule , the stationary current through the drain electrode is _ only _ determined by the fermi function @xmath128 , that is the probability to find an empty state in the leads with just the energy necessary for the vibrational transition @xmath129 .
the maximum of the @xmath99-peak is thus located at the bias voltage for which @xmath130 . within our two - state model , this behavior
is recovered for @xmath131 which corresponds to the absence of any trapping effect due to fast relaxation . when trapping is included , and hence @xmath70 is finite , the maximum of the differential conductance peak is shifted .
this simple model compares remarkably well to the numerical rate - equation results of the previous section .
we focus on a single peak in the @xmath99 diagram and study its line shape for different values of the phenomenological relaxation rate @xmath70 of eq .
( [ eq : rate ] ) .
the effect of increasing @xmath70red to blue curves in fig .
[ fig : relaxation ] ( a)is a reduction of the peak s height and a shift of its position .
we can fit the results of our simple model , eq .
( [ eq : i_model ] ) , to the data in fig .
[ fig : relaxation ] ( a ) , obtaining fig .
[ fig : relaxation ] ( b ) . comparing the fit values for @xmath70 with the parameters of the numerical simulation , see fig . [
fig : relaxation_fit ] , we can see how well this simplified model describes the interplay of the two time scales .
the only deviations occur when both time scales are of comparable magnitude , @xmath132 .
in contrast , for a sufficiently large separation of time scales , the agreement is nearly perfect .
this backs our general considerations of section [ sec : theory ] , where we have argued that the observed phenomena are of a generic nature . , for different values of the relaxation time @xmath70 ( a ) and fit with the simplified model , dots in ( b ) .
the read curves , that is the large , centered peaks have small , the blue ones , damped and shifted to lower @xmath73 very large @xmath70.[fig : relaxation ] ] -values of the simplified model ( dots ) and its application to fig .
[ fig : relaxation ] vs. the numerical values used in the actual simulation .
the curve is a linear fit to these values , @xmath133 , showing the validity of the simplified approach in the regime of strong separation of the time scales.[fig : relaxation_fit ] ] the predominant features of electronic transport through a single jahn - teller molecule are determined by three properties of the system : the spectral distortion and the selection rule fix the position of the peaks in the differential conductivity , their heights are determined by the franck - condon matrix elements .
finally , higher order processes introduce a second , larger time scale , which eventually determines the exact line shape .
we finally discuss several experimental issues .
placing a jahn - teller molecule between two biased electrodes exposes it to a potentially large electric field . for a semiconductor quantum dot ,
weak coupling and good screening implies that the potential drop is concentrated in the contact regions between the electrodes and the dot.@xcite for a single molecule , however , this is less obvious .
thus , the electric field may lift the electronic degeneracy by the stark effect and mix the vibronic levels by breaking the rotational symmetry of the adiabatic potential . to estimate the mixing of vibronic levels
, we note that the relevant vibrational modes @xmath5 and @xmath6 do not possess a dipole moment and are therefore unaffected by the electric field unless anharmonic mode mixing becomes relevant when other , `` dipole - active '' modes are strongly displaced .
one readily estimates , however , that for relevant electric fields ( corresponding to voltage drops of the order of @xmath134 ) the typical displacements of dipole - active vibrational modes are small compared with their oscillator length @xmath135 , so that no anharmonic effects are expected . assuming that the dipole moment is of order @xmath136 , where @xmath137 is the displacement of the dipole - active mode
, we obtain a displacement @xmath138 .
the length @xmath139 is of the order of the size of the molecular junction .
we also note that the electric field does not directly couple the two electronic levels as both are @xmath140 orbitals .
the stark shift is therefore quadratic in the electric field , and we estimate that it is several orders of magnitude smaller than the vibrational frequency .
the coupling of the jahn - teller molecule to fermionic reservoirs , the electronic leads , may also introduce external perturbations that are able to break the symmetries of the system on which our approach relies so heavily . if the perturbation is too strong , for example , the degeneracy of the vibrational modes @xmath5 and @xmath6 is lifted and the energies are shifted by a large amount . the jahn - teller effect is then no longer of @xmath0- but rather of @xmath141-type . a selection rule like
the one used in this paper is not available in such a model , and the current - blockade would not be visible . if , however , the perturbation is only small compared with the tunnel coupling , the mixing of the states induced by the perturbation is able to introduce a second time scale @xmath142 into the problem . as we have discussed in the previous paragraphs , the presence of two widely separated time scales only shifts the positions and alters the shape of the differential conductance peaks , no matter what the detailed physics behind the phenomenology of the conduction peaks are .
we have shown that symmetry - induced degeneracies of electronic and vibrational modes can lead to new transport phenomena in single - molecule junctions . when the molecular junction consists of , say , an octahedral complex with an @xmath0 jahn - teller effect , a combination of berry - phase effects and a spectral trapping mechanism results in a non - trivial current - blockade , even for parameters where the coulomb blockade is lifted .
it is evident from the physics of this blockade mechanism developed in this paper that it is associated with strongly enhanced shot noise of the current .
@xcite we have also shown that the mechanism is quite generic and that the introduction of a second time scale to the problem , for example by higher order relaxation processes , co - tunneling , or external perturbations , has only quantitative effects on our results .
the good agreement of the numerical results of our effective rate model with the general consideration of section [ sec : theory ] shows that the current - blockade that has been found in this context is a generic effect of systems with selection rules and spectral distortions .
et al_.@xcite find a related current - blockade effect in systems where the strong spectral distortion alone damps certain franck - condon matrix elements .
this is , however , no selection rule that derives from symmetries like in this paper ; it rather stems from the almost vanishing overlap of displaced oscillator wave functions .
our discussion of jahn - teller molecules raises several interesting questions .
the most important question concerns a complete understanding of the role of coherences between degenerate states and their implications for the transport properties of the system .
what are the experimentally measurable quantities that are influenced by non - diagonal elements of the density matrix ? and
when is a reduction to rate equations allowed ?
this work was supported in part by spp 1243 of the deutsche forschungsgemeinschaft as well as sfb 658 . | the vibrational modes of jahn - teller molecules are affected by a berry phase that is associated with a conical intersection of the adiabatic potentials .
we investigate theoretically the effect of this berry phase on the electronic transport properties of a single @xmath0 jahn - teller molecule when the tunneling electrons continually switch the molecule between a symmetric and a jahn - teller distorted charge state .
we find that the berry phase , in conjunction with a spectral trapping mechanism , leads to a current - blockade even in regions outside the coulomb blockade .
the blockade is strongly asymmetric in the gate voltage and induces pronounced negative differential conductance . |
automata theory is particularly fruitful in terms of equivalence theorems : regular expressions , deterministic and nondeterministic automata , the myhill - nerode theorem , the regular word logic and the weak second order theory of one successor all are equally expressive in the languages they describe . in this paper ,
i concern myself not with regular predicates ( predicates which hold only for the words in a regular language ) but with regular functions , functions whose behavior can be recognized by an automaton .
this allows a translation of the krohn - rhodes theorem into yet another equivalent .
the krohn - rhodes theorem concerns itself with _
finite state transducers _ , an abstraction of systems that : * accept inputs from a discrete set at discrete times , * retain some memory about previous inputs , which updates whenever an input is read , * for each input read , produce some output from a discrete set based on the input and memory .
these are an abstraction of synchronous ( as opposed to those that update continuously ) , digital ( as opposed to those that deal with analog values ) systems .
originally proved in @xcite , the krohn - rhodes theorem itself gives a decomposition of arbitrary finite state transducers into a cascade of transducers from a small generating set .
computational implementations of this decomposition are available @xcite .
the krohn - rhodes theorem can be used to analyze the rough behavior of automata , providing applications to artificial intelligence @xcite .
finite state transducers are formalized as follows : a _
moore machine _ is a tuple : @xmath0 .
* @xmath1 is a finite set of input characters ( alphabet ) .
* @xmath2 is a finite set of states .
* @xmath3 is the initial state .
* @xmath4 is a finite set of output characters .
* @xmath5 is the transition function .
* @xmath6 is the output function . for convenience
, i will sometimes denote the map @xmath7 by @xmath8 .
given an input word @xmath9 , construct a run @xmath10 and output @xmath11 such that : * @xmath12 = q_0 $ ] .
* @xmath13 = \delta(r[i],w[i])$ ] for @xmath14 . * @xmath15 = \epsilon(r[i])$ ] , for @xmath16 .
where the indexing notation is such that : @xmath17 , \ldots , w[|w|-1].\ ] ] the correct way of thinking about moore machines is that each input acts as a transition between one state and the next , or that each state is the state between inputs .
a proper representation would have input characters half a step offset from states .
outputs are simply a product of the state the automaton is in and so should be in step with the states .
however , representing the sequences of input characters and states as words requires a choice of direction to shift half a step .
a considerable effort has been made to pick the option to result in the cleanest presentation . in this paper ,
the input character @xmath18 $ ] tells the device how to transition from state @xmath19 $ ] to state @xmath13$].$ ] telling the device how to transition from state @xmath20 $ ] to state @xmath19 $ ] .
this alternative is not uncommon . ]
this example below shows how inputs , states , and outputs , respectively , line up according to my notation .
i will typically consider moore machines that simply output their states : a moore machine is said to be _ transparent _ if @xmath21 and @xmath22 is the identity . in this case ,
moore machines are presented as a tuple : @xmath23 .
moore machines can be interpreted as functions from input words to output words : given a moore machine @xmath24 , and word @xmath9 , i denote by @xmath25 the output of @xmath26 on input @xmath27 , with @xmath28 .
moore machines , however , represent only a small subset of those functions on words which can be reasoned about using finite automata . given two alphabets , @xmath29 , a function on words @xmath30 is said to be : * _ length - preserving _ if for any word @xmath27 , @xmath31 .
* _ causal _ if @xmath32 $ ] depends only on @xmath33 , \ldots , w[i]$ ] . * _ strictly causal _ if @xmath32 $ ] depends only on @xmath33 , \ldots , w[i-1]$ ] . * _ character - wise _ if @xmath32 $ ] depends only on @xmath18 $ ] .
define the following useful functions for dealing with words : * let @xmath34 denote the _
_ successor-__@xmath35 function which appends the character @xmath35 to the end of a word .
* let @xmath36 denote the function which removes the last character of a word .
* let @xmath37 denote the function which removes the first character of a word . given a moore machine @xmath26 ,
the function @xmath26 computes is strictly causal , and increases length by 1 .
it is often more convenient to deal with length preserving versions of this function : given a moore machine @xmath24 , and word @xmath9 , denote by : * @xmath38 the output of @xmath26 on input @xmath27 with the last state removed . * @xmath39 the output of @xmath26 on input @xmath27 with the first state ( the start state ) removed .
given a moore machine @xmath26 , the function @xmath40 is strictly causal and length preserving , and the function @xmath41 is causal and length preserving . the notion of finite automaton , or finite state recognizer , is more commonly studied than the finite state transducer . _ finite state automata _ are an abstraction of systems that : * accept inputs from a discrete set at discrete times , * retain some memory about previous inputs , which updates whenever an input is read , * having finished reading a sequence of inputs , either _ accepts _ or _
rejects_. just as finite state transducers can be interpreted as functions on words , finite state automata can be interpreted as predicates on words , returning a boolean value after having read in a word .
their predicative nature means that finite state automata are more convenient to use in applications to formal logic . on the other hand , real world systems are more often interested in transforming inputs , and
so are better represented by finite state transducers . finite automata yield a notion of a regular set or regular event , a collection of words or sequences of inputs which are exactly those which some finite automaton accepts
once i define what it means to represent a function of words with an automaton , this will yield a notion of regular function .
finite state recognizers are formalized as follows : a _ finite state automaton _ is a tuple : @xmath42 .
* @xmath1 is a finite set of input characters ( alphabet ) .
* @xmath2 is a finite set of states .
* @xmath43 is the set of initial states .
* @xmath5 is the transition relation .
* @xmath44 is the set of final states . given an input word @xmath9 , @xmath45 is a run on input @xmath27 if : * @xmath12 \in i$ ] . *
@xmath46,w[i],r[i+1])$ ] for @xmath14 .
@xmath27 is _ accepted _ if there is a run @xmath10 on input @xmath27 such that @xmath47 \in f$ ] .
a finite state automaton @xmath48 is _ deterministic _ if : * @xmath49 is a singleton .
* @xmath50 is a function from @xmath51 , that is , given a @xmath52 , and @xmath53 , there is a unique @xmath54 such that @xmath55 . by default
, @xmath56 is nondeterministic .
a well known theorem of finite automata is that : [ detlemma ] if @xmath57 is the set of accepted inputs of some automaton , then it is also the set of accepted inputs of some deterministic automaton . in either case
, @xmath57 is _
regular_. a common convention in logic is to identify a function @xmath58 with the relation @xmath59 which consists of all pairs of the form @xmath60 , or , for @xmath61-ary functions , @xmath62 for @xmath63 in the domain of @xmath58 . as such , a regular function can be defined as a relation which is regular and also a function .
the question now is how to input multiple words , especially multiple words of different lengths , to a finite automaton i introduce the @xmath64 map to merge words together in parallel so they can be read by an automaton . for words of different lengths ,
i add a dummy character @xmath65 .
define the map @xmath66 which satisfies : @xmath67_j = \begin{cases } w_j[i ] & \text { it exists } \\ \ # & \text { otherwise } \end{cases}\ ] ] and @xmath68 .
below is shown how @xmath64 combines words together into one word : @xmath69 now i can define the notion of regular relation and regular function : an @xmath61-ary relation @xmath70 is _ regular _ if there is a finite automaton @xmath56 with input alphabet @xmath71 such that : @xmath72 an @xmath61-ary function @xmath73 is _ regular _ if there is a finite automaton @xmath56 with input alphabet @xmath74 such that : @xmath75 regular relations and functions are a key part of the analysis of various automaton logics .
for instance : given a finite alphabet @xmath1 , let @xmath76 where : * @xmath77 is the prefix relation on words , * @xmath78 is the equal length relation on words , * @xmath34 is the _
successor-@xmath35 _ unary operation , which appends an @xmath35 onto the end of a word .
i call @xmath79 the _ regular word logic over @xmath1_. operation , and containment . ]
[ wordlogic ] if @xmath57 is a relation on @xmath80 , the following are equivalent : * @xmath57 is regular , * @xmath57 is given by a formula @xmath81 in the language of @xmath82 .
hence my interest in regular functions .
if @xmath83 is a formula in the language of @xmath79 such that : @xmath84 then , since lexicographic ordering @xmath85 is regular and well - founded , the following relation is regular @xmath86,\ ] ] and also a function . restrictions of this sort are called _
skolem functions_. it is also worth noting that a classification of the regular functions also yields a classification of the regular languages , since for any regular language @xmath57 , the characteristic function @xmath87 is regular .
the idea of achieving quantifier elimination on the regular word logic via function composition was inspired in part by a theorem in @xcite , which decomposes regular predicates in terms of shifting operations , character - wise operations , and a univeral regular predicate . finally , in this section , i connect moore machines and finite automata : given a moore machine @xmath88 , the functions @xmath89 and @xmath41 are regular .
construct a deterministic finite automaton that keeps track of two pieces of information : the state the moore machine is expected to be in at any particular point , and a boolean value to keep track of whether the proposed output has so far been correct .
additionally , in the case of @xmath26 , there will be a final character of the form @xmath90 and a small amount of information must be kept track of to handle this correctly .
for example , for @xmath40 , let : @xmath91 where : @xmath92 every strictly causal , length - preserving , regular function is given by @xmath40 for some moore machine @xmath26 .
every causal , length - preserving function is given by @xmath41 for some moore machine @xmath26 .
i prove this for a binary , strictly causal , length - preserving , regular function @xmath58 .
the proof is nearly identical in the general case .
let @xmath93 be given . by proposition [ wordlogic ] ,
there is a formula @xmath94 in the language @xmath95 such that @xmath96 since @xmath58 is a strictly causal function , the first @xmath97 characters of the input determine the @xmath61th character of the output . by simple tricks in @xmath95
, one can construct a formula @xmath98 for each @xmath99 such that @xmath98 is true of exactly those sequences of characters that produce an output of @xmath100 in the next place .
these @xmath98 describe a collection of regular sets which partition all of @xmath80 .
let @xmath101 be a finite automaton that recognizes the corresponding collection .
now to construct the moore machine @xmath26 .
it should run each of the @xmath101 in parallel to determine its state .
since the @xmath101 recognize disjoint collections , exactly one of the @xmath101 will be in an accept state at any time .
the @xmath22 function for the moore machine should take in the tuple of states for the @xmath101 and output the one which is in an accept state .
it suffices now to check that @xmath40 is identically @xmath58 .
the proof for @xmath41 is similar . of course , there are plenty of other regular functions . in this paper
i provide a small set of functions whose closure under multi - ary composition generates all of them . in the next section
, i will interpret the proof of proposition [ detlemma ] to reduce the problem to studying functions given by the actions of moore machines and reverse moore machines .
moore machines allow one to construct functions which transmit information only to the right ( towards the end of the inputs ) , but they need to be combined with a method to transmit information to the left .
consider , for instance , the @xmath37 function , which removes the first character of a string , shifting to the left .
alternately , a function which outputs @xmath102 or @xmath103 depending on whether there are an even or odd number of @xmath104s in the input .
these functions are regular , but not causal .
in this section , i define reverse moore machines , and provide a technique for decomposing a length - preserving regular function as a multivariable composition of a moore machine function , a reverse moore machine function , and character - wise maps to connect them .
this will require several stages .
first , i will briefly discuss some notation for discussing character - wise functions .
second , i will introduce the notion of a reverse moore machine similar to the moore machines introduced in section 1 .
third , i will present the decomposition .
this decomposition is based on the classical powerset determinization construction , viewed from a novel perspective . in the next section
, i will discuss handling general regular functions .
first , some notation for functions which operate character - wise : given two finite alphabets @xmath29 , and a function @xmath93 , call the function @xmath105 which applies @xmath58 to each character of the input , the _ character - wise @xmath58 map_. if @xmath58 is an @xmath61-ary function for @xmath106 , one can also make sense of @xmath107 .
let @xmath108 then one defines the partial function @xmath109 which takes in inputs of all the same lengths and produces an output of the same length , where : @xmath110 noting that @xmath64 does not produce characters with @xmath65 in them for equal - length inputs . of course , @xmath107 is causal and length preserving ( and reverse - causal , when i define the notion ) .
every moore machine function can be written as the action of the corresponding transparent moore machine composed with a bitwise application of its @xmath22 function .
i now define some notation for dealing with reverse moore machines , analogous to the notation established previously .
a _ reverse moore machine _ is a tuple : @xmath111 .
* @xmath1 is a finite set of input characters ( alphabet ) .
* @xmath2 is a finite set of states .
* @xmath112 is the final state .
* @xmath4 is a finite set of output characters .
* @xmath5 is the reverse transition function .
* @xmath6 is the output function .
given an input word @xmath9 , construct a run @xmath10 and output @xmath11 such that : * @xmath47 = q_f$ ] .
* @xmath19 = \delta(r[i+1],w[i])$ ] for @xmath14 .
* @xmath15 = \epsilon(r[i])$ ] , for @xmath16 . here
i show below how inputs , states , and outputs , respectively , line up according to this notation . to prevent type mismatches , i will denote reverse moore machines with letters @xmath113 as opposed to letters @xmath114 for moore machines .
a reverse moore machine is said to be _ transparent _ if @xmath21 and @xmath22 is the identity . in this case , the reverse moore machine is presented as a tuple : @xmath115 . as before , one can interpret reverse moore machines as functions from input words to output words : given a reverse moore machine @xmath116 , and word @xmath9 , denote by : * @xmath117 the output of @xmath57 on input @xmath27 . *
@xmath118 the output of @xmath57 on input @xmath27 with the last state removed . *
@xmath119 the output of @xmath57 on input @xmath27 with the first state ( the start state ) removed .
analogous to the notions of causal and strictly causal , one has notions of reverse causal and strictly reverse causal : given two alphabets , @xmath29 , a function on words @xmath30 is said to be : * _ reverse causal _ if @xmath32 $ ] depends only on @xmath18 , w[i+1 ] , \ldots$ ] . * _ strictly reverse causal _ if @xmath32 $ ] depends only on @xmath120 , w[i+2 ] , \ldots$ ] . of course , a function is character - wise iff it is causal and reverse causal .
one also has a notion of reverse deterministic automaton : a finite automaton @xmath121 is _ reverse deterministic _ if : * @xmath122 is a singleton . * for each @xmath123 , there is a unique @xmath124 such that @xmath125 .
a reverse deterministic automaton can also be viewed as a transparent reverse moore machine .
i now have the notation to state the main result : [ detharvest ] any length - preserving regular function can be written as a multivariable composition of the form : @xmath126 it s worth noting that @xmath127 has the same action here as @xmath64 .
i ve chosen to use @xmath127 here to indicate that i m not using the length - padding features of @xmath64 .
a few lemmas must be proved first : [ automatize ] given a length - preserving regular function @xmath128 , there is an automaton @xmath129 with state set @xmath2 and map @xmath130 such that @xmath131 is @xmath132 applied to any run of @xmath129 on input @xmath27 . given the length - preserving function @xmath30 , let : @xmath133 witness the regularity of @xmath58 . since @xmath134 automatically rejects any inputs with a @xmath65 in them , one can restrict it to : @xmath135 recognizing the same set of inputs .
construct an automaton : @xmath136 where : @xmath137 @xmath129 is nondeterministic .
i assert that runs of @xmath129 on input @xmath27 will necessarily have @xmath4 component @xmath138 , for some @xmath139 .
firstly , it should be clear that if @xmath10 is an accepting run of @xmath56 on input @xmath140 , then @xmath141 is an accepting run of @xmath129 on input @xmath27 for any @xmath139 . now , suppose @xmath142 is an accepting run of @xmath129 on input @xmath27 .
then it is easy to check that @xmath143 is an accepting run of @xmath56 on input @xmath144 .
since @xmath56 witnessed @xmath58 being a regular function , @xmath145 must be @xmath131 .
this completes the proof for @xmath22 taking the @xmath4 component of the states of @xmath129 .
given a nondeterministic automaton @xmath121 , define its _ determinization _ : @xmath146 where : @xmath147 typically throughout this paper i will be concerned with the determinization as a transparent moore machine , so the set of final states does nt matter much .
given a nondeterministic automaton @xmath148 , define its _ harvester _
: @xmath149 where : * the @xmath150 such that @xmath151 ( for @xmath54 ) is given by the least @xmath152 such that @xmath153 or @xmath154 if none exist . * the @xmath150 such that @xmath155 is given by finding the least @xmath156 such that there is a @xmath152 such that @xmath153 and then having @xmath150 be the least @xmath152 such that @xmath153 .
if no such @xmath156 exists , then @xmath150 is given by @xmath154 .
the powerset determinization of an automaton @xmath56 produces an automaton that keeps track of , at every position , the set of states of @xmath56 which are reachable through some sequence of transitions , having read the input up to that point
. however , not every one of these reachable states necessarily shows up in some accepting run : it may be that being in one state now means later on having to be in another state which it can not transition out of , or that being in a state now dooms the automaton to being in a reject state once it has finished reading the input . in order to use the determinization to find an accepting run of the original automaton , provided
there is an accepting run , one needs to start at the end and work backwards , all the while staying within states that one knows can be traced through a sequence of transitions back to a start state at the beginning . specifically ,
if the automaton is in a state @xmath150 which is reachable through some sequence of transitions after having read @xmath157 , then there must be at least one state @xmath124 which is reachable through some sequence of transitions after having read @xmath27 such that reading @xmath35 takes it from state @xmath124 to state @xmath150 .
it is necessary to introduce an additional dummy state @xmath154 to start out in to make sure the harvester automaton is reverse - deterministic .
although this application of the harvester automaton will see only pairs of the form @xmath158 for @xmath35 some symbol being read by the original automaton and @xmath159 the set of states reachable immediately prior to reading that specific @xmath35 , the automaton should be prepared to read in arbitrary input pairs . an invalid input will cause the reverse deterministic automaton to go into the dummy @xmath154 state . additionally , a cheap fix is necessary to account for the fact that one does nt know what state the run of the original automaton ends on and one needs a single `` final '' state for the reverse deterministic automaton to `` start out '' in .
the dummy @xmath154 state represents all final states which are reachable . truncating the run
will remove this dummy state .
[ decomp ] suppose one has a nondeterministic automaton @xmath56 , and valid input @xmath27 .
let @xmath160 be the run of @xmath161 on input @xmath27 .
i claim that @xmath162 on input @xmath163 will have run @xmath10 , an accepting run of @xmath56 on input @xmath27 .
it suffices to show that the only occurrence of @xmath154 in @xmath10 is as the final character . by construction
, @xmath162 will satisfy the transition relations . as mentioned before the construction of @xmath162 also prevents backwards transitioning into the @xmath154 state for this particular input , since a reachable state can always be traced back to a reachable state . by lemma [ automatize ] ,
given a regular , length - preserving function @xmath58 , there is a nondeterministic automaton @xmath129 which takes in a word @xmath27 and has a run @xmath10 which projects to @xmath131 . by lemma [ decomp ] : @xmath164 by modifying the outer reverse moore machine , one can throw in the appropriate projection to its output map to produce @xmath131 .
this completes the proof of theorem [ detharvest ] .
this is a remarkable result .
every length - preserving regular function can be computed in a two - step process : one pass forwards through the input leaving behind some information , then a pass backwards through the input with this additional information to directly produce the output .
two passes suffice ; having more passes does nt increase the expressive power .
the only functions this paper has dealt with so far were length - preserving . in this section , i show that most of the interesting behavior of regular functions was already captured in the length - preserving case .
suppose @xmath93 is a regular function .
then there is a fixed constant @xmath165 associated to @xmath58 such that @xmath131 is no longer than @xmath166 for every @xmath27 . the proof is based on the pumping lemma .
let @xmath56 be a deterministic automaton with @xmath165 states accepting exactly words of the form @xmath140 . choose a specific @xmath27 and
suppose @xmath131 is longer than @xmath166 .
imagine what happens as @xmath56 reads in @xmath140 , specifically , after @xmath27 is finished , and @xmath56 is reading in characters of the form @xmath90 for some @xmath139 . because there are more positions like this than there are states of @xmath56 , by the pigeonhole principle some two positions will have the same state , say at positions @xmath167 and @xmath168 .
note however that if one removes all positions in @xmath140 between @xmath167 and @xmath168 ( including @xmath167 , excluding @xmath168 ) one still has an accepting run , of the form @xmath169 for some @xmath170 strictly shorter than @xmath131 .
this contradicts the assumption that @xmath56 accepted exactly words of the form @xmath140 . from this
, one can also show that for any @xmath61-ary regular function @xmath58 , there is a fixed constant @xmath165 associated to @xmath58 such that @xmath171 is no longer than @xmath165 plus the length of the maximum input .
note that the automaton @xmath56 in the proof never encounters the input character @xmath172 . as such
, one may assume that this character acts as the identity on the states of @xmath56 .
the resulting automaton recognizes all pairs of the form @xmath173 for arbitrary @xmath174 ( @xmath175 simply represents a @xmath165-fold composition of the @xmath176 function ) . by adding in a counter
, one can recognize exactly the pairs of the form @xmath177 .
as such , the function which takes @xmath178 to @xmath179 for @xmath180 is a length - preserving regular function @xmath170 , and thus can be written as a composition of character - wise maps , truncated moore machines , and truncated reverse moore machines as in theorem 1 .
now @xmath58 can be written as : @xmath181 where @xmath182 removes final @xmath65 characters .
note that one does not know how many final @xmath65 characters there will be . for functions which reduce length , the number will be more than @xmath165 and could be as much as @xmath183 .
one might be tempted to try to replace @xmath182 with some function like @xmath184 ( or , even less suited to the task , @xmath36 ) , which either removes a single final @xmath65 or leaves the word alone if it can not . however , since there are regular functions which take words of arbitrary length and reduce them to length 1 , the generating set needs a generator that can produce unbounded shortening as well .
note that if @xmath58 is @xmath61-ary for @xmath106 , it follows that : @xmath185 for some regular , length - preserving function @xmath170 .
as such : any regular function can be written as a multivariable composition of : * truncated moore machines , * truncated reverse moore machines , * character - wise maps , * @xmath64 ( allowing one to generate multiary character - wise maps ) , * @xmath34 for various @xmath35 , * @xmath182 .
it s worth noting here that this generating set is infinite . specifically , there are an infinite number of moore machines and reverse moore machines .
there are also an infinite number of character - wise maps , but this is nt essential
one could use encoding methods to work purely with a single two - character alphabet ( plus , optionally , the dummy character @xmath65 ) .
the infinitude of the generating set , however , is essential .
the easiest way to see this is to talk about period introduction .
provided the input to a regular function has a sufficiently long periodic portion in the middle , the output of the regular function will also have a long periodic portion in the middle ( it suffices to verify this of the generators above ) .
what s more , the period of the periodic portion of the output can only have prime factors which show up either in the periods of the periodic portions of the inputs or which are smaller than the number of states of the associated automaton to the regular function .
however , one can easily build regular functions which introduce any prime factor into the periodicity of their inputs , so there must not be any bound on the sizes of associated automata to regular functions in the generating set . in summary : any set of regular functions which generates all regular functions under multivariable composition must be infinite .
i conclude this paper with a discussion of known results regarding the krohn - rhodes theorem .
first , i will provide a proof of the krohn - rhodes theorem adapted from ginzburg @xcite .
then i will use the krohn - rhodes theorem to decompose the moore machine and reverse moore machine generators into smaller , simpler generators .
the idea of interpreting a the cascade given by the krohn - rhodes theorem as a composition of functions can be found in @xcite .
this paper will spell out this composition precisely and in modern notation , as has been done for previous compositions .
a few sections will be dedicated to cleaning up the resulting set of generators , followed by proposed future research .
in this section , i present the krohn - rhodes theorem as adapted to the context of multivariate composition of regular functions .
the original proof of the krohn - rhodes theorem , in @xcite , was presented in terms of wreath products of semigroups .
more modern presentations of the krohn - rhodes theorem typically present it in terms of the cascade product of finite state transducers .
the cascade product of two transducers @xmath186 is a single system consisting of both machines .
first , machine @xmath187 reads in both the input to the system and the current state of @xmath188 to update its state .
when it has finished , machine @xmath188 updates its state based only on the input to the system . finally , an output is produced based on the states of @xmath188 and @xmath187 .
this reflects the reality of systems where updating the states of machines takes a small but appreciable amount of time . in a well designed system
, @xmath187 should not have to wait for @xmath188 to finish its update before it can update its state .
as such , @xmath187 uses the state of @xmath188 prior to reading the input to update .
the krohn - rhodes theorem separates out two extremes of behavior for finite automata . in general , reading in an input character induces a function on the states of the automaton .
this function may map two states to the same state or to separate states . at one extreme
, it may act as a _
permutation _ in which case it maps all states to separate states . in this case
, it is possible to undo this action .
one can recover the state before reading a character which acts as a permutation , provided one knows which character the automaton read . at the other extreme , an input character may act as a _ reset _ in which case it maps all states to the same state . in this case all information about the previous state is lost . * a _ permutation automaton _ if each of its inputs acts as a permutation on its states , * a _ reset automaton _ if each of its inputs acts as a reset or the identity on its states , * a _ permutation - reset automaton _ if each of its inputs acts as a permutation or a reset on its states .
given a transparent moore machine @xmath26 , one can write its truncated action @xmath40 as a multivariable composition of truncated actions of permutation - reset moore machines @xmath189 for @xmath61 the number of states of @xmath26 , and a final character - wise map @xmath58 .
what s more , this composition takes on a fairly simple form .
let : @xmath190 then : @xmath191 every map in this composition is length - preserving .
of course one can write this as simply one large multivariable composition of a character - wise map and truncated actions of permutation - reset transparent moore machines , but this is unwieldy to write down .
the above also presents an efficient way of computing the composition , although readers concerned with efficiency are encouraged to look into the holonomy decomposition @xcite .
the proof is inductive : i show that for every transparent moore machine @xmath26 , there s another transparent moore machine @xmath192 that keeps track of a state @xmath26 is not in in a permutation - reset way .
this reduces the amount of information that needs to be kept track of by one state , and one can keep doing this until one has kept track of all the information to know what state @xmath26 is in .
the proof in @xcite allows for the possibility that one can keep track of several states the automaton @xmath26 is not in in a permutation - reset way at the same time , as opposed to one at a time in the proof below .
this is more efficient , but adds needless complexity to the proof .
* @xmath170 either acts as : * * a bijection from @xmath49 to @xmath195 ( a permutation on the position indices ) , * * or has singleton image ( a reset on position indices ) , * and for @xmath196 , with @xmath197,@xmath198 . * for any @xmath199 , @xmath58 maps elements of @xmath200 to @xmath201 .
suppose @xmath58 does not act as a bijection .
then @xmath170 which maps everything to the smallest element of @xmath195 which is not in the image of @xmath58 has singleton image and satisfies the inequality condition . given a transparent moore machine @xmath203 , there is a permutation - reset transparent moore machine @xmath192 with the same state set such that on input @xmath27 , the state of @xmath26 at any one time is not the state of @xmath192 . if the action of @xmath35 was a permutation originally , it is still a permutation in the new automaton .
this permutation not only maps the state the automaton is in before reading @xmath35 to the state afterwards , but also from a state the automaton is not in before reading @xmath35 to a state the automaton is not in afterwards . in the second case , note that the choice of @xmath124 does not depend on @xmath150 , so this action is a reset .
obviously , it maps a state the original automaton is not in before reading @xmath35 to a state the automaton is not in after reading @xmath35
. given transparent moore machines @xmath209 as above with state sets @xmath2 , there is a third transparent moore machine : @xmath210 such that for any input @xmath27 , if @xmath26 on reading @xmath27 winds up in state @xmath150 , and @xmath192 on reading @xmath27 winds up in state @xmath211 then @xmath212 on reading @xmath213 will wind up in state @xmath214 , where @xmath150 is the element in positionst element of a set as being in position @xmath167 . to avoid confusion ,
i avoid using the notation `` @xmath167th element '' , instead using notation `` the element in position @xmath167 '' . ]
@xmath214 of @xmath215 .
* computing @xmath220 .
this is the state @xmath192 says that @xmath26 is not in after reading @xmath35 . *
computing @xmath150 , the @xmath167th element of @xmath215 .
this is the state of @xmath26 prior to reading @xmath35 .
* return @xmath168 , the index of @xmath208 in @xmath221 .
* inductive case : * given @xmath26 , one can write : @xmath224 with @xmath192 permutation - reset . by the inductive hypothesis , one can write @xmath225 , which has one fewer state than @xmath26 , as : @xmath226 where : @xmath227 alternately , fiddling with some parentheses in the definitions in the automata : @xmath231 where : @xmath232 in which case : @xmath233 one can combine @xmath160 and @xmath58 to get a single character - wise function on @xmath234 , thus completing the induction .
the proof is still straightforward if one unwinds the induction . in the construction
, @xmath235 is keeping track of a state @xmath26 is not in , but it may as well be keeping track of an index for a state @xmath26 is not in .
@xmath236 is keeping track of an index of a state @xmath26 is not in once one has removed the state in position @xmath235 from @xmath2 .
@xmath237 is keeping track of an index of a state @xmath26 is not in once one has removed the states that @xmath236 and @xmath237 are keeping track of from @xmath2 . and so forth .
one can formalize this indexed removal process as follows : one can interpret an ordinal removal sequence for @xmath61 as a series of commands operating on an ordered set of size @xmath61 of the form `` remove the @xmath241st smallest element remaining . ''
note that as elements are removed , there are fewer elements remaining , hence the decreasing upper limit on @xmath242 in the second constraint .
consistent with the rest of this paper , i begin indexing with 0 , so a 0 means remove the smallest element . | while automata theory often concerns itself with regular predicates , relations corresponding to acceptance by a finite state automaton , in this article i study the regular functions , such relations which are also functions in the set - theoretic sense . here
i present a small ( but necessarily infinite ) collection of ( multi - ary ) functions which generate the regular functions under composition . to this end , this paper presents an interpretation of the powerset determinization construction in terms of compositions of input - to - run maps .
furthermore , known results using the krohn - rhodes theorem to further decompose my generating set are spelled out in detail , alongside some coding tricks for dealing with variable length words .
this will include two clear proofs of the krohn - rhodes theorem in modern notation . |
the close interplay between theory and experiments has always led to fundamental discoveries in condensed matter physics .
this is particularly evident in the search for topological states of matter , starting in the 80 s with the experimental observation of the integer quantum hall ( qh ) effect @xcite , characterized by chiral metallic edge states and insulating bulk states .
few years later , also fractional plateau of the conductance were discovered @xcite , theoretically explained by laughlin as a manifestation of the fractional qh effect @xcite .
after the first theoretical proposal for realizing topological states of matter in the absence of magnetic fields @xcite , bernevig _ et al .
_ suggested that hgte / cdte quantum wells behave as a topological insulator ( ti ) @xcite . realizing the quantum spin hall ( qsh ) effect in the presence of time reversal ( tr ) symmetry @xcite .
soon after the group of molenkamp provided the first experimental report @xcite , showing that nonlocal transport is due to protected helical edge states @xcite , with spin and momentum bound to each other @xcite . with this picture in mind , it is natural to wonder how long the experimental realization of fractional tis , recently proposed by levin and stern @xcite , will take . at the simple intuitive level , fractional tis can be thought as a superposition of two copies of fractional qh states in opposite magnetic fields , with the emergence of tr protected helical edge states with fractional excitations . beyond the emergence of fractional qsh effect in the flat - band lattice models
@xcite , possible experimental realization of such a state have been proposed , ranging from a fractional qh state with a varying @xmath0-factor in the presence of a thin insulating barrier to electron - hole bilayers realized in electron - hole symmetric systems @xcite . + it is thus important to find possible strategies and experimental signatures to identify this novel topological state of matter , which could be exploited also for quantum computing @xcite .
a breakthrough in this direction has been done recently by beri and cooper @xcite , who demonstrated a surprising robustness of fractional tis to magnetic perturbations , which is completely unexpected in integer tis . based on the works carried out in fractional qh systems and integer tis @xcite
, they also showed peculiar transport properties when electrons tunnel through a quantum point contact ( qpc ) , which could be used to probe fractional tis @xcite .
+ in this context , we propose tunneling through a time dependent qpc , realized , for example , by applying time modulated gate voltages , to probe fractional tis . in the case of a static qpc , backscattering at the constriction always decreases the total current , and information about the tunneling processes are encoded in the power - law exponents as a function of bias and temperature .
however , the power laws are often hardly detectable and hidden by several additional mechanisms . here
, we propose a different and more robust method in order to discriminate between fractional and integer tis .
we demonstrate that in fractional tis a time dependent qpc enhances the current , contrary to the decreasing induced in integer tis for not too strong interactions .
this peculiar effect is strictly associated with quasiparticle tunneling , and provides a simple tool to identify fractional tis .
this method takes advantage from the resolving power of finite frequency transport properties in the simple framework of dc current measurements .
different tunneling processes are present , involving one or more excitations depending on the strength of interactions ; we show that the proposed setup also provides a simple way to discriminate between these different processes , without requiring the knowledge of the detailed form of the current .
+ the paper is organized as follows . in sec .
[ model ] we describe the interacting helical edge states of fractional tis and define the relevant tunneling processes . in sec .
[ bscurrent ] we evaluate the backscattering current induced by the time dependent qpc in the weak backscattering regime , and discuss how it is affected by the different tunneling processes , both in integer and in fractional tis .
this analysis leads us to draw the main conclusions of our work .
finally , sec .
[ conclusions ] is devoted to the conclusions .
we consider the upper edge of the fractional qsh bar with right - moving spin up ( @xmath1 ) and left - moving spin down ( @xmath2 ) excitations , and the opposite on the lower edge .
the effective lagrangian density is @xcite ( @xmath3 ) @xmath4 with @xmath5 the propagation velocity , @xmath6 and @xmath7 .
the filling factor is @xmath8 , with @xmath9 for integer tis and @xmath10 for fractional tis in the laughlin sequence .
the bosonic fields @xmath11 obey the commutation relations @xmath12=i\alpha\nu\pi\delta_{\alpha,\alpha ' } \delta_{\sigma,\sigma ' } \mathrm{sign}(x - x')$ ] .
they are related to electron @xmath13 and quasiparticle @xmath14 operators by @xmath15 with @xmath16 the microscopic cutoff length , and respectively correspond to excitations with charge @xmath17 and @xmath18 .
we have omitted the klein factors and the phase factors @xmath19 , which are not relevant for our discussion .
note that the two operators in eq .
coincide in the case of integer tis , since the fundamental excitation is the electron itself .
the hamiltonian is easily obtained from eq . @xmath20
the presence of electron - electron interaction on the edges is due to intra - mode interaction @xmath21 and inter - modes interaction @xmath22 with @xmath23 the electron density associated to the mode with chirality @xmath24 and spin @xmath25 .
note that not all the terms appear in eq .
, since modes with same chirality and different spin are spatially separated , so that we can neglect their mutual interaction . in the following we consider local interactions @xmath26
. the hamiltonian @xmath27 can be diagonalized by introducing charge and spin fields @xcite @xmath28 which satisfy @xmath29=i\frac{\pi}{2}\delta_{\lambda,\lambda ' } \mathrm{sign}(x - x')$ ] , so that it takes the form of a spinful luttinger liquid @xcite @xmath30.\ ] ] the charge and spin parameters are @xmath31 and @xmath32 , with @xmath33 and renormalized velocity @xmath34 here we focus on standard repulsive interactions , which imply @xmath35 ( @xmath36 in the absence of interactions ) @xcite .
+ in the presence of a constriction , tunneling from one edge to the other is allowed . in this work
we focus on the weak backscattering regime , where tunneling is a small perturbation to the decoupled edges configuration , focusing on the four terminal setup depicted in fig .
[ fig:4term ] . in this regime tunneling of quasiparticles dominate over electron tunneling for any interaction @xcite .
the most relevant processes are sketched in fig .
[ fig : processes ] .
they correspond to tunneling of single quasiparticles ( @xmath37 , fig . [
fig : processes](a ) ) , backscattering of two quasiparticles ( @xmath38 , fig .
[ fig : processes](b ) ) and tunneling of two quasiparticles from one edge to the other ( @xmath39 , fig .
[ fig : processes](c ) ) .
note that @xmath40 preserves the charge ( spin ) on each edge , thus one refers to it as the charge ( spin ) process . , @xmath41 and @xmath42 respectively .
solid red ( dashed blue ) lines represent spin up ( down ) excitations.,width=264 ] the corresponding hamiltonians , assuming without loss of generality the qpc located at @xmath43 , are @xmath44+\mathrm{h.c.}\nonumber,\\ h_c&=&2\pi a\lambda_c\left [ \psi^{\dagger}_{l\uparrow}\psi^{\dagger}_{l\downarrow}\psi_{r\uparrow}\psi_{r\downarrow}\right ] + \mathrm{h.c.}\nonumber,\\ h_s&=&2\pi a \lambda_s\left [ \psi^{\dagger}_{l\uparrow}\psi^{\dagger}_{r\downarrow}\psi_{r\uparrow}\psi_{l\downarrow}\right ] + \mathrm{h.c.}\nonumber,\end{aligned}\ ] ] where @xmath45 , with @xmath46 , are the tunneling amplitudes . in the following
we consider a time dependent constriction , with @xmath47 .
this can be achieved through a periodic modulation of the gate voltages that define the qpc .
+ an additional time dependence is introduced by gauging away @xcite the bias voltages @xmath48 , with @xmath49 labeling the four terminals ( see fig .
[ fig:4term ] ) .
finally the bosonized tunneling hamiltonians read @xmath50 from eqs . , and we note that only two combinations of the bias voltages are relevant : the charge @xmath51 and spin bias @xmath52 which couple to the charge and spin sectors respectively . in eq .
the single quasiparticle charge couples to both the charge and spin bias voltages , while in eq .
( eq . ) a double charge couples to the charge ( spin ) bias only
. + the charge and spin bias voltages are also related to the charge ( @xmath53 ) and spin ( @xmath54 ) currents flowing through the topological bar in the absence of the qpc , @xmath55 , with @xmath56 the charge and spin conductances .
the total charge current in the presence of the constriction is @xmath57 , with @xmath58 the backscattering current induced by tunneling at the constriction @xmath59.\ ] ] here , @xmath60 is the total number of right - moving quasiparticles .
the backscattering current can be written as @xmath61 , where different contributions are due to the different tunneling processes @xmath62 $ ] , with @xmath63 note that the spin process does not contribute to the backscattering , since @xmath39 preserves the total number of right - moving excitations @xcite . + we begin focusing on the charge configuration with @xmath64 and @xmath65 , corresponding to a standard two terminal setup with @xmath66 and @xmath67 .
we will comment on alternative configurations in the last part of the work .
we employ the keldysh formalism @xcite in order to evaluate the expectation value of the backscattering currents at lowest order in the tunneling amplitudes @xmath68 with @xmath69 the keldysh time - ordering and @xmath70 labeling the two branches of the keldysh contour @xmath71 . the time dependent qpc generates dc and ac contributions at frequency @xmath72 . in the following
we will focus on the dc components only , which read @xmath73},\label{eq : i1k}\\ \langle i_c^{(\omega)}(v_c)\rangle&=&-ie^*\left |\tilde{\lambda_c}\right |^2\sum_{\eta,\eta_1}\eta_1\int_{-\infty}^{\infty } d\tau \cos\left ( \omega\tau\right ) \nonumber\\ & \times&\sin\left ( 2e^*v_c\tau\right ) e^{8\nu^2\mathcal{g}_c^{\eta,\eta_1}(\tau)}\label{eq : ick } , \end{aligned}\ ] ] with @xmath74 . for clarity
, we explicitly write the dependence on @xmath75 and @xmath76 in the dc component . since the contributions in eqs . , are even ( odd ) functions of @xmath75 ( @xmath76 ) , in the following we restrict the discussion to the case @xmath77 . because of the parity properties of the keldysh green s functions @xmath78 , eqs . and can be written in terms of @xmath79 only , with @xmath80 @xmath81 being the gamma function , @xmath82 the inverse temperature and @xmath83 .
+ for a static qpc ( @xmath84 ) the backscattering currents are ( @xmath85 ) @xmath86 where @xmath87 is the euler beta function and @xmath88 are the backscattered charge and scaling dimensions for the processes .
note that eqs . and are valid both for quasiparticle tunneling in fractional tis ( with @xmath89 ) and for electron tunneling in integer tis ( with @xmath90 and @xmath91 ) .
the smallest scaling dimension in eq . identifies the dominant tunneling process .
in particular one finds that single quasiparticle tunneling ( @xmath92 ) dominates at weak interactions @xmath93 , with @xmath94 . at stronger interactions @xmath95 the charge process ( @xmath96 ) dominates . in any case , for a static qpc , both processes always reduce the total current ( @xmath97 for @xmath98 ) , as can be argued @xcite from eq . .
+ however , when an ac modulated ( @xmath99 ) qpc is considered , the backscattering contributions are no longer driven by the bias only , but the time dependent constriction comes into play .
this can be seen by writing eqs . and
as @xmath100 which shows a striking similarity with the photo - assisted current .
we find that the current is driven by two effective channels @xcite with voltages @xmath101 , and the sign of @xmath102 now depends on how the two terms in eq .
even if there is only one time dependent parameter ( the tunneling amplitude ) , the backscattering current has been interpreted @xcite as originated from a pumping mechanism , since the effect of gauging out the bias voltages is to give rise to the additional time dependent parameter necessary to generate a pumping current . in this case , it is no longer obvious that the constriction leads to a decreasing of the current , but the time dependent qpc may be able to pump a current in the forward direction , leading to a global current enhancement @xcite . at low temperatures ,
one finds from eqs . and the necessary condition for the process @xmath103 to enhance the current to be @xmath104 .
under this condition , the enhancement is expected for @xmath105 .
( units @xmath106 ) as a function of @xmath107 for a fractional ti with @xmath108 , for @xmath109 , @xmath110 and different interaction strength : @xmath36 ( solid black ) , @xmath111 ( dashed blue ) and @xmath112 ( dot - dashed red ) .
( a ) the single quasiparticle contribution enhances the total current at weak interactions for @xmath113 , while ( b ) the charge contribution enhances the total current at stronger interactions for @xmath114.,width=302 ] + let us consider fractional tis , and study whether the current is enhanced or decreased by the time dependent constriction .
positive values of @xmath102 correspond to an enhancement of the current more than the universal value @xmath53 ( remember we are focusing on @xmath98 ) . at weak interactions @xmath93 the dominant contribution
is the single quasiparticle tunneling with scaling dimension @xmath115 ( see eq . ) . in this regime @xmath116 and
the current is enhanced for @xmath117 ( fig . [ fig4b](a ) ) .
the single quasiparticle process decreases the current for stronger interactions @xmath118 ( red curve in fig .
[ fig4b ] ) ; however , in this regime tunneling is dominated by the charge process , which leads to enhancement already at moderate interactions @xmath119 for @xmath120 , as shown in fig . [
fig4b](b ) .
it is worth noticing that at the crossover @xmath121 both the single quasiparticle and charge processes contribute to enhance the current .
we thus expect a _ global enhancement _ of the current in fractional tis in the presence of a time dependent constriction , due either to single quasiparticle ( for weak interaction ) or to charge tunneling ( for strong interaction ) .
+ let us now compare this scenario with the one obtained for an integer ti , where only electrons can tunnel . from eq .
we find that single electron tunneling , which dominates at weak interactions @xmath93 , always decreases the current . at stronger interactions
@xmath95 charge tunneling dominates , which however still leads to a decreasing as long as @xmath122 .
+ we then conclude that , if interactions are not too strong @xmath122 , the enhancement of the current in the presence of an ac modulated qpc is a peculiar feature of fractional tis .
+ other signatures can help to discriminate fractional tis also for @xmath123 .
indeed , it is worth noticing that , as shown in fig .
[ fig4b ] , a transition of @xmath102 from positive to negative values is expected around @xmath124 , and thus depends on the backscattering process through the backscattered charge @xmath125 . since @xmath76 and @xmath75 can be externally controlled , one should be able to extract information about the backscattered charge , and thus on the fractional or integer nature of the system , simply by varying the bias .
in particular , by observing whether the current displays a transition around @xmath126 or @xmath127 , one could also discriminate between single - particle and multi - particle tunneling .
in other words , the position of the transition in the backscattering current has the power to resolve the presence of different fractional excitations . in this perspective
, this quantity has similarities with the photo - assisted differential conductance or finite frequency noise @xcite , since they are similarly affected by the scaling dimensions and by the josephson resonances of the tunneling excitations .
+ furthermore , the current enhancement represents a signature of quasiparticle ( and _ not _ electron ) tunneling . indeed , evaluating the tunneling contribution of electrons ( despite not the dominant process )
would essentially result in replacing @xmath128 and @xmath129 in eqs . and
and one can conclude that both single and double electron tunneling decrease the current for not too strong interactions ( as long as @xmath130 ) .
+ we have thus proposed a simple strategy to identify quasiparticle tunneling in fractional tis .
this method could complement other peculiar responses , such as the quantized conductance , the stability to tr breaking perturbations or the power - law behavior of the backscattering current .
however we stress that , to identify quasiparticle tunneling , one does not need to study the details of the current ( like for example for extracting the exponents of the power laws @xcite ) , but the enhancement of the current in the presence of a time dependent qpc provides by itself a powerful signature of the presence of a fractional ti . for strong interactions
@xmath123 , the current is enhanced both in fractional and in integer tis ; however , the transition from positive to negative values depends on the backscattered charge , allowing to discriminate between fractional and integer excitations .
so we can conclude that the proposed protocol is stable against electron interactions .
+ before concluding , we briefly comment on a different bias voltage configuration , that is the so called spin configuration , with @xmath131 and @xmath132 , corresponding to the presence of a bias voltage between the upper ( at voltage @xmath133 ) and the lower edge ( at voltage @xmath134 ) . here a net charge tunneling current flows across the constriction from one edge to the other , arising from a single quasiparticle tunneling contribution , given by eq . with @xmath135 , and a spin tunneling contribution , given by eq . with @xmath135 and @xmath136 .
the dominant contribution is the single quasiparticle one with scaling dimension @xmath115 . in the presence of a static qpc the tunneling current always flows from positive to negative voltage , as expected .
however , when considering a time dependent qpc in fractional tis the tunneling current may reverse its direction , flowing _ against _ the bias .
this anomalous response , which is expected to occur as long as interactions are not too strong ( @xmath137 ) , represents another peculiar feature of fractional tis with respect to their integer version , where it is not expected to occur at all .
we showed that an enhancement of the current is expected for a fractional topological insulator with a time dependent constriction , allowing for quasiparticle tunneling from one edge to the other .
this effect is not expected to occur neither in integer topological insulators ( for not too strong interactions ) nor in fractional topological insulators when electrons ( and not quasiparticles ) tunnel through the constriction , and thus represents an important tool to identify fractional quasiparticles .
this simple method allows to unambiguously identify fractional topological insulators without requiring the knowledge of the detailed form of the current ( like for example for extracting the power law exponents ) .
furthermore we showed that this effect is robust against electron interaction , and discussed how one can discriminate between single and multi particle processes .
the supports of eu - fp7 via grant no .
itn-2008 - 234970 nanoctm and miur - firb2012 - project hybridnanodev via grant no.rbfr1236vv are acknowledged .
k. von klitzing , g. dorda , m. pepper , phys .
* 45 * , 494 ( 1980 ) .
d. c. tsui , h. l. stormer , and a. c. gossard , phys .
lett . * 48 * , 1559 ( 1982 ) .
r. b. laughlin , phys .
lett . * 50 * , 1395 ( 1983 ) . f. d. m. haldane , phys .
lett . * 61 * , 2015 ( 1988 ) . c. l. kane , and e. j. mele , phys .
* 95 * , 226801 ( 2005 ) .
m. z. hasan and c. l. kane , rev .
* 82 * , 3045 ( 2010 ) .
qi and s .- c .
zhang , rev .
phys . * 83 * , 1057 ( 2011 ) .
b. a. bernevig , t. l. hughes , s .- c .
zhang , science * 314 * , 1757 ( 2006 ) .
m. knig , s. weidmann , c. brne , a. roth , h. buhmann , l. w. molenkampf , x .-
qi , and s .- c .
zhang , science * 318 * , 766 ( 2007 ) .
a. roth , c. brne , h. buhmann , l. w. molenkamp , j. maciejko , x .-
qi , and s .- c .
zhang , science * 325 * , 294 ( 2009 ) . c. wu , b. a. bernevig , and s .- c .
zhang , phys .
* 96 * , 106401 ( 2006 ) .
m. levin and a. stern , phys .
lett . * 103 * , 196803 ( 2009 ) .
m. levin and a. stern , phys .
b * 86 * , 115131 ( 2012 )
. t. neupert , l. santos , s. ryu , c. chamon , and c. mudry , phys .
b * 84 * , 165107 ( 2011 ) .
t. neupert , l. santos , s. ryu , c. chamon , and c. mudry , phys .
lett . * 108 * , 046806 ( 2012 ) .
wei li , d. n. sheng , c. s. ting , and yan chen , arxiv:1407.1563 . n. h. lindner , e. berg , g. refael , and a. stern , phys . rev .
x * 2 * , 041002 ( 2012 ) .
m. cheng , phys .
b * 86 * , 195126 ( 2012 ) .
m. burrello , b. van heck , and e. cobanera , phys .
b * 87 * , 195422 ( 2013 ) .
p. nikolic , t. duric , and z. tesanovic , phys .
lett . * 110 * , 176804 ( 2013 ) .
b. beri and n. r. cooper , phys .
lett . * 108 * , 206804 ( 2012 ) .
j. c. y. teo and c. l. kane , phys .
b * 79 * , 235321 ( 2009 ) .
hou , e .- a .
kim , and c. chamon , phys .
lett . * 102 * , 076602 ( 2009 ) .
g. dolcetto , s. barbarino , d. ferraro , n. magnoli , and m. sassetti , phys .
b * 85 * , 195138 ( 2012 )
. t. l. schmidt , phys .
lett . * 107 * , 096602 ( 2011 ) . c. -x .
liu , j. c. budich , p. recher , and b. trauzettel , phys .
b * 83 * , 035407 ( 2011 ) .
a. strom and h. johannesson , phys .
lett . * 102 * , 096806 ( 2009 ) .
r. de picciotto , m. reznikov , m. heiblum , v. umansky , g. bunin , and d. mahalu , nature * 389 * , 162 ( 1997 ) .
l. saminadayar , d. c. glattli , y. jin , and b. etienne , phys .
79 * , 2526 ( 1997 ) .
y. c. chung , m. heiblum , and v. umansky , phys .
lett . * 91 * , 216804 ( 2003 ) .
m. dolev , m. heiblum , v. umansky , a. stern , and d. mahalu , nature * 452 * , 829 ( 2008 ) .
m. carrega , d. ferraro , a. braggio , n. magnoli , and m. sassetti , phys .
lett . * 107 * , 146404 ( 2011 ) .
huang , s. t. carr , d. gutman , e. shimshoni , and a. d. mirlin , phys .
b * 88 * , 125134 ( 2013 ) .
t. giamarchi , _ quantum physics in one dimension _ ( oxford university press , oxford , 2003 ) . note that when talking about interactions we refer , in analogy with ref .
, to interactions on the edges given by eqs . and , and not to the coulomb interaction which gives rise to the fractional state as generated by the topological order in the bulk .
d. e. feldman and y. gefen , phys .
b * 67 * , 115337 ( 2003 ) . j .-
souquet and p. simon , phys .
b * 86 * , 161410(r ) ( 2012 ) .
lee , y .- l .
lee , and c .- h .
chung , phys .
b * 86 * , 235121 ( 2012 ) .
t. martin , _ in nanophysics : coherence and transport , les houches session lxxxi _ , edited h. bouchiat _
et al_. ( elsevier , amsterdam , 2005 ) .
note that @xmath138 , while @xmath139 .
d. makogon , v. juricic , and c. m. smith , phys .
b * 74 * , 165334 ( 2006 ) .
d. makogon , v. juricic , and c. m. smith , phys .
b * 75 * , 045345 ( 2007 ) .
a. agarwal and d. sen , phys .
rev . * 76 * , 035308 ( 2007 ) .
m. j. salvay , phys .
b * 79 * , 235405 ( 2009 ) .
d. schmeltzer , phys .
b , * 63 * , 125332 ( 2001 ) .
i. safi and e. v. sukhorukov , europhys . lett . * 91 * , 67008 ( 2010 ) .
m. carrega , d. ferraro , a. braggio , n. magnoli , and m. sassetti , new j. phys . * 14 * , 023017 ( 2012 ) .
d. ferraro , m. carrega , a. braggio , and m. sassetti , new j. phys .
* 16 * , 043018 ( 2014 ) . | we analyze the backscattering current induced by a time dependent constriction as a tool to probe fractional topological insulators .
we demonstrate an enhancement of the total current for a fractional topological insulator induced by the dominant tunneling excitation , contrary to the decreasing present in the integer case for not too strong interactions .
this feature allows to unambiguously identify fractional quasiparticles .
furthermore , the dominant tunneling processes , which may involve one or two quasiparticles depending on the interactions , can be clearly determined . |
many jets emanating from active galactic nuclei ( agn ) show narrow and well - collimated structures from less than a parsec ( pc ) up to mpc scales ( e.g. , zensus 1997,references therein ) . from a theoretical point of view ,
several models were proposed concerning the formation and collimation process of a jet . for the formation of jets , radiative pressure models
( e.g. , icke 1980 , 1989 ; sikora et al .
1996 ) and magnetic field models ( e.g. , benford 1978 ; blandford , payne 1982 ; uchida , shibata 1985 , 1986 ; meier et al .
2001 ; koide et al .
2002 ) have been mainly discussed . in the latter models ,
the toroidal component is crucial for the collimation process , and because of the toroidal component both the formation and collimation of jets would be simultaneously explained .
high - resolution observations using very long baseline interferometry ( vlbi ) have recently been providing some evidence in support for the mhd models ( junor et al .
1999 ; gabuzda 2000 ; gabuzda et al .
it is , however , not simple to see the toroidal component or helical structure of the magnetic field by observing the projected direction of the magnetic fields .
the well - known quasar 3c 273 , at a redshift of @xmath0 , is one of the brightest quasars . for a hubble constant of @xmath1 = 100 km s@xmath2 mpc@xmath2 and a deceleration parameter of @xmath3 ,
an angular resolution of one milli - arcsecond ( mas ) corresponds to a linear resolution of 1.86pc .
the jet components in 3c273 have displayed superluminal motions ( cohen et al .
1971 ) , with the speeds of the apparent proper motion being 4.9 to 7.7 @xmath4 derived from vlbi monitoring observations ( abraham et al . 1996 ) .
a vlbi space observatory programme ( vsop ) observation revealed a double - helical structure in the 3c273 jet , which was attributed to a kelvin
helmholtz ( k h ) instability ( lobanov , zensus 2001 ) .
vlbi polarimetric observations have shown that the jet of 3c 273 is highly polarized ( roberts et al .
1990 ) , which makes it suitable to study the rotation measure ( @xmath5 ) .
observations with the very large array ( vla ) showed @xmath5 variations of 5 rad m@xmath6 on the arcsecond scale ( roberts et al .
1990 ) . on the mas scale , however , very long baseline array ( vlba ) observations revealed @xmath5s of up to 2000 rad m@xmath6 towards the nuclear regions , while towards the jet the @xmath5s are around 200 rad m@xmath6 ( taylor 1998 ) .
the high @xmath5s towards the core are probably caused by the magnetic field and plasma in the narrow line region ( nlr ) around the core ( taylor 1998 ) .
to investigate the 3d structure of the magnetic field , a detailed analysis of both the @xmath5 and the polarization angle ( @xmath7 ) should be useful .
the @xmath5 is related to the electron density , @xmath8 , and the magnetic field component parallel to the line of sight , @xmath9 , as @xmath10 , where @xmath11 represents integration along the line of sight . in an optically thin plasma ,
the projected direction of the magnetic field ( a component perpendicular to the line of sight ) is perpendicular to @xmath7 .
since @xmath5 is measured by the polarization angles ( @xmath12 ) at different wavelengths ( @xmath13 ) through the equation @xmath14 , a more accurate @xmath5 can be measured at longer wavelengths . when the @xmath5 is very large , however
, longer - wavelength observations may result in aliasing due to the 180@xmath15 ambiguity of @xmath7 . to avoid this , a combination of observations at either ( i ) both long and short wavelengths or ( ii ) multiple longer wavelengths separated by only a small wavelength difference is required .
therefore , a proper selection of observing the frequencies is essential in deriving the @xmath5 distribution of jets .
vlba polarimetry could thus give one of essential information as to whether or not the magnetic field plays a dominant role in jet formation and/or collimation .
we analyzed polarimetric observations of the 3c273 jet based on archival vlba data .
observations were carried out on 1995 december 9 at 4.702 , 4.760 , 4.890 , and 4.990 ghz in the 5ghz band and 1995 november 22 at 8.102 , 8.240 , 8.420 , and 8.590 ghz in the 8ghz band using all ten stations of the vlba .
each if has an 8-mhz bandwidth .
the target of these observations was m87 , with 3c273 being observed as a calibration source .
both left and right circular polarizations were recorded at each station .
an _ a priori _ amplitude calibration for each station was derived from a measurement of the antenna gain and system temperatures during each run .
fringe fitting was performed on each if and polarization independently using the aips task fring . after deriving the delay and rate difference between parallel - hand cross correlations ,
the cross - hand correlations were fringe fitted to determine the cross - hand delay difference .
once the cross - hand delay difference was determined , the instrumental polarizations of the antennas were determined for each if in the same bands with an unpolarized source , oq 208 , using the aips task pcal .
the polarization angle offset at each station was calibrated using data of oj 287 observed by the university of michigan radio astronomy observatory ( umrao ) .
we note that the difference in the observing dates by umrao and vlba is within 1 day and 5 days for 5 and 8 ghz , respectively ; thus , the chance of variability between the two observations is small . because the total polarized flux of oj287 measured by the vlba is 31.8 @xmath16 1.7 mjy at 5 ghz and 26.4 @xmath16 1.3 mjy at 8 ghz , while the polarized fluxes measured by the umrao are 32.7 @xmath16 6.1 mjy and 24.9 @xmath16 13.7 mjy , respectively , the corresponding values are equal within the errors , confirming that a very high fraction of the integrated polarized flux was present on vlba scales .
we applied the same position angle calibrations for each if within the 5 ghz and 8 ghz bands .
the integrated @xmath5 of oj287 based on vla observations is small , + 31 rad m@xmath6 ( rudnick , jones 1983 ) .
the corresponding differences in the @xmath5 for different frequencies within the 5 ghz and 8 ghz bands are negligible , justifying our use of the same @xmath7 calibration for all ifs at each frequency .
we also note that even if a small offset in the @xmath5 were introduced , it would be constant all over the source , and could not give rise to structure in the @xmath5 distribution .
images were initially obtained using difmap , then imported into aips to self - calibrate the full datasets using the task calib before the final difmap image . in order to obtain the distributions of @xmath5 and @xmath7 , we tapered images at higher frequencies to match the resolution of the lowest frequency observation .
the restored beam size was 3.4 mas @xmath17 1.4 mas with the major axis at a position angle of @xmath183.@xmath157 . in order to register images at different frequencies ,
we refer to optically thin components of jet . as a result ,
the error of alignment was estimated to be within 10
we show images of the polarized intensities overlaid on the total intensity images at 5 and 8 ghz in figures 1a and b , respectively .
the core is located at the northeast end of the jet , and the jet extends along a position angle of 234@xmath15 .
the counter - jet is invisible , presumably due to doppler de - boosting .
the core is weakly polarized at the level of @xmath19 and @xmath20 at 5 and 8 ghz , as seen in previous observations ( leppnen et al .
1995 ; roberts et al . 1990 ; taylor 1998 ) .
the fractional polarization of component c3 at 8 ghz is 8% , while that at 5 ghz is 0.5% .
this suggests that this component is depolarized ; however , there are no obvious signs of depolarization within the 5 ghz and 8 ghz bands .
component c3 has a flat spectrum with @xmath21 ( @xmath22 ) between 5 ghz and 8 ghz .
we show the distribution of the projected magnetic field , which is obtained by de - rotating the @xmath12 to remove the effect of the @xmath5 , superposed on the total intensity contours in figure 2 .
the projected magnetic field initially runs roughly parallel to the jet , then turns oblique to it at the southwestern edge of c2 .
the @xmath5 distribution is shown in pseudo color in figure 3 .
regions of lower fractional polarization could not give a reliable @xmath5 distribution because of the short observation time .
the @xmath5 values on either side of the jet differ significantly from each other .
this gradient clearly appeared in independent observations at 8 ghz and 15 ghz ( zavala , taylor 2001 ) . in figure 4 , we show the @xmath5 distribution along the cross - section slice ab at a position angle of 36@xmath15 .
the @xmath5s differ by up to 210 rad m@xmath6 , where the error is less than 40 rad m@xmath6 and is typically 10 rad m@xmath6 .
the variation can be confidently associated with the jet itself , since it is unlikely that a foreground faraday screen would produce such a strong fine structure comparable to the jet width ( taylor 1998 ) .
it is possible that the distribution of @xmath5 could be due to a gradient in the gas density across the jet .
however , in this case , we would expect to see some evidence of an interaction between the jet and this denser gas , such as the development of a sheath of longitudinal magnetic field due to shear ( cf .
attridge et al . 1999 ) or sharp bending of the jet ( cf .
nan et al .
another interesting possibility is that the variation in the @xmath5 is related to differences in the component of the magnetic field along the line of sight . indeed
, the systematic distribution of @xmath5 across the jet can be very naturally interpreted in terms of a helical magnetic field .
let us consider the simple case of the side view of a straight jet ( a viewing angle is 90@xmath15 ) with a helical magnetic field along the jet .
the sign of @xmath9 differs on the two sides of the jet as the magnetic field reverses direction .
we would see an anti - symmetric distribution of the @xmath5 across the jet . in this simple case
, the @xmath5 would be zero at the center of the jet , because there is no @xmath9 component along the line of sight .
when the viewing angle decreases , an anti - symmetric distribution would remain , with only the addition of an offset in the absolute value of the @xmath5 .
in fact , the viewing angle to the jet of 3c273 must be small , as jet components show superluminal motion , with a viewing angle of @xmath23 16@xmath15 , estimated from a kinematical analysis ( abraham et al .
the distribution is anti - symmetric across the jet with the @xmath5 biased by 330 rad m@xmath6 ( see figure 4 ) .
if the viewing angle is smaller than the pitch angle , it is possible that @xmath9 always has the same sign . to illustrate this , we show the line - of - sight component of the helical magnetic field across the jet at several viewing angles in figure 5 .
the observed bias is ascribed to the longitudinal component of the helical magnetic field with a small viewing angle .
if this is the case , the toroidal component of the helical field is twisted like a right - hand screw as the jet moves downstream .
we note that the helical magnetic field can naturally produce the double - helical structure revealed by the vsop observation ( lobanov , zensus 2001 ) , because the helical magnetic field thread the emitting plasma and are tied to it ( frozen - in ) .
vsop monitoring observations would be able to discriminate between these two possibilities : a helical field or a k h instability .
the observed change in the direction of the projected magnetic field along the jet , as shown in figure 2 , can be explained by a change in the pitch angle of the helical field and incoherent polarization radiation .
again , let us consider the simple case when we look at a helical magnetic field in an optically thin jet from the side view .
the vector addition of the magnetic field integrated across the jet is always directed along the jet , independent of the pitch angle of the helical field .
however , the polarization angle has an ambiguity of 180@xmath15 , because the polarization planes for two magnetic fields pointing in exactly opposite directions would be the same .
therefore , the resultant direction of the polarization angle integrated across the jet would depend on the pitch angle of the field .
when the pitch angle is small , the magnetic field runs almost parallel to the jet , and the @xmath7s on both the near and far sides of the jet are almost perpendicular to the jet , so that the @xmath7 integrated across the jet is also perpendicular to the jet . on the other hand , when the pitch angle is large , the magnetic field at any given point is almost perpendicular to the jet , and the corresponding polarization angles are almost parallel to the jet . since the polarization radiation is incoherent , the vector accumulation of the two polarization radiations is almost parallel , resulting in a sudden jump in the integrated polarization angle of 90@xmath15 as the pitch angle changes .
the spectrum in the region where the projected magnetic field is perpendicular to the jet ( @xmath24 region ) is flatter than that in the region where the magnetic field is parallel to the jet ( @xmath25 region ) .
the fractional polarization is up to 25@xmath26 in the @xmath24 region , compared to @xmath2710@xmath26 in the @xmath25 region .
these facts , combined with the presence of the bright peak at the eastern edge of the @xmath24 region , suggest that this region is compressed by a shock . in fact , in the next bright region , further southwest ( i.e. the c1 )
, the magnetic field is again parallel to the jet , followed by a further knot where the magnetic field is essentially parallel along the jet .
thus , the southwestern edge of c2 probably has a large pitch angle due to compression by the shock and , consequently , the integrated magnetic field appears to be perpendicular to the jet .
in fact , the magnetic field structure is essentially the same as that in the @xmath25 region in terms of the systematic gradient of @xmath5 , which again indicates the helical field structure . on the line 3 mas southwest of the line ab , which is more than one beam apart , the amount of the @xmath5 gradient is 190 rad m@xmath6 , where the error is less than 27 rad m@xmath6 and is typically 10 rad m@xmath6 .
this interpretation may also be applicable to the dichotomous nature of the apparent magnetic field properties in jets between quasars and bl lacs ( gabuzda et al .
1989 , 1992 ) , which could be due simply to different pitch angles for their helical magnetic fields ( cf .
gabuzda et al .
2000 ) .
we believe that the most likely mechanism for the generation of the helical jet magnetic field is that the original magnetic field has been wound up by rotation of the accretion disk ( e.g. uchida , shibata 1985 ) . in this case
, we can identify that the accretion disk is rotating clockwise ( as we see it ) , because we know that the jet is approaching .
taking into account the generation mechanism of the helical magnetic field by mhd model , one would then observe a rotation in which the northern half of the disk is going away , and the southern half is approaching us .
next to m 87 ( ford et al . 1994 ; harms et al . 1994 ; macchetto et al .
1998 ) and the maser observations of a megamaser source , ngc 4258 ( miyoshi et al .
1995 ) , we can predict the direction of rotation in the disk around a possible super - massive black hole . this can be applied to anything that is responsible for winding up the magnetic field , like a possible spinning super - massive black hole ( koide et al .
. it would be interesting to confirm the @xmath5 distribution in a case in which the counter jet could also be observed .
the distribution should be anti - symmetric with respect to the agn core , if the helical field is generated by the rotation of the accretion disk or the spinning black hole .
in order to estimate the 3d structure of the magnetic field in the jet from agn , we have derived distributions of @xmath5 and projected magnetic field of 3c273 jet using archival data from a vlba polarimetry observation .
the systematic gradient across the jet in the distribution of the @xmath5 is interpreted as the helical magnetic field that is suggested by mhd models .
we derived the direction of the twist of the helical magnetic field as a right - hand screw from the sign of @xmath5 .
furthermore , two patterns of parallel and perpendicular to the jet axis in the projected magnetic fields could be explained well by the difference of the pitch angle of the helical magnetic field with incoherent polarization angles .
this interpretation can be applied to the difference of the tendency in the direction of the magnetic field between quasars and bl lacs .
it can be naturally predicted that the direction of the rotation of the accretion disk is clockwise , if the observed helical magnetic field is formed by mhd models .
all these points are in favor of a helical magnetic field operating predominantly in jet formation .
the authors thank d.c .
gabuzda for a critical reading of this paper and valuable comments , and p.g .
edwards and h. kigure for helpful comments .
this research has made use of data from the national radio astronomy observatory ( nrao ) archive , and from the university of michigan radio astronomy observatory , which is supported by funds from the university of michigan .
the national radio astronomy observatory is a facility of the national science foundation operated under cooperative agreement by associated universities , inc . | both in the faraday rotation measure and the intrinsic polarization angle , new features are revealed to indicate a helical magnetic field operating along the jet of the bright active galactic nuclei 3c 273 .
the helical field has been suggested to be related to the formation and collimation of jets by magnetohydrodynamic models .
the distribution of the rm shows a systematic gradient with respect to the jet axis , which is expected by a helical magnetic field .
in addition , the helical field can consistently explain two types in the direction of the projected magnetic field : parallel and perpendicular to the jet axis .
further , if the helical magnetic field is generated by winding up of an initial field by rotation of the accretion disk , we can uniquely determine the direction of the disk rotation , since the jet is approaching us . |
regarding experimental agreements of einstein gravity ( en ) in various area of astrophysics and cosmology , motivates one to consider it as an acceptable theory .
in addition , adding a constant term @xmath1 in the en - hilbert action may lead to agreement between the results of en-@xmath1 gravity with dark energy prediction .
on the other hand , general relativity is consistent with interaction of massless spin @xmath2 fields , in which related gravitons are massless particles with two degrees of freedom .
since the quantum theory of massless gravitons is non - renormalizable @xcite , one may be motivated for modifying general relativity to massive gravity . in order to build up a massive theory with a massive spin @xmath2 particle propagation
, one can add a mass term to the en - hilbert action .
this will result into graviton having a mass of @xmath3 which in case of @xmath4 , the effect of massive gravity will be vanished .
a class of massive gravity theory in flat and curved background which leads to absence @xcite and existence @xcite of ghost , have been investigated .
also , the quantum aspects of the massive gravity and a nonlinear class of massive gravity in ghost - free field @xcite have been explored in refs .
@xcite . generalization to nonlinearly charged massive black holes was done in ref . @xcite .
more details regarding the motivations of massive gravity is given in ref .
@xcite . in this paper
, we are interested in studying the nontrivial ads massive theory that was investigated in @xcite . the motivation for this consideration is due to fact that graviton shows similar behavior as lattice in holographic conductor @xcite . in other words ,
a drude like behavior is observed for the case of massless graviton in this theory which makes the role of graviton similar to lattice .
another interesting subject for study in this theory is metal - insulator transition massiveadscft .
recently , charged massive black holes with consideration of this theory have been investigated in @xcite .
the @xmath5 criticality of these solutions and their geometrical thermodynamic aspects have been studied @xcite .
also , the generalization to gauss - bonnet - maxwell - massive gravity and its stability , geometrical thermodynamics and @xmath5 criticality have been investigated @xcite .
one of the main problems of maxwell s electromagnetic field theory for a point - like charge is that there is a singularity at the charge position and hence , it has infinite self - energy . to overcome this problem in classical electrodynamics , born and infeld in ref .
@xcite introduced a nonlinear electromagnetic field , with main motivation , to solve infinite self - energy problem by imposing a maximum strength of the electromagnetic field .
then , hoffmann in ref .
@xcite investigated en gravity in the presence of born - infeld ( bi ) electrodynamics . in recent two decades , exact solutions of gravitating black objects in the presence of bi theory
have been vastly investigated @xcite .
another interesting property of bi is that , bi type effective action arises in an open superstring theory and d - branes are free of physical singularities @xcite . for a review of aspects of bi theory in the context of string theory see ref .
recently , there has been growing interest in eddington - inspired bi gravity in context of black holes and cosmology @xcite .
also , it was proposed that one can consider bi theory as a gravitational theory born - g .
dualization of the bi theory and some of the special properties of this theory have been investigated in ref .
@xcite .
there are several approaches for studying and obtaining critical behavior and phase transition points of black holes : first method is based on studying heat capacity .
it was pointed out that roots and divergencies of the heat capacity are representing phase transition points . in other words , in place of roots and divergencies of the heat capacity system may go under phase transition .
another important property of the heat capacity is investigation of the thermal stability .
systems with positive heat capacity are denoted to be in thermally stable states .
therefore , the stability conditions are indicated by changes in sign of heat capacity @xcite .
this is known as canonical ensemble . in the second method , by using the renewed interpretation of cosmological constant as thermodynamical variable , one can modify the thermodynamical structure of the phase space @xcite .
one of the most important property of this method is the similarity of critical behavior of the black holes and ordinary thermodynamical van der waals liquid / gas systems vander .
recently , it was pointed out that the extended phase space should be interpreted as an rg - flow in the space of field theories , where isotherm curves codify how the number of degrees of freedom @xmath6 ( or the central charge @xmath7 ) runs with the energy scale @xcite . on the other hand
, it was shown that variation of cosmological constant could be corresponded to variation of number of the colors in yang - mills theory residing on the boundary spacetime @xcite .
the third method is using geometrical concept for studying critical behavior . in other words , by employing a thermodynamical potential and its corresponding extensive parameters , one can construct phase space .
the divergencies of ricci scalar in constructed metric are denoted as phase transition points .
there are several metrics for this method that one can name : weinhold @xcite , quevedo @xcite and hpem @xcite which has mass as thermodynamical potential and ruppeiner @xcite in which entropy is considered as thermodynamical potential .
these metrics are used in context of heat capacity .
another set of metrics was introduced in ref .
@xcite which can be used in context of extended phase space .
finally , a fourth method was introduced in ref .
@xcite which is based on denominator of the heat capacity . in this method by replacing cosmological constant with pressure in denominator of the heat capacity and solving it with respect to pressure , a new relation is obtained for pressure .
the existence of maximum for obtained relation , represents the critical pressure and volume in which phase transition takes place .
the behavior of system in case of this method is consistent with ordinary thermodynamical concepts @xcite .
the outline of the paper will be as follow . in sec .
[ fieldeq ] , we introduce action and basic equations related to en - bi - massive gravity . sec .
[ massivesol ] is devoted to obtain the black hole solutions of this gravity and investigation of the geometrical structure of them . in the next section ,
we calculate conserved and thermodynamic quantities related to obtained solutions and check the validation of the first law of thermodynamics . in section [ stability ] , we study thermal stability of the en - bi - massive black hole solutions in canonical ensemble .
next , we consider cosmological constant as pressure and study the critical behavior of the system .
then we employ the geometrical methods for investigating thermodynamical behavior of the system and extend this study by another method . in the last section we present our conclusions .
the @xmath8-dimensional action of en - massive gravity with negative cosmological constant and a nonlinear electrodynamics is @xmath9 , \label{action}\]]where @xmath10 is the scalar curvature , @xmath11 is the negative cosmological constant and @xmath12 is a fixed symmetric tensor . in eq .
( [ action ] ) , @xmath13 are constants and @xmath14 are symmetric polynomials of the eigenvalues of the @xmath15 matrix @xmath16 which can be written as follows @xmath17 , \;\;\;\;\ ; \mathcal{u}_{2 } = % \left [ \mathcal{k}\right ] ^{2}-\left [ \mathcal{k}^{2}\right ] , \;\;\;\;\ ; \mathcal{u}_{3 } = \left [ \mathcal{k}\right ] ^{3}-3\left [ \mathcal{k}\right ] % \left [ \mathcal{k}^{2}\right ] + 2\left [ \mathcal{k}^{3}\right ] , \notag \\ & & \mathcal{u}_{4 } = \left [ \mathcal{k}\right ] ^{4}-6\left [ \mathcal{k}^{2}% \right ] \left [ \mathcal{k}\right ] ^{2}+8\left [ \mathcal{k}^{3}\right ] \left [ \mathcal{k}\right ] + 3\left [ \mathcal{k}^{2}\right ] ^{2}-6\left [ \mathcal{k}% ^{4}\right ] . \notag\end{aligned}\ ] ] here , we want to study a particular model of nonlinear electrodynamics called bi theory which has attracted lots of attentions due to its relation to effective string actions . the function @xmath18 for bi theory
is given as @xmath19where @xmath20 is the bi parameter , the maxwell invariant is @xmath21 in which @xmath22 is the electromagnetic field tensor and @xmath23 is the gauge potential .
variation of the action ( [ action ] ) with respect to the metric tensor @xmath24 and the faraday tensor @xmath25 , leads to @xmath26@xmath27where @xmath28 is the en tensor and @xmath29 is the massive term with the following form @xmath30
in this section , we obtain static nonlinearly charged black holes in context of massive gravity with ads asymptotes . for this purpose we consider the metric of @xmath8-dimensional spacetime in the following form @xmath31where @xmath32 is a @xmath33 dimension line element for an euclidian space with constant curvature @xmath34 and volume @xmath35 . we should note that the constant @xmath36 , which indicates that the boundary of @xmath37 and @xmath38 , can be a negative ( hyperbolic ) , zero ( flat ) or positive ( elliptic ) constant curvature hypersurface .
we consider the ansatz metric @xcite @xmath39where @xmath7 is a positive constant . using the metric ansatz ( [ f11 ] ) , @xmath40 s are in the following forms @xcite @xmath41 in which @xmath42 . using the gauge potential ansatz @xmath43 in electromagnetic equation ( [ maxwell equation ] ) and considering the metric ( [ metric ] )
, we obtain @xmath44 in which @xmath45 is the following hypergeometric function @xmath46 , \left [ \frac{3d_{7/3}}{2d_{2}}\right ] , -\gamma \right ) , \label{h}\]]where @xmath47 and @xmath48 is an integration constant which is related to the electric charge .
also , the electromagnetic field tensor in @xmath8-dimensions is given by @xmath49 now , we are interested in obtaining the static black hole solutions
. one may use components of eq .
( [ field equation ] ) and obtain metric function @xmath50 .
we use the @xmath51 and @xmath52 components of the eq .
( [ field equation ] ) , which can be written as @xmath53 -2\lambda r^{4}\right .
\notag \\ & & \left .
-d_{2}d_{3}r^{2}f - d_{2}r^{3}f^{\prime } + 4\beta ^{2}r^{4}+d_{2}d_{3}r^{2}k\right\ } \sqrt{1-\left ( \frac{h^{\prime } } { \beta } % \right ) ^{2}}-4\beta ^{2}r^{4}=0 , \label{tteq } \\ & & \notag \\
e_{x_{1}x_{1 } } & = & d_{3}m^{2}c\left [ c_{1}r^{3}+d_{3}d_{4}c_{2}cr^{2}+d_{3}d_{4}d_{5}c_{3}c^{2}r+d_{3}d_{4}d_{5}d_{6}c_{4}c^{3}% \right ] -2\lambda r^{4 } \notag \\ & & -2d_{3}r^{3}f^{\prime } -d_{3}d_{4}r^{2}f - r^{4}f^{\prime \prime } + 4\beta ^{2}r^{4}-4\beta r^{4}\sqrt{\beta ^{2}-h^{\prime 2}}+d_{3}d_{4}r^{2}k=0 .
\label{x1x1eq}\end{aligned}\ ] ] we can obtain the metric function @xmath54 , by using the eqs .
( [ tteq ] ) and ( [ x1x1eq ] ) with the following form @xmath55where @xmath56 is an integration constant which is related to the total mass of the black hole .
it should be noted that , obtained metric function ( f(r ) ) , satisfy all components of the eq .
( [ field equation ] ) , simultaneously .
now , we are in a position to review the geometrical structure of this solution , briefly .
we first look for the essential singularity(ies ) .
the ricci scalar and the kretschmann scalar are @xmath57and so confirm that there is a curvature singularity at @xmath0 .
the ricci and kretschmann scalars are @xmath58 and @xmath59 at @xmath60 .
therefore , the asymptotic behavior of these solutions are ( a)ds for ( @xmath61 ) . on the other hand , in the absence of massive parameter ( @xmath62 ) , the solution ( [ f(r ) ] ) reduces to an @xmath8-dimensional asymptotically ads topological black hole with a negative , zero or positive constant curvature hypersurface in the following form@xmath63 in order to study the effects of the en - bi - massive gravity on metric function , we have plotted various diagrams ( figs .
[ figfr1 ] [ pen ] ) .
@xmath64 by considering specific values for the parameters , metric function has different behaviors .
depending on the choices of the parameters , en - bi - massive black holes can behave like reissner nordstrm black holes .
in other words , these black holes may have two horizons , one extreme horizon and without horizon ( naked singularity ) ( see fig . [ figfr1 ] for more details ) . on the other hand , by adjusting some of the parameters of en - bi - massive black holes , we encounter with interesting behaviors .
the solutions may have three or higher horizons ( figs .
[ figfr2 ] and [ pen ] ) .
the existence of three or higher horizons for black holes is due to the presence of massive gravity @xcite .
in addition to the significant effects of massive term , we should note that the nonlinearity parameter can affect on the number of horizons .
in addition , @xmath65 can change the type of singularity .
in other words , depending on the parameters , one can find a @xmath66 in which singularity is spacelike for @xmath67 , and it would be timelike for @xmath68 ( see bipapers2 for more details ) .
now , we give a brief discussion regarding carter
penrose diagrams . in order to study the conformal structure of the solutions
, one may use the conformal compactification method through plotting the carter
penrose ( conformal ) diagrams . as we mentioned before , depending on the value of nonlinearity parameter , @xmath65
, one may encounter with timelike or spacelike singularity .
penrose diagrams regarding to timelike singularity was discussed in @xcite ( see conformal diagrams in @xcite ) . here
we focus on special case in which singularity is spacelike ( @xmath67 ) . in other words ,
the singularity of this nonlinearly charged black holes behaves like uncharged schwarzschild solutions ( see fig .
[ pen ] ) .
this means that , although massive and nonlinearity parts of the metric function can change the type of singularity and horizon structure of black holes , they does not affect asymptotical behavior of the solutions .
drawing the carter
penrose diagrams , we find the causal structure of the solutions are asymptotically well behaved . @xmath69
in this section , we calculate the conserved and thermodynamic quantities of the static en - bi - massive black hole solutions in @xmath8-dimensions and then check the first law of thermodynamics . by using the definition of hawking temperature which is related to the definition of surface gravity on the outer horizon @xmath70 , one can find@xmath71 + \frac{\left ( 2\beta ^{2}-\lambda \right ) r_{+}}{2\pi d_{2}}+\frac{% d_{3}k}{4\pi r_{+}}-\frac{\beta ^{2}r_{+}}{\pi d_{2}}\sqrt{1+\gamma _ { + } } , \label{totaltt}\]]where @xmath72 . the electric charge , @xmath73 , can be found by calculating the flux of the electric field at infinity , yielding@xmath74 in order to obtain the entropy of the black holes , one can employ the area law of the black holes .
it is a matter of calculation to show that entropy has the following form @xcite @xmath75 it was shown that by using the hamiltonian approach or counterterm method , one can find the mass @xmath76 of the black hole for massive gravity as @xmath77 in which by evaluating metric function on the horizon ( @xmath78 ) , one can obtain @xmath79 + \frac{4d_{2}q^{2}}{d_{1}r_{+}^{d_{3}}}% \mathcal{h}_{+}\right .
\notag \\ & & \left .
+ \frac{cm^{2}r_{+}^{d_{5}}}{d_{2}}\left [ d_{2}d_{3}d_{4}c_{4}c^{3}+d_{2}d_{3}c_{3}c^{2}r_{+}+d_{2}c_{2}cr_{+}^{2}+c_{1}r_{+}^{2}% \right ] \right ) , \label{mass}\end{aligned}\]]where @xmath80 , \left [ \frac{3d_{7/3}}{2d_{2}}\right ] , -\gamma _ { + } \right ) $ ] .
it is notable that , @xmath81 is the electric potential , which is defined in the following form@xmath82 having conserved and thermodynamic quantities at hand , we are in a position to check the first law of thermodynamics for our solutions .
it is easy to show that by using thermodynamic quantities such as charge ( [ totalq ] ) , entropy ( [ totals ] ) and mass ( [ totalm ] ) , with the first law of black hole thermodynamics@xmath83we define the intensive parameters conjugate to @xmath84 and @xmath73 .
these quantities are the temperature and the electric potential @xmath85which are the same as those calculated for temperature ( [ totaltt ] ) and electric potential ( [ totalu ] ) .
here , we study the stability conditions and the effects of different factors on these conditions .
the stability conditions in canonical ensemble are based on the signature of the heat capacity .
the negativity of heat capacity represents unstable solutions which may lead to following results : unstable solutions may go under phase transition and acquire stable states .
this phase transition could happen whether when heat capacity meets a root(s ) or has a divergency .
therefore , the roots of regular numerator and denominator of the heat capacity are phase transition points . in the other scenario ,
the heat capacity is always negative .
this is known as non - physical case .
but there is a stronger condition which is originated from the temperature .
the positivity of the temperature represents physical solutions whereas its negativity is denoted as non - physical one .
therefore , in order to getting better picture and enriching the results of our study , we investigate both temperature and heat capacity , simultaneously .
the heat capacity is given by @xmath86 considering eqs .
( [ totaltt ] ) and ( [ totals ] ) , it is a matter of calculation to show that@xmath87 \notag \\ & & -\frac{4\beta ^{2}}{\pi d_{2}^{2}r_{+}^{d_{3}}}\left ( 1+\gamma _ { + } \right ) ^{\frac{3}{2}}+\frac{4d_{3}q^{2}\left ( d_{1}+\gamma _ { + } \right ) } { \pi d_{2}r_{+}^{3d-7}\sqrt{1+\gamma _ { + } } } .
\label{heat}\end{aligned}\ ] ] in order to study the effects of different parameters on stability conditions and temperature , we have plotted various diagrams ( figs . fig1 - [ fig5 ] ) .
@xmath88 @xmath89 @xmath90 @xmath91 @xmath92 interestingly , in the absence of the massive parameter ( fig .
[ fig1 ] right panel ) , temperature starts from @xmath93 and it is only an increasing function of horizon radius with a root .
therefore , we have two regions of physical and non - physical solutions . adding massive gravity
could modify the behavior of the temperature into an @xmath81 shape diagram starting from @xmath94 without any root .
the extremum is an increasing function of massive parameter ( fig .
[ fig1 ] right panel ) and dimensions ( fig .
fig5 right panel ) , whereas , it is a decreasing function of @xmath36 ( fig .
fig3 right panel ) and electric charge ( fig .
[ fig4 ] right panel ) .
the only exception for this behavior is for strong nonlinearity parameter .
interestingly , for large values of nonlinearity parameter , a massive - less like behavior is observed ( fig .
[ fig2 ] right panel ) . in other words ,
the temperature starts from @xmath93 but the effect of the massive could be seen through two extrema .
the root of temperature and smaller extremum are increasing functions of @xmath20 and larger extremum is a decreasing function of it .
another interesting property of temperature is the effect of dimensions . studying fig .
[ fig5 ] ( right panel ) shows that the temperature for every two sets of dimensions will coincide with each other . in other words , there are places in which despite differences in dimensions , two black holes with two different dimensions and same values for other parameters will have same temperature in a special @xmath70
. in absence of massive gravity , black holes could acquire temperature from zero to @xmath94 , whereas adding massive , will cause the black holes never acquire some temperature .
this effect is vanished in case of large nonlinearity parameter . in other words ,
the strength of nonlinearity parameter has opposing effects to massive s ones .
also , the @xmath81 shape diagram indicates that for every temperature that black holes can acquire two horizons exist except for the extremum .
therefore , considering hawking radiation , one is not able to recognize the size of these black holes by measuring their hawking radiation .
it is worthwhile to mention that extrema and root(s ) of temperature are phase transition points of heat capacity .
regarding stability , it is evident that in absence of the massive gravity , there exists a region of the instability which is located where the temperature is negative .
therefore , this is a non - physical solution ( fig .
[ fig1 ] left panel ) .
interestingly , by adding massive gravity , the non - physical region is vanished and heat capacity acquires divergence point without any root . before divergence point ,
the heat capacity is negative .
therefore , in this region black holes are unstable . in divergence point ,
black holes go under phase transition of smaller unstable to larger stable black holes .
the divergence point is an increasing function of massive parameter ( fig .
[ fig1 ] left panel ) and dimensions ( fig .
[ fig5 ] left panel ) , whereas , it is a decreasing function of @xmath36 ( fig .
[ fig3 ] left panel ) and electric charge ( fig .
[ fig4 ] left panel ) .
interestingly , in strong nonlinearity parameter , the mentioned behavior is modified . in this case
black holes enjoy one root and two divergence points . before root and between two divergencies ,
heat capacity is negative and between root and smaller divergence point and after larger divergence point , heat capacity is positive .
according to thermodynamical concept , systems go under phase transition to acquire stable states .
therefore , following phase transitions take place : non - physical unstable to physical stable ( in root ) , large unstable to smaller stable ( in smaller divergence point ) and smaller unstable to larger stable black holes ( in larger divergency ) .
root and smaller divergence point are increasing functions of @xmath65 ( fig .
[ fig2 ] left panel ) , whereas , larger divergency is a decreasing function of it ( fig .
[ fig2 ] middle panel ) .
it is worthwhile to mention that larger divergency is not highly sensitive to variation of nonlinearity parameter . comparing obtained results for heat capacity ( regarding phase transitions ) and the behavior of the temperature
, one can see that larger to smaller phase transition takes place at maximum ( compare fig .
[ fig3 ] left diagram with right ) and smaller to larger one happens at minimum ( compare fig .
[ fig3 ] middle diagram with right ) of temperature .
therefore , one is able to recognize the type and number of phase transition by only studying temperature s diagrams .
now , we are in a position to study the critical behavior of the system through phase diagrams . using the renewed interpretation of the cosmological constant as thermodynamical pressure , one can use following relation to rewrite thermodynamical relations of the solutions in spherical horizon vander @xmath95 which results into following conjugating thermodynamical variable corresponding to pressure @xcite @xmath96 due to existence of the pressure in obtained relation for total mass of the black holes , one can interpret the total mass as thermodynamical quantity known as enthalpy .
this interpretation will lead to the following relation for gibbs free energy @xcite
@xmath97 now by using eqs .
( [ totalm ] ) and ( [ p ] ) with the relations of volume and gibbs free energy ( eqs .
( [ v ] ) and ( [ g ] ) ) , one finds @xmath98and @xmath99 obtained relation for volume indicates that volume of the black holes is only a function of the topology of the solutions and independent of electrodynamics and gravitational extensions , directly . in order to obtain critical values , one can use @xmath5 diagrams .
in other words , by studying inflection point properties one can obtain critical values in which phase transitions may take place .
therefore , one can use @xmath100 considering obtained values for temperature ( [ totaltt ] ) and pressure ( [ p ] ) , one can obtain pressure as @xmath101 -\frac{d_{2}d_{3}}{16\pi r_{+}^{2}}+\frac{\beta ^{2}}{4\pi } \left ( \sqrt{1+\gamma _ { + } } -1\right ) .
\label{pp}\ ] ] now , by considering eq .
( [ infel ] ) with obtained relation for pressure ( [ pp ] ) , one can obtain two relations for finding critical quantities .
due to economical reasons , we will not present them . regarding the contribution of electromagnetic part ,
it is not possible to obtain critical horizon analytically , and therefore , we use numerical method . considering the variation of @xmath65 and massive parameter
, one can draw following tables [ cols="^,^,^,^,^",options="header " , ] + table ( @xmath2 ) : @xmath102 , @xmath103 , @xmath104 , @xmath105 and @xmath106 .
in addition , we plot following diagrams ( figs .
[ fig6 ] - [ fig9 ] ) to investigate that obtained values are the ones in which phase transition takes place or not .
@xmath107 @xmath108 @xmath109 @xmath110 the formation of swallow tail in @xmath111 diagrams for pressure smaller than critical pressure ( figs .
[ fig6 ] and [ fig8 ] right panels ) , subcritical isobars in @xmath112 diagrams for critical pressure ( figs .
[ fig6 ] and [ fig8 ] middle panel ) and isothermal diagrams in case of critical temperature in @xmath113 diagrams ( figs .
[ fig6 ] and [ fig8 ] left panels ) , show that obtained values are critical ones in which phase transition takes place .
it is evident that critical pressure ( fig .
[ fig7 ] left panel ) and temperature ( fig . [ fig7 ] middle panel ) are increasing functions of the massive parameter , whereas the critical horizon ( fig .
[ fig7 ] left and middle panels ) and universal ratio of @xmath114 are decreasing functions of this parameter .
it is worthwhile to mention that length of subcritical isobars ( which is known as phase transition region ) is a decreasing function of massive parameter ( fig .
[ fig7 ] middle panel ) . in opposite ,
the size of swallow tail and the energy of different phases are increasing functions of @xmath3 ( fig .
[ fig7 ] right panel ) .
interestingly , the effects of variation of nonlinearity parameter is opposite of massive parameter . in other words , critical pressure ( fig .
fig9 left panel ) , temperature ( fig .
[ fig9 ] middle panel ) and the size of swallow tail ( fig .
[ fig9 ] left panel ) are decreasing functions of @xmath115 , whereas , the critical horizon radius ( fig .
[ fig9 ] left and middle panels ) , length of subcritical isobars ( fig .
[ fig9 ] middle panel ) and universal ration of @xmath114 are increasing functions of nonlinearity parameter .
it should be pointed that the length of subcritical isobars affects single regions of different states which in our cases are smaller and larger black holes . in other words , increasing the length of subcritical isobars ( phase transition region ) decreases the single state regions . in this section , by cancelling the electric charge ( @xmath116 ) , we will study the critical behavior of the system .
previously , it was shown that schwarzschild black holes does not have any phase transition in context of extended phase space .
now , we are investigating the effects of massive gravity in case of en - massive gravity .
using obtained relation for calculating critical horizon radius in previous part and setting @xmath116 , one can find following relation for calculating critical horizon radius @xmath117 it is a matter of calculation to show that this relation has following roots which are critical horizon radii @xmath118 } \right ) } { % 2\left ( 1+c_{2}c^{2}m^{2}\right ) } .
\label{root}\ ] ] obtained relation shows that in absence of massive gravity , critical horizon radius will be zero which is not of our interest .
this result consistent with schwarzschild case .
now , for the simplicity , we consider the case of @xmath119 .
this leads into following critical horizon radius @xmath120 it is evident that for the cases of @xmath121 and @xmath122 with vanishing @xmath123 , the critical horizon radius will be zero .
therefore , there is no phase transition for these black holes .
interestingly for case of @xmath122 , the condition for having a positive critical horizon radius will be @xmath124 and @xmath125 . by employing
obtained value for critical horizon radius , one can find critical temperature and pressure in the following forms @xmath126 @xmath127 considering obtained values , one can show that following equality is hold @xmath128which shows that in this case , @xmath129 is a function of massive parameter and coefficients . using obtained critical values ( eqs .
( [ rcc ] ) , ( [ tcc ] ) and ( [ pcc ] ) ) with eqs .
( [ totaltt ] ) , ( [ p ] ) , ( [ g ] ) , ( [ pp ] ) and setting @xmath116 , we plot following diagrams for @xmath130 and @xmath131 dimensions ( figs .
[ q0 ] and [ q06 ] ) .
@xmath132 @xmath133 in ref .
@xcite , it was shown that in context of neutral gauss - bonnet black holes , no phase transition is observed in @xmath131-dimensions . here
, the extension of the massive gravity enables the black holes to enjoy the existence of second order phase transition in @xmath131-dimensions .
also , we observed that contrary to gauss - bonnet case , @xmath129 is a function of massive gravity .
in this section , we employ the geometrical concept for studying thermodynamical behavior of the obtained solutions . in order to do so ,
we employ hpem method . in this method ,
the thermodynamical phase space is constructed by considering mass of the black holes as thermodynamical potential . by doing so
, the components of the phase space will be extensive parameters such as electric charge , entropy and etc .
the general form of hpem metric is @xcite @xmath134where @xmath135 , @xmath136 and @xmath137 ( @xmath138 ) are extensive parameters which are components of phase space .
now , we will investigate whether the phase transition points that were obtained in section ( @xmath139 ) coincide with all divergencies of the ricci scalar of hpem metric .
for economical reasons , we only plot diagrams correspond to variation of massive and nonlinearity parameters .
to do so , we use eqs .
( [ totalq ] ) , ( [ totals ] ) and ( totalm ) with hpem metric ( eq . [ hpem ] ) .
this leads into following diagrams ( fig .
[ fig10 ] ) .
@xmath140 it is evident that employed metric has consistent results with what were found in case of heat capacity ( compare figs .
[ fig1 ] and [ fig2 ] with [ fig10 ] ) . in other words ,
the divergencies of the ricci scalar are matched with phase transition points of the heat capacity .
an interesting characteristic behavior of the diagrams is the different divergencies for different types of phase transition . in case of larger to smaller black holes phase transition , the divergency of the ricci scalar is toward @xmath141 ( compare fig . [ fig2 ] left panel with fig .
[ fig10 ] middle panel ) , whereas in case of smaller to larger phase transition , the divergency is toward @xmath142 ( compare fig .
[ fig2 ] middle panel with fig .
[ fig10 ] right panel ) .
this specific behavior enables us to recognize the type of phase transition independent of heat capacity .
next , we employ another geometrical metric for studying the critical behavior of the system in context of extended phase space . in this metric , due to consideration of the cosmological constant as thermodynamical pressure
, we have three extensive parameters ; electric charge , entropy and pressure .
in order to construct phase space we employ following metric pveinstein @xmath143 considering eqs .
( [ totalm ] ) , ( [ heat ] ) , ( [ p ] ) and ( [ gtdpv ] ) , we plot following diagram ( fig .
[ fig11 ] ) with respect to fig .
[ fig6 ] .
@xmath144 due to existence of a root for heat capacity , in all plotted diagrams , a divergency is observed ( fig .
[ fig11 ] left panel ) .
it is evident that for pressures smaller than critical pressure , system goes under two phase transitions with different horizon radii ( fig .
[ fig11 ] middle panel ) .
this is consistent with what was observed in studying @xmath112 diagrams of fig .
[ fig6 ] . on
the other hand , for critical pressure system goes under a phase transition .
the place of this divergency is exactly located at the critical horizon which is obtainable through @xmath112 diagrams of fig .
fig6 ( fig .
[ fig11 ] right panel ) . finally , for pressures larger than critical pressure no phase transition is observed and the behavior of ricci scalar will be what is plotted in fig .
[ fig11 ] ( left panel ) .
these results are consistent with ordinary thermodynamical concepts and indicates that these three pictures ( phase diagrams , heat capacity and geometrical thermodynamics ) are in agreement .
the final section of this paper is devoted to calculation of the critical pressure in extended phase space by using denominator of the heat capacity .
it was shown that one can calculate critical pressures that were obtained in section ( @xmath145 ) by using denominator of the heat capacity @xcite .
to do so , one should replace the cosmological constant in denominator of the heat capacity ( [ heat ] ) with its corresponding pressure ( [ p ] ) . then , solve the denominator of the heat capacity with respect to pressure .
this will lead into following relation @xmath146 obtained relation for pressure is different from what was obtained through use of temperature ( [ pp ] ) . in this relation , the maximum(s ) of pressure and its corresponding horizon radius are critical pressure and horizon radius in which phase transition takes place .
now , by using indicated values in table @xmath147 and eq .
( [ pnew ] ) , we plot following diagram ( fig .
fig12 ) . @xmath148 it is evident that obtained maximums are critical pressures in which phase transitions take place .
the thermodynamical concept that was mentioned in last section ( pressure being smaller than critical pressure leads to existence of two phase transitions and for pressures larger than critical pressure no divergency is observed ) is also hold in case of this approach .
in other words , this approach is an additional method for studying critical behavior of the system and the results of this approach is consistent with gts , heat capacity and extended phase space .
in this paper , we have considered en - massive gravity in presence of bi nonlinear electromagnetic field .
it was shown that considering this configuration leads to modification of the number and place of horizons that black holes can acquire .
in other words , cases of multiple horizons were observed with different phenomenologies .
next , conserved and thermodynamical quantities were obtained and it was shown that first law of thermodynamics hold for these black holes .
next , we studied the thermodynamical behavior of the system .
it was shown that temperature in the presence and absence of massive gravity presents different behaviors . adding massive put limitations on values that temperature
can acquire , while , there was no limitation for temperature in the absence of it .
interestingly , this behavior was modified in the presence of strong nonlinearity parameter . in strong nonlinearity parameter ,
the behavior of temperature returned to a massive - less like behavior but the effects of the massive were observed in existences of extrema .
it was also seen , that in case of different dimensions , for each pair of dimensions , one can find a point in which temperature for both dimensions are equal .
regarding the stability , it was seen that in the presence of massive gravity , black holes enjoy one phase transition of the smaller unstable to larger stable .
the phase transition was related to the divergency of heat capacity .
then again , in strong nonlinearity parameter , this behavior was modified . in this case ,
black holes had three phase transitions of smaller non - physical unstable to larger physical stable ( in place of root ) , larger unstable to smaller stable ( in place of smaller divergency ) and smaller unstable to larger stable ( in place of larger divergency ) .
clearly , one can conclude that nonlinear electromagnetic field has an opposing effect comparing to massive gravity .
strong nonlinearity parameter modifies the effects of the massive gravity and return the system to the massive - less like behavior , although the effects of massive still observed through extrema .
it was pointed out that at maximums of the temperature , larger unstable to smaller stable and at minimums , smaller unstable to larger stable phase transitions take place .
therefore , studying temperature provides an independent picture for studying phase transitions and stability of the solutions .
next , we extended phase space by considering cosmological constant as thermodynamical variable known as pressure . it was shown that volume of the black holes is independent of generalization of the electromagnetic field and extension of the massive gravity .
obtained values where critical points in which phase transitions took place .
it was shown that the effects of variation of nonlinearity parameter was opposite of the massive parameter .
in other words , these two factors put restrictions on each others effects .
interestingly , in ref .
@xcite variation of massive gravity highly modified the critical temperature and pressure . in case of obtained solutions in this paper ,
the modification was not as considerable as what was observed in case of gauss - bonnet - maxwell - massive black holes .
this shows that generalization of electromagnetic field puts stronger restrictions on the effects of the massive gravity . in other words , in order to have stronger control over contributions of the massive gravity one should increase the nonlinearity of the electromagnetic sector .
in addition , a study in context of neutral solutions was conducted .
it was shown that due to contribution of the massive gravity , the chargeless solutions of this gravity also enjoy the existence of phase transition .
in other words , black holes in en - massive gravity go under phase transitions in extended phase space .
also , it was shown that ratio of @xmath149 was a function of massive gravity .
it was also pointed out that in @xmath150-dimensions , contrary to case of gauss - bonnet black holes , these black holes enjoy second order phase transition .
in addition , it was shown that in case of @xmath121 , no phase transition for massive black holes is observed .
next , geometrical approach was used for studying critical behavior of the system in context of heat capacity and extended phase space .
it was shown that employed metrics for both cases have consistent results and follow the concepts of ordinary thermodynamics .
the characteristic behavior of divergencies in ricci scalar of the geometrical thermodynamical metrics , enabled us to recognize the type of phase transition ( smaller to larger or larger to smaller ) .
finally , another method which was based on denominator of the heat capacity was used to calculate critical pressure and horizon radius .
it was shown that this method has consistent results with extended phase space and follow the concepts of ordinary thermodynamics .
in other words , this method provides an independent approach for investigating critical behavior of the system . due to generalization of born - infeld for electromagnetic sector of the solutions
, it will be worthwhile to study the effects of this generalization on conductivity of these black holes and their corresponding superconductors phase transition .
specially , it will be interesting to see how this generalization will affect the interpretation of graviton as lattice and the drude like behavior .
also , it will be worthwhile to study the metal - insulator transition in context of these solutions . | in this paper , we study massive gravity in the presence of born - infeld nonlinear electrodynamics .
first , we obtain metric function related to this gravity and investigate the geometry of the solutions and find that there is an essential singularity at the origin ( @xmath0 ) .
it will be shown that due to contribution of the massive part , the number , types and places of horizons may be changed .
next , we calculate the conserved and thermodynamic quantities and check the validation of the first law of thermodynamics .
we also investigate thermal stability of these black holes in context of canonical ensemble .
it will be shown that number , type and place of phase transition points are functions of different parameters which lead to dependency of stability conditions to these parameters .
also , it will be shown how the behavior of temperature is modified due to extension of massive gravity and strong nonlinearity parameter .
next , critical behavior of the system in extended phase space by considering cosmological constant as pressure is investigated .
a study regarding neutral einstein - massive gravity in context of extended phase space is done .
geometrical approach is employed to study the thermodynamical behavior of the system in context of heat capacity and extended phase space .
it will be shown that gts , heat capacity and extended phase space have consistent results .
finally , critical behavior of the system is investigated through use of another method .
it will be pointed out that the results of this method is in agreement with other methods and follow the concepts of ordinary thermodynamics . |
we re assuming that you already know the basics of the qspn .
if not , read the qspn document first : @xcite .
the aim of netsukuku is to be a ( physical ) scalable mesh network , completely distributed and decentralised , anonymous and autonomous .
the software , which must be executed by every node of the net , has to be unobtrusive .
it has to use very few cpu and memory resources , in this way it will be possible to run it inside low - performance computers , like access points , embedded devices and old computers .
if this requirements are met , netsukuku can be easily used to build a worldwide distributed , anonymous and not controlled network , separated from the internet , without the support of any servers , isps or control authorities .
node : : we call _ node _ any computer that is hooked up to the netsukuku network .
rnode : : stands for remote node : given a node x , it is any other node directly linked to x , i.e. it s a neighbour of x. map : : a map is a file , kept by each node , which contains all the necessary information about the network , f.e .
routes and nodes status .
example : + a is the rnode of b. + b is the rnode of a and c. + c is the rnode of b.
a simple topology , which does nt impose any structure on the network , can be memorised with a simple map . in this map
, all the information regarding the nodes of the network have to be memorised
. surely , this kind of map can not be utilised by netsukuku , because it would require too much memory .
for example , even if we store just one route to reach one node and even if this route costs one byte , we would need 1 gb of memory for a network composed by @xmath2 nodes ( the current internet ) .
for this reason , it s necessary to structure the network in a convenient topology
. first of all we ll subdivide the network in groups of 256 nodes and we ll use the following definitions : gnode : : means group node . it is a group of nodes , i.e. a set of nodes .
each node of the network belongs to just one gnode .
+ a gnode contains a maximum of 256 nodes .
+ by writing @xmath3 we mean that the node @xmath4 belongs to the gnode @xmath5 .
bnode : : stands for border node .
it is a node which belongs to a gnode g , but that is also directly linked to at least one node of another gnode , i.e. some of its rnodes belongs to different gnodes than its .
+ by writing @xmath6 we mean that the bnode @xmath7 belongs to the gnode @xmath5 .
example : + ] @xmath8 , a is a node belonging to the gnode g , its rnode is b. + @xmath9 , b is a node belonging to the gnode @xmath10 , its rnode is a. + a is a bnode of g , while b is a bnode of @xmath10 .
we further subdivide the network topology in _ groups of 256 groups of nodes _ and we continue to name them as gnode .
+ at this point , we repeat recursively this subdivision process until we can group all the nodes of the network into a single gnode . doing so , we ve structured the network in @xmath11 levels ( from @xmath12 to @xmath4 ) .
+ in the base level ( level 0 ) , there are 256 single nodes . + in the first level ( level 1 ) , there are 256 normal gnodes .
each of them contains 256 single nodes .
+ in the second ( level 2 ) , 256 gnodes of level 1 forms a single _ group of groups of nodes_. + in the third ( level 3 ) , there are 256 groups of 256 groups of 256 groups of 256 nodes
. + continuing in this way , we arrive at the last level ( level @xmath4 ) , where there is a single group which contains the whole network .
+ the qspn algorithm is able to operate independently on any level , considering each gnode as a single node of level 0 .
for this reason , we can view the netsukuku topology as a fractal , where each level is composed by single nodes .
figure [ fig : fract_circle ] is an example of the fractal topology of netsukuku . in this topology
, each gnode contains four nodes , i.e. each group contains four elements .
the network is structured in 6 levels .
+ the red elements , are single nodes ( level 0 ) .
+ four nodes forms a single group of nodes ( level 1 ) .
+ a single bright green circle is a group of groups of nodes ( level 2 )
. + the dark green circles are groups of groups of groups of nodes ( level 3 ) .
+ the dark blue circle are groups of groups of groups of groups of nodes ( level 4 ) . + finally , the bright blue circle is the gnode which contains the whole network ( level 5 ) .
let s assign a numeric i d to each ( g)gnode , starting from the last level : 1 . in the last level ( @xmath4 ) there s only one giant gnode ,
thus we assign to it the i d `` 0 '' .
our global i d will be : @xmath13 2 .
in @xmath14 there are 256 gnodes , which belongs to the gnode 0 of level @xmath4 , thus we assign them the ids from @xmath12 to @xmath15 .
the global i d becomes : @xmath16 3 .
we repeat the step 2 recursively gaining an i d of this form : @xmath17 4 .
since the last level is always @xmath12 , we ll omit it and we ll consider only the first @xmath4 levels . in a network with a maximum of @xmath18 nodes ( the maximum allowed by the ipv4 ) , there would be five levels ( @xmath19 ) , where each gnode will be composed by 256 nodes
. therefore , the i d will be in the usual ip form : @xmath20 for example , a single node of level 0 of the network is : @xmath21 that said , each gnode of the network belongs to only one combination of gnodes of the various levels .
in our previous example we have : @xmath22 where each @xmath23 corresponds to the gnode i d of the level @xmath24 .
note that @xmath25 is the i d attributed to the single node , at level 0 .
the advantages of using a fractal topology are clear .
+ the node @xmath26 , instead of memorising information about each node of the whole network , will keep only that regarding the gnodes where it belongs to .
suppose the node @xmath26 had this i d : @xmath27 it will store in memory information regarding : 1 .
the 256 single nodes which belongs to its same gnode of level 1 , or in other words , the 256 nodes of the gnode @xmath28 , 2 .
the 256 gnodes gnodes which belongs to its same gnode of level 2 , of in other words , the 256 gnodes of the gnode @xmath29 , 3 . finally , the 256 gnodes which belongs to the gnode @xmath30 .
note that doing so , the node @xmath26 will be blind to all the other gnodes . for example , it wo nt know any information regarding the single nodes which belong to all the other gnodes of level 1 different from @xmath28 .
+ even with this lack of knowledge , as we ll see later , the node @xmath26 is still able to reach all the other nodes of the network . in conclusion
, @xmath26 only needs @xmath31 entries in its map , instead of @xmath18 . to clarify
the ideas suppose that each entry costs one byte . in the plain topology we needed @xmath32 , while in the fractal one we just need @xmath33 .
netsukuku is both compatible with ipv4 and ipv6 .
+ in ipv4 there are a maximum of @xmath18 ips , thus we have five levels @xmath19 .
+ in ipv6 there are a maximum of @xmath34 ips , thus @xmath35 . for simplicity
we divide the map of the node @xmath26 , in the _ internal map _ and in the _ external _ one .
the internal map contains information regarding the nodes belonging to @xmath28 .
the external map describes all the other levels of the topology .
the bnode map of the node @xmath26 contains the information regarding the bnodes of each level where @xmath26 belongs .
some examples to clarify the ideas : + suppose that @xmath36 * a bnode of level 0 is a single node linked with two nodes of two different gnodes of level 1 .
* the bnodes of level 0 , known by @xmath26 , are only that which belong to the gnode @xmath25 .
they are all the nodes of @xmath25 which are linked to at least a gnode different from @xmath28 .
the qspn , for each level , will build the routes necessary to connect each ( g)node to all the other ( g)nodes of the same level .
the routes will be saved in the maps of each node .
+ if the node @xmath37 wants to reach a node @xmath38 which belongs to different gnodes , f.e .
@xmath39 , it will add a cidr@xcite route in the routing table of the kernel : + _ all the packets whose destination is @xmath40 will be forwarded to the gateway @xmath41_. + we ll see later how the gateway @xmath41 is chosen .
in the qspn document @xcite , we ve seen how a tracer packet works in a network composed by single nodes , i.e. a gnode of level 0 . + we ll now study its way of working on higher levels . in the abstract sense a single node is an entity which : 1
. receives input from its links 2 .
stores it in its memory 3 .
computes it 4 . and sends the output of the computation over some of its links thus any other entity which performs the same operations can be thought as a single node . + a gnode @xmath5 can act as a single node too . 1 . a bnode @xmath42 , which belongs to @xmath5 , receives an input from its links .
we call @xmath42 the _ ingress _ ( b)node . 2 .
this input is flooded to all the nodes , of any level , of the gnode .
the nodes will memorize the information contained in the input .
+ note that the flood is not a tp .
the flooded pkt will be received only once by each node .
3 . a bnode @xmath43 , of the same gnode , which is different from @xmath42 , receives the flooded input and computes it .
we call @xmath43 the _ egress _ ( b)node . 4 .
the bnode @xmath43 sends the output of its computation to its external links , i.e. avoiding those links which connect it to nodes of @xmath5 or to the same gnode which sent the input to @xmath42 .
consider the network in figure [ fig : qspn_g3 ] .
+ , @xmath44 and @xmath45 @xmath46 , @xmath44 , @xmath47 are gnodes of level 1 .
@xmath48 and @xmath49 belong to @xmath46 .
@xmath50 and @xmath51 belong to @xmath44 .
@xmath52 belongs to @xmath47 .
+ suppose the node @xmath48 sends a tracer packet in level 1 . , which is a bnode , to simplify the example ] the following will happen : 1 .
the tp is flooded in @xmath46 .
2 . @xmath49 receives the tp and appends in it the i d of the gnode @xmath46 .
3 . @xmath49 sends the tp to @xmath50 .
@xmath50 receives it , updates its maps and floods it in @xmath44 .
all the node belonging to @xmath44 will receive the tp , updating their maps .
this same procedure is reiterated from step 2 , i.e. @xmath51 receives the tp , appends the i d of @xmath44 and so on until @xmath52 receives it .
@xmath52 , noticing that its gnode hasnt any links to other gnodes than @xmath44 , will bounce back the tp to @xmath51 and at the same time will flood @xmath47 .
the tp , with the same procedure , will return back to @xmath48 , completing the tp cycle .
in order to use the @xmath53 ( qspn v2 ) in high levels , we need to be sure that a tp , flooded inside a lower gnode , will reach once and only once all the nodes of the same gnode .
moreover , a good metric needs to be defined for the high levels : what is the rtt ( round - trip time ) and bandwidth capacity of the link between two gnodes , how is it measured ?
consider the situation in figure [ fig:3bnodes ] .
a and b are bnodes of the node @xmath46 , while @xmath54 is a bnode of the gnode @xmath44 .
suppose c sends a tp to b. b floods it inside @xmath46 , thus @xmath55 receives it .
at this point , @xmath55 sends the packet to all its external links , and thus to @xmath54 .
@xmath54 will send the packet again to @xmath56 and the cycle will continue .
+ this situation can be avoided if an ingress bnode , before forwarding a packet to a an external gnode @xmath57 , checks if the packet has already traversed @xmath57 itself .
if it has , the bnode wo nt forward it to @xmath57 .
we will refer to rem ( route efficiency measure ) , as a value characterising the quality of a route .
rem can be calculated in various ways , f.e . by taking in account the total rtt and the bw capacity of the route .
we denote the rem of a route @xmath58 as @xmath59 . from the point of view of the qspn v2 ,
the levels are `` flattened '' , because the propagation of ctps ignore the subdivisions of the network in gnodes and levels .
the structure of the network remains unchanged , subdvided fractally in gnodes , but the exploration of the ctp does nt .
a ctp does nt consider a gnode as a single , homogenous entity .
it propagates itself between the single nodes , tracing a precise route formed by single hops , until it is considered interesting . in this way
every node will have its own personal route to reach a specific gnode .
the memory limits are respected , because the ctp , once it jumped over the border of a gnode , deletes the specific details that were useful by the nodes of the gnode .
+ we ll know describe the rules of the flat levels .
+ suppose we want to explore the level @xmath60 ( we already know how to explore @xmath61 ) .
let @xmath62 be the gnode of level @xmath4 .
the generic rule of the qspn v2 in high levels is : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ when an ingress bnode @xmath7 of @xmath5 receives a ctp from a neighbouring bnode of @xmath63 , it appends in the packet its ip . the packet , called _ locked tracer packet _ is then flooded inside @xmath5 .
the ltp of level @xmath4 , is a packet that is propagated with the same rules of the ctp , i.e if it is nt interesting it s dropped , but that is nt updated while it explores the levels inferior to @xmath4 .
when the egress bnode @xmath64 , of @xmath5 receives the ltp , it changes it in a ctp , appends the i d of @xmath5 and forwards it to all its neighbouring gnodes . _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ detailed rules : 1 . when exploring the level @xmath4 , each inferior level must been already explored .
+ to accomplish this , it sufficies that a node , before sending / forwarding a ctp of level @xmath4 has already sent / forwarded at least a ctp of level @xmath14 .
if it hasnt , it will queue the ctp of level @xmath4 .
2 . suppose that the ingress bnode @xmath7 belonging to a gnode @xmath5 of level @xmath4 , has received a ctp of the same level , sent by the gnode @xmath10
. the following will happen : @xmath7 makes copy of the ctp , marking it as a ltp , appends in it its ip and floods it in @xmath5 .
if @xmath7 is also linked to a gnode different from @xmath10 , it will respect the rule [ ruleegressbnode ] for egress bnodes .
[ ruleegressbnode ] suppose that the egress bnode @xmath64 of @xmath5 receives the ltp sent by @xmath7 ( see above ) .
@xmath64 transforms the ltp in a ctp , deletes the ip added by @xmath7 and sets ( in the tp ) the rem of @xmath65 as : @xmath66 finally , it forwards the ctp to all its external rnodes .
4 . a node @xmath4 of @xmath5 , uses the received ltp to learn that @xmath7 is a bnode bordering on @xmath10 and to find out the route of level @xmath4 contained in the packet .
it considers the rem of the route to reach @xmath10 equal to @xmath67 in other words , it saves in its map that @xmath68 .
+ @xmath4 forwards the ltp to its rnodes without modifying it .
5 . the ltp respects the same rule of the qspn v2 :
it is forwarded by a node only if it is carries interesting information , otherwise it is dropped .
a node @xmath4 consider the ltp interesting if one of the following conditions is met : * the ltp contains the ip of a bnode @xmath7 , which was nt already known by @xmath4 . * the ltp contains a bnode already known by @xmath4 , but that borders on a new gnode @xmath69 : with this ltp , the node @xmath4 learns that it can reach @xmath69 by passing through the bnode . *
the ltp contains the information of the death of a bnode which was used by @xmath4 to reach some gnodes . *
the ltp contains a new gnode , which was previously unknown . *
the ltp contains the information of the death of a gnode . *
the ltp contains an improved route to reach a gnode .
+ note : the first three conditions are used by the nodes to build and update the bnode map .
[ rule6 ] recall the situation of rule [ ruleegressbnode ] .
+ when the value @xmath70 changes considerably , both @xmath7 and @xmath64 sends a ctp of level @xmath4 to their external nodes tells @xmath7 and @xmath64 of the rem change ] .
the ctp contains the new @xmath71 value .
every node which used @xmath5 as hop of some routes , will consider the ctp interesting .
+ this rule permits the update of high levels .
[ rule7 ] when a bnode @xmath7 ( of any level ) , which is a gnode of level @xmath72 , loses its external link to the gnode @xmath63 of level @xmath73 , when all its bnodes of level @xmath74 loses their link to @xmath63 .
this is a recursive definition . ]
loses one of its external links of level @xmath4 , a ctp is sent in the level @xmath14 .
the ctp contains the information regarding the lost link .
the ctp is forwarded by all the ( g)nodes that used the bnode @xmath7 in one of their saved routes to reach the lost link .
+ if @xmath7 was the unique bnode of its gnode @xmath5 bordering to gnode @xmath10 of level @xmath4 , then the link @xmath75 becomes broken . in this case a ctp of level @xmath4 is sent by @xmath10 and by the other bnodes of @xmath5 ] . +
a dying bnode is equivalent to a bnode which loses all its external links .
the only difference is that a ctp of level 0 is sent to inform the nodes of its gnode .
we ll now describe some rules by giving an example ( see figure [ figexq2 ] ) .
for simplicity , we ll use the ip notation : @xmath77 indicates the single node @xmath78 belonging to the gnode @xmath79 , belonging to the ggnode @xmath80 .
using the symbolic notation : @xmath81 .
we are also assuming that the routes are symmetric .
_ case 1 : a new gnode join . _
+ we are exploring the level 3 , which is formed by the gnodes @xmath82 .
the level 2 , of each gnode , has been already explored . *
the gnode @xmath83 has just created the link to the gnode @xmath78 .
the new physical link is maintained by the two single bnodes @xmath84 . *
the bnode @xmath85 creates a ctp of level 3 , appends in it the i d `` @xmath78 '' and sends it to @xmath86 .
* the bnode @xmath86 receives it . *
* it reflects back a new ctp to @xmath85 , appending in it the i d `` @xmath83 '' . * * it appends in the received ctp , its ip , marks it as a ltp and sends it to its gnode .
the ltp contains the path @xmath87 .
* * * the node @xmath88 , receiving the ltp , learns the following : @xmath89 is a bnode of level 2 that can be used to reach @xmath78 , and @xmath86 is a bnode of its own gnode which borders on @xmath78
. it also learns its own rem for the route to @xmath78 : @xmath90 @xmath88 continues to forward the ltp ( without modifying it ) . * * * the ltp reaches the node @xmath91 , node of @xmath92 .
@xmath91 , from the ltp learns that : @xmath89 is a bnode of level 2 that can be used to reach @xmath78 .
since the level @xmath93 has been already explored , @xmath91 already knows a route to reach @xmath89 . * * * the ltp having reached all the nodes of @xmath83 ceases to being forwarded . *
the bnode @xmath85 receives the reflected ctp sent by @xmath86 . *
* it appends in it its ip , marks it as a ltp and sends it to its gnode . * * * the ltp is propagated in @xmath78 ( in the same way of the previous ltp ) . * the bnode @xmath94 , which borders on @xmath79 , receives the ltp .
* * it modifies it in a ctp : it removes the ip @xmath85 and appends the i d @xmath78 .
the ctp is now @xmath95 .
finally , it sets the first rem of the ctp : @xmath96 the ctp is then sent to @xmath97 , bnode of @xmath79 .
* the bnode @xmath97 receives the ctp .
it sends a ltp in @xmath79 . *
the bnode @xmath98 , which borders on @xmath80 , receives the ltp .
it appends @xmath79 in it and sends it to @xmath80 .
the ctp is now : @xmath99 . *
a bnode of @xmath80 receives it , and reflect it to @xmath79 .
it sends also the ltp to @xmath80 . *
* the reflected ctp is finally received by the node @xmath86 , which sends the ltp to @xmath83 and reflects the ctp to @xmath85 .
the bnode @xmath85 , drops immediately the ctp , because it does nt contain any interesting information . *
the exploration is completed .
suppose that a tp has been flooded inside the gnode @xmath5 .
suppose also that the node @xmath3 receives two duplicate packets .
+ the @xmath53 instructs the node @xmath4 to keep forwarding only the interesting packets .
a packet , which is a perfect copy of a packet already received , is always uninteresting .
therefore , the node @xmath4 will drop the second copy of the received tp .
however , in this case the node @xmath4 should not consider the rem values saved in the tp , because two tps , which have crossed the same gnodes in the same order , might have different rem values .
+ in conclusion , whenever two or more tps , which have crossed the same high level route , are received by a node of @xmath5 , only one of them will be forwarded . in this way
, the successive nodes will receive only one copy of the same packet .
when a part of the network changes considerably , the maps of the involved levels must be updated .
every node has its own radar , which periodically sends a broadcast request to all its ( physical ) near nodes . by collecting the replies ,
the radar is able to determine the active rnodes of the node and the quality of its links . in level 0
, a ctp is sent every time a node joins the network , dies or every time the change of the quality of a link exceeds a predefined delta value .
the ctp is also is restricted to the gnode where it has been originated : a bnode wo nt send it to external nodes .
+ when a node joins , it wo nt send a ctp , only its rnodes will .
their ctp will be directed to the node . more formally : + if the node @xmath4 joins , that is if the node @xmath4 creates a link with the nodes @xmath100 , with @xmath60 , then a ctp will be sent by each rnode @xmath101 _ only _ to @xmath4 .
this means that if @xmath102 is connected to a node @xmath103 , when @xmath4 joins , @xmath102 will send the ctp only to @xmath4 and not to @xmath103 . +
this saves the propagation of two ctp in the following situation : @xmath104 suppose that the node @xmath56 joins . as a consequence of the rule we ve described ,
the node @xmath54 will send a ctp to @xmath56 and @xmath55 to @xmath56 , thus only two distinct ctp are generated .
their path is @xmath105 and the reverse .
if instead the node @xmath56 sends a ctp to @xmath54 and then @xmath55 to @xmath56 , two distinct ctp will explore the same verse : @xmath106 and @xmath105 , the same is for the reverse , thus , in total , four distinct ctp are propagated . + let @xmath107 be a link .
if its quality changes @xmath56 will send a ctp to @xmath55 and vice - versa .
+ when a node dies , all its rnodes will send a ctp to all their rnodes .
the ctp will include , as the first hop , the i d of the dead node , with a flag which indicates its death .
for example : @xmath108 if @xmath55 dies , then @xmath56 will send the following ctp to @xmath54 : @xmath109 + in order to prevent false positives , the nodes wo nt immediately send the ctp , but will wait a small amount of time . only if the change persists , they will send it . the dynamics for the update of high levels are mainly governed by rule [ rule6 ] and [ rule7 ] of flat levels ( see [ sec : flat ] ) .
in this paragraph , we ll specify some details .
+ rule [ rule6 ] says : `` when the value @xmath70 changes considerably , a new ctp is sent '' . by considerably
we mean that the difference of the current rem value with the previous one exceedes a predefined threshold .
more precisely : let @xmath110 be a route of level @xmath4 . if @xmath111 then the route has changed considerably .
@xmath112 is proportional to @xmath4 , in this way the nodes will be more sensible to minor changes of lower levels and they will consider only relevant changes of higher levels .
+ also in high levels , a bnode wo nt immediately send a new ctp , but it will delay it for a small amount of time .
a new node joins the network when it has been able to create at least one physical link to an active netsukuku node and when it has correctly executed the hooking procedure . in this paragraph , we describe loosely the hooking phase of a new node .
suppose that the node @xmath4 has established a physical link to at least one netsukuku node . in order to become an active netsukuku node ,
@xmath4 has to _ hook _ to its rnodes . during the hook
, @xmath4 will exchange vital information with its rnodes , it will choose its new ip and it will finally become part of a gnode .
the hook procedure is formed by these general steps : 1 .
the node @xmath4 chooses an ip in the range of @xmath113 .
it launches the first radar to see what its rnodes are .
if not a single node is found , it creates a new gnode and ends hooking phase .
3 . at this point
, @xmath4 asks to its nearer rnode the list of all the available free nodes presents inside the gnode of the rnode . + if the rnode rejects the request ( the gnode might be full ) , the node @xmath4 contacts another rnode .
@xmath4 chooses an ip from the received list of free nodes and sets it on its network interface .
@xmath4 will then download the external map from the same rnode .
looking at the external map , it will be able to determine if it has to create a new gnode .
if it has , it creates it and ends the hooking .
@xmath4 gets the internal and the bnode map from the same rnode .
@xmath4 launches a second radar and updates its routing table . 8 .
all the rnodes of @xmath4 send a ctp to update the maps .
when a node creates a new gnode , it will choose a random gnode i d , and thus a random ip .
+ suppose that two isolated gnodes get the same gnode i d .
when they will be linked , they ll enter in conflict .
+ the solution to this problem is to let each new gnode hook as a normal node would . you can find more information about this in the ntk_rfc 001@xcite .
* ` march 2007 ` * * description of the flat levels ( sec . [ sec : flat ] ) * * section [ sec : netdyn - level0 ] `` network dynamics - level 0 '' expanded . *
* section [ sec : netdyn - leveln ] `` network dynamics - level n '' updated : the references to the pre - flatlevels rem metric have been removed .
* ` october 2006 ` + initial release .
99 qspn document : http://netsukuku.freaknet.org/doc/main_doc/qspn.pdf[qspn.pdf ] netsukuku website : http://netsukuku.freaknet.org/ cidr routing : http://en.wikipedia.org/wiki/classless_inter-domain_routing[classless_inter-domain_routing in wikipedia ] ntk_rfc 001 : http://lab.dyne.org/ntk_gnodes_contiguity[gnode contiguity ] | in this document , we describe the fractal structure of the netsukuku topology .
moreover , we show how it is possible to use the qspn v2 on the high levels of the fractal .
this document is part of netsukuku .
+ copyright 2007 andrea lo pumo aka alpt @[email protected]@xmath1 .
all rights reserved .
this document is free ; you can redistribute it and/or modify it under the terms of the gnu general public license as published by the free software foundation ; either version 2 of the license , or ( at your option ) any later version .
this document is distributed in the hope that it will be useful , but without any warranty ; without even the implied warranty of merchantability or fitness for a particular purpose .
see the gnu general public license for more details .
you should have received a copy of the gnu general public license along with this document ; if not , write to the free software foundation , inc .
, 675 mass ave , cambridge , ma 02139 , usa . |
to explain the cosmic acceleration found by the observations of type ia supernovae ( sne ia ) in 1998 @xcite , we usually introduce dark energy with negative pressure which consists about 70% of the total energy density in the universe .
the cosmological constant with the constant equation of state @xmath5 is the simplest candidate of dark energy and it is consistent with current observational data , but its value is too small compared with that estimated from the vacuum energy of quantum field theory .
dynamical fields with scalar field such as quintessence @xcite , phantom @xcite , tachyon @xcite and k - essence @xcite were proposed for dark energy .
the scalar field rolls down a very shallow potential , so its equation of state @xmath2 will approach @xmath6 and it dominates the universe only recently . according to its evolution , the scalar field was classified to several categories .
if the scalar field stuck in a local minimum and starts to roll down to the true minimum until recently , the dark energy models are called thawing models @xcite . for thawing models , @xmath2 starts with the value -1 at early times and then deviates from -1 at the present time .
if the scalar field already rolls down its potential minimum before the onset of acceleration and it slows down as it comes to dominate the universe , the models are called freezing models @xcite . for the freezing models , @xmath2 is different from -1 at early times and approaches -1 at the present time . if @xmath7 initially and @xmath2 decreases as the scalar field rolls down its potential , the models are called cooling models @xcite . for the tracking models , the equation of state @xmath2 tracks below the background equation of state for a wide range of initial conditions ,
the tracker fields become dominate only recently and then it slows to crawl with @xmath8 as @xmath9 asymptotically .
both the freezing and tracking models are special cases of cooling models . due to
the hubble damping and the shallow potential , the variation of @xmath2 is bounded @xcite . for the thawing quintessence models ,
the bound was derived as @xmath10 @xcite , where @xmath11 . for the freezing quintessence models ,
the bound is given as @xmath12 @xcite .
for the tracker quintessence models , a lower bound @xmath13 was given in @xcite .
chiba derived a tighter bound @xmath14 for the tracker quintessence , @xmath15 for the tracker phantom and @xmath16 for the tracker k - essence @xcite .
for the tachyon field , the bound was derived as @xmath17 @xcite .
since the scalar field does not change much for most cases , the dynamics of the potential may be approximated by the linear expansion of @xmath18 as @xmath19 @xcite .
furthermore , for nearly flat potentials , a general approximate relation between @xmath2 and @xmath20 was found @xcite .
the existence of a relation between @xmath2 and @xmath20 was first pointed out for tracker field models in @xcite .
because of the relation , if we take the zeroth order approximation for @xmath20 , then we get approximate @xmath21 .
efstathiou approximated @xmath22 as @xmath23 in the redshift range @xmath24 @xcite , and the approximation was later generalized as the so called chevallier - polarski - linder ( cpl ) parametrization with @xmath25 @xcite . applying the generic relation between @xmath2 and @xmath20 to the cpl parametrization , analytical degenerated relations between @xmath0 and @xmath1 were derived in @xcite and the two - parameter model reduced to one - parameter sslcpl model . in this paper
, we discuss the approximate relation between @xmath2 and @xmath20 and the limits on @xmath3 for tachyon fields . in section
ii , we review the dynamics of tachyon fields .
general bounds along with the bounds for both the thawing and freezing models were derived in section iii . in section iv
, we derive the tighter bounds for tracker fields .
conclusions are drawn in section v.
we start with the dirac - born - infeld type of action for the tachyon field @xcite @xmath26 where @xmath27 , and the negative sign corresponds to the phantom type tachyon fields phenomenologically .
if we take the friedman - robertson - walker metric , then the pressure and energy density of the tachyon field @xmath28 are given by @xmath29 the equation of state for the tachyon field is @xmath30 so @xmath31 when @xmath32 and @xmath7 when @xmath33 . for convenience ,
we introduce the parameter @xmath34 for the tachyon field @xmath28 .
the equation of motion for the tachyon field is @xmath35 where @xmath36 .
the hubble parameter @xmath37 satisfies the friedmann equations : @xmath38 @xmath39,\ ] ] where @xmath40 is the energy density of matter with constant equation of state @xmath41 , @xmath42 , @xmath43 , @xmath44 , and we use the unit @xmath45 . for dust , @xmath46 and @xmath47 .
by using the following dimensionless variables : @xmath48 the evolution equations can be written as @xcite @xmath49 @xmath50,\ ] ] @xmath51 where prime denotes the derivative with respect to @xmath52 . in terms of the variables @xmath53 and @xmath54 , we have @xmath55 with the fractional energy density @xmath20 and the equation of state parameter @xmath56 , the autonomous eqs .
( [ auto1 ] ) , ( [ auto2 ] ) and ( [ auto3 ] ) can be expressed as @xmath57 @xmath58 @xmath59 where the positive sign corresponds to @xmath60 and the negative sign corresponds to @xmath61 . as will be discussed below
, we can always take the positive sign without loss of generality . combining the first two eqs .
( [ auto4 ] ) and ( [ auto5 ] ) , we get @xmath62 since we are interested in the dynamics which is close to a cosmological constant , so @xmath63 and the potential changes very slowly with a nearly constant slope @xmath64 . with these approximations , and take dust @xmath47 , we get @xmath65 the above equation obtained for the tachyon fields is the same as that obtained for the quintessence and phantom fields .
this shows that the dynamics is quite generic for general scalar fields , so it is hard to distinguish different dynamics that caused the cosmic acceleration .
the solution to the above eq .
( [ dgamadomga1 ] ) for general thawing models is @xcite @xmath66 ^ 2,\ ] ] and @xmath67^{-2}.\ ] ] substituting the solution to eq .
( [ auto5 ] ) with the approximation @xmath63 , @xmath68^{-1},\ ] ] into eq .
( [ dynasol1 ] ) , we get the evolution of the equation of state parameter @xmath69 for the tachyon dark energy . if we take further approximation , i.e. , we take taylor expansion of @xmath20 around @xmath70 from eq .
( [ wapprox1 ] ) , we get @xmath71 using eqs .
( [ dynasol1 ] ) , ( [ dynasol2 ] ) and ( [ omegaphieq1 ] ) , we get the sslcpl parametrization @xmath25 with @xcite @xmath72 } { \omega_{\phi 0}^{-1/2}-(\omega_{\phi 0}^{-1}-1)\tanh^{-1}(\sqrt{\omega_{\phi0}})}.\ ] ] we see that the degeneracy relation ( [ waapproxeq1 ] ) among @xmath0 , @xmath1 and @xmath73 is quite general , it holds for both canonical and tachyon scalar fields with @xmath74 and @xmath75 .
for @xmath75 , the scalar field rolls down the potential , @xmath76 , so @xmath77 and @xmath78 have the same sign .
therefore , the second term on the right hand side of eq .
( [ auto4 ] ) is always positive , and we can take the positive sign only .
therefore , @xmath79 . for @xmath75 , one has a lower bound on @xmath3 as @xmath80 for the phantom case @xmath31 , @xmath81 , so @xmath77 and @xmath78 have the opposite sign , and the second term on the right hand side of eq . ( [ auto4 ] )
is always negative .
therefore , for @xmath31 , one gets an upper bound @xmath82 on the other hand ,
the equation of motion for the tachyon field gives @xmath83 the scalar field does not change much during the matter domination and starts to change when it becomes dominant , so the acceleration of the scalar field is limited by the hubble damping as @xmath84 .
for the freezing model , we have @xmath85 .
therefore , we have more stringent bound for some special cases . for the thawing model with @xmath7
, we have @xmath86 in @xcite , the authors derived the bound @xmath17 for thawing tachyon field .
however , for thawing field , @xmath87 , so the lower bound should be @xmath87 .
the thawing solution ( [ waapproxeq1 ] ) for @xmath1 gives @xmath88 . since @xmath89 , so the bound on @xmath3 becomes @xmath90 which is consistent with the bound ( [ wblda ] ) .
for the thawing model with @xmath31 , we have @xmath91 the thawing solution ( [ waapproxeq1 ] ) gives @xmath92 which leads to the bound @xmath93 .
therefore the thawing solution ( [ waapproxeq1 ] ) is consistent with the bound ( [ wbldb ] ) . for the freezing model with @xmath7 , we have @xmath94 since we are interested in the region when @xmath2 is not far away from -1 , @xmath95 , so @xmath96 and this lower bound @xmath97 is more restrictive compared with the general bound ( [ w - exprr ] ) . for some cases ,
the model may stuck in the point with @xmath98 initially , then @xmath2 decreases to @xmath6 very fast and the lower bound ( [ wbldc ] ) can be violated , but the general bound ( [ w - exprr ] ) is still satisfied . for the freezing model with @xmath31 , we have @xmath99 because @xmath100 , so this upper bound @xmath97 is more restrictive compared with the general bound ( [ phantbld1 ] ) .
similar to the quintessence and phantom cases , we derive the tracker equation " @xcite for the tachyon field first .
combining eqs .
( [ auto4 ] ) , ( [ auto5 ] ) and ( [ auto6 ] ) , we obtain @xmath101}{[\gamma_{t } ' + 6\gamma_{t}(1-\gamma_{t})]^2 } + \frac{(1-\frac{7}{2}\gamma_{t})\gamma_{t}^{\prime 2}}{\gamma_{t}[\gamma_{t } ' + 6\gamma_{t}(1-\gamma_{t})]^2 } + \frac{3(1-\gamma_{t})(\gamma_b-\gamma_{t})(1-\omega_{\phi})}{\gamma_{t}'+6\gamma_{t}(1-\gamma_{t})}.\ ] ] the tracker equation " ( [ tracker ] ) is independent of @xmath102 , so it holds for both the phantom and non - phantom cases . to simplify the above tracker equation "
, we introduce the variable @xmath103 as @xmath104}=\ln{\left(\frac{1+w}{-\epsilon w}\right)}.\ ] ] so @xmath105 using the variable @xmath103 , the tracker equation " ( [ tracker ] ) becomes @xmath106 for the tracker solution , initially @xmath107 is negligible and @xmath2 is almost a constant . as the scalar field becomes dominant
, @xmath2 decreases and approaches toward -1 asymptotically , so it belongs to the class of freezing models .
however , because of the @xmath108 relation @xcite for the tracker solution and @xmath109 , @xmath0 is different from @xmath6 which makes the tracking model distinguishable from @xmath110cdm model . by taking @xmath111 and @xmath112 in eq .
( [ tracker1 ] ) , we get @xmath113 for the dust @xmath46 , we get the track solution @xmath114 so the nearly constant @xmath115 , @xmath116 for @xmath7 and @xmath117 for @xmath31 .
since @xmath118 will stop decreasing and then increases toward zero , it has a minimum . to find the minimum , we set @xmath119 in eq .
( [ tracker1 ] ) , then we get @xmath120 note that @xmath121 and @xmath122 , a lower bound on @xmath123 is @xmath124 because the right hand side of eq .
( [ x - minimum1 ] ) is an decreasing function of @xmath2 , and @xmath2 decreases from the tracker solution ( [ w - express1 ] ) , so we get a lower bound on @xmath125 by replacing @xmath2 with @xmath126 , @xmath127 in the last inequality , we notice that the function @xmath128 decreases as @xmath2 decreases . for the case of @xmath7
, we get a lower bound on @xmath3 , @xmath129 where the sound speed of the tachyon field @xmath130 .
the bound was also derived for general tracker k - essence with @xmath7 in @xcite . since @xmath131 , so the lower bound can be also written as @xmath132 the lower bound @xmath133 is more stringent than the bound ( [ wbldc ] ) for general freezing model with @xmath7 . for @xmath31 , we get an upper bound @xmath134 again for the case that @xmath2 is not far away from -1 , this upper bound is more restrictive than the bound ( [ wbldd ] ) for general freezing model with @xmath31 .
the general bounds ( [ w - exprr ] ) and ( [ phantbld1 ] ) , the thawing bounds ( [ wblda ] ) and ( [ wbldb ] ) , the freezing bounds ( [ wbldc ] ) and ( [ wbldd ] ) , and the tracker bounds ( [ w - limit ] ) and ( [ w - expq ] ) are shown in fig . [ bounds ]
. we also numerically solve eqs .
( [ auto4])-([auto6 ] ) for the potential @xmath135 to show explicitly that those bounds are satisfied . for different models .
we also plot the trajectories of different models with the potential @xmath135 by numerically solving eqs .
( [ auto4])-([auto6 ] ) , here @xmath136 can be either positive or negative .
the cyan line corresponds to the general bound , the red lines correspond to the thawing models , the magenta lines correspond to the freezing models , and the blue lines correspond to the tracking models.,scaledwidth=50.0% ] to see whether it is possible to distinguish different dynamical models from the above bounds we derived , we use the observational constraints on the flat cpl model @xcite . for cpl parametrization , @xmath137 . in fig .
[ w0acont ] , we show the marginalized @xmath138 and @xmath139 contours of @xmath0 and @xmath1 from the combined snls3 sne ia , _ planck _ ,
bao and @xmath140 data @xcite .
we also plot the bounds ( [ w - exprr ] ) , ( [ phantbld1 ] ) , ( [ wblda ] ) , ( [ wbldb ] ) , ( [ wbldc ] ) , ( [ wbldd ] ) , ( [ w - limit ] ) and ( [ w - expq ] ) in fig .
[ w0acont ] .
most of the contours are inside the general bounds ( [ w - exprr ] ) and ( [ phantbld1 ] ) , but only a small part of the contours satisfy the bounds ( [ wblda ] ) , ( [ wbldc ] ) and ( [ w - limit ] ) for @xmath7
. the observational constraints slightly favor @xmath141 or @xmath87 when @xmath7 , so the thawing model is slightly favored . and @xmath139 contours of @xmath0 and @xmath1 of the flat cpl model from observational data @xcite .
we also show the bounds we derived.,scaledwidth=60.0% ]
the generic relation ( [ dynasol1 ] ) between @xmath2 and @xmath20 holds for the qintessence , phantom and tachyon thawing fields .
based on the generic relation , we derived the sslcpl model with analytic relation ( [ waapproxeq1 ] ) for @xmath0 and @xmath1 in the context of tachyon field .
the same sslcpl parametrization holds for thawing quintessence , phantom , tachyons with @xmath7 and @xmath31 no matter what the form the potential takes .
this shows that the dynamics of scalar fields become indistinguishable when @xmath2 does not deviate too far away from -1 , and makes the models less distinguishable from @xmath110cdm model . for the tachyon fields
, we also derived the general bounds ( [ w - exprr ] ) and ( [ phantbld1 ] ) , the bounds ( [ wblda ] ) and ( [ wbldb ] ) for the thawing models , and the bounds ( [ wbldc ] ) and ( [ wbldd ] ) for the freezing models . starting from the tracker equation " ( [ tracker1 ] ) we derived for tachyon fields , tighter bounds ( [ w - limit ] ) and ( [ w - expq ] ) were given for the tracking models .
the typical bound on @xmath3 is @xmath142 and the tightest bound is around @xmath143 for the tracker fields . if @xmath2 is measured to be @xmath6 within 20% error , then we require @xmath144 to detect tracking models .
the current observational constraints on the flat cpl model are @xmath145 and @xmath146 @xcite .
so @xmath0 is constrained to be @xmath147 by 10% and @xmath148 by 20% , but the uncertainties on @xmath1 are @xmath149 and @xmath150 . therefore , the current data can not distinguish different classes of models .
we need the data to constrain @xmath1 with the error bar better than @xmath151 , i.e. , we need to improve the constraint on @xmath1 by up to 50% .
this work was partially supported by the national basic science program ( project 973 ) of china under grant no .
2010cb833004 , the nnsf project of china under grant nos .
10935013 and 11175270 , the program for new century excellent talents in university , the fundamental research funds for the central universities and cq cmec under grant no . kj110523 . | we get the same degeneracy relation between @xmath0 and @xmath1 for the tachyon fields as for quintessence and phantom fields .
our results show that the dynamics of scalar fields with different origins becomes indistinguishable when the equation of state parameter @xmath2 does not deviate too far away from -1 , and the time variation @xmath3 satisfies the same bound for the same class of models . for the tachyon fields ,
a limit on @xmath3 exists due to the hubble damping and we derived the generic bounds on @xmath3 for different classes of models .
we may distinguish different models in the phase plane of @xmath4 .
the current constraints on @xmath2 and @xmath3 are consistent with all classes of models .
we need to improve the constraint on @xmath3 by 50% to distinguish different models . |
to understand the inner workings of gamma - ray burst ( grb ) central engines , it is necessary to constrain their true energy release . if a redshift is known , the isotropic energy release in gamma - rays , @xmath4 , is a readily measurable quantity .
the `` energy crisis '' brought on by implied energy releases of @xmath5 erg ( e.g. , * ? ? ?
* ; * ? ? ?
* ) was resolved when it was realized that grb blast waves are collimated with opening angles @xmath6 of 2 - 30@xmath7 @xcite .
thus the beaming - corrected gamma - ray energy release , @xmath8 , is smaller than @xmath9 by a factor @xmath10 .
whatever energy is not released in the prompt emission powers a relativistic blast wave that plows into the circumburst medium ( cbm ) .
measuring the kinetic energy of this outflow , @xmath11 , is challenging because only a fraction of this energy is radiated .
several methods have been used in this endeavor ( see * ? ? ?
* ) . the simplest method , using only the x - ray afterglow light curve ,
was suggested by @xcite , @xcite , and @xcite .
provided that the x - rays are predominately synchrotron and the electrons radiate in the cooling regime , this method yields the energy per unit solid angle and the fraction of the shock energy carried by electrons @xmath12 . for those afterglows with high quality multiwavelength data sets , a fit of the data can be made using models that describe the dynamics of jet / circumburst interaction and that calculate the expected synchrotron and inverse compton emission ( e.g. , * ? ? ? * ; * ? ? ?
* ) . finally , for bright , long - lived afterglows , @xmath11 can be measured more robustly using the late - time radio light curve
. months or in some cases years after the burst , the blast wave becomes sub - relativistic ( i.e. sedov self - similar evolution ) and the outflow is expected to be quasi - spherical ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
early studies , based on these methods , suggested that the energy release of long - duration grbs lies within a narrow range with a mean of @xmath13 erg ( e.g. * ? ? ? * ; * ? ? ? * ) .
there are several problems with this simple picture , however , and the true situation is likely more complicated .
for example , energy input into the blast wave now appears to extend beyond the prompt emission and into the afterglow phase .
this could be due either to a long - lived central engine or slow - moving ejecta that take a long time to reach the shock front @xcite .
the simple pictures of collimated explosions has been challenged by _ swift _ detected grbs .
kocevski & butler ( 2007 ) give an excellent summary of the difficulties and their implications for collimation - corrected energy release @xmath14 .
if ( for whatever reason ) the _ swift _ sample of bursts has no jet breaks then we are faced with a growing number of grbs that ( by virtue of inferred isotropic energy release ) are hyper - energetic i.e. well in excess of the canonical 10@xmath15 erg ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
if accepted , these events would provide stringent tests of existing progenitor models . in this paper
, we present multiwavelength observations of grb070125 .
its bright afterglow has allowed us to follow the grb until day 350 and obtain the most extensive radio data in the _ swift _ era , coupled with well - sampled x - ray and optical light curves indicative of a jet break .
together with the well characterized prompt emission extending beyond 1 mev @xcite , grb070125 is truly a rare event . in [ sec : observation ] we provide details of observations for grb070125 . in [ sec : result ] we describe our results , and find evidence for a jet break in the optical and x - ray data .
the radio spectra show evidence for evolution from an optically thick to an optically thin phase , while the radio light curves show short timescale variability that we ascribe to scintillation .
we carry out a straightforward analytic modeling and derive some physical parameters of the shock and the circumburst medium .
finally , we carry out a detailed model fit of the entire broadband dataset and summarize our results in [ sec : multiwave ] .
we discuss the energetics , environment , relevance of reverse shock emission , efficiency of the grb and the inverse compton scattering effects for grb 070125 in [ sec : discussion ] .
the main conclusions are listed in [ sec : conclusion ] .
in this section , we present multiwavelength observations of grb070125 .
we supplement our data set with the measurements reported in literature ( mostly notices from the gamma - ray burst circulars network ) .
grb070125 was discovered by the inter planetary network ( ipn ) of grb detectors at 07:20:45 ut 25 january 2007 @xcite .
mars odyssey ( hend and grs ) , suzaku ( wam ) , integral ( spi - acs ) , rhessi and konus - wind , all observed this intense , @xmath16 s long , event ( fig .
[ fig : high ] ) .
the burst was not in the field of view of _ swift_. four minutes later the burst fell into the burst alert telescope ( bat ) field of view @xcite .
the _ swift _ bat position was consistent with the position of the grb as triangulated by ipn .
the small @xmath17 error circle from the bat enabled follow - up observations to identify the bright x - ray ( [ sec : xray ] ) and optical ( [ sec : optical ] ) afterglow .
because of the high - energy ( up to 1 mev ) coverage provided by rhessi , konus - wind , and suzaku , tight constraints can be made on the prompt emission spectrum of grb070125 . in particular ,
the gamma - ray fluence was much more accurately measured than a typical _ swift _ grb ( with coverage of only @xmath18kev ) .
a detailed analysis of the high - energy properties of grb070125 has been performed by @xcite , who found that the spectrum is well fit by a band model @xcite with a peak energy @xmath19 kev and the power - law photon indices ( as defined in band model ) of @xmath20 and @xmath21 below and above the peak energy , respectively .
the resulting fluence in the 20 kev 10 mev band was @xmath22 erg @xmath23 .
the x - ray telescope ( xrt ; @xcite ) on board _ swift _ began observing the field of grb 070125 46 ks after the burst trigger @xcite .
due to the relatively low source flux , observations were conducted exclusively in photon counting ( pc ) mode , and so the effects of photon pile - up were negligible . during the first five orbits ,
the x - ray spectrum is well - fit by a power - law model with a photon index @xmath24 ( @xmath25 , where @xmath26 is the photon flux density , * ? ? ?
the inferred absorption is consistent with the galactic absorption value of @xmath27@xmath23 @xcite .
the xrt continued to monitor the x - ray afterglow of grb 070125 over the course of the next two weeks , until the source faded below the xrt sensitivity threshold .
motivated by the proposed lack of x - ray jet breaks observed in the _ swift _ era @xcite , we obtained director s discretionary time on the _ chandra x - ray observatory _ to observe the x - ray afterglow of grb 070125 at very late times .
we obtained a 30 ks exposure using the advanced ccd imaging spectrometer on board @xcite beginning at 21:28 ut 2007 march 5 ( @xmath28d after the burst trigger ) .
no source was detected at the position of the x - ray afterglow at this time .
formally , using a circular aperture with a one - arcsecond diameter , we detect @xmath29 photons in the energy range from @xmath30kev at the location of grb 070125(a @xmath31 upper limit ) .
the log of x - ray observations and the measured fluxes are summarized in table [ tab : xray ] .
the _ swift _ xrt light curve is obtained from the on - line repository @xcite .
we converted counts to flux using spectral parameters derived from the first five xrt orbits , ranging from @xmath32 ks .
there is no evidence for spectral evolution in the x - ray light curve , implying the derived conversion factor is applicable at all times @xcite .
we then converted the @xmath33 kev flux to a flux density ( @xmath34 , where @xmath35 ) at @xmath36kev ( @xmath37hz ) .
this value was so chosen that the flux was equally divided below and above this cutoff . in response to the ipn-_swift _
localization , we began observing the field of grb 070125 with the automated palomar 60-inch telescope ( p60 ; @xcite ) on the night of 2007 january 26 .
inspection of the first images revealed a bright , stationary source at @xmath38 , @xmath39 ( j2000.0 ) , not present in the sloan digital sky survey ( sdss ) images of the field @xcite .
this position was promptly reported as the afterglow of grb 070125@xcite , allowing several groups to obtain spectroscopy of the afterglow while still quite bright ( @xmath40mag ) .
the optical afterglow of grb070125 was observed by a variety of facilities worldwide , as it turned out to be one of the brightest optical afterglows ever detected for a grb ( updike _ et al .
_ , in prep . )
we continued to monitor the afterglow of grb 070125 with the p60 in the kron @xmath41 and sloan @xmath42 filters for the following four nights , until the afterglow faded below our sensitivity limit .
all p60 data were reduced using our custom software pipeline ( see @xcite for details ) using iraf routines .
in addition to our p60 monitoring , we obtained three epochs of late - time optical photometry with the gemini multi - object spectrograph and imager ( gmos ; @xcite ) mounted on the 8-m gemini north telescope .
all three epochs consisted of either single @xmath43 or @xmath42 exposures obtained as acquisition images for spectroscopic observations .
all gmos images were reduced using the iraf ` gemini ` package .
finally , we obtained a single epoch of simultaneous @xmath44- and @xmath41-band imaging on the night of 2007 february 16 with the low - resolution imaging spectrometer ( lris : @xcite ) mounted on the 10-m keck telescope .
individual images were bias - subtracted and flat - fielded using standard iraf routines .
co - addition was performed using swarp .
photometric calibration for all of our optical imaging was performed relative to the sdss , with empirical filter transformations from @xcite applied where necessary .
typical rms photometric uncertainties were @xmath45 mag in all of our images .
@xcite imaged the position of the grb 070125 afterglow with the lbc - blue ccd camera and 8.4-m sx mirror on the lbt on 2007 february 21.1 ( ut ) in r - band .
a faint source is detected at the position of the afterglow with brightness r=@xmath46 mag . since
the source may be contaminated by the host galaxy , this observation represents an upper - limit on the magnitude of the afterglow 26.8 days after the grb . to convert magnitudes to flux densities , we used zeropoint measurements from @xcite .
we have incorporated a modest amount of galactic extinction ( @xmath47 ; @xcite ) into these results .
the results of our optical monitoring of grb 070125 , together with measurements reported by other observatories via the gcn , are shown in table [ tab : opt ] .
we obtained director s discretionary time for observations in the 1.2-mm band ( 250 ghz ) at the max - planck millimeter bolometer array ( mambo ) , installed at the iram 30 m telescope on pico veleta , spain .
we used the mambo-2 version with 117 channels .
the bandwidth used was @xmath48 ghz half power .
our first observations took place on 2007 january 30 and we detected the afterglow of grb 070125 at a flux density of @xmath49 mjy .
we monitored the afterglow regularly until it had dropped below the instrumental sensitivity .
observations were also obtained at 95 ghz using the combined array for research in millimeter - wave astronomy ( carma ) , in good weather on 2007 february 5 , 9 , 12 , and 18 .
individual observations were between @xmath50 hours in length .
system temperatures ranged from 200 to 400 k , scaled to outside the atmosphere .
a single linear polarization was received in a bandwidth of @xmath51 ghz per sideband , depending on the antenna .
carma was in the c configuration , with baselines @xmath52 m. the bright quasar 0748 + 240 was observed as a phase calibrator at 20-minute intervals .
a fainter source nearer grb070125 , 0741 + 312 ( 2 distant ) , was observed for 30 seconds after each observation of the grb . the flux density of this source when self - calibrated was compared to the flux density when calibrated with 0748 + 240 .
the ratio of these measurements provides an estimate of the atmospheric decorrelation affecting the observations .
the measured flux densities of grb 070125 were corrected by this factor , which ranged from 9 to 14% .
the absolute flux density scale was derived from observations of uranus on february 2 and 19 , and transferred by reference to the quasar 3c84 , which was observed on those dates and near the beginning of each observation of grb 070125 .
images were made for each observation using the miriad software package ( @xcite ) .
the quoted uncertainty in the measurement is dominated by the sensitivity of the map , but also includes an estimate for the flux density scale uncertainty of 15% .
we detected the afterglow in the first three observations .
the flux density of the mean observation time 2007 february 5 , 0700 ut was 2.3 mjy the log and flux densities of the mambo and carma observations can be found in table [ tab : radio ] .
the earliest measurement of the radio flux density was taken from the westerbork synthesis radio telescope ( wsrt ) in 5 ghz band by @xcite ) , just 1.5 days after the burst .
a 2@xmath53 upper limit on the 5 ghz flux density of @xmath54@xmath55jy was obtained .
shortly thereafter , we triggered very large array ( vla ) observations of the field .
our first measurement at 8.46ghz , four days after the explosion , resulted in a strong detection ( @xmath56@xmath55jy ; * ? ? ?
encouraged by this detection , we triggered observations with the giant meterwave radio telescope ( gmrt ) in the 610 mhz band on 2007 january 31 .
however , we did not detect the afterglow with the gmrt to a 2@xmath53 limiting flux density of @xmath57@xmath55jy @xcite .
we continued the followup observations of grb 070125 with the vla from 2007 january 29th until 2007 february 2nd at 8.46 ghz and 4.86 ghz .
after 2007 february 5 , we undertook observations in the 1.46 , 4.86 , 8.46 , 14.94 and 22.5 ghz bands .
we followed the grb until day 342 since explosion .
each observation at a single frequency was from 30 minutes to one hour duration .
the bright radio quasar 3c48 ( 0137 + 331 ) was used as a flux calibrator .
we used phase calibrators 0745 + 317 and 0741 + 312 to track the instrumental and/or atmospheric gain and phase variations , as well as to monitor the quality and the sensitivity of the data .
a bandwidth of @xmath58 mhz was used for all the observations .
the data were analyzed using standard data reduction procedures within the astronomical image processing system ( aips ) . during initial observations ,
the vla was in `` c@xmath59d '' and `` d '' array configuration for early observations ; because of the short baselines , there were many sources in the field of view at lower frequencies . for such cases at 4.86 and 1.43 ghz , we did 3-d cleaning with several rounds of self - calibration inside aips .
on 2007 february 7 , 8 and 14 , we made longer observations for durations of @xmath60 , @xmath61 , and @xmath62 hours , respectively , in order to search for variability due to interstellar scintillation ( iss ) ( see [ sec : scintillation ] ) .
we combined every 20 minutes of the data and imaged the afterglow field of view to measure short timescale scintillations .
the results of our radio monitoring of the afterglow of grb070125 can be found in table [ tab : radio ] .
in this section , we carry out a simple analysis with minimum model assumptions to derive robust conclusions . we find strong evidence for a jet break in the combined fitting of the x - ray and the optical light curves ( [ sec : tjet ] ) . in [ sec : scintillation ] we use the radio scintillation data to constrain the size of the emitting region .
together , the jet break time and the inferred radius allow us to constrain the prompt energy release ( [ sec : energetics ] ) and the density of the circumburst medium ( [ sec : density ] ) . in the next section ( [ sec : multiwave ] ) , we undertake a detailed analysis utilizing the full machinery of afterglow models .
we performed a joint fit on the @xmath41 , @xmath42 , and x - ray light curves of grb 070125 using both a single power - law ( spl ) and a broken power - law ( bpl ) model .
the results of these two fits are shown in figure [ fig : jet ] and table [ tab : jet ] .
we also perform the spl and the bpl fits letting the x - ray and optical pre - break indices vary independently .
the results are consistent with the previous fits in which we constrained both the indices to be the same .
altogether , the broken power - law model , indicating a jet break at @xmath63d , is strongly favored ( @xmath64 for 97 d.o.f . vs. @xmath65 for 99 d.o.f . ) .
we note , however , that the distinction between the two models is significantly more pronounced in the optical ( @xmath66 for 52 d.o.f . vs. @xmath67 for 52 d.o.f .
; and @xmath68 for 25 d.o.f . vs. @xmath69 for 52 d.o.f .
) than the x - rays . as first noted by @xcite , without the late - time _
chandra _ data , the x - ray light curve can not distinguish between the spl and bpl models .
formally , we find the single power - law model is actually favored in the x - rays ( @xmath70 for 22 d.o.f . vs. @xmath71 for 20 d.o.f .
) . this could perhaps be due to the denser sampling at early times .
we also perform independent optical and x - ray fitting .
the optical fits are consistent with our joint fits . however
, when we fit the x - ray light curve independently of the optical , the jet break appears much later , @xmath72d ( fig .
[ fig : jet ] ) .
while not formally required by our simple analysis , we consider the possibility that this break may in fact be _
we discuss this further in [ sec : discussion ] . as can be seen in fig .
[ fig : refscint ] and table [ tab : radio ] , there are significant day - to - day deviations in the low frequency radio light curves .
these modulations are likely due to scintillation caused by interstellar propagation effects .
we obtained long - duration observations of the afterglow of grb070125 at 8.5 ghz on three separate occasions : february 7 ( 5.5 hour duration ) , february 8 ( 4 hour duration ) , and february 14 ( 8 hour duration ) .
the data were split into 20minute blocks and imaged in order to extract information on the fast variability of the source .
grb070125 exhibits flux density variations with a significance exceeding 99.8% on each of the three epochs ( see fig . [
fig : scint ] ) .
the reduced @xmath73 values for the hypothesis of no variability are 2.31 ( with 17 degrees of freedom ) , 4.42 ( 10 d.o.f . ) and 2.84 ( 23 d.o.f . ) for the three dates , respectively .
we used intensity structure functions to determine the variability timescale on each date . while the short baseline hindered a definitive determination
, we tentatively identify breaks with @xmath74% accuracy at @xmath75 and @xmath76 s , in the three structure functions .
the breaks are marginally significant in the data from 7 february and 8 february , but not so prominent in the 14 february images .
this can be explained by quenching of the scintillation as the source expands at late times .
these measurements are congruent with the timescales of the peaks and troughs apparent in the corresponding centimeter - wavelength lightcurves ( fig .
[ fig : refscint ] ) .
the interpretation of the variability depends on whether the scintillation occurs in the weak or strong regime .
strong scintillations require that the so - called fried parameter ( coherence length scale , @xmath77 ) be smaller than the fresnel size ( @xmath78 ) .
strong chromatic scintillations is possible only for sources smaller than @xmath79 .
these lead to the following two conditions @xcite : @xmath80 @xmath81 where sm is the scattering measure and @xmath82 is the effective distance to the scattering material , which is essentially the galactic ionized medium @xcite . to determine this
, we first estimate the scattering distance and the scattering measure using the formulation of @xcite . given the galactic coordinates of grb 070125 , @xmath83 ,
the expected sm is @xmath84m@xmath85kpc and the effective distance to the scattering material is @xmath86kpc . for these parameters , the critical frequency obtained from eq .
[ scint1 ] is 12.7 ghz .
these parameters also satisfy the condition of eq .
[ scint2 ] .
hence , the grb is likely to be in the strong scintillation regime .
however , given the paucity of our knowledge of the distribution of scattering material off the galactic plane , these estimates should be taken with caution .
strong scintillation can be diffractive as well refractive in nature .
diffractive scintillation of a point - like source is characterized by variations with a modulation index ( @xmath87 ) close to unity , where the modulation index is a measure of how much the modulated flux varies around its intrinsic flux density level .
the size limit for a source to exhibit diffractive scintillation ( @xmath88 ) is @xcite : @xmath89 for actual size of the source , @xmath90 , @xmath91 .
diffractive scintillation occurs when @xmath92 , which is indeed the case for grb 070125 using the parameters discussed above .
we now attempt to limit the size of the emitting region using the formulation of diffractive scintillation @xcite , but with more robust estimates of @xmath82 and sm .
the timescale of diffractive scintillation can be expressed as : @xmath93 where @xmath94 is the speed of the scattering material transverse to the line of sight .
the decorrelation bandwidth ( @xmath95 ) for diffractive scintillation is then @xcite : @xmath96 using the variability timescale of @xmath97s measured on 8 february , along with @xmath98kms@xmath99 , we infer a scattering measure of @xmath100 . in eq .
[ eq : nu ] we take the decorrelation bandwidth to be roughly half of the frequency of observation , i.e. @xmath101 , because at the boundary of strong and weak scintillation @xmath102 @xcite .
this yields a distance to the scattering screen of @xmath103 kpc .
knowing these two parameters , we can put an upper limit on the angular size of the emitting region of @xmath104 . here
the error in angular size corresponds to 20% error in the determination of @xmath105 . at @xmath106 @xcite , the angular size translates , @xmath107 and @xmath108 . ] to a linear size of @xmath109 cm or a radius of @xmath1 cm .
variability due to refractive scintillation is expected at late stages , even after the source has expanded sufficiently .
refractive scintillations are expected to be broadband in nature and start dominating once diffractive scintillations are quenched .
the source exhibited refractive scintillation during the two months subsequent to 2007 february 8 , as can be seen from the first 70days of the 8.46ghz data plotted in figure [ fig : refscint ] . in a standard grb afterglow model ,
the jet starts to spread sideways after the jet break @xcite . during this stage ,
the fireball size remain constant .
once the jet has become spherical , it reaches the non - relativistic regime . in this regime , the equations of motion follow self - similar sedov - taylor solutions .
thus , this estimate represents the size of the fireball in the post jet break regime until the expansion becomes sub - relativistic , which occurs between days 30 and 50 .
the isotropic gamma ray energy of a grb can be written as : @xmath110 where @xmath111 is the luminosity distance of the grb and @xmath112 is total fluence .
the fluence in the 20 kev 10 mev energy range is @xmath113 erg @xmath23 @xcite , yielding @xmath114 erg .
this is one of the largest isotropic energy releases ( top 1% ) ever reported for a grb @xcite . to determine the true prompt energy release
, this isotropic value needs to be corrected for collimation .
the combined fit to the optical and x - ray data gave a jet break at @xmath115 day ( [ sec : tjet ] ) .
the collimation correction depends on the density profile of the circumburst medium .
we derive corrections for a uniform density ( ism ; @xmath116 constant ) medium and a wind - like ( wind ; @xmath117 ) medium ( * ? ? ?
* and references therein ) .
the collimation angle for a radiative afterglow can be written as @xcite : @xmath118 where @xmath119 is the break in the optical light curve in days , and @xmath120 is the isotropic kinetic energy of the fireball in units of @xmath121 erg .
the term @xmath122 is the initial lorentz factor of the fireball .
similarly , the size of the spherical fireball can be expressed in the ism and wind media as @xcite : @xmath123 which after jet break in grb 070125 translates to : @xmath124 using the emission radius derived from scintillation studies ( @xmath125 ) , we find an opening angle of @xmath126 rad ( @xmath127 ) for the ism model and @xmath128 rad ( @xmath129 ) for the wind model .
the lorentz factor of the shocked ejecta at the time of jet break is @xmath130 in the ism model and @xmath131 in the wind model . from eq .
( [ eq : r_jet1 ] ) and ( [ eq : r_jet ] ) , the circumstellar density can be written in terms of the kinetic energy of the afterglow as @xmath132 @xmath133 and @xmath134 for the ism and the wind models , respectively .
let @xmath135 be the efficiency factor for converting the fireball energy into the radiation energy , i.e. @xmath136 .
for an empirical value of @xmath137 @xcite , the number density of grb 070125 is @xmath138 in the ism model and @xmath139 in the wind model .
this value is quite high , even for grbs , and indicates that the afterglow of grb070125 is expanding into a dense medium .
a natural consequence of a high circumburst density is a high synchrotron self - absorption frequency . to this end , we plot broadband radio spectra ( figure [ fig : radio_spectra ] ) . to improve the signal - to - noise ratio , we binned the spectra into 4 groups ( @xmath140 d , @xmath141 d , @xmath142 d , and @xmath143 d ) .
the division was initially done on the basis of similarly - looking spectra , but later in this section we justify it by demonstrating that the spectral evolution has a weak time dependence . as can be seen , there is a clear turnover in the spectra in the first three epochs .
moreover , there is some indication that the turnover frequency evolves to lower frequencies with time . at the last epoch ,
no turnover frequency is discernible and the spectrum is inverted .
this spectral behavior has been seen in many previously well - studied grb afterglows .
we interpret this behavior in terms of the evolution of the afterglow from an optically thick to optically thin phase , parametrized by a single unknown , the synchrotron self - absorption frequency @xmath144 .
we can measure the value of @xmath144 from the radio spectra using a very simple formulation given below : @xmath145 the above relation is a broken power - law with a break at @xmath144 .
we used the following smooth approximation of eq .
( [ eqn : nua ] ) : @xmath146^{-1}\ ] ] we fit this function to the first three radio spectra and obtain the following values of synchrotron self - absorption frequency : + @xmath147 ghz in the range @xmath148 days , + @xmath149 ghz in the range @xmath150 days , + @xmath151 ghz in the range @xmath152 days .
+ our approximation of a constant @xmath144 within each epoch is justified because of the slow evolution of @xmath144 .
the best fit time dependence to the values given above is @xmath153 , which agrees well with the time dependence predicted in the fireball model ( @xmath154 , @xcite ) .
the spectrum in the final epoch was well into the optically thin phase so we could not determine @xmath144 at this epoch .
we also plot the multiwaveband spectra on day 10.7 and day 23.4 ( fig .
[ fig : analytic ] ) , the two epochs at which we had observations in all the bands simultaneously .
this shows radio data points to be on the optically thick part of the spectra .
the afterglow peaks at 3-mm band and the optical and x - ray data fall in the optically thin regime .
this gives a rough estimate of the break frequencies , @xmath155 , corresponding to the minimum electron lorentz factor , and @xmath156 , the cooling frequency .
we also plot @xmath157 , a measure of energy , against @xmath158 .
the curve peaks at the cooling frequency @xmath156 ( fig .
[ fig : analytic ] ) .
in the previous section , we used simple analytical techniques to estimate four fundamental physical properties of the explosion : the opening angle ( @xmath159 0.25 radian ) , the size of the emitting region ( @xmath160 cm ) , the collimation - corrected prompt energy release ( @xmath161erg ) , and the circumburst density ( @xmath162@xmath133 ; @xmath163 ) . here
we combine _ all _ our observations of the grb 070125 afterglow in an attempt to derive a comprehensive model of the entire afterglow evolution .
our modeling software assumes a standard synchrotron forward shock formulation , with possible contributions from inverse compton ( ic ) emission and radiative losses also included .
this model also includes scintillation uncertainties , and hence gives realistic estimates of various parameters .
further details can be found in @xcite and @xcite . as in [ sec : tjet ] , we ignore all data before @xmath164d due to the possibility of late - time energy injection . based on our results in
[ sec : scintillation ] , we have incorporated scintillation effects into the radio regime .
we used an lmc - like extinction model for the optical data @xcite ; however , the extremely small host contribution to the extinction makes the effects of differing extinction laws negligible . results of the best fit wind model parameters are tabulated in table [ tab : models ] . in terms of @xmath73 and the model - fit statisticconstant , here
@xmath165 is the probability function and @xmath166 is standard deviation ] , the wind model does a slightly better job than the ism model .
however , the resulting best - fit parameters for the wind model are either unphysical ( i.e. @xmath167 , the fractional energy imparted to electrons in the shock , approaching unity ) or quite different from the values we derived in [ sec : result ] .
the extremely small isotropic afterglow kinetic energy ( @xmath168 ) compared to the @xmath0-ray isotropic energy is also troubling ( the same is true in the ism case , but to a lesser extent ) .
we also notice that the wind model is less stable to small changes in the parameter space .
we fit the wind model with and without the inverse compton effects .
the model without inverse compton effect gives even more unphysical values with many of the parameters asymptotically reaching very high values .
the magnetic field fraction required reaches 100% in this model . apparently allowing @xmath12 to be a free parameter is problematic .
microphysics evolution has been considered by @xcite .
it makes everything unconstrained .
we fixed @xmath12 to be 0.4 and obtained a good fit ( see table [ tab : models ] ) .
this exercise demonstrates that microphysical parameters are not constrained by our observations , at least for wind model .
the best - fit results for the ism model are also provided in table [ tab : models ] .
the ism model gives values of various parameters closer to the ones obtained from our simple analytical models .
the jet break time , density scale , and collimation angle are in good agreement with our previous results . the energy quoted in table [ tab : models ] is the isotropic blastwave energy at the time when @xmath169 , i.e. at the time of the transition from fast cooling to slow cooling ( @xmath170d in our model ) .
this isotropic kinetic energy is much smaller than the isotropic gamma - ray energy obtained from the konus - wind / rhessi fluence ( [ sec : energetics ] ) .
this may indicate that either there are high radiative losses at early times or the prompt emission is rather efficient with an extremely high value of @xmath135 .
we plot the results of our broadband modeling in figures [ fig : radio ] , [ fig : bbspec ] , and [ fig : xray ] .
it is difficult to differentiate between the wind and the ism models purely on the basis of these plots .
both models represent the optical data fairly well at early times ( fig .
[ fig : radio ] ) . at late times
( @xmath171d ) , both models over - predict the @xmath41-band flux , though with significantly less discrepancy in the constant density medium .
ic effects are negligible in optical bands . in the radio bands ,
none of the models fit particularly well , especially at early times ( @xmath172 days ) .
this is likely caused partly by the diffractive scintillation ( [ sec : scintillation ] ) .
ic scattering has no influence in this band .
the most puzzling behavior is revealed in the 4.8 ghz band .
an early detection at @xmath173d , both in our data and the wsrt @xcite , is followed by almost two weeks of non - detections .
summing all these non - detections , we can put very strict limits on the 4.8 ghz flux at this time : @xmath174 @xmath55jy .
the reason for this drop in flux is unclear , for it can not be due to scintillation .
we consider the possibility that the early detection is due to the reverse shock emission , in
[ sec : discussion ] . in figure
[ fig : bbspec ] , we plot the broadband spectra from radio to x - ray at various times of the afterglow evolution .
they are represented well with our models .
both the wind and ism environments seem to do an equally good job .
we plot the x - ray light curve of grb070125 in figure [ fig : xray ] .
this is the only band affected by ic emission . in the upper panel of figure [ fig : xray ] , we plot the model assuming only synchrotron radiation , while the lower panel incorporates ic scattering as well .
the wind model in a pure synchrotron fit has very unphysical parameters ( table [ tab : models ] ) ; hence we do not consider it further . in the ism model , the observations at @xmath175d
do not fit the data very well without any ic emission .
the fit improves significantly for both environments when we incorporate ic effects .
the ic component seems to raise the x - ray flux roughly almost the same time the jet break becomes visible in the optical bands .
this could well explain the jet break at a later stage in the x - rays .
we will discuss this further in [ sec : discussion ] .
with our comprehensive broadband models in hand , we now turn to some of the questions raised in the previous sections . here
we examine the ic scattering effect on the grb afterglow lightcurve in the x - ray band .
we adapt an approach in which we use only the synchrotron model for the grb afterglow and derive various parameters such as @xmath176 , density etc from the broadband data fitting . in this approach
, we force the broadband jet break to be fixed on the day of the optical jet break from our analytical fits i.e. on day 3.7 .
we then use these parameters to derive the light curve purely due to ic effect . for reasons noted
earlier we confine our discussion to the ism model . here
, we assume that the spectrum due to ic scattering has the same shape as that of the synchrotron model .
hence , ic spectrum in the x - ray band is @xmath177 here @xmath178 is ic peak flux , @xmath179 is ic cooling frequency and @xmath180 is ic peak frequency .
the best fit parameters from the synchrotron broadband model fit are : @xmath181 .
here @xmath182 is the transition time from the fast cooling to the slow cooling state .
@xmath120 is the isotropic - equivalent kinetic energ at @xmath183 .
ic scattering delays the cooling time by a fraction @xmath184 @xcite , i.e. the cooling time in presence of the ic effect is @xmath185d .
thus , for the time span of our observations , the afterglow remains in the fast cooling state and we will use formulation in this regime . using the formulation described in @xcite and @xcite , we estimate the time ( @xmath186 ) when ic scattering starts becoming important at 1.486 kev ( the x - ray light curve frequency ) . using above best fit parameter values
, we obtain @xmath187 day .
the ic light curve will satisfy the pre - jet break condition in the timerange of @xmath188 , i.e. between day @xmath189 .
the light curve will follow post - jet break formulation from day 3.7 onwards .
the flux density , size , lorentz factor and cooling frequencies derived on day 2.8 are : @xmath190 .
hence , the derived ic parameters on day 2.8 are : @xmath191 .
here @xmath192 and @xmath193 parameters are defined in @xcite .
the time dependences of various parameters in the pre - jet break epoch are : @xmath194 .
our frequency of observation ( @xmath195 hz ) satisfies @xmath196 condition at @xmath197d .
therefore , using eq .
[ eq : ic ] , the 1.5 kev light curve for ic scattering between day @xmath198 becomes @xmath199 from day 3.7 onwards , we use the post - jet break formulation and derive the ic light curve .
the time dependences of various parameters in the post - jet break regime are : @xmath200 . on the jet break day
, we find @xmath201 hz and @xmath202 hz , which still satisfies @xmath196 condition .
therefore , we derive the light curve from day 3.7 onwards using eq . [ eq : ic ] as @xmath203 the ic frequency @xmath204 reaches the x - ray observation frequency of @xmath205 hz on day 5.7 .
hence , the above light curve is valid until day 5.7 .
after day 5.7 , the ic flux density in the @xmath180 regime ( eq .
[ eq : ic ] ) , which gives the light curve in this regime as follows : + @xmath206 combining eqs .
( [ iclc1 ] ) , ( [ iclc2 ] ) and ( [ iclc3 ] ) , we plot the total ic light curve in fig . [
fig : ic ] .
the figure clearly shows that ic scattering flattens the light curve and delays the jet break to later time .
what happens after day @xmath207 is unknown due to the lack of x - ray detection of the afterglow .
while we have strong evidence that ic effects delay the x - ray jet break in grb070125 , we would like to know if this effect could be seen in other grbs as well . in the pre-_swift _ era , the x - ray data were not sufficiently well sampled to search for jet breaks , and so collimation corrections were almost exclusively calculated in the optical bands . in the _ swift _ era , with a plethora of well - sampled xrt light curves , we may be missing the jet break due to ic effects in many grbs .
the radio afterglow is not useful in determining the jet break , since grbs most likely scintillate in radio bands at such early times .
this makes the optical the unique bandpass in which the real jet break can be determined unambiguously .
the inverse compton effect is most prominent in a dense medium .
our radio observations have already shown that grb070125 resides in a very dense medium .
it has been shown by @xcite that ic is important in relativistic ejecta and even in non - relativistic ejecta in high density .
harrison et al . ( 2001 ) found good evidence for ic production of x - rays in grb 000926 and derived a high circumburst density , @xmath208 @xmath133 , comparable to the density we find for grb 070125 .
@xcite have shown that the late time flattening in the x - ray light curve of xrf050406 can be explained as an effect of ic scattering . for inverse compton scattering to play an important role , the electron energy density fraction ( @xmath12 )
must be larger than the magnetic energy density fraction ( @xmath209 ) .
recently , @xcite has discussed chromatic breaks occurring due to scattering of the forward - shock synchrotron emission by a relativistic outflow located behind the leading blast - wave .
this model may have x - ray jet breaks showing up at later times than the optical breaks .
however , this model requires a long - lived central engine , which may not be the case with most of the grbs . with grb070125 ,
we now have found three _ swift _ events with total energy release in excess of @xmath121erg : grb050904 @xcite and grb050820a @xcite .
while both grb050904 and grb050820a appear to have exploded in an dense circum - burst medium , the lack of a bright radio afterglow from grb050820a indicates a more typical environment .
moreover , with the exception of the total energy release , other parameters derived from broadband modeling are in line with previous studies of less energetic grbs @xcite .
it seems likely , therefore , that some factor intrinsic to the progenitor system is responsible for the large energy release . at first blush
, it seems surprising that _ swift _ has detected three of the most energetic grbs ever . with its increase high - energy sensitivity , _
swift_should preferentially select grbs at the low end of the fluence distribution .
we note , however , that a strong selection bias exists . as first noted by @xcite , in many _ swift _
x - ray light curves , the last xrt measurement is not sufficient to rule out a collimation - corrected prompt energy release of @xmath13erg . similarly , in the optical bandpass , @xcite have shown that at least some jet breaks occur at late times beyond the sensitivity of medium aperture facilities . while a detailed discussion of the relative rates of hyper - energetic events is still premature , it is clear at this point that , at the very least , the prompt @xmath0-ray energy distribution is significantly broader than previously believed @xcite . coupled with the recent controversy surrounding the validity of the many high - energy correlations ( e.g. @xcite ) , we believe the future utility of grbs as cosmological probes is significantly lessened .
even more importantly , however , hyper - energetic grbs have important consequences for progenitor models .
sustained engine activity has been seen now in many grbs @xcite .
this poses a problem for the collapsar model , as the duration of the central engine should not significantly exceed the accretion time scale onto the remnant black hole @xcite .
late - time engine activity is naturally accommodates by models in which the central object is a magnetar @xcite .
the existence of hyper - energetic grbs , however , is a direct and severe challenge to the magnetar model . with the current rate of hyper - energetic events ( @xmath211yr@xmath99 ) , coupled with the difficulty in measuring late jet breaks for more typical _ swift_events , future prospects look grim .
however , the impending launch of _ glast _ offers a new hope in the study of grb energetics .
much like blazars , those grbs capable of producing gev photons detectable by the large area telescope should be the most energetic and narrowly beamed events . together , synergistic _
glast _ and _ swift _ observations in the coming years should be able to shed light on the opening angles and energy release of a large sample of grbs .
our observations presented here , particularly the bright , self - absorbed radio afterglow , indicate grb070125 exploded in a dense circum - burst medium . in a separate work , however , @xcite have reported spectroscopic observations indicating an environment almost completely devoid of absorbing material .
@xcite furthermore find no evidence of an underlying host galaxy to deep ( @xmath212mag ) limits .
we briefly reiterate here the resolution of this apparent paradox .
the key to understanding the environment of grb070125 is to note that the broadband afterglow emission and the super - posed spectroscopic absorption features derive from distinct physical regions .
afterglow emission is caused by electrons in the circum - burst medium accelerated by the outgoing blastwave ( e.g. @xcite ) .
these electrons reside relatively close to the explosion center , typically at radii @xmath213pc .
we have strong evidence , however , that the absorption features seen super - posed on grb afterglow spectra derive from material at significantly larger distances from the explosion site : @xmath214kpc , or within the host galaxy ism . evidence in support of this large distance comes primarily from two lines of argument .
first , the presence of indicates a large distance from the explosion site , as the first ionization energy of mg falls below 1 ryd , and thus any mg near the grb would be ionized to at least @xcite .
second , @xcite have reported the detection of variability in the fine structure levels of from uv pumping for grb060418 . by modeling the variability over time , they were able to measure the grb - absorber distance : @xmath215kpc . while this explains the apparent density paradox , we are still left to explain how a dense circum - burst medium could be embedded in such a tenuous ism .
taking a clue from the lack of an underlying host detection , we suggested grb070125 may have exploded in a dense stellar cluster enriched by galaxy interactions .
star formation in such extreme environments can be seen in the local universe ( e.g. tadpole galaxy ; @xcite ) , and under our hierarchical picture of galaxy formation , such interactions should have occurred more frequently at @xmath216 .
the report of the detection of a faint source at the afterglow location at late times @xcite may call this interpretation into question , although it is unclear whether this emission is attributable to the fading afterglow or an underlying host ( or some combination thereof ) .
regardless , future high - resolution imaging ( i.e. _ hst _ ) seems worthwhile to pin down the environment of this truly unique event .
grb 070125 was not detected at @xmath217 d with the wsrt , and then was detected by both the vla and the wsrt around day 5 .
it remained below detection level for the next 15 days , before rebrightening on day 22 .
we explore the possibility that the flux from the grb at @xmath218d could be emission from a reverse shock .
we first consider the possibility that the non - detection was caused by modulations due to scintillation .
the modulation in flux density due to the refractive scintillation can decrease the flux density at the most up to @xmath219jy . to check for this possibility , we combined all the eight observations taken during the 15 day non - detection phase .
this vastly improved the signal - to - noise ratio .
the flux density we obtained at the grb position is @xmath220 @xmath55jy , much lower than scintillation can explain . according to the internal - external shock model for grbs , the prompt emission
is produced by internal shocks within a relativistic outflow , while the afterglow is produced by external shocks with the interstellar medium .
the reverse shock has a much lower temperature than that of the forward shock so it radiates at considerably lower frequencies . in this scenario using the ism model , the reverse shock emission peaks in the optical band at @xcite : @xmath221 for grb 070125 we calculate @xmath222 sec .
this corresponds to the time of peak in the radio band to be : @xmath223 here @xmath224 is the synchrotron self - absorption frequency , which can be written as @xcite : @xmath225^{2/(p+4)}.\ ] ] for values obtained from our multiwavelength analysis , we calculate @xmath226 hz and @xmath227 2.25 days . hence , the reverse shock emission peaks in the radio bands around @xmath228d .
we now estimate the peak radio flux density for the reverse shock . this can be written as : @xmath229 where @xmath230 is the peak optical flux density expressed as @xcite : @xmath231 for our best fit parameters , this value is @xmath211 mjy , which a gives the peak radio flux density to be @xmath232jy , which is two orders on magnitude than the observed one .
this shows that a reverse shock is not strong enough to explain the detection on day 5 .
we can not explain this strange behavior .
one of the major concerns for grb 070125 is the difference between the isotropic @xmath0-ray energy obtained from the high - energy fluence and the isotropic - equivalent blastwave energy obtained from our best fit model on the day when @xmath169 .
the isotropic - equivalent kinetic energy is an order of magnitude smaller than the isotropic @xmath0-ray energy ( @xmath233 erg , [ sec : energetics ] ) .
however , at very early times , the afterglow is in the fast cooling regime , where it undergoes significant loss of the energy because it is highly radiative .
the fireball may lose as much as 80% of its energy during this phase @xcite .
if we incorporate radiative corrections , we may derive a more accurate estimate of the actual isotropic blastwave energy .
from @xcite we find : @xmath234 for the ism model and @xmath235 for the wind model .
for the broadband modeling parameters , the isotropic kinetic energy one hour after the explosion is @xmath236 and @xmath237 erg for the ism and the wind models , respectively . here
, the energy for the wind model is rather unphysical .
the efficiency @xmath135 for the ism model is 0.67 . in terms of reduced @xmath73 for best fit ,
wind model is slightly better .
however , the wind model requires an electron energy density fraction close to 1 , which is very unlikely .
the best fit parameters in the wind model are rather unphysical , with less constrained boundaries . however , fixing the electron energy fraction to be 0.4 also gives reasonable fits .
evidence favoring the ism model comes from the fireball size estimation from the scintillation data .
however , the uncertainties in the diffraction scintillation time estimates may bring large uncertainties in the size estimates .
based on the energetics arguments stated above , the ism model is favored over the wind model , but we can by no means definitely dismiss the latter .
grb 070125 is one of the brightest grbs ever detected , both in terms of its prompt high - energy fluence and its optical and radio afterglows .
the isotropic equivalent energy for the grb is @xmath238 erg .
this is the most extensively followed grb in multiwavebands in the _ swift _ era .
the richness of the data allowed us to derive many important properties of the grb and place useful constraints on many parameters .
grb 070125 was one of the few grbs with sub - mm observations at various epochs .
our 95 ghz and 250 ghz observations with carma and iram , respectively , gave a robust determination of the peak flux density and @xmath155 , and constrained the power ( @xmath239 ) ( fig .
[ fig : analytic ] ) .
simultaneous fitting to optical and x - ray data favors a jet break at day 3.78 than the single powerlaw model .
the evidence for the jet break is indisputable in the optical @xmath41 and @xmath240 bands with pre - break and post - break slopes being 1.73 and 2.49 respectively .
however , the jet break is not very prominent in the x - ray band .
when we do the independent fit to optical and x - ray bands , the optical best fit is consistent with our joint fit .
however , the jet break in the x - ray band is shifted to day @xmath241 . using the best fit parameters of the model ,
we show that the inverse compton effects will dominate throughout our observations with pronounced effects in x - ray frequencies .
these effects delay the jet break in the x - ray band .
we had long observations of the grb at three epochs in the 8 ghz band .
these data gave evidence for diffractive scintillations , which gave an upper limit on the size of the fireball after the jet break , until the sedov - taylor phase started .
this estimate of the fireball size is consistent with the one obtained from the broadband modeling in a constant density medium .
we obtained synchrotron self absorption frequency estimates at various epochs from the vla radio data .
the evolution of the synchrotron self absorption frequency is @xmath242 , which is consistent with the one expected ( @xmath243 ) in the standard afterglow model .
synchrotron self absorption frequency estimates indicate that the grb afterglow is moving in a dense medium .
our model fits could not distinguish between the ism density profile and the wind - like density profile .
the @xmath73 fits were marginally better for the wind model but it needed an unphysically high electron energy fraction ( @xmath211 ) .
when we fixed the electron energy density fraction to 0.4 in the wind model , it did give decent fits and physical parameter values .
however , the parameter values in the ism model are more robust and change little with little change in the input values , unlike the wind model which is rather unstable . in both the ism and the wind models ,
the radiative efficiency of the grb is very high ( @xmath244 ) .
we suggest that ic scattering is a potential candidate in flattening the light curve and delay the jet break in other _ swift _ events and explain the absence of a jet break in x - ray light curves of some of the _ swift _ bursts .
ic effects are more prominent in high density medium .
frequent radio measurements are necessary to measure the circumburst density of the medium .
hence , in absence of good radio data , one can not determine the importance of ic scattering .
grb 070125 is unique because this has the richest radio data in _ swift _ era , having closely spaced x - ray light curve .
even though grb 070125 has rich multiwaveband data , we could not nail down some of the lingering issues , such as wind vs ism density profile .
one reason for this is that much of the evolution is in the jet break phase , when the ism and wind models have similar properties .
very dense samples of sensitive radio , x - ray , @xmath0-ray and optical data from the very beginning until the grb fades below the detection limit are needed . in the future , a combination of _ swift _ glast , alma , evla , and various optical telescopes will provide this opportunity .
pc thanks the vla staff for making radio observations , without which this work was not possible .
pc is a jansky fellow at national radio astronomy observatory .
the national radio astronomy observatory is a facility of the national science foundation operated under cooperative agreement by associated universities , inc .
we thank sarah yost for providing us the grb broadband modeling code and helping out in running the code .
pc used the fussy calculator ( http://fussy.googlecode.com ) for error propagation calculations and wish to thank its author , sanjay bhantagar .
rac was supported in part by nasa grant nng06gj33 g . , d. , matteson , j. , ford , l. , schaefer , b. , palmer , d. , teegarden , b. , cline , t. , briggs , m. , paciesas , w. , pendleton , g. , fishman , g. , kouveliotou , c. , meegan , c. , wilson , r. , & lestrade , p. 1993
, , 413 , 281 , e. c. , hurley , k. , palshin , v. , yamaoka , k. , bandstra , m. e. , boggs , s. e. , hong , s. , kodaka , n. , kozyrev , a. s. , litvak , m. l. , mitrofanov , i. g. , nakagawa , y. e. , ohno , m. , onda , k. , sanin , a. b. , sugita , s. , tashiro , m. , tretyakov , v. i. , urata , y. , & wigger , c. 2007 , arxiv e - prints , 710 , d. n. , hill , j. e. , nousek , j. a. , kennea , j. a. , wells , a. , osborne , j. p. , abbey , a. f. , beardmore , a. , mukerjee , k. , short , a. d. t. , chincarini , g. , campana , s. , citterio , o. , moretti , a. , pagani , c. , tagliaferri , g. , giommi , p. , capalbi , m. , tamburelli , f. , angelini , l. , cusumano , g. , bruninger , h. w. , burkert , w. , & hartner , g. d. 2005 , space science reviews , 120 , 165 , d. n. , romano , p. , falcone , a. , kobayashi , s. , zhang , b. , moretti , a. , obrien , p. t. , goad , m. r. , campana , s. , page , k. l. , angelini , l. , barthelmy , s. , beardmore , a. p. , capalbi , m. , chincarini , g. , cummings , j. , cusumano , g. , fox , d. , giommi , p. , hill , j. e. , kennea , j. a. , krimm , h. , mangano , v. , marshall , f. , mszros , p. , morris , d. c. , nousek , j. a. , osborne , j. p. , pagani , c. , perri , m. , tagliaferri , g. , wells , a. a. , woosley , s. , & gehrels , n. 2005 , science , 309 , 1833 , s. b. , fox , d. b. , moon , d .- s . , harrison , f. a. , kulkarni , s. r. , henning , j. r. , guzman , c. d. , bonati , m. , smith , r. m. , thicksten , r. p. , doyle , m. w. , petrie , h. l. , gal - yam , a. , soderberg , a. m. , anagnostou , n. l. , & laity , a. c. 2006 , , 118 , 1396 , s. b. , kasliwal , m. , harrison , f. a. , palshin , v. , frail , d. a. , cameron , p. b. , berger , e. , fox , d. b. , gal - yam , a. , kulkarni , s. r. , moon , d .- s . , nakar , e. , ofek , e. o. , penprase , b. e. , price , p. a. , sari , r. , schmidt , b. p. , soderberg , a. m. , aptekar , r. , frederiks , d. , golenetskii , s. , burrows , d. n. , chevalier , r. a. , gehrels , n. , mccarthy , p. j. , nousek , j. a. , & piran , t. 2006 , , 652 , 490 , x. , garnavich , p. m. , prieto , j. l. , stanek , k. z. , kochanek , c. s. , bechtold , j. , bouche , n. , buschkamp , p. , diolaiti , e. , fan , x. , giallongo , e. , gredel , r. , hill , j. m. , jiang , l. , mcclellend , c. , milne , p. , pedichini , f. , pogge , r. w. , ragazzoni , r. , rhoads , j. , smareglia , r. , thompson , d. , & wagner , r. m. 2007 , arxiv e - prints , 712 , p. a. , beardmore , a. p. , page , k. l. , tyler , l. g. , osborne , j. p. , goad , m. r. , obrien , p. t. , vetere , l. , racusin , j. , morris , d. , burrows , d. n. , capalbi , m. , perri , m. , gehrels , n. , & romano , p. 2007
, , 469 , 379 , d. a. , kulkarni , s. r. , sari , r. , djorgovski , s. g. , bloom , j. s. , galama , t. j. , reichart , d. e. , berger , e. , harrison , f. a. , price , p. a. , yost , s. a. , diercks , a. , goodrich , r. w. , & chaffee , f. 2001 , , 562 , l55 , a. s. , thorsett , s. e. , metzger , m. r. , sahu , k. c. , petro , l. , livio , m. , ferguson , h. , pian , e. , hogg , d. w. , galama , t. , gull , t. r. , kouveliotou , c. , macchetto , d. , van paradijs , j. , pedersen , h. , & smette , a. 1999 , 519 , l13 , g. p. , bautz , m. w. , ford , p. g. , nousek , j. a. , & ricker , jr .
, g. r. 2003 , in presented at the society of photo - optical instrumentation engineers ( spie ) conference , vol . 4851 , x - ray and gamma - ray telescopes and instruments for astronomy .
edited by joachim e. truemper , harvey d. tananbaum .
proceedings of the spie , volume 4851 , pp .
28 - 44 ( 2003 ) . ,
j. e. truemper & h. d. tananbaum , 2844 , p. , fan , x. , jiang , l. , dai , x. , kuhn , o. , bouche , n. , buschkamp , p. , smith , p. , milne , p. , bechtold , j. , stanek , k. z. , prieto , j. , wagner , r. m. , rhoads , j. , hill , j. , baruffolo , a. , desantis , c. , diolaiti , e. , dipaola , a. , farinato , j. , fontana , a. , gallozzi , s. , gasparo , f. , giallongo , e. , grazian , a. , pasian , f. , pedichini , f. , ragazzoni , r. , smareglia , r. , speziali , r. , testa , v. , & vernet , e. 2007 , grb coordinates network , 6165 , 1 , f. a. , bloom , j. s. , frail , d. a. , sari , r. , kulkarni , s. r. , djorgovski , s. g. , axelrod , t. , mould , j. , schmidt , b. p. , wieringa , m. h. , wark , r. m. , subrahmanyan , r. , mcconnell , d. , mccarthy , p. j. , schaefer , b. e. , mcmahon , r. g. , markze , r. o. , firth , e. , soffitta , p. , & amati , l. 1999 , 523 , l121 , f. a. , yost , s. a. , sari , r. , berger , e. , galama , t. j. , holtzman , j. , axelrod , t. , bloom , j. s. , chevalier , r. , costa , e. , diercks , a. , djorgovski , s. g. , frail , d. a. , frontera , f. , hurley , k. , kulkarni , s. r. , mccarthy , p. , piro , l. , pooley , g. g. , price , p. a. , reichart , d. , ricker , g. r. , shepherd , d. , schmidt , b. , walter , f. , & wheeler , c. 2001 , , 559 , 123 , i. , allington - smith , j. r. , beard , s. m. , crampton , d. , davies , r. l. , dickson , c. g. , ebbers , a. w. , fletcher , j. m. , jorgensen , i. , jean , i. , juneau , s. , murowinski , r. g. , nolan , r. , laidlaw , k. , leckie , b. , marshall , g. e. , purkins , t. , richardson , i. m. , roberts , s. c. , simons , d. a. , smith , m. j. , stilburn , j. r. , szeto , k. , tierney , c. , wolff , r. j. , & wooff , r. 2003 , in instrument design and performance for optical / infrared ground - based telescopes .
edited by iye , masanori ; moorwood , alan f. m. proceedings of the spie , volume 4841 , pp .
1645 - 1656 ( 2003 ) . ,
m. iye & a. f. m. moorwood , 16451656 , k. , cline , t. , mitrofanov , i. , kozyrev , a. , litvak , m. , sanin , a. , tretyakov , v. , parshukov , a. , boynton , w. , fellows , c. , harshman , k. , shinohara , c. , starr , r. , yamaoka , k. , ohno , m. , fukazawa , y. , takahashi , t. , tashiro , m. , terada , y. , murakami , t. , makishima , k. , smith , d. m. , lin , r. p. , mctiernan , j. , schwartz , r. , wigger , c. , hajdas , w. , zehnder , a. , von kienlin , a. , lichti , g. , rau , a. , cummings , j. , krimm , h. , barthelmy , s. , & gehrels , n. 2007 , grb coordinates network , 6024 , 1 , t. h. , polletta , m. , fournon , i. p. , stacey , g. , xu , k. , siana , b. , farrah , d. , berta , s. , hatziminaoglou , e. , rodighiero , g. , surace , j. , domingue , d. , shupe , d. , fang , f. , lonsdale , c. , oliver , s. , rowan - robinson , m. , smith , g. , babbedge , t. , gonzalez - solares , e. , masci , f. , franceschini , a. , & padgett , d. 2006 , , 131 , 261 , s. r. , djorgovski , s. g. , odewahn , s. c. , bloom , j. s. , gal , r. r. , koresko , c. d. , harrison , f. a. , lubin , l. m. , armus , l. , sari , r. , illingworth , g. d. , kelson , d. d. , magee , d. k. , dokkum , p. g. v. , frail , d. a. , mulchaey , j. s. , malkan , m. a. , mcclean , i. s. , teplitz , h. i. , koerner , d. , kirkpatrick , d. , kobayashi , n. , yadigaroglu , i .- a . , halpern , j. , piran , t. , goodrich , r. w. , chaffee , f. h. , feroci , m. , & costa , e. 1999 , , 398 , 389 , r. j. , teuben , p. j. , & wright , m. c. h. 1995 , in astronomical society of the pacific conference series , vol . 77 , astronomical data analysis software and systems iv , ed .
r. a. shaw , h. e. payne , & j. j. e. hayes , 433+ , p. m. , ledoux , c. , smette , a. , ellison , s. l. , jaunsen , a. o. , andersen , m. i. , fruchter , a. s. , fynbo , j. p. u. , hjorth , j. , kaufer , a. , mller , p. , petitjean , p. , savaglio , s. , & wijers , r. a. m. j. 2007 , , 468 , 83 llll 0.54203 & @xmath245 & @xmath246 & @xmath247 + 0.54468 & @xmath248 & @xmath249 & @xmath250 + 0.54630 & @xmath251 & @xmath252 & @xmath253 + 0.54868 & @xmath254 & @xmath255 & @xmath256 + 0.55235 & @xmath257 & @xmath258 & @xmath259 + 0.55447 & @xmath260 & @xmath261 & @xmath262 + 0.55787 & @xmath263 & @xmath264 & @xmath265 + 0.60723 & @xmath266 & @xmath267 & @xmath268 + 0.60991 & @xmath269 & @xmath270 & @xmath271 + 0.61257 & @xmath272 & @xmath273 & @xmath274 + 0.61587 & @xmath275 & @xmath276 & @xmath277 + 0.61921 & @xmath278 & @xmath279 & @xmath280 + 0.62327 & @xmath281 & @xmath282 & @xmath283 + 0.67575 & @xmath284 & @xmath285 & @xmath286 + 0.67845 & @xmath287 & @xmath288 & @xmath289 + 0.68217 & @xmath290 & @xmath291 & @xmath292 + 0.68565 & @xmath293 & @xmath294 & @xmath295 + 0.68898 & @xmath296 & @xmath297 & @xmath298 + 0.69205 & @xmath299 & @xmath300 & @xmath301 + 0.73354 & @xmath302 & @xmath303 & @xmath304 + 0.73787 & @xmath305 & @xmath306 & @xmath307 + 1.34534 & @xmath308 & @xmath309 & @xmath310 + 1.35747 & @xmath311 & @xmath312 & @xmath313 + 1.41052 & @xmath314 & @xmath315 & @xmath316 + 1.42675 & @xmath317 & @xmath318 & @xmath319 + 1.47713 & @xmath320 & @xmath321 & @xmath322 + 1.48861 & @xmath323 & @xmath324 & @xmath325 + 1.78359 & @xmath326 & @xmath327 & @xmath328 + 1.89040 & @xmath329 & @xmath330 & @xmath331 + 1.94946 & @xmath332 & @xmath333 & @xmath334 + 2.05462 & @xmath335 & @xmath336 & @xmath337 + 2.18650 & @xmath338 & @xmath339 & @xmath340 + 2.30888 & @xmath341 & @xmath342 & @xmath343 + 2.47372 & @xmath344 & @xmath345 & @xmath346 + 2.68019 & @xmath347 & @xmath348 & @xmath349 + 2.83975 & @xmath350 & @xmath351 & @xmath352 + 3.07971 & @xmath353 & @xmath354 & @xmath355 + 3.38860 & @xmath356 & @xmath357 & @xmath358 + 3.59126 & @xmath359 & @xmath360 & @xmath361 + 4.04549 & @xmath362 & @xmath363 & @xmath364 + 5.97272 & @xmath365 & @xmath366 & @xmath367 + 9.08277 & @xmath368 & @xmath369 & @xmath370 + 10.5438 & @xmath371 & @xmath372 & @xmath373 + 14.5620 & @xmath374 & @xmath375 & @xmath376 + 39.6190 & @xmath377 & @xmath378 & @xmath379 + cccclccll 0.54034 & v & 0.55 & 3590 & @xmath380 & 0.17 & 18.37 & @xmath381 & gcn6041 , 6036 + 0.55035 & b & 0.45 & 4020 & @xmath382 & 0.23 & 18.69 & @xmath383 & gcn6036 , 6041 + 0.67441 & v & 0.55 & 3590 & @xmath384 & 0.17 & 18.57 & @xmath385 & gcn6036 , 6041 + 0.68433 & b & 0.45 & 4020 & @xmath386 & 0.23 & 18.80 & @xmath387 & gcn6036 , 6041 + 0.79149 & r & 0.66 & 3020 & @xmath388 & 0.14 & 18.45 & @xmath389 & p60 + 0.79379 & r & 0.66 & 3020 & @xmath390 & 0.14 & 18.37 & @xmath391 & p60 + 0.79609 & r & 0.66 & 3020 & @xmath392 & 0.14 & 18.43 & @xmath393 & p60 + 0.79848 & r & 0.66 & 3020 & @xmath394 & 0.14 & 18.47 & @xmath395 & p60 + 0.80087 & r & 0.66 & 3020 & @xmath396 & 0.14 & 18.39 & @xmath397 & p60 + 0.80229 & r & 0.66 & 3020 & @xmath398 & 0.14 & 18.46 & @xmath399 & gcn6028 + 0.80326 & r & 0.66 & 3020 & @xmath400 & 0.14 & 18.41 & @xmath401 & p60 + 0.89965 & r & 0.63 & 3631 & @xmath402 & 0.14 & 18.89 & @xmath403 & gmos + 0.90833 & r & 0.66 & 3020 & @xmath404 & 0.14 & 18.66 & @xmath405 & gcn6044 + 1.08990 & r & 0.66 & 3020 & @xmath406 & 0.14 & 18.56 & @xmath407 & p60 + 1.09277 & r & 0.66 & 3020 & @xmath408 & 0.14 & 18.49 & @xmath409 & p60 + 1.09564 & r & 0.66 & 3020 & @xmath410 & 0.14 & 18.55 & @xmath411 & p60 + 1.10146 & r & 0.66 & 3020 & @xmath412 & 0.14 & 18.52 & @xmath413 & p60 + 1.10450 & i & 0.77 & 3631 & @xmath394 & 0.10 & 18.51 & @xmath414 & p60 + 1.10742 & i & 0.77 & 3631 & @xmath394 & 0.10 & 18.51 & @xmath415 & p60 + 1.11337 & i & 0.77 & 3631 & @xmath416 & 0.10 & 18.43 & @xmath417 & p60 + 1.11630 & i & 0.77 & 3631 & @xmath388 & 0.10 & 18.49 & @xmath418 & p60 + 1.11929 & r & 0.66 & 3020 & @xmath419 & 0.14 & 18.46 & @xmath420 & p60 + 1.12222 & r & 0.66 & 3020 & @xmath419 & 0.14 & 18.46 & @xmath421 & p60 + 1.12516 & r & 0.66 & 3020 & @xmath422 & 0.14 & 18.50 & @xmath423 & p60 + 1.12811 & r & 0.66 & 3020 & @xmath419 & 0.14 & 18.46 & @xmath420 & p60 + 1.13108 & r & 0.66 & 3020 & @xmath424 & 0.14 & 18.48 & @xmath425 & p60 + 1.13405 & i & 0.77 & 3631 & @xmath422 & 0.10 & 18.54 & @xmath426 & p60 + 1.13698 & i & 0.77 & 3631 & @xmath427 & 0.10 & 18.52 & @xmath428 & p60 + 1.14743 & r & 0.66 & 3020 & @xmath429 & 0.14 & 18.59 & @xmath430 & gcn6035 + 1.15466 & r & 0.66 & 3020 & @xmath431 & 0.14 & 18.49 & @xmath432 & p60 + 1.15705 & r & 0.66 & 3020 & @xmath433 & 0.14 & 18.50 & @xmath434 & p60 + 1.16182 & r & 0.66 & 3020 & @xmath435 & 0.14 & 18.49 & @xmath436 & p60 + 1.16421 & r & 0.66 & 3020 & @xmath437 & 0.14 & 18.52 & @xmath438 & p60 + 1.16661 & i & 0.77 & 3631 & @xmath439 & 0.10 & 18.55 & @xmath440 & p60 + 1.16900 & i & 0.77 & 3631 & @xmath433 & 0.10 & 18.54 & @xmath441 & p60 + 1.17378 & i & 0.77 & 3631 & @xmath412 & 0.10 & 18.56 & @xmath442 & p60 + 1.17617 & i & 0.77 & 3631 & @xmath443 & 0.10 & 18.58 & @xmath444 & p60 + 1.17857 & r & 0.66 & 3020 & @xmath439 & 0.14 & 18.51 & @xmath445 & p60 + 1.18096 & r & 0.66 & 3020 & @xmath412 & 0.14 & 18.52 & @xmath446 & p60 + 1.18336 & r & 0.66 & 3020 & @xmath422 & 0.14 & 18.50 & @xmath447 & p60 + 1.18576 & r & 0.66 & 3020 & @xmath448 & 0.14 & 18.58 & @xmath449 & p60 + 1.18815 & r & 0.66 & 3020 & @xmath450 & 0.14 & 18.52 & @xmath451 & p60 + 1.19060 & i & 0.77 & 3631 & @xmath452 & 0.10 & 18.61 & @xmath453 & p60 + 1.19300 & i & 0.77 & 3631 & @xmath454 & 0.10 & 18.59 & @xmath455 & p60 + 1.19539 & i & 0.77 & 3631 & @xmath456 & 0.10 & 18.57 & @xmath457 & p60 + 1.20019 & i & 0.77 & 3631 & @xmath458 & 0.10 & 18.60 & @xmath459 & p60 + 1.20261 & r & 0.66 & 3020 & @xmath460 & 0.14 & 18.53 & @xmath461 & p60 + 1.20500 & r & 0.66 & 3020 & @xmath439 & 0.14 & 18.51 & @xmath462 & p60 + 1.20740 & r & 0.66 & 3020 & @xmath463 & 0.14 & 18.51 & @xmath464 & p60 + 1.20980 & r & 0.66 & 3020 & @xmath452 & 0.14 & 18.57 & @xmath465 & p60 + 1.21220 & r & 0.66 & 3020 & @xmath448 & 0.14 & 18.58 & @xmath466 & p60 + 1.21464 & i & 0.77 & 3631 & @xmath467 & 0.10 & 18.61 & @xmath468 & p60 + 1.21704 & i & 0.77 & 3631 & @xmath469 & 0.10 & 18.58 & @xmath470 & p60 + 1.21944 & i & 0.77 & 3631 & @xmath452 & 0.10 & 18.61 & @xmath471 & p60 + 1.22185 & i & 0.77 & 3631 & @xmath472 & 0.10 & 18.64 & @xmath473 & p60 + 1.22425 & i & 0.77 & 3631 & @xmath456 & 0.10 & 18.57 & @xmath457 & p60 + 1.22521 & rc & 0.66 & 3020 & @xmath474 & 0.14 & 18.76 & @xmath475 & gcn6050 + 1.22910 & r & 0.66 & 3020 & @xmath467 & 0.14 & 18.57 & @xmath476 & p60 + 1.23151 & r & 0.66 & 3020 & @xmath477 & 0.14 & 18.56 & @xmath478 & p60 + 1.23391 & r & 0.66 & 3020 & @xmath479 & 0.14 & 18.61 & @xmath480 & p60 + 1.23632 & r & 0.66 & 3020 & @xmath481 & 0.14 & 18.62 & @xmath482 & p60 + 1.27104 & rc & 0.66 & 3020 & @xmath483 & 0.14 & 18.86 & @xmath484&gcn6050 + 1.31410 & ic & 0.81 & 2380 & @xmath485 & 0.10 & 17.90 & @xmath486 & gcn6050 + 1.31410 & g & 0.49 & 3631 & @xmath487 & 0.06 & 19.54 & @xmath488 & gcn6050 + 1.31410 & rc & 0.66 & 3020 & @xmath489 & 0.14 & 18.66 & @xmath490 & gcn6050 + 1.35377 & v & 0.55 & 3590 & @xmath491 & 0.17 & 19.09 & @xmath492 & gcn6041 + 1.37590 & rc & 0.66 & 3020 & @xmath493 & 0.14 & 18.56 & @xmath494 & gcn6050 + 1.37605 & r & 0.66 & 3020 & @xmath495 & 0.14 & 18.95 & @xmath496 & gcn6039 + 1.40437 & rc & 0.66 & 3020 & @xmath497 & 0.14 & 19.26 & @xmath498 & gcn6050 + 1.79844 & r & 0.66 & 3020 & @xmath499 & 0.14 & 19.29 & @xmath500 & p60 + 1.81062 & i & 0.77 & 3631 & @xmath501 & 0.10 & 19.39 & @xmath502 & p60 + 1.84702 & r & 0.66 & 3020 & @xmath503 & 0.14 & 19.37 & @xmath504 & p60 + 1.85907 & i & 0.77 & 3631 & @xmath505 & 0.10 & 19.42 & @xmath506 & p60 + 1.90324 & r & 0.66 & 3020 & @xmath507 & 0.14 & 19.45 & @xmath508 & p60 + 1.91542 & i & 0.77 & 3631 & @xmath509 & 0.10 & 19.48 & @xmath510 & p60 + 1.96396 & rc & 0.66 & 3020 & @xmath511 & 0.14 & 19.57 & @xmath512 & gcn6096 + 1.97868 & r & 0.66 & 3020 & @xmath513 & 0.14 & 19.48 & @xmath514 & p60 + 1.99132 & i & 0.77 & 3631 & @xmath515 & 0.10 & 19.59 & @xmath516 & p60 + 2.64396 & rc & 0.66 & 3020 & @xmath517 & 0.14 & 20.09 & @xmath518 & gcn6047 + 2.65995 & rc & 0.66 & 3020 & @xmath519 & 0.14 & 20.12 & @xmath520 & gcn6047 + 2.67696 & rc & 0.66 & 3020 & @xmath521 & 0.14 & 20.07 & @xmath522&gcn6047 + 2.78096 & rc & 0.64 & 3020 & @xmath523 & 0.14 & 20.11 & @xmath524&gcn6047 + 2.79596 & rc & 0.66 & 3020 & @xmath525 & 0.14 & 20.21 & @xmath526&gcn6047 + 2.81479 & ks & 2.22 & 670 & @xmath527 & 0.02 & 17.84 & @xmath528 & gcn6054 + 2.81479 & j & 1.26 & 1600 & @xmath529 & 0.05 & 18.77 & @xmath530 & gcn6054 + 2.81479 & h & 1.66 & 1024 & @xmath531 & 0.01 & 18.32 & @xmath532 & gcn6054 + 2.99296 & r & 0.66 & 3020 & @xmath533 & 0.14 & 20.30 & @xmath534&gcn6096 + 3.02006 & r & 0.66 & 3020 & @xmath535 & 0.14 & 20.30 & @xmath536 & p60 + 3.03328 & i & 0.77 & 3631 & @xmath537 & 0.10 & 20.37 & @xmath538 & p60 + 3.62096 & rc & 0.66 & 3020 & @xmath539 & 0.14 & 20.66 & @xmath540 & gcn6064 + 3.91797 & r & 0.63 & 3631 & @xmath541 & 0.14 & 21.08 & @xmath542 & gmos + 4.03995 & r & 0.66 & 3020 & @xmath543 & 0.14 & 20.93 & @xmath544&gcn6096 + 4.07106 & i & 0.77 & 3631 & @xmath545 & 0.10 & 20.93 & @xmath546 & p60 + 4.08226 & r & 0.66 & 3020 & @xmath547 & 0.14 & @xmath548 & @xmath549 & p60 + 8.85767 & r & 0.66 & 3020 & @xmath550 & 0.14 & @xmath551 & @xmath552 & p60 + 10.0099 & i & 0.77 & 3631 & @xmath553 & 0.10 & 23.65 & @xmath554 & gmos + 11.82110 & r & 0.66 & 3020 & @xmath555 & 0.14 & @xmath556 & @xmath557 & p60 + 12.00096 & r & 0.66 & 3020 & @xmath558 & 0.14 & @xmath559 & @xmath560 & gcn6096 + 21.99400 & r & 0.66 & 3020 & @xmath561 & 0.14 & @xmath562 & @xmath563 & lris + 21.99400 & g & 0.49 & 3631 & @xmath564 & 0.06 & @xmath565 & @xmath563 & lris + 26.79396 & r & 0.67 & 3631 & @xmath566 & 0.14 & 26.16 & @xmath567 & gcn6165 + cccccc 2007 jan 26.82 & 1.51 & wsrt & 4.86 & @xmath568 & 87 + 2007 jan 29.32 & 4.01 & vla & 8.46 & 360 & 42 + 2007 jan 30.24 & 4.93 & vla & 8.46 & 454 & 38 + 2007 jan 30.95 & 5.64 & wsrt & 4.86 & 102 & 26 + 2007 jan 31.07 & 5.76 & vla & 4.86 & < 141 & 58 + 2007 jan 31.11 & 5.80 & vla & 8.46 & 382 & 52 + 2007 jan 31.76 & 6.45 & gmrt & 0.61 & @xmath569 & 150 + 2007 jan 31.83 & 6.52 & mambo2 & 250 & 3140 & 590 + 2007 feb 01.92 & 7.61 & mambo2 & 250 & 1910 & 720 + 2007 feb 02.06 & 7.75 & vla & 8.46 & 563 & 62 + 2007 feb 04.92 & 10.61 & mambo2 & 250 & @xmath570 & 710 + 2007 feb 05.03 & 10.72 & vla & 8.46 & 482 & 52 + 2007 feb 05.04 & 10.73 & vla & 22.50 & 1594 & 70 + 2007 feb 05.07 & 10.76 & vla & 4.86 & @xmath571 & 42 + 2007 feb 05.29 & 10.98 & carma & 95 & 2300 & 700 + 2007 feb 06.24 & 11.93 & vla & 8.46 & 489 & 43 + 2007 feb 06.26 & 11.95 & vla & 4.86 & @xmath572 & 34 + 2007 feb 07.11 & 12.80 & vla & 22.50 & 1603 & 235 + 2007 feb 07.12 & 12.81 & vla & 8.46 & 405 & 20 + 2007 feb 07.38 & 13.07 & vla&14.94 & 1159 & 234 + 2007 feb 08.21 & 13.90 & vla & 14.94 & 917 & 293 + 2007 feb 08.23 & 13.92 & vla & 4.86 & @xmath573 & 50 + 2007 feb 08.24 & 13.93 & vla & 8.46 & 559 & 26 + 2007 feb 08.40 & 14.09 & vla & 22.50 & 1621 & 151 + 2007 feb 09.21 & 14.90 & vla & 4.86 & @xmath574 & 49 + 2007 feb 09.25 & 14.94 & vla & 8.46 & 399 & 69 + 2007 feb 09.27 & 14.96 & vla & 22.50 & 1343 & 162 + 2007 feb 09.27 & 14.96 & carma & 95 & 2300 & 800 + 2007 feb 09.28 & 14.97 & vla & 14.94 & 891 & 129 + 2007 feb 10.37 & 16.06 & vla & 14.94 & 1410 & 137 + 2007 feb 10.79 & 16.48 & mambo2 & 250 & 2670 & 930 + 2007 feb 11.08 & 16.77 & vla & 8.46 & 385 & 42 + 2007 feb 11.12 & 16.81 & vla & 22.50 & 1367 & 104 + 2007 feb 11.16 & 16.85 & vla & 4.86 & @xmath575 & 71 + 2007 feb 12.08 & 17.77 & vla & 1.46 & @xmath576 & 460 + 2007 feb 12.16 & 17.85 & vla & 8.46 & 596 & 109 + 2007 feb 12.20 & 17.89 & vla & 14.94 & 1226 & 155 + 2007 feb 12.21 & 17.90 & carma & 95 & 2100 & 700 + 2007 feb 12.26 & 17.95 & vla & 22.50 & 1222 & 167 + 2007 feb 12.79 & 18.48 & mambo2 & 250 & @xmath577 & 930 + 2007 feb 13.07 & 18.76 & vla&1.46 & @xmath578 & 600 + 2007 feb 13.11 & 18.80 & vla & 4.86 & @xmath579 & 52 + 2007 feb 13.17 & 18.86 & vla & 8.46 & 660 & 39 + 2007 feb 13.21 & 18.90 & vla & 14.94 & 1217 & 197 + 2007 feb 13.25 & 18.94 & vla & 22.50 & 1248 & 83 + 2007 feb 14.18 & 19.87 & vla & 8.46 & 581 & 14 + 2007 feb 16.17 & 21.86 & vla & 1.46 & @xmath580 & 540 + 2007 feb 16.19 & 21.88 & vla & 4.86 & @xmath581 & 47 + 2007 feb 17.02 & 22.71 & vla & 1.46 & @xmath582 & 766 + 2007 feb 17.04 & 22.73 & vla & 4.86 & @xmath583 & 64 + 2007 feb 18.16 & 23.85 & vla & 22.50 & 1168 & 118 + 2007 feb 18.18 & 23.87 & vla & 8.46 & 303 & 63 + 2007 feb 18.21 & 23.90 & vla & 14.94 & 798 & 182 + 2007 feb 18.27 & 23.96 & carma & 95 & 2400 & 700 + 2007 feb 21.05 & 26.74 & vla & 14.94 & 1101 & 148 + 2007 feb 22.21 & 27.90 & vla & 4.86 & @xmath584 & 78 + 2007 feb 23.07 & 28.76 & vla & 1.46 & @xmath585 & 804 + 2007 feb 25.16 & 30.85 & vla & 22.50 & 839 & 73 + 2007 feb 27.20 & 32.89 & vla & 1.46 & @xmath586 & 639 + 2007 mar 01.22 & 34.91 & vla & 4.86 & @xmath587 & 95 + 2007 mar 01.24 & 34.93 & vla & 14.94 & 432 & 149 + 2007 mar 01.28 & 34.97 & vla & 8.46 & 414 & 66 + 2007 mar 01.30 & 34.99 & vla & 22.50 & 788 & 74 + 2007 mar 13.18 & 46.87 & vla & 1.46 & @xmath588 & 739 + 2007 mar 15.20 & 48.89 & vla & 4.86 & 229 & 49 + 2007 mar 15.22 & 48.91 & vla & 8.46 & 443 & 59 + 2007 mar 21.22 & 54.91 & vla & 4.86 & 262 & 52 + 2007 mar 21.26 & 54.95 & vla & 8.46 & 473 & 44 + 2007 mar 23.08 & 56.77 & vla & 22.50 & 559 & 161 + 2007 mar 25.15 & 58.84 & vla & 1.46 & @xmath589 & 280 + 2007 apr 02.08 & 66.77 & vla & 4.86 & 226 & 57 + 2007 apr 02.12 & 66.81 & vla & 8.46 & 345 & 48 + 2007 apr 02.16 & 66.85 & vla & 1.46 & @xmath590 & 480 + 2007 apr 02.20 & 66.89 & vla & 14.94 & 530 & 137 + 2007 apr 02.24 & 66.93 & vla & 22.50 & 568 & 137 + 2007 apr 21.13 & 85.82 & vla & 1.46 & @xmath591 & 570 + 2007 apr 21.18 & 85.87 & vla & 4.86 & 302 & 55 + 2007 apr 22.18 & 86.87 & vla & 8.46 & 403 & 56 + 2007 apr 22.99 & 87.68 & vla & 22.50 & 615 & 167 + 2007 apr 23.02 & 87.71 & vla & 14.94 & @xmath592 & 226 + 2007 may 15.03 & 109.72 & vla & 8.46 & 267 & 42 + 2007 may 17.00 & 111.69 & vla & 1.46 & @xmath593 & 278 + 2007 may 18.81 & 113.50 & vla & 4.86 & @xmath594 & 145 + 2007 may 19.91 & 114.60 & vla & 14.94 & @xmath595 & 720 + 2007 may 20.04 & 114.73 & vla & 22.50 & @xmath596 & 88 + 2007 jul 04.76 & 160.45 & vla&8.46 & 145 & 33 + 2007 jul 04.80 & 160.49 & vla&4.86 & 174 & 45 + 2007 jul 04.85 & 160.54 & vla & 1.46 & @xmath597 & 160 + 2007 aug 03.87 & 190.56 & vla & 22.50 & @xmath598 & 223 + 2007 aug 11.68 & 198.37 & vla & 4.86 & @xmath599 & 42 + 2007 aug 11.72 & 198.41 & vla & 8.46 & 166 & 46 + 2007 aug 13.71 & 200.40 & vla & 14.94 & @xmath600 & 169 + 2007 sep 08.64 & 226.33 & vla & 4.86 & @xmath601 & 57 + 2007 oct 04.52 & 252.21 & vla & 8.46 & 148 & 60 + 2007 oct 04.60 & 252.29 & vla & 4.86 & 203 & 34 + 2007 nov 18.39 & 297.08 & vla & 8.46 & 111 & 19 + 2007 nov 18.48 & 297.17 & vla & 4.86 & 161 & 23 + 2008 jan 02.27 & 341.96 & vla & 8.46 & 64 & 18 + 2008 jan 02.35 & 342.04 & vla & 4.86 & 133 & 21 + lcc|cc & 1.80 & @xmath602 , @xmath603 & @xmath604 & @xmath605 , @xmath606 + & & & @xmath607 & @xmath608 + & & & @xmath609 & @xmath609 + & 2.23 ( 99 d.o.f . ) & 2.25 ( 98 d.o.f . ) & 1.30 ( 97 d.o.f . ) & 1.31 ( 96 d.o.f . ) + & 0.88 ( 22 d.o.f . ) & 0.87 ( 22 d.o.f . ) & 1.32 ( 20 d.o.f . ) & 1.31 ( 20 d.o.f . ) + & 2.22 ( 52 d.o.f . ) & 2.21 ( 52 d.o.f . ) & 1.54 ( 52 d.o.f . ) & 1.54 ( 52 d.o.f . ) + & 3.43 ( 25 d.o.f . ) & 3.59 ( 24 d.o.f . ) & 0.77 ( 25 d.o.f . ) & 0.81 ( 24 d.o.f . ) + [ tab : jet ] l|ll|ll|ll @xmath610(@xmath121erg ) & @xmath611 & @xmath612 & @xmath613 & @xmath614 & @xmath615 & @xmath616 + @xmath6 ( rad ) & @xmath617 & @xmath618 & @xmath619 & @xmath620 & @xmath621 & @xmath622 + @xmath623 & @xmath624 & @xmath625 & @xmath626 & @xmath627 & @xmath628 & @xmath629 + @xmath12 & @xmath630 & @xmath631 & @xmath632 & @xmath633 & 0.4 & 0.4 + @xmath209(% ) & @xmath634 & @xmath635 & @xmath636 & @xmath637 & @xmath638 & @xmath639 + @xmath640 or @xmath641 & @xmath642 & @xmath643 & @xmath644 & @xmath645 & @xmath646 & @xmath647 + @xmath648 ( days ) & @xmath649 & @xmath650 & @xmath651 & @xmath652 & @xmath653 & @xmath654 + @xmath655 ( days ) & @xmath656 & @xmath657 & @xmath658 & @xmath659 & @xmath660 & @xmath661 + @xmath662(days ) & @xmath663 & @xmath664 & @xmath665 & @xmath666 & @xmath667 & @xmath668 + @xmath669 & @xmath670 & 0.06 & @xmath671 & 0.11 & @xmath672 & 0.08 + fit statistic & 592.02 & 633.02 & 556.26 & 575.36 & 584.85 & 588.96 + @xmath673 & @xmath674 & @xmath675 & @xmath676 & @xmath677 & @xmath678 & @xmath679 + - and @xmath42-band ) light curves of grb 070125as a result of joint optical and x - ray fits .
best - fit single power - law models are shown with dashed lines , while the broken power - law models are shown in solid lines .
it is clear that in the optical bands , a broken power - law ( indicating a jet break ) is strongly favored .
_ bottom _ : x - ray light curve of grb 070125 , the the joint fit to optical and x - ray data . again
the single power - law model is shown as a dashed line , while the broken power - law model is shown as a solid line .
grey solid line indicates the independent fit to x - ray data .
the independent fit is consistent with the joint fit in the optical bands but shifts the jet break to @xmath680 @xmath207 days .
we discuss this in [ sec : tjet ] and [ sec : discussion ] .
, title="fig:",scaledwidth=70.0% ] - and @xmath42-band ) light curves of grb 070125as a result of joint optical and x - ray fits .
best - fit single power - law models are shown with dashed lines , while the broken power - law models are shown in solid lines .
it is clear that in the optical bands , a broken power - law ( indicating a jet break ) is strongly favored .
_ bottom _ : x - ray light curve of grb 070125 , the the joint fit to optical and x - ray data . again
the single power - law model is shown as a dashed line , while the broken power - law model is shown as a solid line .
grey solid line indicates the independent fit to x - ray data .
the independent fit is consistent with the joint fit in the optical bands but shifts the jet break to @xmath680 @xmath207 days .
we discuss this in [ sec : tjet ] and [ sec : discussion ] . , title="fig:",scaledwidth=70.0% ] , @xmath681 , @xmath152 and @xmath682 and extracted the best fit @xmath144
. between day @xmath682 the spectrum is in the optically thin phase , @xmath144 is below the observed frequencies .
the grey area indicates the uncertainity region for the powerlaw index slope .
, title="fig:",scaledwidth=49.0% ] , @xmath681 , @xmath152 and @xmath682 and extracted the best fit @xmath144
. between day @xmath682 the spectrum is in the optically thin phase , @xmath144 is below the observed frequencies .
the grey area indicates the uncertainity region for the powerlaw index slope .
, title="fig:",scaledwidth=49.0% ] , @xmath681 , @xmath152 and @xmath682 and extracted the best fit @xmath144
. between day @xmath682 the spectrum is in the optically thin phase , @xmath144 is below the observed frequencies .
the grey area indicates the uncertainity region for the powerlaw index slope .
, title="fig:",scaledwidth=49.0% ] , @xmath681 , @xmath152 and @xmath682 and extracted the best fit @xmath144
. between day @xmath682 the spectrum is in the optically thin phase , @xmath144 is below the observed frequencies .
the grey area indicates the uncertainity region for the powerlaw index slope . , title="fig:",scaledwidth=49.0% ] , @xmath144 , @xmath155 , and @xmath156 parameters .
the plots at the bottom are @xmath683 plots . for slow cooling
the flux should peak at @xmath155 and the total power ( @xmath683 ) should peak at @xmath156 .
, title="fig:",scaledwidth=49.0% ] , @xmath144 , @xmath155 , and @xmath156 parameters .
the plots at the bottom are @xmath683 plots . for slow cooling
the flux should peak at @xmath155 and the total power ( @xmath683 ) should peak at @xmath156 .
, title="fig:",scaledwidth=49.0% ] , @xmath144 , @xmath155 , and @xmath156 parameters .
the plots at the bottom are @xmath683 plots . for slow cooling
the flux should peak at @xmath155 and the total power ( @xmath683 ) should peak at @xmath156 .
, title="fig:",scaledwidth=49.0% ] , @xmath144 , @xmath155 , and @xmath156 parameters .
the plots at the bottom are @xmath683 plots . for slow cooling
the flux should peak at @xmath155 and the total power ( @xmath683 ) should peak at @xmath156 .
, title="fig:",scaledwidth=49.0% ] band , and @xmath240 band dataset .
the darker line is for the ism model and the wind model is plotted with a thin line .
the fits in the radio bands are not good , probably because of scintillation effects.,title="fig:",scaledwidth=49.0% ] band , and @xmath240 band dataset .
the darker line is for the ism model and the wind model is plotted with a thin line .
the fits in the radio bands are not good , probably because of scintillation effects.,title="fig:",scaledwidth=49.0% ] band , and @xmath240 band dataset .
the darker line is for the ism model and the wind model is plotted with a thin line .
the fits in the radio bands are not good , probably because of scintillation effects.,title="fig:",scaledwidth=49.0% ] band , and @xmath240 band dataset .
the darker line is for the ism model and the wind model is plotted with a thin line .
the fits in the radio bands are not good , probably because of scintillation effects.,title="fig:",scaledwidth=49.0% ] band , and @xmath240 band dataset .
the darker line is for the ism model and the wind model is plotted with a thin line .
the fits in the radio bands are not good , probably because of scintillation effects.,title="fig:",scaledwidth=49.0% ] band , and @xmath240 band dataset .
the darker line is for the ism model and the wind model is plotted with a thin line .
the fits in the radio bands are not good , probably because of scintillation effects.,title="fig:",scaledwidth=49.0% ] | we present a comprehensive multiwavelength analysis of the bright , long duration gamma - ray burst grb 070125 , comprised of observations in @xmath0-ray , x - ray , optical , millimeter and centimeter wavebands .
simultaneous fits to the optical and x - ray light curves favor a break on day 3.78 , which we interpret as the jet break from a collimated outflow .
independent fits to optical and x - ray bands give similar results in the optical bands but shift the jet break to around day 10 in the x - ray light curve .
we show that for the physical parameters derived for grb 070125 , inverse compton scattering effects are important throughout the afterglow evolution .
while inverse compton scattering does not affect radio and optical bands , it may be a promising candidate to delay the jet break in the x - ray band .
radio light curves show rapid flux variations , which are interpreted as due to interstellar scintillation , and are used to derive an upper limit of @xmath1 cm on the radius of the fireball in the lateral expansion phase of the jet .
radio light curves and spectra suggest a high synchrotron self absorption frequency indicative of the afterglow shock wave moving in a dense medium .
our broadband modeling favors a constant density profile for the circumburst medium over a wind - like profile ( @xmath2 ) .
however , keeping in mind the uncertainty of the parameters , it is difficult to unambiguously distinguish between the two density profiles .
our broadband fits suggest that grb 070125 is a burst with high radiative efficiency ( @xmath3 ) . |
ionization cooling of a muon beam can be accomplished by passing the beam through energy - absorbing material and accelerating structures , both embedded within a focusing magnetic lattice .
the rate of change of the normalized transverse emittance with path length is then approximately described by @xcite @xmath0 where @xmath1 is the path length , @xmath2 the muon energy , @xmath3 the radiation length of the absorber medium , @xmath4 , and @xmath5 is the betatron function of the beam ( where the size of the beam is given by @xmath6 ) . in eq .
[ eq : cool ] we see , in addition to the @xmath7 transverse cooling term , a transverse heating term due to multiple coulomb scattering of the muons in the absorbers . since cooling ceases once the heating and cooling terms are equal , eq . [ eq : cool ] implies an equilibrium emittance , which in principle ( neglecting other limiting effects ) would be reached asymptotically were the cooling channel continued indefinitely .
since the heating term is proportional to @xmath5 and inversely proportional to the radiation length of the absorber medium , the goal of achieving as small an equilibrium emittance as possible requires us to locate the absorber only in low-@xmath5 regions and to use a medium with the longest possible radiation length , namely hydrogen . to achieve low @xmath5
, we want the strongest possible focusing elements .
we are thus led to superconducting solenoids filled with liquid hydrogen as possibly the optimal solution .
a variety of solenoidal focusing lattices have been considered for muon - beam cooling , including the so - called alternating - solenoid , fofo , dfofo , sfofo , and single- and double - flip designs @xcite . in all of these , liquid - hydrogen ( lh@xmath8 ) absorbers are used to minimize multiple scattering .
the specifications of the absorbers for some representative cases are given in table [ tab : abs ] .
.specifications of typical absorbers ( from ref .
@xcite ) . [ cols="<,^,^,^,<",options="header " , ] @xmath9 assuming @xmath10 muons / bunch at 15hz at 200 mev/@xmath11 .
+ @xmath12 operation at 1 atm will probably require a somewhat lower operating temperature to give more headroom " with respect to the boiling point ; this will be iterated once the fluid - flow and thermal models have been fully calculated and verified in bench tests .
a key problem is handling the heat deposited in the hydrogen by the muon beam . as shown in table [
tab : abs ] , this can exceed 100 watts per absorber .
lh@xmath8 targets have been successfully operated in such a heat - deposition regime , notably the target for the sample experiment at bates @xcite .
these high - power target designs have followed the approach developed at slac @xcite , in which the liquid is pumped around an external cooling loop ( see fig .
[ fig : loop ] ) .
the loop includes a heat exchanger , within which the deposited heat is transferred to a supply of cold helium gas , as well as a heater for fast regulation of temperature .
the difficulty in such designs is assuring flow of the lh@xmath8 within the absorber cell that is predominantly transverse to the beam direction , since any little parcel of lh@xmath8 that flowed along the beam would be quickly heated above the boiling point . in the sample target ( fig .
[ fig : loop]b ) transverse flow was accomplished using a specially - designed perforated baffle , and for the proposed slac experiment 158 target ( fig .
[ fig : loop]a ) the use of asymmetrically - perforated screens is under investigation @xcite .
we are developing a design in which transverse flow within the absorber cell is produced by variously - angled nozzles placed around the periphery .
in addition to the external - loop approach , we are considering as an alternative heat exchange internal to the absorber vessel , using driven convection to provide mixing and transverse flow ( fig .
[ fig : convection ] ) .
this has the obvious virtue of fewer moving parts .
moreover , the transverse flow arises naturally , rather than being imposed by a clever but complicated design .
the key question is whether there will be sufficient convection at the anticipated power level to give the necessary rate of heat exchange via the cooling tubes at the periphery of the cell .
we are addressing this by a series of computational - fluid - dynamics ( cfd ) calculations that numerically solve the navier - stokes equation on a suitable two - dimensional grid to evaluate the nusselt number ( the dimensionless parameter that characterizes the rate of convective heat transport ) _ vs. _ the rayleigh number ( the dimensionless parameter , proportional to the power dissipation , that characterizes the degree of turbulence ) .
preliminary indications are that there will be sufficient convection ; this will be investigated further numerically and tested experimentally on the bench .
the thickness of the absorber windows is a critical parameter
. they must be thick enough to sustain the pressure from the lh@xmath8 , yet as thin as possible so as to minimize multiple scattering .
the window thicknesses in table [ tab : abs ] have been chosen based on the asme standard for pressure vessels @xcite .
this choice also satisfies the fermilab safety code for liquid - hydrogen targets @xcite . as illustrated in fig .
[ fig : windows ] , asme @xcite specifies three standard window profiles : hemispherical , ellipsoidal , and torispherical .
the minimum thickness required in each case is @xmath13 where @xmath14 is the pressure differential , @xmath15 the radius of curvature ( for hemispherical windows ) , @xmath16 the length of the major axis ( for ellipsoidal or torispherical windows ) , _ i.e. _ the absorber diameter , @xmath17 the maximum allowable stress , and @xmath18 the weld efficiency . in the above equations we give also @xmath1 , the sagitta of the window at its center , which determines which window shapes can be used for absorbers of given dimensions : since each absorber has two windows , the absorber length @xmath19 must satisfy @xmath20 .
( note that for ellipsoidal windows the asme code considers only the case where the length of the major axis is twice that of the minor axis , and for the torispherical case the radius of curvature of the knuckle " is 6% that of the main cap ; see fig .
[ fig : windows ] . )
the maximum allowable stress is the smaller of ( ultimate stress)/4 or ( yield stress)/1.5 @xcite . in practice
we find that it is the ultimate stress that is the limit .
the hemispherical window shape minimizes the needed thickness .
however , a hemispherical window is practical only for absorbers whose length exceeds their diameter . since our absorbers are typically wider in diameter than they are long , we choose either the ellipsoidal or torispherical window profile . as an example , we show in fig .
[ fig : fofo ] the mechanical design for the fofo1 " absorber of table [ tab : abs ] .
given the 15 cm radius of the absorber , both hemispherical and ellipsoidal windows are ruled out .
torispherical windows with 5.1 cm sagitta leave just 2.4 cm for the flange and manifold assembly that joins the two windows and connects to the hydrogen inlets and outlets .
our solution is to machine each window with an integral flange out of a single block of material .
the two flanges bolt together in
clamshell " fashion to form the manifold . in the neutrino factory feasibility study
@xcite we consider fofo and single - flip options for the cooling channel . in both designs the cooling performance
is significantly limited by scattering in the absorber windows . to minimize this effect we propose to operate the absorbers at 1atm pressure .
absorbers at 1atm pressure is not an established technique , 2atm having been used in previous high - power target designs @xcite .
there may be safety concerns that will prevent such operation @xcite ; this is a topic for r&d . ] in the single - flip design the absorbers are long enough to permit the use of 300@xmath21 m ellipsoidal windows . in the fofo design two
cooling sections are used having two different absorber sizes . in the first ( fofo1 " ) section , 400@xmath21 m torispherical windows are used as discussed above . by the second section , the beam has become small enough to permit a reduction in absorber diameter , allowing use of 200@xmath21 m ellipsoidal windows . in specifying these window thicknesses ,
however in the integral - flange approach discussed above , @xmath22 , allowing 10% thinner windows , or alternatively , operation at 1.1atm with the thicknesses given . ] the reduction in window thickness results in a lowering of the equilibrium emittance from 2.6@xmath23 to 2.2@xmath23mm@xmath24rad and a corresponding increase in the cooling rate .
we have also begun to explore the option of customizing the thickness profile of the window in order to minimize the thickness at the center while maximizing strength .
an ansys finite - element calculation has been carried out that shows that the stresses in a torispherical window are greatest near the edge , in the region in which the window curvature under pressure exhibits a point of inflection ( fig .
[ fig : new - window ] ) . by thickening the material near the edge one can reduce
the maximum stresses substantially , allowing the material near the center to be thinner by perhaps as much as a factor of 5 . of course manufacturability
will also impose a limit on how thin the center of a machined window can be ; we will explore this soon by building and testing prototypes .
beryllium or albemet ( a composite of 62% beryllium/38% aluminum ) could reduce the impact of the windows on the cooling performance . however , based on the cea bubble - chamber accident , beryllium is believed to be incompatible with liquid hydrogen , and an r&d program will be required to establish safe design parameters for these materials .
with 40% greater strength than aluminum and 2.1 times the radiation length , albemet has the potential to lower the total radiation - length fraction per absorber from 2.4% to 1.8% or less , depending on the detailed optimization of absorber dimensions .
( while beryllium windows may also be feasible , there may be little additional gain in going beyond albemet . ) other cooling scenarios ( _ e.g. _ sfofo ) use absorbers that are thicker compared to their diameter . here
effects of windows on cooling performance are reduced , and aluminum windows may be adequate .
whether r&d on exotic window materials is worthwhile may thus depend on which cooling approach prevails . in all scenarios
the specific power dissipation in the absorbers is large and represents a substantial portion of the cryogenic load of the cooling channel .
handling this heat load is a significant design challenge .
an r&d program is already in place at iit to understand the thermal and fluid - flow aspects of maintaining a constant temperature within the absorber volume despite the large spatial and temporal variations in power density .
this program is beginning with cfd studies and is planned to proceed to bench tests and high - power beam tests of absorber prototypes over the next year . in some scenarios ( especially those with emittance exchange ) , lithium hydride ( lih ) absorbers
may be called for .
since it is a solid , lih in principle can be fabricated in arbitrary shapes . in emittance - exchange channels ,
dispersion in the lattice spatially separates muons according to their energies , whereupon specially shaped absorbers can be used to absorb more energy from muons of higher energy and less from those of lower energy .
however , solid lih shapes are not commercially available , and procedures for their fabrication would need to be developed .
such an effort is challenging since lih reacts with water , releasing hydrogen gas and creating an explosion hazard .
we thank m. boghosian of iit who carried out the cfd calculations and z. tang of fermilab who carried out the ansys window calculations . this work was supported in part by the u.s . dept . of energy , the national science foundation , and the iit research institute . | we describe the program of research and development being undertaken by a consortium of illinois university groups to develop liquid - hydrogen energy absorbers for muon - beam cooling . |
many quasars show short - term or / and long - term variability ( ulrich , maraschi & urry 1997 ) .
these changes in the source flux help to constrain the physics and the size of the central engine .
astronomers came to an early understanding that the central engines of the agns and qsos can not be resolved easily , if at all , but that the variability timescales measure the size of the emitting regions ( i.e. blandford & mckee 1982 ) . later on , this became the basis of reverberation studies ( for a recent review see peterson et al .
2004 and the references therein ) that could map the innermost broad line regions ( hereafter blr ) .
nevertheless , the exact variability mechanisms remain unclear . typically , the variability is aperiodic , but it does show some dependencies on time lag , luminosity , wavelengths and redshift .
for example , the variability amplitude increases with time lag ( vanden berk et al .
2004 ) , more luminous quasars are less variable ( vanden berk et al . 2004 ; de vries et al . 2005 ; giveon et al .
1999 ) , and the variability increases toward the blue part of the spectrum ( vanden berk et al . 2004 ; de vries et al . 2005 ) .
it is particularly difficult to constrain the variability versus redshift dependence because of the inevitable biases at high redshift , limiting the quasar luminosity range , the probed time - baseline , etc . of the more distant quasars .
the radio properties of qsos also seem to be related to the optical variability : the radio - loud ones are relatively more variable that the radio quiet ones , and the blazars show even stronger variability because of beaming , which is quite different than the long - term variability studied in this work . for a more comprehensive review of the qso variability properties we refer the reader to the summary of wold , brotherton & shang ( 2007 ) . as of now
, there is no commonly accepted theory of the qso variability and the existence of some strange objects , albeit rare ones , complicates the picture even further .
a good example is the radio - quiet qso sdssj153259.96 - 003944.1 ( stalin & srianand 2005 ) who is the prototype of the rare class of weak ( or absent ) emission - line quasars ( wlqs ) .
this object shows strong long - term variability and flat optical spectrum like bllacs but there is no radio emission , no optical polarization and no x - ray ( shemmer et al .
these properties may be explained by a deficit of line - emitting gas in the vicinity of the central continuum source , similar to the x - ray weak quasar phl1811 ( leighly et al .
2007 and the references therein ) .
the most comprehensive quasar variability studies , especially the ones covering long time scales , are focused either on bright low - redshift ones ( hook et al . 1994 ; hawkins 2002 ) , or they study the behavior of the structure function for large samples ( de vries , becker & white 2003 ; de vries et al . 2005 ; hovatta et al .
the literature is lacking well sampled light curves of individual distant quasars , with the notable exception of kaspi et al .
( 2007 ) with whom we share five objects from our extended program .
while challenging , the distant qsos are also rewarding because the variability can help constrain the size of the qso accretion disks ( hawkins 2007 ) at early times .
furthermore , identifying variable high - redshift quasars is a necessary first step for reverberation mapping of their broad line regions ( i.e. kaspi et al .
these arguments motivated us to start an optical monitoring study of qsos with @xmath1@xmath54 . here
we present a photometric sequence for q2203 + 292 , which is one of the first detected quasars within that redshift range ( dickinson & mccarthy 1987 ) and it is radio - quiet ( schneider et al .
1992 ; schmidt et al . 1995 ; omont et al .
surdej et al .
( 1993 ) , crampton , mcclure & fletcher ( 1992 ) and kochanek ( 1993 ) reported that it is not gravitationally lensed .
the only study in the literature on the q2203 + 292 variability comes from mccarthy et al .
( 1988 ) who concluded from three lick 3 m plates and two plates from the 5 m hale telescope ( longair & gunn 1975 ; riley , longair & gunn 1980 ) that the quasar did not vary strongly over a period of the 15 years since its discovery .
a few additional broadband observations of q2203 + 292 are reported elsewhere , but they are in different photometric systems ( see table [ table_literature ] ) .
turner ( 1991 ) calculated for q2203 + 292 a mass of the central black hole of ( 0.69 - 9.6)@xmath610@xmath7 m@xmath3 and a mass accretion rate of 1.6 - 22 m@xmath3/yr , depending on the adopted cosmological model .
this values are typical for the quasars at the same redshift ( turner 1991 ; dietrich & hamann 2004 ) .
.broad band photometry of q2203 + 292 since 1987 .
[ cols="^,^,^,^ " , ]
the quasar shows ( fig . [ fig5 ] ) a brightness increase of @xmath20.75 mag at the beginning of the coverage but it is nearly constant later . due to the gaps in the lightcurve , any non - linear fluctuations ( such as flares ) can not be ruled out .
we verify the variability properties of the q2203 + 292 by means of differential photometry and monte carlo simulation .
the differential light curves ( fig . [ fig2 ] ) of the quasar were generated with respect to the reference stars a and b , two of the most stable stars near the quasar .
the rms of the relative light curve a b is 0.035 mag , and since the two stars have similar magnitudes , their individual errors are @xmath20.025 mag .
the rms for the quasar light curve with respect to star a ( marked as q a ) is 0.083 and with respect to star b ( marked as q a ) is 0.086 mag .
the maximum peak - to - peak variation of the quasar is @xmath20.92 mag over the entire monitoring period ( table .
[ table_uniflightcurge ] ) but it is reduced to 0.21 mag if we consider only the rozhen observations .
the rms of all 18 qso measurements is 0.20 mag , and if only our 14 measurements are considered , it decreases to 0.08 mag . to test further the variability of q2203 + 292
, we carried out a monte carlo simulation drawing 18 measurements from a constant source with the measured mean magnitude of q2203 + 292 .
each of these points was generated from a gaussian distribution with the observational error of the corresponding measurement , so that the artificial datasets more faithfully represent the properties of the real observations .
if we consider all data , including the ones from the literature , none of one million simulated data sets exceeded the observed rms however , the colour transformations can be a source of extra uncertainty , so we carried the same simulation only for our 14 measurements to obtained that in 98.5 per cent of the cases the data are inconsistent with a constant source . excluding the vlt point lowers this probability down to 86.4 per cent .
we conclude that if the colour transformation of the historical observations can be considered reliable , the quasar have undergone a brightening episode in the past but the unaccounted systematic effects stop is from making a strong statement about this .
all the literature data consistently deviate from ours in one direction , albeit by different amount , hinting that the variation may be real .
next , our own data show that during the recent years the quasar is in relatively stable state . the @xmath8 band in the rest - frame of the quasar correspond to the uv flux between 970 and 1420 , including the ly@xmath4 emission line .
to compare the variability properties of q2203 + 292 with those of lower redshift quasars , we calculated the structure function s(@xmath9 ) , which is commonly used to characterize the variability of large quasar samples ( i.e. hughes , aller & aller 1992 ) : @xmath10 ^ 2 \rangle.\ ] ] here , @xmath11 is the magnitude at time @xmath12 and @xmath9 is the time interval between the two measurements in the qso rest frame .
the broken brackets express ensemble average over measurements with the same time interval .
the structure function is less sensitive to the inhomogeneity of the observational coverage , and it can be applied to both individual objects and to samples of objects .
note that sometimes in the literature the structure function is defined as a square root of the ensemble average .
the structure function for q2203 + 292 is shown in fig .
the small number of measurements that form each bin lead to large uncertainties making it difficult to draw definite conclusions . the overall shape of s(@xmath9 ) for q2203 + 292 is similar to that of other qsos studied in the literature ( i.e. vanden berk et al .
it is dominated by observational errors for short time intervals and by the intrinsic qso variability for the longer ones .
there is indication that s(@xmath9 ) may reach a plateau at time scale just above 1 yr .
however , quasars are known to vary on a much longer time - scale , so we interpret this as poor sampling .
note that the structure functions of some qsos may show intrinsic structure in s(@xmath9 ) that is often interpreted as variability driven by more than one physical mechanism ( see hughes et al .
1992 for examples ) .
we can not exclude that this can be the case with q2203 + 292 .
further observations over longer time span are necessary to address this question . a quick comparisons with the literature , typically for @[email protected] - 2 yr , shows that q2203 + 292 is relatively more variable than the average qso in the samples of cristiani et al .
( 1996 ) and wold et al .
( 2007 ) and comparable with those of di clemente et al .
( 1996 ) and de vries et al .
( 2003 ) .
recently , wold et al .
( 2007 ) explored the dependence of the quasar variability from the mass of the central black hole . for @xmath13=10@xmath1410@xmath15 m@xmath3 as estimated for q2203 + 292 by turner ( 1991 ) they obtain @xmath0-band amplitude of 0.3 - 0.4 mag , which is similar to the value reported here . to calculate the absolute luminosity of q2203 + 292 , we adopted the following cosmological parameters : @xmath16=0.7 , @xmath17=0.3 , and @xmath18=70 km s@xmath19 mpc@xmath19 .
the absolute @xmath0-band magnitude , @xmath20 , is related to the apparent magnitude , @xmath0 , by @xmath21 where @xmath22 is galactic absorption , and @xmath23 is the luminosity distance for a flat universe .
the k - correction @xmath24 is calculated by convolving a quasar spectrum with a sensitivity curve for a standard @xmath25 filter ( fig .
[ fig7 ] ) .
we used our composite spectrum ( see sec .
2.4 ) for the sdss quasars observed in @xmath26 ( schneider et al . 2005 ) and the q2203 + 292 spectrum from constantin et al .
( 2002 ) for our target . in the latter case @xmath27=0.25 mag , 0.06 mag smaller than the value derived for @xmath1=4.40 from the composite spectrum .
although this is a systematic rather than random error , we added this difference in quadrature to the @xmath20 uncertainty ( equal to the rms of all q2203 + 292 measurements ) to obtain a conservative error estimate . assuming @xmath22=0.25 mag ( schlegel , finkbeiner & davis 1998 ) and using the average @xmath8=20.46 mag we obtained @[email protected] mag , typical for the sdss quasar distribution at that redshift ( fig .
[ fig8 ] ) .
the minimum mass of the emitting gas in the quasar broad line region ( blr ) can be estimated following baldwin et al .
( 2003 ) , assuming case b recombination , when all ly@xmath4 photons escape and the electron temperature is @xmath29=20,000 k : @xmath30 where @xmath31 is the ly@xmath4 luminosity and @xmath32 is the electron density .
we measured @xmath31 from the spectra of schneider et al .
( 1989 ) and constantin et al .
( 2002 ) : 8.6@xmath610@xmath33 and 5.6@xmath610@xmath33 erg s@xmath19 , respectively .
naturally , these are only lower limits because of the ly@xmath4 self - absorption . in both cases , we fitted the quasar continuum with a power law @xmath34 , and fixing @[email protected] . assuming @xmath32=10@xmath35 @xmath36
, we obtain @xmath37=0.44 and 0.29 m@xmath3 for the measurements from the two spectra .
narrow band ly@xmath4 imaging down to 25.5 mag per sq .
arcsec ( mccarthy et al . 1988 ) yielded no other emission line sources at the same redshift within 2@xmath62 arcmin@xmath38 from q2203 + 292 .
thompson , djorgovski & beckwith ( 1994 ) failed to find associated [ oiii ] emitters in [email protected] arcsec@xmath38 field centred at the quasar .
we used the photometricsat200a camera at the 2 m telescope at the rozhen observatory to carry out a search for associated emission line objects on 2006 august 19 .
the observations were obtained through a narrow band ( fwhm=32 ) interference filter @xmath39 centred at 6572 , corresponding to ly@xmath4 at @[email protected] .
our field of view was 5 arcmin@xmath38 , which is much bigger than that in the earlier studies .
the exposure time was 2 h , which was split into six 1200 s exposures .
we found no evidence for sources with emission lines falling into the bandpass of our narrow band filter down to a surface brightness level of @xmath224.5 mag per sq .
arcsec , in agreement with the previous attempts .
we carried out multi - year photometric @xmath8-band monitoring of the @xmath1=4.40 radio quiet quasar q2203 + 292 , and we found that it exhibits maximum peak - to - peak difference between two points on the light curve of @xmath20.3 mag for our data and @xmath20.9 mag when combined with older literature data .
the rms amplitude of the lightcurve is 0.08 mag and 0.20 mag , respectively .
the detected variability is at @xmath23@xmath40 level when the photometric accuracy of the both data sets are taken into account .
the monte carlo simulation can not reproduce the observed variation with a constant source in 10@xmath41 simulations , if we consider all the data but it does in 2.5 per cent of the simulations if we exclude the literature data and in 13.6 per cent is we consider only the rozhen and sao observations .
these results lead us to the conclusion that during the recent years the quasar is in a stable state but we refrain from making a strong statement about the earlier variability because of possible unaccounted systematic effects in the transformation between the different photometric systems .
unlike previous works , which used large samples of quasars to determine their variability properties , our goal was to assemble a well sampled light curve of individual quasars .
the structure function analysis concluded that the object shows variability properties similar to those of the lower redshift quasars .
we also found that narrow - band imaging at the redshifted ly@xmath4 shows no other emission line objects within 5@xmath65 arcmin@xmath38 field .
this work was partially supported by the following grants : @xmath42 , @xmath43 , @xmath44 of the bulgarian science foundation and @xmath45 with the sofia university .
the authors thank dr .
david crampton for giving us access to unpublished data and to dr .
anca constantin for providing the q2203 + 292 spectrum and the median composite spectrum of high redshift quasars .
we thank dr .
haralambi markov for his help with the spatial correction of the instrumental magnitudes , dr .
tanyu bonev for his assistance with the forero2 observations and dr .
ilia roussev for editing the manuscript .
we also wish to thank the anonymous referee for the useful comments that helped to improve the paper greatly . | we report cousins @xmath0-band monitoring of the high - redshift ( @xmath1=4.40 ) radio quiet quasar q2203 + 292 from may 1999 to october 2007 .
the quasar shows maximum peak - to - peak light curve amplitude of @xmath20.3 mag during the time of our monitoring , and @xmath20.9 mag when combined with older literature data .
the rms of a fit to the light curve with a constant is 0.08 mag and 0.2 mag , respectively .
the detected changes are at @xmath23-sigma level .
the quasar was in a stable state during the recent years and it might have undergone a brightening event in the past .
the structure function analysis concluded that the object shows variability properties similar to those of the lower redshift quasars .
we set a lower limit to the q2203 + 292 broad line region mass of 0.3 - 0.4 m@xmath3 .
narrow - band imaging search for redshifted ly@xmath4 from other emission line objects at the same redshift shows no emission line objects in the quasar vicinity .
[ firstpage ] quasar : general - quasars : individual : q2203 + 292 |
recently , there has been a lot of interest in quantitative estimates for isoperimetric @xcite , sobolev @xcite , gagliardo - nirenberg @xcite , and brunn - minkowski @xcite type inequalities .
the aim of all of these results is to show that if a set / function almost attains the equality in one of these inequalities , then it is close ( in a quantitative way ) to a minimizer .
these results have natural applications in the study of the asymptotic behavior of solutions to evolution equations @xcite , and to show stability for minimizers of perturbed variational problems , see for instance @xcite .
our goal is to investigate stability for the relative isoperimetric inequality inside convex cones .
this inequality has been used , for instance , to characterize isoperimetric regions inside convex polytopes for small volumes ( * ? ? ?
* corollary 3 ) .
hence , as in @xcite , one may use our stability result to prove quantitative closeness to such isoperimetric regions in perturbed situations .
let @xmath2 , and @xmath3 be an open , convex cone .
we denote the unit ball in @xmath4 centered at the origin by @xmath5 ( similar notation is used for a generic ball ) and the de giorgi perimeter of @xmath6 relative to @xmath7 by @xmath8 the relative isoperimetric inequality for convex cones states that if @xmath9 is a borel set with finite lebesgue measure @xmath10 , then @xmath11 if @xmath6 has a smooth boundary , the perimeter of @xmath6 is simply the @xmath12-hausdorff measure of the boundary of @xmath6 inside the cone ( i.e. @xmath13 ) .
we also note that if one replaces @xmath7 by @xmath4 , then the above inequality reduces to the classical isoperimetric inequality for which there are many different proofs and formulations ( see e.g. @xcite , @xcite , @xcite , @xcite , @xcite , @xcite )
. however , ( [ i1 ] ) is ultimately due to lions and pacella @xcite ( see also @xcite for a different proof using secord order variations ) .
their proof is based on the brunn - minkowski inequality which states that if @xmath14 are measurable , then @xmath15 ) can be seen as an immediate corollary of the anisotropic isoperimetric inequality .
this fact suggested to us that there should also be a direct proof of ( [ i1 ] ) using optimal transport theory ( see theorem [ 7 ] ) , as is the case for the anisotropic isoperimetric inequality @xcite .
the aim of this paper ( in the spirit of @xcite ) is to exploit such a proof in order to establish a quantitative version of ( [ i1 ] ) . to make this precise , we need some more notation .
we define the _ relative isoperimetric deficit _ of a borel set @xmath6 by @xmath16 note that ( [ i1 ] ) implies @xmath17 .
the equality cases were considered in @xcite for the special case when @xmath18 is smooth ( see also @xcite ) .
we will work out the general case in theorem [ 7 ] with a self - contained proof .
however , the ( nontrivial ) equality case is not needed in proving the following theorem ( which is in any case a much stronger statement ) : [ t0 ] let @xmath3 be an open , convex cone containing no lines , @xmath19 , and @xmath9 a set of finite perimeter with @xmath20 .
suppose @xmath21 satisfies @xmath22 .
then there exists a constant @xmath23 such that @xmath24 the assumption that @xmath7 contains no lines is crucial . to see this ,
consider the extreme case when @xmath25 .
let @xmath26 be any unit vector and set @xmath27 so that @xmath28 . however , @xmath29 so that in this case theorem [ t0 ] can only be true up to a translation , and this is precisely the main result in @xcite and @xcite . similar reasoning can be applied to the case when @xmath7 is a proper convex cone containing a line ( e.g. a half space ) .
indeed , if @xmath7 contains a line , then by convexity one can show that ( up to a change of coordinates ) it is of the form @xmath30 , with @xmath31 an open , convex cone .
therefore , by taking @xmath6 to be a translated version of @xmath32 along the first coordinate , the symmetric difference will be positive , whereas the relative deficit will remain @xmath33 . in general ( up to a change of coordinates )
, every convex cone is of the form @xmath34 , where @xmath35 is a convex cone containing no lines . indeed ,
theorem [ t0 ] follows from our main result : [ t1 ] let @xmath36 , where @xmath37 and @xmath35 is an open , convex cone containing no lines .
set @xmath19 , and let @xmath9 be a set of finite perimeter with @xmath20 .
suppose @xmath21 satisfies @xmath22 .
then there exists a constant @xmath23 such that @xmath38 let us remark that if @xmath39 , then @xmath25 and the theorem reduces to the main result of @xcite , the only difference being that here we do not attempt to find any explicit upper bound on the constant @xmath40 .
however , since all of our arguments are constructive , " it is possible to find explicit upper bounds on @xmath40 in terms on @xmath41 and the geometry of @xmath7 ( see also section 1.4 ) .
as we will show below , our result is strictly related to the quantitative version of the anisotropic isoperimetric inequality proved in @xcite . to show this link ,
we first introduce some more notation .
suppose @xmath32 is an open , bounded , convex set , and let , @xmath32 is assumed to contain the origin .
however , this is not needed ( see lemma [ 6 ] ) . ]
latexmath:[\[\label{2 } perimeter of a set @xmath6 of finite perimeter ( i.e. @xmath43 ) is defined as @xmath44 where @xmath45 is the reduced boundary of @xmath6 , and @xmath46 is the measure theoretic _
unit normal ( see section 2 ) . note that for @xmath47 , @xmath48 and @xmath49 for all @xmath50 . if @xmath6 has a smooth boundary , @xmath51 so that for @xmath52 we have @xmath53 in general , one can think of @xmath54 as a weight function on unit vectors .
indeed , @xmath55 has been used to model surface tensions in the study of equilibrium configurations of solid crystals with sufficiently small grains ( see e.g. @xcite , @xcite , @xcite ) and also in modeling surface energies in phase transitions ( see @xcite ) .
the anisotropic isoperimetric inequality states @xmath56 this estimate ( including equality cases ) is well known in the literature ( see e.g. @xcite ) . in particular , gromov @xcite uses certain properties of the knothe map from @xmath6 to @xmath32 in order to establish ( [ b2 ] ) .
however , as pointed out in @xcite and @xcite , the argument may be repeated verbatim if one uses the brenier map instead .
this approach leads to certain estimates which are helpful in proving a sharp stability theorem for ( [ b2 ] ) ( see ( * ? ? ?
* theorem 1.1 ) ) . using the anisotropic perimeter
, we now introduce the isoperimetric deficit of @xmath6 @xmath57 note that @xmath58 and @xmath59 for all @xmath47 and @xmath50 .
thanks to ( [ b2 ] ) and the associated equality cases , we have @xmath60 with equality if and only if @xmath6 is equal to @xmath32 ( up to a scaling and translation ) . note also the similarity between @xmath61 and @xmath62 .
indeed , they are both scaling invariant ; however , @xmath61 may not be translation invariant ( depending on @xmath7 ) .
we denote the asymmetry index of @xmath6 by @xmath63 the general stability problem consists of proving an estimate of the form @xmath64 where @xmath65 and @xmath66 . in the euclidean case ( i.e. @xmath52 )
, hall conjectured that ( [ b3 ] ) should hold with @xmath67 , and this was confirmed by fusco , maggi , pratelli @xcite . indeed , the @xmath68 exponent is sharp ( see e.g. ( * ? ? ?
* figure 4 ) ) .
their proof depends heavily on the full symmetry of the euclidean ball . for the general case when @xmath32 is a generic convex set ,
non - sharp results were obtained by esposito , fusco , trombetti @xcite , while the sharp estimate was recently obtained by figalli , maggi , pratelli @xcite .
their proof uses a technique based on optimal transport theory . for more information about the history of ( [ b3 ] ) , we refer the reader to @xcite and @xcite . we now provide a short sketch of the proof of theorem [ t1 ] for the case when @xmath69 and @xmath6 has a smooth boundary .
the first key observation is that the relative isoperimetric inequality inside a convex cone is a direct consequence of the anisotropic isoperimetric inequality with @xmath19 . indeed , as follows from the argument in section [ sect : isop ineq ] , @xmath70 , so follows immediately from .
this observation suggests that one may exploit gromov s argument in a similar way as in the proof of ( * ? ? ?
* theorem 1.1 ) to obtain additional information on @xmath6 .
indeed , we can show that there exists a vector @xmath71 such that such that ( [ 12r ] ) holds is exactly the main result in @xcite .
however , here we need to show that we can find a vector such that both ( [ 12o ] ) and ( [ 12r ] ) hold simultaneously . ]
@xmath72 @xmath73 , with @xmath74 and @xmath75 .
moreover , let @xmath76 . then using that @xmath77 , we obtain @xmath78 ; therefore , @xmath79 @xmath80 ( see corollary [ 8 ] ) , ( [ 12a ] ) and ( [ 12e ] ) hold with @xmath81 in place of @xmath82 ( see lemmas [ 15 ] & [ l0 ] ) .
thanks to ( [ 12e ] ) , we see that our result would readily follow if we can show @xmath83 ( see lemmas [ 17 ] & [ 18 ] ) , @xmath84 which , of course , implies theorem [ t1 ] .
therefore , we are left with proving ( [ t16 ] ) .
firstly , assume that @xmath81 and @xmath85 are sufficiently small ( i.e. smaller than a constant depending only on @xmath41 and @xmath32 ) . by ( [ 12a ] ) and the fact that ( see section @xmath86 ) @xmath87 we have @xmath88 but since @xmath85 is small , @xmath89 ; hence , @xmath90 moreover , thanks to the relative isoperimetric inequality inside @xmath91 ( see e.g. ( * ? ?
* inequality ( 3.43 ) ) ) , we have that for @xmath81 small enough , @xmath92 where in the last step we used that @xmath93 is close to @xmath94 ( see ( [ 12e ] ) ) and @xmath85 is small .
therefore , using ( [ 12e ] ) and ( [ t14 ] ) , @xmath95 since @xmath96 contains no lines , by some simple geometric considerations one may reduce the problem to the case when @xmath97 ( see lemma [ 19 ] ) , and then it is not difficult to prove @xmath98 which combined with ( [ t17 ] ) establishes ( [ t16 ] ) , and hence , the theorem .
next , we briefly discuss the assumptions for the sketch of the proof above .
indeed , one may remove the size assumption on @xmath85 by showing that if @xmath81 is small enough , then @xmath85 will be automatically small ( see proposition [ 29 ] ) . controls @xmath99 ; hence , we want to avoid any compactness argument .
however , even using compactness , we do not know any simple argument which shows that @xmath100 as @xmath101 .
] furthermore , we may freely assume that @xmath81 is small since if @xmath102 , then the theorem is trivial : @xmath103 the regularity of @xmath6 was used in order to apply the sobolev - poincar type estimate ( * ? ? ?
* lemma 3.1 ) which yields ( [ 12a ] ) ( see lemma [ 15 ] ) . if @xmath6 is a general set of finite perimeter in @xmath7 with finite mass and small relative deficit , then lemma [ 21a ] tells us that @xmath6 has a sufficiently regular subset @xmath104 so that @xmath105 and @xmath106 are controlled by @xmath81 .
combining this fact with the argument above yields the theorem for general sets of finite perimeter ( see proposition [ 21 ] ) .
lastly , the assumption on the mass of @xmath6 ( i.e. @xmath69 ) can be removed by a simple scaling argument .
we now discuss the sharpness of the estimate in theorem [ t1 ] .
indeed , it is well known that there exists a sequence of ellipsoids @xmath107 , symmetric with respect to the origin and converging to the ball @xmath5 , such that @xmath108 where @xmath109 ( see e.g. ( * ? ? ? *
consider the cone @xmath110 and set @xmath111 . by symmetry
, it follows that @xmath112 and @xmath113 .
we also note that @xmath114 @xmath115 therefore , @xmath116 and we have @xmath117 this example shows that the @xmath68 exponent in the theorem can not , in general , be replaced by something larger . one may wonder whether it is possible for theorem [ t1 ] to hold with a constant depending only on the dimension and not on the cone .
indeed , in ( * ? ? ?
* theorem 1.1 ) , the constant does not depend on the convex set associated to the anisotropic perimeter .
however , this is not so in our case . to see this ,
consider a sequence of open , symmetric cones in @xmath118 indexed by their opening @xmath119 .
let @xmath120 be a unit half - ball along the boundary of the cone @xmath121 disjoint from @xmath122 ( see figure 1 ) , where @xmath123 .
note that @xmath124 . therefore , @xmath125
endow the space @xmath126 of @xmath127 tensors with the metric @xmath128 , where @xmath129 denotes the transpose of @xmath130 .
let @xmath131 and denote the distributional derivative of @xmath132 by @xmath133 , i.e. @xmath133 is an @xmath134 matrix of measures @xmath135 in @xmath4 satisfying @xmath136 if @xmath137 is a borel set , then @xmath138 is the @xmath127 tensor whose entries are given by @xmath139 , and @xmath140 is the total variation of @xmath133 on @xmath141 with respect to the metric defined above , i.e. @xmath142 let @xmath143 be the set of all @xmath144 with @xmath145 .
for such a @xmath132 , decompose @xmath146 , where @xmath147 is the density with respect to the lebesgue measure and @xmath148 is the corresponding singular part .
denote the distributional divergence of @xmath132 by @xmath149 , and let @xmath150 .
then we have @xmath151 .
if @xmath133 is symmetric and positive definite , note that @xmath152 if @xmath6 is a set of finite perimeter in @xmath4 , then the reduced boundary @xmath153 of @xmath6 consists of all points @xmath154 such that @xmath155 for all @xmath156 and the following limit exists and belongs to @xmath157 : @xmath158 we call @xmath159 the measure theoretic _
unit normal to @xmath6 . by the well - known representation of the perimeter in terms of the hausdorff measure
, one has @xmath160 ( see e.g. ( * ? ? ?
* theorem 3.61 ) and ( * ? ? ?
* equation ( 3.62 ) ) ) .
this fact along with one of the equality cases in ( [ i1 ] )
@xmath161 yields the following useful representation of the relative deficit ( recall that @xmath21 satisfies @xmath22 ) : @xmath162 next , if @xmath163 , then for @xmath164- a.e .
@xmath165 there exists an inner trace vector @xmath166 ( see ( * ? ? ?
* theorem 3.77 ) ) which satisfies @xmath167 furthermore , @xmath168 denotes the set of points in @xmath4 having density @xmath169 with respect to @xmath6 ; i.e. @xmath170 means @xmath171
having developed the necessary notation , we are ready to state the following general version of the divergence theorem ( see e.g. ( * ? ? ?
* theorem 3.84 ) ) which will help us prove the isoperimetric inequality for convex cones ( i.e. theorem [ 7 ] ) : @xmath172 now we develop a few more tools that will be used throughout the paper .
fix @xmath173 , and let @xmath174 to apply the techniques in @xcite , we need a convex set that contains the origin .
therefore , let us translate @xmath32 by the vector @xmath175 which minimizes the ratio @xmath176 , where @xmath177 , @xmath178 and @xmath179 is defined as in ( [ 2 ] ) .
next , we introduce the _ minkowski gauge _ associated to the convex set @xmath180 : latexmath:[\[\label{1 } note that the convexity of @xmath180 implies the triangle inequality for @xmath182 so that it behaves sort of like a norm ; however , if @xmath180 is not symmetric with respect to the origin , @xmath183 .
hence , this
norm " is in general not a true norm .
nevertheless , the following estimates relate this quantity with the standard euclidean norm @xmath184 ( see ( * ? ? ?
* equations ( 3.2 ) and ( 3.9 ) ) ) : @xmath185 @xmath186 is scaling and translation invariant in its argument . the next lemma states that it is also translation invariant in @xmath32 ( observe that if @xmath187 does not contain the origin , then @xmath188 can also be negative in some direction ) . [ 6 ] let @xmath189 .
then @xmath190 for all @xmath191 .
it suffices to prove @xmath192 .
@xmath193 by using the divergence theorem for sets of finite perimeter ( * ? ? ?
* equation ( 3.47 ) ) , we obtain @xmath194 which proves the result . here
we show how to use gromov s argument to prove the relative isoperimetric inequality for convex cones .
as discussed in the introduction , the first general proof of the inequality was due to lions and pacella @xcite and is based on the brunn - minkowski inequality .
the equality cases were considered in @xcite for the special case when @xmath18 is smooth .
our proof of the inequality closely follows the proof of ( * ? ? ?
* theorem @xmath195 ) with some minor modifications .
[ 7 ] let @xmath7 be an open , convex cone and @xmath196
. then @xmath197 [ 7aaa ] moreover , if @xmath7 contains no lines , then equality holds if and only if @xmath198 .
* proof of ( [ revised ] ) . * by rescaling , if necessary , we may assume that @xmath199 ( i.e. @xmath200 ) .
define the probability densities @xmath201 and @xmath202 . by classical results in optimal transport theory ,
it is well known that there exists an a.e .
unique map @xmath203 ( which we call the brenier map ) such that @xmath204 where @xmath205 is convex , @xmath206 and @xmath207 for a.e .
@xmath208 ( see e.g. @xcite ) . moreover , since @xmath132 is the gradient of a convex function with positive jacobian , @xmath209 is symmetric and nonnegative definite ; hence , its eigenvalues @xmath210 are nonnegative for a.e .
@xmath211 . as a result
, we may apply the arithmetic - geometric mean inequality to conclude that for a.e .
@xmath208 , @xmath212 therefore , @xmath213 where we recall that @xmath168 denotes the set of points with density @xmath169 ( see section 2.1 ) .
next , we use ( [ las1 ] ) and ( [ 7b ] ) : @xmath214 by the convexity of @xmath32 and the fact that @xmath215 for a.e .
@xmath208 , it follows that @xmath216 , so by the definition of @xmath217 , @xmath218 hence , @xmath219 furthermore , note that if @xmath220 , then @xmath221 ; therefore , @xmath222 moreover , observe that by the definition of @xmath223 , it follows easily that @xmath224 for @xmath164 -a.e .
@xmath225 ; therefore , @xmath226 for @xmath164 -a.e . @xmath227 .
thus , @xmath228 and this proves the inequality . * equality case . * if @xmath229 , then @xmath230 and it is easy to check that equality holds in each of the inequalities above .
conversely , suppose there is equality .
in particular , @xmath231 . by @xcite ( see also ( * ? ? ?
* theorem a.1 ) ) , we obtain that @xmath232 with @xmath233 .
next , we will use the following identity which is valid for any @xmath234 : @xmath235 where @xmath236 by the previous equality case , we know @xmath237 , therefore , @xmath238 and we obtain @xmath239 .
this implies @xmath240 for @xmath164-almost every @xmath241 .
hence , @xmath242 in the direction defined by @xmath243 , which gives @xmath244 for all @xmath245
. however , by assumption , @xmath7 contains no lines ; thus , @xmath246 and we conclude the proof . [ 8 ] if @xmath189 , then @xmath247 .
since the inequality is scaling invariant , we may assume that @xmath69 . from ( [ las2 ] ) and
the fact that @xmath248 we obtain @xmath249 dividing by @xmath250 and using the representation of @xmath81 given by ( [ fin20 ] ) yields the result .
[ 13 ] let @xmath189 with @xmath69 , and let @xmath251 be the brenier map from @xmath6 to @xmath180 .
then @xmath252 let @xmath253 be the brenier map from @xmath6 to @xmath32 so that @xmath132 is the a.e .
unique gradient of a convex function @xmath205 .
then @xmath254 ( this follows easily from the fact that @xmath255 and @xmath256 is still convex ) .
therefore , by and @xmath257 , @xmath258 next , we recall from the proof of theorem [ 7 ] that @xmath259 .
hence , @xmath260 for @xmath164 a.e . @xmath261 and @xmath262 for @xmath164 a.e .
@xmath263 . therefore , using ( [ 13a ] ) , @xmath264 the fact that @xmath265 finishes the proof .
we split the proof in several steps . in section @xmath266 , we collect some useful technical tools . then in section @xmath267 , we prove theorem [ t1 ] under the additional assumption that @xmath6 is close to @xmath32 ( up to a translation in the first @xmath268 coordinates ) . finally , we remove this assumption in section @xmath269 to conclude the proof of the theorem .
let @xmath270 be the standard orthonormal basis for @xmath4 . recall that @xmath77 , where @xmath271 is an open , convex cone containing no lines .
hence , up to a change of coordinates , we may assume without loss generality that @xmath272 with this in mind and a simple compactness argument , we note that @xmath273 where @xmath274 is the ball in @xmath275 .
next , we introduce the trace constant of a set of finite perimeter .
recall the definition of @xmath180 given in section @xmath276 , so that ( [ 10 ] ) and ( [ 10a ] ) hold .
given a set @xmath189 , let @xmath277 denote the trace constant of @xmath6 , where @xmath278 note that @xmath279 is scaling invariant , and in general @xmath280 .
the trace constant contains valuable information about the geometry of @xmath6 .
for example , if @xmath6 has multiple connected components or outward cusps , then @xmath281 . in general , sets for which @xmath282 enjoy a nontrivial sobolev - poincar type inequality ( see ( * ?
* lemma 3.1 ) ) . in
what follows , we list all the technical tools needed in order to prove theorem [ t1 ] .
we decided to move some of the proofs to the appendix in order to make this section more accessible .
the following two lemmas are general facts about sets of finite perimeter .
[ 17 ] let @xmath283 be a bounded set of finite perimeter .
then there exists @xmath284 such that for any @xmath285 , @xmath286 .
[ 18 ] let @xmath283 be a bounded set of finite perimeter .
then there exist two constants @xmath287 such that if @xmath288 , then @xmath289 lemma [ 17 ] is well - known , and follows by applying ( * ? ? ?
* remark 3.25 ) to @xmath290 .
also lemma [ 18 ] should be known , but we have been unable to find a reference .
therefore , we provide a proof in the appendix .
[ geo ] there exists a bounded , convex set @xmath291 so that for all @xmath292 with @xmath293 , we have @xmath294 [ geo1 ] furthermore , if @xmath295 , then @xmath296 we will show that one may pick @xmath297 small enough so that @xmath298 has the desired properties .
we will establish ( [ geo1 ] ) first .
since @xmath299 , it suffices to prove @xmath300 . if ( for contradiction ) there exists @xmath301 , then @xmath302 and @xmath303 . since @xmath302 and @xmath293 , it follows that @xmath304 .
also , @xmath305 for @xmath306 .
now , @xmath307 , hence , @xmath308 . let @xmath309 be the constant from ( [ 19z ] ) , and assume without loss of generality that @xmath310 . if @xmath311 is such that @xmath312 , where @xmath313 , then @xmath314 ( see figure 2 ) .
let @xmath315 , @xmath316 , and recall that @xmath308 .
since @xmath96 is a cone , we have @xmath317 with @xmath318 .
hence , @xmath319 , but since @xmath320 we obtain @xmath321 , where @xmath322 denotes the ball in dimension @xmath323 .
therefore , @xmath324 next , pick @xmath325 so that @xmath326 . thus , by letting @xmath327 , we obtain @xmath328 .
therefore , we conclude @xmath329 , a contradiction . hence , ( [ geo1 ] ) is established . since @xmath330 and @xmath331 , ( [ geo2 ] ) follows from ( [ geo1 ] ) .
the next lemma ( whose proof is postponed to the appendix ) tells us that a set with finite mass , perimeter , and small relative deficit has a subset with almost the same mass , good trace constant , and small relative deficit ( compare with ( * ? ? ?
* theorem 3.4 ) ) .
[ 21a ] let @xmath189 with @xmath332 then there exists a set of finite perimeter @xmath333 and constants @xmath334 such that if @xmath335 then @xmath336 @xmath337 the big advantage of using @xmath104 in place of @xmath6 is that ( [ 24 ] ) implies a nontrivial trace inequality for @xmath104 which allows us to exploit gromov s proof in order to prove ( [ 12o ] ) with @xmath104 in place of @xmath6 . indeed ,
if @xmath6 is smooth with a uniform lipschitz bound on @xmath338 , one may take @xmath339 .
[ 15 ] let @xmath189 , @xmath69 , and assume @xmath340 , with @xmath341 and @xmath342 as in lemma [ 21a ] .
moreover , let @xmath156 satisfy @xmath343 .
then there exists @xmath344 and a constant @xmath345 such that @xmath346 .
let @xmath347 be the brenier map from @xmath348 to @xmath180 , and denote by @xmath349 the @xmath350 component of @xmath351 . for all @xmath352 , we apply ( * ? ? ?
* lemma @xmath266 ) to the function @xmath349 and the set @xmath348 to obtain a vector @xmath353 such that @xmath354 where we have used ( [ 10 ] ) in the second inequality
. next , recall that @xmath279 is scaling invariant .
hence , using ( [ 24 ] ) we have @xmath355 but by ( * ? ? ?
* corollary @xmath356 ) and corollary [ 8 ] , @xmath357 therefore , by summing over @xmath358 we obtain @xmath359 let @xmath360 , with @xmath361 as in the definition of @xmath180 ( see ( [ 9 ] ) ) .
the triangle inequality implies @xmath362 hence , by corollary [ 13 ] , ( [ 24aa ] ) , and ( [ 24d ] ) we have @xmath363 as @xmath364 , the result follows .
the translation @xmath365 from lemma [ 15 ] can be scaled so that it enjoys some nice properties which we list in the next lemma .
the proof is essentially the same as that of ( * ? ? ?
* theorem 1.1 ) , adapted slightly in order to accommodate our setup .
however , for the sake of completion , we include it in the appendix .
[ l0 ] suppose @xmath189 with @xmath69 .
let @xmath366 , @xmath104 , and @xmath367 be as in lemma [ 15 ] .
define @xmath368 .
then there exists a positive constant @xmath369 such that for @xmath340 , with @xmath342 as in lemma [ 21a ] , we have @xmath370 next , define @xmath371 ( @xmath372 is defined in a similar manner ) . in the case
@xmath373 , the following lemma tells us that the last @xmath374 components of @xmath375 are controlled by the relative deficit . [ 19 ] let @xmath189 with @xmath69 , and let @xmath376 be as in lemma [ l0 ] . there exist positive constants @xmath377 such that if @xmath378 and @xmath379 , then @xmath380 .
let @xmath381 be the bounded , convex set given by lemma [ geo ] .
an application of lemma @xmath382 and ( [ geo2 ] ) yields @xmath383 now , note that @xmath384 ; hence , by using this fact and ( [ 16a ] ) we obtain @xmath385 therefore , there exists @xmath386 such that for @xmath379 , @xmath387 thus , the result follows with @xmath388 ( note that @xmath32 completely determines @xmath389 ) . [ 21 ] let @xmath189 with @xmath69 , and let @xmath391 be as in lemma [ l0 ] .
then there exist positive constants @xmath392 , @xmath393 , and @xmath394 such that if @xmath395 and @xmath396 , then @xmath397 .
thanks to lemma @xmath398 , we may assume without loss of generality that @xmath399 .
let @xmath400 , with @xmath104 as in lemma [ 21a ] and @xmath156 such that @xmath343 . by lemma
[ 15 ] and the fact that @xmath77 , @xmath401 however , thanks to ( [ l3 ] ) , @xmath402 , so for @xmath85 and @xmath81 sufficiently small we have @xmath403 , and this implies @xmath404 next , by using the relative isoperimetric inequality ( apply ( * ? ? ?
* inequality ( 3.41 ) ) to @xmath405 and the set @xmath91 ) , we have that for @xmath81 small enough , @xmath406 furthermore , @xmath407 and by using ( [ l2 ] ) , lemma [ 17 ] , and ( [ l3 ] ) , @xmath408 therefore , we can select @xmath409 such that if @xmath395 and @xmath396 , then @xmath410 thus , using ( [ 21e ] ) we obtain @xmath411 hence , by ( [ 21h ] ) , ( [ l2 ] ) , and lemma [ 17 ] it follows that @xmath412 but @xmath413 , and from ( [ l3 ] ) it readily follows that @xmath414 combining this fact with ( [ zz2 ] ) yields a positive constant @xmath415 such that @xmath416 next , let @xmath417 be the bounded , convex set given by lemma [ geo ] . we note that since @xmath399 , ( [ geo1 ] ) implies @xmath418 therefore , using lemma
[ 18 ] and ( [ z2 ] ) we have @xmath419 thus , for @xmath392 sufficiently small we can take @xmath420 to conclude the proof . [ 22 ] let @xmath189 with @xmath69 , @xmath392 and @xmath393 be as in proposition [ 21 ] , and @xmath421 be as in lemma [ l0 ]
. then there exists a positive constant @xmath422 such that if @xmath395 and @xmath396 , then @xmath423 note that by proposition [ 21 ] we obtain @xmath397 .
next , by applying lemma [ 17 ] and ( [ 16a ] ) we have @xmath424 therefore , we may take @xmath425 to conclude the proof . in proposition
[ 29 ] below , we refine corollary [ 22 ] .
namely , we show that if @xmath81 is small enough , then the assumption on the size of @xmath426 is superfluous .
however , to prove proposition [ 29 ] we need to reduce the problem to the case when @xmath427 ( recall @xmath428 ) .
this is the content of lemma [ 20 ] . for arbitrary @xmath429 ,
decompose @xmath430 as @xmath431 where @xmath432 is the closest point on the boundary of the cone @xmath96 to @xmath430 and @xmath433 ( see figure 3 ) .
note that @xmath434 is perpendicular to @xmath435 .
[ 20 ] let @xmath189 with @xmath69 , and let @xmath391 be as in lemma [ l0 ] . there exist constants
@xmath436 such that if @xmath437 and @xmath438 , then @xmath439 .
firstly , observe that @xmath440 since @xmath441 , it follows that @xmath442 has a nontrivial intersection with @xmath443 .
let @xmath444 and note that , by convexity , @xmath445 next , pick @xmath446 .
observe that @xmath367 is the smallest radius for which @xmath447 so that it contains some @xmath448 ( see figure 3 ) . since @xmath7 is convex
, there exists a constant @xmath449 such that @xmath450 . but @xmath451 .
therefore , @xmath452 , and since @xmath446 we have that @xmath453 .
[ 29 ] let @xmath189 with @xmath69 , and let @xmath391 be as in lemma [ l0 ]
. then there exist @xmath454 such that if @xmath455 , then @xmath396 with @xmath393 as in proposition @xmath456 .
if @xmath373 , then the result follows from lemma [ 19 ] . if @xmath457 , then write @xmath458 as in ( [ 18a ] ) with the understanding that @xmath459 if and only if @xmath460 .
in the case where @xmath461 ( i.e. @xmath438 ) , thanks to lemma [ 20 ] , we have @xmath439 .
therefore , it suffices to prove that for @xmath81 sufficiently small , @xmath462 .
we split the proof into three steps .
the idea is as follows : firstly , we assume by contradiction that @xmath463 .
this allows us to translate @xmath6 by a suitable vector @xmath464 so that @xmath465 is a distance @xmath466 from the origin ( see figure 4 ) .
the second step consists of showing that up to a small mass adjustment , the relative deficit of this new set is controlled by @xmath467 .
lastly , we show that the new set satisfies the hypotheses of proposition [ 21 ] ; therefore , we conclude that it should be a lot closer to the origin than it actually is
. * step 1 . * assume for contradiction that @xmath468 .
select @xmath469 such that for @xmath470 we have @xmath471 by ( [ 16a ] ) , lemma [ 17 ] , and lemma [ 20 ] , @xmath472 next , we set @xmath473 and @xmath474 so that @xmath475 where @xmath476 is chosen to satisfy @xmath477 .
therefore , @xmath478 ) : using ( [ 29ca ] ) , @xmath479 but , @xmath480 , therefore , @xmath481 , and hence , @xmath482 . thus , combining the previous fact with ( [ 29b ] ) and ( [ 29c ] ) , @xmath483 and @xmath484 , we deduce @xmath485 so that for @xmath486 , by @xmath484 again and the fact that @xmath487 , @xmath488 since @xmath7 is a convex cone , it follows that @xmath489 and @xmath490 . thus , @xmath491 recall that @xmath492 and @xmath493 ( see section 2 ) . upon subtracting @xmath494 from both sides of ( [ 31c ] ) , dividing by @xmath250 ( recall @xmath495 , and using that @xmath496
we have @xmath497 let @xmath498 . then , assuming without loss of generality that @xmath499 , @xmath500 * step 3 .
* using lemma [ 17 ] and ( [ fin5 ] ) , for @xmath81 small enough we have @xmath501 next , we claim @xmath502 ) we deduce that @xmath503 but since @xmath504 , @xmath505 as before , @xmath506 ( since @xmath481 ) .
therefore , ( [ 31d ] ) and ( [ 31e ] ) imply the claim ( i.e. ( [ 31 ] ) ) .
furthermore , by using ( [ 30a ] ) and @xmath507 , it follows that for some constant @xmath508 , @xmath509 be the translation as in lemma [ l0 ] for the set @xmath510 , so that @xmath511 . by lemma @xmath382 and @xmath512 , @xmath513 moreover , ( [ 33 ] ) and ( [ dm ] )
imply that if @xmath81 is sufficiently small , then there exists a constant @xmath514 so that @xmath515 and @xmath516 )
. furthermore , using ( [ 33l ] ) and ( [ 33 ] ) , we deduce that for @xmath81 small enough @xmath517 where @xmath518 is as in proposition [ 21 ] .
thus , by applying proposition [ 21 ] to @xmath519 and using ( [ 33 ] ) again , it follows that @xmath520 ) , ( [ 33ne ] ) , and ( [ 34 ] ) we obtain @xmath521 which is impossible if @xmath81 is sufficiently small .
this concludes the proof .
we are now in a position to prove theorem [ t1 ] .
firstly , we assume that @xmath69 .
indeed , let @xmath518 and @xmath522 be the constants given in proposition [ 21 ] and [ 29 ] , respectively , and set @xmath523 .
if @xmath524 , then it follows from proposition [ 29 ] and corollary [ 22 ] that @xmath525 let @xmath526 and suppose now that @xmath527 .
pick @xmath528 such that @xmath529 and apply the previous estimate to the set @xmath530 to obtain @xmath531 and this implies @xmath532 since @xmath533 , this yields the theorem for the case when @xmath534 if @xmath535 , then @xmath103 therefore , we obtain the theorem with @xmath536 .
for @xmath537 let @xmath538 note that @xmath539 our strategy is as follows : first , we show that the incremental ratios of @xmath540 at @xmath541 have a positive lower bound .
then we prove that @xmath542 is uniformly bounded away from zero when @xmath430 is away from zero .
these two facts readily yield the result . +
* we claim there exists @xmath543 such that @xmath544 indeed , let @xmath545 be any sequence converging to @xmath33 . by lemma
[ 17 ] we know that the family of measure @xmath546 defined by @xmath547 satisfy @xmath548 . moreover , up to choosing a subsequence ( which we do not relabel ) so that @xmath549 for some @xmath550 , it is immediate to check that @xmath546 converge weakly to @xmath551 hence , by the lower semicontinuity of the total variation ( see for instance ( * ? ? ?
* corollary 1.60 ) ) , @xmath552 we now observe that , again by the lower semicontinuity of the total variation , the right hand side attains a minimum for some @xmath553 .
hence , by the arbitrariness of @xmath554 , @xmath555 to conclude it suffices to observe that @xmath556 , as otherwise @xmath557 for any @xmath558 ( up to sets of measure zero ) , which contradicts the boundedness of @xmath130 .
* step 2 . *
we claim that there exists @xmath559 such that @xmath560 with @xmath561 as in step 1 .
the proof is by compactness : if @xmath562 then @xmath563 . on the other hand , by continuity
@xmath540 attains a minimum over the compact set @xmath564 .
let @xmath565 be a vector where such a minimum is attained .
then this minimum can not be zero as otherwise @xmath566 ( up to a set of measure zero ) . by iterating the estimate
, this implies that @xmath567 for any @xmath568 , contradicting again the boundedness of @xmath130 .
let @xmath569 . if @xmath570 , then by ( * ? ? ?
* theorem 3.4 ) there exists a set of finite perimeter @xmath341 satisfying @xmath571 we claim that @xmath104 is the desired set . indeed , since @xmath572 ( see corollary [ 8 ] ) , ( [ lma ] ) and ( [ lmb ] ) yield ( [ 23 ] ) and ( [ 24 ] ) ; therefore , it remains to prove ( [ 24d ] ) . from the construction of @xmath104 in (
* theorem 3.4 ) , we have @xmath573 , where @xmath574 is the maximal element given by ( * ? ? ?
* lemma 3.2 ) that satisfies @xmath575 to prove ( [ 24d ] ) , we first claim that for some positive constant @xmath40 , @xmath576 note from the definitions that @xmath577 moreover , by ( * ? ? ?
* equation ( 2.10 ) ) and ( * ? ? ?
* equation ( 2.11 ) ) we may write @xmath578 therefore , @xmath579 next , we note that @xmath580 and by ( * ? ? ?
* lemma 2.2 ) , @xmath581 at @xmath164 a.e .
point of @xmath582 furthermore , taking into account ( [ 10a ] ) and ( [ 24e ] ) , we have @xmath583 hence , ( [ 24f ] ) and ( [ 24 g ] ) yield ( observe that @xmath584 on @xmath585 ) @xmath586 by the anisotropic isoperimetric inequality ( see ( * ? ? ?
* theorem 2.3 ) or ( [ las2 ] ) ) , @xmath587 moreover , by ( [ lma ] ) , @xmath588 combining ( [ 25d ] ) , ( [ fin6 ] ) , and ( [ 24h ] ) it follows that @xmath589 therefore ( recall @xmath572 and @xmath69 ) , @xmath590 using the definition of @xmath591 , ( [ 24 g ] ) , and ( [ 24i ] ) we obtain @xmath592 and this proves our claim ( i.e. ( [ 25 ] ) ) .
our next task is to use ( [ 25 ] ) in order to prove ( [ 24d ] ) , thereby finishing the proof of the lemma .
let @xmath156 be such that @xmath593 .
note that by ( [ 25 ] ) , @xmath594 but since @xmath595 and @xmath596 , we have @xmath597 so that two applications of ( [ fin7 ] ) yield @xmath598 and by using @xmath599 we have @xmath600 upon combining ( [ 28k ] ) and ( [ 28l ] ) , ( [ 24d ] ) follows easily
. recall that by definition @xmath601 , where @xmath569 . since @xmath572 , by taking @xmath340 , @xmath602 will be sufficiently small in order for us to assume the setup of ( * ? ? ?
* inequality ( 3.30 ) ) , with the understanding that the set @xmath32 in the equation corresponds to our @xmath180 , and whenever @xmath32 appears in our estimates , it is the same set that we defined in the introduction ( i.e. @xmath19 ) . note that in ( * ? ? ? * proof of theorem 1.1 ) the authors dilate the sets @xmath104 and @xmath6 by the same factor @xmath156 so that @xmath603 ; however , they still denote the resulting dilated sets by @xmath104 and @xmath6
. we will keep the scaling factor so that our @xmath348 and @xmath604 correspond , respectively , to their @xmath104 and @xmath6 . with this in mind , note that ( * ? ? ? * inequality ( 3.30 ) ) is valid up to a translation .
indeed , this translation is obtained by applying ( * ? ? ?
* lemma 3.1 ) to the functions @xmath349 and the set @xmath348 , where @xmath351 , and @xmath605 is the brenier map between @xmath348 and @xmath180 . since @xmath606 in lemma @xmath607 was obtained by the same exact process , @xmath243 satisfies ( * ? ? ?
* inequality ( 3.30 ) ) .
thus , by the estimates under ( * ? ? ?
* inequality ( 3.30 ) ) it follows that @xmath608 therefore , we have @xmath609 and this implies @xmath610 recalling that @xmath611 ( see ( [ 23 ] ) ) , @xmath69 , and @xmath81 is small , it readily follows that ( see ( [ 28a ] ) ) @xmath612 and we obtain ( [ l3 ] ) . also , ( [ 101 ] ) , ( [ 102 ] ) , ( [ rbs ] ) , and @xmath613 ( see ( [ 24d ] ) ) imply the existence of a positive constant @xmath40 so that @xmath614 ) and ( [ 104 ] ) imply @xmath615 by combining this together with ( [ 105 ] ) , we readily obtain ( [ 16a ] ) and ( [ l2 ] ) .
* acknowledgments . *
we wish to thank eric baer for his careful remarks on a preliminary version of the paper .
we are also grateful to an anonymous referee for his valuable comments and suggestions that significantly improved the presentation of the paper .
the first author was supported by nsf grant dms-@xmath616 .
the second author was supported by an nsf - rtg fellowship for graduate studies at the university of texas at austin .
any opinions , findings , and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the us national science foundation ( nsf ) .
99 a. alberti and l. ambrosio , a geometrical approach to monotone functions in @xmath4 , _ math . z. _ 230 ( 1999 ) , no . 2 , 259 - 316
. l. ambrosio , n. fusco , and d. pallara , functions of bounded variation and free discontinuity problems .
oxford mathematical monographs . the clarendon press , _ oxford university press _
, new york , 2000 .
g. bianchi and h. egnell , a note on the sobolev inequality , _
_ 100 ( 1991 ) , pp .
w. blaschke , kreis und kugel , berlin .
_ de gruyter _ , 2nd edn .
y. brenier , polar factorization and monotone rearrangement of vector - valued functions , _ comm .
pure appl .
_ 44 ( 4 ) ( 1991 ) 375 - 417 .
brothers and f. morgan , the isoperimetric theorem for general integrands , _ michigan math . j. _ 41
( 1994 ) , no .
3 , 419 - 431 .
burago and v.a .
zalgaller , geometric inequalities .
berlin : _ springer verlag _ ( 1988 ) .
original russian edition : geometricheskie neravenstva , leningrad .
i. chavel , isoperimetric inequalities .
cambridge tracts in mathematics , 145 , _ cambridge university press _ ( 2001 ) .
e. carlen and a. figalli , stability for a gns inequality and the log - hls inequality , with application to the critical mass keller - segel equation , preprint , 2011 .
a. cianchi , n. fusco , f. maggi , and a. pratelli , the sharp sobolev inequality in quantitative form , _ journal of the european mathematical society _ , ( 2009 ) 11 , n.5 , 1105 - 1139 .
d. cordero - erausquin , b. nazaret , and c. villani , a new approach to sharp sobolev and gagliardo - nirenberg inequalities , _ adv .
_ 182 , 2 ( 2004 ) , 307 - 332
. b. dacorogna and c.e .
pfister , wulff theorem and best constant in sobolev inequality , _ j. math .
pures appl . _
( 9 ) 71 ( 2 ) ( 1992 ) 97 - 118 .
a. dinghas , ber einen geometrischen satz von wulff fr die gleichgewichtsform von kristallen , _ ( german ) z. kristallogr . ,
_ 105 , ( 1944 )
. l. esposito , n. fusco , and c. trombetti , a quantitative version of the isoperimetric inequality : the anisotropic case , _ ann . sc .
( 5 ) 4 ( 2005 ) , no . 4 , 619 - 651 .
a. figalli , f. maggi , on the shape of liquid drops and crystals in the small mass regime , _ arch .
_ , to appear .
a. figalli , f. maggi , and a. pratelli , a mass transportation approach to quantitative isoperimetric inequalities , _ invent .
_ , 182 ( 2010 ) , no . 1 , 167 - 211
a. figalli , f. maggi , and a. pratelli , a refined brunn - minkowski inequality for convex sets , _ ann .
poincar anal .
non lineaire _ , 26 ( 2009 ) , no . 6 , 2511 - 2519 .
a. figalli , f. maggi , and a. pratelli , a note on cheeger sets , _ proc .
_ , 137 ( 2009 ) , 2057 - 2062 .
a. figalli , f. maggi , and a. pratelli , sharp stability theorems for the anisotropic sobolev and log - sobolev inequalities on functions of bounded variation .
preprint , 2011 .
i. fonseca and s. mller , a uniqueness proof for the wulff theorem , _ proc .
edinburgh sect .
_ a 119 ( 1991 ) , no .
1 - 2 , 125 - 136 .
f. morgan , riemannian geometry .
a beginner s guide .
second edition .
_ a. k. peters , ltd .
_ , wellesley , ma , 1998 .
f. morgan , the levy - gromov isoperimetric inequality in convex manifolds with boundary , _ j. geom .
_ 18 ( 2008 ) , no . 4 , 1053 - 1057
. n. fusco , the classical isoperimetric theorem , _ rend .
sci fis . mat .
vol 71 , ( 2004 ) , 63 - 107 .
n. fusco , f. maggi , and a. pratelli , the sharp quantitative isoperimetric inequality , _ ann . of math .
_ 168 ( 2008 ) , 941 - 980 .
gurtin , on a theory of phase transitions with interfacial energy , _ arch .
rational mech .
_ 87 ( 1985 ) , no . 3 , 187 - 212 . c. herring , some theorems on the free energies of crystal surfaces , _ phys .
_ 82 ( 1951 ) , 87 - 93 .
p. l. lions and f. pacella , isoperimetric inequalities for convex cones , _ proc .
_ 109 ( 1990 ) 477 - 485 .
f. maggi , some methods for studying stability in isoperimetric type problems , _ bull .
_ , 45 ( 2008 ) , 367 - 408 .
mccann , a convexity principle for interacting gases , _ adv .
_ 128 ( 1 ) ( 1997 ) 153 - 179 .
mccann , existence and uniqueness of monotone measure - preserving maps , _ duke math .
j_. 80 ( 2 ) ( 1995 ) 309 - 323 .
v.d . milman and g.
schechtman , asymptotic theory of finite - dimensional normed spaces . with an appendix by m. gromov .
lecture notes in mathematics , 1200 . _
springer - verlag _ , berlin , 1986 .
viii+156 pp .
muratov , droplet phases in non - local ginzburg - landau models with coulomb repulsion in two dimensions .
_ 299 ( 2010 ) , no . 1 , 45 - 87 . r. osserman , the isoperimetric inequality , _ bull .
_ , 84 ( 1978 ) , 1182 - 1238
. m. ritor and c. rosales , existence and characterization of regions minimizing perimeter under a volume constraint inside euclidean cones , _ trans .
_ 356 ( 2004 ) , no . 11 , 4601 - 4622 .
a. segal , remark on stability of brunn - minkowski and isoperimetric inequalities for convex bodies , preprint , 2011 .
g. talenti , the standard isoperimetric theorem , _ handbook of convex geometry _ , vol .
a , p.m. grueber and j.m .
willis , eds . , 73 - 123 , amsterdam , north holland ( 1993 ) .
taylor , crystalline variational problems , _ bull .
_ 84 ( 1978 ) , no . 4 , 568 - 588 . j. van schaftingen , anisotropic symmetrization , _ ann .
h. poincar anal .
non lineaire _ 23 ( 2006 ) , no . 4 , 539 - 565 . g. wulff , zur frage der geschwindigkeit des wachstums und der auflsung der kristallflchen , _ z. kristallogr .
_ 34 , 449 - 530 . | the relative isoperimetric inequality inside an open , convex cone @xmath0 states that , at fixed volume , @xmath1 minimizes the perimeter inside @xmath0 . starting from the observation that this result can be recovered as a corollary of the anisotropic isoperimetric inequality ,
we exploit a variant of gromov s proof of the classical isoperimetric inequality to prove a sharp stability result for the relative isoperimetric inequality inside @xmath0 .
our proof follows the line of reasoning in @xcite , though several new ideas are needed in order to deal with the lack of translation invariance in our problem . * a sharp stability result for the relative isoperimetric inequality inside convex cones * a. figalli and e. indrei |
the torque on an electric dipole * p * , at rest in a uniform electric field * e * , is @xmath0 the torque on a magnetic dipole * m * , at rest in a uniform magnetic field * b * , is @xmath1 but what if the dipole is moving , at a constant velocity * v * ?
it is well known that a moving electric dipole acquires a _
dipole moment @xcite @xmath2
so one might guess that the torque on a moving electric dipole in uniform electric and magnetic fields would be @xcite @xmath3 similarly , a moving magnetic dipole acquires an _ electric _ dipole moment @xcite @xmath4 and one might guess that the torque on a moving magnetic dipole would be @xmath5 but these formulas are incorrect ; in each case there is a third term:@xcite @xmath6 in general , then , @xmath7 our purpose in this note is to derive that result
. @xcite -.1in[.7 ] -.5 in
an electric dipole consists of positive and negative charges @xmath8 separated by a displacement @xmath9 ( figure 1 ) .
the force on a point charge is @xmath10 so the torque @xcite on a dipole of moment @xmath11 is @xmath12\nonumber\\ & = & q{\bf d}\times\left({\bf e}+{\bf v}\times { \bf b}\right)\nonumber\\ & = & ( { \bf p}\times { \bf e})+{\bf p}\times({\bf v}\times { \bf b}).\end{aligned}\ ] ] using the vector identity @xmath13 , the previous result becomes @xmath14 that is surely the _ easiest _ way to obtain eq .
( 7 ) ( it borrows from an argument by v. namias@xcite ) .
but the magnetic analog is not so straightforward,@xcite
so let s repeat the derivation , this time treating the electric dipole as the point limit of a uniformly polarized object.@xcite polarization ( * p * ) and magnetization ( * m * ) together constitute an antisymmetric second - rank tensor : @xmath15 and the transformation rule gives@xcite @xmath16 we use primes for the proper " ( rest ) system of the dipole ( @xmath17 ) ; no prime means the lab " frame ( @xmath18 ) , through which the dipole is moving with velocity @xmath19 . notice that if @xmath20 , then @xmath21 , confirming eq .
( 3 ) , and if @xmath22 , then @xmath23 , confirming eq . ( 5 ) . now , suppose we have an electric dipole @xmath24 at rest at the origin in @xmath25 : @xmath26 assume first that @xmath24 is perpendicular to * v * , say @xmath27 then @xmath28 and all other components are zero . according to the lorentz law , the force density is @xmath29 where @xcite @xmath30 in our case,@xcite @xmath31 @xmath32,\ ] ] @xmath33 so @xmath34 and @xmath35.\ ] ] the torque is @xmath36,\end{aligned}\ ] ] where @xmath37 .
the @xmath38 and @xmath39 components of the integral are zero ; the @xmath40 component is @xcite @xmath41 so @xmath42= ( { \bf p}_0\times { \bf e } ) + { \bf p}_0 \times ( { \bf v } \times { \bf b}),\ ] ] which agrees with eq . (
11 ) , and hence confirms eq .
( 7 ) for the case @xmath24 perpendicular to * v*. ( for this orientation the electric dipole moments in @xmath18 and @xmath25 are the same there is no lorentz contraction .
this can be confirmed by integrating eq . (
18 ) : @xmath43 . ) if @xmath24 is parallel to * v * , @xmath44 giving @xmath45 with all other components of * p * and * m * being zero . in this case ,
@xmath46 and @xmath47 so that @xmath48.\ ] ] the torque is @xmath49.\ ] ] because of lorentz contraction , @xmath50as is confirmed by integrating eq .
( 31)and again we recover eq .
now , suppose we have a magnetic dipole @xmath51 , at rest at the origin , in @xmath25 : @xmath52 assume first that @xmath51 is perpendicular to * v * , say @xmath53 from the transformation rule , @xmath54 this time @xmath55 @xmath56,\ ] ] @xmath57 so @xmath58,\ ] ] and @xmath59\nonumber\\ & & \quad\quad\quad\times { \bf b}.\end{aligned}\ ] ] the torque is @xmath60 in agreement with eq .
( 8) , for @xmath51 perpendicular to * v*. ( in this orientation , the magnetic dipole moments in @xmath18 and @xmath25 are the same , as we confirm by integrating eq .
( 39 ) : @xmath61 . ) if @xmath51 is parallel to * v * , @xmath62 giving @xmath63 with all other components of * p * and * m * being zero . in this case
@xmath64 and @xmath65 vanish , leaving @xmath66,\ ] ] so @xmath67\times{\bf b},\ ] ] and the torque is @xmath68 which is again consistent with eq .
( 8) because in this orientation @xmath69 and , integrating eq .
( 47 ) , @xmath70 .
in the presence of an electric field , a magnetic dipole ( a loop of electric current ) carries
hidden momentum"@xcite @xmath71 and hence also hidden momentum @xmath72 it is hidden " in the sense that it is not associated with overt motion of the object as a whole ; it occurs in systems with internally moving parts , such as current loops.@xcite the _ total _ angular momentum is the sum of overt " and hidden components : @xmath73 and the torque is its rate of change:@xcite @xmath74 thus , combining eqs .
( 9 ) and ( 53 ) , the overt " torque on a moving dipole is @xmath75 by contrast , a ( stationary ) _ electric _ dipole has no internally moving parts , and it harbors no hidden momentum.@xcite however , when it is in _ motion _ it picks up a magnetic dipole moment @xmath76 [ eq .
( 3 ) ] that _ does _ carry hidden momentum,@xcite so eq . ( 54 ) ( with @xmath77 ) applies to this case as well.@xcite
co - authored this note in his private capacity ; no official support or endorsement by the centers for disease control and prevention is intended or should be inferred .
see , for instance , v. hnizdo , magnetic dipole moment of a moving electric dipole , " am . j. phys .
* 80 * , 645 - 647 ( 2012 ) . in this paper , except for eqs .
( 1 ) and ( 2 ) , * p * and *
m * are the dipole moments of the _ moving _ dipoles_not _ the proper " moments in the dipoles rest frame .
see , for example , k. w. h. panofsky and m. phillips , _ classical electricity and magnetism _
( dover , new york , 2005 ) , sec .
18 - 4 . in a recent article [ d. j. griffiths and v. hnizdo ,
mansuripur s paradox , " am . j. phys . * 81 * , 570 - 574 ( 2013 ) ] we stated that there is no third term " in the torque formula .
this was misleading : there _ is _ a third term ( in general ) , but it was zero ( in that context ) because @xmath78 at the location of the dipole .
if the fields are _ non_uniform , the torque ( 9 ) is supplemented by @xmath79 $ ] , where @xmath80 is the dipole s displacement from the point with respect to which torques are calculated .
the term inside the brackets comprises the force on a moving dipole [ g. e. vekstein , on the electromagnetic force on a moving dipole , " eur .
* 18 * , 113117 ( 1997 ) ] .
this additional term vanishes when torques are calculated with respect to the position of the dipole .
we shall calculate all torques and angular momenta with respect to the origin .
v. namias , electrodynamics of moving dipoles : the case of the missing torque , " am . j. phys .
* 57 * , 171 - 177 ( 1989 ) . of course , you could model * m * as separated monopoles ( this is what namias did in ref . 7 ) , but it turns out the torque is different for gilbert " dipoles and ampre " dipoles ( tiny current loops ) .
see ref .
this approach is based on an argument by m. mansuripur ,
trouble with the lorentz law of force : incompatibility with special relativity and momentum conservation , " phys .
lett . * 108 * , 193901 - 1 - 4 ( 2012 ) .
a simpler version starts by representing stationary electric and magnetic by their charge and current densities : @xmath81 and lorentz transforming the 4-vector @xmath82 to get @xmath64 and @xmath83 for _ moving _ dipoles [ eqs .
( 22 ) , ( 25 ) , ( 40 ) , and ( 43 ) ] .
see , for example , d. j. griffiths , _ introduction to electrodynamics _
( pearson , boston , 2013 ) , eq .
( 12.118 ) .
10 , section 7.3.5 .
when attached to a delta function , the prime denotes differentiation with respect to its argument .
here we integrate by parts .
w. shockley and r. p. james , ` try simplest cases ' discovery of ` hidden momentum ' forces on ` magnetic currents ' , " phys .
lett . * 18 * , 876 - 879 ( 1967 ) ; w. h. furry , examples of momentum distributions in the electromagnetic field and in matter , " am . j. phys . * 37 * , 621 - 636 ( 1969 ) ; l. vaidman , torque and force on a magnetic dipole , " am . j. phys . * 58 * , 978 - 983 ( 1990 ) .
hidden momentum is mechanical in nature , even though it arises ( typically ) in electromagnetic contexts .
for details see ref . 4 , or ref . 10 ( example 12.13 ) .
if the electric field is a function of time , there will be an additional term of the form @xmath84 $ ] , but it vanishes if the torque is calculated with respect to the dipole s location .
there is no term involving @xmath85 , since the dipole moments are assumed to be constant .
d. j. griffiths,dipoles at rest , " am .
* 60 * , 979 - 987 ( 1992 ) .
note that a moving electric dipole carries a component of circulating current ; see hnizdo , ref . 1 .
a. l. kholmetskii , o. v. missevitch , and t. yarman , torque on a moving electric / magnetic dipole , " prog .
b * 45 * , 83 - 99 ( 2012 ) . these authors claim that the last two terms are new discoveries , but both these terms were obtained already by namias ( ref . 7 , eqs . ( 1 ) and ( 3 ) ; the 3rd term is implicit in the former equation ) .
incidentally , the same result [ eq .
( 54 ) ] holds for @xmath86 and @xmath87 constructed from magnetic monopoles and their currents , though ( curiously ) the roles of the last two terms are reversed : the fourth appears already in the analog to eq .
( 9 ) , while the third is due to hidden momentum . | we calculate the torque on an ideal ( point ) dipole moving with constant velocity through uniform electric and magnetic fields . |
the vista variables in the va lctea ( vvv ) eso public survey @xcite is a time - series , near - infrared ( ir ) survey of the galactic bulge and an adjacent portion of the inner disk , covering 562 square degrees of the sky ( fig .
[ fig : area ] ) .
the survey has provided multi - color photometry in 5 broadband filters ( @xmath0 , @xmath1 , @xmath2 , @xmath3 , and @xmath4 ) , but its main goal is to provide , for the first time , a homogeneous database for a variability study of the observed regions in the @xmath4-band .
vvv has much improved photometric precision compared with , and extends much deeper than , 2mass @xcite .
in addition , in contrast to single - epoch surveys , which only allow the construction of 2-dimensional ( 2d ) maps , with the addition of temporal information for well - established distance indicators such as rr lyrae stars ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ?
* ) , the vvv survey will enable us to resolve the 3d structure not only of the milky way , but also of the sagittarius dwarf spheroidal galaxy ( sgr dsph ; * ? ? ?
* ) , parts of which are also included in the vvv fields , and possibly even detect previously unknown galactic structures and streams .
the survey is carried out at the visible and infrared survey telescope for astronomy ( vista ) , a 4m - class telescope operated by eso and located at cerro paranal , chile .
the heart of the vista / vircam instrument is a @xmath6 array of raytheon virgo ir detectors ( @xmath7 pixels ) , with a pixel size of @xmath8 @xcite .
the size of a uniformly covered field ( also called a `` tile '' ) is 1.501 deg@xmath9 , hence the vvv survey requires a total of 348 such `` tiles '' to cover the survey area and include a small overlap between neighboring tiles ( fig .
[ fig : area ] ) .
the data reduction is carried out at the cambridge astronomy survey unit ( casu ) in collaboration with the uk wide - field astronomy unit ( wfau ) in edinburgh .
details about the data pipeline and the various steps of the calibrations can be found in @xcite , @xcite , and @xcite .
briefly , vvv images are pipeline - processed by the vista data flow system ( vdfs ; @xcite ) , including all steps of data reduction from image processing to photometry and its calibration .
individual exposures are subjected to standard steps of pre - processing , such as flat - fielding , dark subtraction , and non - linearity correction .
science frames are composed of two dithered images , i.e. subsequent exposures taken with an offset of typically @xmath10 pixels , in order to remove detector artifacts , cosmic rays , and other cosmetic defects .
the resulting detector frame stacks , a.k.a .
`` pawprints , '' have non - contiguous areal coverage due to the large gaps between the 16 vircam chips . at each observational epoch ,
a sequence of six pawprints is acquired , and these are further combined to form a contiguous mosaic image , a.k.a . `` tile . '' further steps in the reduction consist of source extraction and aperture photometry .
sets of small circular apertures with increasing radii are used in order to maximize the signal - to - noise ratio in highly time - varying seeing conditions , and to suppress systematics due to source crowding .
furthermore , the pipeline includes point - spread function ( psf ) estimation and psf fits for each object .
flux loss in the wings of the psfs is remedied by aperture corrections @xcite .
sources are classified based on the shape of the psf ( see , e.g. , * ? ? ?
source positions are astrometrized using 2mass @xcite stars as reference , with a median accuracy of @xmath11 mas , depending on magnitude @xcite .
magnitudes are corrected for detector distortion , and in the case of the @xmath12 filters are zero - point ( zp ) calibrated on a frame - by - frame basis using local 2mass secondary standards . for the zp calibration of the @xmath0 and @xmath1 filters , the procedure described in @xcite is followed .
the zp accuracy of @xmath12 magnitudes depends on the number of available non - saturated local standards ( i.e. , the sky conditions ) , and is usually within @xmath13 , while the @xmath14 photometry is typically accurate to @xmath15 mag .
we note that all vvv photometric data are on the vista magnitude system .
the vvv survey started its data - gathering phase in february 2010 , with the first semester focusing on complete multi - band coverage of the whole survey area ( 562 square degrees ) in its 5 broadband filters ( see [ sec : over ] ) .
this was completed in 2011 .
the @xmath4-band observations that comprise the vvv variability study started in parallel to the multi - color observations in 2010 , and are still ongoing .
once completed , the bulge region will contain between 60 and 100 epochs for each tile , whereas the disk tiles will have @xmath16 epochs each .
the first two vvv public data releases have already become available from eso , and @xcite contains a detailed description of the first vvv data release . as of this writing , the various bulge tiles have been observed between 31 and 70 times , whereas the disk tiles have been observed 17 to 36 times ( see also fig .
[ fig : epochs - delays ] ) .
only observations which were carried out under the observing constraints as specified in the survey proposal are given the status `` completed . '' with respect to the submitted and scheduled observations , the vvv survey is @xmath17 complete . including observations scheduled for years 5 and 6 of the survey
, this translates to an overall completeness of @xmath18 . since
several vvv tiles have already been observed a few dozen times , a first look at stellar variability in the vvv data is already possible and this is precisely the main goal of this paper .
the vvv survey has been monitoring the bulge and the southern disk in the @xmath4-band since 2010 .
it will provide , for the first time , a homogeneous database with long - baseline time - series photometry with up to @xmath19 epochs for close to @xmath20 point sources .
a brief overview of stellar variability in the vvv data was provided in @xcite . at its current status , after the extensive monitoring of the bulge fields has started , vvv has already provided a considerable number of epochs , suitable for analyses of stellar variability .
vvv provides a sparse sampling of the time - domain , usually a single epoch for a few fields on a night ( with an occasional second visit ) , distributed close to randomly over the seasonal visibility period of the area .
figure [ fig : epochs - delays ] shows the time sampling of bulge and disk fields .
most of the currently available time - series data ( see [ sec : status ] ) were taken in the third year , and epochs for some tiles have a moderate clustering in the time sampling .
the reason for the similarity in the sampling between different tiles is that observations are carried out in concatenations of nearby fields , in order to provide information on the near - ir atmospheric foreground emission , as is necessary in the data reduction process . in the following ,
we provide a brief overview of the basic properties of the @xmath4-band light curves based on the data of the bulge tile b293 ( @xmath21 , @xmath22 ) .
this tile has 64 epochs with available vdfs photometry , and lies over a moderately crowded stellar field , allowing us to give the current best assessment of the quality of the time - series data .
the limiting magnitude in the @xmath4-band varies between @xmath23 and @xmath24 mag , depending on the galactic latitude , due to differences in extinction and crowding @xcite , and also showing strong nightly variations , mainly due to the highly variable brightness of the near - ir sky , and partly because of the variable seeing or / and cloud coverage during observations .
we note that these conditions also cause significant variations in the photometric properties of bright objects , due to the changing saturation level .
figure [ fig : epochs ] shows the number of @xmath4 detections for vdfs 1.2 photometry as a function of the @xmath4-band brightness for objects fainter than @xmath25 mag in field b293 .
the drop beyond @xmath26 mag is due to the non - detection of faint objects when observing conditions are sub - optimal .
the overwhelming majority of stars brighter than @xmath27 mag are detected under almost all observing conditions .
most of the bright objects with fewer than 60 detections are crowded and suffer from seeing - dependent merging with nearby sources , while some epochs for the faintest objects are missing mostly due to the high variation in the limiting magnitude caused by the variable near - ir atmospheric foreground .
the secondary falling ridge starting at @xmath28 mag is due to the lower sensitivity of certain vircam chips .
figure [ fig : sigma - ks ] shows the density distribution of vvv variable objects on the root - mean - square ( rms ) scatter average @xmath4 magnitude plane for sources on tile b293 with at least 20 epochs , after applying a general @xmath29 threshold - rejection procedure to the light curves .
note that the @xmath4 limiting magnitude of this field is about 0.5 mag fainter than the visible limit of the distribution in figure [ fig : sigma - ks ] @xcite , but the majority of the faintest objects was detected only at a few epochs .
figure [ fig : sigma - ks ] demonstrates well the high quality of the vdfs photometry note that the photometric zp is calibrated at each epoch , thus the scatter shown contains not only the formal photometric errors , but also the dispersion from the uncertainties of the nightly zps . the low noise at faint magnitudes ( @xmath30 mag at @xmath31 mag ) will allow us to detect rr lyrae stars even beyond the bulge . for bright stars ( @xmath32 mag ) ,
the photometric precision is around @xmath33 mag , which will allow us to study low - amplitude variables , including planetary transits around nearby k and m dwarfs @xcite , and to investigate the detailed near - ir light curve properties of pulsating and chromospherically active stars alike .
there are various strategies for improving the photometric quality of the vvv light curves with respect to the vdfs catalogs ( see [ sec : var - over ] ) , by means of more sophisticated source extraction and photometric algorithms which may be better suited for dense stellar fields .
the vvv science team is developing highly automated software pipelines which implement both of the two main approaches that currently deal with crowded - field photometry : psf - fitting photometry ( e.g. , daophot / allframe , @xcite ; dophot , @xcite ) and difference image analysis ( dia ; e.g. , @xcite ) .
one of these pipelines , based on daophot , is described in @xcite .
ultimately , some of these codes will be incorporated into the vdfs , providing a highly robust and versatile data analysis pipeline , equally well suited to handle all types of target fields .
in addition to using the vdfs photometry catalogs from casu and daophot/ dophot pipelines ( see [ sec : techn ] ) , vvv is also developing its own dia pipeline . dia is particularly effective for obtaining photometry of faint objects in highly crowded fields .
we employ a modified version of @xcite s optimal image subtraction package isis @xcite .
the standard approach taken by isis is to convolve and photometrically match a reference image with good seeing to a target image of poorer seeing in order to obtain a difference image showing only flux changes between the two epochs .
a light curve can then be obtained through photometry of a time series of difference images .
the advantage of dia for variability studies is that the difference images contain only varying objects and are therefore unaffected by blended non - varying sources .
the convolution kernel to transform the reference image psf to that of the target image is determined through a least - squares minimization using corresponding sub - regions of the two images .
difference image sub - region .
most of the image is gray , reflecting non - varying regions of the image .
varying regions show up as white ( black ) for flux increases ( decreases ) .
saturated stars typically show up as large blackened structures , sometimes with masked central regions as shown by the example within the rectangle .
genuine variations have a more compact psf such as the two circled examples.,scaledwidth=80.0% ] whilst the standard approach is to convolve a good seeing reference image to a poorer seeing target image , this may not work well when the psf is undersampled on the reference image .
ideally , good difference image quality requires typically at least 2.5 pixels / fwhm ( i.e. , @xmath34 seeing , at the average vircam pixel size ) , whereas the seeing at the vista site in cerro paranal is often below this the best - seeing images may have less than 2 pixels / fwhm . for this reason
, we choose to use a relatively poor - seeing image with a well - sampled psf as the reference image , and then convolve a target image with superior seeing to it . for a reference image @xmath35 , a geometrically registered target image @xmath36 and convolution kernel @xmath37 ,
we minimize the sum @xmath38 ^ 2 \,\ , , \label{newdia}\ ] ] where the sum is over image pixels @xmath39 and @xmath40 represents a smooth differential background model .
@xcite s isis package minimizes @xmath41 with respect to a set of linear coefficients to basis functions which represent both the convolution kernel @xmath37 and the differential background model @xmath40 . for @xmath37
the basis functions are a superposition of gaussian - like functions , and for @xmath40 the basis functions are 2d polynomials .
figure [ fig : dia ] shows an example of a subtracted image that was obtained following this approach . whilst using a poor - seeing image for the reference may result in a difference image with reduced signal - to - noise ratio , this can be compensated by the fact that all difference images will have the same psf ( unlike @xcite s original method ) and one can iterate the procedure using the stack of difference images to produce a new , high signal - to - noise reference image .
the vvv dia pipeline uses a modified version of isis which uses equation ( [ newdia ] ) in place of the standard approach .
once a difference image sequence has been produced , variable objects are identified from a squared stack of difference images , and psf photometry is performed for all epochs at the locations of all variable objects ( defined above some signal - to - noise threshold ) . resulting raw light curves are written to files .
we can search these files using specific criteria for e.g. periodic variability , or use some sutiable filter for specific transient objects such as microlenses . at present
we are performing searches for regular periodic variables in order to calibrate the performance of the pipeline and , ultimately , catalog target populations such as rr lyrae stars . using dia
we should be able to extend the sensitivity of traditional photometry offered by the vdfs catalogs and detect much fainter variables and transients , as well as variable sources that are blended and thus missing from the casu catalogs .
dia provides only _ relative _ photometry for variable light curves it does not tie the photometry onto a calibrated magnitude scale . however , absolute calibration for a dia photometry time - series reduces to calibrating just one epoch as all other epochs are then calibrated through the dia relative photometry .
we calibrate our dia photometry using vdfs aperture photometry catalogs for one of the best - seeing epochs in the time series .
we refer to this epoch as our _ photometric reference _ , which is generally different to the reference image used for image subtraction discussed in
as we have seen , the image subtraction reference image is typically one of poor seeing , in order to maximize sampling of the psf and ensure good image subtraction .
for photometric calibration , however , we choose a good - seeing image , as this provides the best signal - to - noise photometry and minimizes blending effects . in order to tie in the dia photometry catalogs to the vdfs photometry catalogs we have to cross - match them . prior to cross - matching
the catalogs are first filtered for unsaturated objects which are classified as `` stellar '' by the vdfs pipeline .
the ra and dec coordinates for objects in the dia source lists are obtained using the world coordinate system information in the dia reference image .
each object is then cross - matched to the vdfs object with the smallest angular separation from the dia position , within a maximum radius of @xmath42 , where @xmath43 is the dia reference image seeing .
the factor of 2.5 was chosen by inspection of a histogram of first- and second - nearest matches for some of the most crowded fields , assuming the second - nearest matches are tracers of potential contaminants .
figure [ fig : xmatchhist ] shows an example histogram of radial distances for first- and second - nearest vdfs catalog matches to dia objects within the four galactic center tiles .
is given above the plot .
the cross - match maximum radius is shown as a dashed line .
the first- and second - nearest matches are shown as the blue and red histograms , respectively .
the radial distances , @xmath44 , are given in units of @xmath45 , where @xmath43 is the dia reference image seeing .
the fraction of potential contaminants ( second - nearest matches ) within the maximum radius , @xmath46 , is shown beneath the legend.,scaledwidth=80.0% ] the vdfs pipeline at casu performs crowded - field aperture photometry @xcite . for each object ,
flux counts are obtained for a series of aperture sizes , and used to build a curve - of - growth model .
this model is used as an approximation to a psf profile to calculate the necessary corrections for the conversion of aperture flux to object magnitude . for each vvv dia object and cross - matched vdfs catalog object , it is necessary to select the largest aperture which contains the majority of the psf while minimizing the risk of contaminating flux from blended stars .
the largest vdfs aperture useful for point - source measurements has a radius of 4 arcsec , which is large enough to guarantee encircling at least 99% of the total flux under all expected seeing conditions .
however , within highly crowded fields such an aperture may be strongly affected by blending , and under typical seeing 99% of the flux will be encircled by much smaller apertures . within the range of 6 vdfs apertures up to this largest aperture
, we select the largest aperture for which blending effects remain small and we perform this selection for each object individually . to this end
we compute the forward difference magnitude between neighbor vdfs apertures ( which have curve - of - growth aperture corrections applied to them ) , @xmath47 , and we then determine the aperture @xmath48 , at which @xmath49 exceeds some limit .
this is currently set to @xmath50 mag , based on an inspection of the aperture magnitude series of a few thousand objects in the galactic center tiles , but we are seeking a more robust way of determining this limit . having established the dia object magnitude at the photometry reference epoch , @xmath51 , we can then transform the difference flux measurements for all epochs onto a calibrated magnitude scale .
the magnitude @xmath52 at each epoch @xmath39 is calibrated from the difference flux , @xmath53 , and the difference flux at the photometry reference epoch , @xmath54 . taking these together with the ( optimal ) aperture flux ( corrected to the full psf flux via the curve - of - growth model ) , @xmath55 , the magnitude zp , @xmath56 , and exposure time , @xmath57 ,
from the vdfs cross - matched object , we have : @xmath58 where @xmath59 @xmath60 and @xmath61 the random error on each individual epoch magnitude is simply the error on its difference magnitude @xmath62 .
the systematic error in the baseline is obtained from the quadrature sum of the errors in the photometric reference epoch magnitude and difference magnitude : @xmath63 -band light curves for two example cepheids ( nos .
10 , _ top panel _ , and 26 , _ bottom panel _ ) in @xcite .
the gray line is a smoothed radial basis function approximation to the vvv dia data ( black points ) , with the light gray region representing the maximal systematic offset error in the light curve baseline ( equation [ sys ] ) .
the @xmath4 light curve of @xcite ( red points ) is also shown , offset - corrected to the vvv magnitude calibration for ease of comparison.,title="fig:",scaledwidth=75.0% ] -band light curves for two example cepheids ( nos .
10 , _ top panel _ , and 26 , _ bottom panel _ ) in @xcite .
the gray line is a smoothed radial basis function approximation to the vvv dia data ( black points ) , with the light gray region representing the maximal systematic offset error in the light curve baseline ( equation [ sys ] ) .
the @xmath4 light curve of @xcite ( red points ) is also shown , offset - corrected to the vvv magnitude calibration for ease of comparison.,title="fig:",scaledwidth=75.0% ] the errors @xmath64 , @xmath65 , and @xmath62 are all obtained from equations ( [ eq : photmag]-[eq : epochdmag ] ) through standard propagation of errors .
figure [ fig : matsunaga ] shows calibrated vvv dia light curves for two cepheids found by @xcite , with their @xmath4-band photometry also shown for comparison .
we typically find an offset between the calibrated vvv dia photometry and the @xcite calibrated photometry , though it is within our computed maximal systematic uncertainty ( determined from equation [ sys ] and shown by the light gray region ) . in figure
[ fig : matsunaga ] we have subtracted off these offsets and we have folded the vvv photometry to the @xcite periods for direct comparison .
the gray lines represent smooth radial basis function approximations to the vvv dia data ( black points ) .
figure [ fig : matsunaga2 ] , in turn , compares vvv dia photometry with vdfs aperture photometry ( aperture number 3 ) for the same @xcite objects as in figure [ fig : matsunaga ] .
vdfs aperture photometry also typically shows an offset ( subtracted off in figure [ fig : matsunaga2 ] ) within the systematic uncertainty of the dia , for reasons which we are still investigating .
the random errors from dia compare favorably with those from aperture photometry .
-band light curves for the same cepheids as in figure [ fig : matsunaga ] , compared with light curves obtained directly from vdfs aperture photometry ( aperture number 3 ; blue points ) and offset - corrected .
the dia light curves are identical to those in figure [ fig : matsunaga ] . ,
title="fig:",scaledwidth=75.0% ] -band light curves for the same cepheids as in figure [ fig : matsunaga ] , compared with light curves obtained directly from vdfs aperture photometry ( aperture number 3 ; blue points ) and offset - corrected .
the dia light curves are identical to those in figure [ fig : matsunaga ] .
, title="fig:",scaledwidth=75.0% ] the vvv survey aims at reaching a highly complete census of variable stars that can be used as population tracers , such as radially pulsating stars and eclipsing binary systems .
while the old and metal - poor rr lyrae stars will be used to trace the 3d structure of the bulge , classical cepheids and eclipsing binaries will be employed to map the spiral arm structure on the ( largely unexplored ) far side of the galaxy , both in the disk region and behind the bulge . in order to get an unbiased picture on galactic structures traced by these stars , both the rate and precision in the detection of periodic signals have to be sufficiently high .
the time - domain coverage at the present stage of the vvv survey is not yet sufficiently large for conducting variability searches at high completeness , therefore our current investigations are limited to already known samples of distance indicators ( @xcite ) .
however , we are already in the position to give estimates on the future signal detection performance in vvv based on the time - series data of the 8 vvv bulge fields with more than 60 epochs , which have a partial overlap with the highly complementary optical time - domain survey ogle - iii @xcite . in the following , we present a simple and preliminary assessment of signal detection rates based on the data for the ab - type ( i.e. , fundamental - mode ) rr lyrae stars in this region that are known from ogle - iii .
we performed a positional cross - matching procedure between the ogle - iii catalog of rr lyrae stars @xcite and the vdfs _ pawprint _ photometric catalogs , including data up to 2012 november ( 61 epochs ) .
figure [ fig : ksep ] ( _ top panel _ ) shows the density plot of the angular separations between the best - matching sources as a function of the average @xmath4 magnitude , for all the images .
the average cross - matching accuracy is @xmath66 , and more than @xmath67 of the data points are concentrated in a small locus with separations less than @xmath68 , which indicates that both the precision and accuracy in the astrometry of both surveys is very high ( the average vircam pixel size is @xmath8 ; see [ sec : over ] ) .
the small clusters of points represent complete light curves with poorer cross - matching accuracy , while the points with more diffuse distribution are due to intermittent cross - matching inaccuracies due to , e.g. , source merging at poor seeing , elongated sources , spurious signal contamination close to saturated stars , etc . in total , 1558 out of 1832 rrab stars
were successfully cross - matched with near - ir counterparts with separations not larger than @xmath69 , meaning @xmath70 completeness in on - chip detections with respect to ogle - iii . the main limiting factor for the vdfs photometry is source crowding , since many close objects can not be separated even using small apertures .
psf photometry and dia ( see [ dia ] ) are expected to significantly improve this figure .
figure [ fig : ksep ] ( _ bottom panel _ ) shows the cumulative distribution of average @xmath71 magnitudes of those rrab stars where near - ir counterparts were not found in vdfs catalogs within the above constraints .
the distribution is close to uniform , which means that a similar fraction of objects is missing at bright and faint magnitudes .
thus the limiting magnitude does not affect our completeness compared to ogle - iii .
on the contrary , for stars lying close to the galactic plane , where extinction is severe in optical bands , we expect to have a much higher completeness in the census of variable stars , due to the advantage of the employed near - ir wavebands in penetrating high - extinction regions .
to evaluate the signal detection rate , we compute the lomb - scargle periodograms @xcite of the vvv @xmath4 light curves of successfully cross - matched rrab stars that have at least 50 epochs ( 1076 objects ) .
then , we characterize the completeness in signal detection by the cumulative detection efficiency ( cde ) :
@xmath72 where @xmath73 is the number of successfully detected rrab stars , and @xmath74 is the total number of rrab stars , respectively both including all stars with magnitude @xmath75 up to a certain magnitude @xmath4 .
we consider a detection successful if the signal is within the @xmath76 significance level evaluated on white noise , and its frequency , refined by non - linear least - squares fitting , does not differ from the one reported by ogle - iii by more than the nominal frequency resolution ( i.e. , the inverse of the baseline , or @xmath77 ) . the result is shown in figure [ fig : cdp ] , compared with detection efficiencies computed from synthetic data .
for the latter , a typical @xmath4-band rrab light curve ( similar to sw and ; see @xcite , and also fig .
[ fig : felipe ] below ) with low and high amplitude ( @xmath78 , @xmath79 mag ; and @xmath80 , @xmath81 mag , respectively ) was sampled randomly within the visibility periods of the bulge in a vvv - like scenario , random noise was added based on the scatter diagram shown in figure [ fig : sigma - ks ] , and the cde was computed from a few thousand realizations . we conclude that a signal detection rate of @xmath82 can already be reached on vvv data for rrab stars in the bulge , in light curves sampled with @xmath83 epochs . this number will be pushed significantly higher by raising the number of epochs to @xmath84 ( i.e. , increasing the spectral @xmath85 and resolution , as well as decreasing the effect of aliasing ) , and improving the photometric precision by either profile - fitting photometry or dia .
another important aspect of signal detection efficiency is the relative number of significant signals with incorrect periods , e.g. , due to aliasing .
ideally , the number of such cases should not exceed a few percent because , for instance , wrong periods imply wrong distances through the period - luminosity ( pl ) relations , and if these errors are too frequent , they can blur the 3d structures that we wish to trace .
the relative frequency of signals with incorrectly recovered periods can be quantified by the cumulative false detection rate ( cfr ) , defined analogously to cde ( eq . [ cde ] ) : @xmath86 where @xmath87 is the cumulative number of detections where the frequency is incorrect , and @xmath88 is the total number of light curves with significant signals , both again including sources up to a certain magnitude .
the top panel of figure [ fig : cdp ] shows the cfr measured in our rrab sample , once again in comparison with the values computed from synthetic data ( see above ) .
the cfr is about @xmath89 for bulge rrab stars , which is already sufficient to rather sharply trace the 3d structure of the underlying stellar population , and we expect to reach @xmath13 by the end of the survey .
we note that the peak close to 12 mag is produced by very few stars , and is caused by heavy aliasing introduced by the intermittent saturation of these bright objects .
current estimates based on the analysis of the available vista datasets as delivered by the casu vdfs pipeline have suggested that the final number of variable stars observed in the vvv survey may be in the range between 10@xmath90 and 10@xmath91 stars .
these large numbers of objects require new approaches for the data analysis and selection , including artificial intelligence algorithms .
machine learning techniques applied to variable star classification have become particularly popular in recent years .
for instance , in @xcite the authors explored several classification techniques , quantitatively comparing performance ( e.g. , in terms of computational time ) and final results ( e.g. , in terms of accuracy ) of different classifiers with their corresponding learning algorithms .
more recently , a few other studies have focused on specific methodologies , with the implicit goal of finding the best compromise between robustness and speed . as an example ,
@xcite and @xcite have independently presented tree - based methods for the automated classification of hipparcos and ogle variable stars ; @xcite have employed machine learning techniques to detect quasi - stellar objects in the macho and eros-2 databases ; while @xcite opted for algorithms based on multivariate bayesian statistics in order to explore the variable star content of the tres lyr1 sky field .
whichever specific algorithms preferred , the general idea behind these supervised machine learning methodologies is to create a function , the _ classifier _ , able to infer the most probable _ label _ of an object ( in our case , the variability class to which an unclassified variable star belongs to ) on the basis of what is learned by the analysis of inputs ( light curve features ) from a _ training set _ ( a collection of high - quality light curves of previously classified variable stars ) . in the most general framework , the sequence of steps to be performed can be thought as : 1 ) build a training set ( template database ) ; 2 ) determine the input feature representation ( e.g. , periods and harmonics , as derived from fourier analysis ) of the learned function ; 3 ) determine the nature of the classifier with its corresponding learning algorithm ( e.g. , artificial neural network , support vector machines ( svms ) , tree - based methods , etc . ) ; 4 ) run the classifier on the gathered training set , using the information on well - known variables stored in the training set for searching and labeling unknown variable stars in the test set , i.e. , in the data archive that one is dealing with ; 5 ) finally , evaluate the accuracy of the learned function , i.e. , evaluate the fraction of correctly classified variable stars ( e.g. , through a so - called confusion matrix ; * ? ? ?
* ) . in what follows , we describe in some detail the first three points of this proposed working strategy . until now
, stellar variability in the near - ir has been a relatively ill - explored research field : in particular , the number of high - quality light curves was very limited and , even worse , many variability classes have not yet been observed in a sufficiently extensive way in the near - ir , so that good light curves are entirely lacking for some such classes .
since vvv is the first ever large survey dedicated to stellar variability in the near - ir , the first problem we had to face has thus been the construction of a proper training set , i.e. , a database of high - quality ( `` template '' ) near - ir light curves taken to be representative of the different variability classes under study .
our effort in building such a template database has included four main routes , which are described in the following . * _ a ) near - ir light curves from the literature .
* firstly , we have extensively explored the available literature , searching for papers that presented high - quality near - ir time - series data . when these data were not published in machine - readable form and we could not obtain the data directly from the authors , we digitized the data tables and/or plots , and used optical character recognition software to convert those into ascii files . among the types of variables for
which data could be retrieved in this way are rr lyrae , classical cepheids , miras , eclipsing binaries , and wolf - rayet stars , among others .
examples of high - quality data obtained in this way are shown in figure [ fig : felipe ] . * _ b ) near - ir light curves from public archives . _
* we have also searched publicly available archival databases for high - quality near - ir data , finding the 3.8 m united kingdom infrared telescope ( ukirt ) wide - field camera ( wfcam ; @xcite ) calibration archive ( wfcamcal ) especially appropriate for our purposes .
wfcamcal s current data release ( dr8 ) contains data from 364,905 individual pointings on both the northern and southern hemispheres , spread over nearly half of the sky .
the majority of the fields are observed repeatedly , with a rather irregular sampling that has however provided for many sources a reasonably extended time coverage .
the selection of variable sources from the wfcamcal catalog and the corresponding light curves will be presented in ferreira lopes et al .
( 2013 , in preparation ) ; an example of a light curve derived in this way is provided in figure [ fig : wfcam ] . * _ c ) vvv templates observing project .
* in addition to data from the literature and from public archives , we have also carried out time - series near - ir observations of our own . in this way
, we have monitored hundreds of ( optically well - studied ) variable stars in the @xmath2 , @xmath3 , and @xmath4 bands , using several different facilities located at different observatories across the globe ( see table 3 in @xcite for a listing of all the telescopes and instruments used ) .
an example of a light curve obtained in this way is shown in figure [ fig : wysco ] ; additional examples can be found in @xcite . a full description of the project will be provided in angeloni et al .
( 2013 , in preparation ) . * _ d ) vvv light curves . _ * last but not least , vvv itself has started delivering light curves of variable stars that had already been classified by previous optical surveys , most notably macho @xcite and ogle @xcite .
as an example , figure [ fig : vvv_blher ] shows a bl her ( type ii cepheid ) light curve , obtained using ogle - iii @xmath71-band data @xcite and vvv @xmath4 data .
as more and more light curves of this quality become available , our template database will be augmented accordingly , thus enabling increasingly more accurate classification of previously unclassified vvv variable sources .
the problem of automated classification of variable stars using temporal series has been around for some time .
for example , @xcite have successfully implemented algorithms for automated classification which are based on extensive light curve databases , containing thousands of entries , from hipparcos , ogle , and other projects all providing light curves in the visible bandpasses only . in the case of near - ir projects like vvv
, special care must be taken both in the feature extraction process and in the implementation of classification algorithms , because the number of training examples ( i.e. , high - quality light curve templates ) that is available at the outset is not as large .
how does one optimally identify and extract the most informative features that best describe a temporal series ?
the usual model for a light curve of a given variable star , where @xmath92 represents the magnitude as a function of time @xmath93 ( usually julian days ) , is @xmath94 where @xmath95 here @xmath48 and @xmath96 represent the intercept and slope respectively of the linear trend in the time series , @xmath97 is the harmonic sum ( which tries to describe the star s actual variability ) , @xmath98 is the number of harmonics for a given frequency @xmath99 , @xmath100 is the number of frequencies obtained from the light curve , and @xmath101 is the ( possibly correlated ) noise . from here , obtaining features of the light curve is straightforward : the frequencies could be estimated via the `` classical '' lomb - scargle periodogram @xcite , first detecting and subtracting the linear trend , then detecting and subtracting the frequencies one at a time ( i.e. , performing prewhitening ) , and finally performing a full least - squares fit with the model in equation ( [ fullfit ] ) . from here , one can decide to extract some useful features for classification . following the work of @xcite ,
these features are ( i ) the parameters of the linear trend ; ( ii ) the amplitudes , which are given by @xmath102 ( iii ) the ( time - invariant ) phases , given by @xmath103 } \,\
, , \ ] ] where @xmath104 ( with @xmath105 ) ; and ( iv ) the ratio @xmath106 of two variances .
@xmath107 corresponds to the variance of the raw photometric measurements , while @xmath108 is the variance from the residual _ after _ the first frequency fit ( with the corresponding @xmath109 harmonics of this first frequency ) .
the above prescribed procedure , though somewhat `` standard , '' has nonetheless some issues that must be carefully addressed .
the first issue with which we must be careful is whether classical periodogram searches , such as the lomb - scargle method , constitute the optimum solution for the frequency / period searches .
for example , @xcite claimed that the classical periodogram method can be thought of as trying to fit sine curves ( with the corresponding phase ) of different frequencies , i.e. , a fit of the form @xmath110 where the periodogram is just a function of the amplitude of this sinusoid for each frequency @xmath111 . according to these authors
, this procedure does not take into account the fact that the sinusoid may be `` floating '' around a different mean value than that of the linear trend in equation ( [ fullfit ] ) .
they claim , in addition , that this procedure does not weigh the data points according to their respective variances , as they should in order to have unbiased estimators for the amplitudes in the presence of uncorrelated noise . to solve these issues ,
@xcite add an extra term to the function @xmath112 above to account for this `` floating mean , '' and write the least - squares solution assuming a diagonal covariance structure for the data , where each entry of this diagonal covariance matrix corresponds to the variance on each data point .
@xcite , on the other hand , disagree with this procedure .
they actually claim that performing least - square fits to sinusoids is different from calculating the periodogram , which in essence is the squared amplitude of the discrete fourier transform of the data .
moreover , they claim that their formalism is capable of dealing with all types of colored noise .
the second issue that one has to take care of is aliasing , i.e. , spurious peaks due to sampling .
for example , @xcite have shown that the problem of aliasing can trick not only period - finding algorithms , but also the human eye , misidentifying a `` real '' peak in the periodogram by a fake peak produced by sampling ( ir)regularities .
this is a very serious and critical problem for classification purposes : if we fail to identify the true period of a periodic variable star , then we lose perhaps its most important feature .
basically , @xcite showed that it is not always true that the highest peak in the periodogram is actually a ( physical ) period of the object of study .
the third issue is how to select @xmath100 , the number of frequencies in the harmonic fit and , given that number , how to select @xmath98 , the number of harmonics in each frequency fit .
even further , one may ask whether or not the number of harmonics should be the same for all the frequencies .
physically , there is no a priori reason to assume that the number of harmonics should be the same , nor that all variability types are characterized by the same number of frequencies .
as can be seen , the issue of feature identification and extraction remains subject to considerable debate .
we are currently conducting extensive tests using vvv data and our templates database ( [ sec : templates ] ) , in order to obtain an optimum strategy for the purpose of classifying vvv light curves .
as a first approach to solve some of the issues mentioned above , we propose to initially over - fit the model in equation ( [ fullfit ] ) , using more features than possibly necessary to fit the light curves . in more detail
, we force our model to fit for each light curve up to five frequencies and up to five harmonics for each frequency to obtain the features with which we represent the light curves .
therefore , each light curve can now be represented as a vector @xmath113 containing the frequencies , amplitudes , phases , slope and intercept from the linear trend , and ratio of variances . in a second step , we fit a binary classifier , @xmath114 , for each class @xmath115 which can learn how to discriminate between classes and perform feature selection _ simultaneously_. this classifier is called lasso .
the lasso classifier estimates the parameters @xmath116 by minimizing a cost function subject to a constraint on the size of the @xmath117 norm of the parameter vector given by @xmath118 .
this constraint on the size of the parameter vector shrinks the parameter estimates towards zero , and the use of the @xmath117 norm forces some of the parameters to be equal to zero .
the features associated with parameters that are estimated as being equal to zero are irrelevant for the classifier .
therefore , this procedure allows us to have a different set of features for each classifier , and is convenient in the present setting in which we need to classify variable stars which can be mono - periodic , multi - periodic , or non - periodic .
for example , mono - periodic light curves should need fewer frequencies in the harmonic fit than multi - periodic stars , which implies that fewer features are needed for the classifier and this is borne out naturally by lasso . in order to improve our classifiers
as new light curve data become available , we plan to augment our templates training set ( [ sec : templates ] ) by selecting the most informative light curves through an active learning methodology .
active learning procedures are very helpful to efficiently select the most informative instances to be included in training sets @xcite . in the astronomical context ,
@xcite recently used active learning and showed its benefits for handling astronomical data .
these authors show that active learning techniques can reduce the bias of the training process and increase the classification accuracy .
most active learning models are composed of two phases .
the first is the exploration phase , where the model explores the most informative instances to select .
the second is the exploitation phase , where the model includes the feedback after the new selected instance and updates itself in order to repeat the exploitation phase @xcite .
usually the exploration phase requires a lot of computations , in most cases passing through the data many times , making it very difficult to be directly applied to large astronomical catalogs
. the exploitation phase may be less costly , depending on the model used .
as in @xcite , most active learning models assume the existence of an oracle , an entity that can correctly classify any query instance . unfortunately , such an oracle may not be available , or / and may be too expensive to implement . trying to deal with the absence of an oracle , semi - supervised active learning techniques attempt to use only the available labels @xcite .
these models use expectation maximization to estimate the best prediction for the missing labels using the current ones , but unfortunately most of these models can not handle huge datasets because of the computational cost of the algorithm .
recent work introduced active learning for large datasets @xcite , based on a similar approach as in @xcite but using hashing techniques to speed up the process .
this provides a very interesting approach to the large datasets " problem , but still assumes the existence of an oracle . in this context
, we are currently developing an active learning framework for large datasets , modelling the partial absence of the oracle . after developing this semi - supervised active learning framework
, we will iterate between automated classification of the available vvv light curves and increasing the size of the training set by incorporating new template - quality light curves ( i.e. , light curves with extremely high classification probability , as judged by the code and/or the oracle , in this case a vvv astronomer ) to the latter .
this will allow us to refine the classification procedure , and then re - run the whole process with increased confidence as additional data are incorporated into the main vvv light curve database . in this way
, our final vvv light curve database will contain not only periods and magnitudes , but also the variability classes to which the stars are associated which in some cases could / should be confirmed a posteriori using additional data , such as spectra and photometric observations in other bandpasses .
one of the main scientific goals of the vvv survey is to complete the census of classical radially pulsating variable stars , such as rr lyrae stars and cepheids , in the galactic bulge and disk .
these stars provide very important means for the study of the 3d structure of the milky way because they follow precise pl relations in the near - ir ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , thus they can be used as standard candles , and employed to infer the spatial distribution of their underlying stellar populations .
this is particularly efficient if photometric data are available in multiple bands , because in this case the interstellar reddening can be computed on a star - by - star basis , by comparing the measured color index to the intrinsic one , as predicted by the pl relations in different bandpasses .
this information can be used both for the distance determinations and for mapping the extinction , and even to trace the large - scale variations in the reddening law in the bulge that has been reported by various studies ( see , e.g. , * ? ? ?
* ; * ? ? ?
the vvv near - ir data on pulsating stars is particularly powerful in combination with optical ( e.g. , @xmath71-band ) light curves of complementary time - domain surveys such as ogle - iii , because the accurate mean magnitudes and the large wavelength separation allow a very precise determination of the absolute extinction from the color excess , and the result will be highly independent from the value of the selective - to - total extinction ratio @xmath119 . in @xcite , we used vvv near - ir photometry of known bulge rrab stars in combination with optical light curves from ogle - iii , to study the 3d structure of the bulge .
our results showed that the spatial distribution of the rr lyrae stars is significantly different form the x - shaped distribution of the red clump stars @xcite ; rather , it is rather spheroidal , and does not show a strong bar .
this finding implies that the milky way may have retained a classical bulge component , with a high fraction of stars in non - cylindrical stellar orbits .
the new rr lyrae stars discovered by the vvv survey will allow us to extend our analysis to a much larger sample .
based on our signal detection tests discussed in
[ sec : cross ] , we can conclude that the vvv survey is capable of yielding a highly complete census of fundamental - mode rr lyrae stars in the bulge .
figure [ fig : rrab ] shows one of the several thousand new rr lyrae stars discovered so far by vvv in bulge areas that are not covered by any other time - domain surveys .
the high - quality photometry will allow us to unambiguously classify the majority of the new objects , even relatively faint stars , based on their light - curve features ( see [ sec : class ] ) .
by greatly expanding the areal coverage as compared to previous optical surveys , and reaching much deeper in highly obscured areas , the vvv survey will provide unique data that are essential for understanding the structure and formation history of the milky way .
the same can be expected for classical cepheids either behind the bulge or in the disk region , since these stars have much larger amplitudes .
these objects will provide important means to trace the spiral arm structure on the far side of the galaxy , which has been out of the reach of optical surveys due to high interstellar extinction , and thus present vast uncharted territories of the milky way .
as we have seen , one of the main goals of the vvv survey is to describe in detail the inner milky way structure , which is already being accomplished using distance indicators such as rr lyrae stars and red clump giants ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . however , in addition to rr lyrae and cepheid variables , long - period variables ( lpvs ) are also expected to be detected in large numbers .
since lpvs are very bright and also follow families of pl relations ( see * ? ? ?
* ; * ? ? ?
* for recent reviews and extensive references ) , they are potentially very powerful distance indicators as well . lpv stars , including miras and semi - regular variables ( srvs ) , have been studied previously in the galactic bulge by different teams ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . however , these were mostly limited to bright objects ( @xmath120 mag ) .
vvv represents a significant progress in the search and study of lpvs in the inner bulge , due to its higher spatial resolution and variability campaign carried out in the near - ir , which allows one to pierce deeply into the most crowded and highly obscured regions of the galaxy , reaching up to @xmath121 mag in most observed fields @xcite .
a search for miras and srvs in the inner galactic bulge is already in progress , using vvv @xmath4-band data . in this first analysis
, we selected 12 vvv fields , namely b327 to b338 ( see fig .
[ fig : area ] ) , covering about 18 square degrees in the inner bulge . in a massive variability survey such as vvv ,
miras are relatively easy to find , due to the large amplitude of their light curves , which exceed 0.35 mag @xcite and are on average 0.6 mag @xcite in the @xmath37-band , and periods in the range @xmath122 .
about 100 reliable mira candidates were identified from a total of @xmath123 million light curves that were produced , with 23 - 32 data points per object .
each light curve covers the first two years of observations ( 2010 - 2012 ) with different cadence depending on the vvv field ( see [ sec : status ] , [ sec : var - over ] ) . in order to find the period and amplitude of each variable ,
the light curves of the mira candidates were fitted by applying both fast @xmath124 @xcite and lomb - scargle methods @xcite .
an illustrative example of our results is shown in figure [ fig : lpv ] ; naturally , the derived periods will become better defined as more data covering a longer baseline become available . a first catalog will be presented in gran et al .
( 2013 , in preparation ) .
in addition to miras and srvs , this includes a large number of variable agb stars ( previously identified as oh / ir or maser sources , see @xcite ) , besides young stellar objects and other still unclassified sources . the majority of pre - main sequence ( pms ) stars have measurable variability , which is usually attributable to magnetic activity .
hot spots formed at the base of accretion funnel flows can also cause variability , though this would be more important at blue and near - uv wavelengths .
some pms stars , e.g. kh15d @xcite , have variable extinction , which can lead to high - amplitude variability even in the ir . however , the highest amplitudes are observed in eruptive pms variables , where the mechanism is believed to be large variations in the accretion rate which directly change the luminosity of the star and the inner parts of the circumstellar disk .
eruptive pms variables are usually divided into fu orionis types ( fuors ) and ex lupi types ( exors ) , which show occasional variability in excess of 2 to 6 magnitudes on timescales of months ( exors ) or years to decades ( fuors ; for a review , see @xcite ) .
fuors show a rapid rise in luminosity followed by a slow decay over decades .
many fuors have associated molecular outflows or jets @xcite in which the effects of historical eruptions on the outflow rate can sometimes be seen @xcite .
fewer than 20 eruptive variables are known in each of the fuor and exor classes , but it is possible that this type of variability is common amongst pre - main sequence stars , though intermittent in nature with long intervals of quiescence . if it is common , this would be important for two reasons .
firstly , it may explain the commonly observed scatter in hertzsprung - russell diagrams of pms clusters , a phenomenon that hampers the assignation of masses to pms stars with evolutionary tracks , with consequences for measurements of the initial mass function .
secondly , they may explain the long - standing `` luminosity problem '' ( e.g. , * ? ? ?
* ) , which consists of the fact that low - mass pms stars are typically less luminous than expected for objects that should be above the main sequence while descending a hayashi track .
as far as eruptive pms variables go , our work on the vvv data has the potential to ( i ) precisely quantify the incidence of exors and
( ii ) unveil variability in optically - obscured protostars , an area where only a small amount of work has been done to date .
we might expect the high - amplitude variability to be more common in these obscured , generally younger systems , in which the average accretion rate is higher .
we are also beginning to explore the potential of the common low - amplitude variability as a method for tracing dispersed pms stellar associations .
this may allow us to investigate the duration of pms evolution ( which will be mass - dependent ) , by comparing the number of variables with the number of stars with disks .
our first searches for pms variables were based on vvv data release 1 ( dr1 ; * ? ? ?
* ) , which had a typical time baseline of only a few months in 2010 and only 5 epochs of @xmath4 photometry .
we searched in the disk region of vvv ( @xmath125 , see fig .
[ fig : area ] ) for candidate variables with @xmath126 mag , which is the approximate upper limit for magnetic variability .
we also did a more general search for variables with rms @xmath4 variability @xmath127 mag , in order to see whether that would pick out star formation regions ( sfrs ) .
both searches of the catalogs produced a large number of false positive candidates requiring visual inspection , but their number was reduced by requiring a detection in the _ spitzer _ glimpse legacy survey @xcite , and in particular by focusing on the subset of 18,949 very red sources ( @xmath128-[8.0]>1 $ ] mag ) identified in glimpse by @xcite .
most of the robitaille sources are candidate protostars .
the @xmath126 mag search identified 1881 candidates in vvv , of which 47 are detected in glimpse and 12 were known as red robitaille sources .
these 12 were all confirmed as genuine variables by inspection of the images .
they have @xmath129 between 1.02 and 1.76 mag , and the majority are undetected in the @xmath2 and @xmath3 passbands , although they are well above the @xmath4 detection limit .
a low - resolution spectrum was obtained for one of these red variable sources ( g314 , see fig .
[ fig : g314 ] ) with magellan / fire @xcite in march 2012 .
the spectrum in figure [ fig : g314 ] is based on a preliminary reduction , but it clearly shows several very strong emission lines of h@xmath130 , which are marked with vertical lines .
moreover , the emission was spatially extended by a few arcseconds along the slit , demonstrating the presence of a large - scale outflow .
an additional search was made for lower - amplitude variables with rms magnitude variations @xmath127 mag in the dr1 data .
this returned 24,798 candidates , of which 703 are in glimpse and 41 objects are robitaille red sources ( including the 12 with @xmath126 mag ) .
we cross - matched these candidate variables which have _ not _ been visually inspected with the @xcite catalog of sfrs to see whether there was a significant spatial association .
we found that 4% ( 937/24,798 ) of the candidate variables are within 10 arcminutes of a known sfr and the histogram of candidate to sfr separations shows a rising number of matches with decreasing separation .
this indicates a clear asociation of at least a significant minority of vvv variables with sfrs ( see fig .
[ fig : sfr ] ) .
similarly , we found that 30% ( 214/703 ) of the glimpse sources are within 10 arcminutes of a known sfr , and the separation histogram has a similar trend . to establish the true fraction of vvv variables that are pms stars will require visual inspection of a large sample and careful checking of the photometry to avoid the errors that can arise in crowded fields .
subsequently , we undertook a search for very high - amplitude variables in the combined 2010 to 2012 datasets for the disk region , focusing on the region at @xmath131 ( contreras pea et al .
2013 , in preparation ) .
following visual inspection of a few hundred candidates , a total of 77 genuine variables were identified with @xmath132 mag , which is the level more typically associated with eruptive variability when long time baselines are available .
spectroscopy from 0.8 to 2.5 @xmath133 m has now been obtained for 6 of these systems in april 2013 , using magellan / fire in echelle mode at @xmath134 .
priority was given to sources ( i ) with a likely association with clusters or sfrs of known distance ; ( ii ) with highest amplitude ; ( iii ) with signs of a recent outburst in the light curve ; ( iv ) bright enough for quick observation .
a further 4 systems with slightly lower amplitude in the g305 star forming complex were also observed , and g314 was re - observed .
a few of these 11 systems also showed strong , spatially extended h@xmath130 emission , similar to that seen in g314 .
data reduction and analysis is currently ongoing .
eclipsing binary systems in general , but detached systems especially , are extremely important objects in astrophysics , for they provide one of the most robust means of deriving stellar radii , masses , and even ages ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , thus providing a `` royal road '' to stellar astrophysics ( * ? ? ?
* ; * ? ? ?
* and references therein )
. the vvv survey will provide a huge number of eclipsing systems of all types , and we have already started to travel along this `` royal road '' using vvv data .
examples of vvv light curves for two eclipsing binary systems can be found in figure [ fig : ebs ] . as an example of their use in astrophysics , a method to use the vvv data for tracing the structure of the milky way with detached eclipsing binaries ( debs ) was introduced in @xcite .
the idea behind the method is that one can calculate the distance by comparing the derived absolute magnitudes from the model with the observed ones .
this normally requires spectroscopic observations to calculate radial velocities of both components , used for calculating the masses and orbital parameters , crucial for obtaining the true absolute values of the stellar parameters ( like the radii ) .
one also needs accurate theoretical atmosphere models and/or calibrations to infer surface brightness on the basis of colors , line ratios or other observables @xcite . with the increasing availability of extensive databases of state - of - the - art stellar evolution models , it is now possible to derive a complete set of physical parameters of a binary s components without time - consuming spectroscopy , from their light curves only . in our approach
we implemented two codes , which can analyze large numbers of deb light curves in a short time .
the first one the detached eclipsing binary light curve fitter ( debil ) is a program which rapidly fits deb light curves to a simple , geometric model @xcite .
the second code
method for eclipsing component identification ( meci ; * ? ? ?
* ; * ? ? ?
* ) fits a physical model to each deb using readily available photometric data only .
it is designed to work from the debil model as a starting point , building an improved physical model of the deb therefrom .
meci assumes that the binary s stellar components formed together and evolved along their respective evolutionary tracks , without any mass transfer .
if observed magnitudes in different bands are given , meci returns the absolute magnitudes , so the distance may be calculated directly .
finally , to deal with the strong and variable interstellar reddening , we implemented the idea of reddening - free indices , introduced in @xcite .
these indices are combinations of magnitudes in 3 bands , in the form of @xmath135 where @xmath136 are the apparent magnitudes in the available bands , and
@xmath115 is a multiplication coefficient dependent on the extinction law assumed .
the coefficient @xmath115 is given in such way that , within a given extinction law , defined in terms of the ratios of extinction values in the given bands @xmath137 , the following equation is true : @xmath138 where the `` 0 '' subscripts indicate quantities corrected for reddening , and thus @xmath139 , @xmath140 , @xmath141 . combining equations ( [ eq : mx ] ) and ( [ eq : mx0 ] ) , we have : @xmath142 = m_{1,0 } - c(m_{2,0 } - m_{3,0 } ) .
\label{eq : mxd1}\ ] ] thus , @xmath143 naturally , the definition in equation ( [ eq : mx ] ) is also applicable to _ absolute _ magnitudes , and so @xmath144 where the capitalized @xmath145 s denote absolute magnitudes .
on this basis , we obtain , for the apparent distance modulus , @xmath146 - \left[m_1 - c(m_2 - m_3)\right ] , \label{eq : mx - mx}\ ] ] which implies @xmath147 - \left[m_1 - c(m_2 - m_3)\right ] .
\label{eq : mx - mxd2}\ ] ] this can be rewritten as @xmath148 - \left[m_1 - c(m_2 - m_3)\right ] + \left[a_1 - c ( a_2 - a_3)\right ] .
\label{eq : mx - mxd3}\ ] ] thus , @xmath149 , \label{eq : mx - mxd4}\ ] ] and so , using equation ( [ eq : cdef ] ) , @xmath150 where @xmath151 is the distance in pc .
therefore , _ reddening - free magnitudes are ideally suited for the calculation of distances , without the need for prior knowledge of extinction values towards individual stars_. the vvv variability campaign is still ongoing ( [ sec : var - over ] ) , and a catalog of eclipsing binaries has thus not yet been prepared .
therefore , we used the eclipsing binaries from the ogle - ii variable stars catalog @xcite , for which @xmath71-band light curves are available .
@xcite presented models for about 10,000 of them , obtained with the debil code , and identified 3170 of them as `` detached . '' to build the reddening - free indices we used @xmath152 photometry from vvv , in addition to @xmath153 data from ogle - ii , and assumed a canonical extinction law with @xmath154 . for consistency
, we also built a set of reddening - free isochrones , based on the padova models @xcite .
the procedure just described successfully yielded physical parameters , ages , and distances to 23 deb systems . in figure [ fig_2.4_lc ]
we present examples of the utilized ogle - ii light curves , superimposed on the best - fitting meci models . in figure [ fig_2.4_rec ]
we present the positions of the 16 closest systems in the galactic plane @xmath155 , with the sun located at ( 0,0 ) and the galactic center at ( 8,0 ) kpc .
they are plotted over a reconstruction of the milky way from @xcite .
the locations of at least 13 of the studied targets coincide well with major structures of the milky way , such as the bulge and spiral arms , including scutum - centaurus , norma , far 3 kpc , and perseus .
this shows that our approach is suitable for tracing the structure of the milky way .
it is notable how many objects were found in a poorly studied area behind the bulge , which proves that the combination of near - ir vvv data with optical photometry from other sources can be a powerful tool for studying this part of the milky way .
another 7 systems ( not shown on the reconstruction ) were found at larger distances , corresponding to the sagittarius stream a structure of stars related to the sgr dsph galaxy ( e.g. , * ? ? ?
these sagittarius deb candidates , the first to be identified in that galaxy s stream , will be the subject of detailed spectroscopic follow - up by our team . cataclysmic variable stars ( cvs ) , including novae and dwarf novae , consist of tight binary systems in which one of the components the so - called _ primary _ is a wd star , and the other component the _ secondary _ is a low - mass , k- or m - dwarf ( i.e. , main sequence ) star . in cv systems ,
the dwarf star has filled its roche lobe , and is transferring matter onto the primary .
cvs are thus eruptive binary systems , with the outbursts being detectable over a wide wavelength range , including the near - ir . however , the quiescent light curves of dwarf novae and related objects are dominated by the orbital motion , and their behavior in the near - ir differs from that seen at optical wavelengths . while in the optical the variations are produced by the hotter parts of the system , namely the primary star and/or the accretion disk , in the near - ir the light curve is dominated by the emission of the cool , late - type secondary star , whose spectra peak at longer wavelengths ( e.g. , * ? ? ?
* ) . in close binaries
the secondary star will be distorted by tidal effects when close to or filling its roche lobe .
the orbital motion thus causes ellipsoidal variations which are more prominent in near - ir light curves .
this modulation enables us to determine the orbital parameters of the system and even to map the surface brightness distribution of these stars ( e.g. , * ? ? ?
* ; * ? ? ?
the recent discovery of dozens of new dwarf novae cvs by the ogle team demonstrates the ability of large surveys to search for new variable sources even in the most crowded regions of the galaxy @xcite .
long - period symbiotic systems , in which the secondary star is a late - type giant rather than a main - sequence star , have also been discovered in both the ogle and macho data @xcite . in like vein
, we expect large numbers of interacting binary systems to show up in the vvv data .
indeed , in the case of novae , preliminary results have already been presented in @xcite .
there are about 400 known novae in the galaxy , with @xmath156 of them falling in the vvv survey area .
interestingly , the spatial distribution of novae shows a `` zone of avoidance , '' with just a few objects belonging to the innermost regions .
moreover , the comparison with nearby galaxies suggests that we lose many nova eruptions every year ( * ? ? ? * and references therein ) .
not surprisingly , the regions with a lack of objects are the regions most heavily obscured by dust , beyond the capabilities of the current searches for novae , which are mostly carried out in the optical . @xcite
provide @xmath12 data for 93 galactic novae .
for some of these objects , colors have been reported for the first time in a homogeneous dataset , since a large fraction of these novae were beyond the detection limit of previous near - ir surveys ( e.g. , 2mass ; * ? ? ?
* ) . the low novae rate observed in the galaxy encouraged us to start a search for the hidden novae in the aforementioned `` zone of avoidance '' region .
we thus selected 24 vvv fields covering about 36 square degrees towards the inner galactic bulge . in this first attempt , we made use of the first variability tables available , limited to @xmath157 epochs in the inner bulge at the time of writing .
even with a limited number of epochs , the vvv data allowed us to discover two new galactic novae candidates ( vvv - nov-001 and vvv - nov-002 ; * ? ? ?
* ; * ? ? ?
* ) , and a few more objects are currently under study .
vvv light curves of vvv - nov-001 and vvv - nov-002 are presented in figure [ fig : lcurve ] .
spectra taken with the soar telescope show the presence of emission lines which are typical of novae in vvv - nov-001 ( saito et al .
2013c , in preparation ) , while optical data from ogle confirmed vvv - nov-002 as a d - class nova @xcite .
while vvv - nov-001 was discovered on the galactic plane , with @xmath158 deg , vvv - nov-002 is one of the closest novae to the galactic center known , with @xmath159 deg .
a search for novae will be extended to the high - extinction regions of the inner disk in the near future .
gravitational microlensing events can be of enormous astrophysical significance ( e.g. , * ? ? ?
* ; * ? ? ?
* and references therein ) .
models of the spatial dependence of the microlensing optical depth @xmath160 @xcite show that ir surveys like vvv can be very efficient in the search for microlensing events , and can probe directly the mass distribution contained in the inner regions of the galaxy .
unfortunately , existing optical microlensing searches based on ccd detectors do not cover the whole bulge or the plane ( although ogle - iv already represents an important step forward in this regard ) , and in particular , they miss the inner regions where this optical depth is higher , thus poorly constraining the models .
therefore , one of the main goals of the vvv survey is to search for microlensing events in the inner milky way .
we are especially interested in looking for rare events in our data , such as : * very reddened events ; * short - timescale events , due to planetary or brown dwarf microlensing .
the advantage of using microlensing to search for planets is that there is no preference for nearby objects or bright stars , contrary to what occurs with other techniques .
thus , microlensing allows us to further probe the planet parameter space , searching , for instance , for planets with periods that are too long to be detected by other techniques , or not sufficiently close to the star as to produce detectable doppler shifts in their spectra @xcite ; * very long - timescale events , due to massive black holes .
such a long - term variability study is possible with our @xmath161-year - long survey , adequately covering the baseline for these long - timescale microlensing events ; * binary microlensing events ; * parallax events ; * high - magnification events in obscured dense fields ; * microlensing of source stars in the sgr dsph galaxy ( e.g. , * ? ? ?
* ) .
so far we have concentrated our searches mostly on the regions that have not been previously searched for microlensing .
these fields are located in the innermost regions of the bulge , where the microlensing optical depth is expected to be high .
a quick search has so far revealed about two dozen bulge microlensing events with high amplification .
the light curves and the parameters fitted show a range of timescales from a few days to several months , consistent with previous ( complementary ) results from the ogle survey @xcite .
some examples of microlensing events are provided in figure [ fig : ml ] .
the galactic bulge is known to contain a significant number of chromospherically active and rotating variable stars , including stars showing ellipsoidal light curve modulation , fk com stars , and rs cvn systems @xcite .
we have cross - matched the @xcite catalog , which is based on macho observations , with the vvv photometric catalogs as provided by the casu vdfs 1.2 pipeline , successfully obtaining @xmath4-band light curves for numerous stars previously classified as rotating variables .
some examples of the corresponding vvv light curves are provided in figure [ fig : rscvn1 ] .
these light curves confirm that vvv near - ir data can be successfully used even to study systems whose light curve amplitude does not exceed @xmath162 mag in @xmath4 , thus also helping properly classify objects whose previous classification may be unclear or ambiguous .
a significant number of galactic star clusters fall inside the area surveyed by the vvv : 36 known globular clusters ( gcs ) and 355 known open clusters ( ocs ; * ? ? ?
* ) , in addition to newly discovered gc @xcite and oc @xcite candidates .
the inner galactic star systems are generally deeply buried behind a curtain of dust and gas that hides them in optical observations , and so our knowledge of their variable star populations in particular is at present very incomplete ( e.g. , * ? ? ?
fortunately , the highly diminished extinction in @xmath4 ( @xmath163 ) will help us unveil the variable stars that are present in these poorly studied objects . also fundamental in our variability studies
is vvv s temporal coverage , with @xmath164 epochs over a 5-year period ; the conditions of the observations , with most of them taken with seeing under 1 arcsec ; the excellent spatial resolution of the vircam camera , @xmath8 per pixel ; and a complete spatial coverage of the inner regions of the galaxy .
vvv data thus place us in a privileged position , as far as the study of variable stars in heavily obscured star clusters goes , allowing us to probe deep into the centers of these objects and out to their tidal radii and beyond .
we have already started the search for variable stars in ocs and gcs alike , focusing initially on the vvv fields that lie in the bulge region of the survey , since the latter have been observed more often than the disk fields ( [ sec : var - over ] ) . in what follows ,
we describe some of our initial results , as far as the gc variability search is concerned ( see also * ? ? ?
gcs contain large numbers of rr lyrae stars , which show a tight relation between periods and absolute magnitude in the near - ir @xcite , thus helping constrain the vvv gcs distances and extinctions . for many of these clusters , especially the faintest ones ,
the other photometric technique extensively used to derive their parameters , the study of their color - magnitude diagrams ( cmd ) , is highly complicated by the presence of elevated field stellar contamination and differential reddening , even at near - ir wavelengths ( see fig . [ fig_cmd_ter ] ) .
when present , rr lyrae variables can provide us with better means to obtain these physical parameters .
unfortunately , rr lyrae analysis in the near - ir has its own difficulties : the amplitude of the rr lyrae decreases towards the near - ir when compared to the optical , and the shape of the rrab light curves becomes more sinusoidal ( see fig .
[ fig_var_6441 ] ) .
even so , the quality of our light curves will allow us to look for these variables and to properly characterize their amplitudes , periods , mean magnitudes , and colors . in @xcite
, we present our first attempts to tackle the study of variable stars in the vvv gcs , and show our efforts to characterize , in the near - ir , variables that had already been previously identified in optical studies ( as in the case of ngc 6441 ) , or to discover new variables in extremely obscured gcs where optical observations are very complicated or unfeasible ( as in the cases of terzan 10 or 2mass - gc02 ) .
an intriguing characteristic that only galactic gcs seem to show is the so - called oosterhoff dichotomy ( e.g. , * ? ? ?
* ; * ? ? ?
* and references therein ) , i.e. , galactic gcs divide themselves in two main groups according to the mean period of their rrab variables , oosterhoff i systems having shorter periods ( @xmath165 days ) and oosterhoff ii systems with longer periods ( @xmath166 days ) . in the milky way , but not in nearby satellite galaxies , very few systems are found in the so - called `` oosterhoff gap '' zone , with @xmath167 .
interestingly , the periods of the rr lyrae candidates we found in 2mass - gc02 and in terzan 10 make these two gcs outliers in the established picture , with terzan 10 being an oosterhoff ii gc but seemingly having too high a metallicity ( @xmath168 } = -1.0 $ ] dex ; @xcite , feb
. 2010 update ) to belong to the group , and 2mass - gc02 falling in the almost empty oosterhoff gap region @xcite .
terzan 10 could thus be a less extreme example of the new `` oosterhoff iii '' group proposed by @xcite ( @xcite ; see also * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , which so far contains exclusively bulge gcs with even higher [ fe / h ] and longer @xmath169 , namely ngc 6388 and ngc 6441 .
any new rr lyrae stars that we may be able to find in the inner gcs , and especially those with very few or no rr lyrae known , will be particularly useful to further explore and understand the oosterhoff dichotomy , and thus help us place bulge gcs in the wider picture of milky way formation ( * ? ? ? * and references therein ) .
in addition to the many different types of well - classified variable stars discussed in the previous sections , the vvv survey data have already resulted in the detection of various transient objects @xcite , some of which appear to defy commonly adopted classification schemes .
one example is provided by vvv - wit-01 @xcite , an extreme transient event whose light curve is shown in figure [ fig : wit01 ] .
its color , as shown by the vvv multi - color data , is also extremely red , with @xmath170 mag . as discussed by @xcite
, the source could be an eruptive pre - main sequence star , a reddened luminous blue variable , a nova , or even a highly obscured supernova event or a previously unknown kind of variable star .
like vvv - wit-01 , many additional transient events are expected to be detected in the vvv data , thus opening an exciting path towards the discovery and analysis of energetic events towards the innermost and most heavily obscured regions of the milky way . at the other extreme , long - term photometric monitoring of ultra - cool dwarf ( ucd ) stars
can also be carried out with the vvv survey .
indeed , the variability of ucds is key to understand the atmospheric conditions and cloud formation in sub - stellar objects .
several groups have monitored ucds on a daily basis , over timescales ranging from a few days to a few months ( e.g. , * ? ? ?
* ; * ? ? ?
using vvv data we will be able not only to discover new ucds , but also probe a new regime of variability for sub - stellar objects , thus helping to constrain long - term atmospheric changes in these very low - mass objects . in figure [ fig : bd ] is given an example of the long - term behavior of the newly discovered brown dwarf @xcite .
as discussed by @xcite , the available data do not present signs of periodicity , and any variability is restricted to amplitudes @xmath171 mag .
the vvv eso public survey provides a treasure trove of scientific data that can be exploited in numerous different scientific contexts . in terms of stellar variability
, the project will provide up to several million calibrated @xmath4-band light curves for genuinely variable sources , including pulsating stars , eclipsing systems , rotating variables , cataclysmic stars , microlenses , planetary transits , and even transient events of unknown nature . at the present point in time , with the data - gathering phase of the vvv survey having just crossed its half - way mark , we are really just taking the first steps in what will certainly be a long and exciting journey , during which it will be possible to address a myriad of time - domain astronomical applications , including not only research on variable stars as such but also their use as distance indicators and tracers of galactic structure , origin , and evolution .
vvv is a public survey , and so the data will quickly be made available to the entire astronomical community as we move along , thus opening the door to many additional applications and synergies with other ongoing and future projects that target the same fields as those covered by vvv
. * acknowledgements . *
the vvv survey is supported by the european southern observatory , the basal center for astrophysics and associated technologies ( pfb-06 ) , and the chilean ministry for the economy , development , and tourism s programa iniciativa cientfica milenio through grant p07 - 021-f , awarded to the milky way millennium nucleus .
we gratefully acknowledge the support provided by fondecyt through grants # 1110326 ( m.c .
, j.a .- g . ) , 1120601 ( j.b . ) , 1130140 ( r.k . ) , 3130320 ( r.c.r ) , and 3130552 ( j.a .- g . ) .
c. navarrete acknowledges grant conicyt - pcha / magster nacional/2012 - 22121934 .
, s.e . , and k.p .
acknowledge additional support by project vri - puc 25/2011 .
alonso - garca , j. , dkny , i. , catelan , m. , contreras ramos , r. , & minniti , d. 2013 , in fifty years of wide field studies in the southern hemisphere : resolved stellar populations in the galactic bulge and the magellanic clouds , asp conf .
, in press ( arxiv:1307.0419 ) | the vista variables in the va lctea ( vvv ) eso public survey is an ongoing time - series , near - infrared ( ir ) survey of the galactic bulge and an adjacent portion of the inner disk , covering 562 square degrees of the sky , using eso s vista telescope .
the survey has provided superb multi - color photometry in 5 broadband filters ( @xmath0 , @xmath1 , @xmath2 , @xmath3 , and @xmath4 ) , leading to the best map of the inner milky way ever obtained , particularly in the near - ir .
the main variability part of the survey , which is focused on @xmath4-band observations , is currently underway , with bulge fields having been observed between 31 and 70 times , and disk fields between 17 and 36 times .
when the survey is complete , bulge ( disk ) fields will have been observed up to a total of 100 ( 60 ) times , providing unprecedented depth and time coverage .
here we provide a first overview of stellar variability in the vvv data , including examples of the light curves that have been collected thus far , scientific applications , and our efforts towards the automated classification of vvv light curves .
[ 1996/06/01 ] 2@xmath5 2c ii 4c iv 2fe ii 3fe iii 1 mg i 2 mg ii 2si ii 4si iv 2al ii 3al iii 1o
i 1n i 1h i = |
metamaterials are artificial materials that can be engineered to exhibit fascinating electromagnetic properties that do not occur in nature , such as negative refractive index , `` perfect '' imaging and electromagnetic cloaking @xcite , chiral metamaterials have shown giant optical activity @xcite , circular dichroism @xcite , negative refraction @xcite , and possible reversal of the casimir force @xcite .
most of the applications critically required an negative index of refraction in three dimensions ( 3d ) .
it is an open question , whether there is any 3d metamaterial design working for optical frequencies still feasible to fabricate . here
, we theoretically proposed a new design that provides negative index of refraction with one surface and can be used to fabricate 3d metamaterials .
this is an important step towards the realization of bulk 3d at optical wavelengths .
previous theoretical designs @xcite obtain nim in 3d at ghz and thz .
two groups independently proposed fully isotropic bulk magnetic metamaterial designs , based on srrs arranged in a cubic lattice by employing spatial symmetries @xcite . on the other hand ,
some designs of 3d isotropic nims exist , but fabricating them has remained a challenging task and virtually impossible at optical frequencies . for example , koschny et al .
@xcite , designed an early example of an isotropic nim . however , high - constant dielectric assumed across the gaps of the srrs rendered the experimental realization impractical .
alternative approaches have also been investigated , including direct lasing fabrication @xcite , and high refractive index spheres @xcite .
double negative index materials exhibit simultaneously negative magnetic permeability , @xmath1 , and electric permittivity , @xmath0 , over a common frequency range @xcite .
negative permeabilities are the result of a strong resonant response to an external magnetic field ; negative permittivity can appear by either a plasmonic or a resonant response ( or both ) to an external electric field .
pendry et al .
@xcite suggested a double metallic srr design for negative @xmath1 and a parallel metallic wire periodic structure to give negative @xmath0 .
several variations of the initial design have been studied @xcite .
for example , a single ring resonator with several cuts has been proven capable of reaching negative @xmath1 at a higher frequency @xcite , and cut wire pairs and fishnet structures @xcite allow for a negative magnetic response at optical frequencies . in these old designs ,
it needs two particles , one particle gives negative @xmath1 and the other gives negative @xmath0 at the same frequency range .
later , rockstuhl and lederer proposed a new design@xcite , nanoapertures embedded in a continuous metal film , that gives @xmath2 from the metallic surface and a resonant @xmath1 , which , however , did not reach @xmath3 because of the weak resonance .
since the imaginary part of @xmath1 is large , @xmath4 and @xmath2 will give negative index of refraction@xcite .
it@xmath5s very interesting to develop a new design that will use only one `` particle '' on one surface that will provide negative index of refraction and also @xmath3 and @xmath2 , with parallel direction on the surface .
we will use babinet@xmath5s principle @xcite to obtain a complementary structure in a metallic surface that will give @xmath3 for the first time , instead of @xmath6 resulting from babinet - complementary resonators studied in the previous literature , with perpendicular direction on the surface .
in addition , the remaining continuous metallic surface will give @xmath2 .
1 2 babinet@xmath5s principle has been used @xcite to design artificial planar metamaterials , which can provide only negative electric permittivity .
such inverse elements , like complementary srrs ( csrrs ) , have been proposed as an alternative to conventional metallic wires for the design of planar metamaterials , which give @xmath2 .
complementary metamaterials show @xcite similar properties as their inverse structures . the transmission coefficient , @xmath7 , for the csrr is related to the transmission coefficient , @xmath8 , for the srr by @xmath9 .
in addition , according to babinet@xmath5s principle , their transmission and reflection behavior , as well as their scattered electric and magnetic fields , are interchanged @xcite .
these complementary designs provide an effective negative permittivity , @xcite with perpendicular direction . in this manuscript
, we will use babinet@xmath5s principle to design complementary structures that will provide an effective negative permeability and the metallic surface to give negative permittivity . therefore , we will obtain negative @xmath10 with only one planar metamaterial , with parallel direction of the surface of planar metamaterial . 3
babinet@xmath5s principle relates the fields scattered by two complementary planar structures of arbitrary shape made of infinitely thin , perfectly conducting sheets .
if we consider an incident electromagnetic ( em ) field , @xmath11 , then its complementary em field , @xmath12 , is defined @xcite as @xmath13 and @xmath14 , which corresponds to a rotation around the propagation axis .
if we have a thin metallic surface , the reflected ( @xmath15 ) and the transmitted ( @xmath16 ) electric and magnetic fields satisfy the following conditions .
if we have charge on the thin metallic surface , the perpendicular reflected electric field is opposite to the transmitted electric field , i.e. , @xmath17 , where @xmath18 is the axis vector perpendicular to the metallic surface .
the parallel electric field is continuous to both sides of the thin metallic surface , i.e. , @xmath19 .
if the incident em wave generates current in the thin metallic surface , the perpendicular magnetic field is continuous , i.e. , @xmath20 and the parallel magnetic field is opposite to the two sides of the metallic surface , i.e. , @xmath21 .
if we have a perfectly conducting surface , no magnetic dipole occurs in the plane due to this condition @xmath21 .
in addition , no electric dipole occurs perpendicular to the plane , due to the condition @xmath22 .
however , in - plane electric dipoles and magnetic dipoles perpendicular to the plane are possible .
due to the babinet principle , we have two related complementary plane structures .
the surface charge in one structure is related to the magnetic flux in the complement and the opposite occurs . the magnetic flux is related to the surface charge .
also , the parallel electric field in the plane can be generated , surface electric current perpendicular to the electric field ( see examples below ) .
in particular , this means a resonance in one structure is related to a corresponding resonance in the dual fields in the complementary structure . in a
, we have a thin metallic rectangular patch .
the propagation direction of the em field is perpendicular to the metallic patch and the electric field , @xmath23 , is parallel to the long side of the rectangular patch .
we have created an electric dipole in the y - direction of a. the electric current , @xmath24 , is along the direction of @xmath23 and the magnetic flux is shown in a. in b , we presented the babinet - complementary structure , where we have a thin metallic surface and a rectangular hole .
as discussed above , the surface charge shown in a can be transformed to the magnetic flux .
the positive ( negative ) charge , shown in a , can be related to the magnetic flux going inside ( outside ) the metallic surface ( see b ) .
the magnetic flux of a can be related with the charge of b. another example , shown in a , is the metallic srr .
we present the current , the charge and the magnetic flux . in a
, we have a strong magnetic dipole along the propagation direction of the em wave .
there is also a minor electric dipole in the srr gap .
this dipole moment is responsible for the bianisotropy of the srr . in b
, we present the complementary structure , and used the idea of exchanging surface charge and magnetic flux .
we do not have a magnetic dipole in the complementary srr ( csrr ) , but there is an electric dipole , due to the neck of csrr , parallel to the neck . 4 + ( a)(b ) + + ( a)(b ) + + ( c)(d ) + in a , we have a new metallic design that gives an electric dipole and the complementary design gives a residual magnetic dipole shown in b , for perpendicular direction on the metallic surface .
in the previous section we reviewed the babinet principle for different designs ( figures [ fig1][fig3 ] ) and the complementary designs . these designs ( figures [ fig1][fig3 ] ) can help us understand how the electric dipole on the metallic film is related to he magnetic dipole of the `` complementary '' design .
we can use these ideas of babinet@xmath5s principle to use metallic surface to give @xmath2 and @xmath3 and will have negative index of refraction , @xmath25 . in a
, we present a new metallic design that has no electric dipole but a resonant electric quadrupole moment . considering at first normal incidence , the complementary structure show in b will possess a resonant mode with a residual magnetic dipole moment perpendicular to the plane as consequence of babinet@xmath5s principle .
we have thus established the correspondence of a resonant electric quadrupole mode in the `` octopus '' structure to a resonant magnetic dipole mode in the complementary structure .
the babinet principle allows us to map fields between the corresponding resonances of mutually complementary structures .
therefore , after indentifying the resonances of the babinet complement , we can elect to excite these resonances in a different geometry of incidence . now leaving the original babinet setup behind
, we realize that the correspondence of the two resonant modes in the direct and complementary structure , i.e. , the derived `` translation rule '' of local charges , currents and flux distributions on the surface , has to hold for any external excitation . in
order to couple to the magnetic dipole of the complementary structure ( b ) we thus change the propagation direction to be parallel with the surface such that the magnetic field of the incident wave can couple to the ring current ( i.e. the magnetic dipole ) and we can obtain a negative @xmath1 . notice the charges are shown in the `` legs '' of this design .
the current goes around and provides a magnetic dipole , which can give a negative permeability .
in addition the metallic surface ( b ) will give negative permittivity , @xmath0 .
this complementary design can help us to understand how to obtain negative @xmath1 and negative @xmath0 .
the design shown in overlap with @xmath2 and @xmath3 and the impedance @xmath26 is very large and we have a lot of reflection and low transmission . deforming the structure while maintaining the same principle topology of , we did a lot of simulations for different designs to have very low impedance ( @xmath27 for air ) and to have very high transmission .
we find a new design ( see ) which satisfies the criterion @xmath2 and @xmath3 , and they have @xmath28 and @xmath29 the same magnitude .
we will use an alternative design to obtain negative @xmath1 and the metallic surface will give negative @xmath0 at the same frequency region , and the magnitude of @xmath0 and @xmath1 will be comparable , and the impedance to be close to 1 . in a , we present a complementary design , which has a big hole in the metallic surface that will give a negative @xmath1 .
this structure is on a @xmath30 thick copper sheet with lattice constant @xmath31 and @xmath32 . in b , we plot the current density , which gives a circular current and provides magnetic dipole momentum .
we have used the design of a , to obtain transmission , reflection , and absorption @xmath33 with the cst .
we simulated a periodic array of the proposed structures .
with propagation direction parallel to the metallic surface , we have used the amplitude and the phase of transmission and reflection to obtain the frequency dependence of @xmath0 , @xmath1 , and @xmath10 through the retrieval procedure @xcite . in , we plot the transmission , t , and reflection , r , and absorption , a , versus the frequency . at the frequency
9.9 ghz , there is dip in the reflection . in the retrieval results ,
we obtain negative @xmath10 with only one metallic surface with the appropriate hole design .
the hole design gives negative @xmath1 and the metallic surface gives negative @xmath0 , the impedance is close to one , and the transmission is close to @xmath34 .
the figure of merit ( fom = @xmath35re(n)/im(n)@xmath35 ) is 9 with @xmath36 with frequency 9.9ghz .
the @xmath37 with @xmath38 with frequency 9.95ghz .
this is a unique design that gives negative @xmath0 and @xmath1 , with only one metallic surface .
in conclusion , we have shown it is possible to construct babinet - complement structures ( holes in planar metamaterials ) that provide a resonant magnetic moment perpendicular to the screen .
we use such a magnetic resonance providing @xmath3 together with the @xmath2 coming from the electric background provided by the continuous regions of the screen , to obtain a double negative index metamaterial for propagation direction parallel to the screen . in this material , both @xmath2 and @xmath3
are provided by the same single component .
these designs can be used to fabricate three - dimensional metamaterials and obtain negative index of refraction at microwave and infrared wavelengths . _
acknowledgement : _ work at ames laboratory was partially supported by the department of energy ( basic energy sciences , division of materials sciences and engineering ) under contract no.de-ac02-07ch11358 ( computational studies ) this work was partially supported by the office of naval research , award no .
n00014 - 10 - 1 - 0925 . | metamaterials are patterned metallic structures which permit access to a novel electromagnetic response , negative index of refraction , impossible to achieve with naturally occurring materials . using the babinet principle , the complementary split ring resonator ( srr )
is etched in a metallic plate to provide negative @xmath0 , with perpendicular direction . here
we propose a new design , etched in a metallic plate to provide negative magnetic permeability , @xmath1 , with perpendicular direction .
the combined electromagnetic response of this planar metamaterial , where the negative @xmath1 comes from the aperture and the negative @xmath0 from the remainder of the continuous metallic plate , allows achievement of a double negative index metamaterial ( nim ) with only one metasurface and strong transmission .
these designs can be used to fabricate nims at microwave and optical wavelengths and three dimensional metamaterials . |
most known interstellar molecules have been detected in the millimeter wavelength range .
this is clearly related to the fact that , in general , the simplest and hence the lightest molecules are the most abundant cosmic molecular species .
their strongest rotational lines at the temperatures characteristic of the molecular interstellar medium arise at mm and submm wavelengths .
heavier molecules , often with large permanent dipole moments , have detectable rotational lines at lower frequencies in the microwave range .
transitions between low energy levels are more favorable in terms of line intensities , especially in regions of low or modest density and low temperature .
the latter include , in particular , the inner parts of dark clouds , which are among the coldest regions of the interstellar medium , because they have no internal energy sources and are shielded from external sources of radiation by their own gas and dust .
another reason why observations at low frequencies are needed is the fact that the line intensities of molecules having fractional abundance below approximately @xmath11 relative to h@xmath1 fall below the confusion level in the millimeter wavelength range . in such cases , sensitive observations at low frequencies , which are less `` contaminated '' by the emission of simple molecules ,
in addition , hyperfine splitting is often negligible at high frequencies , but detectable at low frequencies ( e.g. for species hc@xmath2n , hc@xmath6n , etc . ) .
hyperfine splitting may be used to determine line opacities and thus to obtain more accurate molecular column densities and abundances .
spectral scans , covering wide ranges of frequencies , are powerful tools for molecular searches .
typically many lines of the same molecule fall within the observed spectral window , making it possible to more reliably identify observed spectral features .
a number of spectral scans of molecular clouds have so far been carried out in atmospheric windows between 17 ghz and 700 ghz .
a selection of the frequency ranges covered and references are 17.622 ghz @xcite ; 72144 ghz @xcite ; 330360 ghz @xcite ; 455507 ghz , @xcite ; 607725 ghz , @xcite ; 780900 ghz , @xcite .
recently , @xcite carried out a spectral scan of the dark cloud tmc-1 in the frequency range between 8.8 and 50 ghz .
however , no spectral scan has been performed at frequencies below 8.8 ghz .
therefore we undertook a spectral survey of tmc-1 at 46 ghz .
this is the lowest frequency spectral survey carried out toward a molecular cloud .
tmc-1 has a shape of a ridge elongated in the northwest - southeast direction .
the structure and chemical composition of tmc-1 have been studied by @xcite , @xcite , @xcite , and @xcite .
the temperature of this object is about 10 k , which is typical for dark clouds ; the h@xmath1 density estimates towards the `` cyanopolyyne peak '' ( see below ) vary between @xmath12 @xmath13 @xcite and @xmath14 @xmath13 @xcite . the study by peng et al .
( 1998 ) revealed 45 clumps which are grouped into 3 cylindrical features oriented along the ridge .
the lsr velocities of these features are approximately 5.7 , 5.9 , and 6.1 km s@xmath15 .
tmc-1 has proven to be an excellent object for the study of chemistry in dark clouds under quiescent conditions .
it shows a carbon rich chemistry with chemical gradients across the densest part of the ridge extending over 0.2 pc x 0.6 pc .
the most prominent chemical feature of tmc-1 is the high abundance of various carbon chain molecules , such as cyanopolyynes hc@xmath16n and radicals c@xmath17h .
the peak position of the line intensities of cyanopolyynes ( the cyanopolyyne peak ) was found to be located @xmath18 southeast from another prominent position in tmc-1 , the peak of the ammonia line emission .
the heaviest known interstellar molecule , hc@xmath19n , as well as a number of other heavy molecules have been found exactly towards the cyanopolyyne peak , making it an especially interesting object for observations at low frequencies .
the observations were performed with the 305-m radio telescope of the arecibo observatory . the entire arecibo c - band ( 46 ghz ) was observed towards the cyanopolyyne peak position ( @xmath20 , @xmath21 ) .
the c - band observations were performed in total power mode which is effective as a result of the excellent system stability , and which maximizes the signal to noise ratio .
two senses of linear polarization were independently observed and the two outputs averaged together .
the four correlator boards of the spectrometer were offset in frequency by 5.25 mhz , and each was set to a bandwidth of 6.25 mhz , thus yielding an instantaneous continuous frequency coverage of 21 mhz with a frequency resolution of 3 khz ( 0.183 km s@xmath15 at 5 ghz ) .
each frequency setup was observed for 5 minutes .
the entire c - band was observed 3 times , with one observation shifted by 3 mhz , yielding an rms sensitivity of about 1718 mk ( about 22.3 mjy ) .
in addition , a number of selected frequency ranges within the c - band and x - band ( 810 ghz ) were observed with longer integration times and rms sensitivities 78 mk or higher .
the x - band observations were performed in two senses of circular polarization .
the beam size was about @xmath22 in c - band and about @xmath23 in x - band .
the gain was about 8 k / jy in c - band and about 4 k / jy in x - band .
the frequency ranges were centered at the frequencies of different molecular transitions taken from the jpl catalog of line positions and intensities ( http://spec.jpl.nasa.gov/ftp/pub/catalog/catform.html ) and the cologne database for molecular spectroscopy ( http://www.ph1.uni-koeln.de/vorhersagen/ ) , as well as from the catalog of known molecular lines by f. lovas ( http://physics.nist.gov/cgi-bin/micro/table5/start.pl ) .
each scan of the total power observations consisted of 50 6second on source observations followed by two 10second calibration scans .
the first of these calibration scans was taken with a noise diode switched on , and the second with the noise diode off .
the data were calibrated by dividing the on source spectra by the calibration spectra ; the latter were calculated as the difference of the two calibration signal scans . to optimize the accuracy of the calibration while minimizing the effect of noise in the calibration scans , the calibration scans were approximated by polynomials of order 20 to 29 .
polynomoials of low order ( up to 3 ) were applied to remove residual baselines from the calibrated spectra . both the calibration and subsequent data reduction were performed with the class software package .
lines with resolved hyperfine structure ( i.e. , when the hyperfine components are presented as different entries in table [ gauss ] ) were fitted using the standard class method hfs .
this method enables us to derive the optical depth of the strongest component from the relative intensities of the observed components assuming that the excitation temperatures , lsr velocities , and linewidths of individual components are the same .
tmc-1 has a very narrow lines of less than 0.4 km s@xmath15 width . in order to achieve high velocity resolution of 0.18 km s@xmath15 with the available number of spectrometer channels the entire band of 46 ghz was divided into 381 spectra .
a portion of one of the c band spectrum is shown in fig .
[ scan ] . as a result of the spectral scan itself
, we detected only lines of h@xmath1co and hc@xmath2n .
however , in more sensitive observations at selected frequencies in c and x bands we detected lines of c@xmath1s , c@xmath3s , c@xmath4h , c@xmath4h@xmath1 , hc@xmath3n and its @xmath5c substituted isotopic species , hc@xmath2n , hc@xmath6n , and hc@xmath7n .
the @xmath24 line of hc@xmath6n and all lines of hc@xmath7n , c@xmath4h@xmath1 , ccs , and c@xmath3s are new , i.e. , they were not included in the catalog of known interstellar lines by f. lovas at the time of writing this paper .
the list of the detected lines with their fitted gaussian parameters is presented in table [ gauss ] .
most of our results were analysed with rotational diagrams . to construct rotational diagrams , we derived a rotational level population from the total intensity of relevant hyperfine components . for optically thin emission , the integrated line intensity , @xmath25 ( k cm s@xmath26 ,
is related to the upper level column density @xmath27 by the well - known relation : @xmath28 where @xmath29 is the line strength , @xmath30 is the permanent electric dipole moment , and @xmath31^{-1}$ ] . at low frequencies , where @xmath32 , @xmath33 is approximately equal to @xmath34 .
@xmath35 and @xmath36 are the excitation temperature and the background brightness temperature , respectively .
if @xmath37 , the term @xmath38/[j(\nu , t_{ex } ) - j(\nu , t_{bg})]$ ] approaches unity . in this case , assuming lte and applying boltzmann s equation , one can obtain from eq .
( 1 ) : @xmath39 where @xmath40 is the energy of the upper level of the transition , @xmath41 is the partition function , and @xmath42 is the excitation temperature characterizing the rotational transitions of interest . using eq . ( 2 ) , one can obtain the total molecular column density @xmath43 and rotational temperature @xmath42 from observations of several lines of the same molecule .
however , in the case of hc@xmath2n and some other molecules the rotational temperature is fairly low ; hence , the assumption @xmath44 is not valid , and the term @xmath38/[j(\nu , t_{ex } ) - j(\nu , t_{bg})]$ ] must be taken into account .
therefore we derived @xmath43 and @xmath42 applying the following iterative approach : 1 .
determination of `` initial guesses '' of @xmath42 and @xmath43 from eq .
( 2 ) 2 .
calculation of @xmath45/[j(\nu , t_{rot } ) - j(\nu , t_{bg})]$ ] from the estimate of @xmath42 3 .
determination of the next estimate of @xmath42 and @xmath43 using the calculated value of + @xmath45/[j(\nu , t_{rot } ) - j(\nu , t_{bg})]$ ] steps 2 and 3 were repeated until convergence was achieved .
the calculated rotational temperatures and column densities are given in table [ trot ] for c@xmath4h and c@xmath4h@xmath1 , no rotational diagrams were constructed .
to estimate column densities of these species , we assumed that the energy level populations are thermalized at a temperature taken from the literature and presented in table [ trot ] .
first , the optical depths of the observed lines were determined from the relation @xmath46 , where @xmath47 is the main beam brightness temperature .
then , using the prescription from the secton `` intensity units and conventions '' of the jpl catalog documentation ( http://spec.jpl.nasa.gov/ftp/pub/catalog/doc/catdoc.pdf ) absorption coefficients per molecule per km s@xmath15 , @xmath48 , were found .
column densities @xmath43 were obtained using the relation @xmath49 .
the same technique was applied to obtain the upper limits to the column density presented in table [ neg ] .
the spectra of the @xmath50 hc@xmath3n emission are shown in fig [ hc3n ] .
one sees a complex line profile , already found in a number of other molecular lines , and explained in terms of the existence of small - scale structure in tmc-1 @xcite . following @xcite ,
who distinguished three main kinematic components towards tmc-1 , we approximated the hyperfine
split line by a sum of three kinematic components with their intensities , lsr velocities , and linewidths free .
the lsr velocities of the components proved to be 5.69 , 5.85 , and 6.06 km s@xmath15 , which are approximately coincident with the velocity components determined by peng et al . and by dickens et al .
a very limited number of offset measurements were performed which showed that ( 1 ) the peak of the 5.69 km s@xmath15 emission is located approximately 40 arcsec south ; ( 2 ) the peak of the 5.85 km s@xmath15 emission is located approximately 30 arcsec southeast ; and ( 3 ) the peak of the 6.06 km s@xmath15 emission is located approximately 40 arcsec north of the nominal position . however , both the extent and the sampling of our `` map '' are insufficient to determine the spatial distribution of the hc@xmath3n emission .
the optical depth of the main hyperfine component is fairly small ( about 0.2 for the 5.69 and 5.85 km s@xmath15 components ) ; however , it is in agreement with optical depths about unity for the higher frequency transitions reported by , e.g. , @xcite
. therefore , simple methods of analysis , such as rotation diagrams , are not adequate ( the formal use of the basic rotation diagram method yields a rotational temperature of 2 k ) , and a statistical equilibrium ( se ) analysis which allows for effects of finite optical depth should be used ( see e.g. * ? ? ?
in addition to the main species , we detected the @xmath51 hcc@xmath5cn emission and marginally detected two other @xmath5c substituted isotopomers , hc@xmath5ccn and h@xmath5cccn .
the weakness of the lines of the @xmath5c subsituted species did not allow us to measure reliably the isotopic abundance ratios , but the results seem to be in a qualitative agreement with the previous results of @xcite , obtained from the observations of the @xmath52 , @xmath53 , and @xmath54 lines at higher frequencies , i.e. , the hcc@xmath5cn emission is the strongest and the hc@xmath5ccn and h@xmath5cccn emission is weaker by a factor of 1.3 .
the ratio of the hc@xmath3n / hcc@xmath5cn line intensities , @xmath55 is also in agreement with the ratio of hc@xmath3n / hcc@xmath5cn abundances of 55 found by takano et al . , indicating a small optical depth for the lines observed using arecibo .
we detected @xmath56 emission of hc@xmath2n ( fig . [ cyanop ] ) .
the ratio between the intensities of different hyperfine components corresponds to lte with small optical depth . combining our line intensity with the intensities of different lines of hc@xmath2n published by @xcite , in a rotational diagram ( fig .
[ rot ] ) we derive a hc@xmath2n rotational temperature of 4.3 k , in agreement with e.g. , @xcite , and a column density of @xmath8 @xmath9 .
the @xmath57 , @xmath58 , and @xmath59 lines were observed and spectra are shown in fig .
the rotation diagram ( fig .
[ rot ] ) was constructed assuming hc@xmath6n source size of @xmath60 @xcite , and yields a rotational temperature of 7.7 k and a hc@xmath6n column density of @xmath61 @xmath9 .
we detected the @xmath62 and @xmath63 transitions of hc@xmath7n , and marginally detected the @xmath64 , @xmath65 , and @xmath66 lines .
the @xmath67 line was not found because of the weakness of the rotational transition and its hyperfine splitting ( fig .
[ hc9n ] ) .
rotational diagrams were constructed assuming a source size equal to @xmath60 ( fig .
[ rot ] , upper right ) and @xmath68 , suggested by @xcite ( fig . [ rot ] , lower right ) .
one can see that the best fit to our data is obtained for the former source size .
thus , we derived a rotational temperature equal to 8.4 k and a hc@xmath7n column density equal to @xmath10 @xmath9 . our results on cyanopolyyne rotational temperatures are in agreement with the previous results by @xcite , who found that the rotational temperatures of the cyanopolyynes increase with the number of heavy atoms over the range hc@xmath2n hc@xmath7n and explained this effect in terms of less efficient radiative decay in the longer cyanopolyyne chains .
an unexpected result is the weakness of the tentatively detected @xmath69 line of ccs at 5402.6 mhz relative to the 10 line of c@xmath3s at 5780.8 mhz ( fig .
[ other ] ) .
the ccs abundance is higher by an order of magnitude @xcite , and usually ccs lines are much stronger than c@xmath3s lines at approximately the same frequencies ( e.g. , @xcite ) .
the ccs line is intrinsically weaker than the c@xmath3s line ( @xmath70 for the c@xmath3s line and only 6.71 for the ccs line ) .
the relevant rotational diagram ( fig .
[ rot ] ) yields a rotational temperature of 5.7 k and a ccs column density of @xmath71 @xmath9 , which is in good agreement with that derived by means of a lvg analysis by @xcite .
we derived the c@xmath3s column density from lvg calculations , assuming @xmath72 k and n@xmath73 @xmath13 @xcite .
our value , @xmath74 @xmath9 , is in good agreement with that derived by fuente et al .
thus , the weakness of the @xmath75 line of ccs is a consequence of its intrinsic weakness , the high location of its energy levels ( @xmath76 equals 10.84 k for the @xmath77 ccs level ) , and the fairly low rotational temperature of ccs .
we detected a strong @xmath78 , @xmath79 , @xmath80 transition of c@xmath4h ( fig .
[ hc3n ] ) , already observed in tmc-1 by bell et al .
( 1982 , 1983 ) .
the arecibo line is narrow as a result of the weakness of the 6.06 km s@xmath15 component towards the observed position .
since this component is fairly strong in the @xmath81 line of hc@xmath3n , our data suggest that there is a difference between the small scale ( about @xmath82 ) distributions of hc@xmath3n and c@xmath4h .
note that the same c@xmath4h line , observed with a much larger @xmath83 beam ( see fig . 2 in @xcite ) , is broader , having a width of about 0.4 km @xmath84 , which is typical of tmc1 .
we believe that the detection of the @xmath85 para - c@xmath4h@xmath1 line is real , since the c@xmath4h@xmath1 column density ( derived from the line intensity under the assumptions that the c@xmath4h@xmath1 rotational temperature is 4.2 k and the ortho - to - para ratio is 4.2 @xcite ) is equal to @xmath86 @xmath9 , which is comparable to , but somewhat smaller than , the value of @xmath87 @xmath9 derived by kawaguchi et al .
the @xmath88 h@xmath1co line , detected using arecibo ( fig . [ scan ] ; table [ gauss ] ) is rather similar but not identical to that observed previously by e.g. , @xcite with a @xmath89 beam .
the relative intensities of hyperfine components yielded the optical depth of the main component to be 1.6 and the line excitation temperature to be 1.2 k. no @xmath88 h@xmath90co and h@xmath1c@xmath91o lines were detected .
we estimate the lower limits to the ratios of the @xmath92 level populations for h@xmath1co / h@xmath90co and h@xmath1co / h@xmath1c@xmath91o assuming equal excitation temperatures for the @xmath88 transition of all isotopic species .
the limits are 24 and 76 , which are below the @xmath93c/@xmath5c and @xmath94o/@xmath91o abundance ratios by factors of approximately 3 and 6 , respectively .
thus , the non detection of the @xmath88 h@xmath90co and h@xmath1c@xmath91o lines does not contradict to the known values of the @xmath93c/@xmath5c and @xmath94o/@xmath91o abundance ratios .
table [ neg ] presents the upper limits to the antenna temperatures of observed but undetected lines together with the relevant column density limits .
the column density limits were calculated from eq.(1 ) under the assumption that the line populations are thermalized at 10 k. most of the molecular species from table [ neg ] have already been found in tmc-1 at higher frequencies .
comparison of the published column densities with the upper limits in table [ neg ] shows that an increase in sensitivity by a factor of 25 could lead to the detection of a much larger number of molecular lines .
thus , molecular observations at low frequencies ( 410 ghz ) may be useful for studying dark clouds if the sensitivity ( at @xmath95 level ) is about 510 mk or better .
we did not estimate h@xmath90co and h@xmath1c@xmath91o column densities because of well - known strong deviations from lte of the level populations of these species .
note that some other molecules may remain undetected in dark clouds due to excitation particularities rather than low abundances .
for instance , ch@xmath1nh was not detected in tmc-1 and l134 by @xcite ; they suggested that it may be an excitation effect , since the observed transitions have critical densities of about @xmath96 @xmath13 , which is well below the density of tmc-1 .
our non detection of the @xmath88 ch@xmath1nh emission line is probably a result of the @xmath97 transition at 167.4 ghz ( @xmath98 cm@xmath15 ) , which can radiatively depopulate the @xmath92 level at the moderate densities found in tmc-1 .
the results of the lowest frequency spectral survey carried out toward a dark cloud , studying tmc-1 in the 46 ghz range using the arecibo radio telescope , and those from sensitive observations at selected frequencies in c and x bands can be summarized as follows : * a number of molecular lines were detected .
the majority of the detected lines belong to the cyanopolyynes hc@xmath3n , hc@xmath2n , hc@xmath6n , and hc@xmath7n . *
the rotational temperatures of the detected molecules fall in the range 49 k. cyanopolyyne column densities vary from @xmath8 @xmath9 for hc@xmath2n to @xmath10 @xmath9 for hc@xmath7n .
* molecular observations at low frequencies ( 410 ghz ) can be useful for studying dark clouds , but to be really effective , the sensitivity must be increased by a factor of 2 to 3 . to date , observations at around 10 ghz have been more fruitful than those at lower frequencies .
the national astronomy and ionosphere center is operated by cornell university under a cooperative agreement with the national science foundation .
we are grateful to the staff of the arecibo observatory , and especially jeff hagen , mike nolan , phil perillat , chris salter , and arun venkataraman for help during the observations and useful discussions .
we thank the anonymous referee for important comments .
svk s visit to arecibo was supported by the naic director s office .
the work was partly supported by the russian foundation for basic research ( grant no .
01 - 02 - 16902 ) .
araya , e. , hofner , p. , goldsmith , p. f. , slysh , v. i. , & takano , s. 2003 , , 596 , 556 bell , m. b. , sears , t. j. , & matthews , h. e. , 1982 , , 255 , l75 bell , m. b. , matthews , h. e. , & sears , t. j. , 1983 , , 127 , 241 bell , m. b. , avery , l. w. , & watson , j. k. g. , 1993 , , 86 , 211 bell , m. b. , watson , j. k. g. , feldman , p. a. , & travers , m. j. , 1998 , , 508 , 286 comito , c. , schilke , p. , lis , d. c. , motte , f. , phillips , t. g. , & mehringer , d. m. , 2002 , sfchem 2002 : chemistry as a diagnostic of star formation , c. l. curry & m. fish , to be published by nrc press , ottava , canada , 18 cummins , s. e. , linke , r. a & thaddeus , p. , 1986 , , 60 , 819 dickens , j. e. , langer , w. d. , & velusamy , t. , 2001 , , 558 , 693 dickens , j. e. , irvine , w. m. , & devries , c. h. , 1997 , , 479 , 307 fuente , a. , cernicharo , j. , barcia , a. , & gomez - gonzales , j. , 1990 , , 231 , 151 goldsmith , p.f . & langer , w.d .
1999 , , 517 , 209 guelin , m. , friberg , p. , & mezaoui , a. , 1982 , , 109 , 23 henkel , c. , wilson , t. l. , & pankonin , v. , 1981 , , 99 , 270 hirahara , y. , suzuki , h. , yamamoto , s. , kawaguchi , k. , kaifu , n. , ohishi , m. , takano , s. , ishikawa , s .-
, & masuda , a. , 1992 , , 394 , 539 jewell , p. r. , hollis , j. m. , lovas , f. j. , & snyder , l. e. , 1989 , , 70 , 833 johansson , l. e. b. , andersson , c. , ellder , j. , friberg , p. , hjalmarson , . ,
hoglund , b. , irvine , w. m. , olofsson , h. , & rydbeck , g. , 1984 , , 130 , 227 kaifu , n. , ohishi , m. kawaguchi , k. , saito , s. , yamamoto , s. , miyaji , t. , miyazawa , k. , ishikawa , s. , noumaru , c. , harasawa , s. , and okuda , m. 2004 , , 56 , no.1 , 69 kawaguchi , k. , kaifu , n. , ohishi , m. , ishikawa , s .-
i . , hirahara , y. , yamamoto , s. , saito , s. , takano , s. , murakami , a. , vrtilek , j. m. , gottlieb , c. a. , & thaddeus , p. , 1991
, , 43 , 607 ohishi , m. , & kaifu , n. , 1998 , faraday discussion , 109 , 205 peng , r. , langer , w. d. , velusamy , t. , kuiper , t. b. h. , & levin , s. , 1998 , , 497 , 842 pratap , p. , dickens , j. e. , snell , r. l. , miralles , m. p. , bergin , e. a. , irvine , w. m. , & schloerb , f. p. , 1997
, , 486 , 862 schilke , p. , benford , d. j. , hunter , t. r. , lis , d. c. , & phillips , t. g. , 2001 , , 132 , 281 snell , r. l. , schloerb , f. p. , young , j. s. , hjalmarson , . , &
friberg , p. , 1981
, , 244 , 45 sorochenko , r. l. , tolmachev , a. m. , & winnewisser , g. , 1986 , , 155 , 237 turner , b.e .
1991 , , 76 , 617 turner , b. e. , herbst , e. , & terzieva , r. , 2000 , , 126 , 427 takano , s. , masuda , a. , hirahara , y. , suzuki , h. , ohishi , m. , ishikawa , s. , kaifu , n. , kasai , y. , kawaguchi , k. , & wilson , t. l. , 1998 , , 329 , 1156 white , g. j. , araki , m. , greaves , j. s. , ohishi , m. , & higginbottom , n. s. , 2003 , , 407 , 589 llllllr hc@xmath3n & 9097.0346 & @xmath100 & 0.138(0.003 ) & 5.689(0.002 ) & 0.167(0.001 ) & 0.684 + & & & 0.091(0.002 ) & 5.851(0.003 ) & 0.168(0.006 ) & 0.512 + & & & 0.071(0.001 ) & 6.061(0.003 ) & 0.167(0.007 ) & 0.400 + hc@xmath3n & 9089.3321 & @xmath101 & 0.180(0.012 ) & 5.689(0.002 ) & 0.167(0.001 ) & 1.085 + & & & 0.136(0.007 ) & 5.851(0.003 ) & 0.168(0.006 ) & 0.814 + & & & 0.115(0.003 ) & 6.061(0.003 ) & 0.167(0.007 ) & 0.693 + hc@xmath3n & 9100.2727 & @xmath102 & 0.042(0.004 ) & 5.689(0.002 ) & 0.167(0.001 ) & 0.240 + & & & 0.032(0.003 ) & 5.851(0.003 ) & 0.168(0.006 ) & 0.180 + & & & 0.024(0.001 ) & 6.061(0.003 ) & 0.167(0.007 ) & 0.137 + h@xmath5cccn & 8817.096 & @xmath51 @xmath80 & 0.006(0.002 ) & 5.66(0.14 ) & 0.40(0.00 ) & 0.014 + hc@xmath5ccn & 9059.736 & @xmath51 @xmath80 & 0.007(0.002 ) & 5.91(0.06 ) & 0.40(0.00 ) & 0.015 + hcc@xmath5cn & 9060.608 & @xmath51 @xmath80 & 0.009(0.002 ) & 5.72(0.04 ) & 0.40(0.00 ) & 0.021 + hcc@xmath5cn & 9059.318 & @xmath51 @xmath103 & 0.008(0.002 ) & 6.11(0.15 ) & 0.40(0.00 ) & 0.018 + hc@xmath2n & 5325.421 & @xmath52 @xmath104 & 0.12(0.003 ) & 5.91(0.01 ) & 0.41(0.01 ) & 0.28 + hc@xmath2n & 5325.330 & @xmath52 @xmath80 & 0.06(0.002 ) & 5.91(0.01 ) & 0.41(0.01 ) & 0.15 + hc@xmath2n & 5324.270 & @xmath52 @xmath105 & 0.03(0.001 ) & 5.91(0.01 ) & 0.41(0.01 ) & 0.07 + hc@xmath6n * & 5640.0316 & @xmath106 & 0.027(0.004 ) & 6.00(0.02 ) & 0.34(0.04 ) & 0.075 + hc@xmath6n * & 5640.0091 & @xmath107 & 0.021(0.004 ) & 6.00(0.02 ) & 0.34(0.04 ) & 0.061 + hc@xmath6n * & 5639.9580 & @xmath108 & 0.018(0.004 ) & 6.00(0.02 ) & 0.34(0.04 ) & 0.049 + hc@xmath6n & 7896.023 & @xmath109 & 0.028(0.012 ) & 5.92(0.04 ) & 0.33(0.04 ) & 0.080 + hc@xmath6n & 7896.010 & @xmath110 & 0.024(0.012 ) & 5.92(0.04 ) & 0.33(0.04 ) & 0.069 + hc@xmath6n & 7895.989 & @xmath111 & 0.021(0.012 ) & 5.92(0.04 ) & 0.33(0.04 ) & 0.060 + hc@xmath6n & 9024.020 & @xmath112 & 0.028(0.006 ) & 5.94(0.02 ) & 0.39(0.03 ) & 0.069 + hc@xmath6n & 9024.009 & @xmath113 & 0.025(0.006 ) & 5.94(0.02 ) & 0.39(0.03 ) & 0.061 + hc@xmath6n & 9023.994 & @xmath114 & 0.022(0.006 ) & 5.94(0.02 ) & 0.39(0.03 ) & 0.053 + hc@xmath7n * & 5229.3272 & @xmath115 & 0.016(0.004 ) & 5.87(0.14 ) & 1.11(0.23 ) & 0.014 + hc@xmath7n * & 5810.362 & @xmath116 & 0.017(0.005 ) & 6.30(0.23 ) & 1.51(0.39 ) & 0.012 + hc@xmath7n * & 8134.503 & @xmath117 & 0.016(0.005 ) & 5.90(0.11 ) & 0.57(0.21 ) & 0.027 + hc@xmath7n * & 8715.538 & @xmath62 & 0.022(0.003 ) & 5.96(0.02 ) & 0.39(0.05 ) & 0.053 + hc@xmath7n * & 9296.5721 & @xmath118 & 0.022(0.003 ) & 5.91(0.03 ) & 0.33(0.06 ) & 0.061 + ccs * & 5402.6175 & @xmath69 & 0.009(0.003 ) & 5.70(0.07 ) & 0.40(0.00 ) & 0.021 + c@xmath3s * & 5780.759 & @xmath51 & 0.041(0.005 ) & 5.91(0.01 ) & 0.49(0.03 ) & 0.077 + c@xmath4h & 9497.616 & @xmath119 & 0.098(0.01 ) & 5.77(0.01 ) & 0.21(0.03 ) & 0.447 + & & @xmath80 & & & & + c@xmath4h@xmath1 * & 8931.92 & @xmath85 & 0.009(0.003 ) & 5.72(0.01 ) & 0.11(0.30 ) & 0.079 + h@xmath1co & 4829.6594 & @xmath88 @xmath120 & @xmath121 & 5.84(0.01 ) & 0.71(0.02 ) & @xmath122 + lcc hc@xmath2n & @xmath123 & @xmath124 + hc@xmath6n & @xmath125 & @xmath126 + hc@xmath7n & @xmath127 & @xmath128 + ccs & @xmath129 & @xmath130 + c@xmath4h & 5.6 & @xmath131 + c@xmath4h@xmath1 & 4.2 & @xmath132 + lllccc hc@xmath7n & 4067.2695 & @xmath67 & 0.026 & & col + hc@xmath19n & 5071.8851 & @xmath62 & 0.027 & @xmath133 & jpl + hc@xmath19n & 5410.0103 & @xmath118 & 0.012 & @xmath134 & jpl + hc@xmath19n & 8115.007 & @xmath135 & 0.033 & @xmath136 & jpl + hc@xmath19n & 8791.257 & @xmath137 & 0.042 & @xmath138 & jpl + hc@xmath19n & 9129.376 & @xmath139 & 0.018 & @xmath140 & jpl + hc@xmath141n & 5348.622 & @xmath142 & 0.012 & @xmath143 & col + hc@xmath141n & 5562.566 & @xmath137 & 0.013 & @xmath144 & col + c@xmath4d & 8868.9005 & @xmath78 & 0.033 & @xmath145 & col + & & @xmath146 & & & + c@xmath2o & 5467.388 & @xmath52 & 0.030 & @xmath147 & jpl + c@xmath4s & 5912.1755 & @xmath148 & 0.027 & @xmath149 & col + c@xmath4s & 8868.7075 & @xmath150 & 0.033 & @xmath151 & col + c@xmath2s & 5536.222 & @xmath152 & 0.029 & @xmath153 & jpl + c@xmath2s & 9227.028 & @xmath24 & 0.036 & @xmath154 & jpl + c@xmath2n & 5607.1377 & @xmath155 & 0.030 & @xmath156 & col + & & @xmath157 & & & + c@xmath158h & 4201&@xmath159 @xmath79&0.026 & @xmath160 & jpl + ch@xmath3c@xmath4h & 4071.4935 & @xmath161 & 0.030 & @xmath162 & col + ch@xmath1nh & 5289.813&@xmath88 & 0.030 & @xmath163 & lov + hcooh & 4916.312 & @xmath164 & 0.039 & @xmath165 & lov + ch@xmath1co & 5660.9476 & @xmath166 & 0.024 & @xmath167 & jpl + h@xmath1c@xmath91o & 4388.8011 & @xmath88 & 0.021 & & lov + h@xmath90co & 4593.0865 & @xmath88 & 0.072 & & lov + | the results of the lowest frequency spectral survey carried out toward a molecular cloud and sensitive observations at selected frequencies are presented . the entire arecibo c - band ( 46 ghz )
was observed towards the cyanopolyyne peak of tmc-1 with an rms sensitivity of about 1718 mk ( about 22.5 mjy ) .
in addition , a number of selected frequency ranges within the c - band and x - band ( 810 ghz ) were observed with longer integration times and rms sensitivities 78 mk ( @xmath0 2 mjy ) or higher . in the spectral scan itself , already known h@xmath1co and hc@xmath2n lines were detected .
however , in more sensitive observations at selected frequencies , lines of c@xmath1s , c@xmath3s , c@xmath4h , c@xmath4h@xmath1 , hc@xmath3n and its @xmath5c substituted isotopic species , hc@xmath2n , hc@xmath6n , and hc@xmath7n were found , about half of them detected for the first time .
the rotational temperatures of the detected molecules fall in the range 49 k. cyanopolyyne column densities vary from @xmath8 @xmath9 for hc@xmath2n to @xmath10 @xmath9 for hc@xmath7n .
our results show that for molecular observations at low frequencies ( 410 ghz ) to be useful for studying dark clouds , the sensitivity must be of the order of 510 mk or better . to date , observations at around 10 ghz have been more productive than those at lower frequencies . |
air transportation systems have been traditionally described as graphs with vertices representing airports and edges direct flights during a fixed time period @xcite .
these graphs are called airport networks and have been studied at different geographical resolution scales , restricted , for instance , to a single country ( usually the u.s .
( usan ) @xcite but also china @xcite or europe @xcite ) , or for the whole world ( wan ) @xcite .
these networks show high heterogeneity in the distribution of connections per airport and in the traffic sustained by each connection .
a non linear relation between the number of connections of the airports ( topology ) and the number of passengers ( traffic ) has been observed in ref .
@xcite and used later for modelling @xcite .
furthermore , airport networks are structured in clusters of highly interconnected airports that reflect the geographical areas in which the traffic is naturally divided @xcite .
the dynamics of the connections and the traffic levels have been also analyzed for the usan @xcite .
all of these are aspects of the graphs that influence their capability to transport persons , goods and even other less desirable passengers .
for example the propagation of infectious diseases at a global scale that occurs when infected persons travel across the network @xcite .
the modelling and forecasting of disease spreading patterns using air traffic data is a story of a notable success @xcite .
one can , thus , wonder if this success can be extended to the propagation of other phenomena .
in particular , we are interested in considering here flight delays and the way in which congestion can become a systemic risk .
according to the @xmath0 report of the congress joint economic committee , flight delays have an economic impact in the u.s .
equivalent to @xmath1 billions of dollars per year @xcite , while a similar cost is expected in europe @xcite .
the situation can turn even grimmer in the next decade since the air traffic is envisaged to increase @xcite .
delays damage companies balances due to enhanced operation costs contributing to deteriorate their image with costumers @xcite .
passengers suffer a loss of time , even more acute in case of missing connections , that translates into decreased productivity , missed business opportunities or leisure activities .
additionally , attempts to recover delays lead to excess fuel consumption and larger co@xmath2 emissions . as a consequence of this challenging situation ,
a considerable effort has been invested in the area of air traffic management to characterize the sources of initial ( primary ) delays @xcite and the way in which they may be transferred and amplified by consequent operations , the so - called reactionary delays @xcite .
the concept of delay itself implies a time difference with respect to the baseline provided by a predefined schedule @xcite .
the propagation of delays thus corresponds to the spreading of a malfunction across the system .
the mechanisms responsible for it reflect the complexity of air traffic operations .
apart from the airport networks structure and dynamics , other factors contributing to the delay propagation are airport congestion @xcite , plane rotation or crew and passenger connection disruptions @xcite .
airline schedules typically include a buffer time to deal with all these issues .
however , when this time is not enough , the departure of the next flight gets delayed and can affect further operations in a cascade - like effect @xcite
. there has been several attempts to model delay spreading @xcite .
these studies differ in the level of detail included but in general they consider the effects of delays or disruptions in the operations of a few major airports ( hubs ) . in this work , we take instead a network - wide perspective to analyze the performance of a transportation system .
we define metrics able to quantify the level of spread of the delays in the network .
we then apply these metrics to a database with information on the operations in the u.s during @xmath3 , and introduce a model that reproduces the delay propagation patterns observed in the data .
the model shows also a notable capacity to evaluate the risk of development of system - wide congestion and to assess the resilience of daily schedules to service disruptions .
the data was downloaded from the web page of the bureau of transport statistics ( bts ) @xcite . in particular , we used the airline on - time performance data , which is built with flight statistics provided by air carriers that exceed one percent of the annual national revenue for domestic regular service .
the database comprehends @xmath4 scheduled flights operated by @xmath5 carriers connecting @xmath6 different commercial airports .
the total flights operated in the us in @xmath3 , not only those that report on - time performance data , sum up @xmath7 @xcite .
therefore , the database comprises information accounting for @xmath8 of them .
the information per flight includes real and scheduled departure ( arrival ) times , origin and destination airport , an identification code ( tail number ) for each aircraft , airline , etc .
this data enables us to represent the us airport network and furthermore replicate the scheduled flights for every day of @xmath3 .
a detailed description can be found in appendix a. it is important to note that this schedule is based on real events , which in some occasions may differ from the original planned schedule of the companies .
if a flight gets canceled , diverted or even rescheduled the airline may introduce changes in the original schedule that are not possible to trace back .
however , given that these flights represent , respectively , the @xmath9 and @xmath10 of all flights in the database , one can expect these changes not to be of large magnitude .
the database was obtained using the information available at the bureau of transport statistics @xcite . in particular , we used the airline on - time performance data , which is built with flight data provided by air carriers that exceed one percent of the annual national revenue for domestic scheduled service .
the database comprehends @xmath4 scheduled domestic flights operated by @xmath5 carriers connecting @xmath6 different commercial airports .
considering all flights in @xmath3 , not only those that report on - time performance data , the number of scheduled domestic flights totalizes @xmath7 @xcite and so our dataset includes information for @xmath8 of the total .
it is worth noting that this schedule is based on real events , which is not necessarily the original schedule in hands of the companies at the beginning of the day .
if a flight gets canceled , diverted or even in milder situations the managers of an airline can introduce changes in the original schedule that we can not trace back . however , given that these flights represent , respectively , the @xmath9 and @xmath10 of all flights in the database , one can expect these changes not to be of large magnitude . among the available data fields , we consider the following as the most relevant for our work : _ tail number , airline i d , airports of origin and destination , date of the flight , scheduled departure and arrival times , real departure and arrival times , and whether the flight was canceled or diverted_.
the tail number is a code that identifies the aircraft and that allows us to follow it along the daily plane rotation .
arrival and departure times ( real or scheduled ) considered refer to when the flight actually reaches or departs from the gate , we are not taking taxing or take - off and landing times as departure ( arrival ) times .
we will exclude of the coming analysis diverted and canceled flights since they difficult the characterization of the delay propagation and are a small fraction of the total operations .
a network between airports can be built based on the data : airports are the vertices and edges represent direct flights from one airport to another . note that this is a directed graph and that depends on the time - scale of data aggregation . on an annual basis ,
the resulting us air - transportation network comprises @xmath6 commercial airports and @xmath32 connections .
figure [ fig : dist ] depicts the complementary cumulative distribution of the number of flights and different connections for all the airports of the network in 2010 .
both distributions confirm the presence of high heterogeneities in the airport network .
the most active airports for @xmath3 are represented in table [ tab : airportdegree ] , showing that the maximal degree corresponds to atlanta international airport ( atl ) with @xmath33 different connections . for the analysis of the clusters of congested airports
, we will considered networks aggregated only during one day . besides the evolution of the size of the largest cluster per hour
, dynamics can be characterized by the evolution of the number of clusters during the day ( see figure [ fig : numberclusters ] ) . while in a satisfactory day ( figure [ fig : numberclusters ] c and d ) the number of clusters varies in each hour without a recognizable pattern , in an unsatisfactory day ( figure [ fig : numberclusters ] a and b ) the number of clusters increase in the first hours of the morning and from then on decay merging into fewer clusters , in most cases , in the afternoon ( eastern time ) .
this high - level interaction dynamics between clusters appears to be crucial in the evolution of an unsatisfactory day , where high - degree nodes play an important role to make this merging event come about .
however events involving individual nodes occur and varies dramatically from time to time .
figure [ fig : growthdynamics ] displays how nodes that belong to the largest cluster of the day vary their condition rapidly . one hour they are above the @xmath18 minutes threshold and next they recover and vice versa .
most nodes switch from one state to the other very quickly , although some few nodes repeat their condition at least two time steps ( red series ) . in order to study the temporal persistence of airports in the largest congested cluster across the whole database
, we display in table [ tab : topdelay ] the list of the top 10 airports in days in the largest congested cluster and in days with problems .
although some airports appear in both lists , the order changes and both sets are not exactly equal . in these lists
, there is a strong component of airports located in the west coast .
we think that this is due to the time difference between east and west coasts .
flight operations initiate before in the east coast and so the delays can propagate westwards toward the end of the day . in the results in the main paper , we show that the largest congested clusters are not persistent , at least not in more than @xmath34 of the airports between different days .
the airports in table [ tab : topdelay ] are those most persistent in the largest congested cluster .
it is interesting to notice that only two major hubs , newark ( ewr ) and san francisco ( sfo ) , are present in the top 10 list .
the modeling approach followed is agent - based at the level of aircrafts and is data - driven in the sense that the daily schedules and the primary delays are obtained directly from real records in the database .
this level of realism is necessary to confront the model predictions with the real unfolding of the delay events during each day .
concretely , the model dynamics simulates three main subprocesses : aircraft rotation , flight connectivity and airport congestion .
the latter two are independent from each other , and can be turned on / off to explore the relevance of each subprocess in leading to network - wide congestion .
aircraft rotation , on the other hand , is intrinsic to the schedule and can not be suppressed .
the basic time unit of the simulations is one minute , every aircraft state is tracked at this temporal resolution .
we assume that the flights are not able to recover delays on air , and so the departure delays are equal to those at arrival to destination . throughout a day
, each aircraft follows the connections given in the schedule , the so - called plane rotations .
the airports are supposed to have a capacity per hour proportional to the scheduled airport arrival rate with a proportionality factor @xmath12 .
further arrivals produce delays .
passengers ( crew ) of incoming flights have a certain probability of connecting with other flights within a time window of @xmath13 hours from the scheduled arrival .
the probability of connection is proportional , with a factor @xmath14 , to flight connectivity levels provided by the bts for each u.s .
a more precise description of the model is included in appendix b. this model has , thus , two free parameters : @xmath14 , controlling passenger connectivity , and @xmath12 , accounting for airport capacity . in the following section
, we will examine the effect of these parameters on the systemic spread of delays .
flight delays are defined as the difference between the scheduled and real departure ( arrival ) times @xcite .
actually most of the flights operated in @xmath3 were on time , even some before schedule , but @xmath15 of those reporting performance arrived or departed late .
their delays do not show a characteristic value : the delay distribution displays a broad tail as can be seen in figure 1a .
this implies that most flights arrived late by just a few minutes , while others were hours behind schedule .
the shape of the distributions is similar regardless of the arrival or departure nature of the operations .
the planned buffer time on ground for each aircraft should help absorb part of the delays , specially those mildest as will be discussed next , thus altering the shape of the departure delay distribution . however , this factor is not able to substantially modify the characteristics of the distributions .
interestingly , the shape of the delay distribution does not change either when the season of the year is considered .
summer concentrates the major part of the year traffic , so the total delay is higher but when the distribution of delay per flight is taken into account both summer and winter behave similarly ( see figure 1b ) .
the overall distributions of delays are thus quite robust .
some small differences can be only observed when focusing on particular airports . in figure 1c ,
the departure delay distribution is plotted for atlanta , jfk new york and honolulu airports .
while the distributions in atlanta and new york are similar , the honolulu airport shows a bias toward larger delays due to its isolation from the continent .
the effect of the buffer time in the airports for absorbing delays can be measured using the turn around time ( @xmath16 ) .
the @xmath16 stands for the time spent by an aircraft on ground from arrival to departure from the gate .
this measure is associated with airport operational efficiency and is used to improve the planning of flight connectivity and aircraft rotational sequence stability @xcite .
we refer as @xmath11 to the difference between scheduled and real times at the gate . on the one hand
, a negative value of @xmath11 means that an aircraft stayed at the gate longer than expected and so fresh delay was introduced . on the other hand
, a positive @xmath11 shows that the operation was quicker than scheduled and that part of the delay was recovered . in figure 2a
, we depict @xmath11 for each flight along a day in the most trafficked airport of the network : hartsfield - jackson in atlata ( atl ) .
that day , march @xmath17 , happened to be one of the worst in the database in terms of average flight delay .
the abundance of positive values of @xmath11 is a prove in favor of the capacity of the airport to recover delays .
the distributions of @xmath11 for all the operations in @xmath3 separated in positive an negative values are displayed in figure 2b .
these distributions , as those for the delays , show long tails , which is a marker of the complex nature of delay spreading mechanisms .
the focus so far has been on individual flight delays .
we define now a metric of congestion for the full network .
to do so , the average delay of all delayed flights during the year is taken as baseline and amounts to @xmath18 minutes .
an airport is considered as congested whenever the average delay of all its departing flights over a certain period of time exceeds @xmath18 minutes .
additionally , a daily airport network is built using the flights of the day to assess whether congested airports are organized in connected clusters or not .
note that being in the same cluster is a measure of spatio - temporal correlation of congestion but not necessarily a sign of a cause - effect relation .
we apply the same metric in the simulations in order to compare empirical and model results .
maps with the congested airports and the connections between them are shown for different days of the database in figures 3a-3c .
as can be seen , the scenario dramatically changes from day to day : in some days a large cluster surges covering @xmath19 of all airports , while in others only one or two airports cluster together . this
is confirmed when the size of the largest connected cluster is depicted as a function of the day in figure 3d .
a strong variability is thus the main characteristic of the dynamics of the size of the largest congested cluster .
the cumulative distribution of the cluster size is displayed in figure 3e and it seems compatible with an exponential decay .
even if the fluctuations are large , there exists a well defined characteristic cluster size .
given the cluster variability , an important question to answer is whether the congested airports are recurrent . in panel 3f , we calculate the jaccard index to compare the sets of airports in the largest cluster in consecutive days or for the top @xmath20 worst and best days .
this index is @xmath21 if the clusters are equal and @xmath22 if they are strictly different .
interestingly , the index is relatively low for days with large clusters , which implies that the same airports are not consistently part of the cluster . in order to compare empirical results and model predictions regarding the evolution of the cluster of congested airports , we run the model fixing the airport capacity parameter @xmath23 and fitting the flight connectivity factor @xmath14 to obtain a maximum cluster size similar to the one observed as the data . by fixing @xmath12 to @xmath21
, we are assuming the same airport capacity as originally scheduled .
the results for the temporal evolution of the congested cluster size hour by hour can be seen in figure 4 for march @xmath17 and april @xmath24 .
similar plots for other days of the year are included in appendix c. note that the fit of @xmath14 is essential to get the maximum of these curves , however all the cluster size evolution predicted by the model follows strikingly well that of the real data .
actually , almost @xmath25 of the airports in the real cluster are correctly identified by the model since they are top ranking when airports are ordered by probability of congestion .
furthermore , by fixing @xmath14 , without any fitting , the model can predict with @xmath26 accuracy if a day will develop or not a large congested cluster ( see appendix c for further details ) .
the model allows us also to explore which are the contributions of the main three ingredients ( plane rotation , flight connectivity or airport congestion ) to propagate delays . from figures 4b - c
, we can conclude that flight connectivity is the most important factor .
one may still wonder if the picture changes when the capacity of the airports is modified .
actually , the model exhibits weak sensitivity to variations on the @xmath12 coefficient as shown in figure 19 . slightly increasing
the airport capacity will not ease off the propagation of delays since the main cause of the spreading , flight connections , is independent of it .
conversely , a very strong decrease on the airports capacity , around @xmath27 , is needed to trigger new primary delays that later on will spread in a cascading effect .
this might be the case when generalized severe weather conditions or labor conflicts occur .
the initial delays affect the outcome of the model . in the results of figure 4
, we take the primary delays for each aircraft from the data as initial conditions for the model . introducing different initial conditions
, we can assess the resilience of a day schedule to an increase of unexpected incidences .
this question is explored in figure 5 where a fraction of randomly selected flights are delayed .
the size of the largest cluster is estimated as a function of the fraction of delayed flights and of the intensity of the initial delays .
for the sake of simplicity , we set all the initial delays in the simulation equal to a fixed value ( delay intensity in figure 5 ) .
the results are displayed for the schedules of two days : april @xmath24 and march @xmath17 , which respectively show a very small and very large cluster in the real data .
in particular , the average flight delay on march @xmath17 was the second largest in @xmath3 .
the congestion on the worst day of the year , october @xmath28 , can be explained due to extreme meteorological conditions @xcite , while on march @xmath17 no major external event was reported .
therefore , the network - wide propagation of delays in that day was likely caused and driven by internal mechanisms of the system .
comparing in figure 5 the curves for march @xmath17 and april @xmath29 , one notices that the surface representing the largest cluster size for march @xmath17 are displaced toward smaller values of the initial delay intensity or fraction of flights with primary delay .
this shows a higher susceptibility of the schedule of this day to disruptive perturbations .
another interesting feature of the curves of figure 5 is that , given enough primary delays , they show a non - negligible risk of systemic failure regardless of the schedule .
the curves in figure 5 for different values of @xmath14 also confirm the relevance of connections and crew rotations for the spreading of delays .
the primary flight delays in a day of real operations do not necessarily localize randomly in the network .
if the causes are bad weather , technical or labor issues are more prone to concentrate in a few airports . in figure 6 , this issue is explored by comparing the intra - day evolution of the cumulative size of the largest congested cluster when the initial delays are introduced in the model in two different ways . the first one is by using the primary delays given in the database .
the second procedure is by randomly shuffling the flights affected by the primary delays .
the values of the real delays in the database are maintained but they are assigned to flights selected at random .
the comparison of the curves for the two cases with the real data shows that random perturbations are way more efficient to collapse the system .
while airports in general have some capacity to recover delays , the random selection of delayed flights affect a larger number of them and besides concentrate a heavier burden on smaller airports which have less capacity to react .
this result evinces that the method followed for schedule evaluation in figure 5 is conservative in the sense that it considers the schedule under a non favorable scenario for the distribution of primary delays .
in summary , we analyze the spreading of delays in an air traffic network . in particular , our results focus on the us airport network in @xmath3 but the concepts and techniques employed can be easily extrapolated to the analysis of the performance of a generic transport system .
we introduce a measure for the level of network - wide extension of the delays by defining when an airport is considered as congested and studying how congested airports form connected clusters in the network .
the size of the largest congested cluster displays in the data a high variability from one day to the next .
this feature is due to the re - start that the system suffers at the end of each day and points toward the relevance of the daily schedule to define the delay propagation patterns .
in addition we introduce a data - driven model able to reproduce the delay evolution observed in the data .
the model includes three main mechanisms to spread delays : plane rotation , flight connections of either passenger or crews and airport congestion .
the last two processes can be modulated at will to understand the role that each one of them plays in delay propagation .
our simulations evidence that passenger and crew connections is the most effective single mechanism to induce network congestion .
we show how the model can be used to assess the daily schedule ability to deal with an increase in the number of disruptive events and also study the relevance of primary delay localization for the evolution of congestion in the network .
furthermore the model offers the possibility of evaluating the effects of interventions in the system before their real implementation .
flight delays represent failures to meet constraints imposed by a daily schedule .
its propagation in the network is a paradigmatic example of the way in which a distributed transport system moves toward collapse .
the framework develop in this work is thus of easy extension to system with dynamics regulated by predefined schedules .
its translation to other airport networks is , of course , straightforward , and even though the modeling of other transportation systems may require some particular details , the applicability of the metrics defined to measure network - wide congestion based on clustering is universal .
pf receives support from the network complex world within the wpe of sesar ( eurocontrol and eu commission ) .
jjr acknowledges funding from the ramn y cajal program of the spanish ministry of economy ( mineco )
. partial support from mineco was also received through project modass ( fis2011 - 24785 ) and from the eu commission through projects eunoia ( fp7-dg.connect-318367 ) and lasagne ( fp7-ict-318132 ) .
guimer , r. , mossa , s. , turtschi , a. & amaral , lan .
the worldwide air transportation network : anomalous centrality , community structure , and cities global roles .
_ proc natl acad sci usa _ * 102 * , 7794 - 7799 ( 2005 ) .
colizza , v. , barrat , a. , barthlemy , m. & vespignani , a. the role of the airline transportation network in the prediction and predictability of global epidemics .
_ proc natl acad sci usa _ * 103 * , 2015 - 2020 ( 2006 ) .
balcan , d. , colizza , v. , gonalves , b. , hu , h. , ramasco , j.j . & vespignani , a. multiscale mobility networks and the large scale spreading of infectious diseases .
_ proc natl acad sci usa _ * 106 * , 21484 - 21489 ( 2009 ) .
joint economic committee of us congress , _ your flight has been delayed again : flight delays cost passengers , airlines and the u.s .
economy billions_. available online at http://www.jec.senate.gov ( may 22 , 2008 ) .
rupp , n.g .
_ further investigations into the causes of flight delays _
, working paper , department of economy , east carolina university .
available online at http://www.ecu.edu/cs-educ/econ/upload/ecu0707.pdf ( 2007 ) .
schaefer , l. & millner , d. flights delay propagation analysis with the detailed policy assessment tool .
_ procs . of the 2001 ieee international conference on systems , man and cybernetics _
* 2 * , 1299 - 1303 ( 2001 ) .
severe weather occurred on october 27 , see noaa http://www.spc.noaa.gov/climo/reports/101027_rpts.html and cnn http://articles.cnn.com/2010-10-27/us/us.weather_1_tornado-damage-tornado-sightings-airport-delays?_s=pm%3aus | technologically driven transport systems are characterized by a networked structure connecting operation centers and by a dynamics ruled by pre - established schedules .
schedules impose serious constraints on the timing of the operations , condition the allocation of resources and define a baseline to assess system performance . here
we study the performance of an air transportation system in terms of delays .
technical , operational or meteorological issues affecting some flights give rise to primary delays .
when operations continue , such delays can propagate , magnify and eventually involve a significant part of the network .
we define metrics able to quantify the level of network congestion and introduce a model that reproduces the delay propagation patterns observed in the u.s . performance data .
our results indicate that there is a non - negligible risk of systemic instability even under normal operating conditions .
we also identify passenger and crew connectivity as the most relevant internal factor contributing to delay spreading . |
understanding anomalous transport in magnetically confined plasmas is an outstanding issue in controlled fusion research .
a satisfactorily understanding of the non - local features as well as the non - gaussian probability distribution functions ( pdfs ) found in experimental measurements of particle and heat fluxes is still lacking . in particular , experimental observations of the edge turbulence in the fusion devices @xcite show that in the scrape of layer ( sol ) the plasma fluctuations are characterized by non - gaussian pdfs .
it has been recognized that the nature of the cross - field transport through the sol is dominated by turbulence with a significant ballistic or non - local component where a diffusive description is improper @xcite .
moreover , the scaling of the confinement time @xmath0 with @xmath1 @xcite is typical in low - confinement mode discharges , instead of the diffusion induced result @xmath2 , where @xmath3 is the system size .
there is a considerable amount of experimental evidence @xcite and recent numerical gyrokinetic @xcite and fluid @xcite simulations that plasma turbulence in tokamaks is highly non - local .
in addition , intermittent turbulence is characterized by patchy spatial structure that is bursty in time .
the pdfs of these intermittent events shows unimodal structure with `` elevated '' tails that deviates from a gaussian prediction .
the understanding of these events are at best limited @xcite .
moreover , the high possibility of confinement degradation by intermittency strongly calls for a predictive theory . a prominent candidate for explaining the suggestive non - local features of plasma turbulence is the inclusion of a fractional velocity derivative in the fokker - planck ( fp ) equation leading to an inherently non - local description as well as giving rise to non - gaussian pdfs of e.g. densities and heat flux .
the non - locality is introduced through the integral description of the fractional derivative @xcite and the non - maxwellian distribution function drives the observed pdfs of densities and heat flux far from gaussian .
the aim of this study is to elucidate the effects of a non - maxwellian distribution function induced by the fractional velocity derivative in the fokker - planck equation .
some previous papers on plasma transport have used models including a fractional derivative where the fractional derivative is introduced on phenomenological premises @xcite . in the present work
we introduce the levy statistics into the langevin equation thus yielding a fractional fp description .
this approach is similar to that of ref .
@xcite resulting in a phenomenological description of the non - local effects in plasma turbulence . using fractional generalizations of the liouville equation ,
kinetic descriptions have been developed previously @xcite .
it has been shown that the chaotic dynamics can be described by using the fp equation with coordinate fractional derivatives as a possible tool for the description of anomalous diffusion @xcite .
much work has been devoted on investigation of the langevin equation with levy white noise , see references @xcite , or related fractional fp equation @xcite .
furthermore , fractional derivatives have been introduced into the fp framework in a similar manner as the present work @xcite but a study including drift waves is still called for . to this end
we quantify the effects of the fractional derivative in the fp equation in terms of a modified dispersion relation for density gradient driven linear plasma drift waves where we have considered a case with constant external magnetic field and a shear - less slab geometry . in order to calculate an equilibrium pdf we use a model based on the motion of a charged levy particle in a constant external magnetic field obeying non - gaussian , levy statistics .
this assumption is the natural generalization of the classical example of the motion of a charged brownian particle with the usual gaussian statistics @xcite .
the fractional derivative is represented with the fourier transform containing a fractional exponent .
we find a relation for the deviation from maxwellian distribution described by @xmath4 through the quasi - neutrality condition and the characteristics of the plasma drift wave are fundamentally changed , i.e. the values of the growth - rate @xmath5 and real frequency @xmath6 are significantly altered .
a deviation from the maxwellian distribution function alters the dispersion relation for the density gradient drift waves such that the growth rates are substantially increased and thereby may cause enhanced levels of transport . the paper is organized as follows . in sec .
2 the mathematical framework of the fractional fp equation ( ffpe ) is introduced . in sec .
3 a dispersion relation for the density gradient driven drift modes using the ffpe are derived . in sec 4
the deviations from a maxwellian distribution function are investigated and the dispersion relation is solved in sec .
we conclude the paper with a results and discussion in sec .
following the theory of brownian motion we write an equation of motion for a colloidal particle in a background medium as a langevin equation of the following form @xcite @xmath7 here , we assumed that the influence of the background medium can be split into a dynamical friction , @xmath8 , and a fluctuating part , @xmath9 which is a gaussian white noise . the gaussian white noise assumption is usually imposed in order to obtain a maxwellian velocity distribution describing the equilibrium of the brownian particle .
this connection is due to the relation between the gaussian central limit theorem and classical boltzmann - gibbs statistics @xcite .
however , the gaussian central limit theorem is not unique and a generalization of the gaussian central limit theorem to the case of summation of independent identically distributed random variables described by long tailed distributions is performed by lvy @xcite , and khintchine @xcite . in this case
the lvy distributions replace the gaussian in a generalized central limit theorem . the simplest case of generalized brownian motion considered by west and seshadri @xcite is to assume for fluctuation part , @xmath9 , in equation ( [ eq:1.1 ] ) to be a white lvy noise . following the approach used by barkai @xcite
we find the fractional fokker - planck equation ( ffpe ) with fractional velocity derivatives for shear - less slab geometry in the presence of a constant external force as @xmath10 where @xmath11 represents the particle species and @xmath12 . here , the term @xmath13 is the fractional riesz derivative .
the fractional differentiation may be represented through singular integrals or by its fourier transform as we will see later in equation ( [ eq:1.4 ] ) .
note that the connection to the integral representation indicates that the model is inherently non - local in velocity space .
the diffusion coefficient , @xmath14 , is related to the damping term @xmath15 , according to a generalized einstein relation @xcite @xmath16 here , @xmath17 is a generalized temperature , and taking force @xmath18 to represent the lorentz force ( due to a constant magnetic field and a zero - averaged electric field ) acting on the particles of species @xmath19 with mass @xmath20 and @xmath21 is the euler gamma function .
we find the solution by using the fourier representation of equation ( [ eq:1.2 ] ) above as @xmath22 where @xmath23 is the larmor frequency of species @xmath19 , @xmath24 is the unit vector in the direction of magnetic field and @xmath25 is the characteristic function @xmath26 where we have denoted the wave - vector by @xmath27 and the corresponding wave vector for the velocity as @xmath28 .
we can rewrite the kinetic equation by identification of time derivatives of the wave vectors as @xmath29 we use the method of characteristics on the equation ( [ eq:1.4 ] ) and ( [ eq:1.6 ] ) whereby we find that the characteristics are @xmath30 following the method used in ref .
@xcite the solution corresponding to the homogenous and steady state system in fourier space is @xmath31 in order to find the solution in real space we compute the inverse fourier transform of equation ( [ eq:2.14 ] ) @xmath32 we define a new variable @xmath33 where coefficient @xmath14 is given by the expression in equation ( [ eq:1.3 ] ) .
@xmath34 is a normalization factor which remains to be defined . taking the inverse fourier transform of the equation ( [ eq:2.15 ] ) for @xmath35
we get @xmath36 the unknown normalization factor @xmath37 can be determined by comparing the integrals of the maxwellian distribution and our distribution . in comparison
the maxwellian distribution is defined as @xmath38 where @xmath39 is the thermal velocity of species @xmath19 . by integrating the maxwellian distribution over the velocity space
we find the density as @xmath40 whereas performing the same integration of the expression in equation([eq:2.16 ] ) we obtain @xmath41 we can now compare the two results obtained in equations ( [ eq:2.18 ] ) - ( [ eq:2.19 ] ) and we find the following relation @xmath42 the distribution function can now be determined by replacing this expression into equation ( [ eq:2.16 ] ) for @xmath34 yielding @xmath43 we can easily recover the maxwellian distribution in equation ( [ eq:2.17 ] ) by setting @xmath35 in the definition for @xmath14 in equation ( [ eq:1.3 ] ) and using that @xmath44 . note that for a general @xmath45 , the equilibrium distribution is as follows @xmath46 where @xmath47 and we have introduced a generalized thermal velocity as @xmath48 the generalized equilibrium distribution including the effects of the fractional velocity derivative in equation ( [ eq:2.24.1 ] ) becomes @xmath49 we will now determine the dispersion relation for density gradient driven drift waves including the effects of the fractional velocity differential operator .
in order to quantify the non - local effects on drift waves induced by the fractional differential operator we will determine the dispersion relation for density gradient driven drift modes .
we start by formulating the linearized gyro - kinetic theory where the particle distribution function , averaged over gyro - phase is of the form ( see ref .
@xcite ) @xmath50 we assume that the turbulence is purely electrostatic and neglect magnetic field fluctuations @xmath51 . for small deviations from the local equilibrium we find the linearized gyro - kinetic equation of the form @xmath52 where @xmath53 is the drift wave frequency of species @xmath19 , and we assumed that the space dependence of @xmath54 is only in the @xmath55 direction perpendicular to the magnetic field as well as for the density gradient . in the equation above
, @xmath56 is the bessel function of order zero , @xmath57 is the parallel velocity , @xmath58 is the perpendicular velocity and hence we write the total speed as @xmath59 . inserting the expression for @xmath54 from the equation ( [ eq:2.24 ] ) in equation ( [ eq:2.27 ] ) and rearranging the terms we find the perturbed distribution @xmath60 as @xmath61j_{0}(|\omega_{s}|^{-1}k_{\bot}v_{\bot})\delta \phi_{\mathbf{k},\omega } \frac{n_{s}(\mathbf{r})}{2\pi^{3/2}(\gamma(1+\alpha))^{-1/2}\sqrt{2v_{t , s}^{\alpha } } } \times\nonumber\\ \int \frac{d\mathbf{k}_{\bot}^{v}d\mathbf{k}_{\parallel}^{v}}{(2\pi)^{3/2}}e^{-i(\mathbf{k}_{\bot}^{v}\mathbf{v}_{\bot}+\mathbf{k}_{\parallel}^{v}\mathbf{v}_{\parallel})}e^{-\frac{v_{t , s}^{\alpha}}{\gamma(1+\alpha)\alpha}(|\mathbf{k}^{v}_{\bot}|^{\alpha}+|\mathbf{k}^{v}_{\parallel}|^{\alpha})}.\end{aligned}\ ] ] here
, the wave vector perpendicular to magnetic field is @xmath62 .
the gyro - kinetic equation ( [ eq:2.28 ] ) is complemented with poisson equation for the electric potential . for fluctuations with wave vectors
much smaller than the debye wave vector , the poisson equation becomes the quasi - neutrality condition @xmath63 where the density fluctuation is related to the distribution function through @xmath64 in the above equation we have separated the adiabatic response ( first term on the right hand side ) from the non - adiabatic response ( second term on the right hand side ) .
we have to keep in mind that the density @xmath65 coming from the @xmath66 in the adiabatic response is also given by equation ( [ eq:2.24 ] ) and for a general @xmath67 the adiabatic response can be different than that calculated by maxwellian distribution of equation ( [ eq:2.17 ] ) . using the quasi - neutrality condition ( [ eq:2.29 ] )
we find the dispersion equation which determines the eigenfrequencies as a function of the wave vector , @xmath68 . in the simplest case we consider a plasma consisting of electrons and a single species of singly charged ions with the equal temperatures .
for the density fluctuation therefore we have @xmath69.\end{aligned}\ ] ] therefore , the dispersion equation as in the ref .
@xcite is @xmath70 where @xmath71 gives the adiabatic contribution , and @xmath72j_{0}(b_{s}v_{\bot}/v_{ts})\times\nonumber\\ \frac{1}{2\pi^{3/2}(\gamma(1+\alpha))^{-1/2}\sqrt{2v_{t , s}^{\alpha } } } \int \frac{d\mathbf{k}_{\bot}^{v}d\mathbf{k}_{\parallel}^{v}}{(2\pi)^{3/2}}e^{-i(\mathbf{k}_{\bot}^{v}\mathbf{v}_{\bot}+\mathbf{k}_{\parallel}^{v}\mathbf{v}_{\parallel})}e^{-\frac{v_{t , s}^{\alpha}}{\gamma(1+\alpha)\alpha}(|\mathbf{k}^{v}_{\bot}|^{\alpha}+|\mathbf{k}^{v}_{\parallel}|^{\alpha})},\end{aligned}\ ] ] gives the non - adiabatic contribution . here , @xmath73 .
if we take @xmath35 in the equation([eq:2.33 ] ) we recover the dispersion equation for a maxwellian distribution as in the ref .
@xcite .
first , we may analyze the contribution from the adiabatic parts of the dispersion relation only by ignoring all fluctuations , yielding @xmath74 , therefore @xmath75 and @xmath76 becomes connected through equation ( [ eq:2.37 ] ) .
this indicates that the deviation from a maxwellian distribution described by @xmath45 for electrons and ions becomes dependent on each other .
we will get back to this relation in later sections .
we will now turn our attention to the problem of solving the dispersion relation described by equation ( [ eq:2.33 ] ) . in order to solve this dispersion equation
we use the method proposed in ref .
@xcite with the difference that here we have to perform additional integrations over @xmath77 .
we have @xmath78 where the plasma dispersion function is @xmath79,\end{aligned}\ ] ] with @xmath80 , @xmath81 and the function @xmath82 is @xmath83 the integral over @xmath84 can be written in a general way as @xmath85 where @xmath86 , @xmath87 and , @xmath88 the analytical solutions for integrals over @xmath77 with an arbitrary @xmath45 in the equations ( [ eq:3.3 ] ) and ( [ eq:3.5 ] ) requires rather tedious calculations .
instead we consider an infinitesimal deviation of the form @xmath89 , where @xmath90 and expand the terms depending on @xmath45 in the equations ( [ eq:3.3 ] ) and ( [ eq:3.5 ] ) around @xmath91 as follows @xmath92,\end{aligned}\ ] ] where @xmath93 + 2v_{t , s}^2 { \log}[v_{t , s}]|k^{v}|^2 + 2v_{t , s}^2|k^{v}|^2 { \log}[|k^{v}|]\}.\end{aligned}\ ] ] here , we have used the euler - mascheroni constant @xmath94 . the first term in equation ( [ eq:3.6 ] ) will produce @xmath95 which give the maxwellian adiabatic response @xmath96 by using the expansion defined by the expression ( [ eq:3.6 ] ) in equations ( [ eq:2.34 ] ) and ( [ eq:2.34 ] ) , the adiabatic and non - adiabatic part of the dispersion relation @xmath97 and @xmath98 are
as follows @xmath99 ^ 2=1+\epsilon w^{ad , s}.\nonumber\\\end{aligned}\ ] ] and @xmath100\psi_{s}(b_{s}v_{\bot}/v_{ts})\times\nonumber\\ \frac{1}{(\sqrt{\pi}v_{t , s}(\mathbf{r}))^3}e^{-(v_{\bot}^2+v_{\parallel}^2)/v_{t , s}^2(\mathbf{r})}+\nonumber\\ ( 2\pi\int_{-\infty}^{\infty}dv_{\parallel}\int_{0}^{\infty}dv_{\bot}v_{\bot}[\frac{\omega-\omega_{*s}}{k_{\parallel}v_{\parallel}-\omega}]\psi_{s}(b_{s}v_{\bot}/v_{ts})\times\nonumber\\ \frac{1}{2\sqrt{2}\pi^{3/2 } } \int \frac{d\mathbf{k}_{\bot}^{v}d\mathbf{k}_{\parallel}^{v}}{(2\pi)^{3/2}}e^{-i(\mathbf{k}_{\bot}^{v}\mathbf{v}_{\bot}+\mathbf{k}_{\parallel}^{v}\mathbf{v}_{\parallel})}\lambda(k_{\bot}^{v})\lambda(k_{\parallel}^{v}))\epsilon+\mathcal{o}[\epsilon]^2=n^{s}_{\mathbf{k},\omega}+\epsilon w^{s}_{\mathbf{k},\omega}.\nonumber\\\end{aligned}\ ] ] inserting these relations we may rewrite the dispersion relation ( [ eq:2.33 ] ) in the form @xmath101 the first terms on the right and left hand sides generate the usual contributions to the dispersion equation as in ref . @xcite and
the terms proportional to @xmath4 generate the non - maxwellian contributions where we have @xmath102 with the usual plasma dispersion function @xmath103 written as @xmath104,\end{aligned}\ ] ] and @xmath105 the effects of the fractional velocity derivative can be boiled down to a non - maxwellian contribution of the form @xmath106 where the non - maxwellian plasma dispersion function is given by @xmath107,\end{aligned}\ ] ] with the function @xmath82 being @xmath108 it is important to note that the deviation from maxwellian is different for the different species ( electrons and ions ) . in the rest of sec .
4 , we will quantify the deviations .
the non - maxwellian contribution to equation ( [ eq:3.4 ] ) is @xmath109 where @xmath110 to extimate the non - maxwellian contribution we need to determine the inverse fourier transforms of the equations ( [ eq:3.16 ] ) and ( [ eq:3.18 ] ) resulting in @xmath111 + 2e^{z^2}{_1 f_1}[\frac{3}{2},\frac{1}{2},-z^2]\right\}\end{aligned}\ ] ] with @xmath112 and @xmath113 $ ] denoting kummer s confluent hypergeometric function .
therefore we can write @xmath114 by inserting typical values for the plasma parameters from ref .
@xcite we find the velocities as @xmath115 $ ] and @xmath116 $ ] and we obtain @xmath117 following the adiabatic condition in equation ( [ eq:2.37 ] ) and the expanded dispersion relation in equation ( [ eq:3.10 ] ) we obtain the following ratio between the non - maxwellian contributions @xmath118 this relation means that if there is a deviation of the distribution function from the maxwellian for plasma electrons , the deviation from the maxwellian for ions will be @xmath119 larger .
we will solve the dispersion relation in terms of expansions of the plasma dispersion function by noting that the drift waves are defined in the frequency range @xmath120 in evaluating equations ( [ eq:3.12 ] ) and ( [ eq:3.15 ] ) .
we define the expansion parameter for electrons in powers of @xmath121 and for ions we expand it in powers of @xmath122 , respectively .
the maxwellian dispersion function @xmath103 has the same definition as in ref .
@xcite @xmath123 = -2\xi_{e}+\frac{4\xi_{e}^3}{3}+i\sqrt{\pi}(1-\xi_{e}^2)+\mathcal{o}[\xi_{e}^4],\end{aligned}\ ] ] whereas the non - maxwellian plasma dispersion function @xmath124 becomes @xmath125 = \frac{v_{t , e}}{\sqrt{\pi}}lim_{\sigma\rightarrow 0}\nonumber\\ \int_{-\infty}^{\infty}du\phi(u)[\frac{1}{u - i\sigma}+\frac{\xi_{e}}{(u - i\sigma)^2}+\frac{\xi_{e}^2}{(u - i\sigma)^3}+\frac{\xi_{e}^3}{(u - i\sigma)^4}+\mathcal{o}[\xi_{e}^4]].\end{aligned}\ ] ] for ions , using the expansion in powers of @xmath126 we can rewrite the above integrals as a function of the expansion parameter as @xmath127 = -\xi_{i}^{-1}-\frac{1}{2}\xi_{i}^{-3}+\mathcal{o}[\xi_{i}^{-5}],\end{aligned}\ ] ] and the non - maxwellian @xmath128 becomes @xmath129 = \frac{v_{t , i}}{\sqrt{\pi}}lim_{\sigma\rightarrow 0}\nonumber\\ \int_{-\infty}^{\infty}du\phi(u)[\frac{1}{(-\xi_{i}-i\sigma)}-\frac{u}{(\xi_{i}+i\sigma)^2}+\frac{u^2}{(-\xi_{i}-i\sigma)^3}-\frac{u^3}{(\xi_{i}+i\sigma)^4}+\mathcal{o}[\xi_{i}^{-5}]].\end{aligned}\ ] ] we can now evaluate he maxwellian integrals of the forms @xmath130 and @xmath131 assuming @xmath132 , @xmath133 we get @xmath134 and @xmath135 where @xmath136 denotes modified bessel function of the zeroth order .
the final result will be found after evaluating the non - maxwellian @xmath137 and @xmath138 are given as @xmath139 and @xmath140 where we have used @xmath141 $ ] , @xmath142 $ ] and @xmath143 .
finally we can summarize different terms in the dispersion relation ( [ eq:3.10 ] ) as @xmath144 where @xmath145 . note that the non - maxwellian contributions in equations ( [ eq:3.20 ] ) , ( [ eq:3.22 ] ) , ( [ eq:3.25 ] ) and ( [ eq:3.26 ] ) have been calculated numerically . by utilizing the found values of the integrals above we rewrite the dispersion relation ( [ eq:3.10 ] ) as follows @xmath146 where @xmath147 are given in equation ( [ eq:3.18.3 ] ) and we will use the ratio between @xmath148 and @xmath149 from equation ( [ eq:3.18.4 ] ) .
we have derived a dispersion relation for drift waves driven by a density gradient in a shear - less slab geometry with constant magnetic field where the small deviation from a maxwellian distribution is described by @xmath4 . here
we will determine the quantitative effects on the real frequency and growth rate as a function of this deviation .
we start by assuming that we have adiabatic electrons for which the dispersion equation ( [ eq:3.28 ] ) is , @xmath150 after rearranging the terms in the above equation we finally get the following relation for @xmath149 : @xmath151e^{-b_{i}/2}\mathcal{i}_{0}(b_{i})}{w^{ad , tot}\xi_{i}^3+(\bar{\omega}_{*,i}\xi_{i}^3-\xi_{i}^4)z_{\epsilon}(\xi_{i})\gamma_{\epsilon}(b_{i})}\end{aligned}\ ] ] where @xmath152 .
this relation gives the possible deviation of the equilibrium pdf from the maxwellian pdf for a given plasma turbulence , i.e @xmath153 .
one has to remember that only positive values of @xmath154 $ ] are physically meaningful . using the same plasma parameters
as was used in equations ( [ eq:3.18.3 ] ) and ( [ eq:3.24],[eq:3.26 ] ) we compute the term @xmath128 , and from equation ( [ eq:3.22 ] ) we get @xmath155 here , those integrations omitted resulted in zero contributions and rewriting equation ( [ eq:4.1 ] ) by using these explicit values results in the expression for the deviation in equation ( [ eq:4.1.1 ] ) we obtain @xmath156e^{-b_{i}/2}\mathcal{i}_{0}(b_{i})}{66.3\xi_{i}^3 - 39.2\xi_{i}+39.2\bar{\omega}_{*,i}+23.0 \bar{\omega}_{*,i}\xi_{i}^2}\end{aligned}\ ] ] figure [ fig1 ] shows @xmath149 from equation ( [ eq:4.4 ] ) where @xmath157 . here , the values of @xmath158 are normalized to @xmath159 . we have assumed parameter values @xmath143 , @xmath160 and @xmath161 with @xmath162 .
it is found that there is a threshold in the growth rate @xmath5 close to @xmath163 and that increasing to @xmath164 only increases the deviation from a maxwellian from 0 to 0.03 .
it should be noted that @xmath4 increases the excess kurtosis of the distribution function by a similar amount thus a quite small deviation from a maxwellian can have a rather significant impact . in figure [ fig2 ] ,
the mode growth rate as a function of @xmath149 is shown .
note that in this figure the values of growth rate are the solutions of the equation ( [ eq:4.4 ] ) for a given @xmath149 while in the figure [ fig1 ] we solve equation ( [ eq:4.4 ] ) for @xmath4 at a given @xmath153 . as the dispersion equation is of 3rd order in @xmath165 three possible solutions exist , however we are only interested in the solutions with non - zero imaginary value , @xmath166 corresponding to unstable situations .
it is shown in figure [ fig2 ] that a deviation of @xmath167 yield an increase of about @xmath168 in the growth rate .
furthermore , the growth rate increases almost linearly with increasing @xmath149 and such an increase in the growth rate will lead to a significant increase in the level of anomalous flux . in summary
, we have derived a dispersion relation for density gradient driven linear drift waves including the effects coming from the inclusion of a fractional velocity derivative in the fokker - planck equation in the case of constant magnetic field and a shear - less slab geometry .
the solutions of this fokker - planck equation are the alpha - stable distributions .
it has not yet been shown that in a direct way one can derive the alpha - stable distribution function @xcite from the classical form of collision operator @xcite .
one way may be to construct a new type of collisional operator by considering a fractal phase space and reformulate the collision operator on this new space . however , such a discussion is outside the scope of the present paper . interestingly enough
, we note that non - local effects are observed in non - linear collisionless fluid simulations of plasma turbulence where the non - local transport showing levy features are induced by the interaction of the non - linear terms in the dynamical equations @xcite .
the non - local features of non - linear fluid models are indicated by recent analytical theories using path - integral methods to derive probability density fucntions of fluxes @xcite .
the fractional derivative is represented with the fourier transform containing a fractional exponent that we are able to connect to the deviation from a maxwellian distribution described by @xmath4 .
the characteristics of the plasma drift wave are fundamentally changed , i.e. the values of the growth - rate @xmath5 and real frequency @xmath6 are significantly altered .
a deviation from the maxwellian distribution function alters the dispersion relation for the density gradient drift waves such that the growth rates are substantially increased and thereby may cause enhanced levels of transport .
_ acknowledgements _ the authors would like to thank professor t. flp for her helpful comments .
this work was funded by the european communities under association contract between euratom and _
11 s. j. zweben , j. a. boedo , o. grulke , c. hidalgo , b. labombard , r. j. maqueda , p. scarin and j. l. terry _ plasma phys .
control . fusion _ * 49 * s1-s23 ( 2007 ) v. naulin _ journal of nucl .
mater . _ * 363 - 365 * 24 - 31 ( 2007 ) s. m. kaye , cris w. barnes , m. g. bell , j. c. deboo , m. greenwald , k. riedel , d. sigmar , n. uckan , and r. waltz _ phys .
fluids b _ * 2 * 2926 ( 1990 ) n. lopez cardozo _ plasma phys .
contr . fusion _ * 37 * 799 ( 1995 ) k. w. gentle , r. v. bravenec , g. cima , h. gasquet , g. a. hallock , p. e. phillips , d. w. ross , w. l. rowan , and a. j. wootton _ phys .
plasmas _ * 2 * 2292 ( 1995 ) j. d. callen and m. w. kissick _ plasma phys .
control . fusion _ * 39 * b173 ( 1997 ) p. mantica , p. galli , g. gorini , gmd hogeweij , j. de kloe , and nj lopes cardozo _ phys .
lett . _ * 82 * 5048 ( 1999 ) b. ph .
van - milligen , e. de la luna , f.l .
tabars , e. ascasbar , t. estrada , f. castejn , j. castellano , i. garcia - cortes , j. herranz , c. hidalgo , j.a .
jimenez , f. medina , m. ochando , i. pastor , m.a .
pedrosa , d. tafalla , l. garca , r. sanchez , a. petrov , k. sarksian and n. skvortsova _ nucl .
fusion _ * 42 * 787 ( 2002 ) r. balescu _ aspects of anomalous transport in plasmas _ ( iop publishing ) ( 2005 ) g. dif - pradalier , p. h. diamond , v. grandgirard , y. sarazin , j. abiteboul , x. garbet , ph .
ghendrih , a. strugarek , s. ku and c. s. chang _ phys .
e _ * 82 * 025401 ( 2010 ) r. sanchez , d. e. newman , j .-
n leboeuf , v. k. decyk and b. a. carreras _ phys .
* 101 * , 205002 ( 2008 ) d. del - castillo - negrete , b. a. carreras and v. e. lynch _ phys .
* 94 * , 065003 ( 2005 ) s. t. bramwell _ nat .
_ 5 443 ( 2009 ) b. a. carreras , c. hidalgo , e. sanchez , m. a. pedrosa , r. balbin , i. garcia - corts , b. van milligen , d. e. newman , and v. e. lynch _ phys .
plasmas _ * 3 * 2664 ( 1996 ) b. a. carreras , b. ph .
van milligen , c. hidalgo , r. balbin , e. sanchez , i. garcia - cortes , m. a. pedrosa , j. bleuel and m. endler _ phys .
lett . _ * 83 * 3653 ( 1999 ) j. anderson and e. kim _ plasma phys .
contr . fusion _ * 52 * 012001 ( 2010 ) j. anderson and p. xanthopoulos _ phys .
plasmas _ * 17 * 110702 ( 2010 ) g. m. zaslavsky , _ hamiltonian chaos and fractional dynamics _
( oxford university press , oxford ) ( 2005 ) r. sanchez , b. ph .
van - milligen and b. a. carreras _ phys .
plasmas _ * 12 * 056105 ( 2005 ) d. del - castillo - negrete _ phys . plasmas _ * 13 * 082308 ( 2006 ) r. sanchez , b. a. carreras , d. e. newman , v. e. lynch and b. ph .
van milligen _ phys .
e _ * 74 * , 016305 ( 2006 ) g. m. zaslavsky _ phys .
_ * 371 * 461 ( 2002 ) v. e. tarasov _
conference series _ * 7 * 17 ( 2005 ) g. m. zaslavsky _ physica d _ * 76 * 110 ( 1994 ) b. j. west and v. seshadri _ physica a _ * 113 * 203 ( 1982 ) r. metzler , e. barkai and j. klafter _ phys
lett . _ * 82 * 3563 ( 1999 ) h. c. fogedby _ phys .
lett . _ * 73 * 2517 ( 1994 ) m. o. vlad , j. ross and f. w. schneider _ phys .
e _ * 62 * 1743 ( 2000 ) a. v. chechkin and v. yu gonchar _ open sys . & information dyn . _
* 7 * 375 - 390 ( 2000 ) a. v. chechkin and v. yu gonchar _ phys . of plasmas _
v9 , * 1 * 78 ( 2002 ) s. chandrasekhar _ rev .
modern phys . _
* 21 * 383 ( 1949 ) a. y. khintchine _ the mathematical foundation of statistical mechanics _
dover ( new york ) ( 1948 ) p. lvy _ theorie del addition des variables _ ( gauthier - villiers , paris ) ( 1937 ) b. j. west and v. seshadri _ linear systems with levy fluctuations _
, physica a , * 113 * 203 ( 1982 ) e. barkai _ stable equilibrium based on lvy statistics : stochastic collision models approach _ , phys .
e. rapid communication , * 68 * 055104(r ) ( 2003 ) r. balescu _ phys .
fluids _ * b 4 * 91 ( 1992 ) e. w. montroll and j. t. bendler _ j. stat . phys .
_ * * 34**129 ( 1984 ) r. gatto and h. e. mynick _ theory and application of the generalized balescu - lenard transport formalism , nova publishers _ ( 2008 ) | transport events in turbulent tokamak plasmas often exhibit non - local or non - diffusive action at a distance features that so far have eluded a conclusive theoretical description . in this paper a theory of non - local transport is investigated through a fokker - planck equation with fractional velocity derivatives . a dispersion relation for density gradient driven linear drift modes
is derived including the effects of the fractional velocity derivative in the fokker - planck equation .
it is found that a small deviation ( a few percent ) from the maxwellian distribution function alters the dispersion relation such that the growth rates are substantially increased and thereby may cause enhanced levels of transport . |
one of the constraints that lead to the nielsen - ninomiya @xcite no - go theorem , is the chiral invariance of the dirac action @xmath2 for massless fermions on the lattice . under certain conditions @xcite , the spectrum of fermions @xmath3 would suffer from doubling , and
the axial anomaly would cancel .
this has been , from the onset , a recurrent problem of lattice qcd @xcite . in order to recover the axial anomaly , at least at the pertubative level , ginsparg and wilson @xcite derived a relation , the ginsparg - wilson relation ( gwr ) which explicitly breaks the standard chiral symmetry , @xmath4 where @xmath5 is a matrix proportional to the lattice spacing @xmath6 ( in most of this
talk we will consider the case where @xmath5 is a simple number ) .
recently lsher @xcite proved that chiral invariance of an action @xmath2 which complies with the gwr can be recovered in an extended form , @xmath7 thus a conjecture appeared in the literature suggesting that it may be possible to overcome the nielsen - ninomiya no - go theorem on the lattice and to fully simulate , without doubling , the chiral symmetry of qcd ( see @xcite for a recent review ) .
this includes pcac @xcite and the axial anomaly @xcite .
in particular the atiyah - singer @xcite index theorem has a hasenfratz - laliena - niedermayer @xcite version on the lattice , @xmath8 where @xmath9 is the number of zero modes of the dirac action @xmath2 with negative ( positive ) chirality .
@xmath10 is an _ integer _
topological charge of the gauge field configuration , defined with the charge density , a function of the lattice site @xmath11 , @xmath12{l } \ \ \ { \gamma_5 \ , r \ , d(\vec n , \vec n)\ } \\ _ { colour ,
\ , dirac } \end{array } \ .
\label{gwr density}\ ] ] therefore each gwr dirac action _ defines _ the density ( [ gwr density ] ) and the index ( [ index ] ) .
moreover examples of lattice actions have been found where this index is nontrivial and coincides with a charge independently defined from a topological gauge density .
for instance the fixed point topological density @xcite coincides , on all gauge configurations , with the density ( [ gwr density ] ) defined with the fixed point dirac operator .
the recent overlap gwr solution of neuberger @xcite also solves the gwr and complies with the index theorem .
the neuberger solution was checked in dimension @xmath0 for smooth gauge configurations @xcite , and in dimension @xmath13 in the continuum limit @xcite .
thus the gwr , together with these recent results , essentially solve the problem of the topological index theorem on the lattice .
nevertheless we remark that the definition of the dirac operator and the definition of the topological charge on the lattice are not unique .
moreover the topological charge on the lattice is not conserved @xcite .
this suggests that different gwr dirac operators and topological densities produce different indexes when we depart from the continuum limit , to a finite lattice with finite charge density . in this
talk we address the _ extremum _ problem of finding the lattice dirac operator which provides the best index ( [ index ] ) . because this is a difficult problem , we specialize in lattice actions which are constructed with the wilson action @xcite . in order to compare the index of different dirac operators , we first choose a single topological charge density .
although this is an arbitrary definition , let us choose the simplest near neighbor lattice topological charge density which discretizes the continuum density @xmath14 and which sums to an integer total topological charge .
then we search for the gwr dirac operator which index ( [ index ] ) is closer to this simplest topological charge .
the arbitrariness in the choice of our lattice operators is partly constrained by locality .
locality is a crucial issue to connect the lattice to the quantum field theories in the continuum .
only local actions are universal .
moreover it has been conjectured @xcite that local gwr solutions have better topological properties than non - local ones . only topological configurations which are larger than the range of the couplings in @xmath2 would produce a number of zero modes @xmath15 different from @xmath16 in agreement with eq .
( [ index ] ) .
this conjecture is certainly correct in the case of the zenkin @xcite action .
the zenkin action is non - local @xcite , and although it complies with eq .
( [ gwr 5 ] ) , it does not comply with eq .
( [ index ] ) .
thus we will restrain from using non - local definitions for the lattice action and for the lattice charge density . in this talk
we investigate in detail the eigenvalues of some lattice actions , with the aim to clarify how the index theorem is extrapolated from the continuum to finite lattices with finite topological charge densities .
we also propose a new criterion for gwr solutions constructed with the wilson action . in section [ sec - eigen wilson ]
we study in detail the effects of topological gauge configurations on the eigenvalues of the wilson action for the discrete version of the schwinger model . in section [ sec - properties ]
we review properties of lattice actions . in section [ sec - topology ]
we address the topological abelian charge in @xmath0 . in section [ sec - wilson ]
we review the wilson action . in section [ sec - criterion ]
we present our criterion . finally in section [ sec - conclusion ] we conclude .
a plausible property of lattice actions is @xmath17 hermiticity . for instance
the wilson action is @xmath17 hermitean .
when an action verifies @xmath18 , we can show that @xmath2 and @xmath19 have the same eigenvalues @xmath20 , @xmath21 and that the eigenvalues of @xmath2 and @xmath19 are mapped by @xmath17 .
when we conjugate eq ( [ sameig ] ) , we find that for every right eigenvalue there is a left eigenvalue @xmath22 with eigenvector @xmath23 , and so the conjugate transformation @xmath24 leaves the spectrum invariant .
the spectrum of eigenvalues @xmath20 is symmetric with respect to the real axis of the argand plot .
we proceed to define chirality in the lattice .
( [ sameig ] ) can be used to show that the @xmath17 is involved in a extended orthogonality condition , @xmath25 a possible definition for the chirality of an eigenvector @xmath26 is , @xmath27 and eq .
( [ orthogonal ] ) shows that the complex eigenvalues have a vanishing chirality , @xmath28 .
only the real eigenvalues may have a non vanishing chirality .
we now study properties of gwr solutions .
the gwr implies that if @xmath26 is a zero mode of the gwr action @xmath29 , then @xmath30 .
thus we can use @xmath31 and @xmath32 to decompose the kernel in a set of left vectors and a set of right vectors .
this shows that the zero modes have a chirality @xmath33 .
moreover we can verify that if @xmath2 is a gwr action , then @xmath34 also complies with the gwr .
thus the eigenvalues @xmath35 of @xmath2 also have a chirality @xmath33 .
it is also convenient to define an intermediate matrix , @xmath36 because the gwr is equivalent to @xmath37 . in the case
that @xmath38 is @xmath17 hermitean , we also obtain that @xmath39 , and this shows that the eigenvectors of @xmath38 belong to the unitary circle of the complex argand plane .
the eigenvalues of @xmath2 are in a circle with center @xmath40 and radius @xmath41
. we will be interested in the @xmath42 eigenvalues of @xmath38 which correspond to zero modes of @xmath2 .
moreover the crucial property for the index theorem has been show , @xmath43 both with analytical @xcite and algebraic @xcite methods .
the @xmath0 case is very convenient for the simple study of discrete topology .
the euclidean dirac matrices @xmath44 of @xmath13 are now replaced by the pauli matrices @xmath45 .
the @xmath0 charge density is similar to the rotational of the vector potential , @xmath46 it is similar to a magnetic field . on the lattice this magnetic field density
is extracted from the plaquette , @xmath47 and a possible definition of topological charge density is , @xmath48 where @xmath49 .
this definition of the topological density is not unique , the necessary condition is that it must reproduce the correct continuum limit .
the definition ( [ topo density ] ) of the topological charge is particularly interesting because it is quantized .
it is clear that the total topological charge of the lattice , which is defined with , @xmath50 is an integer .
the charge @xmath51 is the sum of the decimal part of a set of numbers , and these numbers have a vanishing sum , therefore @xmath51 is an integer .
this is similar to @xmath0 compact qed , where the magnetic monopoles arise naturally on the lattice . here
the topological charge is equivalent to a magnetic flux through the toroidal @xmath0 lattice .
we now construct non - trivial gauge configurations . in this work
we are particularly interested in extrapolating continuum properties to finite lattices .
we specialize to gauge configurations with uniform topological charge density .
these configurations are unique , except for gauge transformations . in the @xmath52 lattice
a simple configuration is built with the prescription @xcite of chiu , @xmath53 \
, \nonumber \\
\pi \ , q { n_2 -1 \over n^2 } \ , \nonumber
\\ a_2(n_1,n)&=&2 \pi \ , q { n_1 -1 \over n } \ , \label{gauge}\end{aligned}\ ] ] where the undefined @xmath54 are zero . with this definition ( [ gauge ] ) and in the particular case of integer @xmath55 ,
the topological density is constant . if @xmath56 we find that @xmath57 and that the topological charge is @xmath58 .
the case where q is an arbitrary continuous parameter can be used when we are interested in an interpolation between the cases of uniform density .
then the density is uniform , except at the single point ( n , n ) .
the gauge configuration ( [ gauge ] ) is periodic in @xmath55 , with a period of @xmath59 . in the relevant range @xmath56
we find that the topological charge @xmath51 of eq .
( [ topo density ] ) is a step - like function of @xmath55 , @xmath60
the wilson @xcite action is the simplest and most widely used action in lattice field theory
. it also provides a good example to observe roots and doubles , @xmath61 where @xmath62 is a parameter of the order of @xmath63 . the vector @xmath64 term of eq .
( [ wilsoneg ] ) is the naive massless dirac action on the lattice , where the derivative @xmath65 is computed by means of finite differences . in the limit of free fermions , which corresponds to @xmath66 , all these operators commute . in this case
the fourier transform of the above matrices depend on a single momentum , @xmath67 in the free case , and in @xmath68 dimensions , @xmath69 has @xmath70 roots at all possible combinations of @xmath71 and @xmath72 , see eq .
( [ trigonometric ] ) .
this is the source of the doubler problem because only the root at vanishing @xmath73 is physical .
wilson @xcite included the scalar @xmath74 term in eq .
( [ wilsoneg ] ) to remove the unwanted doubles .
however this scalar term @xmath74 is not chiral invariant .
chiral invariance can be implemented with the lscher transformation .
however horvath showed that the @xmath75 operator can at most @xcite be exponentially local , therefore the @xmath76 of the gwr is not a simple number in this case . from the gwr relation ,
( [ gwr 5 ] ) we get , @xmath77 where the wilson action @xmath78 is in general invertible , see section [ sec - eigen wilson ] .
we can remark that the operator in eq .
( [ simple again ] ) is a solution of the gwr , however it is not exponentially local . in this case
the nontrivial operator of the lscher transformation , @xmath79 is also non - local .
moreover @xmath80 is not @xmath17 hermitean .
this excludes both the wilson action and the @xmath80 as the best candidate to study exact chiral symmetry on the lattice .
here we depart from the free case and study the eigenvalues of the wilson action for topological gauge configurations .
this continues reference @xcite .
the wilson action @xmath81 is the simplest and most used lattice action , see section [ sec - wilson ] .
our framework is the massless wilson action with parameters @xmath82 , see eqs ( [ wilsoneg ] ) , in a simple 2-dimensional u(1 ) gauge theory which is a discrete version of the schwinger model , see section [ sec - topology ] . in the continuum
the wilson action @xcite exactly complies with the index theorem ( [ index ] ) , but for finite lattices it does not .
our aim in this section is to extrapolate the topological charge from the continuum to the lattices of different size @xmath52 , and to study how the wilson action deviates from the index theorem .
we encounter the problem that the topological charge and the topological density on the lattice are not uniquely defined .
this problem is not present in the continuum , where the topological density and its derivatives are well defined .
this implies that a lattice action close to the continuum limit is only allowed to have a topological charge density both small and nearly constant .
however we are interested , both physically and mathematically , in extrapolating the index theorem to the case of finite lattices , with finite topological charge .
and then we want to investigate how far the topological density can be increased without spoiling the index theorem .
to remain as close as possible to the continuum definitions , we momentarily specialize to very smooth gauge configurations .
in particular we study the wilson action in configurations with constant plaquette density , see section [ sec - topology ] , which is equivalent to a constant topological charge density .
first we study some general properties of the wilson action . because it is @xmath17 hermitean , the spectrum of its eigenvalues , here denoted @xmath20 ,
is symmetric with respect to the real axis of the argand plot , see section [ sec - properties ] .
moreover in @xmath0 it turns out that in the case of even lattices , the spectrum of @xmath83 in the argand plot is also symmetric with respect to the imaginary axis .
the transformation @xmath84 leaves the spectrum invariant ( see for instance figs . [ eigenwil1],[eigenwil2],[eigenwil3 ] ) .
so we will use even @xmath52 lattices with @xmath85 from now on , for simplicity .
we also remark that @xmath86 real@xmath87 for any non vanishing gauge configuration , thus @xmath81 is invertible ( except in the free quark case ) . in what concerns @xmath88 we observe that its eigenvalues are exactly symmetric with respect to the origin .
this implies for instance that the traces tr@xmath89 $ ] and tr@xmath90 $ ] both vanish for any finite @xmath91 .
this is not the case of the trace tr@xmath92/{\cal z}$ ] which is non vanishing and has been proposed as a topological index @xcite , where @xmath93 is a constant normalizing factor .
however we will not further study this proposed topological charge because it clearly is neither local nor integer .
we now study quantitatively the real eigenvalues and the topological properties of the wilson action @xmath81 .
we are interested in eigenvalues @xmath94 with chirality @xmath95 , see section [ sec - properties ] .
the non vanishing chirality implies that the relevant eigenvalues are the real eigenvalues .
the @xmath74 term was introduced in the free action by wilson to separate the four roots of the @xmath65 , see section [ sec - wilson ] . in the free limit
this results in four real eigenvalues : one root at @xmath96 , two degenerate eigenvalues equal to @xmath97 and one eigenvalue equal to @xmath98 . here
we find that quite a similar separation exists in the quadruplets of real eigenvalues that occur in gauge configurations with a finite topological charge ( see figs .
[ eigenwil1],[eigenwil2],[eigenwil3 ] ) .
moreover the chirality of these real eigenvalues approximately agree with the index theorem .
in particular we observe that , + - the number of real eigenvalues is always a multiple of four , @xmath99 .
the integer @xmath100 sign@xmath101 is a possible definition of the topological charge of the gauge configuration , where @xmath102 is the chirality of the smallest real eigenvalue .
+ - in the present case of a constant topological density , @xmath103 if @xmath104 is smaller than half of the maximum possible charge density . the integer topological charge @xmath51 and its topological density @xmath105 are defined in section [ sec - topology ] .
+ - the real eigenvalues and the corresponding chiralities are close to integers .
the difference is proportional to a small number , @xmath106 .
the @xmath99 eigenvalues can be divided in the following 3 sets of eigenvectors , with eigenvalues , @xmath107 sign(q_2 ) , \nonumber \\
\lambda= 2 + o(\epsilon^2 ) \ & , & \ \chi = [ 1 -2.0 \epsilon + o(\epsilon^2 ) ] sign(q_2 ) , \nonumber \\
\lambda= 4 - 3.0 \epsilon + o(\epsilon^2 ) \ & , & \
\chi = [ -1 + o(\epsilon^2 ) ] sign(q_2 ) , \end{aligned}\ ] ] which contain respectively @xmath108 , @xmath109 and @xmath108 eigenvalues .
+ - when @xmath108 approaches the maximum value of the order of @xmath110 , these 3 sets of eigenvalues spread out to 3 finite intervals , and eventually they overlap .
it turns out that right before these intervals overlap , they just leave a gap at @xmath111 $ ] .
+ - in the continuum limit of @xmath112 , and for very smooth topological objects , all the different @xmath113 .
so we find that in this limit of large n , and from the view point of zero modes , the index theorem is verified , @xmath114 this result agrees with the proofs of fujikawa @xcite and adams @xcite .
if we now leave the case of constant topological density , it occurs that @xmath115 ( we mean that the absolute value of the topological density @xmath105 is smaller than 0.22 in any plaquette of the lattice ) is a sufficient condition to enforce that @xmath103 .
we verified this crucial result with a large number of randomly generated gauge configurations , for all possible @xmath51 .
all the remaining properties , that we just detailed , of the eigenvalues of the wilson action are also maintained when @xmath116 .
this suggests that it is possible to extrapolate the continuum limit to a significant and finite subinterval of the range @xmath117 . to illustrate how the complex eigenvalues transform in real ones ,
we arbitrarily interpolate between different constant charge densities with a continuous variation of the topological parameter @xmath55 ( see section [ sec - topology ] ) .
it turns out that for some particular values of @xmath55 , which produce a large @xmath104 at the point @xmath118 , two opposite pairs of complex eigenvalues ( with @xmath96 chirality ) quite suddenly transform into two opposite pairs of real eigenvalues , and then these real eigenvalues continuously increase their chirality @xmath119 .
this transition point is continuous , but at this precise value of @xmath55 the velocity of the eigenvalues in the argand plot is infinite . at this point
the topological number @xmath120 steps up ( down ) .
the opposite also happens , at some other particular values of @xmath55 , where @xmath98 real eigenvalues suddenly transform into complex eigenvalues .
the different topological charges can also be computed ( see section [ sec - topology ] ) . and the observed pattern is quite general .
the charge @xmath120 , defined with the real eigenvalues of the wilson action , and the charge @xmath51 , defined with the plaquette density @xmath121 in section [ sec - topology ] , are quite close only up to half of the maximum possible topological @xmath51 .
this verifies our empirical rule of @xmath122 . in figs .
[ eigenwil1],[eigenwil2],[eigenwil3 ] we show a sequence of the eigenvalues of the wilson action for consecutive values of the interpolating topological parameter @xmath55 .
a simple @xmath123 lattice is used and the eigenvalues are displayed in the argand plan .
we aim to understand how the index theorem extrapolates from the continuum limit to finite lattices with finite topological charge densities . in d=2 , and providing that the topological density @xmath122 , we show that the wilson action @xmath124 has the correct number of small real eigenvalues to comply with the index theorem .
the problem is that these eigenvalues are not exactly vanishing and that their chirality is not exactly @xmath125 .
the deviations to the index theorem are of order @xmath126 , where @xmath51 is the integer topological charge .
we also prove that the gwr neuberger solutions either project these approximate zero modes of @xmath124 on the origin , with correct chirality , or send them close to the remote other end @xmath35 of the unitary circle .
this motivates a new criterion for gwr solutions @xmath127 with index identical to @xmath51 ( and @xmath17 hermitean , constructed with the wilson action ) , which relies in the function @xmath128 .
this function @xmath128 is straightforwardly defined by replacing @xmath124 and @xmath78 by the same real number @xmath20 in the expression for @xmath127 .
we map @xmath129 .
the @xmath128 is a trivial case of a gwr solution .
moreover the @xmath17 hermiticity implies that @xmath128 is real .
we thus conclude that @xmath130 or @xmath131 . in the free case @xmath124 and @xmath78 commute , and have the correct real eigenvalues at @xmath96 , while the doubles correspond to the eigenvalues @xmath97 or @xmath98 .
we remark that the doubles correspond to pathological eigenvectors , with alternating signs in neighboring lattice sites .
we conclude that , unless the fermion fields are redefined , a correct gwr solution must produce @xmath132 , @xmath133 and @xmath134 . studying small topological densities
@xmath105 with finite @xmath51 , we can further show that @xmath135)=0 $ ] , @xmath133 , @xmath136)=2/r$ ] . in the case of small topological densities ,
@xmath124 and @xmath78 have the same eigenvalues , but they do not exactly have the same eigenvectors .
nevertheless the relevant small real eigenvalues @xmath20 of the wilson action , have a chirality @xmath137$]sign@xmath138 and this implies that the difference between the @xmath124 eigenvector @xmath26 and the @xmath78 eigenvector @xmath139 is very small , of order @xmath140 .
this shows that @xmath141 , the correct eigenvector @xmath142 of @xmath127 is quite close to @xmath26 ( and to @xmath139 ) . _
if @xmath127 complies with the index theorem , _ we can assume that this eigenvector @xmath142 is a zero mode , because @xmath26 is already quite close to the correct fermionic zero mode .
therefore @xmath143 .
moreover we can only have @xmath144 or @xmath35 .
this implies that @xmath130 . because @xmath145 we find that _ @xmath146 projects the interval @xmath147 $ ] on the origin @xmath96 .
inversely , if @xmath130 , _ @xmath127 has @xmath148 eigenvectors close to @xmath130 .
moreover they all have similar chirality @xmath137$]sign@xmath138 .
even if these eigenvectors mix , which is natural because they are very close , their chirality will remain close to sign@xmath149 , and finite .
_ therefore these @xmath148 eigenvalues are real , vanishing , with the correct chirality , and in the right number to comply with the index theorem _ .
similarly we can show that @xmath146 projects on @xmath35 both a small neighborhood of the point @xmath97 and the interval @xmath150 $ ] .
if we assume that the complex eigenvalues of the wilson action are irrelevant to the index because they have vanishing chirality , we conclude that the condition @xmath135)=0 $ ] , @xmath133 , @xmath136)=2/r$ ] is both a necessary and a sufficient condition for any @xmath17 hermitean gwr solution , constructed with the wilson action , and free from doubling , to comply with the index theorem in the case of small topological density .
in the case of a finite topological charge density the three sectors of real eigenvectors of the wilson that include @xmath96 , @xmath97 and @xmath98 spread out .
analytically it is hard to find how far the exact criterion that we just derived for small @xmath151 can be extended to finite @xmath151 .
nevertheless our numerical study of the wilson action shows that up to moderate densities @xmath122 , the sectors including @xmath96 , @xmath97 and @xmath98 remain separated .
for instance the sector of @xmath96 , with chiralities @xmath152 sign@xmath153 , and the sector of @xmath97 , with chiralities @xmath154 sign@xmath153 , are separated by a gap located at @xmath155 .
it is thus quite plausible that a correct gwr solution should project the smaller @xmath108 real eigenvalues @xmath20 of @xmath124 on @xmath96 , and that the remaining @xmath156 real eigenvalues should be projected onto the other end of the circle , close to @xmath35 .
we may now determine which is the function @xmath128 that maximally complies with the index theorem , up to the highest possible topological density @xmath157 .
this @xmath128 is a step function that jumps precisely at the gap @xmath155 .
finally when the density is large , @xmath158 or more at one or many plaquettes of the lattice , the index theorem is spoiled .
this is natural on the lattice , because for large densities the different definitions of topological charge density diverge and the topological charge is not conserved . in what concerns the three sectors of real eigenvectors of the wilson action
, that include respectively @xmath96 , @xmath97 and @xmath98 , they swell so much that they start to overlap .
the real eigenvalues either mix and transform into complex eigenvalues ( with vanishing chirality ) , or they carry their chirality to the wrong sector .
we remark that this is not a physical problem , providing the gauge configurations with large plaquette densities are suppressed by the gauge part of the action .
we now present our criterion , which is very simple to use . *
`` a @xmath17 hermitean gwr solution @xmath127 with constant @xmath5 and constructed with the wilson @xmath124 action complies maximally with the index theorem if and only if @xmath159 when we replace @xmath160 , where @xmath20 is a real number , @xmath161 and @xmath162 belongs to a well determined and narrow subinterval of @xmath1630,2[$ ] ' ' * .
our criterion also applies to gwr solutions constructed with any other action @xmath164 ( other than the wilson action ) , that approximately complies with the index theorem and that is @xmath17 hermitian , @xmath165 .
this talk is devoted to clarify how far the index theorem can be extrapolated from the continuum to finite lattices with finite topological charge densities .
the @xmath0 wilson action @xmath124 is examined , and we find that it approximately complies with the index theorem for finite topological charge densities lower @xmath166 .
we also study the gwr solution of neuberger which exactly complies with the index theorem .
finally we produce a criterion for gwr solutions , constructed with the wilson action , that maximally comply with the index theorem . with our criterion , it is simple to show that locality constitutes a sufficient condition for a dirac action to comply with the index theorem together with the simplest topological charge @xmath51 .
it is clear that any local action is differentiable in the free limit , and therefore the condition @xmath167 can be extended to a neighborhood of these points . using our criterion
, this proves that any local action ( gwr solution and with the correct free limit ) complies with the index theorem at least in the case of small topological charge densities .
the locality conjecture is correct in the case of actions built from the wilson action , and it is probably correct in the general case .
moreover the locality condition can be improved .
we aim at finding the action with the highest convergence radius in the neighborhood of the points @xmath168 , and in a sense this is comparable with finding the action with the highest locality .
our criterion complies with the chiu and zenkin criterion @xcite which states that the correct gwr solutions must have at least one eigenvalue at @xmath35 .
however we observe that it is possible to find a non - local action with eigenvalues at @xmath35 , which does not have zero modes outside the free limit .
therefore , unlike the locality condition , the chiu - zenkin criterion is not a sufficient condition to force an action to comply with the index theorem , although it certainly constitutes a necessary condition .
our criterion can be applied to the neuberger gwr solution defined with @xmath169 . with the @xmath20 substitution
we find that @xmath170 .
this complies with our criterion when @xmath171 , and it is quite evident that this is the simplest action which complies with it .
the criterion shows that the neuberger overlap action is the simplest gwr solution constructed with the wilson action that maximally complies with the index theorem . moreover , in the schwinger model and using the dimensionless units of @xmath82 , we find that @xmath172 .
this determines @xmath171 with a value which is not very far from the @xmath173 that provides the highest locality in the free limit .
this choice of @xmath174 increases the precision of the observation of chandrasekharan @xcite , stating that the neuberger action only complies with the index theorem if @xmath175 , ( @xmath175 can also be derived from the root and pole structure of the free neuberger action ) .
the choice of @xmath174 completely fixes the gwr parameter @xmath5 because the free continuum limit of the lattice implies that @xmath176 .
so the best choice seems to be @xmath177 .
finally we outline possible continuations of this study .
we are researching the analytical extension of the proof of our criterion to finite topological densities @xmath178 .
the repetition of the present study in four dimensions would also be physically relevant .
i acknowledge misha polikarpov for explaining computing techniques and topology on the lattice .
i am also grateful to herbert neuberger for explaining the locality conjecture and for pointing to a numerical error .
i thank dimitri diakonov , isabel salavessa and emilio ribeiro for discussions on the topological index .
00 nielsen , h. , and ninomiya , m. , _ nuc .
_ * b185 * , 20 ( 1981 ) ; + _ nuc .
* b193 * , 173 ( 1981 ) .
wilson , k. g. , _ phys .
_ * d10 * 2445 ( 1974 ) ; + quarks and strings on a lattice , in _ new phenomena in subnuclear physics , erice lectures-1975 _ , edited by zichichi , a. , plenum , new york , 1977 .
ginsparg , p.h . , and wilson , k.g .
rev . _ * d25 * , 2649 ( 1982 ) .
lscher , m. , _ nucl .
phys . _ * b538 * , 515 ( 1999 ) [ arxiv : hep - lat/9808021 ] ; + _ phys
. lett . _ * b428 * , 342 ( 1998 ) [ arxiv : hep - lat/9802011 ] .
niedermayer , f. , _ nucl .
* 73 * , 105 ( 1999 ) [ arxiv : hep - lat/9810026 ] .
chandrasekharan , s. , _ phys .
_ * d60 * , 074503 ( 1999 ) [ arxiv : hep - lat/9805015 ] .
zinn - justin , j. the regularization problem and anomalies in quantum field theory , in _ topology of strongly correlated systems , xviii lisbon autumn school- 2000 _ , edited by bicudo , p. , ribeiro , j. , sacramento , p. , seixas , j. , vieira , v. , world scientific , singapore , 2001 ; + h. neuberger , regulated chiral gauge theory , ibid .
atiyah , m. f. , singer , i. m. , _ annals math . _
* 87 * , 485 ( 1968 ) ; + _ annals math . _
* 87 * , 546 ( 1968 ) ; + atiyah , m.f . , and segal , g. b. , _ annals math . _
* 87 * , 531 ( 1968 ) ; + nakahara , m. , _ geometry , topology and physics _ , graduate student series in physics , institute of physics publishing , bristol and philadelphia , 1990 .
hasenfratz , p. , laliena , v. , and niedermayer , f. , _ phys .
_ * b427 * , 125 ( 1998 ) [ arxiv : hep - lat/9801021 ] .
hasenfratz , p. , and niedermayer , f. , _ nucl .
phys . _ * b414
* , 785 ( 1994 ) [ arxiv : hep - lat/9308004 ] ; + bietenholz , w. , and wiese , u. j. , _ nucl . phys . _ * b464 * , 319 ( 1996 ) [ arxiv : hep - lat/9510026 ] .
neuberger , h. , _ phys .
_ * b417 * , 141 ( 1998 ) , [ arxiv : hep - lat/9707022 ] ; + _ phys .
lett . _ * b427 * , 353 ( 1998 ) , [ arxiv : hep - lat/9801031 ] .
chiu , t. , _ phys .
rev . _ * d58 * , 074511 ( 1998 ) [ arxiv : hep - lat/9804016 ] .
chandrasekharan , s. , _ phys .
_ * d59 * , 094502 ( 1999 ) [ arxiv : hep - lat/9810007 ] .
fujikawa , k. , _ phys .
rev . _ * d60 * , 074505 ( 1999 ) [ arxiv : hep - lat/9904007 ] ; + _ nucl .
* b546 * , 480 ( 1999 ) hep - th/9811235 .
adams , d. h. , [ arxiv : hep - lat/9812003 ] .
t hooft , g. , _ phys .
_ * b349 * , 491 ( 1995 ) [ arxiv : hep - th/9411228 ] .
for a recent study of other classes of gwr solutions see , + gattringer , g. , and hip , i. , _ phys .
_ * b480 * , 112 ( 2000 ) [ arxiv : hep - lat/0002002 ] .
hernandez , p. , jansen , k. , and luscher , m. , _ nucl .
* b552 * , 363 ( 1999 ) [ arxiv : hep - lat/9808010 ] .
zenkin , bull .
lebedev phys .
inst . * * ( 1988 ) no.910 ; + zenkin , s. v. , _ mod .
_ * a6 * , 151 ( 1991 ) ; chiu , t. , wang , c. , and zenkin , s. v. , _ phys . lett . _ * b438 * , 321 ( 1998 ) [ arxiv : hep - lat/9806031 ] ; bicudo , p. , _ phys
. lett . _ * b478 * , 379 ( 2000 ) [ arxiv : hep - lat/9909157 ] .
horvath , i. , _ phys .
rev . _ * d60 * , 034510 ( 1999 ) [ arxiv : hep - lat/9901014 ] .
farchioni , f. , hip , i. , and lang , c. b. , _ phys .
lett . _ * b443 * , 214 ( 1998 ) [ arxiv : hep - lat/9809016 ] .
hernndez , p. , _ nucl .
_ * b536 * , 345 ( 1998 ) [ arxiv : hep - lat/9801035 ] .
smit , j. vink , j. , _ nucl .
phys . _ * b286 * , 485 ( 1987 ) ; + alls , b. et al . ,
_ phys . rev . _ * d58 * , 071503 ( 1998 ) .
chiu , t. , and zenkin , s.v .
rev . * d59 * , 074501 ( 1999 ) [ arxiv : hep - lat/9806019 ] .
( 0,0)(0,0 ) ( 10,10 ) .
we show the argand plot of the eigenvalues @xmath20 of the wilson @xmath179 , on a @xmath123 lattice and with parameters @xmath82 .
, title="fig : " ] ( 30,170)im@xmath180 ( 150,30)re@xmath180 ( 140,170)@xmath181 ( 0,0)(0,0 ) ( 10,10 ) .
we show the argand plot of the eigenvalues @xmath20 of the wilson @xmath179 , on a @xmath123 lattice and with parameters @xmath82 .
, title="fig : " ] ( 30,170)im@xmath180 ( 150,30)re@xmath180 ( 140,170)@xmath182 ( 0,0)(0,0 ) ( 10,10 ) .
we show the argand plot of the eigenvalues @xmath20 of the wilson @xmath179 , on a @xmath123 lattice and with parameters @xmath82 .
, title="fig : " ] ( 30,170)im@xmath180 ( 150,30)re@xmath180 ( 140,170)@xmath183 ( 0,0)(0,0 ) ( 10,10 ) .
we show the argand plot of the eigenvalues @xmath20 of the wilson @xmath179 , on a @xmath123 lattice and with parameters @xmath82 .
, title="fig : " ] ( 30,170)im@xmath180 ( 150,30)re@xmath180 ( 140,170)@xmath184 ( 0,0)(0,0 ) ( 10,10 ) .
we show the argand plot of the eigenvalues @xmath20 of the wilson @xmath179 , on a @xmath123 lattice and with parameters @xmath82 .
, title="fig : " ] ( 30,170)im@xmath180 ( 150,30)re@xmath180 ( 140,170)@xmath185 ( 0,0)(0,0 ) ( 10,10 ) .
we show the argand plot of the eigenvalues @xmath20 of the wilson @xmath179 , on a @xmath123 lattice and with parameters @xmath82 . , title="fig : " ] ( 30,170)im@xmath180 ( 150,30)re@xmath180 ( 140,170)@xmath186 | we study how far the index theorem can be extrapolated from the continuum to finite lattices with finite topological charge densities . to examine how the wilson action approximates the index theorem , we specialize in the lattice version of the schwinger model .
we propose a new criterion for solutions of the ginsparg - wilson relation constructed with the wilson action .
we conclude that the neuberger action is the simplest one that maximally complies with the index theorem , and that its best parameter in @xmath0 is @xmath1 .
address = dep .
fsica and cfif , inst .
sup .
tc .
, 1049 - 001 lisboa , portugal |
the concept of exchange interactions in almost - localized correlated fermions first appeared in the pioneering papers of heisenberg@xcite and was formulated in a more general mathematical way by dirac@xcite .
although the early papers by dirac already contain the general expression of the hamiltonian in terms of n - particle permutations , no attention was paid , up to the sixties , to higher orders than the `` pair - exchange '' heisenberg term .
thouless@xcite was the first to point out that higher order exchanges as three- and four - spin cyclic permutations might be important in quantum solids like @xmath1he .
but only ten years later , striking experimental results on nuclear magnetism in the bcc phase of solid @xmath1he , in the millikelvin range@xcite , were interpreted by hetherington , delrieu and roger through a ring - exchange model with two- three- and four - particle interactions of comparable magnitude@xcite . from general considerations put forward by thouless , permutations of even parity ( like three - particle cycles ) induce ferromagnetism while odd permutations ( pair and four - particle exchanges ) favor antiferromagnetism , and the striking phase diagram of bcc solid @xmath1he corresponds to a highly frustrated quantum - spin system with competing three- and four - spin permutations . during the last two decades a lot of progress has been accomplished in the investigation of solid @xmath1he films adsorbed on graphite , a simple model system exhibiting even more frustration since the frustrated nature of ring - exchange interactions is enhanced by the frustrated geometry of the triangular lattice . the conceptual beauty of solid @xmath1he lies in the fact that the system is simple enough ( the pair interactions between @xmath1he atoms are mainly of hard core nature ) to allow the calculation of exchange frequencies from first principles@xcite and a quantitative comparison with experimental results . the relevance of the multi - spin exchange concept is not restricted to the physics of nuclear @xmath1he spins .
delrieu@xcite suggested that three - spin exchange might be dominant in the two - dimensional electron wigner solid near the quasi - classical limit , and this has been corroborated through wkb calculations@xcite .
more recent monte - carlo simulations have proved that competing three and four - spin exchange interaction should occur near melting@xcite .
the first fourth - order @xmath4 expansion of the one - band hubbard hamiltonian in terms of four - spin interactions has been published in 1977 by takahashi@xcite . soon after the discovery of high - t@xmath5 superconductors , roger and delrieu@xcite suggested , on the basis of an expansion of the three - band hubbard model , that four - spin exchange might be significant in the cuo@xmath2 planes of cuprates . during the last ten years
many experimental results have revealed the presence of four - spin exchange interactions in the cu - o plaquettes of cuprates@xcite and copper - based spin - ladder materials@xcite .
dirac formalism is introduced in the framework of degenerate perturbation theory . the hamiltonian is written @xmath6 , where @xmath7 describes independent particles and v is a perturbation .
the `` unperturbed '' degenerate ground - states for n distinguishable particles can be written as products of independent particle states : @xmath8 .
each of these states is itself a product of two kets corresponding to the orbital and spin variables respectively : @xmath9 @xmath10=@xmath11 appears as a product of an orbital wavefunction : @xmath12 and a spin wavefunction : @xmath13 simply represents the site occupied by the particle ( i ) .
the permutation p of two particles can be expressed as a product of two operators : @xmath14 @xmath15 acting on orbitals and @xmath16 acting on spin variables .
if the hamiltonian does not depend explicitly on the spin , we can as a first step concentrate on the orbital part of the wave function and solve the eigenvalue problem : @xmath17 for the orbital wavefunction @xmath18 describing n _
distinguishable _ particles . the ground - state of the unperturbed part @xmath7 of the hamiltonian is n !
fold degenerated and the corresponding subspace @xmath19 is spanned by the n ! states : @xmath20 represents a permutation @xmath21 of the n integers @xmath22 .
we now apply degenerate perturbation theory@xcite to the perturbed hamiltonian @xmath23 . at first order , the splitting of the n !
degenerated energy levels is given by the eigenvalues of the hamiltonian @xmath24 defined by its matrix elements : @xmath25 where @xmath26 and @xmath27 are two permutations of the symmetric group @xmath28 . since @xmath29 is invariant with respect to any permutation ,
we can write : @xmath30 where @xmath31 , and the eigenvalue problem , restricted to the subspace @xmath19 can be formally represented by the hamiltonian : @xmath32 where the summation runs over permutations @xmath15 of the symmetric group @xmath28 .
this result extends straightforwardly to higher order degenerate perturbation theory : the higher orders are expressed in terms powers of @xmath29 and projection operator @xmath33 on @xmath19@xcite , and these operators commute with permutation operators .
hence , at arbitrary order in degenerate perturbation theory , we can write : @xmath34 we now have to introduce the spin degrees of freedom an express that the global wavefunction is completely antisymmetric . as a general result from group theory a completely antisymmetric wave function
can be expressed by the following bilinear expression@xcite : @xmath35 where @xmath36 represents a linear combination of different permutations @xmath37 corresponding to a given irreducible representation of the symmetry group schematized by a young diagram , while @xmath38 represents a linear combination of permutation @xmath39 corresponding to the representation associated with the `` complementary '' young diagram , obtained by exchanging the lines and the columns .
for spin-1/2 , a complete antisymmetrisation of the spins can not be realised over more than 2 variables , hence the corresponding young diagrams have at most two lines and each diagram corresponds to a given value of the total spin @xmath40 .
it is then possible to establish a correspondence between the expression [ eq .
( 10 ) ] of the hamiltonian acting only on the orbital variables with an equivalent hamiltonian acting only on the spin variables . expressing the antisymmetry of the wave function : @xmath41 where @xmath42 is the parity of the permutation , we can write : @xmath43 and multiplying to the left by @xmath44 : @xmath45 taking into account that in eq .
( 10 ) @xmath15 and the inverse permutation @xmath44 appear with the same weight @xmath46 , the hamiltonian is written equivalently in spin space : @xmath47 where the sum is over all permutations of the symmetric group @xmath28 acting on spin variables .
large u expansions of the one - band hubbard model@xcite have generally been performed within the framework of second quantization ( i.e working on the total antisymmetric wavefunction ) .
dirac formalism allows a simpler and more physical derivation .
we start with n distinguishable particles on n discrete sites , each site containing zero , one or two particles .
there is an onsite coulomb energy @xmath48 for putting two particles on the same site , and each particle has a probability @xmath49 to hop from one site to the nearest neighbor .
the hamiltonian does not depend explicitly on the spin variables . in perturbation theory ,
we start from the infinite @xmath48 limit : each site is occupied by one particle .
since the particles are considered as distinguishable , the ground state is
n ! fold degenerated .
we apply degenerate perturbation theory to express the energy splitting as an expansion in @xmath4 when @xmath48 is large but finite@xcite . up to fourth
order , the effective hamiltonian restricted to the subspace @xmath19 spanned by the n !
fold degenerated ground states is written@xcite : @xmath50\ ] ] where @xmath51 is the projection operator onto the subspace @xmath19 and @xmath52 in the fourth order expansion contribute to the cyclic exchange of the four particle coordinates @xmath53 , @xmath54 , @xmath55 and @xmath56 .
each of these four particle hops is accomplished by one of the four @xmath29 in the previous term and at each step a factor @xmath57 where @xmath58 is the number of doubly occupied sites , in the corresponding intermediate virtual state , has to be taken into account ( this factor arises from the operator @xmath40 ) .
there are 4!=24 terms depending on the order in which the four hops 1 , 2 , 3 , 4 , are successively realised .
eight of them [ corresponding to the sequences ( 1324 ) , ( 1342 ) , ( 3124 ) , ( 3142 ) , ( 2413 ) , ( 4213 ) , ( 2431 ) , ( 4231 ) ] contribute to a factor @xmath59 , the sixteen others contribute to @xmath60 .
the total contribution is then @xmath61 . ]
this expression can be expanded in term of two , three and four - particle permutations @xmath15 acting on coordinates : @xmath62- & \nonumber \\ k \sum_{<i , j , k , l > } [ p^r_{ijkl}+ ( p^r_{ijkl})^{-1 } ] & \nonumber \end{aligned}\ ] ] with : @xmath63 , @xmath64 , @xmath65 , @xmath66 and @xmath67 for the square lattice .
the sums in the first line are on first , second and third neighbor pairs respectively . the sum run over triangles in the second line and over square plaquettes in the last line . as a typical example , the fig .
1 illustrates the evaluation of the four - particle cyclic permutation term . since cyclic two and
four - particle permutations are of odd parity , while three - particle permutations are even , we obtain the corresponding exchange hamiltonian , with permutation operators @xmath16 acting on spin variables : : @xmath68 + \nonumber \\ & k \sum_{<i , j , k , l > } [ p^\sigma_{ijkl}+ ( p^\sigma_{ijkl})^{-1 } ] \end{aligned}\ ] ] using the identity : @xmath69 and the decomposition of any permutation in a product of transpositions , it is easy to check that this eq .
16 is identical to the result given by takahashi .
however our previous expression clearly separate the ferromagnetic contributions arising from three - particle exchange and the antiferromagnetic ones coming from pair- and four - particle exchange .
the formulation in terms of permutation operators is also generally more convenient for the calculation of the thermodynamic properties of a n - particle exchange hamiltonian .
dirac formalism has been generalized by thouless for the exchange - problem in solid @xmath1he@xcite and by herring for many - electron systems @xcite .
this has been applied later to ab - initio calculations of multi - spin exchange frequencies in solid @xmath1he and in the two - dimensional wigner - solid .
a pioneering multidimensional wkb calculation of the exchange frequencies has been applied to the two - dimensional wigner solid near the quasi - classical limit@xcite .
it has proved the conjecture by delrieu that three - particle exchange should dominate in that regime .
although , at physical densities , solid @xmath1he is far from the quasi - classical regime , the same wkb approach was extrapolated with qualitative results concerning the hierarchy of various cyclic - exchange frequencies . with the rapid increase of the performances of computers , ab - initio calculations of ring - exchange frequencies through path
integral monte carlo ( pimc ) became possible .
in contrast to wkb evaluation they are relevant to the whole range of physical densities , including solid phases near melting where quantum fluctuations are large .
the exchange frequencies for bcc solid @xmath1he@xcite near melting , are in excellent agreement with the experimental thermodynamic properties , which can be inferred through exact high - temperature series expansions .
although the long - range coulomb potential is completely different from the hard core potential between helium atoms , wkb calculations near the classical limit@xcite and pimc in the whole density range@xcite give striking analogies between solid @xmath1he films and the electron wigner solid , on the same triangular lattice .
in both systems , three - particle exchange dominates near the quasi - classical limit , leading to ferromagnetism , while near melting , antiferromagnetic two- and four - particle exchange compete with ferromagnetic three - particle exchange leading to highly frustrated antiferromagnetic systems .
close to melting , both systems might be prototypes of two - dimensional `` spin liquids''@xcite with no long range order at t=0 .
experiments on the magnetism of the wigner solid are now at the very beginning@xcite , but the nuclear magnetism of solid @xmath1he films has now been extensively studied .
exfoliated graphite offers large ( 10x10 nm ) flat crystallites with a strong adsorption potential for he .
up to two solid he films can be adsorbed on this substrate , within a wide density - range .
the magnetism of 2d helium films has been thoroughly studied , down to 10 - 100 @xmath70k in : 1 .
submonolayer solid @xmath1he films on graphite@xcite 2 .
solid @xmath1he films on graphite coated by a high density @xmath1he@xcite or @xmath71he@xcite layer or by a hd bilayer@xcite .
the main difference between ( i ) and ( ii ) is that , due to a stronger adsorption potential , the first layer is closer to the pure 2d case . at the same density exchange frequencies in the first layer
are 10 times lower than in the second layer where @xmath1he atom permutations are made easier through excursion in the third dimension .
however the same general trends are observed : * a low - density commensurate solid phase is stabilized by the graphite potential in case ( i ) and by the periodic potential of the first layer in case ( ii ) .
its susceptibility has an antiferromagnetic character with a negative curie - weiss constant .
both susceptibility and specific heat present anomalous deviations with respect to the usual asymptotic behavior at high temperature ( curie weiss law for the susceptibility and t@xmath72 behavior for the specific heat ) .
these features are due to the frustrated nature of the system and can be quantitatively fit through the ring - exchange model@xcite . down to 10
@xmath70k no long - range ordering has been detected , which seems to confirm the presence of a spin - liquid state .
an ultra - low temperature specific - heat measurement@xcite and susceptibility measurements@xcite put a higher limit of 10 - 100 @xmath70k on a possible gap in the excitation spectrum .
hence solid @xmath1he films could be the first experimental realization of a gapless spin liquid .
* at higher densities , an incommensurate solid phase is observed .
the susceptibility has a positive curie weiss constant , but the unusual features of the specific heat indicate that we have a `` frustrated ferromagnetic '' phase with relatively larger three - spin exchange but significant competing four - particle exchange@xcite .
the frustration is easily `` tunable '' in the sense that when the density is further increased , three - spin exchange dominates leading to a more conventional ferromagnet .
some anomalies in the specific heat are present only in the commensurate phases .
a very interesting interpretation has been recently proposed by h. fukuyama@xcite , in terms of `` ground - state vacancies '' surrounded by magnetic polarons .
this exciting idea established a close link with strongly interacting almost - localized electron systems close to a mott - hubbard transition ( e.g the physics of cuprates ! ) .
in the insulating phases of cuprates , the first experimental results in contradiction with a pure heisenberg model were the anomalously broad multi - magnon raman spectrum@xcite and the infrared optical absorption by phonon assisted multimagnon excitations@xcite in la@xmath2cuo@xmath73 .
both results can be understood with the occurrence of four - spin exchange@xcite .
however the key experiments which really invalidates the pure heisenberg model is the inelastic neutron study of the magnon spectrum near the brillouin zone@xcite .
it is in agreement with the presence of ferromagnetic three - spin exchange terms @xmath74 as those appearing in the expansion of the one - band hubbard model ( eq .
however , due to accidental cancellations occurring for the nel phase of the square lattice , the linear spin - wave spectrum is insensitive to the presence of four - spin exchange operators@xcite .
an excellent way to probe directly the presence of four - spin terms is the study of spin - spin correlations in the paramagnetic phase .
these correlation have been recently measured through diffuse magnetic scattering of polarized neutrons in la@xmath2cuo@xmath73 at @xmath75@xcite .
the results are compared to predictions from high - temperature series expansions of the hamiltonian .
they prove definitively the presence of four - spin interaction with a ratio @xmath76 of about 20% .
the same cu - o plaquettes are present in copper based quasi-1d spin ladder materials sr@xmath77cu@xmath78o@xmath79 and la@xmath80ca@xmath81cu@xmath78o@xmath79 , with approximately the same cu - o bonds and distances as in la@xmath2cuo@xmath82 . the first neighbor pair exchanges along rungs and legs are expected to be roughly equal @xmath83 .
in that case the theory predict a gap @xmath84 in the excitation spectrum .
the experimental value of the gap is substantially lower : @xmath85@xcite . a relatively small four - spin exchange @xmath86 accounts for the decrease of the gap@xcite .
this effect is related to the frustration due to four - spin exchange , which increases the density of low lying eigenstates . _
the concept of multi - particle exchange , pioneered by dirac at the beginning of the 20@xmath87 century , now appears as essential in the physics of strongly - correlated fermion systems which have raised considerable interest during the last decades . _
w. heisenberg , z. phys , * 38 * ( 1926 ) 411 ; _ ibid _ * 49 * ( 1928 ) 619 .
a. m. dirac , proc .
soc . * a 123 * ( 1929 ) 714 .
p. a. m dirac , _ the principles of quantum mechanics _
( clarendon , oxford ) ch .
ix , ( 1947 ) .
thouless , proc .
london * 86 * ( 1965 ) 893 .
kummer et al , phys .
lett . * 34 * ( 1975 ) 517 ; e.d adams , j.m .
a. landesman , j. phys . lett .
* 39 * ( 1978 ) l190 .
h. godfrin et al . , phys .
* 44 * ( 1980 ) 1695 .
osheroff , m.c .
cross , d.e .
fisher , phys .
* 44 * ( 1980 ) 792 .
hetherington , phys .
* 35 * ( 1975 ) 1442 .
delrieu et al .
* 40 * ( 1980 ) 71 .
m. roger , j.m .
delrieu , j.h .
hetherington , rev .
* 55 * ( 1983 ) 1 . d.m .
ceperley , g. jacucci , phys .
rev . lett .
* 58 * ( 1987 ) 1648 .
delrieu , private communication ( 1980 ) .
m. roger , phys .
b*30 * ( 1984 ) 6432
. b. bernu , l. candido , d.m .
ceperley , phys .
* 86 * ( 2001 ) 870
. m. takahashi , j. phys . c.
* 10 * ( 1977 ) 1289 .
m. roger , j.m .
delrieu , phys .
rev . b*39 * ( 1989 ) 2299 .
s. sugai et al .
rev . b*42 * ( 1990 ) 1405 .
r. coldea et al .
* 86 * ( 2001 ) 5377 .
a. toader et al . to be published j. lorenzana , j. eroles , s. sorella , phys .
* 83 * ( 1999 ) 5122 .
imai et al .
phys . rev .
* 81 * ( 1998 ) 220 .
eccleston et al .
* 81 * ( 1998 ) 1702 .
s. brehmer et al .
b * 60 * ( 1999 ) 329 . c. herring in _ magnetism vol .
ii b _ ed .
rado and suhl , ( academic press , london ) ch .
i , ( 1966 ) .
t. kato , progress of theoretical physics * iv * , 514 ( 1949 ) ; ibid * v * ( 1950 ) 207 . c. bloch , nuclear physics * 6 * ( 1958 ) 329 .
l. landau , e. lifchitz , _ quantum mechanics _ , ( pergamon press , london ) ch .
ix , ( 1959 ) .
m. katano , d.s .
hirashima , phys .
b * 62 * ( 2000 ) 2573 .
g. misguich et al . , phys .
* 81 * ( 1998 )
1098 . t. okamoto , s. kawaji , phys . rev .
b * 57 * ( 1998 ) 9097
. m. morishita , h. nagatani , h. fukuyama , phys .
b * 65 * ( 2002 ) 104524 .
the density of the completed first @xmath1he layer is so high that its magnetism is negligible . d.s .
greywall , phys .
b * 41 * ( 1990 ) 1842
. m. siquera et al . , phys .
* 78 * ( 1997 ) 2600 .
k. ishida et al . , phys .
* 79 * ( 1997 ) 3451 .
m. roger et al .
* 80 * ( 1998 ) 1308 .
h. ikegami et al . , phys . rev .
* 85 * ( 2000 ) 5146 .
e. collin et al .
* 86 * ( 2001 ) 2447 .
r. masutomi , y. karaki , h. ishimoto , phys .
* 92 * ( 2004 ) 025301 j.d .
perkins et al .
, phys . rev
* 71 * ( 1993 ) 1621 . | the concept of exchange in strongly - correlated fermions is reviewed with emphasis on the generalization of the heisenberg pair exchange to higher order @xmath0-particle permutations .
the `` frustration '' resulting from competing ferromagnetic three - spin exchange and antiferromagnetic two- and four - spin exchanges is illustrated on a two - dimensional model system : solid @xmath1he films .
recent experimental results proving the presence of significant four - spin exchange interactions in the cuo@xmath2 plaquettes of high @xmath3 cuprates are reported .
quantum spin systems , @xmath1he , cuprates . |
the diffuse @xmath0-ray emission , both galactic and extragalactic , is of significant interest for astrophysics , particle physics , and cosmology . the diffuse galactic emission ( dge ) is produced by interactions of crs , mainly protons and electrons , with the interstellar gas ( via @xmath3-production and bremsstrahlung ) and radiation field ( via inverse compton [ ic ] scattering ) .
it is a direct probe of cr fluxes in distant locations , and may contain signatures of physics beyond the standard model , such as dark matter annihilation or decay . the dge is a foreground for point - source detection and hence influences the determination of their positions and fluxes .
it is also a foreground for the much fainter extragalactic component , which is the sum of contributions from unresolved sources and truly diffuse emission , including any signatures of large scale structure formation , emission produced by ultra - high - energy crs interacting with relic photons , and many other processes ( e.g. , @xcite and references therein ) .
therefore , understanding the dge is a necessary first step in all such studies .
the excess diffuse emission @xmath1 gev in the egret data @xcite relative to that expected from dge models consistent with the directly measured cr nucleon and electron spectra @xcite led to the proposal that this emission was the long - awaited signature of dark matter annihilation @xcite .
more conventional interpretations included variations of cr spectra in the galaxy @xcite , contributions by unresolved point sources @xcite , and instrumental effects @xcite .
a model of the dge depends on the cr spectra throughout the galaxy as well as the distribution of the target gas and interstellar radiation field ( isrf ) .
starting from the distribution of cr sources and particle injection spectra , the distribution of crs throughout the galaxy is determined taking into account relevant energy losses and gains , then the cr distributions are folded with the target distributions to calculate the dge @xcite . defining the inputs and calculating the models are not trivial tasks and involve analysis of data from a broad range of astronomical and astroparticle instruments @xcite .
the _ fermi _ lat was launched on june 11 , 2008 .
it is over an order of magnitude more sensitive than its predecessor , egret , with a more stable response due to the lack of consumables .
the lat data permit more detailed studies of the dge than have been possible ever before . in this paper ,
analysis and results for the dge are shown for the galactic mid - latitude range @xmath2 measured by the lat in the first 5 months of the science phase of the mission .
this region was chosen for initial study since it maximises the fraction of signal from dge produced within several kpc of the sun and hence uncertainties associated with cr propagation , knowledge of the gas distribution , etc .
, should be minimised . at lower galactic latitudes
the large - scale dge is dominant while the emission at higher latitudes is more affected by contamination from charged particles misclassified as photons and uncertainties in the model used to estimate the dge .
further details of the present analysis are given in @xcite
the lat is a pair - conversion telescope with a precision tracker and calorimeter , each consisting of a @xmath4 array of 16 modules , a segmented anti - coincidence detector ( acd ) that covers the tracker array , and a programmable trigger and data acquisition system .
full details of the instrument , onboard and ground data processing , and other mission - oriented support are given in @xcite .
the data selection used in this paper is made using the standard lat ground processing and background rejection scheme .
this consists of two basic parts : first a simple accept - or - reject selection ( prefiltering ) followed by a classification tree ( ct ) @xcite based determination of the relative probability of being background or signal .
the prefiltering phase screens particles entering the lat for their charge neutrality using the tracker and acd .
the direction reconstruction software extrapolates found particle trajectories in the tracker back to the scintillation tiles of the acd covering the lat and only accepts events in which the intersected tile shows no significant signal .
in addition , considerations such as how well the found tracks project into the measured energy centroid in the calorimeter , and the shape of the shower energy deposition , are also used .
the overall background rejection of the prefiltering phase is @xmath5 depending on energy , yielding a @xmath0-ray efficiency @xmath6 relative to @xmath0-rays that convert in the lat .
classification trees , which afford an efficient and statistically robust method for distinguishing signal from noise , are used to reduce backgrounds further .
the cts are trained using monte carlo data which have passed the prefiltering described above .
multiple cts are used to make the procedure robust against statistical fluctuations during the training procedure . the result from averaging these cts is the probability for an event to be a photon or background .
this final rejection parameter allows the background levels to be set according to the needs of the analysis . for the analysis of diffuse emission ,
the ct generated probability is set to allow a monte carlo predicted orbit - averaged background rate of @xmath7 hz integrated over the full instrument acceptance @xmath8 mev .
this yields a @xmath0-ray efficiency @xmath9 , and the remaining background is at a level where the majority of the contamination arises from irreducible sources such as @xmath0-rays produced by cr interactions in the passive material outside the acd , e.g. , the thermal blanket and micrometeroid shield for the lat .
the events corresponding to the above criterion are termed `` diffuse '' class and are the standard low - background event selection .
the analysis presented here uses post - launch instrument response functions ( irfs ) .
these take into account pile - up and accidental coincidence effects in the detector subsystems that are not considered in the definition of the pre - launch irfs .
cosmic rays , primarily protons , pass through the lat at a high rate and sufficiently near coincidences with @xmath0-rays leave residual signals that can result in @xmath0-rays being misclassified , particularly at energies @xmath10 mev .
the post - launch irfs were derived using lat events read out at regular intervals as a background overlay on the standard simulations of @xmath0-rays and provide an accurate accounting for the instrumental pile - up and accidental coincidence effects .
the on - axis effective area for the event selection used in this paper is @xmath11 @xmath12 at 1 gev and is energy dependent ; this is approximately 10% lower than the pre - launch effective area corresponding to the same event selection .
the systematic uncertainties of the effective area , evaluated by comparing the efficiencies of analysis cuts for data and simulation of observations of vela , are also energy dependent : 10% below 100 mev , decreasing to 5% at 560 mev , and increasing to 20% at 10 gev and above .
the point spread function ( psf ) and energy resolution are as described in @xcite .
the lat nominally operates in a scanning mode that covers the whole sky every two orbits ( i.e. , 3 hrs ) .
we use data taken in this mode from the commencement of scientific operations in mid - august 2008 to the end of december 2008 .
the data were prepared using the lat science tools package , which is available from the _ fermi _ science support centre .
events satisfying the diffuse class selection and coming from zenith angles @xmath13 ( to greatly reduce the contribution by earth albedo @xmath0-rays ) were used . to further reduce the effect of earth albedo backgrounds , the time intervals when the earth was appreciably within the field of view ( specifically , when the centre of the field of view was more than 47@xmath14 from the zenith ) were excluded from this analysis .
this leaves 9.83 ms of total livetime in the data set .
the energy - dependent exposure was calculated using the irfs described above .
the photon counts and exposure were further processed using the _ gardian _ package , part of a suite of tools we have developed to analyse the dge @xcite .
gamma - ray skymaps with 5 bins per decade in energy from 100 mev to 10 gev were generated .
for each energy bin the intensity was obtained by dividing the in - bin counts by the spectrally - weighted exposure over the bin .
we used two methods for the spectral weighting : a power law with index @xmath15 and the spectral shape of the assumed dge model ( described below ) . with
the energy binning used in this paper the differences in the derived intensities were @xmath16% between these two weighting schemes .
figure [ fig1 ] ( left ) shows the lat data averaged over all galactic longitudes and the latitude range @xmath2 .
the hatched band surrounding the lat data indicates the systematic uncertainty in the measurement due to the uncertainty in the effective area described above .
also shown are the egret data for the same region of sky derived from count maps and exposures available via the _ cgro _ science support centre and processed following the procedure described in @xcite and we have included the standard systematic uncertainty of 13% @xcite . for both data
sets the contribution by point sources has not been subtracted .
the lat - measured spectrum is significantly softer than the egret measurement with an integrated intensity @xmath17 @xmath18 s@xmath19 sr@xmath19 compared to the egret integrated intensity @xmath20 @xmath18 s@xmath19 sr@xmath19 where the errors are statistical only .
not included in the figure is the systematic uncertainty in the energy scale , which is conservatively estimated from comparison between monte carlo and beam test data as @xmath21% for 100 mev to 1 gev , and @xmath22% above 1 gev where it is believed that if any bias is present energies are overestimated . taking the uncertainty on the energy scale into account ,
the lat spectrum could be softer , increasing the discrepancy with the egret spectrum further .
figure [ fig1 ] ( right ) compares the lat spectrum with the spectra of an _ a priori _ dge model , and a point - source contribution and unidentified background ( uib ) component derived from fitting the lat data that are described below .
the dge model is an updated version of the `` conventional '' model from galprop @xcite .
major improvements include use of the formalism and corresponding code for pion production in @xmath23-interactions by @xcite , a complete recalculation of the isrf @xcite , updated gas maps , and an improved line - of - sight integration routine .
the source and uib components were obtained by fitting the lat data using _ gardian _ with the dge model fixed .
point source locations were taken from the 3 month _ fermi _ lat source list down to sources with 5-@xmath24 significance . due to the limited statistics of all but the very brightest sources
, we used 3 bins per energy decade in the fitting procedure .
source positions were fixed but the spectra were fit using one free parameter per energy bin .
the uib component was determined by fitting the data and sources over all galactic longitudes for the high - latitude region @xmath25 for the full lat energy range shown in the figure .
using this high - latitude region minimises the effect of contamination by the bright galactic ridge which can be significant even up to @xmath26 from the plane due to the long tails of the psf at low energies . to determine the uncertainty of the source and uib components
, we modified the effective area to the extremes of its systematic uncertainty defined before and refitted the data . since the dge model is fixed , the absolute change in intensity caused by the modification to the effective area propagates directly to the source and uib components .
the systematic uncertainty on these components is energy dependent and due to several effects . for energies
@xmath27 gev the psf is @xmath28 ( 68% containment ) and the sources are well - localised spatially . since the model is fixed and the sky maps are sparser at high latitudes for the data taking period in this paper , the uib component absorbs almost all of the intensity from the modification to the effective area . at low energies
the psf is wider , @xmath29 ( 68% containment ) at 100 mev for @xmath0-ray conversions in the front section of the lat , and the sources are less well - localised spatially . in addition , the sky maps are well populated even at high latitudes and display spatial structure .
the psf broadening of the sources provides spatial structure and because the dge model is fixed , more intensity is assigned to the source component to compensate in the fit .
these effects lead to the systematic error in the source component being relatively larger than the isotropic at low energies and vice versa at high energies .
note , this applies for the high - latitude region from where the uib component is derived , and also for the mid - latitude range for which we show the combined contribution by sources in fig .
[ fig1 ] ( right ) .
because the uncertainties in the source and uib components are not independent we have conservatively added their systematic uncertainties for the total intensity band shown in fig .
[ fig1 ] ( right ) .
the uib component comprises the true extragalactic diffuse @xmath0-ray emission , emission from unresolved galactic and extragalactic sources , and residual particle backgrounds ( crs that pass the @xmath0-ray classification analysis and @xmath0-rays produced by cr interactions in the passive material outside the acd ) in the lat data . in addition
, other relevant foreground components that are not completely modelled , such as emission from the solar disk and extended emission @xcite and other potentially relevant `` diffuse '' sources @xcite are included .
hence , the uib component does _ not _ constitute a measurement of the extragalactic diffuse emission .
furthermore , comparison with the egret estimate of the extragalactic diffuse emission @xcite is problematic due to the different dge models used and analysis details that are beyond the scope of the current paper and will be addressed in a subsequent publication .
the intensity scales of the lat and egret have been found to be different with the result that the lat - measured spectra are softer . in our early study of the vela spectrum @xcite , which was made using pre - launch irfs ,
the difference was apparent already above 1 gev .
following on - orbit studies new irfs have been developed to account for inefficiencies in the detection of @xmath0-rays in the lat due to pile - up and accidental coincidence effects in the detector subsystems .
the inefficiency increases at lower energies , with the result that the irfs used in the present analysis indicate greater intensities in the range below 1 gev , with the magnitude of the effect ranging up to 30% at 100 mev .
our confidence that the irfs used in the present analysis accurately represent our knowledge of the instrument comes from detailed instrument simulations that were validated with beam tests of calibration units , and to post - launch refinements based on actual particle backgrounds .
the systematic uncertainty on the effective area gives an energy dependent measure of our confidence in the irfs used in the present analysis . as a consequence ,
the lat - measured dge spectrum averaged over all galactic longitudes for the latitude range @xmath2 is systematically softer than the egret - measured spectrum .
the spectral shape is compatible with that of an _ a priori _ dge model that is consistent with directly measured cr spectra .
the excess emission above 1 gev measured by egret is not seen by the lat in this region of the sky .
while the lat spectral shape is consistent with the dge model used in this paper , the overall model emission is too low thus giving rise to a @xmath30% excess over the energy range 100 mev to 10 gev .
however , the dge model is based on pre-_fermi _ data and knowledge of the dge .
the difference between the model and data is of the same order as the uncertainty in the measured cr nuclei spectra at the relevant energies @xcite .
in addition , other model parameters that can affect the @xmath0-ray production rate ( e.g. , @xmath31- the conversion between co line intensity and molecular hydrogen column density in the interstellar medium ) have not been modified in the present paper .
overall , the agreement between the lat - measured spectrum and the model shows that the fundamental processes are consistent with our data , thus providing a solid basis for future work understanding the dge . _
acknowledgements : _ the _ fermi _ lat collaboration acknowledges support from a number of agencies and institutes for both development and the operation of the lat as well as scientific data analysis .
these include nasa and doe in the united states , cea / irfu and in2p3/cnrs in france , asi and infn in italy , mext , kek , and jaxa in japan , and the k. a. wallenberg foundation , the swedish research council and the national space board in sweden . additional support from inaf in italy for science analysis during the operations phase is also gratefully acknowledged . | the diffuse galactic @xmath0-ray emission is produced by cosmic rays ( crs ) interacting with the interstellar gas and radiation field .
measurements by the energetic gamma - ray experiment telescope ( egret ) instrument on the _ compton gamma - ray observatory ( cgro ) _ indicated excess @xmath0-ray emission @xmath1 gev relative to diffuse galactic @xmath0-ray emission models consistent with directly measured cr spectra ( the so - called `` egret gev excess '' ) .
the excess emission was observed in all directions on the sky , and a variety of explanations have been proposed , including beyond - the - standard - model scenarios like annihilating or decaying dark matter .
the large area telescope ( lat ) instrument on the _ fermi _ gamma - ray space telescope has measured the diffuse @xmath0-ray emission with unprecedented sensitivity and resolution .
we report on lat measurements of the diffuse @xmath0-ray emission for energies 100 mev to 10 gev and galactic latitudes @xmath2 . the lat spectrum for this region of the sky is well reproduced by the diffuse galactic @xmath0-ray emission models mentioned above and inconsistent with the egret gev excess .
gamma rays , cosmic rays , fermi gamma - ray space telescope |
experiments to date exploit wide feshbach resonances and are thus well described by the simplest single - channel hamiltonian , where the two fermion species interact via an attractive contact potential @xmath2 here , @xmath3 ( we set @xmath4 and @xmath5 ) , @xmath6 is the volume , and we define the chemical potential @xmath7 and ` zeeman ' field @xmath8 such that @xmath9 and @xmath10 . at present , only pairing between different hyperfine species of the _ same _ atom has been explored experimentally , so we restrict ourselves to a single mass @xmath11 . the interaction strength @xmath12 is expressed in terms of the s - wave scattering length @xmath13 using the prescription : @xmath14 we also derive the fermi momentum using the average density @xmath15 , so that @xmath16 .
throughout our calculations , we will keep @xmath17 fixed .
the full phase diagram is parameterised by just a few observables : the temperature @xmath18 , the interaction strength @xmath19 , and the density difference or ` magnetisation ' @xmath20 . to determine the position of the phase boundaries
, we must minimise the mean - field free energy density @xmath21,\end{gathered}\ ] ] with respect to the bcs order parameter @xmath22 , where @xmath23 and @xmath24 .
such a mean - field analysis provides a reasonable description of the zero temperature phase diagram , but at finite temperature , it neglects the contribution of non - condensed pairs to both the density @xmath25 and magnetisation @xmath26 .
this contribution is necessary to approach the transition temperature of an ideal bose gas in the molecular limit , and can be included in the non - condensed phase ( @xmath27 ) through the nozires - schmitt - rink ( nsr ) fluctuation correction to the energy @xcite @xmath28 with @xmath29 + \tanh[\frac{\beta}{2}(\xi_{{{\mathbf p}}+{{\mathbf q } } } - h)]}}{i\omega + \xi_{{{\mathbf p } } } + \xi_{{{\mathbf p}}+{{\mathbf q } } } } \ ; .\end{gathered}\ ] ] this gives an estimate of the effect of pair fluctuations on the second order phase boundary ( but not the first order boundary , where @xmath30 ) . and magnetisation @xmath31 ( inset ) versus interaction @xmath19 .
there are four different phases : the normal ( n ) state , the phase - separated ( ps ) state , the ordinary superfluid ( sf ) and the magnetised superfluid ( sf@xmath32 ) . above the line @xmath33 , the normal state is completely polarised ( @xmath34 ) .
the red and black lines enclosing the ps state are both first - order phase boundaries , while the sf@xmath32-n transition is second - order , and the sf - sf@xmath32 transition ( green line ) is at least third - order .
the tricritical point is represented by orange circles at @xmath35 with @xmath36 or @xmath34 .
[ fig : zerot],scaledwidth=45.0% ]
considerable insight can be gained by first examining the zero temperature mean - field phase diagram , as shown in fig .
[ fig : zerot ] .
the general structure parallels that of the two - channel case found in ref . @xcite .
since there is a gap in the quasiparticle excitation spectrum @xmath37 of the unpolarised superfluid , the superfluid ground state will remain unchanged for @xmath38 .
we see that the @xmath39 superfluid line in the inset of fig .
[ fig : zerot ] corresponds to an _ area _ in the @xmath40 versus @xmath19 diagram , which expands as @xmath19 increases .
a key feature of the strong coupling side is that for @xmath41 the superfluid state is able to sustain a finite population of majority quasiparticles .
this `` gapless '' @xcite superfluid phase is only stable for @xmath42 and it thus possesses only one fermi surface . in the extreme bec limit , this state is straightforwardly understood as an almost ideal mixture of bosonic pairs and fermionic quasiparticles . however , as we move towards unitarity , the bosons and fermions begin to interact more strongly , leading eventually to a first - order phase transition to the normal state . here , a system with fixed @xmath43 will undergo phase separation into normal and superfluid regions if @xmath44 , where @xmath45 denotes the magnetisation in the normal and superfluid phases at @xmath46 , the critical field for the first - order transition . in the bcs limit ( @xmath47 ) , @xmath48 which is less than the quasiparticle gap , so the superfluid state is unmagnetised @xmath49 , and phase separation occurs for arbitrarily low magnetisation , consistent with ref .
@xcite . for the moment we neglect the fflo state , but will return to this point later .
a crucial observation is that the line @xmath34 to the right of the region of phase separation can be thought of as a continuous zero temperature transition at which the condensate is totally depleted .
it is thus natural to identify the point on @xmath34 where phase separation starts as a tricritical point . indeed a landau expansion of the free energy both confirms this and identifies the tricritical point at @xmath35 . and interaction @xmath19 .
the plane at temperature @xmath50 is the phase diagram in fig .
[ fig : zerot ] .
the yellow line represents the locus of tricritical points calculated in the mean - field approximation , while the orange tricritical line corresponds to mean - field theory plus pair fluctuations .
the fluctuation correction breaks down in the unitarity regime @xmath51 , and is thus shown as a dotted line .
the slice at @xmath52 is based on a mean - field calculation and it shows the region of phase separation terminating in a tricritical point ( yellow circle ) at finite temperature , followed by a second - order phase transition from the superfluid to normal state .
note that the boundary between the fflo and normal states ( blue line ) defines a small region of fflo phase confined to the bcs side of the crossover , as explained in the text.,scaledwidth=45.0% ] .
the effective interaction is parameterised by the detuning @xmath53 .
the colour scheme for tricritical lines is the same as in fig .
[ fig : tricritical].,scaledwidth=45.0% ] with this background , we now turn to the analysis of the fate of the tricritical point when temperature is finite , beginning with the mean - field description .
it is well known that there exists a finite temperature tricritical point in the bcs limit @xmath54 , which is a natural consequence of having a first - order transition from the superfluid to normal state at @xmath50 and a second - order transition at @xmath39 .
first studied by sarma in the context of superconductivity in the presence of a magnetic field @xcite , the bcs tricritical point is located at @xmath55 @xcite , where @xmath56 $ ] ( i.e. at weak coupling all energies scale with @xmath22 ) .
this corresponds to a magnetisation @xmath57 , where @xmath58 is the fermi surface density of states . to investigate how the bcs tricritical point is related to the one at zero temperature ,
we must develop a perturbative expansion of eq .
( [ eq : energy ] ) for small @xmath22 and general @xmath19 .
doing so , one finds ( fig .
[ fig : tricritical ] ) that the tricritical point at @xmath34 is connected to that in the bcs limit by a line of tricritical points that passes through a maximum somewhere in the ` unitarity ' regime @xmath59 .
moreover , for any given value of @xmath60 , the @xmath61 phase diagram is highly reminiscent of the @xmath0he-@xmath1he system , with @xmath31 playing the role of the fraction of @xmath0he .
this is not surprising , as the finite @xmath43 system corresponds in general to a mixture of bosonic pairs and fermionic quasiparticles .
note that even the gapped superfluid can be magnetised at finite temperature due to thermal excitation of quasiparticles . of course , at @xmath39
the transition into the superfluid state is second order at any point in the bcs - bec crossover .
it is interesting to examine how the fflo phase fits in with the basic topology of the phase diagram . in the bcs limit
, we already know that the point where the fflo - normal phase boundary meets the normal - superfluid boundary asymptotes to the tricritical point @xcite .
assuming that the transition from the fflo state to the normal state is second - order ( although ref .
@xcite found it to be weakly first order , this will make a relatively small difference ) , and performing a mean - field analysis , we find that the fflo point of intersection leaves the finite temperature tricritical point with increasing interaction ( see fig .
[ fig : tricritical ] ) , leading eventually to the extinction of the fflo phase at @xmath62 .
note that although this treatment is somewhat approximate , as we have taken the sf - fflo boundary to be the same as the sf - n boundary in the absence of fflo , the point of intersection will coincide with that derived from a complete mean - field analysis .
moreover , despite all our assumptions , we expect the detachment of the point of intersection from the tricritical point and the eventual disappearance of fflo to be robust features , since in the bec regime we essentially have a mixture of bosons and fermions . the inclusion of the fluctuation contribution eq .
( [ fluct ] ) is crucial for recovering the extreme bec limit , where it is clear that the ( second - order ) transition temperature asymptotes to @xmath63 ( with @xmath64 ) , the ideal bec temperature of a gas of bosons of density @xmath65 and mass @xmath66 .
more importantly , we find that fluctuations shift the mean - field tricritical line to lower temperatures and magnetisations on the bec side , while leaving the tricritical points on the bcs side largely unchanged , as expected .
however , in a broad region around unitarity , we find that the approximation underlying eq .
( [ fluct ] ) generally leads to non - monotonic behavior of @xmath67 , with @xmath68 for small @xmath8 .
we interpret this behaviour as a breakdown of the nsr treatment , yielding an unphysical compressibility matrix @xmath69 that is not positive semi - definite . to address this problem
, we note that the nsr scheme is a controlled approximation when we introduce resonant scattering with a finite width , with the width being a small fraction of the fermi energy @xcite .
the simplest such description is provided by the two - channel model @xcite .
the two - channel description of scattering depends upon two parameters : a detuning @xmath53 describing the distance from the resonance , and a width @xmath70 of the resonance measured in units of the fermi energy .
the one - channel description is recovered in the @xmath71 limit , while the treatment of gaussian fluctuations is essentially perturbative in @xmath70 , with @xmath72 in eq .
( [ fluct ] ) being replaced with @xmath73 , so in this case the nsr treatment is expected to be accurate .
the resulting phase diagram is shown in figure [ fig : tricritical_2chan ] .
the zero temperature phase diagram coincides with the result of ref . @xcite . with fluctuations accounted for , and for sufficiently small @xmath70
, we now find a well - behaved line of tricritical points spanning the crossover region .
we expect that the true phase boundary at @xmath71 is qualitatively similar . in the @xmath74-@xmath75 plane .
the red and black lines are first- and second - order phase boundaries , respectively .
the arrows at constant @xmath75 represent the trajectories followed when going from the centre to the edges of a trapped gas .
the two trajectories correspond to two different magnetisations of the gas : one greater and one less than the tricritical point @xmath76.,scaledwidth=42.0% ]
we now discuss the consequences of our results for trapped gases studied in experiment . modeling the trapped gas by the local density approximation ( lda )
, the spatial dependence of the density induced by the trapping potential @xmath77 is accounted for by a spatially - varying chemical potential @xmath78 , with @xmath8 kept constant . in the @xmath74-@xmath75 plane , we thus move on a horizontal line ( see fig . [
fig : tri_trap ] ) . at sufficiently low temperatures ,
a trapped gas will consist of a superfluid core surrounded by the normal state .
the transition between normal and superfluid states in the trap can be either second or first order , depending on whether @xmath75 is above or below the tricritical point .
moreover , as long as the temperature is non - zero , we can always find a sufficiently small @xmath8 so that @xmath75 lies above the tricritical point .
this leads us to a key point : if a trapped gas at a given temperature and magnetisation has a first - order transition between its normal and superfluid phases , then we will _
always _ cross the tricritical point by decreasing the magnetisation at fixed temperature .
we emphasise that there are qualitative differences between first and second order transitions in a trap : the former yields a discontinuity in the density and magnetization at the phase interface , resulting in a form of phase separation as seen in recent experiments @xcite , while the latter possesses a density that varies smoothly in space .
therefore , the magnetisation and temperature at which a tricritical point is crossed should be detectable experimentally .
in fact , a critical magnetisation for the onset of phase separation in a trap has been observed experimentally @xcite , and a calculation by chevy supports the idea that this coincides with crossing a tricritical point @xcite .
in addition , the order of the transition will have an impact on experiments that use phase separation as a signature of superfluidity @xcite .
the presence of a first - order transition in the trap can be even more pronounced if the density discontinuities result in a breakdown of lda .
experiments on highly elongated traps already provide evidence for such a breakdown @xcite , and one requires the addition of surface energy terms at the phase interface to successfully model the trapped density profiles @xcite .
an outstanding issue is the experimental detection of the gapless sf@xmath32 phase .
while optically probing the momentum distribution of the minority species is one promising method for detecting sf@xmath32 @xcite , another possibility is to study density correlations using , for example , shot noise experiments as suggested in ref .
a simple mean - field calculation gives ( for the uniform system ) : @xmath79 ^ 2\end{aligned}\ ] ] where @xmath80 is the fermi - dirac distribution . at @xmath50 ,
the result is a ` hole ' in the correlation function for momenta less than the fermi wavevector of the majority quasiparticles .
such a measurement would therefore constitute both a confirmation of the sf@xmath32 phase and a vivid demonstration of the blocking effect of quasiparticles on @xmath81 pairing . in conclusion
, we have determined the structure of the finite temperature phase diagram of the two component fermi gas , as a function of both interaction strength and population imbalance , finding a region of phase separation terminating in a tricritical point for general coupling in the bcs - bec crossover .
a secondary result of our work is the demonstration that the nsr scheme yields unphysical results in a broad region around unitarity .
this is significant , as it is widely viewed as offering a smooth , albeit uncontrolled approximation throughout the crossover .
we emphasize that there is no _ a priori _ reason to believe in the accuracy of the nsr scheme without introducing an additional parameter , as we have done here .
the ginzburg criterion governing the smallness of fluctuation corrections is satisfied in both the bcs limit where it takes the form @xmath82 , and in the bec limit where @xmath83 is the relevant criterion . but at unitarity the shift in the transition temperature relative to the mean field value will be of order @xmath84 . at the same time
the upper critical dimension at the tricritical point is three , so we may expect that our results there will be little changed . finally , we have argued that these tricritical points play an important role in experiments on trapped fermi gases ( see , also , the subsequent related work on trapped gases at unitarity by gubbels et al .
indeed , a recent comprehensive study of the temperature dependence of the phase - separated state at unitarity has yielded experimental results consistent with the phase diagram outlined here @xcite . | the two - component fermi gas is the simplest fermion system displaying superfluidity , and as such is relevant to topics ranging from superconductivity to qcd .
ultracold atomic gases provide an exceptionally clean realisation of this system , where interatomic interactions and atom spin populations are both independently tuneable . here
we show that the finite temperature phase diagram contains a region of phase separation between the superfluid and normal states that touches the boundary of second - order superfluid transitions at a tricritical point , reminiscent of the phase diagram of @xmath0he-@xmath1he mixtures .
a variation of interaction strength then results in a line of tricritical points that terminates at zero temperature on the molecular bose - einstein condensate ( bec ) side . on this basis
, we argue that tricritical points are fundamental to understanding experiments on polarised atomic fermi gases
. over the past decade , experimental progress in the field of cold atomic gases has resulted in unprecedented control over pairing phenomena in two - component fermi gases . the ability to vary the effective interaction between atoms using magnetically tuned feshbach resonances
has already permitted the experimental investigation of the crossover from a bec of diatomic molecules to the bardeen - cooper - schrieffer ( bcs ) limit of weakly - bound cooper pairs of fermionic atoms @xcite .
a natural extension of these studies is an exploration of the fermi gas with imbalanced spin populations , especially since this system has a far richer phase diagram than the equal spin case . as well as exhibiting a quantum phase transition between the superfluid and normal states ,
the polarized fermi gas has been predicted to possess exotic superfluid phases such as the inhomogeneous fulde - ferrell - larkin - ovchinnikov ( fflo ) state @xcite , where the pairing of fermions occurs at finite centre - of - mass momentum , and the deformed fermi surface state @xcite .
the exact nature of the superfluid states for the polarised fermi gas is still the subject of considerable debate .
however , atomic gases provide an ideal testing ground for this system , since the particle numbers can be varied independently from all other experimental parameters , and pioneering experiments have recently been performed @xcite .
contrast atomic gases with the case of superconductors , where the magnetic field used to generate a spin imbalance ( via the zeeman effect ) also couples to orbital degrees of freedom . in this work , we elucidate the finite temperature phase diagram of a polarised fermi gas . while much insight has been gained from previous theoretical studies @xcite , so far a key ingredient of the phase diagram has been overlooked : the tricritical point , at which the phase transition between superfluid and normal states switches from first to second order . by determining the behaviour of the tricritical point as a function of interaction strength , we can completely characterise the topology of the phase diagram without recourse to an extensive numerical treatment .
specifically , we shall focus on the uniform , infinite system , and concern ourselves almost exclusively with the phase boundary between the normal and homogeneous superfluid states .
we will , however , discuss the ramifications of the inferred phase diagram for the trapped system . |
we consider the gaussian regression framework @xmath4 where @xmath5 is an unknown vector of @xmath6 and the @xmath7 are independent centered gaussian random variables with common variance @xmath2 . throughout the paper
, @xmath8 is assumed to be unknown which corresponds to the practical case .
our aim is to estimate @xmath0 from the observation of @xmath1 . for specific forms of @xmath0 , this setting allows to deal simultaneously with the following problems .
[ ex - sd ] the vector @xmath0 is of the form @xmath9 where @xmath10 are distinct points of a set @xmath11 and @xmath12 is an unknown mapping from @xmath11 into @xmath13 .
[ ex - vs ] the vector @xmath0 is assumed to be of the form @xmath14 where @xmath15 is a @xmath16 matrix , @xmath17 is an unknown @xmath18-dimensional vector and @xmath18 some integer larger than 1 ( and possibly larger than @xmath19 ) .
the columns of the matrix @xmath15 are usually called predictors .
when @xmath18 is large , one may assume that the decomposition ( [ sv ] ) is sparse in the sense that only few @xmath20 are non - zero .
estimating @xmath0 or finding the predictors associated to the non - zero coordinates of @xmath17 are classical issues .
the latter is called variable selection .
our estimation strategy is based on estimator selection .
more precisely , we start with an arbitrary collection @xmath21 of estimators of @xmath0 based on @xmath1 and aim at selecting the one with the smallest euclidean risk by using the same observation @xmath1 . the way the estimators @xmath22 depend on @xmath1 may be arbitrary and possibly unknown . for example
, the @xmath22 may be obtained from the minimization of a criterion , a bayesian procedure or the guess of some experts .
the problem of choosing some best estimator among a family of candidate ones is central in statistics .
let us present some examples .
[ ex - ctp ] many statistical procedures depend on a ( possibly multi - dimensional ) parameter @xmath23 that needs to be tuned in view of obtaining an estimator with the best possible performance .
for example , in the context of linear regression as described in example [ ex - vs ] , the lasso estimator ( see tibshirani @xcite and chen _ et al . _
@xcite ) defined by @xmath24 with @xmath25}\ ] ] depends on the choice of the parameter @xmath26 .
selecting this parameter among a grid @xmath27 amounts to selecting a ( suitable ) estimator among the family @xmath28 .
another dilemma for statisticians is the choice of a procedure to solve a given problem . in the context of example [ ex - ctp ] , there exist many competitors to the lasso estimator and one may alternatively choose a procedure based on ridge regression ( see hoerl and kennard @xcite ) , random forest or pls ( see tenenhaus @xcite , helland @xcite and helland @xcite ) .
similarly , for the problem of signal denoising as described in example [ ex - sd ] , popular approaches include spline smoothing , wavelet decompositions and kernel estimators
. the choice of a kernel may be possibly tricky .
consider the problem described in example [ ex - sd ] with @xmath29 .
for a kernel @xmath30 and a bandwidth @xmath31 , the nadaraya - watson estimator ( see nadaraya @xcite and watson @xcite ) @xmath32 is defined as @xmath33 where for @xmath34 @xmath35 there exist many possible choices for the kernel @xmath30 , such as the gaussian kernel @xmath36 , the uniform kernel @xmath37 , etc .
given a ( finite ) family @xmath38 of candidate kernels @xmath30 and a grid @xmath39 of possible values of @xmath40 , one may consider the problem of selecting the best kernel estimator among the family @xmath41 . a common way to address the above issues is to use some cross - validation scheme such as leave - one - out or @xmath42-fold .
even though these resampling techniques are widely used in practice , little is known on their theoretical performances . for more details
, we refer to arlot and celisse @xcite for a survey on cross - validation technics applied to model selection .
compared to these approaches , as we shall see , the procedure we propose is less time consuming and easier to implement .
moreover , it does not require to know how the estimators depend on the data @xmath1 and we can therefore handle the following problem .
a statistician is given a collection @xmath28 of estimators from a family @xmath43 of experts @xmath23 , each of which keeping secret the way his / her estimator @xmath22 depends on the observation @xmath1 .
the problem is then to find which expert @xmath23 is the closest to the truth .
given a selection rule among @xmath3 , an important issue is to compare the risk of the selected estimator to those of the candidate ones .
results in this direction are available in the context of model selection , which can be seen as a particular case of estimator selection .
more precisely , for the purpose of selecting a suitable model one starts with a collection @xmath44 of those , typically linear spaces chosen for their approximation properties with respect to @xmath0 , and one associates to each model @xmath45 a suitable estimator @xmath46 with values in @xmath47 . selecting a model
then amounts to selecting an estimator among the collection @xmath48 . for this problem ,
selection rules based on the minimization of a penalized criterion have been proposed in the regression setting by yang @xcite , baraud @xcite , birg and massart @xcite and baraud _ et al _ @xcite .
another way , usually called lepski s method , appears in a series of papers by lepski @xcite and was originally designed to perform model selection among collections of nested models . finally , we mention that other procedures based on resampling have interestingly emerged from the work of arlot @xcite and clisse @xcite .
a common feature of those approaches lies in the fact that the proposed selection rules apply to specific collections of estimators only .
an alternative to _ estimator selection _ is _ aggregation _ which aims at designing a suitable combination of given estimators in order to outperform each of these separately ( and even the best combination of these ) up to a remaining term .
aggregation techniques can be found in catoni @xcite , juditsky and nemirovski @xcite , nemirovski @xcite , yang @xcite , @xcite , @xcite , tsybakov @xcite , wegkamp @xcite , birg @xcite , rigollet and tsybakov @xcite , bunea , tsybakov and wegkamp @xcite and goldenshluger @xcite for @xmath49-losses .
most of the aggregation procedures are based on a sample splitting , one part of the data being used for building the estimators , the remaining part for selecting among these .
such a device requires that the observations be i.i.d . or at least that one has at disposal two independent copies of the data .
from this point of view our procedure differs from classical _ aggregation _ procedures since we use the whole data @xmath1 to build and select . in the gaussian regression setting that is considered here , we mention the results of leung and barron @xcite for the problem of mixing least - squares estimators .
their procedure uses the same data @xmath1 to estimate and to aggregate but requires the variance to be known .
giraud @xcite extends their results to the case where it is unknown .
our approach for solving the problem of estimator selection is new .
we introduce a collection @xmath44 of linear subspaces of @xmath6 for approximating the estimators in @xmath3 and use a penalized criterion to compare them . as already mentioned and as we shall see , this approach requires no assumption on the family of estimators at hand and is easy to implement , an r - package being available on .
a general way of comparing estimators in various statistical settings has been described in baraud @xcite .
however , the procedure proposed there is mainly abstract and inadequate in the gaussian framework we consider .
we prove a non - asymptotic risk bound for the estimator we select and show that this bound is optimal in the sense that it essentially can not be improved ( except for numerical constants maybe ) by any other selection rule .
for the sakes of illustration and comparison , we apply our procedure to various problems among which aggregation , model selection , variable selection and selection among linear estimators . in each of these cases ,
our approach allows to recover classical results in the areas as well as to establish new ones . in the context of aggregation
we compute the aggregation rates for the unknown variance case .
these rates turn out to be the same as those for the known variance case . for selecting an estimator among a family of linear ones
, we propose a new procedure and establish a risk bound which requires almost no assumption on the considered family .
finally , our approach provides a way of selecting a suitable variable selection procedure among a family of candidate ones .
it thus provides an alternative to cross - validation for which little is known .
the paper is organized as follows . in section [
sect - main ] we present our selection rule and the theoretical properties of the resulting estimator . for illustration , we show in sections [ sect - aggreg ] , [ sect - linear ] and [ sect - vs ] respectively , how the procedure can be used to aggregate preliminary estimators , select a linear estimator among a finite collection of candidate ones , or solve the problem of variable selection .
section [ sec : numerique ] is devoted to two simulation studies .
one aims at comparing the performance of our procedure to the classical @xmath42-fold in view of selecting a tuning parameter among a grid . in the other
, we evaluate the performance of the variable selection procedure we propose to some classical ones such as the lasso , random forest , and others based on ridge and pls regression .
finally , the proofs are postponed to section [ sect - proof ] . throughout the paper @xmath50 denotes a constant that may vary from line to line .
given a collection @xmath51 of estimators of @xmath0 based on @xmath1 , the selection rule we propose is based on the choices of a family @xmath44 of linear subspaces of @xmath6 , a collection @xmath52 of ( possibly random ) subsets of @xmath44 , a weight function @xmath53 and a penalty function @xmath54 , both from @xmath44 into @xmath55 .
we introduce those objects below and refer to sections [ sect - aggreg ] , [ sect - linear ] and [ sect - vs ] for examples .
the collection @xmath56 can be arbitrary .
in particular , @xmath3 need not be finite nor countable and it may consist of a mix of estimators based on the minimization of a criterion , a bayes procedure or the guess of some experts . the dependency of these estimators with respect to @xmath1 need not be known .
nevertheless , we shall see on examples how we can use this information , when available , to improve the performance of our estimation procedure .
let @xmath44 be a family of linear spaces of @xmath6 satisfying the following .
[ h0 ] the family @xmath44 is finite or countable and for all @xmath45 , @xmath58 . to each estimator @xmath59
, we associate a ( possibly random ) subset @xmath60 .
typically , the family @xmath44 should be chosen to possess good approximation properties with respect to the elements of @xmath3 and @xmath57 with respect to @xmath22 specifically .
one may take @xmath61 but for computational reasons it will be convenient to allow @xmath57 to be smaller .
the choices of @xmath57 may be made on the basis of the observation @xmath22 .
we provide examples of @xmath44 and @xmath57 in various statistical settings described in sections [ sect - aggreg ] to [ sect - vs ] .
we consider a function @xmath53 from @xmath44 into @xmath55 and assume [ h1 ] @xmath63 whenever @xmath44 is finite , inequality ( [ sigma ] ) automatically holds true .
however , in practice @xmath64 should be kept to a reasonable size .
when @xmath65 , @xmath66 can be interpreted as a prior distribution on @xmath44 and gives thus a bayesian flavor to the procedure we propose . to the weight function @xmath53 ,
we associate the function @xmath62 mapping @xmath44 into @xmath55 and defined by @xmath67}=e^{-\delta(s ) } \label{edk.eq}\ ] ] where @xmath68 denotes the positive part of @xmath69 and @xmath70 are two independent @xmath71 random variables with respectively @xmath72 and @xmath73 degrees of freedom .
this function can be easily computed from the quantiles of the fisher distribution as we shall see in section [ calcpen.st ] . from a more theoretical point of view , it is shown in baraud _
et al _ @xcite that under assumption [ h4 ] below , there exists a positive constant @xmath50 ( depending on @xmath74 only ) such that @xmath75 [ h4 ] there exists @xmath76 such that for all @xmath77 , @xmath78 the selection procedure we propose involves a penalty function @xmath54 from @xmath44 into @xmath55 with the following property .
[ h2 ] the penalty function @xmath54 satisfies for some @xmath79 , @xmath80 whenever equality holds in ( [ pen ] ) , it derives from ( [ eq : oom ] ) that @xmath81 measures the complexity of the model @xmath47 in terms of dimension and weight . denoting @xmath82 the projection operator onto a linear space @xmath83 , given the families @xmath57 , the penalty function @xmath54 and some positive number @xmath84
, we define @xmath85},\ ] ] where @xmath86 for all @xmath87 let us set @xmath88}.\ ] ] this quantity corresponds to an accuracy index for the estimator @xmath22 with respect to the family @xmath57 .
the following result holds .
[ main ] let @xmath89 .
assume that assumptions [ h0 ] , [ h1 ] and [ h2 ] hold .
there exists a constant @xmath50 ( given by ( [ cste.eq ] ) ) depending on @xmath30 and @xmath84 only such that for any @xmath90 in @xmath3 satisfying @xmath91 we have the following bounds @xmath92 } & \le & \e{\left[{\inf_{\lambda\in\lambda}{\left[{{\left\|{f-\widehat f_{\lambda}}\right\|}^{2}+a(\widehat f_{\lambda},\ss_{\lambda})}\right]}}\right]}+\sigma\sigma^{2}+\delta \label{eqbase}\\ & \le & \inf_{\lambda\in\lambda}{\left\{{\e{\left[{{\left\|{f-\widehat f_{\lambda}}\right\|}^{2}}\right]}+\e{\left[{a(\widehat f_{\lambda},\ss_{\lambda})}\right]}}\right\}}+\sigma\sigma^{2 } + \delta \;\;\label{eq3}\end{aligned}\ ] ] ( provided that the quantity involved in the expectation in ( [ eqbase ] ) is measurable ) .
furthermore , if equality holds in ( [ pen ] ) and assumption [ h4 ] is satisfied , for each @xmath87 * if the set @xmath57 is non - random , @xmath93}}\nonumber\\ & \le&\!\!\!\!\e{\left[{{\left\|{f-\widehat f_{\lambda}}\right\|}^{2}}\right]}+\inf_{s\in\ss_{\lambda}}{\left[{\e{\left[{{\left\|{\widehat f_{\lambda}-\pi_{s}\widehat f_{\lambda}}\right\|}^{2}}\right]}+(\dim(s)\vee \delta(s))\sigma^{2}}\right ] } \label{bornea1}\end{aligned}\ ] ] * if there exists a ( possibly random ) linear space @xmath94 such that @xmath95 with probability 1 , @xmath96}\le \e{\left[{{\left\|{f-\widehat f_{\lambda}}\right\|}^{2}}\right]}+\e{\left[{\dim(\widehat s_{\lambda})\vee \delta(\widehat s_{\lambda})}\right]}\sigma^{2},\ ] ] where @xmath97 is a positive constant only depending on @xmath74 and @xmath30 .
let us now comment theorem [ main ] .
it turns out that inequality ( [ eqbase ] ) leaves no place for a substantial improvement in the sense that the bound we get is essentially optimal and can not be improved ( apart from constants ) by any other selection rule among @xmath3 . to see this ,
let us assume for simplicity that @xmath3 is finite so that a measurable minimizer of @xmath98 always exists and @xmath99 can be chosen as 0 .
let @xmath100 , @xmath101 ( to fix up the ideas ) , @xmath44 a family of linear spaces satisfying the assumptions of theorem [ main ] and @xmath54 , the penalty function achieving equality in ( [ pen ] ) .
besides , assume that @xmath44 contains a linear space @xmath47 such that @xmath102 and associate to @xmath47 the weight @xmath103 . if @xmath61 for all @xmath23 , we deduce from ( [ eq : oom ] ) and ( [ eqbase ] ) that for some universal constant @xmath97 , whatever @xmath3 and @xmath104 @xmath105}}\nonumber\\ & \le & \e{\left[{\inf_{\lambda\in\lambda}{\left[{{\left\|{f-\widehat f_{\lambda}}\right\|}^{2}+\inf_{s\in\ss}{\left({{\left\|{\widehat f_{\lambda}-\pi_{s}\widehat f_{\lambda}}\right\|}^{2}+{\mathop{\rm pen}\nolimits}(s)\widehat \sigma^{2}_{s}}\right)}}\right]}}\right]}\nonumber\\ & \le & \e{\left[{\inf_{\lambda\in\lambda}{\left[{{\left\|{f-\widehat f_{\lambda}}\right\|}^{2}+{\left\|{\widehat f_{\lambda}-\pi_{s}\widehat f_{\lambda}}\right\|}^{2}+\dim(s)\widehat \sigma^{2}_{s}}\right]}}\right]}.\label{1000borne}\end{aligned}\ ] ] in the opposite direction , the following result holds .
[ prop - opt ] there exists a universal constant @xmath50 , such that for any finite family @xmath28 of estimators and any selection rule @xmath106 based on @xmath1 among @xmath43 , there exists @xmath107 such that @xmath108}\ge \e{\left[{\inf_{\lambda\in\lambda}{\left[{{\left\|{f-\widehat f_{\lambda}}\right\|}^{2}+{\left\|{\widehat f_{\lambda}-\pi_{s}\widehat f_{\lambda}}\right\|}^{2}+\dim(s)\sigma^{2}}\right]}}\right]}.\ ] ] we see that , up to the estimator @xmath109 in place of @xmath2 and numerical constants , the left - hand sides of ( [ 1000borne ] ) and ( [ borne1000 ] ) coincide . in view of commenting ( [ eq3 ] )
further , we continue assuming that @xmath3 is finite so that we can keep @xmath110 in ( [ eq3 ] ) .
a particular feature of ( [ eq3 ] ) lies in the fact that the risk bound pays no price for considering a large collection @xmath3 of estimators .
in fact , it is actually decreasing with respect to @xmath3 ( or equivalently @xmath43 ) for the inclusion .
this means that if one adds a new estimator to the collection @xmath3 ( without changing neither @xmath44 nor the families @xmath57 associated to the former estimators ) , the risk bound for @xmath90 can only be improved .
in contrast , the computation of the estimator @xmath90 is all the more difficult that @xmath111 is large .
more precisely , if the cardinalities of the families @xmath57 are not too large , the computation of @xmath90 requires around @xmath111 steps .
the selection rule we use does not require to know how the estimators depend on @xmath1 .
in fact , as we shall see , a more important piece of information is the ranges of the estimators @xmath112 as @xmath1 varies in @xmath6 .
a situation of special interest occurs when each @xmath22 belongs to some ( possibly random ) linear space @xmath113 in @xmath44 with probability one . by taking @xmath57 such that @xmath94 for all @xmath23
, we deduce from theorem [ main ] by using ( [ eq3 ] ) and ( [ bornea2 ] ) the following corollary .
[ maincor ] assume that the assumptions of theorem [ main ] are satisfied , that assumption [ h4 ] holds and that equality holds in ( [ pen ] ) . if for all @xmath87 there exists a ( possibly random ) linear space @xmath114 such that @xmath95 with probability 1 , then @xmath90 satisfies @xmath115}\le \inf_{\lambda\in\lambda}{\left[{\e{\left[{{\left\|{f-\widehat f_{\lambda}}\right\|}^{2}}\right]}+\e{\left[{\dim(\widehat s_{\lambda})\vee \delta(\widehat s_{\lambda})}\right]}\sigma^{2}}\right]}+\delta,\ ] ] for some @xmath50 depending on @xmath30 and @xmath74 only .
one may apply this result in the context of model selection .
one starts with a collection of models @xmath116 and associate to each @xmath117 an estimator @xmath118 with values in @xmath117 . by taking @xmath119 ( here @xmath120 ) and @xmath121 for all @xmath122
, our selection procedure leads to an estimator @xmath123 which satisfies @xmath124}\le \inf_{m\in\m}{\left[{\e{\left[{{\left\|{f-\widehat f_{m}}\right\|}^{2}}\right]}+{\left({\dim(s_{m})\vee \delta(s_{m})}\right)}\sigma^{2}}\right]}.\ ] ] when @xmath125 for all @xmath122 , our selection rule becomes @xmath126}\ ] ] and turns out to coincide with that described in baraud _
interestingly , corollary [ maincor ] shows that this selection rule can still be used for families @xmath3 of ( non - linear ) estimators of the form @xmath127 where the @xmath128 are chosen randomly among @xmath44 on the basis of @xmath1 , doing thus as if the linear spaces @xmath128 were non - random .
an estimator of the form @xmath127 can be interpreted as resulting from a model selection procedures among the family of projection estimators @xmath129 and hence , ( [ critsm ] ) can be used to choose some best model selection rule among a collection of candidate ones .
in this section , we consider the problems of _ model selection aggregation _ ( ms ) , _ convex aggregation _ ( @xmath130 ) and _ linear aggregation _ ( l ) defined below .
given @xmath131 preliminary estimators of @xmath0 , denoted @xmath132 , our aim is to build an estimator @xmath133 based on @xmath1 whose risk is as close as possible to @xmath134 where @xmath135 and , according to the aggregation problem at hand , @xmath43 is one of the three sets @xmath136 when @xmath137 , @xmath138 is the set @xmath139 consisting of the initial estimators .
when @xmath140 , @xmath138 is the convex hull of the @xmath141 . in the literature , one may also find @xmath142^{m},\ \sum_{j=1}^{m}\lambda_{j}\le 1}\right\}}\ ] ] in place of @xmath143 in which case @xmath138 is the convex hull of @xmath144 .
finally , when @xmath145 , @xmath138 is the linear span of the @xmath141 .
each of these three aggregation problems are solved _ separately _ if for each @xmath146 one can design an estimator @xmath147 satisfying @xmath148}-c\inf_{g\in\ff_{\lambda}}{\left\|{f - g}\right\|}^{2}\le c'\psi_{n,\lambda}\sigma^{2}\ ] ] with @xmath149 , @xmath150 free of @xmath151 and @xmath152 these problems have only been considered when the variance is known .
the quantity @xmath153 then corresponds to the best possible upper bound in ( [ aggreg1 ] ) over all possible @xmath104 and preliminary estimators @xmath141 and is called the _ optimal rate of aggregation_. for a more precise definition , we refer the reader to tsybakov @xcite .
@xcite considered the problem of solving these three problems _ simultaneously _ by building an estimator @xmath133 which satisfies ( [ aggreg1 ] ) simultaneously for all @xmath146 and some constant @xmath154 .
this is an interesting issue since it is impossible to know in practice which aggregation device should be used to achieve the smallest risk bound : as @xmath43 grows ( for the inclusion ) , the bias @xmath134 decreases while the rate @xmath153 increases .
the aim of this section is to show that our procedure provides a way of solving ( or nearly solving ) the three aggregation problems both _ separately _ and _ simultaneously _ when the variance is unknown . throughout this section
, we consider the family @xmath155 consisting of the @xmath117 defined for each @xmath156 and @xmath157 as the linear span of the @xmath141 for @xmath158 . along this section
, we shall use the weight function @xmath53 defined on @xmath155 by @xmath159},\ ] ] take @xmath101 and @xmath160 taking thus @xmath100 .
the choices of @xmath84 and @xmath30 is only to fix up the ideas .
note that @xmath53 satisfies assumption [ h1 ] with @xmath161 . to avoid trivialities , we assume all along @xmath162 .
problem ( l ) is the easiest to solve .
let us take @xmath163 with @xmath145 and @xmath164 and @xmath165 for all @xmath166 .
minimizing @xmath167 over @xmath168 amounts to minimizing @xmath169 over @xmath170 and hence , the resulting estimator is merely @xmath171 .
the risk of @xmath172 satisfies @xmath173}\le \inf_{g\in \ff_{\lambda}}{\left\|{f - g}\right\|}^{2}+ m\sigma^{2}.\ ] ] whatever @xmath19 and @xmath174 which solves the problem of _ linear aggregation_. to tackle problem ( ms ) , we take @xmath163 with @xmath137 , that is , @xmath175 , @xmath176 and associate to each @xmath177 the collection @xmath57 reduced to @xmath178 .
note that @xmath179 and @xmath180 for all @xmath181 , so that under the assumption that @xmath182 we may apply corollary [ maincor ] with @xmath110 ( since @xmath138 is finite ) , @xmath183 and get that for some constant @xmath184 the resulting estimator @xmath185 satisfies @xmath186}\le \inf_{g\in \ff_{\lambda}}{\left\|{f - g}\right\|}^{2}+ \log(m)\sigma^{2}.\ ] ] this risk bound is of the form ( [ aggreg1 ] ) except for the constant @xmath50 which is not equal to 1 .
we do not know whether problem ( ms ) can be solved or not with @xmath149 when the variance @xmath2 is unknown and @xmath174 is large ( possibly larger than @xmath19 ) . for this problem
, we emphasize the aggregation rate with respect to the quantity @xmath187 if @xmath188 , take again the estimator @xmath172 . since the convex hull of the @xmath141 is a subset of the linear space @xmath189 , for @xmath140 we have @xmath190}\le \inf_{g\in \ff_{\lambda}}{\left\|{f - g}\right\|}^{2}+ m\sigma^{2}.\ ] ]
let us now turn to the case @xmath191 .
more precisely , assume that @xmath192 and set @xmath193 .
we consider the family of estimators @xmath163 with @xmath140 and @xmath194 the set @xmath143 being compact , @xmath195 admits a minimum @xmath196 over @xmath143 and we set @xmath197 .
[ prop - convex ] there exists a universal constant @xmath154 such that @xmath198}-c\inf_{g\in\ff_{\lambda}}{\left\|{f - g}\right\|}^{2}\le c\sqrt{nl^{2}\log(em/\sqrt{nl^{2}})}\sigma^{2}.\ ] ] this risk bound is of the form ( [ aggreg1 ] ) except for the constant @xmath50 which is not equal to 1 .
again , we do not know whether problem ( @xmath130 ) can be solved or not with @xmath149 when the variance @xmath2 is unknown and @xmath174 possibly larger than @xmath19 .
consider now three estimators @xmath199 with values respectively in @xmath189 , @xmath200 and the convex hull @xmath201 of the @xmath141 ( we use a new notation for this convex hull to avoid ambiguity ) .
one may take the estimators defined in section [ aggregm ] but any others would suit .
the aim of this section is to select the one with the smallest risk to estimate @xmath0 .
to do so , we apply our selection procedure with @xmath202 , taking thus @xmath203 , and associate to each of these three estimators the families @xmath204 defined by ( [ mla ] ) , ( [ mmsa ] ) and ( [ mca ] ) respectively and choose @xmath205 . [ prop - simult ] assume that ( [ hm ] ) holds and that @xmath182 .
there exists a universal constant @xmath184 such that whatever @xmath206 and @xmath207 with values in @xmath189 , @xmath200 and @xmath201 respectively , the selected estimator @xmath90 satisfies for all @xmath104 , @xmath208}\le \inf_{\lambda\in { \left\{{{\rm l},{\rm ms},\cv}\right\ } } } { \left [ { \e{\left[{{\left\|{f-\widehat f_{\lambda}}\right\|}^{2}}\right ] } + b_{\lambda}}\right]},\ ] ] where @xmath209}.\ ] ] in particular , if @xmath206 and @xmath207 fulfills ( [ aggreg1 ] ) , then @xmath208}\le \inf_{\lambda\in { \left\{{{\rm l},{\rm ms},\cv}\right\ } } } { \left [ { \inf_{g\in\ff_{\lambda}}{\left\|{f - g}\right\|}^{2}+b_{\lambda}}\right]},\ ] ] where @xmath210 stands for @xmath138 when @xmath211 .
in this section , we consider the situation where the estimators @xmath22 are linear , that is , are of the form @xmath212 for some known and deterministic @xmath213 matrix @xmath214 . as mentioned before , this setting covers many popular estimation procedures including kernel ridge estimators , spline smoothing , nadaraya estimators , @xmath23-nearest neighbors , projection estimators , low - pass filters , etc . in some cases
@xmath215 is symmetric ( e.g. kernel ridge , spline smoothing , projection estimators ) , in some others @xmath215 is non - symmetric and non - singular ( as for nadaraya estimators ) and sometimes @xmath215 can be both singular and non - symmetric ( low pass filters , @xmath23-nearest neighbors ) .
a common feature of those procedures lies in the fact that they depend on a tuning parameter ( possibly multidimensional ) and their practical performances can be quite poor if this parameter is not suitably calibrated .
a series of papers have investigated the calibration of some of these procedures . to mention a few of them ,
cao and golubev @xcite focus on spline smoothing , zhang @xcite on kernel ridge regression , goldenshluger and lepski @xcite on kernel estimators and arlot and bach @xcite propose a procedure to select among symmetric linear estimator with spectrum in @xmath216 $ ] .
the procedure we present can handle all these cases in an unified framework . throughout the section ,
we assume that @xmath43 is finite . to apply our selection procedure , we need to associate to each @xmath215 a suitable collection of approximation spaces @xmath57 . to do so
, we introduce below a linear space @xmath217 which plays a key role in our analysis . for the sake of simplicity , let us first consider the case where @xmath215 is non - singular .
then @xmath217 is defined as the linear span of the right - singular vectors of @xmath218 associated to singular values smaller than 1 .
when @xmath215 is symmetric , @xmath217 is merely the linear span of the eigenvectors of @xmath215 associated to eigenvalues not smaller than 1/2 . if none of the singular values are smaller than 1 , then @xmath219 .
let us now extend the definition of @xmath217 to singular operators @xmath215 .
let us recall that @xmath220 where @xmath221 stands for the transpose of @xmath215 and @xmath222 for its range .
the operator @xmath215 then induces a one to one operator between @xmath223 and @xmath224 .
write @xmath225 for the inverse of this operator from @xmath224 to @xmath222 .
the orthogonal projection operator from @xmath6 onto @xmath222 induces a linear operator from @xmath224 into @xmath222 , denoted @xmath226 .
then @xmath217 is defined as the linear span of the right - singular vectors of @xmath227 associated to singular values smaller than 1 . again
if this set is empty , @xmath219 .
when @xmath215 is non - singular or symmetric , we recover the definition of @xmath217 given above . for each @xmath87 , take @xmath57 such that @xmath228 . from a theoretical point of view
, it is enough to take @xmath229 but practically it may be wise to use a larger set and by doing so , to possibly improve the approximation of @xmath22 by elements of @xmath57 .
one may for example take @xmath230 where @xmath231 is the linear span of the right - singular vectors associated to the @xmath232 smallest singular values of @xmath227 .
take @xmath233 and @xmath53 of the form @xmath234 where @xmath235 satisfies assumption [ h1 ] with @xmath236 .
one may take @xmath237 even though this choice is not necessarily the best . finally , for some @xmath79 , take @xmath238 for all @xmath45 and select @xmath90 by minimizing the criterion given by ( [ eq : criterion ] ) , taking thus @xmath110 in ( [ def - est ] ) .
the following holds .
[ linear estimators ] let @xmath79 , @xmath239 and @xmath240 .
if assumption [ h0 ] holds and @xmath241 for all @xmath45 , the estimator @xmath90 satisfies @xmath242}\le \inf_{\lambda}\e{\left[{{\left\|{f-\widehat f_{\lambda}}\right\|}^{2}}\right]}+\sigma^{2},\ ] ] for some @xmath50 depending on @xmath243 and @xmath74 only .
the problem of selecting some best linear estimator among a family of those have also been considered in arlot and bach @xcite in the gaussian regression framework , and in goldenshluger and lepski @xcite in the multidimensional gaussian white noise model .
arlot and bach proposed a penalized procedure based on random penalties . unlike ours , their approach requires that the operators be symmetric with eigenvalues in @xmath216 $ ] and that the cardinality of @xmath43 is at most polynomial with respect to @xmath19 .
goldenshluger and lepski proposed a selection rule among families of kernel estimators to solve the problem of structural adaptation .
their approach requires suitable assumptions on the kernels while ours requires nothing .
nevertheless , we restrict to the case of the euclidean loss whereas goldenshluger and lepski considered more general @xmath49 ones .
throughout this section , we consider the problem of variable selection introduced in example [ ex - vs ] and assume that @xmath244 in order to avoid trivialities .
when @xmath18 is small enough ( say smaller than 20 ) , this problem can be solved by using a suitable variable selection procedure that explores all the subsets of @xmath245 .
for example , one may use the penalized criterion introduced in birg and massart @xcite when the variance is known , and the one in baraud _
@xcite when it is not .
when @xmath18 is larger , such an approach can no longer be applied since it becomes numerically intractable . to overcome this problem , algorithms based on the minimization of convex criteria
have been proposed among which are the lasso , the dantzig selector of cands and tao @xcite , the elastic net of zou and hastie @xcite .
an alternative to those criteria is the forward - backward algorithm described in zhang @xcite , among others .
since there seems to be no evidence that one of these procedures outperforms all the others , it may be reasonable to mix them all and let the data decide which is the more appropriate to solve the problem at hand . as enlarging @xmath3
can only improve the risk bound of our estimator , only the cpu resources should limit the number of candidate estimators .
the procedure we propose could not only be used to select among those candidate procedures but also to select the tuning parameters they depend on . from this point of view
, it provides an alternative to the cross - validation techniques which are quite popular but offer little theoretical guarantees .
start by choosing a family @xmath246 of variable selection procedures .
examples of such procedures are the lasso , the dantzig selector , the elastic net , among others .
if necessary , associate to each @xmath247 a family of tuning parameters @xmath248 .
for example , in order to use the lasso procedure one needs to choose a tuning parameter @xmath31 among a grid @xmath249 .
if a selection procedure @xmath250 requires no choice of tuning parameters , then one may take @xmath251 .
let us denote by @xmath252 the subset of @xmath245 corresponding to the predictors selected by the procedure @xmath250 for the choice of the tuning parameter @xmath40 . for @xmath253
, let @xmath117 be the linear span of the column vectors @xmath254 for @xmath158 ( with the convention @xmath255 ) . for @xmath256 and @xmath257 ,
associate to the subset @xmath258 an estimator @xmath259 of @xmath0 with values in @xmath260 ( one may for example take the projection of @xmath1 onto the random linear space @xmath260 but any other choice would suit ) .
finally , consider the family @xmath21 of these estimators by taking @xmath261 and set @xmath262 .
all along we assume that @xmath43 is finite ( so that we take @xmath110 in ( [ def - est ] ) ) . throughout
, we shall restrict ourselves to subsets of predictors with cardinality not larger than some @xmath263 . in view of approximating the estimators @xmath22
, we suggest the collection @xmath44 given by @xmath264 we associate to @xmath44 the weight function @xmath53 defined for @xmath45 by @xmath265}+\log(1+d)\ \ { \rm with}\ \
d=\dim(s ) . \label{delta.eq}\ ] ] since @xmath266 assumption [ h1 ] is satisfied with @xmath267 .
let us now turn to the choices of the @xmath268 .
the criterion given by ( [ eq : criterion ] ) can not be computed when @xmath61 for all @xmath23 as soon as @xmath18 is too large .
in such a case , one must consider a smaller subset of @xmath44 and we suggest for @xmath269 @xmath270 ( where the @xmath117 are defined above ) , or preferably @xmath271 whenever this latter family is not too large . note that these two families are random .
our choices of @xmath53 and @xmath57 ensure that @xmath272 for all @xmath87 and that @xmath273 hence , by applying corollary [ maincor ] with @xmath274 , we get the following result .
[ cor - vs1 ] let @xmath79 , @xmath239 and @xmath275 be some positive integer satisfying @xmath276 .
let @xmath277 be a ( finite ) collection of random subsets of @xmath245 with cardinality not larger than @xmath275 based on the observation @xmath1 and @xmath278 a family of estimators @xmath0 , also based on @xmath1 , such that @xmath279 . by applying our selection procedure
, the resulting estimator @xmath90 satisfies @xmath208}\le \inf_{\lambda\in\lambda}{\left[{\e{\left[{{\left\|{f-\widehat f_{\lambda}}\right\|}^{2}}\right]}+\e{\left[{\dim(s_{\widehat m(\lambda)})}\right]}\log(p)\sigma^{2}}\right]},\ ] ] where @xmath50 is a constant depending on the choices of @xmath30 and @xmath74 only . again , note that the risk bound we get is non - increasing with respect to @xmath43 .
this means that if one adds a new variable selection procedure or considers more tuning parameters to increase @xmath43 , the risk bound we get can only be improved . without additional information on the estimators
@xmath22 it is difficult to compare @xmath280}\sigma^{2}$ ] and @xmath281}$ ] .
if @xmath22 is of the form @xmath282 for some deterministic subset @xmath45 it is well - known that @xmath283}={\left\|{f-\pi_{s}f}\right\|}^{2}+\dim(s)\sigma^{2}\ge \dim(s)\sigma^{2}.\ ] ] under the assumption that @xmath284 and that @xmath285 belongs to @xmath286 with probability close enough to 1 , we can compare the risk of the estimator @xmath90 to the cardinality of @xmath287 .
[ cor - vs2 ] assume that the assumptions of corollary [ cor - vs1 ] hold and that @xmath288 for all @xmath87 .
if @xmath284 for some non - void subset @xmath289 with cardinality not larger than @xmath275 , then @xmath208}\le \log(p)|m^*|\sigma^2+r_{n}(m^{*})\ ] ] where @xmath50 is a constant depending on @xmath30 and @xmath74 only , and @xmath290}}\right)}^{1/2}.\ ] ] @xcite gives sufficient conditions on the design @xmath15 to ensure that @xmath291}$ ] is exponentially small with respect to @xmath19 when the family @xmath286 is obtained by using the lars - lasso algorithm with different values of the tuning parameter .
in the linear regression setting described in example [ ex - vs ] , we carry out a simulation study to evaluate the performances of our procedure to solve the two following problems .
we first consider the problem , described in example [ ex - ctp ] , of tuning the smoothing parameter of the lasso procedure for estimating @xmath0 .
the performances of our procedure are compared with those of the @xmath42-fold cross - validation method .
secondly , we consider the problem of variable selection .
we solve it by using our criterion in view of selecting among a family @xmath246 of candidate variable selection procedures .
our simulation study is based on a large number of examples which have been chosen in view of covering a large variety of situations .
most of these have been found in the literature in the context of example [ ex - vs ] either for estimation or variable selection purposes when the number @xmath18 of predictors is large .
the section is organized as follows .
the simulation design is given in the following section .
then , we describe how our procedure is applied for tuning the lasso and performing variable selection . finally , we give the results of the simulation study .
one example is determined by the number of observations @xmath19 , the number of variables @xmath18 , the @xmath292 matrix @xmath15 , the values of the parameters @xmath17 , and the ratio signal / noise @xmath293 .
it is denoted by @xmath294 , and the set of all considered examples is denoted @xmath295 . for each example , we carry out 400 simulations of @xmath1 as a gaussian random vector with expectation @xmath296 and variance @xmath297 , where @xmath298 is the @xmath213 identity matrix , and @xmath299
. the collection @xmath295 is composed of several collections @xmath300 for @xmath301 where each collection @xmath300 is characterized by a vector of parameters @xmath302 , and a set @xmath303 of matrices @xmath15 : @xmath304 where @xmath305 and @xmath306 consists of pairs @xmath307 such that @xmath18 is smaller , equal or greater than @xmath19 .
the examples are described in further details in section [ simex.st ] .
they are inspired by examples found in tibshirani @xcite , zou and hastie @xcite , zou @xcite , and huang et al .
@xcite for comparing the lasso method to the ridge , adaptive lasso and elastic net methods .
they make up a large variety of situations .
they include cases where * the covariates are not , moderately or strongly correlated , * the covariates with zero coefficients are weakly or highly correlated with covariates with non - zero coefficients , * the covariates with non - zero coefficients are grouped and correlated within these groups , * the lasso method is known to be inconsistent , * few or many effects are present .
we consider here the problem of tuning the smoothing parameter of the lasso estimator as described in example [ ex - ctp ] . instead of considering the lasso estimators for a fixed grid @xmath43 of smoothing parameters @xmath23 , we rather focus on the sequence @xmath308 of estimators given by the @xmath275 first steps of the lars - lasso algorithm proposed by efron _ et al . _
hence , the tuning parameter is here the number @xmath309 of steps . in our simulation study
, we compare the performance of our criterion to that of the @xmath42-fold cross - validation for the problem of selecting the best estimator among the collection @xmath310 .
we recall that our selection procedure relies on the choices of families @xmath44 , @xmath311 for @xmath312 , a weight function @xmath53 , a penalty function @xmath54 and two universal constants @xmath79 and @xmath240 .
we choose the family @xmath44 defined by ( [ sm ] ) .
we associate to @xmath313 the family @xmath314 where the @xmath315 are defined in section [ sect - irm ] and @xmath316 is the set of indices corresponding to the predictors retuned by the lars - lasso algorithm at step @xmath317 .
we take @xmath318 with @xmath319 defined by ( [ delta.eq ] ) and @xmath100 .
this value of @xmath30 is consistent with what is suggested in baraud _
_ @xcite .
the choice of @xmath84 is based on the following considerations .
first , choosing @xmath84 around one seems reasonable since it weights similarly the term @xmath320 which measures how well the estimator fits the data and the approximation term @xmath321 involved in our criterion ( [ eq : criterion ] ) .
second , simple calculation shows that the constant @xmath322 involved in theorem [ main ] is minimum for @xmath84 close to 0.6 .
we therefore carried out our simulations for @xmath84 varying from 0.2 to 1.5 .
the results being very similar for @xmath84 between 0.5 and 1.2 , we choose @xmath323 .
we denote by @xmath324 the resulting estimator of @xmath0 . for each @xmath325 ,
the prediction error is estimated using a @xmath42-fold cross - validation procedure , with @xmath326 .
the estimator @xmath327 is chosen by minimizing the estimated prediction error .
the simulations were carried out with r ( www.r-project.org ) using the library elasticnet .
for each example @xmath328 , we estimate on the basis of 400 simulations the oracle risk @xmath329 and the euclidean risks @xmath330 and @xmath331 of @xmath324 and @xmath327 respectively .
the results presented in table [ tb1 ] show that our procedure tends to choose a better estimator than the cv in the sense that the ratios @xmath332 are closer to one than @xmath333 .
.[tb1 ] mean , standard - error and quantiles of the ratios @xmath334 calculated over all @xmath335 such that @xmath336 .
the number of such examples equals 654 , see section [ simex.st ] .
[ cols= " > , > , > , > , > , > , > , > " , ]
throughout this section , we use the following notations . for all @xmath87 and @xmath337 , we write @xmath338 where @xmath339 for
all @xmath87 , let @xmath340 be such that @xmath341 we also write @xmath342 and @xmath343 for the linear space generated by @xmath47 and @xmath0 .
it follows the facts that for all @xmath87 and @xmath337 @xmath344 and simple algebra that @xmath345 for @xmath346 and @xmath45 , let us set @xmath347 if @xmath348 and @xmath349 otherwise . for all @xmath23 and @xmath47
, we have @xmath350 and @xmath351 hence , by using ( [ pen ] ) and ( [ penn ] ) we get @xmath352 where @xmath353 for each @xmath45 , @xmath354 and since the variable @xmath355 is independent of @xmath356 and is stochastically larger than @xmath357 , we deduce from the definition of @xmath358 and ( [ sigma ] ) , that on the one hand @xmath359 . on the other hand , since @xmath47 is arbitrary among @xmath57 and since @xmath360 we deduce from ( [ eq1 ] ) that for all @xmath87 , @xmath361}\ ] ] with @xmath362 and ( [ eq3 ] ) follows by taking the expectation on both sides of ( [ eq10 ] ) .
note that provided that @xmath363}\ ] ] is measurable , we have actually proved the stronger inequality @xmath364}\le \e{\left[{\inf_{\lambda\in\lambda}{\left\{{{\left\|{f-\widehat f_{\lambda}}\right\|}^{2}+a(\widehat f_{\lambda},\ss_{\lambda})}\right\}}}\right]}+\sigma^{2}\sigma + \delta.\ ] ] let us now turn to the second part of the theorem , fixing some @xmath87 .
since equality holds in ( [ pen ] ) , under assumption [ h4 ] by ( [ eq : oom ] ) @xmath365 if @xmath57 is non - random , for some @xmath366 and all @xmath337 , @xmath93}}\\ & \le&\e{\left[{{\left\|{\widehat f_{\lambda}-\pi_{s}\widehat f_{\lambda}}\right\|}^{2}}\right]}+(\dim(s)\vee \delta(s))\e{\left[{\widehat\sigma_{s}^{2}}\right]},\\ & = & \e{\left[{{\left\|{\widehat f_{\lambda}-\pi_{s}\widehat
f_{\lambda}}\right\|}^{2}}\right]}+{\dim(s)\vee \delta(s)\over n-\dim(s)}{\left[{{\left\|{f-\pi_{s}f}\right\|}^{2}+(n-\dim(s))\sigma^{2}}\right]}.\end{aligned}\ ] ] since @xmath367 , we have @xmath368}\le 2\e{\left[{{\left\|{f-\widehat f_{\lambda}}\right\|}^{2}}\right]}+2\e{\left[{{\left\|{\widehat f_{\lambda}-\pi_{s}\widehat f_{\lambda}}\right\|}^{2}}\right]},\ ] ] and under assumption [ h4 ] , @xmath369 , and hence for all @xmath337 @xmath370}&\le & { \left({1+{2\kappa\over 1-\kappa}}\right)}\e{\left[{{\left\|{f-\widehat f_{\lambda}}\right\|}^{2}}\right]}+{2\kappa\over 1-\kappa}\e{\left[{{\left\|{\widehat f_{\lambda}-\pi_{s}\widehat f_{\lambda}}\right\|}^{2}}\right]}\\ & & \ \ + \ \ { \left({\dim(s)\vee \delta(s)}\right)}\sigma^{2}.\end{aligned}\ ] ] which leads to ( [ bornea1 ] ) .
let us turn to the proof of ( [ bornea2 ] ) .
we set @xmath371 . since with probability one @xmath372 , @xmath373}\le \e{\left[{{\mathop{\rm pen}\nolimits}(\widehat s_{\lambda})\widehat \sigma_{\lambda}^{2}}\right]}\ ] ] and it suffices thus to bound the right - hand side . since equality holds in ( [ pen ] ) and
since @xmath95 @xmath374}\\ & \le & 2k{\ed(\widehat s_{\lambda})\over n-\dim(\widehat s_{\lambda})}{\left[{{\left\|{f-\widehat f_{\lambda}}\right\|}^{2}+({\left\|{\eps}\right\|}^{2}-2n\sigma^{2})_{+}+2n\sigma^{2}}\right]}.\end{aligned}\ ] ] under assumption [ h4 ] , @xmath375 and we deduce from ( [ eq : oom ] ) that for some constant @xmath50 depending only on @xmath30 and @xmath74 @xmath376 and the result follows from the fact that @xmath377\le 3\sigma^{2}$ ] for all @xmath19 . for all @xmath87 and @xmath378 , @xmath379 and hence , @xmath380}.\ ] ] besides , since the minimax rate of estimation over @xmath47 is of order @xmath381 , for some universal constant @xmath50 , @xmath382}\ge \dim(s)\sigma^{2}.\ ] ] putting these bounds together lead to the result . under ( [ hm ] )
, it is not difficult to see that @xmath383 so that @xmath44 is not empty and since for all @xmath384 @xmath385}\le |m|(1+\log m)\le { n\over 2},\ ] ] assumptions [ h0 ] to [ h2 ] are satisfied with @xmath183 .
besides , the set @xmath143 being compact , @xmath195 admits a minimum over @xmath143 ( we shall come back the minimization of this criterion at the end of the subsection ) and hence we can take @xmath110 . by applying theorem [ main ] and using ( [ bornea1 ] )
, the resulting estimator @xmath197 satisfies for some universal constant @xmath184 @xmath386}\le \inf_{g\in\ff_{\lambda}}{\left\{{{\left\|{f - g}\right\|}^{2}+\overline a(g,\ss)}\right\}},\ ] ] where @xmath387}.\ ] ] we bound @xmath388 from above by using the following approximation result below the proof of which can be found in makovoz @xcite ( more precisely , we refer to the proof of his theorem 2 ) . for all @xmath389 in the convex hull @xmath138 of the @xmath141 and all @xmath390
, there exists @xmath156 such that @xmath391 and @xmath392 by using this lemma and the fact that @xmath393 for all @xmath394 , we get @xmath395}\sigma^{2}.\ ] ] taking for @xmath396 the integer part of @xmath397 which belongs to @xmath398 $ ] under ( [ hm ] ) , we get @xmath399 for some universal constant @xmath150 which together with ( [ conv1 ] ) leads to the risk bound @xmath198}-c\inf_{g\in\ff_{\lambda}}{\left\|{f - g}\right\|}^{2}\le c\sqrt{nl^{2}\log(em/\sqrt{nl^{2}})}\sigma^{2}.\ ] ] concerning the computation of @xmath400 , note that @xmath401}\\ & = & \inf_{s\in\ss_{\cv}}{\left\{{{\left[{\inf_{\lambda\in\lambda}{\left ( { { \left\|{y-\pi_{s}f_{\lambda}}\right\|}^{2}+\alpha{\left\|{f_{\lambda}-\pi_{s } f_{\lambda}}\right\|}^{2}}\right)}}\right]}+{\mathop{\rm pen}\nolimits}(s)\,\widehat\sigma^2_{s}}\right\}},\end{aligned}\ ] ] and hence , one can solve the problem of minimizing @xmath167 over @xmath87 by proceeding into two steps .
first , for each @xmath47 in the finite set @xmath402 minimize the convex criterion @xmath403 over the convex ( and compact set ) @xmath143 . denote by @xmath404 the resulting minimizers . then , minimize the quantity @xmath405 for @xmath47 varying among @xmath402 .
denoting by @xmath406 such a minimizer , we have that @xmath407 . by applying theorem
[ main ] , we obtain that the selected estimator @xmath90 satisfies @xmath208}\le \inf_{\lambda\in { \left\{{{\rm l},{\rm ms},\cv}\right\ } } } { \left[{\e{\left[{{\left\|{f-\widehat f_{\lambda}}\right\|}^{2}}\right]}+\e{\left[{a(\widehat f_{\lambda},\ss_{\lambda})}\right]}}\right]}.\ ] ] let us now bound @xmath408}$ ] for each @xmath87 . if @xmath409 , by using ( [ bornea1 ] ) and the fact that @xmath410 , we have @xmath411}\le \e{\left[{{\left\|{f-\widehat f_{{\rm l}}}\right\|}^{2}}\right]}+m\sigma^{2}.\ ] ]
if @xmath412 , we may use ( [ bornea2 ] ) since with probability one @xmath413 and since @xmath414 for all @xmath181 , we get @xmath415}\le \e{\left[{{\left\|{f-\widehat f_{{\rm ms}}}\right\|}^{2}}\right]}+\log(m)\sigma^{2}.\ ] ] finally , let us turn to the case @xmath416 and denote by @xmath389 the best approximation of @xmath0 in @xmath201 . since @xmath417 , for all @xmath418 , @xmath419 and hence by using ( [ bornea1 ] ) @xmath420}\le \e{\left[{{\left\|{f-\widehat f_{\cv}}\right\|}^{2}}\right]}+\overline a(g,\ss_{\cv})\ ] ] where @xmath421 is given by ( [ def - abar ] ) . by arguing as in section ( [ sect - ca ] )
, we deduce that under ( [ hm ] ) @xmath420}\le \e{\left[{{\left\|{f-\widehat f_{\cv}}\right\|}^{2}}\right]}+\sqrt{nl^{2}\log(em/\sqrt{nl^{2}})}\sigma^{2}.\ ] ] by putting these bounds together we get the result .
since assumptions [ h0 ] to [ h2 ] are fulfilled and @xmath3 is finite , we may apply theorem [ main ] and take @xmath110 . by using ( [ bornea1 ] ) , we have for some @xmath50 depending on @xmath243 and @xmath74 , @xmath422}}\\ & \leq & \inf_{\lambda\in\lambda } { \left\{{\e{\left[{{\left\|{f-\widehat f_{\lambda}}\right\|}^2}\right]}+\e{\left[{{\left\|{\widehat f_{\lambda}-\pi_{s_{\lambda}}\widehat f_{\lambda}}\right\|}^2}\right]}+a(1+\dim(s_{\lambda}))\sigma^2}\right\}}.\end{aligned}\ ] ] for all @xmath87 , @xmath423}&=&{\left\|{f - a_{\lambda}f}\right\|}^2+\e{\left[{{\left\|{a_{\lambda}\eps}\right\|}^2}\right]}\\ & = & { \left\|{f - a_{\lambda}f}\right\|}^2+{\rm tr}(a_{\lambda}^*a_{\lambda})\sigma^2\\ & \ge & \max{\left\{{{\left\|{f - a_{\lambda}f}\right\|}^2,{\rm tr}(a_{\lambda}^*a_{\lambda})\sigma^2}\right\}}\end{aligned}\ ] ] and @xmath424}&=&{\left\|{(i-\pi_{s_{\lambda}})a_{\lambda}f}\right\|}^2+\e{\left[{{\left\|{(i-\pi_{s_{\lambda}})a_{\lambda}\eps}\right\|}^2}\right]},\\ & \le & 2\max{\left\{{{\left\|{(i-\pi_{s_{\lambda}})a_{\lambda}f}\right\|}^2,\e{\left[{{\left\|{a_{\lambda}\eps}\right\|}^2}\right]}}\right\}}\\ & = & 2\max{\left\{{{\left\|{(i-\pi_{s_{\lambda}})a_{\lambda}f}\right\|}^2,{\rm tr}(a_{\lambda}^{*}a_{\lambda})\sigma^{2}}\right\}}\end{aligned}\ ] ] and hence , corollary [ linear estimators ] follows from the next lemma .
[ lemme lineaire ] for all @xmath87 we have @xmath425 writing @xmath426 and using the fact that @xmath427 and the definition of @xmath226 , we obtain @xmath428 where @xmath429 are the singular values of @xmath227 counted with their multiplicity and @xmath430 is an orthonormal family of right - singular vectors associated to @xmath431 .
if @xmath432 , then @xmath433 and we have @xmath434
. otherwise , @xmath435 , we may consider @xmath436 as the largest @xmath232 such that @xmath437 and derive that @xmath438 which proves the assertion @xmath439 .
for the bound @xmath440 , we set @xmath441 and note that @xmath442 induces a semi - positive quadratic form on @xmath222 . as a consequence the quadratic form @xmath443 is dominated by the quadratic form @xmath444 on @xmath222 .
furthermore @xmath445 where @xmath446 is the inverse of the linear operator @xmath447 induced by @xmath448 restricted on @xmath223 .
we then have that the quadratic form induced by @xmath446 is dominated by the quadratic form @xmath449 on @xmath222 .
in particular the sequence of the eigenvalues of @xmath446 is dominated by the sequence @xmath450 so @xmath451 which conclude the proof of lemma [ lemme lineaire ] . along the section , we write @xmath452 for @xmath453 and @xmath113 for @xmath454 for short . by using ( [ eqbase ] ) with @xmath110 and since @xmath455 , we have @xmath456}\le \e{\left[{\inf_{\lambda\in\lambda}\|f-\pi_{\widehat s_{\lambda}}y\|^2+{\mathop{\rm pen}\nolimits}(\widehat s_{\lambda})\widehat \sigma^2_{\widehat s_{\lambda}}}\right]}+(1+\log(p+1))\sigma^2,\ ] ] for some constant @xmath184 depending on @xmath30 only . writing @xmath457 for the event @xmath458
, we have @xmath459}\le a_{n}+r'_{n}\ ] ] where @xmath460}\\ r'_{n}&= & \e{\left[{\inf_{\lambda\in\lambda}{\left\{{\|f-\pi_{\widehat s_{\lambda}}y\|^2+{\mathop{\rm pen}\nolimits}(\widehat s_{\lambda})\widehat \sigma^2_{\widehat s_{\lambda}}}\right\}}{\bf 1}_{b}}\right]}.\end{aligned}\ ] ] let us bound @xmath461 from above .
note that @xmath462 and @xmath463 and since @xmath464 , by using ( [ eq : oom ] ) we get @xmath465 for some constant @xmath150 depending on @xmath30 and @xmath74 only .
let us now turn to @xmath466 . for all @xmath87 , @xmath467 and @xmath468 since for all @xmath45 , @xmath469 , by using ( [ eq : oom ] )
again , there exists some positive constant @xmath470 depending on @xmath30 and @xmath74 only such that for all @xmath87 , @xmath471 and hence , @xmath472 some calculation shows that @xmath473}\le{\left({{\left\|{f}\right\|}^{2}+2n\sigma^{2}}\right)}^{2}$ ] and hence , by cauchy - schwarz inequality @xmath474 the result follows by putting the bounds on @xmath461 and @xmath466 together .
arlot , s. ( 2007 ) . .
phd thesis , university paris xi .
arlot , s. ( 2009 ) .
model selection by resampling penalization .
, 3:557624 .
arlot , s. and bach , f. ( 2009 ) .
data - driven calibration of linear estimators with minimal penalties . , 22:4654 .
arlot , s. and celisse , a. ( 2010 ) . a survey of cross - validation procedures for model selection .
, 4:4079 .
baraud , y. ( 2000 ) .
model selection for regression on a fixed design .
, 117(4):467493 .
baraud , y. ( 2010 ) .
estimator selection with respect to hellinger - type risks . .
baraud , y. , giraud , c. , and huet , s. ( 2009 ) .
gaussian model selection with an unknown variance .
, 37(2):630672 .
birg , l. ( 2006 ) .
model selection via testing : an alternative to ( penalized ) maximum likelihood estimators .
, 42(3):273325 .
birg , l. and massart , p. ( 2001 ) .
gaussian model selection .
, 3(3):203268 .
boulesteix , a. and strimmer , k. ( 2006 ) . .
, 8(1):3244 .
breiman , l. ( 2001 ) .
random forests . , 45:532 .
bunea , f. , tsybakov , a. b. , and wegkamp , m. h. ( 2007 ) .
aggregation for gaussian regression .
, 35(4):16741697 .
cands , e. and tao , t. ( 2007 ) .
the dantzig selector : statistical estimation when @xmath18 is much larger than @xmath19 .
, 35(6):23132351 .
cao , y. and golubev , y. ( 2006 ) . on oracle inequalities related to smoothing splines .
, 15(4):398414 ( 2007 ) .
catoni , o. ( 1997 ) . mixture approach to universal model selection . technical report , ecole normale suprieure , france .
catoni , o. ( 2004 ) . statistical learning theory and stochastic optimization . in _
lecture notes from the 31st summer school on probability theory held in saint - flour , july 825 , 2001_. springer - verlag , berlin .
celisse , a. ( 2008 ) . .
phd thesis , university paris xi .
chen , s. s. , donoho , d. l. , and saunders , m. a. ( 1998 ) .
atomic decomposition by basis pursuit .
, 20(1):3361 ( electronic ) .
daz - uriarte , r. and alvares de andrs , s. ( 2006 ) . .
, 7(3 ) .
efron , b. , hastie , t. , johnstone , i. , and tibshirani , r. ( 2004 ) .
least angle regression .
, 32(2):407499 . with discussion , and a rejoinder by the authors .
genuer , r. , poggi , j .-
m . , and tuleau - malot , c. ( 2010 ) .
variable selection using random forests .
, to appear .
giraud , c. ( 2008 ) .
mixing least - squares estimators when the variance is unknown .
, 14(4):10891107 .
goldenshluger , a. ( 2009 ) . a universal procedure for aggregating estimators .
, 37(1):542568 .
goldenshluger , a. and lepski , o. ( 2009 ) .
structural adaptation via @xmath475-norm oracle inequalities .
, 143(1 - 2):4171 .
helland , i. ( 2006 ) .
partial least squares regression . in kotz ,
s. , balakrishnan , n. , read , c. , vidakovic , b. , and johnston , n. , editors , _ encyclopedia of statistical sciences ( 2nd ed . )
_ , volume 9 , pages 59575962 , new york .
helland , i. ( 2001 ) .
some theoretical aspects of partial least squares regression . , 58:97107 .
hoerl , a. and kennard , r. ( 1970 ) .
ridge regression : bayes estimation for nonorthogonal problems .
, 12:5567 .
hoerl , a. and kennard , r. ( 2006 ) . ridge regression . in kotz , s. , balakrishnan , n. ,
read , c. , vidakovic , b. , and johnston , n. , editors , _ encyclopedia of statistical sciences ( 2nd ed . )
_ , volume 11 , pages 72737280 , new york .
huang , j. , ma , s. , and zhang , c .- h .
( 2008 ) . , 4(1603 - 1618 ) .
juditsky , a. and nemirovski , a. ( 2000 ) .
functional aggregation for nonparametric regression .
, 28(3):681712 .
lepski , o. v. ( 1990 ) . a problem of adaptive estimation in gaussian white noise .
, 35(3):459470 .
lepski , o. v. ( 1991 ) .
asymptotically minimax adaptive estimation . i.
upper bounds . optimally adaptive estimates . , 36(4):645659 .
lepski , o. v. ( 1992a ) .
asymptotically minimax adaptive estimation .
schemes without optimal adaptation . adaptive estimates . , 37(3):468481 .
lepski , o. v. ( 1992b ) . on problems of adaptive estimation in white gaussian noise . in _ topics in nonparametric
estimation _ , volume 12 of _ adv .
soviet math .
_ , pages 87106 .
soc . , providence , ri .
leung , g. and barron , a. r. ( 2006 ) .
information theory and mixing least - squares regressions .
, 52(8):33963410 .
makovoz , y. ( 1996 ) .
random approximants and neural networks . , 85(1):98109 .
nadaraya , e. a. ( 1964 ) . on estimating regression .
, 9(1):141142 .
nemirovski , a. ( 2000 ) .
topics in non - parametric statistics . in _
lectures on probability theory and statistics ( saint - flour , 1998 ) _ , volume 1738 of _ lecture notes in math .
_ , pages 85277 .
springer , berlin .
rigollet , p. and tsybakov , a. b. ( 2007 ) .
linear and convex aggregation of density estimators .
, 16(3):260280 .
strobl , c. , boulesteix , a .-
l . , kneib , t. , augustin , t. , and zeileis , a. ( 2008 ) .
conditional variable importance for random forests . , 9(307 ) .
strobl , c. , boulesteix , a .-
, zeileis , a. , and hothorn , t. ( 2007 ) .
bias in random forest variable importance measures : illustrations , sources and a solution . , 8(25 ) .
tenenhaus , m. ( 1998 ) . .
ditions technip , paris .
thorie et pratique .
[ theory and application ] .
tibshirani , r. ( 1996 ) .
regression shrinkage and selection via the lasso .
, 58(1):267288 .
tsybakov , a. b. ( 2003 ) .
optimal rates of aggregation . in _ proceedings of the 16th annual conference on learning theory ( colt ) and 7th annual workshop on kernel machines _ , pages 303313 .
lecture notes in artificial intelligence 2777 , springer - verlag , berlin .
watson , g. s. ( 1964 ) .
smooth regression analysis .
, 26:359372 .
wegkamp , m. ( 2003 ) .
model selection in nonparametric regression .
, 31:252273 .
yang , y. ( 1999 ) .
model selection for nonparametric regression .
, 9:475499 .
yang , y. ( 2000a ) . combining different procedures for adaptive regression .
, 74(1):135161 .
yang , y. ( 2000b ) . mixing strategies for density estimation . , 28(1):7587 .
yang , y. ( 2001 ) . adaptive regression by mixing .
, 96(454):574588 .
zhang , t. ( 2005 ) .
learning bounds for kernel regression using effective data dimensionality .
, 17(9):20772098 .
zhang , t. ( 2008 ) .
adaptive forward - backward greedy algorithm for learning sparse representations .
technical report , rutgers university , nj . zhao , p. and yu , b. ( 2006 ) . on model selection consistency of lasso .
7(nov):25412563 .
zou , h. ( 2006 ) .
the adaptive lasso and its oracle properties . , 101(476):14181429 .
zou , h. and hastie , t. ( 2005 ) .
regularization and variable selection via the elastic net . , 67(2):301320 .
the penalty @xmath476 , defined at equation ( [ edk.eq ] ) , is linked to the edkhi function introduced in baraud _ al _ @xcite ( see definition 3 ) , via the following formula : @xmath477 therefore , according to the result given in section 6.1 in baraud _
@xcite , @xmath476 is the solution in @xmath478 of the equation @xmath479 the collection @xmath295 is composed of several collections @xmath480 that are detailed below .
the collections @xmath481 to @xmath482 are composed of examples where @xmath15 is generated as @xmath19 independent centered gaussian vectors with covariance matrix @xmath50 . for each @xmath483 , we define a @xmath484 matrix @xmath485 and a @xmath18-vector of parameters @xmath302 .
we denote by @xmath303 the set of 5 matrices @xmath15 simulated as @xmath19-i.i.d @xmath486 .
the collection @xmath300 is then defined as follows : @xmath487 where @xmath305 and @xmath488 in section [ cvedkhi.st ] , and @xmath489 in section [ selvar.st ] .
the matrix @xmath50 equals the @xmath484 identity matrix denoted @xmath490 .
the parameters @xmath17 satisfy @xmath491 for @xmath492 , @xmath493 for @xmath494 , @xmath495 for @xmath496 , @xmath497 for @xmath498 .
the matrix @xmath50 satisfies @xmath513 for @xmath514 , @xmath515 for @xmath516 , @xmath517 for @xmath518 , with @xmath519 and @xmath520 .
the parameters @xmath17 satisfy @xmath491 for @xmath521 , @xmath522 for @xmath523 . in this last example , we denote by @xmath545 the set of 5 matrices @xmath15 simulated as follows . for @xmath546 , we denote by @xmath547 the column @xmath548 of @xmath15 .
let @xmath549 be generated as @xmath19 i.i.d .
@xmath550 and let @xmath551 be generated as @xmath19 i.i.d . @xmath552 .
then for @xmath553 , @xmath554 , for @xmath555 , @xmath556 , for @xmath557 , @xmath558 , for @xmath492 , @xmath559 .
the parameters @xmath17 are as in collection @xmath510 . the collection @xmath544 is defined as the set of examples @xmath560 for @xmath561 , @xmath562 , and @xmath563 . the collection @xmath295 is thus composed of 660 examples for @xmath306 chosen as in ( [ i3.eq ] ) , and 825 for @xmath306 chosen as in ( [ i2.eq ] ) . for some of the examples ,
the lasso estimators were highly biased leading to high values of the ratio @xmath564 , see equation ( [ oracl.eq ] ) .
we only keep the examples for which the lasso estimator improves the risk of the naive estimator @xmath1 by a factor at least @xmath565 .
this convention leads us to remove 171 examples over 825 .
these pathological examples are coming from the collections @xmath481 , @xmath510 and @xmath512 for @xmath566 and @xmath567 , and from collections @xmath499 and @xmath507 when @xmath568 .
the examples of collection @xmath512 were chosen by zou to illustrate that the lasso estimators may be highly biased .
all the other examples , correspond to matrices @xmath15 that are nearly orthogonal . _ the lasso procedure _ is described in section [ cvedkhi.st ] .
the collection @xmath571 where @xmath572 is the set of indices corresponding to the predictors returned by the lars - lasso algorithm at step @xmath573 ( see section [ cvedkhi.st ] ) . _
the ridge procedure _ is based on the minimization of @xmath574 with respect to @xmath17 , for some positive @xmath40 , see for example hoerl and kennard @xcite .
tibshirani @xcite noted that in the case of a large number of small effects , ridge regression gives better results than the lasso for variable selection . for each @xmath575 ,
the regression coefficients @xmath576 are calculated and a collection of predictors sets is built as follows .
let @xmath577 be such that @xmath578 then , the collection @xmath579 is defined as @xmath580 .
_ the elastic net procedure _ proposed by zou and hastie @xcite mixes the @xmath581 and @xmath582 penalties of the lasso and the ridge procedures .
let @xmath583 be a grid of values for the tuning parameter @xmath40 of the @xmath582 penalty .
we choose @xmath584 where @xmath585 denotes the collection of the active sets of cardinality less than @xmath275 , selected by the elastic net procedure when the @xmath582-smoothing parameter equals @xmath40 . for each @xmath575
the collection @xmath585 can be conveniently computed by first calculating the ridge regression coefficients and then applying the lars - lasso algorithm , see zou and hastie @xcite . _ the partial least squares regression _
( plsr1 ) aims to reduce the dimensionality of the regression problem by calculating a small number of components that are usefull for predicting @xmath1 .
several applications of this procedure for analysing high - dimensional genomic data have been reviewed by boulesteix and strimmer @xcite .
in particular , it can be used for calculating subsets of covariates as we did for the ridge procedure .
the plsr1 procedure constructs , for a given @xmath40 , uncorrelated latent components @xmath586 that are highly correlated with the response @xmath1 , see helland @xcite .
let @xmath587 be a grid a values for the tuning parameter @xmath40 .
for each @xmath588 , we write @xmath589 for the pls regression coefficients calculated with the first @xmath40 components .
we then set @xmath590 , where @xmath591 is build from @xmath589 as for the ridge procedure . _
the adaptive lasso procedure _ proposed by zou @xcite starts with a preliminary estimator @xmath592 .
then one applies the lasso procedure replacing the parameters @xmath593 in the @xmath581 penalty by the weighted parameters @xmath594 for some positive @xmath595 .
the idea is to increase the penalty for coefficients that are close to zero , reducing thus the bias in the estimation of @xmath0 and improving the variable selection accuracy .
zou showed that , if @xmath592 is a @xmath596-consistent estimator of @xmath17 , then the adaptive lasso procedure is consistent in situations where the lasso is not .
a lot of work has been done around this subject , see huang et al .
@xcite for example . - using the ridge estimator , @xmath598 as preliminary estimator . for each @xmath599
, the adaptive lasso procedure is applied for calculating the active sets , @xmath600 , of cardinality less than @xmath275 .
the collection @xmath601 is thus defined as @xmath602 . _
the random forest algorithm _ was proposed by breiman @xcite for classification and regression problems .
the procedure averages several regression trees calculated on bootstrap samples .
the algorithm returns measures of variable importance that may be used for variable selection , see for example daz - uriarte and alvares de andrs @xcite , genuer et al .
@xcite , strobl et al .
@xcite .
let us denote by @xmath40 the number of variables randomly chosen at each split when constructing the trees and @xmath605 for each @xmath606 , we consider the set of indices @xmath607 where @xmath608 are the ranks of the variable importance measures .
two importance measures are proposed . the first one is based on the decrease in the mean square error of prediction after permutation of each of the variables .
it leads to the collection @xmath609 .
the second one is based on the decrease in node impurities , and leads similarly to the collection @xmath610 .
we have to choose @xmath275 , the largest number of predictors considered in the collection @xmath613 . for all methods , except the exhaustive method
, @xmath275 may be large , say @xmath614 . nevertheless , for saving computing time , we chose @xmath275 large enough such that the dimension of the estimated subset is always smaller than @xmath275 .
for the exhaustive method , @xmath275 must be chosen in order to make the calculation feasible : @xmath615 for @xmath616 , @xmath617 for @xmath618 and @xmath619 for @xmath620 . | we consider the problem of estimating the mean @xmath0 of a gaussian vector @xmath1 with independent components of common unknown variance @xmath2 .
our estimation procedure is based on estimator selection . more precisely , we start with an arbitrary and possibly infinite collection @xmath3 of estimators of @xmath0 based on @xmath1 and , with the same data @xmath1 , aim at selecting an estimator among @xmath3 with the smallest euclidean risk .
no assumptions on the estimators are made and their dependencies with respect to @xmath1 may be unknown .
we establish a non - asymptotic risk bound for the selected estimator .
as particular cases , our approach allows to handle the problems of aggregation and model selection as well as those of choosing a window and a kernel for estimating a regression function , or tuning the parameter involved in a penalized criterion .
we also derive oracle - type inequalities when @xmath3 consists of linear estimators . for illustration
, we carry out two simulation studies .
one aims at comparing our procedure to cross - validation for choosing a tuning parameter .
the other shows how to implement our approach to solve the problem of variable selection in practice . , |
the interstellar medium ( ism ) of the milky way ( mw ) and other galaxies is a complex mix of gas and dust .
the processes involved in maintaining the mass , energy , and ionization balance of the ism are not properly understood .
high resolution ultraviolet ( _ uv _ ) spectra provides information about these processes as many absorption lines of atoms , ions and molecules are present in the _ uv _ band of the electromagnetic spectrum , one of the most important of which is o@xmath2 ( o vi ) , a diagnostic of temperatures of about @xmath3 k @xcite .
such temperatures are found at the interface of hot ( t @xmath4 k ) and warm ( t @xmath5 k ) ionized gas in the ism .
thus , o vi absorption lines at 1031.9 and 1037.6 are crucial diagnostics of the energetic processes of interface environments in the ism of galaxies
. the gas at such temperatures is cooling radiatively and the cooling is essentially independent of density , metallicity and the heating mechanism @xcite .
o vi formation by photo - ionization is unable to explain the observed abundances , given the energy of photons needed to get such high ionization ( 114 ev ) .
o vi is mostly produced by shock heating and is collisionally ionized @xcite .
previous studies of o vi have been limited to observations by the _ copernicus _ satellite @xcite and the _ hopkins ultraviolet telescope _ @xcite .
@xcite concluded that the hot gas exists in discrete regions rather than being continuously present in the ism .
the launch of _ far ultraviolet spectroscopic explorer ( fuse ) _ ; @xcite enabled a wider and more descriptive study of o vi absorption and emission in the ism and the intergalactic medium ( igm ) . with a spectral resolution of 20,000 , _
fuse _ has been able to resolve fine details of o vi in many different environments .
_ fuse _ observed o vi absorption lines in the local ism of the mw @xcite , disk of the mw @xcite , halo of the mw @xcite , the large magellanic cloud ( lmc ; @xcite ) , the small magellanic cloud ( smc ; @xcite ) , starburst galaxies @xcite , igm @xcite , etc . apart from absorption studies , _ fuse _ has also recorded o vi spectra in emission from observations of diffuse ism in the mw @xcite and superbubbles in the lmc @xcite . these studies have not only augmented our knowledge about the formation and distribution of o vi in the mw and the magellanic clouds but have also helped in better understanding of the complexities of the ism .
owing to the contiguity to the mw ( @xmath6 50 kpc ; @xcite ) and being nearly a face - on galaxy with a low inclination angle ( @xmath6 35@xmath7 ; @xcite ) , the lmc has been the subject of numerous studies to understand and interpret the properties of ism .
recent studies of the ism in the lmc include observations of diffuse uv emission @xcite , the _ spitzer _ infrared dust survey ( sage ; @xcite ) , the hi survey @xcite , the h@xmath1 survey @xcite , the o vi distribution @xcite , and the survey of hot gas in the x - ray bands @xcite .
more recently , @xcite analyzed the absorption of o vi , c iv , si iv and n v ions towards four early type stars in the lmc and provide crucial details about the environments that may be probed through the study of these ions .
@xcite surveyed the distribution and kinematics of o vi towards 12 early type stars in the lmc , which was very selective and the targets were restricted to wolf - rayet stars and o - type stars of spectral types o7 and earlier . here
we report an extensive survey of o vi absorption at 1032 in the lmc . the observation log is given in table [ tab1 ] .
the results presented in this survey become important as they not only provide a detailed mapping of the interface gas traced by o vi but also throw light on the small scale structure of the o vi distribution in various regions of the lmc .
[ fig1 ] shows the lines of sight and different regions of the lmc including the superbubbles ( marked by circles ) that have been studied in this survey .
the _ fuse _ mission and its operations are described by @xcite and @xcite .
observations are made through one of 3 apertures : the hirs ( @xmath8 @xmath9 @xmath10 ) ; the mdrs ( @xmath11 @xmath9 @xmath10 ) ; and the lwrs ( @xmath12 @xmath9 @xmath12 ) , with all three obtaining data simultaneously . depending on the coating of the spectrograph , observations are possible through sic and lif channels that are further divided into eight different segments ; the sic 1a , sic 2a , sic 1b and sic 2b segments covering the wavelength range 905 to 1105 and lif 1a , lif 2a , lif 1b and lif 2b segments covering the wavelength range 1000 to 1187 .
the data from sic 2b segment have been known to suffer from a fixed noise .
the sensitivity of lif 1a segment near 1032 is almost double that of other segments and therefore , we have used only the lif 1a observations @xcite .
most of the data are from the large aperture lwrs with a few from the mdrs aperture .
the fully calibrated _ fuse _ spectra were downloaded from the multimission archive at stsci ( mast ) processed by the latest _ fuse _ data reduction pipeline ( calfuse version 3.2 ; @xcite ) .
_ fuse _ made more than 600 pointings in and around the lmc .
several of the observations were rejected initially by just looking at the spectra ( with non - existent or low signal - to - noise of o vi absorption ) .
70 unique _ fuse _ targets were selected based mainly on the simplicity of the continuum fitting in the vicinity of o vi ( at 1031.9 ) .
o vi absorption for 1 of the sightline has been reported by @xcite , 11 have been covered by @xcite survey , 3 by @xcite and 1 by @xcite .
spectral types and other stellar information was taken from @xcite and @xcite .
we have downgraded all the spectra reported here to 35 km s@xmath13 to have a higher signal - to - noise .
this has been done for all the spectra irrespective of the quality to maintain uniformity in the data analysis procedure .
fitting the stellar continuum in the neighbourhood of the o vi absorption profile has been discussed in detail by @xcite and @xcite .
@xcite have limited their study to early type stars based on the fact that these stars have completely developed o vi p cygni profiles that are easier to fit as the early type stars have a high mass loss rate and have minimal wind variations . @xcite
have shown that for galactic sight lines , the stellar wind variability has negligible effect on o vi absorption but may introduce substantial errors towards targets in the magellanic clouds .
@xcite did find variation in the equivalent width and column density towards a lmc star when estimated at two different times .
this warrants for extra care while fitting the continuum for magellanic cloud targets . as discussed above , for the lmc , the estimation of stellar continuum in the vicinity of o vi absorption is not trivial .
our background targets are mostly early o and b type stars with several wolf - rayets .
these have been selected based on the fact that the the continuum near the o vi absorption at 1031.9 is simple to fit .
a few of the lines of sight do show a complex behaviour near o vi absorption ( for e.g. , sk-67d05 , sk-67d168 , sk-70d115 , etc . ) .
for such exceptional lines of sight , the complexity in the continuum fitting is due to a local dip or a sudden rise near the o vi absorption and these targets needed a comparatively higher order polynomial for the fitting . for a flavour of the complexity involved ,
sample continuum fitting for three lines of sight , bi13 , sk-70d115 and sk-67d168 are shown in fig .
[ fig2 ] . following the fitting procedure of @xcite and @xcite ,
the local stellar continuum was estimated for all the targets and were fitted by a legendre polynomial fit of low order ( @xmath14 ) .
several continua were tested and the uncertainties involved for the complete data set were used in the measurement of the o vi column densities .
the normalized spectra in the vicinity of o vi absorption are presented in fig .
[ fig3 ] .
the o vi absorption may be contaminated by absorption from molecular hydrogen , however , it is minimal in the lmc velocity range .
this is because the molecular fraction of h@xmath15 in the lmc is only about 12% of the molecular fraction in the galactic disk @xcite .
the closest absorption line of h@xmath15 to o vi absorption in the lmc lies at 1032.35 ( at v@xmath16 + 123 km s@xmath13 ) , which is due to ( 60 ) r(4 ) transition .
an estimation of possible contamination by h@xmath15 towards 12 lines of sight in the lmc has been done by @xcite where they find that h@xmath15 absorption does not affect the lmc o vi column densities significantly .
the h@xmath15 absorption feature are usually not a problem for measurements of o vi absorption at velocities more than 20 km s@xmath13 apart ( @xcite and references therein ) .
we estimate the contamination from h@xmath15 lines by fitting the ( 60 ) p(3 ) h@xmath15 transition for the mw and ( 60 ) r(4 ) h@xmath15 transition for the lmc ( fig .
[ fig4 ] ) . from this , we are able to estimate the contamination due to ( 60 ) r(4 ) mw h@xmath15 line and
( 60 ) p(3 ) lmc h@xmath15 line .
noticeable contamination is seen for the sightlines sk-71d45 and sk-67d250 which is well within the error bar of the derived column densities .
an example of a sightline is also shown in fig
. [ fig4 ] in which the lmc o vi absorption is free from any contamination from molecular hydrogen .
owing to the negligible contribution , we have not excluded the contamination by h@xmath15 in the o vi measurements for the lmc .
another possible contaminant is the ( 60 ) r(0 ) transition of the hd molecule at 1031.91 that overlaps with the mw o vi absorption .
several lines of hd molecule is covered by _ fuse _ ( see @xcite for a complete list ) .
column densities are significantly low in the lmc and contamination due to molecular hydrogen lines is non - significant for o vi measurements in the lmc .
absorption due to hd is about 4 orders weaker compared to h@xmath15 transitions .
therefore , we neglect any contribution from hd molecule to the column densities of o vi in the lmc .
the measurement of equivalent widths and column densities of o vi absorption for all the lines of sight were done following @xcite and @xcite .
this apparent optical depth technique @xcite is now commonly used in the analysis of interstellar absorption lines and is applicable to cases with non - saturated absorption profiles .
briefly the technique uses an apparent optical depth in terms of velocity , i.e. , an instrumentally blurred version of the true optical depth , given as @xmath17,\ ] ] where @xmath18 is the estimated continuum intensity and @xmath19 is the intensity of the absorption line as a function of velocity .
if the resolution of the instrument is very high compared to the fwhm of the absorption line , the apparent optical depth is a very good representation of the true optical depth .
the apparent column density ( @xmath20 [ atoms @xmath0 ( km s@xmath13)@xmath13 ] ) is calculated by the following relation @xmath22 where @xmath23 is the wavelength ( in ) and _ f _ is the oscillator strength of the atomic species ( for o vi , _ f _ value of 0.1325 has been adopted from @xcite ) .
similar to @xcite , we find that the 1032 o vi profile is broad and is fully resolved by _
fuse_. however , the weaker absorption of the o vi doublet at 1037.6 is found to be inseparable from the cii * and h@xmath15 absorption .
the details of the apparent optical depth measurements are listed in table [ tab2 ] .
the overlap of the o vi absorption of the mw and the lmc does not allow a precise measurement of column densities for the lmc .
for the lines of sight ( sk-65d21 , sk-67d69 , sk-67d105 , hv2543 , sk-66d100 , hv5936 , sk-67d211 , sk-69d220 , sk-66d172 , sk-68d137 , d301 - 1005 , and d301-nw8 ) where the lmc o vi is distinct from the mw o vi , the limits of velocities over which integration was performed were easy to obtain . taking cue from these lines of sight , we estimate the lower velocity limit for the lines of sight where the lmc o vi was not separated from the mw o vi . the errors in the equivalent widths and column densities are 1@xmath24 uncertainties derived using the uncertainty in the _ fuse _ data and the fitting procedures . the principal contributor to these errors is the ambiguity in the continuum fitting of several of the spectra .
the lmc spectra were selected based on the quality of the spectra around o vi spectral feature .
the survey presented here has more - or - less complete coverage of the well known regions of the lmc such as 30 doradus , n11 , lmc 4 , etc . , and thus , is very useful in studying the variation of o vi at small scales as well .
we find a significant amount of variation in the o vi column densities in a single region and in different regions of the lmc .
our data cover a wide range of targets ; o type stars , b type stars , wolf - rayet objects , supernova remnants , etc .
the complete list of the background targets is provided in table [ tab1 ] .
these observations give fine details of the small scale structure of o vi column density in the lmc probing to a scale of @xmath6 10 pc in some regions .
we find the o vi absorption in the lmc to be very patchy as is true for the mw .
the o vi column density varies from the lowest value of 5.28 @xmath25 atoms @xmath0 near a hii region ( dem 7 ) to the highest value of 3.74 @xmath26 atoms @xmath0 just north of the lmc bar .
@xcite report the distribution of o vi in the lmc to be patchy as well with the log n(o vi ) value in the lmc varying from 13.9 to 14.6 atoms @xmath0 and a mean of 14.37 atoms @xmath0 .
their data did not provide any small scale variation in the column density of o vi with the smallest scale probed is about 450500 pc .
given the coverage of our data , we have studied the variation in o vi absorption in different regions of the lmc .
we find that the 30 doradus and sn 1987 a regions dominate in terms of the abundance of o vi .
the mean and median values of log n(o vi ) for targets in and around 30 doradus and snr 1987a are 14.30 and 14.27 atoms @xmath0 respectively .
30 doradus is the largest hii region of lmc with a dense concentration of early type massive stars and has thus , attracted extensive research in different wavelength bands @xcite .
we discuss in detail about the o vi distribution and its properties in 30 doradus region in section 5 .
n11 region , which is associated with several ob associations and has a superbubble at its center , shows a high value of o vi column density .
the mean of log n(o vi ) in the n11 region is 14.21 atoms @xmath0 .
another interesting region is the lmc 4 supergiant shell that includes the shapley constellation iii , which is one of the largest region associated with star formation @xcite .
the log n(o vi ) value in lmc 4 supershell varies from a minimum of 13.86 atoms @xmath0 to a maximum of 14.45 atoms @xmath0 with a mean value of 14.20 atoms @xmath0 and a median value of 14.25 atoms @xmath0 .
other regions of the lmc also show patchiness in the o vi distribution .
statistically , the mean of the o vi column density in the lmc is 1.88 @xmath26 atoms @xmath0 which is lower than 2.34 @xmath26 atoms @xmath0 given by howk et al .
( 2002 ) for 12 lines of sight .
the difference is most likely due to the wide coverage of background targets in our data .
the median value of the o vi column density is 1.66 @xmath26 atoms @xmath0 which also is less than the value reported by @xcite .
overall we find that there is an ubiquitous presence of o vi throughout the lmc .
the kinematics of o vi in the lmc is difficult to study because of the ambiguity in separating the lmc o vi absorption from the mw absorption .
for some of the _ fuse _ targets , i.e. , for sk-65d21 , sk-67d69 , sk-67d105 , hv2543 , sk-66d100 , hv5936 , sk-67d211 , sk-69d220 , sk-66d172 , sk-68d137 , d301 - 1005 , and d301-nw8 , o vi absorption at lmc velocities are distinct from the mw ( fig .
[ fig3 ] ) .
linewidth for the lmc o vi profiles are obtained with relatively less error for these lines of sight .
the lmc o vi absorption profiles have all been fitted with a single gaussian .
the corresponding fwhm for these lines of sight are 55 , 85 , 78 , 90 , 90 , 113 , 100 , 85 , 105 , 94 , 107 , and 91 km s@xmath13 respectively .
the temperature range estimated from these widths is t @xmath6 1 @xmath9 10@xmath27 5 @xmath9 10@xmath27 k. o vi abundance is maximum at a temperature of 3 @xmath9 10@xmath28 k , thus , higher fwhm of these profiles is probably due to other broadening mechanisms such as more than one velocity component and/or collision and turbulence .
this also represents the kinematic flow structure of o vi in the lmc .
the linewidth for the mw o vi profiles are narrower than the lmc profiles @xcite . for the galactic halo ,
@xcite report a range for @xmath24 ( linewidth ) from 16 to 65 km s@xmath13 ( corresponding fwhm range is 38153 km s@xmath13 ) .
@xcite , for the local ism , report average @xmath24 to be 16 km s@xmath13 ( fwhm = 38 km s@xmath13 ) while @xcite report @xmath24 values ranging from 15 km s@xmath13 to 36 km s@xmath13 ( fwhm ranging from 35 km s@xmath13 to 85 km s@xmath13 ) .
the linewidth of the smc o vi absorption profile are comparable to that of the lmc . the fwhm range for the smc o vi absorption is from 82 to 115 km s@xmath13 with a mean of 94 km s@xmath13 @xcite .
we obtained a mean fwhm value of 91 km s@xmath13 for the lmc selected lines of sight . @xcite
have compared o vi absorption profiles of the lmc with fe ii absorption and find that the fe ii profiles are much narrower suggesting that the thermal broadening effect for o vi absorption is much more significant .
the mw and the smc offer a different ism environment compared to the lmc especially due to the difference in the metallicity .
the absorption profiles of the mw o vi are different from the lmc o vi and sometimes these profiles are difficult to separate in an unambiguous manner , thus , a comparison of the kinematics is not possible . @xcite
have compared the o vi absorption with fe ii absorption at 1125.45 ( fig .
9 in @xcite ) .
while , the fe ii profiles for the mw and the lmc are clearly separated from each other , the o vi absorption suffers an overlap .
this is due to the difference in width of the two absorption .
the overlap of the mw o vi absorption profile with the lmc o vi absorption constrained @xcite to arrive at a definite conclusion about the existence of outflows from the lmc . following the discussion about the distribution of o vi in the lmc in the previous section
, we now present an overview of o vi in the mw and the smc .
o vi in the mw has been extensively studied since the launch of _ fuse _ @xcite .
@xcite were the first to study the o vi absorption in the disk and halo of the mw as seen towards 11 extragalactic objects ( active galactic nuclei ) using _ fuse _ , confirming the large scale presence of hot gas in the halo @xcite .
the authors find that log n@xmath29(o vi ) varies from 13.80 to 14.64 atoms @xmath0 and the distribution of o vi is quite patchy . to compare with the projected o vi column density on the plane of the mw ,
we calculated the projection of o vi column on to the plane of the lmc . taking the inclination angle of the lmc to be 33@xmath7 , we find the mean value of log n@xmath29(o vi ) @xmath30 log n(o vi ) cos @xmath31 = 14.16 atoms @xmath0 .
@xcite quote a mean value of 14.29 atoms @xmath0 for their sample .
the median value of our sample is 14.14 atoms @xmath0 , while for the @xcite sample , it is 14.21 atoms @xmath0 . @xcite report _ fuse _ observations of o vi absorption towards 100 extragalactic lines of sight to study the properties and distribution of o vi in the galactic halo .
the average log n ( o vi ) for the complete sample is 14.36 atoms @xmath0 while log n ( o vi ) sin @xmath32 value for the complete sample is 14.21 atoms @xmath0 .
the results reveal that there are substantial differences in the values of log n(o vi ) and log n(o vi ) sin@xmath32 in the northern galactic hemisphere compared to the southern galactic hemisphere .
the patchiness in the distribution of o vi absorption is found to be similar over angular scales extending from @xmath33 1@xmath7 to 180@xmath7 .
an extensive survey of o vi in the mw disk has been reported by @xcite in which the authors have studied o vi column density towards 148 early type stars . the correlation between o vi column density and effective distance to a star establishes the fact that the o vi is interstellar in nature and points to the universal presence of the interstellar phenomena that gives rise to o vi throughout the galaxy .
@xcite have surveyed o vi absorption towards 18 early type stars in the smc and report a widespread presence of o vi .
the mean value of log n(o vi ) in the smc is 14.53 , which is higher than the lmc and the mw values . the o vi column density in the smc correlates with the distance from ngc 346 , a star forming region that shows the highest abundance of o vi in the smc .
the lmc has been the focus of h@xmath1 and x - ray surveys to search for ionized structures in the ism .
h@xmath1 surveys have revealed the presence of hii regions , supernova remnants , and large scale structures including superbubbles and super - shells @xcite , whereas , x - ray observations have been done to study bright x - ray sources @xcite , the hot gas in the ism @xcite and diffuse x - ray emission @xcite . since o
vi traces the ism gas with temperatures @xmath6 10@xmath28k , which is at the interface between hot gas ( temperature @xmath35 10@xmath27k ) traced by x - rays and warm gas ( temperature @xmath6 10@xmath36k ) traced by h@xmath1 , correlation between o vi abundance and x - ray and h@xmath1 emissions is expected .
the o vi observations cover specific regions of the lmc with varying environmental conditions .
to get an idea about the variation of o vi column densities with different environments in the lmc , we have overlaid o vi column densities as circles on h@xmath1 image ( @xcite ; fig .
[ fig5 ] ) .
the diameter of the circle is linearly proportional to log n(o vi ) .
interestingly , it is noted that o vi abundance is high in regions with low h@xmath1 and x - ray emissions , i.e. , regions that are relatively inactive .
however , regions like superbubbles are o vi rich , for e.g. , 30 doradus c and n11 .
the gross picture suggests that o vi does not correlate with either h@xmath1 or x - ray emissions .
we have plotted log n(o vi ) against log relative h@xmath1 and x - ray surface brightnesses to have a better insight .
the x - ray surface brightness is obtained from _ rosat _
pspc mosaic image of the lmc covering the energy range 0.52.0 kev @xcite .
figures [ fig6 ] and [ fig7 ] show the correlation of log n(o vi ) with h@xmath1 and x - ray surface brightnesses for the lmc barring five lines of sight in the x - ray correlation plot .
four of the excluded lines of sight ( sk-67d250 , d301 - 1005 , d301-nw8 and sk-65d63 ) are not covered by the x - ray observations and one ( sk-69d257 ) has extremely high x - ray emission ( about 2 orders of magnitude higher than other lines of sight ) .
the h@xmath1 and x - ray surface brightness are measured by re - binning the corresponding images to increase the signal to noise .
the lack of correlation is evident in both plots , as also found by @xcite .
since , h@xmath1 traces warm ism and star formation , there seems to be no direct relation between o vi formation and these processes .
a better correlation with x - ray is expected as the hot gas traced by x - ray is presumed to cool through temperatures where o vi formation takes place .
there is a high abundance of o vi in supernova remnants and superbubbles of the lmc , whereas , bright x - ray emission is observed mostly from the supernova remnants of the lmc .
x - ray emission associated with supernova remnants in the lmc is a factor of 2 to 3 times greater than the x - ray emission associated with supergiant shells and the bar in the lmc @xcite .
superbubbles may be formed due to the strong stellar wind from young stars and/or supernova explosions forming a local cavity in the surrounding ism . these are excellent examples of the interaction of young hot stars and the ism .
superbubbles contain hot gas ( @xmath37 ) that is heated by shocks created by stellar winds @xcite , which is an ideal condition for the formation of o vi . the lmc hosts more than 20 superbubbles and recently o vi has been detected in a superbubble n70 @xcite , where they have reported around 60% excess in abundance of o vi in comparison to non - superbubble lines of sight in the proximity of n70 .
the authors conclude that superbubbles act as local o vi reservoirs and have a different absorption profile compared to the non - superbubble o vi absorption profiles .
they find that the o vi is formed by the thermal conduction between the interior hot x - ray producing gas and the cool photo - ionized shell of n70 .
@xcite have detected o vi in emission in several superbubbles in the lmc further emphasizing that superbubbles are important contributors to the overall o vi budget .
we have 22 o vi observations covering 10 superbubbles . of the 22 observations , 3 are in n70 ( sk-67d250 , d301 - 1005 and d301-nw8 ) which have already been reported by @xcite and 4 are in n144 , n204 , n206 and n154 respectively which have been discussed by @xcite .
we report 15 new observations of o vi absorption in the superbubbles 30 doradus c , n158 , n11 , n51 , and n57
. significant variation in o vi abundance exists towards these lines of sight .
the minimum value of log n(o vi ) is 14.04 atoms @xmath0 in superbubble n206 and the maximum value of log n(o vi ) is 14.57 atoms @xmath0 in superbubble n70 . the properties of o vi in superbubbles reported here are tabulated in table [ tab3 ] . comparing the abundance of o vi for the superbubble and non - superbubble lines of sight
, we find that the mean log n(o vi ) for superbubble lines of sight is @xmath38 = 14.35 atoms @xmath0 while for the non - superbubble lines of sight this is @xmath39 = 14.19 atoms @xmath0 . thus , an excess o vi abundance of about 40% in superbubbles of the lmc is found in comparison to non - superbubble regions .
combining the @xcite and @xcite data for superbubble and non - superbubble lines of sight ( excluding sk-67d05 sightline ; see @xcite for details ) , the o vi excess in superbubbles is about 46% , which is comparable to our results .
thus , results reported here support and confirm that superbubbles do show higher o vi abundance in comparison to the general halo absorption seen in other lmc lines of sight . some of the non - superbubble lines of sight show an enhanced o vi column density owing to local effects .
@xcite compared superbubble and non - superbubble lines of sight for the lmc and found that some quiescent environments showed an enhanced o vi abundance , sometimes even larger than that of superbubbles . even in a single superbubble , there are significant variations in the o vi column densities .
we have four observations each for the superbubbles 30 doradus c and n11 . in the case of 30 doradus c , we find that the variation in n(o vi ) is more than a factor of 2 ( from the minimum value of n(o vi ) to the maximum value ) . for n11
, this variation is about a factor of 2.5 .
the variation in n(o vi ) for n70 is not much ( for the three lines of sight included here ) and this corroborates with the @xcite data .
the 30 doradus region of the lmc is ideally suited to study the interaction between a high rate of star formation and the surrounding ism .
this region is dominated by the star forming cluster ngc 2070 that contains the very interesting star cluster r136 at the center .
the proximity to 30 doradus has allowed extensive research on the stellar content @xcite and initial mass function ( @xcite and references therein ) .
30 doradus has also been investigated in detail in the infrared bands to study the dust properties @xcite .
@xcite studied the 30 doradus in the mid - infrared wavelength band to determine the physical conditions of the ionized gas and found that the local effects of hot stars in 30 doradus appear to dominate over any large - scale trend with distance from the central cluster r136 .
the o vi abundance in 30 doradus is higher than in other regions of the lmc with the highest value of log n(o vi ) being 14.56 atoms @xmath0 near the center of the cluster r136 .
[ fig8 ] shows the change in the log n(o vi ) values from the center of r136 up to an angular distance of 1 degree .
we find an overall decrease in the o vi abundance away from the center of r136 .
an interpretation of this plot could be that the processes involved in the formation of o vi are likely to be associated with stellar radiation field but local effects can not be neglected on a large angular scale .
it should be noted that we do not find any correlation between o vi absorption and h@xmath1 emission for lmc and thus , no relation between star formation is established on a larger scale . as noticed earlier , log n(o vi )
does not correlate with x - ray emission for the lmc as a whole but surprisingly there is a good correlation for the 30 doradus lines of sight ( fig .
[ fig9 ] ) .
the correlation is between log n(o vi ) and x - ray emission from 30 doradus considering all the lines of sight within 1 degree around r136 except for one sightline ( sk-69d257 ) that has exceptionally high value of x - ray surface brightness .
one of the reasons for such high x - ray emission from sk-69d257 is its proximity to a high mass x - ray binary lmc x-1 @xcite .
the correlation in 30 doradus suggests that o vi observed in this region is present in the ism gas surrounding the x - ray emitting plasma , which may be due to the compactness , high density and non - uniform structure of 30 doradus .
the correlation confirms that the x - ray emitting gas cools through temperatures where o vi is being formed and supports the general consensus about the formation of o vi by collisional ionization in the interface regions between cooler photo - ionized ism gas and hot exterior ism gas @xcite .
we have presented o vi column density measurements for the lmc using _ fuse _ data for 70 lines of sight .
this is the widest coverage of the lmc to date .
the results reported here reveal significant variation in o vi column densities over a very small angular scale thus confirming the patchiness of o vi distribution in the lmc .
the most important inferences drawn from this work are following : 1 .
this survey probes 70 lines of sight with varying environmental conditions .
we find strong o vi absorption in the lmc that is not restricted to active regions .
high o vi abundance is present even in relatively inactive regions of the lmc .
2 . there are significant variations in the velocity profiles of o vi absorption .
the o vi absorption profile is broader than the mw absorption for many lines of sight but it should be noted that unambiguous separation of the mw and the lmc components is not possible for most of the lines of sight .
this proves to be a significant hurdle in interpreting exact kinematics for o vi in the lmc ( for e.g. outflows from the lmc , etc . ) .
the maximum column density measured for the lmc is log n(o vi ) = 14.57 atoms @xmath0 and minimum value is log n(o vi ) = 13.72 atoms @xmath0 . the mean value of o vi column density is @xmath40log n(o vi)@xmath41 = 14.23 atoms @xmath0 , which is slightly lower than the earlier reported value .
the median value of o vi column density in the lmc comes to be 14.22 atoms @xmath0 .
the results corroborates with the previous finding that the distribution of o vi in the lmc is patchy .
4 . despite the fact that the lmc has lower metallicity than the mw , the abundance of o vi and properties of o vi absorption are similar in both the galaxies .
the mean of log n@xmath29(o vi ) value for the mw is 14.29 atoms @xmath0 while the projected column density for the lmc , i.e. , log n@xmath29(o vi ) is 14.15 atoms @xmath0 . a more extensive study for the mw suggests this value to be 14.21 atoms @xmath0 .
the smc with even lower metallicity has higher o vi abundance with mean log n(o vi ) = 14.53 atoms @xmath0 . 5 .
o vi absorption in the lmc does not correlate with h@xmath1 ( warm gas ) or x - ray ( hot gas ) emission but we find a good correlation between o vi absorption and x - ray emission in the 30 doradus region .
it is also seen that the o vi absorption is decreasing with increasing angular distance from the star cluster r136 suggesting some correlation with stellar radiation field .
the observations reported cover 10 superbubbles of the lmc and for 5 superbubbles ( 30 doradus c , n158 , n51 , n11 and n57 ) o vi absorption is reported for the first time .
superbubbles are o vi rich and have 40% excess compared to non - superbubble lines of sight .
ap and nvs acknowledge fellowships from the department of science and technology , new delhi . this research has made use of _ far ultraviolet spectroscopic explorer _ data .
_ fuse _ was operated by johns hopkins university for nasa .
we acknowledge the use of multimission archive at space telescope science institute ( mast ) and nasa astrophysics data system ( ads ) .
absorption overplotted on the normalized o vi absorption profiles for 3 lines of sight . for sk-65d21 ,
h@xmath15 absorption does not contaminate the lmc o vi absorption , while sk-71d45 shows the maximum contamination .
sk-67d250 shows moderate contamination . ] ) and for other individual regions in the lmc .
the plot includes all lines of sight from fig .
[ fig8 ] except for 1 sightline ( sk-69d257 ) as it shows an exceptionally high x - ray emission due to its proximity to x - ray binary lmc - x1 ( points et al .
2001 ) . ]
p1030703 & sk-67d05 & lwrs & 04 50 18.96 & -67 39 37.9 & o9.7ib & 11.34 & 1 + a0490402 & sk-68d03 & lwrs & 04 52 15.56 & -68 24 26.9 & o9i & 13.13 & 2 + e5111901 & bi13 & lwrs & 04 53 06.48 & -68 03 23.1 & o6.5v & 13.75 & 3 + d0980801 & sk-67d18 & mdrs & 04 55 14.90 & -67 11 24.5 & o6 - 7n - nm+wn5 - 6a & 11.95 & 1 + p1174401 & sk-67d20 & lwrs & 04 55 31.50 & -67 30 01.0 & wn4b & 13.86 & 4 + b0100201 & pgmw-3070 & mdrs & 04 56 43.25 & -66 25 02.0 & o6v & 12.75 & 5 + d0300302 & lh103102 & mdrs & 04 56 45.40 & -66 24 45.9 & o7vz & 13.55 & 5 + d0300101 & lh91486 & mdrs & 04 56 55.58 & -66 28 58.0 & o6.5vz & 14.2 & 5 + b0100701 & pgmw-3223 & mdrs & 04 57 00.80 & -66 24 25.3 & o8.5iv & 12.95 & 5 + f9270604 & hv2241 & lwrs & 04 57 15.83 & -66 33 54.8 & o7-iii & 13.5 & 5 + p1174901 & hd32402 & lwrs & 04 57 24.19 & -68 23 57.2 & wc & & 3 + z9050801 & sk-67d32 & lwrs & 04 59 52.30 & -67 56 55.0 & wn & 14.48 & 3 + p1030901 & sk-65d21 & lwrs & 05 01 22.33 & -65 41 48.1 & o9.7iab & 12.02 & 6 + b0770201 & hv2274 & lwrs & 05 02 40.96 & -68 24 21.3 & b0-b2iii - i & 14.2 & 5 + p1174501 & sk-66d51 & lwrs & 05 03 10.20 & -66 40 54.0 & wn8h & 12.71 & 4 + d1380201 & ngc1818-d1 & lwrs & 05 04 32.56 & -66 24 51.0 & b0-b2v - iv & 14.93 & 5 + p1172101 & sk-70d69 & lwrs & 05 05 18.73 & -70 25 49.8 & o5v & 13.94 & 6 + p1171703 & sk-67d69 & lwrs & 05 14 20.16 & -67 08 03.5 & o4iii(f ) & 13.09 & 7 + c1030101 & macho78 - 6097 & lwrs & 05 18 04.70 & -69 48 19.0 & b0-b2v - iv & 14.4 & 5 + e5112902 & bi130 & lwrs & 05 18 06.06 & -69 14 34.5 & o8.5v((f ) ) & 12.53 & 3 + z9050201 & sk-69d94 & lwrs & 05 18 14.53 & -69 15 01.0 & b0-b2iii - i & 9.72 & 5 + d9044701 & snr0519 - 697 & lwrs & 05 18 44.20 & -69 39 12.4 & supernova - remnant & & + d0880104 & brey22 & mdrs & 05 19 16.40 & -69 39 19.5 & wc / o9.5ib & 12.3 & 1 + b0270101 & hd269445 & lwrs & 05 22 59.87 & -68 01 46.6 & wn & 11.45 & 1 + p1173602 & sk-69d124 & lwrs & 05 25 18.37 & -69 03 11.1 & o9ib & 12.81 & 2 + d1530104 & sk-67d105 & lwrs & 05 26 06.37 & -67 10 57.6 & of & 12.42 & 5 + a1110101 & sk-67d106 & mdrs & 05 26 15.20 & -67 29 58.3 & b0-b2iii - i & 11.78 & 8 + a1110201 & sk-67d107 & mdrs & 05 26 20.67 & -67 29 55.4 & b0-b2iii - i & 12.5 & 8 + c1510101 & hd36521 & lwrs & 05 26 30.32 & -68 50 25.4 & wc & 12.42 & 9 + p1031402 & sk-68d80 & lwrs & 05 26 30.43 & -68 50 26.6 & wc4+ob & 12.40 & 10 + p2030104 & sk-68d82 & lwrs & 05 26 45.31 & -68 49 52.8 & wn5?b+(b3i ) & 9.86 & 5 + p1173701 & bi170 & lwrs & 05 26 47.79 & -69 06 11.7 & o9.5ib & 13.09 & 11 + c1550201 & sk-67d111 & lwrs & 05 26 47.95 & -67 29 29.9 & o6:iafpe & 12.57 & 11 + f9270701 & hv2543 & lwrs & 05 27 27.40 & -67 11 55.4 & o8+o9 & 12.92 & 5 + p1172501 & sk-70d91 & lwrs & 05 27 33.74 & -70 36 48.3 & o6.5v & 12.78 & 11 + p1172303 & sk-66d100 & lwrs & 05 27 45.59 & -66 55 15.0 & o6ii(f ) & 13.26 & 6 + e9570301 & hde269599 & lwrs & 05 28 22.68 & -69 08 32.2 & be & 10.18 & 5 + m1142001 & sk-65d63 & lwrs & 05 28 39.50 & -65 39 01.1 & o9.7i & 12.56 & 12 + c1030201 & hv982 & lwrs & 05 29 52.50 & -69 09 22.0 & b0-b2v - iv & 14.6 & 5 + e5113602 & sk-70d97 & lwrs & 05 30 11.35 & -70 51 42.2 & o9iii & 13.33 & 5 + p1175001 & sk-67d144 & lwrs & 05 30 12.22 & -67 26 08.4 & wc4+ob & 13.6 & 13 + p2170301 & bi184 & lwrs & 05 30 30.60 & -71 02 31.3 & b0-b2v - iv & 13.84 & 3 + d1380301 & ngc2004-b15 & mdrs & 05 30 36.58 & -67 17 42.3 & b0-b2v - iv & 14.18 & 5 + z9050601 & sk-71d38 & lwrs & 05 30 38.77 & -71 01 47.9 & wc & 13.1 & 5 + p1031502 & sk-71d45 & lwrs & 05 31 15.55 & -71 04 08.9 & o4 - 5iii(f ) & 11.51 & 1 + a1330123 & sk-67d166 & lwrs & 05 31 44.31 & -67 38 00.6 & o4if+ & 12.27 & 1 + b0860901 & sk-67d168 & mdrs & 05 31 52.10 & -67 34 20.8 & o8iaf & 12.08 & 12 + d9042002 & snr0532 - 675 & lwrs & 05 32 23.00 & -67 31 02.0 & supernova - remnant & & + p1173101 & sk-67d191 & lwrs & 05 33 34.12 & -67 30 19.6 & o8v & 13.46 & 2 + b0770302 & hv5936 & lwrs & 05 33 39.00 & -66 37 39.8 & b0-b2v - iv & 14.8 & 5 + p1175101 & sk-69d191 & lwrs & 05 34 19.39 & -69 45 10.0 & wc4 & 13.35 & 13 + f3210301 & j053441 - 693139 & lwrs & 05 34 41.30 & -69 31 39.0 & o2-o3.5if * & 13.7 & 14 + p1171603 & sk-67d211 & lwrs & 05 35 13.92 & -67 33 27.0 & o2iii(f ) * & 12.28 & 14 + z9051001 & brey64 & lwrs & 05 35 54.45 & -68 59 07.4 & wn9h & 13.21 & 3 + d9043101 & snr0536 - 692 & lwrs & 05 36 07.70 & -69 11 52.6 & supernova - remnant & & + z9050701 & sk-69d220 & lwrs & 05 36 43.83 & -69 29 47.4 & wnoroiafpe & 10.58 & 1 + p1172201 & sk-66d172 & lwrs & 05 37 05.56 & -66 21 35.7 & o2iii(f)*+ob & 13.13 & 14 + c0020601 & bi253 & lwrs & 05 37 34.49 & -69 01 09.8 & of & 13.76 & 3 + e5114201 & sk-68d137 & lwrs & 05 38 24.77 & -68 52 32.8 & o3iiif & 13.29 & 3 + p1171803 & mk42 & lwrs & 05 38 42.10 & -69 05 54.7 & o3if / wn6-a & 10.96 & 15 + p1031706 & sk-69d243 & lwrs & 05 38 42.57 & -69 06 03.2 & wn5+ob & 9.5 & 16 + f9140101 & 30dor - s - r136 & lwrs & 05 38 51.70 & -69 06 00.0 & hii - region & & + p1031802 & sk-69d246 & lwrs & 05 38 53.50 & -69 02 00.7 & wn & 11.16 & 4 + p1174601 & hde269927 & lwrs & 05 38 58.25 & -69 29 19.1 & wn & 12.63 & 4 + p2170101 & sk-69d257 & lwrs & 05 39 58.91 & -69 44 03.2 & o9ii & 12.53 & 5 + d9042801 & snr0543 - 689 & lwrs & 05 43 07.20 &
-68 58 52.0 & supernova - remnant & & + d0981401 & d301 - 1005 & mdrs & 05 43 08.30 & -67 50 52.4 & o9.5v & 14.11 & 3 + d0981201 & sk-67d250 & mdrs & 05 43 15.48 & -67 51 09.6 & o7.5ii(f ) & 12.68 & 17 + d0981501 & d301-nw8 & mdrs & 05 43 15.96 & -67 49 51.0 & d301-nw8 & 14.37 & 3 + i8120701 & sk-70d115 & lwrs & 05 48 49.76 & -70 03 57.5 & o6.5iaf & 12.24 & 5 + notes .
units of right ascension are hours , minutes , and seconds ; units of declination are in degrees , arcminutes , and arcseconds
. + references ( 1 ) @xcite ; ( 2 ) @xcite ; ( 3 ) @xcite ; ( 4 ) @xcite ; ( 5 ) @xcite ; ( 6 ) @xcite ; ( 7 ) @xcite ( 8) @xcite ; ( 9 ) @xcite ; ( 10 ) @xcite ; ( 11 ) @xcite ( 12 ) @xcite ; ( 13 ) @xcite ; ( 14 ) @xcite ; ( 15 ) @xcite ; ( 16 ) @xcite ; ( 17 ) @xcite sk-67d05 & 73@xmath426 & 13.72@xmath43 & 175 , 330 + sk-68d03 & 112@xmath428 & 14.02@xmath44 & 180 , 330 + bi13 & 88@xmath424 & 13.94@xmath45 & 175 , 315 + sk-67d18 & 132@xmath427 & 14.16@xmath46 & 165 , 330 + sk-67d20 & 147@xmath428 & 14.08@xmath47 & 175 , 335 + pgmw-3070 & 132@xmath4223 & 14.10@xmath48 & 180 , 345 + lh103102 & 132@xmath4223 & 14.16@xmath49 & 180 , 330 + lh91486 & 266@xmath4229 & 14.47@xmath50 & 175 , 385 + pgmw-3223 & 129@xmath4215 & 14.10@xmath51 & 175 , 315 + hv2241 & 271@xmath4218 & 14.50@xmath52 & 165 , 365 + hd32402 & 271@xmath428 & 14.48@xmath53 & 160 , 320 + sk-67d32 & 129@xmath427 & 14.11@xmath49 & 165 , 360 + sk-65d21 & 94@xmath4223 & 13.94@xmath47 & 225 , 340 + hv2274 & 160@xmath4211 & 14.14@xmath54 & 165 , 345 + sk-66d51 & 185@xmath4210 & 14.26@xmath55 & 170 , 370 + ngc1818-d1 & 280@xmath4213 & 14.52@xmath56 & 150 , 340 + sk-70d69 & 182@xmath4212 & 14.18@xmath57 & 150 , 295 + sk-67d69 & 232@xmath4228 & 14.40@xmath58 & 170 , 340 + macho78 - 6097 & 151@xmath4225 & 14.15@xmath56 & 165 , 320 + bi130 & 94@xmath427 & 13.89@xmath55 & 165 , 320 + sk-69d94 & 150@xmath423 & 14.22@xmath59 & 160 , 340 + snr0519 - 697 & 97@xmath424 & 13.97@xmath57 & 160 , 300 + brey22 & 126@xmath4224 & 14.07@xmath51 & 160 , 330 + hd269445 & 200@xmath4210 & 14.29@xmath60 & 175 , 365 + sk-69d124 & 331@xmath428 & 14.57@xmath60 & 190 , 430 + sk-67d105 & 110@xmath428 & 13.92@xmath61 & 180 , 320 + sk-67d106 & 180@xmath427 & 14.30@xmath43 & 175 , 345 + sk-67d107 & 254@xmath4210 & 14.45@xmath60 & 160 , 360 + hd36521 & 126@xmath429 & 14.07@xmath62 & 175 , 340 + sk-68d80 & 303@xmath4216 & 14.56@xmath52 & 145 , 335 + sk-68d82 & 141@xmath4210 & 14.17@xmath63 & 165 , 320 + bi170 & 235@xmath4220 & 14.43@xmath64 & 165 , 365 + sk-67d111 & 214@xmath4219 & 14.34@xmath56 & 175 , 365 + hv2543 & 156@xmath4242 & 14.23@xmath52 & 160 , 365 + sk-70d91 & 256@xmath427 & 14.43@xmath52 & 160 , 365 + sk-66d100 & 214@xmath4221 & 14.34@xmath59 & 160 , 340 + hde269599 & 123@xmath426 & 14.05@xmath59 & 165 , 320 + sk-65d63 & 184@xmath428 & 14.21@xmath63 & 180 , 375 + hv982 & 208@xmath4258 & 14.33@xmath65 & 175 , 360 + sk-70d97 & 226@xmath424 & 14.39@xmath49 & 175 , 375 + sk-67d144 & 194@xmath4210 & 14.25@xmath51 & 165 , 335 + bi184 & 118@xmath4210 & 14.04@xmath66 & 165 , 330 + ngc2004-b15 & 92@xmath427 & 13.86@xmath67 & 175 , 330 + sk-71d38 & 94@xmath427 & 13.92@xmath57 & 165 , 315 + sk-71d45 & 194@xmath429 & 14.26@xmath68 & 160 , 345 + sk-67d166 & 206@xmath429 & 14.32@xmath56 & 165 , 390 + sk-67d168 & 186@xmath4212 & 14.26@xmath56 & 165 , 375 + snr0532 - 675 & 147@xmath429 & 14.16@xmath57 & 165 , 345 + sk-67d191 & 229@xmath4216 & 14.42@xmath53 & 165 , 340 + hv5936 & 174@xmath428 & 14.18@xmath57 & 175 , 375 + sk-69d191 & 185@xmath4225 & 14.22@xmath51 & 165 , 340 + j053441 - 693139 & 182@xmath4233 & 14.22@xmath56 & 165 , 330 + sk-67d211 & 144@xmath428 & 14.11@xmath43 & 160 , 350 + brey64 & 139@xmath428 & 14.20@xmath69 & 180 , 330 + snr0536 - 692 & 116@xmath427 & 14.03@xmath67 & 165 , 320 + sk-69d220 & 128@xmath429 & 14.09@xmath70 & 160 , 315 + sk-66d172 & 173@xmath429 & 14.20@xmath56 & 175 , 360 + bi253 & 259@xmath4246 & 14.46@xmath60 & 160 , 300 + sk-68d137 & 234@xmath427 & 14.45@xmath71 & 165 , 330 + mk42 & 228@xmath4224 & 14.41@xmath57 & 160 , 330 + sk-69d243 & 307@xmath4215 & 14.56@xmath58 & 150 , 345 + 30dor - s - r136 & 185@xmath4225 & 14.25@xmath58 & 165 , 320 + sk-69d246 & 211@xmath427 & 14.37@xmath59 & 155 , 325 + hde269927 & 245@xmath428 & 14.42@xmath49 & 160 , 320 + sk-69d257 & 171@xmath428 & 14.20@xmath57 & 160 , 315 + snr0543 - 689 & 186@xmath4234 & 14.27@xmath72 & 160 , 360 + d301 - 1005 & 284@xmath4257 & 14.53@xmath64 & 165 , 385 + sk-67d250 & 316@xmath4233 & 14.57@xmath59 & 165 , 375 + d301-nw8 & 228@xmath4230 & 14.42@xmath64 & 175 , 365 + sk-70d115 & 186@xmath4211 & 14.23@xmath49 & 165 , 330 + | we present a survey of interstellar o vi absorption in the large magellanic cloud ( lmc ) towards 70 lines of sight based on _ far ultraviolet spectroscopic explorer ( fuse ) _ observations .
the survey covers o vi absorption in a large number of objects in different environmental conditions of the lmc .
overall , a high abundance of o vi is present in active and inactive regions of the lmc with mean log n(o vi ) = 14.23 atoms @xmath0 .
there is no correlation observed between o vi absorption and emissions from the hot gas ( x - ray surface brightness ) or the warm gas ( h@xmath1 surface brightness ) .
o vi absorption in the lmc is patchy and the properties are similar to that of the milky way ( mw ) . in comparison to the small magellanic cloud ( smc ) , o vi is lower in abundance even though smc has a lower metallicity compared to the lmc and the mw .
we present observations in 10 superbubbles of the lmc of which we detect o vi absorption in 5 superbubbles for the first time and the superbubbles show an excess o vi absorption of about 40% compared to non - superbubble lines of sight .
we have also studied the properties of o vi absorption in the 30 doradus region .
even though o vi does not show any correlation with x - ray emission for the lmc , a good correlation between log n(o vi ) and x - ray surface brightness for 30 doradus region is present .
we also find that o vi abundance decreases with increasing distance from the star cluster r136 .
[ firstpage ] galaxies : ism ism : structure ism : atoms galaxies : individual : large magellanic cloud ultraviolet : ism |
the design and the analysis of caching systems , a very traditional and widely studied topic in computer science , has recently drawn again attention by the networking research community .
this interest revival is mainly due to the important role that caches play today in the distribution of contents over the internet .
massive content delivery networks , indeed , represent the standard solution adopted by content and network providers to reach large populations of geographically distributed users in an effective way .
mcdn allow providers to cache contents close to the users , achieving the twofold goal of reducing network traffic while minimizing the latency suffered by users .
unfortunately , the performance evaluation of caching systems is very hard , as the computational cost to analyze the behavior of a cache is exponential in both the cache size and the number of contents @xcite .
for this reason , the effort of the research community has mainly focused on the development of accurate and computationally efficient approximate techniques for the analysis of caching systems , under various traffic conditions .
che s approximation @xcite , proposed for the analysis of least recently used ( lru ) caches under the independent reference model ( irm ) , has emerged as one of the most powerful methods to obtain accurate estimates of the hit " probability at limited computational costs @xcite .
the main idea of this technique is to summarize the response of a cache to the requests arriving for any possible content by a single primitive quantity , which is assumed to be deterministic and the same for any content .
this approximation simplifies the analysis of caching systems because it allows to decouple the dynamics of different contents . in @xcite che
s approximation for lru caches under the irm found a theoretical justification . in this paper
we extend the mathematical analysis of che s approximation to lru caches operating under the shot noise model ( snm ) @xcite .
this model provides a simple , flexible and accurate description of the temporal locality found e.g. in video on demand ( vod ) traffic , capturing today traffic characteristics in a more natural and precise way than traditional traffic models .
inspired by the seminal paper @xcite , we investigate the validity of che s approximation by means of an asymptotic analysis of the cache eviction time .
specifically , we provide a large deviation principle , a law of large numbers and a central limit theorem for the cache eviction time , as the cache size grows large .
furthermore , to the best of our knowledge , we give for the first time a non - asymptotic analytical upper bound for the error estimate of the hit `` probability entailed by che s approximation .
we provide also upper and lower bounds for the hit '' probability , under che s approximation , for a tandem network of caches . finally , we present some numerical illustrations .
our results show that che s approximation is a provable , highly accurate and scalable tool to assess the performance of lru caching systems under the snm .
we consider a cache , whose size ( or capacity ) , expressed in number of objects ( or contents ) , is denoted by @xmath0 .
the cache is fed by an exogenous arrival process of objects requests generated by users .
requests which find the object in the cache are said to produce a hit `` , whereas requests that do not find the object in the cache are said to produce a ' ' miss " .
an important performance index is the hit " probability , which is the fraction of the requests producing a hit " .
the miss stream of a cache , i.e. the process of requests which are not locally satisfied by the cache , is forwarded to either other caches or to a common repository containing all the objects , i.e. the entire objects catalogue . in the literature
it is common to neglect all propagation delays . in this paper
we focus on caches implementing the lru policy : upon the arrival of a request , an object not already stored in the cache is inserted into it . if the cache is full , to make room for a new object the least recently used item is evicted , i.e. the object which has not been requested for the longest time is expunged from the cache .
several models have been proposed to describe the process of requests arriving at a cache .
the simplest and still the most widely adopted is certainly the irm @xcite , which makes the following two fundamental assumptions : @xmath1 the catalogue consists of a fixed number of objects , which does not change over the time ; @xmath2 the process of requests of a given object is modeled by a homogeneous poisson process . as a consequence , the irm completely ignores all temporal correlations in the sequence of requests and does not take into account a key feature of real traffic referred to as temporal locality , which means that if an object is requested at a given time , then it is more likely that the same object will be requested again in the near future .
it is well - known that the temporal locality has a beneficial effect on the cache performance , as it increases the hit " probability @xcite .
several extensions of the irm have been proposed to incorporate the temporal locality into the traffic model .
existing generalizations @xcite typically assume that the process of requests is time - stationary , usually either a renewal process or a markov modulated poisson process .
however , these models do not capture the kind of temporal locality encountered in traces related to video on demand ( vod ) traffic , which is instead well described by the snm as shown in @xcite .
the basic idea of the snm is to represent the requests process as the superposition of many independent processes , each one referring to a specific object .
the requests process of a fixed content @xmath3 is specified by two physical ( random ) parameters : @xmath4 and @xmath5 .
@xmath4 represents the time instant at which the content enters the system ( i.e. it becomes available to the users ) ; mark @xmath5 describes some attribute of the content @xmath3 , which summarizes its main characteristics ( content type , volume , etc . ) .
we assume that the set of times @xmath6 at which contents become available to users ( i.e. , they are introduced in the common repository ) is distributed according to a homogeneous poisson process on @xmath7 with intensity @xmath8 .
here , @xmath9 is supposed to be an unordered set of times . we suppose that
, after the introduction into the catalogue of the content @xmath3 , the requests for this content arrive at the cache according to a cox process @xmath10 on @xmath7 whose stochastic intensity @xmath11 is defined by @xmath12 see e.g. @xcite .
we assume that @xmath13 is a sequence of independent and identically distributed random variables , independent of @xmath9 , with values on some measurable space @xmath14 .
furthermore , we suppose that @xmath15 is a measurable non - negative function such that @xmath16 for any @xmath17 and @xmath18 .
finally , we suppose that , for any @xmath19 , @xmath20 almost surely and we assume that the cox processes @xmath21 are independent , given @xmath22 .
we denote by @xmath23 a tagged content introduced into the catalogue at the deterministic time @xmath24 , @xmath25 .
moreover , we denote by @xmath26 , @xmath27 , the number of contents different from @xmath23 that have been requested in the time interval @xmath28 $ ] , i.e. @xmath29,\,\xi_m\in ( -\infty , t]\}.\nonumber\end{aligned}\ ] ] throughout this paper we shall consider the random variable @xmath30 which plays an important role in the dynamics of an lru cache because the cache eviction time may be expressed in terms of @xmath31 .
indeed , under the lru replacement policy , given the event @xmath32 , we have that the content @xmath23 is expunged from the cache ( provided it has not requested in the meanwhile ) as soon as the @xmath0th content , different from @xmath23 , is requested .
so , under the lru replacement policy , the so - called cache eviction time for the content @xmath23 is given by the random variable @xmath33 define the quantity @xmath34\,\mathrm{d}u,\quad t>0.\ ] ] the following proposition holds .
[ prop : lap ] if @xmath35 , then the random variable @xmath31 is poisson distributed with mean @xmath36 .
note that the condition @xmath35 is fairly general : for example , it is satisfied whenever the popularity profile is of multiplicative form i.e. ( with a little abuse of notation ) @xmath37 and @xmath38<\infty$.}\ ] ] indeed , in such a case , for @xmath27 , we have @xmath39\,\mathrm{d}u \label{eqg}\\ & \leq{\mathbb{e}}[z_1]\int_0^\infty\left(\int_{u - t}^{u}h(s)\,\mathrm{d}s\right)\,\mathrm{d}u\leq{\mathbb{e}}[z_1]t
< \infty,\nonumber\end{aligned}\ ] ] where @xmath40 $ ] , @xmath41 , and we used the elementary inequality @xmath42 , @xmath43 . _
proofofproposition[prop : lap ] .
_ for any @xmath44 , we define the restriction of @xmath26 to contents that have been introduced in the model in the time interval @xmath45 $ ] by @xmath46,\,\xi_m\in [ t - t_0,t]\}. \nonumber\end{aligned}\ ] ] by the slivnyak theorem ( see e.g. @xcite ) , the law of @xmath47 given the event @xmath32 coincides with the law of @xmath9 and so , for any @xmath41 , @xmath48 = \mathbb{e}\left[\mathrm{e}^{\theta \widetilde{\bold{x}}_{t_0}(t)}\right],\ ] ] where @xmath49,\,\xi_m\in [ t - t_0,t]\}.\ ] ] letting @xmath50)$ ] denote the number of points @xmath9 in the time interval @xmath51 $ ] and @xmath52)$ ] denote the number of points @xmath53 in the time interval @xmath28 $ ] , we rewrite @xmath54 as @xmath55)}\ind\{n^{(m)}([0,t])\geq 1\}.\ ] ] since , given @xmath4 and @xmath5 , @xmath56 is a poisson process with intensity function @xmath57 , we have @xmath58)\geq 1\,|\,\xi_m , z_m)= 1-\mathrm{e}^{-\int_{0}^{t}h(s-\xi_m , z_m)\,\mathrm{d}s}.\nonumber \ ] ] recalling that , given @xmath59)=k\}$ ] , the @xmath60 points of @xmath61 on @xmath51 $ ] are independent and uniformly distributed over @xmath51 $ ] ( see e.g. @xcite ) , for any @xmath41 , we have @xmath62)=k\right]&=\mathbb{e}\left[\prod_{m=1}^{k}\mathrm{e}^{\theta\ind\{n^{(m)}([0,t])\geq 1\}}\,\mid\ , n([t - t_0,t])=k\right]\nonumber\\ & = \prod_{m=1}^{k}\left(1+(\mathrm{e}^\theta-1)\frac{1}{t_0}\int_{t - t_0}^t\mathbb{e } [ p_t(u , z_1)]\,\mathrm{d}u\right)\nonumber\\ & = \left[\frac{1}{t_0}\int_{-t_0}^0\mathbb{e}\left[1+(\mathrm{e}^\theta-1 ) p_t(u+t , z_1)\right]\,\mathrm{d}u\right]^{k}.\nonumber\end{aligned}\ ] ] therefore ,
@xmath63&= \mathrm{e}^{-\lambda t_0 } \sum_{k\geq 0 } \left[\frac{\lambda}{k!}\int_{-t_0}^0\mathbb{e}\left[1+(\mathrm{e}^\theta-1 ) p_t(u+t , z_1)\right]\,\mathrm{d}u\right]^{k}\nonumber\\ & = \exp\left(\lambda(\mathrm{e}^\theta-1)\int_{-t_0}^0\mathbb{e } \left[p_t(u+t , z_1)\right]\,\mathrm{d}u\right).\label{eq:2}\end{aligned}\ ] ] the claim follows by and , letting @xmath64 tend to @xmath65 .
+ @xmath66 in the context of an lru cache under the snm , che s approximation consists in replacing the cache eviction time @xmath67 by the deterministic constant @xmath68=c\}.\ ] ] note that , if @xmath35 for any @xmath27 , then by proposition [ prop : lap ] we have @xmath69 so , if moreover @xmath70 is strictly increasing , we deduce @xmath71 since the law of @xmath31 ( and therefore of @xmath67 ) does not depend on @xmath23 and @xmath72 , hereafter we simply write @xmath73 , @xmath74 and @xmath75 in place of @xmath31 , @xmath67 and @xmath76 . in this section
we investigate the validity of che s approximation for large values of @xmath0 .
we shall do this by analyzing the behavior of @xmath74 as @xmath77 .
intuitively , che s approximation finds a theoretical justification if we may show that , as @xmath77 , @xmath78 almost surely .
this is indeed achieved in proposition [ cor : lln ] .
proposition [ cor : ldp ] and the inequalities and provide some estimates on the probability that @xmath74 deviates from its most probable value @xmath75 , as @xmath0 grows large .
finally , the gaussian approximation for @xmath74 in proposition [ prop : normalcaching ] gives some insights on how the random variable @xmath79 scales to zero , as @xmath80 .
hereafter , we shall consider the function @xmath81 defined by .
the following tail estimates and law of large numbers hold .
[ cor : ldp ] if @xmath70 is strictly increasing and @xmath82 are bijective and continuous ( i.e. @xmath81 is a homeomorphism of @xmath83 ) , then @xmath84 and @xmath85 [ cor : lln ] under the assumptions of proposition [ cor : ldp ] , if in addition for any divergent sequences @xmath86 , @xmath87 of positive numbers we have @xmath88 then @xmath89 the proofs are based on a large deviation principle for the process @xmath90 stated in lemma [ prop : ldp ] .
we recall , see e.g. @xcite , that a non - negative stochastic process @xmath91 obeys a large deviation principle on @xmath92 with speed @xmath93 and rate function @xmath94 if @xmath95 is a function which increases to infinity and @xmath96 $ ] is a lower semi - continuous function such that , for all borel sets @xmath97 , @xmath98 where @xmath99 denotes the interior of @xmath100 and @xmath101 denotes the closure of @xmath100 . for later purposes , we recall that a rate function @xmath94 on @xmath92 has no peaks if @xmath102 there exists @xmath103 such that @xmath104 ; @xmath105 @xmath94 is non - increasing on @xmath106 and non - decreasing on @xmath107 .
[ prop : ldp ] under the assumptions of proposition [ cor : ldp ] , we have that the family of random variables @xmath90 obeys a large deviation principle on @xmath92 with speed @xmath108 and rate function @xmath109 , @xmath25 , @xmath110
. * remark1 .
* under the assumptions of proposition [ cor : ldp ] , by , and it follows that for an arbitrarily small @xmath111 , for any @xmath112 there exists @xmath113 so that for any @xmath114 @xmath115 and @xmath116 @xmath66 * remark2 . * for later purposes ,
we remark that the rate function @xmath94 in lemma [ prop : ldp ] is continuous , @xmath117 and @xmath94 decreases on @xmath118 and increases on @xmath119 .
so , in particular , @xmath94 has no peaks .
+ @xmath66 * remark3 . * here , we assume that @xmath81 defined by is a strictly increasing homeomorphism of @xmath83 , and we give sufficient conditions which guarantee .
+ @xmath102 if @xmath120 is ultimately lipschitz continuous , i.e. @xmath121 then holds .
indeed , let @xmath112 be arbitrarily fixed and @xmath122 so large that @xmath123 for all @xmath124 . by the lipschitz property of @xmath120
we have @xmath125 therefore @xmath126 by assumption @xmath127 , as @xmath128 .
therefore there exists @xmath129 such that for any @xmath130 @xmath131 the claim follows combining the inequalities and .
+ @xmath105 if there exists @xmath132 such that @xmath81 is differentiable on @xmath133 and @xmath134 then holds . indeed , for all
@xmath135 we have @xmath136 therefore @xmath137 is ultimately bounded and the claim follows by the previous point @xmath102 .
+ @xmath66 * example1 . *
consider the snm defined by a multiplicative popularity profile of the form and assume .
in such a case , @xmath81 is given by and it clearly satisfies the assumptions of proposition [ cor : ldp ] .
we check that @xmath81 satisfies . setting @xmath138 , @xmath27
, we have @xmath139\,\mathrm{d}u \nonumber\\ & = \int_{0}^t\left[1- \phi_{z_1}\left(-h(u)\right ) \right]\,\mathrm{d}u + \int_{0}^\infty\left[1- \phi_{z_1}\left(-h(u+t ) + h(u)\right ) \right]\,\mathrm{d}u.\label{eq : secondbounded}\end{aligned}\ ] ] since @xmath140 is differentiable on @xmath141 , we easily have that @xmath142 is differentiable on @xmath83 and @xmath143 where the latter inequality follows by the nonnegativity of the third addend in the right - hand - side of .
the claim follows by remark 3 noticing that if @xmath144 is such that @xmath145 , then @xmath146 in the particular case , when @xmath147}(t),\quad\text{for some constant $ l>0$}\ ] ] we have @xmath148,\label{eq : g1}\end{aligned}\ ] ] where for @xmath149 we set @xmath150 and @xmath151 . indeed , for @xmath27 , we have @xmath152}(s)\,\mathrm{d}s\big)\big)\,\mathrm{d}u = \int_{0}^\infty\big(1-\phi_{z_1}(-\eta(t , u))\big)\,\mathrm{d}u,\nonumber\end{aligned}\ ] ] where @xmath153}(s)\,\mathrm{d}s=\frac{(u\wedge l-(u - t)^+)^+}{l}\nonumber \ ] ] and for @xmath154 we set @xmath155 .
we distinguish two cases : @xmath156 and @xmath157 . if @xmath156 , then if @xmath158 then @xmath159
. so , for @xmath160 $ ] , @xmath161}(u)+(t - u+l)\ind_{(l , l+t]}(u)\big)\big)\,\mathrm{d}u \nonumber\\ & = t-\mathbb{e}\left[\frac{l}{z_1}\left(1-\mathrm{e}^{-\frac{z_1}{l}t}\right)\right]+(l - t)\big(1-\phi_{z_1}\big(-\frac{t}{l}\big)\big)+ \int_{l}^{l+t}\big(1-\phi_{z_1}\big(-\frac{1}{l}(t - u+l)\big)\big)\,\mathrm{d}u\nonumber\\ & = 2t+(l - t)\big(1-\phi_{z_1}\big(-\frac{t}{l}\big)\big)-2\mathbb{e}\left[\frac{l}{z_1}\left(1-\mathrm{e}^{-\frac{z_1}{l}t}\right)\right].\nonumber\end{aligned}\ ] ] if @xmath157 , then if @xmath162 then @xmath163
. so , for @xmath157 , @xmath164 + ( t - l)(1-\phi_{z_1}(-1))+l-{\mathbb{e}}\left[\frac{l}{z_1}\left(1-\mathrm{e}^{-z_1}\right)\right]\nonumber\\ & = 2l + ( t - l)(1-\phi_{z_1}(-1))-2\mathbb{e}\left[\frac{l}{z_1}\left(1-\mathrm{e}^{-z_1}\right)\right].\nonumber\end{aligned}\ ] ] + @xmath66 @xmath165 by remark 2 we have that , for any @xmath166 , @xmath167 , and , for any @xmath168 , @xmath169 .
relations and follow by applying the large deviation principle of lemma [ prop : ldp ] considering , respectively , the borel sets @xmath170 and @xmath171 ( note that @xmath120 is strictly increasing since @xmath81 is such ) . + @xmath66 @xmath172 it is well - known that @xmath173 if and only if @xmath174 therefore follows by the borel - cantelli lemma if we check that @xmath175 let @xmath176 be arbitrarily fixed and take @xmath177 .
by lemma [ prop : ldp ] we have that there exists a non - negative integer @xmath178 such that for any @xmath179 @xmath180 and @xmath181 where we used remark 2 .
therefore @xmath182 which proves .
then the claim follows by assumption noticing that by and relation @xmath183 we easily get @xmath184 almost surely , as @xmath80 .
+ @xmath66 the proof of lemma [ prop : ldp ] uses a result from @xcite , which we recall for commodity of the reader .
we first introduce a few notation and terminology .
let @xmath91 be a non - negative stochastic process whose sample paths are right - continuous , non - decreasing and such that @xmath185 almost surely .
we define the inverse process of @xmath91 as @xmath186 the following theorem holds ( see theorem 1@xmath102 in @xcite ) .
[ thm : duffield ] let @xmath91 and @xmath187 be as above and let @xmath188 be a strictly increasing homeomorphism of @xmath83 .
we have that if @xmath189 obeys a large deviation principle on @xmath92 with speed @xmath93 and rate function @xmath94 which has no peaks , then @xmath190 obeys a large deviation principle on @xmath92 with speed @xmath191 and rate function @xmath192 , @xmath193 , @xmath194 .
@xmath195 let @xmath196 be arbitrarily fixed . by proposition [ prop : lap ]
@xmath73 is poisson distributed with mean @xmath36 . therefore @xmath197&=
\lim_{t\to\infty}\frac{1}{g(t)}\log\mathrm{e}^{\lambda g(t)(\mathrm{e}^\theta-1)}=\lambda(\mathrm{e}^\theta-1):=\lambda(\theta).\nonumber\end{aligned}\ ] ] so by the grtner - ellis theorem ( see e.g. @xcite ) the stochastic process @xmath198 satisfies a large deviation principle on @xmath92 with speed @xmath81 and rate function @xmath199 , @xmath25 , @xmath200 .
note that @xmath201 is the inverse process of @xmath202 .
the claim then follows by theorem [ thm : duffield ] .
indeed the rate function @xmath203 has no peaks since @xmath204 and @xmath203 decreases on @xmath205 and increases on @xmath206 .
+ @xmath66 hereafter , we denote by @xmath207 a standard normal random variable and by @xmath208 the convergence in distribution . following the ideas in @xcite ( see propositions 1 and 3 therein ) , we derive a central limit theorem for the cache eviction time of the snm .
[ prop : normalcaching ] assume @xmath70 bijective , strictly increasing and such that there exists a positive function @xmath209 such that @xmath210\quad\text{and}\quad \lim_{y\to\infty}\frac{g(y)-g(y+xf(y))}{\sqrt{g(y+xf(y))}}=-\frac{x}{\sqrt\lambda}. \label{eq : hyp1}\end{aligned}\ ] ] then @xmath211 under the assumptions of proposition [ cor : lln ] , by and we have @xmath212 , as @xmath213 .
therefore , by we deduce that the random variable @xmath214 scales to zero as @xmath215 , as @xmath0 grows large .
* example2 . *
consider the snm defined by a multiplicative popularity profile of the form and assume and @xmath216 in such a case , @xmath81 is given by and it is clearly strictly increasing . in the following we use the same notation as in example 1 .
we check that holds with @xmath217 indeed , this claim follows if we verify @xmath218 by lhopital s rule we have @xmath219 so follows if we check that the second term in is bounded . by the elementary inequality @xmath42 , @xmath43
, we have @xmath220\int_t^\infty(h(u)-h(u - t))\,\mathrm{d}u\nonumber\\ & \leq{\mathbb{e}}[z_1]\int_0^\infty\left(1-h(u)\right)\,\mathrm{d}u={\mathbb{e}}[z_1]\int_0^\infty u h(u)\,\mathrm{d}u<\infty,\nonumber\end{aligned}\ ] ] where the latter equality is a consequence of the fact that @xmath221 is a probability density on @xmath83 .
as shown in example 1 , under the foregoing assumptions the hypotheses of proposition [ cor : lln ] hold and so @xmath214 scales to 0 as @xmath222 , as @xmath80 .
+ @xmath66 the proof of proposition [ prop : normalcaching ] uses lemma [ prop : normal ] below , which is of its own interest .
we denote by @xmath223 the class of real - valued lipschitz functions from @xmath7 to @xmath7 with lipschitz constant less than or equal to one . given two real - valued random variables @xmath224 and @xmath225
, the wasserstein distance between the laws of @xmath224 and @xmath225 , written @xmath226 , is defined as @xmath227-\mathbb{e}[\varphi(u')]|.\ ] ] we recall that the topology induced by @xmath228 on the class of probability measures over @xmath7 is finer than the topology of weak convergence ( see e.g. @xcite ) . [
prop : normal ] if @xmath35 , then @xmath229 * remark4 .
* lemma [ prop : normal ] provides a gaussian approximation for @xmath73 in the wasserstein distance .
on the other hand , proposition 1 in @xcite provides a gaussian approximation , in the kolmogorov distance @xmath230 , for the corresponding quantity under the irm .
we note that a gaussian approximation of @xmath73 , in the kolmogorov distance , under the snm may be easily obtained by lemma [ prop : normal ] using the relation @xmath231 where @xmath232 is a real - valued random variable , see e.g. @xcite .
+ @xmath66 @xmath233 . by the assumptions on @xmath81 we have @xmath234 ,
@xmath235 and @xmath236 , as @xmath77 . for any @xmath237 , @xmath238 by lemma [ prop : normal ]
we have @xmath239 so , letting @xmath0 tend to infinity in and using condition we deduce @xmath240 @xmath66 @xmath241 .
define the borel measure @xmath242 over @xmath28 $ ] ( note that @xmath81 increases on @xmath28 $ ] and so @xmath243 is a lebesgue - stieltjes measure ) and the function @xmath244}(x)/\sqrt{\lambda g(t)}$ ] , @xmath245 $ ] . by corollary 3.4 in @xcite and proposition [ prop :
lap ] , we have @xmath246}|h(x)|^2\,\mu(\mathrm{d}x)\big|+\int_{[0,t]}|h(x)|^3\,\mu(\mathrm{d}x ) = \frac{1}{\sqrt{\lambda g(t)}}.\nonumber\end{aligned}\ ] ] @xmath66 the `` in '' probability is defined as the probability of finding at time @xmath247 a tagged content @xmath23 in the cache , given that @xmath248 and @xmath249 .
thus : + @xmath102 under che s approximation , the `` in '' probability is given by @xmath250)\geq 1\,\mid\,(\xi_{m_0},z_{m_0})=(x , z))\nonumber\\ & = 1-\mathrm{e}^{-\int_{t - x - t_c}^{t - x}h(u , z)\,\mathrm{d}u},\label{eq : pinche}\end{aligned}\ ] ] where @xmath251 denotes the number of points @xmath252 in @xmath253 ; + @xmath105 without relying on che s approximation , the conditional `` in '' probability is given by @xmath254)\geq 1\,|\,(\xi_{m_0},z_{m_0})=(x , z),t_{c})\nonumber\\ & = p_{\mathrm{in},\mathrm{che}}^{(t - x)}(z , t_c).\nonumber\end{aligned}\ ] ] the hit `` probability is defined as the ratio between the average rate at which ' ' hits " of a tagged content occur and the average rate at which requests of the tagged content are observed .
thus : + @xmath255 under che s approximation , the hit " probability is given by @xmath256 } { { \mathbb{e}}[h(t-\xi_{m_0},z_{m_0})]}\nonumber\\ & = \frac{{\mathbb{e}}[h(t-\xi_{m_0},z_{m_0})p^{(t-\xi_{m_0})}_{\mathrm{in},\mathrm{che}}(z_{m_0},t_c ) ] } { { \mathbb{e}}[h(t-\xi_{m_0},z_{m_0})]},\label{eq : refdonoyund}\end{aligned}\ ] ] with the convention @xmath257 . the equality is a consequence of the fact that , under che s approximation , the probability ( denoted by @xmath258 ) that the tagged content @xmath23 , introduced into the catalogue at time @xmath248 and with mark @xmath249 , is found in the cache by an arriving request at time @xmath247 is equal to @xmath259 . indeed @xmath260}(t_n^{(m_0)})\ge
1\,\big|\,t\in n^{(m_0)},\ , ( \xi_{m_0},z_{m_0})=(x , z)\big)\nonumber\\ & = { \mathbb{p}}\big(n^{(m_0)}((t - t_c , t])\geq 1)\big| ( \xi_{m_0},z_{m_0})=(x ,
z)\big)\label{eq : sliv}\\ & = p^{(t - x)}_{\mathrm{in},\mathrm{che}}(z , t_c),\nonumber\end{aligned}\ ] ] where the equality is a consequence of the slivnyak theorem ( see e.g. @xcite ) . note that the probability @xmath261 does not depend on @xmath23 and @xmath247 . indeed , for an arbitrary @xmath262 we have @xmath263}{{\mathbb{e}}[h(t-\xi_{m_0},z_{m_0})\ind\{s<\xi_{m_0}<t\ } ] } & = \frac{(t - s)^{-1}\int_{s}^{t}{\mathbb{e}}[h(t - u , z_{1})p_{\mathrm{in},\mathrm{che}}^{(t - u)}(z_{1},t_c)]\,\mathrm{d}u}{(t - s)^{-1}\int_{s}^{t}{\mathbb{e}}[h(t - u , z_{1})]\,\mathrm{d}u}\nonumber\\ & = \frac{\int_{s}^{t}{\mathbb{e}}[h(t - u , z_{1})p_{\mathrm{in},\mathrm{che}}^{(t - u)}(z_{1},t_c)]\,\mathrm{d}u}{\int_{s}^{t}{\mathbb{e}}[h(t - u , z_{1})]\,\mathrm{d}u},\nonumber\end{aligned}\ ] ] and so letting @xmath264 tend to @xmath265 we deduce @xmath266\,\mathrm{d}u}{\int_{0}^{\infty}{\mathbb{e}}[h(u , z_{1})]\,\mathrm{d}u};\label{phitche}\end{aligned}\ ] ] @xmath267 without relying on che s approximation , the conditional hit " probability is given by @xmath268 } { { \mathbb{e}}[h(t-\xi_{m_0},z_{m_0})]}.\nonumber\end{aligned}\ ] ] arguing as above , one may easily check that @xmath269\,\mathrm{d}u}{\int_{0}^{\infty}{\mathbb{e}}[h(u , z_{1})]\,\mathrm{d}u}.\nonumber\end{aligned}\ ] ] being @xmath270 and @xmath74 independent , the ( unconditional ) hit " probability is given by @xmath271 where @xmath272 denotes the law of @xmath74 and @xmath273 is defined as @xmath261 with @xmath274 in place of @xmath75 . by using the above relations and classical estimates for the tail of a poisson distribution
, we can evaluate the error committed by approximating @xmath275 with @xmath261 .
the following proposition holds .
[ prop : error0 ] if @xmath70 is strictly increasing , then , for any @xmath111 and @xmath276 , we have @xmath277 where @xmath278 , @xmath25 .
this proposition allows an assessment of the accuracy of che s approximation in different scenarios . as shown by the numerical simulations in @xcite ,
see also section [ sec : numerical ] , in most cases exploiting proposition [ prop : error0 ] , we can show that che s approximation leads to surprisingly accurate predictions of caching performance .
_ proofofproposition[prop : error0 ] .
_ we preliminary note that , for any @xmath111 and @xmath276 , we have @xmath279 indeed , since @xmath81 is strictly increasing is equivalent to @xmath280 , which holds since @xmath183 .
note that , due to , @xmath281 is a non - decreasing function .
so , for all @xmath111 , we have @xmath282}(p_{\mathrm{hit},\mathrm{che}}(t_c)-p_{\mathrm{hit},\mathrm{che}}(\theta))\,{\mathbb{p}}_{t_c}(\mathrm{d}\theta)\nonumber\\ & \,\,\,\,\,\ , \,\,\,\,\,\ , + \int_{(t_c(1+\delta),\infty)}(p_{\mathrm{hit},\mathrm{che}}(\theta)-p_{\mathrm{hit},\mathrm{che}}(t_c))\,{\mathbb{p}}_{t_c}(\mathrm{d}\theta)\nonumber\\ & \,\,\,\,\,\ , \,\,\,\,\,\ , + \int_{(t_c(1-\delta),t_c(1+\delta)]}|p_{\mathrm{hit},\mathrm{che}}(\theta)-p_{\mathrm{hit},\mathrm{che}}(t_c)|\,{\mathbb{p}}_{t_c}(\mathrm{d}\theta)\nonumber\\ & \leq{\mathbb{p}}(t_c\leq
t_c(1-\delta))+{\mathbb{p}}(t_c > t_c(1+\delta))\nonumber\\ & \,\,\,\,\,\ , \,\,\,\,\,\ , + \max_{\theta\in\{t_c(1-\delta),t_c(1+\delta)\}}|p_{\mathrm{hit},\mathrm{che}}(\theta)-p_{\mathrm{hit},\mathrm{che}}(t_c)|.\nonumber\end{aligned}\ ] ] the claim follows noticing that by the definition of @xmath74 , the inequality and the properties of the poisson distribution ( see e.g. lemma 1.2 in @xcite , formulas ( 1.10 ) and ( 1.11 ) ) we have @xmath283 and @xmath284 @xmath66
the analysis of networks of caches is a difficult task , indeed an exact characterization of the miss stream of an lru cache is in general prohibitive . under the irm a standard ,
rather crude approach , proposed in the literature ( see e.g. @xcite ) consists in : @xmath1 approximating the miss stream of a content at a cache with a homogeneous poisson process whose rate matches the miss stream rate ; @xmath2 assuming the state of caches to be independent . however , significant errors may be experienced .
an alternative approach , that has been recently proposed for feed - forward networks of lru caches ( such as networks with linear topologies or trees ) consists in approximating the real miss stream with that of a cache operating under che s approximation , ( see @xcite and @xcite ) .
this approach has been experimentally shown to be potentially fairly accurate , but , unfortunately , at the same time , it is computationally highly expensive @xcite .
recently a more efficient procedure has been proposed in @xcite , where further approximations are considered to simplify the computation of the hit probability . however , in this latter case , the accuracy of the estimate is in part sacrificed . here
we show how the approach of @xcite can be adapted to the snm .
our study reveals that the exact computation of the `` hit '' probability , under che s approximation , for a simple tandem network of caches ( i.e. a network of two lru caches in series ) is computationally hard ( see remark 5 ) .
this is mainly due to the effect of the complex dependencies between the states of the two caches .
since the analysis at the first cache can be carried on as in the previous section , here we focus on the second cache .
note that an arriving request for content @xmath23 can produce a `` hit '' at the second cache only if it misses the content @xmath23 at the first cache .
so , under che s approximation , the hit " probability for content @xmath23 , introduced into the catalogue at time @xmath248 and with mark @xmath249 , is given by @xmath285}(t_n^{(m_0)})=0,\nonumber\\ & \,\,\,\,\,\,\,\,\,\,\,\,\ , \sum_{n}\ind_{(t - t_{c_2},t)}(t_n^{(m_0)})\ind\{t_n^{(m_0)}-t_{n-1}^{(m_0)}>t_{c_1}\}\ge 1\,\big|\,t\in n^{(m_0 ) } , ( \xi_{m_0},z_{m_0})=(x , z)\big).\label{eq : phittandem}\end{aligned}\ ] ] where @xmath286 denotes the cache eviction time at the cache @xmath287 under che approximation .
the following proposition holds .
[ prop : tandem ] we have that @xmath288 and @xmath289 here @xmath290 @xmath291 and @xmath292 _ proof .
_ by and the slivnyak theorem ( see e.g. @xcite ) , we have @xmath293)=0,\nonumber\\ & \,\,\,\,\,\,\,\,\,\,\,\,\ , \sum_{n}\ind_{(t - t_{c_2},t - t_{c_1}]}(t_n^{(m_0)})\ind\{t_n^{(m_0)}-t_{n-1}^{(m_0)}>t_{c_1}\}\ge 1\,\big|\ , ( \xi_{m_0},z_{m_0})=(x , z)\big),\nonumber\end{aligned}\ ] ] and this quantity is equal to zero if @xmath294 , which proves .
otherwise , there exists an integer @xmath295 such that @xmath296 .
we consider the partition of the set @xmath297 $ ] formed by the intervals @xmath298 $ ] , @xmath299 , where the @xmath300 s are defined in the statement , and define the r.v.s @xmath301 .
since , by construction @xmath302 , for any @xmath303 , provided that @xmath304 , for some @xmath305 , and @xmath306 , then necessarily @xmath307 .
therefore , setting @xmath308)=0\}$ ] and @xmath309 we deduce @xmath310 the bonferroni inequality and the union bound yield @xmath311 for any @xmath312 , @xmath313 , we have @xmath314 where the latter equality follows noticing that , given @xmath315 , @xmath316 . by the independence of the increments of the poisson process we deduce @xmath317 and
so @xmath318 consequently , by and we have the following bounds on the hit " probability @xmath319 relation follows by and the following computation : @xmath320 @xmath66 * remark5 . * proposition [ prop : tandem ] provides the exact value of the hit " probability when @xmath321 .
as it is clear from the proof , one might exactly compute the
hit " probability even when @xmath322 by applying the inclusion - exclusion formula .
however , the resulting computational cost would be very high since one has to compute the probability of any intersection of the events @xmath323 defined by . in conclusion
, we can say that any computationally efficient approach to the performance analysis of a tandem network of caches must resort to some extra approximations ( in addition to che s approximation ) , which affect inevitably the accuracy of the method .
as shown by several recent experimental works , many video contents ( such as youtube contents ) exhibits few typical normalized temporal popularity profiles , each profile corresponding to a large class of contents with similar characteristics ( e.g. contents in the same youtube category ) @xcite .
hence , restricting the analysis to a single class @xmath3 of contents , we may assume that : @xmath1 @xmath5 represents the content volume , i.e. the total number of requests it typically attracts ; @xmath2 all contents of the class exhibit the same normalized popularity profile .
this justifies the choice of a snm with a multiplicative popularity profile such as .
recall that for this model the function @xmath81 is given by .
assuming , by and ( with @xmath274 in place of @xmath75 ) we easily have @xmath324)^{-1}\int_{0}^\infty h(u ) \phi_{z_1}'\left(-\int_{u-\theta}^{u}h(s)\,\mathrm{d}s
\right ) \mathrm{d } u,\end{aligned}\ ] ] where @xmath325 is the first order derivative of @xmath326 .
relations and provide a computationally efficient tool to estimate the hit " probability , under che s approximation , of lru caches under the snm . indeed , we may estimate @xmath75 by numerically inverting and using the relation @xmath327 . replacing @xmath274 in with such estimate of @xmath75
, we finally have an estimate of the hit " probability under che s approximation .
we now assess the accuracy of the che approximation for the evaluation of the hit " probability by describing some numerical results .
we suppose that the arrival rate of new contents @xmath328 is equal to @xmath329 units per day ; we assume that the content volume @xmath330 follows a pareto distribution with probability density @xmath331 and mean @xmath332=\frac{\alpha a}{\alpha-1}=3 $ ] ( we refer the reader to @xcite and @xcite for a practical justification on the choice of a pareto distribution ) ; we consider a multiplicative popularity profile of the form with @xmath333 , where the parameter @xmath334 has to be interpreted as the content life - span .
figures [ fig : figura1new ] and [ fig : figura2new ] report the hit " probability , as predicted by che s approximation , vs the cache size for different values of the exponent @xmath338 and the content life - span @xmath334 , respectively . for each estimate , the figures report also the interval in which the exact value of the hit " probability falls as given by proposition [ prop : error0 ] . all computations have been carried out while guaranteeing relative numerical errors smaller than @xmath339
. note that in all cases of practical relevance ( i.e. for values of the hit " probability exceeding @xmath339 ) che s approximation leads to negligible errors .
the surprisingly good degree of accuracy entailed by che s approximation , which has been already experimentally ( i.e. against simulations ) observed by several authors @xcite , is now confirmed even for the snm .
g. bianchi et al . check before storing : what is the performance price of content integrity verification in lru caching ?
, 43(3):5967 , 2013 . h. che , y. tung and z. wang .
hierarchical web caching systems : modeling , design and experimental results .
, 20(7):13051314 , 2002 . | in this paper we analyze least recently used ( lru ) caches operating under the shot noise requests model ( snm ) .
the snm was recently proposed in @xcite to better capture the main characteristics of today video on demand ( vod ) traffic .
we investigate the validity of che s approximation @xcite through an asymptotic analysis of the cache eviction time .
in particular , we provide a large deviation principle , a law of large numbers and a central limit theorem for the cache eviction time , as the cache size grows large .
finally , we derive upper and lower bounds for the hit " probability in tandem networks of caches under che s approximation . caching systems , performance evaluation , asymptotic analysis .
primary : 68m20 ; secondary : 60f10 , 60f05 . |
duality is a sort of symmetry that exists between two or more very special theories .
the remarkable fact about these kinds of symmetries is that when we make the duality transformations over the important fields , we change the regime of couplings in the theories involved . in other words ,
if we have a theory whose coupling constant is @xmath4 , it is possible to build another theory with coupling constant @xmath5 , i.e. , we can do strong coupling calculations in one model in perturbation approach if we change it by its dual .
the most basic example is the duality between the sine - gordon model and the thirring model . in this case , the solitons of the first model are exchanged by the particles of the former @xcite .
another examples are the network of dualities relating the five superstring theories and eleven - dimensional m - theory @xcite .
the most recent duality found , still in the superstring context , is the @xmath6 correspondence , where a @xmath7-dimensional type @xmath8 superstring in the @xmath9 background is conjectured to be dual to a @xmath10 super - yang mills theory in @xmath11 @xcite . in this piece of work we study tensorial fields in extra - dimensional spaces and find an intriguing characteristic that is , in some sense , related to duality
the fact is that we find evidences of an structure of resonances that have the same characteristics but in different higher dimensional spaces with different regimens of coupling constants.this last fact is the one that made us call such a behavior as a `` duality '' .
antisymmetric tensor fields or simply differential forms arise naturally in string theory @xcite and supergravity @xcite and play an important role in dualization @xcite . in particular
they appear in the @xmath12 sector of each of the type ii string theories .
these tensor fields couple naturally to higher - dimensional extended objects , the @xmath13branes and are important for their stability .
besides this , they are related to the linking number of higher dimensional knots @xcite .
the rank of these antisymmetric tensors is defined by the dimension of the manifold@xcite . beyond this , these kind of fields play an important role in the solution of the moduli stabilization problem of string theory@xcite . because of these aspects , it is important to study higher rank tensor fields in membrane backgrounds . in this direction
antisymmetric tensor fields have already been considered in models of extra dimension . generally , the @xmath0forms of highest rank do not have physical relevance .
this is due to the fact that when the rank increases , also increases the number of gauge freedom .
such fact can be used to cancel the dynamics of the field in the brane@xcite .
the mass spectrum of the two and three - form have been studied , for example , in refs . @xcite and @xcite in a context of five dimensions with codimension one .
later , the coupling between the two and three - forms with the dilaton was studied , in different contexts , in @xcite . in another direction
soliton - like solutions are studied with increasing interest in physics , not only in condensed matter , as in particle physics and cosmology . in brane models ,
they are used as mechanism of field localization , avoiding the appearance of the troublesome infinities .
several kinds of defects in brane scenarios are considered in the literature @xcite . in these papers ,
the authors consider brane world models where the brane is supported by a soliton solution to the baby skyrme model or by topological defects available in some models . as an example , in a recent paper , a model is considered for coupling fermions to brane and/or antibrane modeled by a kink antikink system @xcite .
the localization of fields in a framework that consider the brane as a kink has been studied for example in @xcite . in this context
the present authors have studied the issue of localization and resonances of a three - form field in @xcite and in a separate paper @xmath0form fields in a scenario of a @xmath1brane with codimension two @xcite . in these scenarios ,
some facts about localization of fields are known : the scalar field(@xmath14form ) is localizable , but the vector gauge fields ( @xmath15form ) , the kalb - ramond field(@xmath16form ) and the three - form field are not .
the reason why this happens with the vector field is that , in four dimensions , it is conformal and all information coming from warp factors drops out , necessarily rendering a non - normalizable four dimensional effective action .
however , in the work of kehagias and tamvakis @xcite , it is shown that the coupling between the dilaton and the vector gauge field produces localization of the vector field . in analogy with the work of kehagias and tamvakis@xcite , as cited above , the authors have also considered these coupling with a three - form field@xcite , where a condition for localization is found .
the facts above are even clear if we consider , in a general way , a @xmath0form field in an arbitrary dimension .
this can be considered with and without the dilaton coupling . in this reasoning ,
general conditions for localization can be found in an elegant way .
for example , for branes with space dimensions bigger than three , one should expect that the gauge vector field is localizable even without the dilaton coupling .
this is due to the fact that this field is not conformal in these cases .
in fact what we find is that , bigger is the space - time dimension , bigger will be the order of the @xmath0form localized without the need of the dilaton . for the other cases the dilaton coupling is needed .
the main goal of this paper is to present the structure of dual resonance spaces .
however , standard calculations related to aspects of @xmath0form fields living in the background of a codimension one @xmath1brane are necessary .
therefore , the paper is organized as follows .
section 1 is devoted to discuss a solution of einstein equation with source given by a kink . in section 2
we analyze how the gauge freedom can be used to cancel the angular component of an arbitrary @xmath0form .
yet in this section we discuss aspects of localization of the fields involved . in the third section , we study massive modes , resonances and show aspects of dual spaces containing same resonance structures . at the end , we discuss the conclusions and perspectives .
we start our analysis studying the space - time background . before analyzing the coupling of the @xmath0forms with the dilaton ,
it is necessary to obtain a solution of the equations of motion for the gravitational field in the background of the dilaton and the membrane . for such
, we introduce the following action , similar of the one used in @xcite : @xmath17,\ ] ] where @xmath18 , @xmath19 is the metric determinant and @xmath20 is the ricci scalar .
note that we are working with a model containing two real scalar fields .
the field @xmath21 plays the role of membrane generator of the model while the field @xmath22 represents the dilaton .
the potential function depends on both scalar fields .
it is assumed the following ansatz for the spacetime metric : @xmath23 where @xmath24 is the metric of the @xmath25-brane , @xmath26 is the codimension coordinate , and @xmath27 is a deformation parameter .
the deformation parameter controls the kind of deformed topological defect we want in order to mimic different classes of membranes .
the deformation method is based in deformations of the potential of models containing solitons in order to produce new and unexpected solutions @xcite . as usual ,
capital latin index represent the coordinates in the bulk and greek index , those on the @xmath25-brane .
the equations of motion are given by @xmath28 @xmath29 @xmath30 and @xmath31 in order to solve this system , we use the so - called superpotential function @xmath32 , defined by @xmath33 , following the approach of kehagias and tamvakis @xcite .
the particular solution regarded follows from choosing the potential @xmath34 and superpotential @xmath35 as @xmath36 and @xmath37 where @xmath38 and @xmath39 are parameters to adjust the dimensionality .
as pointed out in @xcite , this potential give us the desired soliton - like solution.in this way it is easy to obtain first order differential equations whose solutions are solutions of the equations of motion above , namely @xmath40 @xmath41 and @xmath42 the solutions for these new set of equations are the following : for @xmath43 : @xmath44 @xmath45 and for @xmath46 : @xmath47 @xmath48- \sum_{n=1}^{s-1}\frac1{2n}\tanh^{2n}\left(\frac{ay}{s}\right)\biggr{\ } } \nonumber\end{aligned}\ ] ] an important observation to be made here is concerned to the appearance of a spacetime singularity due to the dilaton field .
the ricci scalar is given by @xmath49 and the plot is given in the figure below ( lined ) , @xmath50 ( dashed ) and @xmath51 ( dotted ) . ] this is a naked singularity emerging at the infinity and in principle could jeopardize the issue of fields localization .
however this singularity can be understood from a higher dimensional viewpoint . as an example , a @xmath52-brane in type ii - a theory is singular because the presence of a diverging dilaton field . however , that singularity disappears when it is lifted to the eleventh dimensional supergravity viewpoint . here
, we changed our solution to six dimensions @xmath53 the dilaton turns out to be the radius of an extra @xmath54 direction ( that is why it is called dilaton ) .
the above higher dimensional metric is free of singularities .
this interpretation also explain why , in this manuscript , we consider the same coupling between all q - forms and the dilaton .
this kind of solutions have been studied , for example , in @xcite and @xcite .
in fact , we used the string like solution of @xcite to analyze the localization q - forms in co - dimension two brane word @xcite .
although the objective of the present manuscript is not to deal with this singularity problem , this issue is widely discussed in the literature .
it has been argued that the imposition of unitary boundary conditions renders these singularities harmless from the physical viewpoint @xcite,@xcite , @xcite,@xcite . besides this aspect , there are reasons to believe that considering these kind of spaces as physically meaningful
. there are proposals @xcite that five - dimensional bulk gravity in the thin domain - wall case has an equivalent description in terms of a cutoff four - dimensional conformal field theory on the domain - wall , on the lines of the @xmath6 correspondence @xcite . from the viewpoint of @xmath6 correspondence ,
some singularities may have physical significance and it is very interesting to see the role the dilaton field plays in this framework @xcite .
in this section we study @xmath0form fields in the gravitational background of the last section .
we consider these fields with and without the dilaton coupling .
it is important to note that , when we exclude this coupling , the metric do not has the @xmath55 factor . with these considerations , we must look for localization of forms in these frameworks . the first important thing to be understood is the gauge symmetry of these fields .
the fact is that , when the number of dimension increases , the number of gauge freedom also increases , and this can be used to cancel the degrees of freedom in the visible brane .
this analysis was performed in details by the authors in a recent paper @xcite . due the importance of this for our present work ,
we give here a sketch of it .
first , note that the number of degrees of freedom ( d.o.f . ) of an antisymmetric tensor field with @xmath56 indices is given by @xmath57 for the one form we have @xmath58 where @xmath21 is a scalar field .
therefore we have for the number of physical d.o.f .
@xmath59 in the above expression we have used the stiefel relation . after this we claim that the number of physical d.o.f . of a @xmath0form
is given by @xmath60 and proves that this is valid for a @xmath61form .
in this case the gauge parameter will be an @xmath0form and therefore , using the stiefel relation , we have for the physical d.o.f .
of the @xmath61form @xmath62 the d - form has no dynamics . using gauge symmetries , from the above result , we can see that the d.o.f . of the @xmath63 form can be made all null at the visible brane .
therefore we must analyze only the cases @xmath64 .
the full action for this field is given by @xmath65,\ ] ] where @xmath66}$ ] is the field strength for the @xmath56-form @xmath67 .
we can use gauge freedom to fix @xmath68 and we are left with the following two type of terms @xmath69},\ ] ] @xmath70 = \partial_y x_{\ind{\mu_1}{\mu_q}}=x'_{\ind{\mu_1}{\mu_q}}.\ ] ] using the above facts we obtain for the action @xmath71\},\end{aligned}\]]and the equation of motion are given by @xmath72=0\end{aligned}\ ] ] we can now separate the @xmath26 dependence of the field using @xmath73 defining @xmath74 , where @xmath75 stands for the four dimensional field strength , we get @xmath76 and @xmath77 in the above expression @xmath78 is a constant of separation of variable that represents the mass of @xmath56-form @xmath55 in the @xmath25-brane .
it is easy to see that @xmath79 , with @xmath80 being a constant , solves the above equation for @xmath81 . in this particular case
, the effective action can be found easily to give us @xmath82.\ ] ] using now our solution for @xmath83 we have that the integration in the extra dimension is finite for @xmath84 .
it becomes clear now that only the scalar field is localized for @xmath85 .
we also see that for @xmath86 , for example , the vector field is localized . as commented in the introduction , this was expected because this field is not conformal in this space dimension . despite of this , the dilaton is yet needed for the localization of higher order forms .
the coupling is inspired in string theory and we have for the action @xmath87,\ ] ] with the same definitions used above . repeating the same steps as before we obtain @xmath88\}.\end{aligned}\ ] ] the equation of motion takes the form @xmath89=0,\end{aligned}\ ] ] and performing the same separation of variable we get @xmath90 again we can see that @xmath79 , with @xmath80 being a constant , solves the above equation for @xmath81 .
the effective action can again be found easily and is given by @xmath91.\ ] ] using now our solution for @xmath83 , @xmath92 and @xmath93 we have that , when @xmath94 , the integration in the extra dimension is finite for all @xmath56-forms in any @xmath25-brane .
it is important to stress that for @xmath85 , the above expression reproduces the same conditions found for the one , two and three form found in the literature .
for the massive modes , the best way to make the analysis is to transform the equation ( [ udilaton ] ) in a schrdinger type equation .
it is easy to see that an equation of the form @xmath95 can be put in a schrdinger form @xmath96 through the transformations @xmath97 with @xmath98 and @xmath99 where the prime is a derivative with respect to @xmath26 . now using the equation ( [ udilaton ] ) we obtain @xmath100 where @xmath101 , and we take @xmath26 of function of @xmath102 through @xmath103 .
we show bellow the graphic of @xmath104 by considering @xmath105 and @xmath106 for various @xmath107 values .
( lined ) scaled by 1/10 , @xmath50 ( dashed ) , @xmath51 ( dotted ) for @xmath108 ( top ) and @xmath109 ( bottom ) . ] in order to analyze resonances , we must compute transmission coefficients(@xmath110 ) numerically , which gives a clear physical interpretation about what happens to a free wave that interact with the membrane .
the idea of the existence of a resonant mode is that for a given mass the potential barrier is transparent to the particle , i.e. , the transmission coefficient has a peak at this mass value .
that means that the amplitude of the wave - function has a maximum value at @xmath111 and the probability to find this kk mode inside the membrane is higher . in order to obtain these results numerically ,
we have developed a computational code to compute transmission coefficients for the above potential profile .
a more extensive and detailed analysis of resonances with transmission coefficients will be given in a separate paper by the authors @xcite . in fig .
[ fig2 ] we give the plots of @xmath112 for @xmath113 and considering the case @xmath114 ( top ) , @xmath50 ( middle ) , @xmath51 ( bottom ) for @xmath108 . ] as mentioned before , we find resonances for this value of @xmath107 only for @xmath43 . in fig .
[ fig3 ] we show the plot for @xmath109 ( top ) , @xmath50 ( middle ) , @xmath51 ( bottom ) for @xmath109 ] and we can see how this alter the existence of resonances . for this values
we have an interesting structure of resonances , witch is similar to that found in quantum mechanical problems .
as we stress in our separate paper , the existence of resonances depends strongly of the shape of the potential . in our case , therefore , the resonance is driven by the choice of the parameter @xmath107 .
the case for @xmath85 is doubtless the more important one and several issues deserve a careful analysis .
first , the generalization employed here simplifies the problem and clearly shows how the condition @xmath94 , necessary for localization of the zero mode , changes for different @xmath25 and @xmath56 . for comparison , previous articles found this condition for @xmath115 separately @xcite@xcite@xcite .
therefore , we gain a considerable simplification and understanding for all those cases with our generalization .
moreover , we found resonances that was never observed before in @xcite and @xcite .
this is mainly due the fact that we realized that the resonances depends strongly of the parameter @xmath107 . in @xcite
the authors study a one form and found only one resonance peak . by changing the parameter we discovered much more resonant modes fig .
[ fig2 ] , indicating that the higher the parameter more resonances we get .
the same study was performed by the same authors in ref .
@xcite , but for the two form .
they founded the localization condition for the zero mode , that agrees with our expression .
after this , they studied the possibility of resonances and found just one resonance peak . again , with our generalization , besides the peak found by them we found a lot of other peaks when we change the parameter @xmath107 as in fig .
the case @xmath116 was presented in a recent paper by the present authors@xcite , where the rich structure of resonances was first found .
it is important to note here that all analysis pointed above can be visualized in the two graphics given below fig .
[ fig2 ] and fig .
this was only possible due the general expression for @xmath107 that we are presenting here . as a byproduct
the present authors has observed that for a fixed value of @xmath107 and @xmath117 , a @xmath56-form in a @xmath25-brane has the same resonance structure of a @xmath118-from in a @xmath119-brane if @xmath120 .
all graphics above are plotted for @xmath121 and are the same as that with a three - form in 3-brane @xcite .
also for a fixed @xmath117 and @xmath56 the resonance peaks increase for larger values of the dimension of the brane .
there are an interesting situation with @xmath107 constant .
we can have models with small values of @xmath117 displaying the same resonance structure .
we see in table bellow some values of @xmath25 , @xmath56 and @xmath122 with @xmath109 .
.in the left @xmath123 , @xmath124 and in the rigth @xmath125 , @xmath126 .
[ cols="^,^",options="header " , ] as we can see , we have dual spaces of resonance(dsr ) : they have the same resonance structure .
it is interesting to note that these dsr admits situations with large and low coupling constant .
this is similar to the @xmath127 duality of string theory .
the difference is that here we have an infinite class of spaces , with a certain @xmath117 , which has the same resonant mass peaks as another infinite class , with a different @xmath117 .
it is not clear to the authors if there is some structure or symmetry behind these dsr that could explain why they appear .
we let this question for a future work .
in this paper we have studied the issue of localization of @xmath0form fields in a @xmath1brane embedded in a @xmath2 space - time .
the membrane was described by a kink , a codimension one topological defect , and related topological objects called deformed defects .
furthermore , we have considered a gravitational background where the dilaton field plays an important role .
we have obtained the localization of the zero modes of several antisymmetric fields regarding specific peculiarities for each dimension .
for example , for branes with space dimensions bigger than three , we found that the gauge vector field is localizable even without the dilaton coupling .
this is due to the fact that this field is not conformal in these cases . for other types of forms ,
the dilaton coupling is still needed .
we have calculated massive states through the related quantum mechanical problem and have analyzed the appearance of possible resonances . in this case
we have made use of numerical computations of transmission coefficients through the membrane .
the existence of resonances depends strongly of the shape of the potential , as is well known .
we have been able to find several peaks of resonant modes for specific parameters that models the potential and for specific dimensionality of the spacetime . in particular , as an important and intriguing result , we have found a class of resonances that have the same behavior , but living in spacetimes with different dimensions and with different dilaton coupling constant .
because of this last fact , we call this set of resonances in this model as dual spaces of resonances(dsr ) . however , it is not at all understood how this appears from the usual way to treat dualities in field theories .
we let these sort of discussions for a future work .
we would like to thank : prof .
francisco pimentel for usefull discussions , and the laboratrio de leos pesados for the hospitality .
we also acknowledge the financial support provided by fundao cearense de apoio ao desenvolvimento cientfico e tecnolgico ( funcap ) , the conselho nacional de desenvolvimento cientfico e tecnolgico ( cnpq ) and funcap / cnpq / pronex .
j. m. maldacena , adv .
* 2 * ( 1998 ) 231 [ int .
j. theor . phys .
* 38 * ( 1999 ) 1113 ] [ arxiv : hep - th/9711200 ] .
j. polchinski , `` string theory .
1 : an introduction to the bosonic string , '' spires entry _ cambridge , uk : univ . pr .
( 1998 ) 402 p _
p. van nieuwenhuizen , phys .
rept . * 68 * , 189 ( 1981 ) .
a. smailagic and e. spallucci , phys .
d * 61 * , 067701 ( 2000 ) [ arxiv : hep - th/9911089 ] .
a. smailagic and e. spallucci , phys .
b * 489 * , 435 ( 2000 ) [ arxiv : hep - th/0008094 ] .
i. oda and s. yahikozawa , `` linking numbers and variational method , '' phys .
b * 238 * , 272 ( 1990 ) .
m. nakahara , `` geometry , topology and physics , '' _ boca raton , usa : taylor _ & _ francis ( 2003 ) 573 p _ s. kachru , m. b. schulz and s. trivedi , `` moduli stabilization from fluxes in a simple iib orientifold , '' jhep * 0310 * , 007 ( 2003 ) [ arxiv : hep - th/0201028 ]
. i. antoniadis and t. maillard , `` moduli stabilization from magnetic fluxes in type i string theory , '' nucl .
b * 716 * , 3 ( 2005 ) [ arxiv : hep - th/0412008 ] .
v. balasubramanian , p. berglund , j. p. conlon and f. quevedo , `` systematics of moduli stabilisation in calabi - yau flux compactifications , '' jhep * 0503 * , 007 ( 2005 ) [ arxiv : hep - th/0502058 ] .
s. s. gubser , phys .
d * 63 * ( 2001 ) 084017 [ arxiv : hep - th/9912001 ] .
s. b. giddings , e. katz and l. randall , jhep * 0003 * , 023 ( 2000 ) [ arxiv : hep - th/0002091 ] .
m. gremm , phys .
d * 62 * , 044017 ( 2000 ) [ arxiv : hep - th/0002040 ] . t. gherghetta and m. e. shaposhnikov , phys .
lett . * 85 * , 240 ( 2000 ) [ arxiv : hep - th/0004014 ] .
g. alencar , r. r. landim , m. o. tahim and k. c. mendes , `` antisymmetric tensor fields in codimension two brane - world , '' arxiv:1009.1183 [ hep - th ] .
a. herrera - aguilar , d. malagon - morejon , r. r. mora - luna and u. nucamendi , mod .
a * 25 * , 2089 ( 2010 ) [ arxiv:0910.0363 [ hep - th ] ] .
a. g. cohen and d. b. kaplan , phys .
b * 470 * , 52 ( 1999 ) [ arxiv : hep - th/9910132 ] .
m. gremm , phys .
b * 478 * , 434 ( 2000 ) [ arxiv : hep - th/9912060 ] .
b. mukhopadhyaya , s. sen , s. sen and s. sengupta , `` bulk kalb - ramond field in randall sundrum scenario , '' phys .
d * 70 * , 066009 ( 2004 ) [ arxiv : hep - th/0403098 ] b. mukhopadhyaya , s. sen and s. sengupta , `` bulk antisymmetric tensor fields in a randall - sundrum model , '' phys .
d * 76 * , 121501 ( 2007 ) [ arxiv:0709.3428 [ hep - th ] ] .
g. de risi , `` bouncing cosmology from kalb - ramond braneworld , '' phys .
d * 77 * , 044030 ( 2008 ) [ arxiv:0711.3781 [ hep - th ] ] .
b. mukhopadhyaya , s. sen and s. sengupta , `` a randall - sundrum scenario with bulk dilaton and torsion , '' phys . rev .
d * 79 * , 124029 ( 2009 ) [ arxiv:0903.0722 [ hep - th ] ] .
d. bazeia , a. r. gomes , l. losano and r. menezes , `` braneworld models of scalar fields with generalized dynamics , '' phys .
b * 671 * , 402 ( 2009 ) [ arxiv:0808.1815 [ hep - th ] ] .
d. bazeia , a. r. gomes and l. losano , `` gravity localization on thick branes : a numerical approach , '' int .
j. mod .
a * 24 * , 1135 ( 2009 ) [ arxiv:0708.3530 [ hep - th ] ] .
d. bazeia , f. a. brito and a. r. gomes , `` locally localized gravity and geometric transitions , '' jhep * 0411 * , 070 ( 2004 ) [ arxiv : hep - th/0411088 ] .
d. bazeia , c. furtado and a. r. gomes , `` brane structure from scalar field in warped spacetime , '' jcap * 0402 * , 002 ( 2004 ) [ arxiv : hep - th/0308034 ] .
h. -t .
li , y. -x .
liu , z. -h .
zhao _ et al .
_ , phys .
rev . * d83 * , 045006 ( 2011 ) .
[ arxiv:1006.4240 [ hep - th ] ] .
i. olasagasti and a. vilenkin , `` gravity of higher - dimensional global defects , '' phys .
d * 62 * , 044014 ( 2000 ) [ arxiv : hep - th/0003300 ] .
a. kehagias and k. tamvakis , `` localized gravitons , gauge bosons and chiral fermions in smooth spaces generated by a bounce , '' phys .
b * 504 * , 38 ( 2001 ) [ arxiv : hep - th/0010112 ] .
d. bazeia , l. losano and j. m. c. malbouisson , `` deformed defects , '' phys . rev .
d * 66 * , 101701 ( 2002 ) [ arxiv : hep - th/0209027 ] . | in this work we consider @xmath0form fields in a @xmath1brane embedded in a @xmath2 space - time .
the membrane is generated by a domain wall in a randall - sundrum - like scenario .
we study conditions for localization of zero modes of these fields .
the expression agrees and generalizes the one found for the zero , one , two and three - forms in a @xmath3brane . by a generalization
we mean that our expression is valid for any form in an arbitrary dimension with codimension one .
we also point out that , even without the dilaton coupling , some form fields are localized in the membrane .
the massive modes are considered and the resonances are calculated using a numerical method .
we find that different spaces have identical resonance structures , which we call dual spaces of resonances(dsr ) . |
the determination of the signal template of the radiating companion of a black hole is of great importance for the forthcoming interferometric gravitational observatories . an inspiralling neutron star or a several solar - mass black hole orbiting a massive black hole , as well as a particle constituent contributing to the radiation of the debris trapped by the black hole are notable examples of these companions . either is the case , one wants to describe the evolution of the orbit of a test particle in the neighborhood of a black hole under the influence of gravitational radiation backreaction . in a highly idealized picture ,
the orbit of the particle is a carter geodesic@xcite characterized by four constants of the motion , the energy , the total angular momentum , the rest - mass of the particle and the separation constant . in the spirit of perturbation theory , the radiation backreaction effects may be taken into account by evolving these constants of the motion , thus picturing the trajectory of the particle by a sequence of geodesic orbits .
the history of the orbit may be divided in separate epoches .
for example , the first thing that happens to a particle trapped by a nonrotating _ schwarzschild _ black hole on an arbitrary but distant orbit is that it loses eccentricity due to radiation losses , and the radius shrinks until the close region of the last stable circular orbit is approached .
there one has to replace the adiabatic picture with one valid in the strong field regime . the description of a generic orbit in the strong field region of a spinning _ kerr _
black hole is difficult ( as has been pointed out by thorne@xcite ) due to the special nature of the separation constant which lacks a straightforward geometrical interpretation .
recently , progress in treating the separation constant has been reported by ori@xcite .
there is a considerable ease in the description of special orbits such as equatorial or quasicircular , and in fact , several papers have obtained the signal templates for such special orbits . in order that any of these special orbits can be credited as significant contributor to the gravitational signal , one has to make sure that the carter constant evolves to the envisaged special values .
several authors have recently published results on radiative losses by systems of spinning masses .
kidder , will and wiseman@xcite formulate the problem by using the lagrangian ( proposed by barker and oconnell@xcite ) for bodies with finite masses and spins .
they and kidder @xcite compute the spin - orbit and spin - spin contributions to the momentary radiative loss of energy , linear and angular momenta . in the test particle
limit their results have subsequently been used by shibata@xcite , in his study of equatorial orbits , and by ryan@xcite . by computing the backreaction forces
, ryan obtains the power and angular momentum losses for circular and generic orbits . in our work
, we have taken up the evolution of the orbit in the far field region of the black hole .
the trajectory of the idealized , nonradiating particle is described by the lense - thirring picture@xcite .
this picture emerges in the test particle limit of the scenario described in @xcite .
the spin @xmath5 of the black hole is fixed to the @xmath6 axis .
the orbit is viewed approximately as an ellipse precessing in a plane which rotates about @xmath7 .
the motion is characterized by three constants : the energy @xmath8 , the magnitude @xmath9 and spin projection @xmath10 of the orbital angular momentum , and by two slowly changing angle parameters @xmath2 and @xmath11 .
the lense - thirring approximation is linear in the spin terms , thus the ( additive ) schwarzschild effects may be neglected in this treatment . in an earlier paper@xcite ,
we have computed the radiation losses from the mass quadrupole tensor in the lense - thirring approximation , by the orbit smoothing method of landau and lifshitz@xcite .
however , one may be rightly concerned about the degradation of accuracy of this computation , as a price of the relative simplicity introduced by the smoothing method . in the present paper , we drop the orbit smoothing , and we include the velocity quadrupole tensor @xmath12 in the description of the radiative losses . in the test particle limit , this ( formally ) corresponds to the 1pn approximation@xcite .
the rest of the paper is organized as follows . in sec .
2 we describe the orbital motion in the lense - thirring approximation . in terms of a set of new angular variables ( related to euler rotations ) ,
the equations of motion take a particularly simple form : all time derivatives depend on the radial variable @xmath13 alone . in sec .
3 , we explore the advantages of having a pure radial equation of motion .
we find the turning points at @xmath14 which allow us to parameterize the radial motion by trigonometric functions .
integrations over the time can be replaced by integrations over the parameter , which are straightforward in principle , but can be messy in practice .
we evade such complications by finding suitable parametrizations .
the first parameter @xmath15 we introduce is a generalization of the eccentric anomaly of the keplerian orbits .
it is employed for computing the half period defined as the time spent by the particle between consecutive turning points . for the purpose of averaging the radiative losses , and for determining the change in the new angular variables , a generalized true anomaly parameter @xmath16
is properly defined . in terms of this parameter ,
when integration by the residue theorem is performed , the only pole of the integrands will be in the origin . in sec .
4 , we compute the rates of change of the constants of motion due to radiation backreaction , to first order in @xmath7 , by employing the radiative multipole tensors of kidder , will and wiseman@xcite , which originate in the blanchet - damour - iyer formalism@xcite . as we are interested in the cumulative effects of the secular motion , we compute averaged losses over the period of radial motion .
we give the losses in an unambiguous manner in terms of the constants of motion @xmath17 and @xmath10 .
as has been pointed out by ryan @xcite , the definition of the orbit parameters @xmath18 and @xmath19 is subject to ambiguities . in sec . 5
, we define the semimajor axis and eccentricity . with advantages explained in the main text ,
our definitions differ both from ryan s and the ones used in the lense - thirring paper .
we rewrite all losses in terms of our orbital parameters @xmath20 and @xmath19 . in the concluding remarks ,
we complete the characterization of the changes caused by gravitational radiation by giving the change in the @xmath3 angle parameters over one orbit .
we then compare our results with those of other authors .
we wish to make some comments on the expansion procedure we have adopted .
earlier works employ either the expansion parameter @xmath21 or the inverse @xmath22 of the speed of light .
since we eliminate @xmath23 and @xmath24 by use of the first integrals of the motion , we choose to retain @xmath25 and @xmath26 for bookkeeping of the orders in our formalism .
the correction terms in the lagrangian , in the equations of motion and radiation losses will carry an extra @xmath27 factor . in our computation , we keep only the terms containing the spin @xmath0 from among these correction terms , thus we may as well choose @xmath0 as our expansion parameter .
our treatment of radiation backreaction effects can be generalized beyond the lense - thirring approximation . in a follow - up paper@xcite
we will apply our averaging method and parameterization for computing the backreaction effects on a bound system of two finite masses .
we consider the orbit of a small particle of mass @xmath28 about a body with mass @xmath29 and angular momentum vector @xmath30 . in the lense - thirring approximation@xcite
, the lagrangian of the system has the form@xcite @xmath31 where @xmath32 , a dot denotes @xmath33 and the term describing ( perturbative ) rotation effects is @xmath34 the cartesian coordinates of the test particle are @xmath35 , with origin chosen at the black hole .
hence the orbital momentum @xmath36 and runge - lenz vector @xmath37 satisfy the equations of motion , respectively @xmath38 @xmath39 here @xmath40 is the momentum of the orbiting mass . note that the angular momentum @xmath41 undergoes a pure rotation about the spin with angular velocity : @xmath42 from ( [ l ] ) and ( [ a ] ) , the vector @xmath43 can be written in terms of the basis @xmath44 in the plane with normal @xmath41 : @xmath45 where @xmath46 . inserting this expression of @xmath47 in ( [ adot ] )
, the direction @xmath48 of the runge - lenz vector is found to rotate about a linear combination of the vectors @xmath7 and @xmath41 : @xmath49\times{{\bf a}\over a } \ .\label{averst}\ ] ] thus the direction vector rotates about @xmath7 with the angular velocity @xmath50 ( as does @xmath41 ) and about @xmath41 with the angular velocity : @xmath51 although the orbital momentum and the runge - lenz vector are not conserved ( unlike in the zeroth order in @xmath7 ) , here also there are constants of motion . because of the fact that the potential is axisymmetric and stationary , the orbital momentum component @xmath10 and the energy @xmath52 are conserved .
since the motion of @xmath41 is a pure rotation , its magnitude @xmath9 is also constant : @xmath53 the constants of the motion are @xmath54}-{g \mu m\over
r}\label{e}\\ \label{eq : lz } l_z&=&\mu r^2\sin^2\theta\ \dot\varphi-2\frac{g\mu}{c^2r}s\sin^2\theta \label{lz } \\ \label{eq : l2 } l^2&=&{\mu^2}r^4 ( \dot\theta^2+\sin^2\theta\ \dot\varphi^2)-4{g\mu^2s\over c^2}r\sin^2\theta\ \dot\varphi \ .
\label{l2 } \\end{aligned}\ ] ] ( here we are using polar coordinates , @xmath55 , @xmath56 , @xmath57 ) . as a consequence , the angle @xmath58 subtended by @xmath41 and @xmath7
is also conserved : @xmath59 the first integrals of the equations of motion follow from ( [ eq : e])-([eq : l2 ] ) by simple algebra ( we need not recourse to the hamilton - jacobi formalism ) .
thus the equation of the radial motion takes the first - order form , decoupled from the angular degrees of freedom : @xmath60 the equations for the angular variables @xmath61 , however , are coupled among themselves and with the @xmath13 motion : @xmath62 there is an alternative description of the motion in the lense - thirring approximation which bears a more intimate geometrical relation to the picture provided by perturbation theory .
we shall introduce this formalism by performing time - dependent euler rotations about the origin so as to place the kepler ellipse in its momentary orientation : @xmath63 here @xmath64 is the initial position of the particle on the plane perpendicular to @xmath41 ( although @xmath41 is different from @xmath65 , the condition @xmath66 is fulfilled ) .
a rotation about the cartesian @xmath67 and @xmath6 axes is given , respectively , by @xmath68 the coordinates of the particle are written in terms of the euler angles @xmath69 and @xmath70 as @xmath71 hence we get the relations @xmath72 which allow us to express the time derivatives of the new variables : @xmath73 note that @xmath74 and @xmath75 carries the interpretation of angular velocity of rotation about @xmath7 and in the plane perpendicular to @xmath41 respectively .
an important feature of the new coordinates @xmath76 is that their derivatives appear linearly in the equations of motion ( [ psit ] ) and ( [ phit ] ) .
this is not the case with the polar coordinates ( [ polar ] ) because the root of @xmath77 can not be extracted .
the geometrical interpretation of the angle variable @xmath78 assures that it is monotonously changing .
we choose @xmath70 to increase with the motion evolving .
there is a freedom in the precise way one introduces the three euler - angle variables in place of the polar angles .
even though the equations of motion are made simple by use of the angular variables ( [ newvar]-[newvar2 ] ) , the description of the momentary plane of orbit is tied to a different choice of the euler angles @xmath79 , where the initial @xmath6 axis is chosen along @xmath80 rather than along @xmath41 .
these angular variables appear ( with the appropriate change in notation ) in the lense - thirring paper .
the angle @xmath81 is not a constant , and the resulting equations are less simple .
the primed angles will not be employed in our computations , and we relegate their discussion to the appendix .
we now proceed with the solution of the equations of motion .
first consider the radial equation ( [ radial ] ) .
for the turning points @xmath14 , we get the cubic equation @xmath82 in the no - spin limit , @xmath83 , the solutions are ( the unphysical ) @xmath84 and @xmath85 where @xmath86 is the length of the runge - lenz vector to zeroth order : @xmath87 when a small spin term is present , the roots of ( [ cubic ] ) will have a slightly different form from ( [ rpm ] ) : @xmath88 and @xmath89 , where @xmath90 and @xmath91 are assumed to be small . note that the constants @xmath17 and @xmath10 will not coincide with their zeroth order limits .
the constant @xmath86 is still defined by ( [ a0 ] ) , although it is not identical with the length @xmath92 of the runge - lenz vector ( which is not constant ) .
writing the cubic equation as @xmath93 , we get @xmath94r+2e\mu[-r^2(\epsilon_++\epsilon_- + \epsilon)+r(\epsilon_+r_- + \epsilon_-r_++\epsilon(r_++r_- ) ) -\epsilon r_+r_-]=0\ .\ ] ] the first term is the zeroth - order part of the original equation , written in terms of the roots , thus it equals @xmath95 . inserting the chosen form of the roots we get @xmath96 the orbit
is supposed to lie at large distance , thus the first root will not occur . using the above results , a parameter @xmath15 generalizing the eccentric anomaly of the kepler orbits
is readily introduced : @xmath97 such that @xmath98 is at @xmath99 and @xmath100 , respectively . from here : @xmath101 for equatorial orbits , this parametrization has been employed by shibata@xcite .
expressing @xmath102 from ( [ radial ] ) and linearizing in @xmath0 , the expression for @xmath103 is found : @xmath104 integration of ( [ txi ] ) from 0 to @xmath105 gives the orbital period : @xmath106 however the parametrization ( [ eq : xi ] ) is inconvenient for carrying out the averaging of the losses , because it yields a complicated array of poles of the integrands .
hence in the sequel we will , in place of ( [ xi ] ) , find a parametrization @xmath107 satisfying the following two criteria : @xmath108 where @xmath109 are constants .
that is , we want to keep the properties of the true anomaly parametrization of the kepler orbit .
property ( b ) generalizes kepler s second law for the area@xcite .
the unique parametrization satisfying both ( a ) and ( b ) is : @xmath110 hence @xmath111r^2\sin\chi\ ] ] and to first order in @xmath0 : @xmath112 \ . \label{tchi}\ ] ] we will need also the relation of the type @xmath113 .
this can be obtained from integration of ( [ psit ] ) : @xmath114 the constant of integration @xmath2 measures the angle subtended by the semiminor axis ( @xmath115 ) and the node line at @xmath116 . in a perturbative picture ,
the orientation of the momentary ellipse of the orbit is reinterpreted at each revolution , and the value of the integration constant is subject to a corresponding shift .
our parametrization ( [ chi ] ) makes especially simple the integration over one period of all expressions @xmath117 containing in the denominator no @xmath70 dependence and no other @xmath13 dependence than @xmath118 , with @xmath119 a positive integer .
note that ( [ omegal ] ) , ( [ psit ] ) and ( [ phit ] ) have this property .
further , as will be shown in the next section , the instantaneous radiative losses of the constants of motion share this property . in all of these cases one has : @xmath120 where the @xmath13 dependence of the denominator of the integrand is especially simple : @xmath121 .
the _ average _ over one period is @xmath122 with @xmath123 given by eq .
( [ period ] ) .
the integrals may conveniently be evaluated by use of the residue theorem .
one introduces the complex variable @xmath124 , and the integral is taken over the unit circle in the @xmath125 plane .
the value of the integral is given by the sum of the residues of the poles inside the circle .
in all relevant cases our parametrization assures that the only pole is at @xmath126 .
simple examples are the integrations of ( [ phit ] ) , ( [ psit ] ) and ( [ omegal ] ) : @xmath127 here ( [ dephi ] ) is the precession angle of the orbital angular momentum @xmath128 about @xmath7 .
the interpretation of the variable @xmath70 as the polar angle in the plane perpendicular to @xmath41 is in accordance with the identical first order terms in ( [ eq : depsi ] ) and ( [ eq : shift ] ) .
the ellipse is found to counterrotate ( for @xmath129 .
the periastron shift is a combination of ( [ eq : dephi ] ) and ( [ eq : shift ] ) .
both @xmath130 and @xmath131 will be unchanged in the averaged - motion approximation of landau and lifsitz @xcite .
our angles are related to the polar angle @xmath132 for equatorial orbits ( @xmath133 ) by @xmath134 .
this relation helps comparison with shibata s @xcite expression ( 2.19 ) for the periastron shift @xmath135 .
ryan@xcite has introduced a parametrization @xmath136 which , when rewritten in terms of the constants of motion @xmath17 and @xmath10 ( employing his definitions for the orbital parameters ) , will read : @xmath137 this differs from our parametrization ( [ chi ] ) by the last @xmath138 term , which shows that ryan s parameter @xmath139 and our @xmath16 reach the turning points simultaneously .
in fact both ( [ xi ] ) and ( [ chi ] ) can be supplemented by arbitrary terms involving a @xmath140 squared factor , and they still will satisfy ( a ) . however none of these parameters , and among those ryan s , will satisfy the generalized second kepler law ( b ) .
the mass and current quadrupole tensors of a test particle with coordinates @xmath141 are @xmath142 @xmath143^{stf } + { \frac { 3 \mu}{2 m } } ( r_i s_j)^{stf } \ , \ ] ] where stf means symmetrization and trace removal in the free indices .
the rates of the energy and the angular momentum losses in the test particle 1pn approximation are , respectively@xcite , @xmath144 @xmath145 evaluating the right - hand sides by use of ( [ radial ] ) , ( [ psit ] ) and ( [ phit ] ) and dropping higher - order terms in @xmath0 , the total power is @xmath146 this is independent of the angles @xmath70 and @xmath147 .
the loss in @xmath9 is evaluated by @xmath148 and it has the form @xmath149 the momentary loss in @xmath10 is @xmath150\sin^2\iota+3\,r\dot r \ , l\mu\,\sin2\psi\sin^2\iota \left ( 2\,gm\mu^{2}r-3\,l^{2}+6\,e\mu\,r^{2}\right ) \nonumber\\ & & -6\,g^{2}m^{2}\mu^{4}r^{2}+22\,gl^{2}m\mu^{2}r-24\,\mu^{3}er ^{3}gm-36\,\mu\,er^{2}l^{2 } \ } \ .
% -{\frac { 8\,g^{3}l^{3}m^{2}\cos\iota}{45\,\mu\,r^{6}c^{7 } } } \ .\end{aligned}\ ] ] after parameterizing by @xmath16 and averaging over the period @xmath123 as described in the previous section , we obtain @xmath151 among these losses averaged over one period of the radial motion , the only quantity depending on the initial angle @xmath2 is @xmath152 .
this is to be interpreted such that the rate of loss of the angular momentum component @xmath10 varies with the position of the periastron . in ( [ eq : avelosses ] ) , the terms with @xmath153 average to zero when the precession time scale is short compared to the radiation reaction time scale .
( for a more extensive discussion of this subject , _ cf
. _ @xcite . )
the major axis @xmath18 and the eccentricity @xmath19 of the kepler motion are constants and together with @xmath58 they determine the full set of the constants of motion @xmath17 and @xmath10 . in the lense - thirring picture of a perturbed kepler motion @xmath17 and @xmath10 contain first - order terms , and the corresponding constants @xmath18 and @xmath19 should also contain first - order terms . there is an ambiguity in how to introduce these orbit parameters , as noted by ryan@xcite .
he chooses these parameters by referring to the quasicircular orbits in the kerr metric .
this method is passing when wishing to approximate carter orbits .
however , there appears to be no reason for invoking phenomena tied to black holes when describing arbitrary weakly bound spinning bodies .
in fact , we want to keep our treatment general enough to cover all axially symmetric gravitational fields with first order contributions from the spin - orbit interaction term in the lagrangian ( [ delag ] ) .
thus we choose as the definition of the semimajor axis @xmath18 and eccentricity @xmath19 : @xmath154 which implies : @xmath155 these relations can be inverted : @xmath156\ .\label{ee}\\end{aligned}\ ] ] the averaged losses in terms of @xmath18 and @xmath19 then read : @xmath157 differentiating ( [ iota ] ) one has : @xmath158\ .\ ] ] those terms in the brackets which are zero - order in @xmath0 , cancel .
the averaged losses of the ellipse parameters can be obtained by differentiating ( [ aa ] ) and ( [ ee ] ) and from ( [ eq : aviota ] ) : @xmath159 \ .\end{aligned}\ ] ] let us take an example of a binary system contaning a massive black hole with mass @xmath160 and spin @xmath161 , and where the companion s orbit has the semimajor axis @xmath162 km , _
i.e. _ , one - hundredth of the hulse - taylor system
. then the increase in the angle @xmath58 per loss of @xmath163 ( a dimensionless quantity ) is @xmath164 for a circular orbit .
the enhancement factor for an eccentric orbit is @xmath165 the enhancement factor is monotonously increasing ( fig.1 ) from @xmath166 to make the change in @xmath58 comparable to that in @xmath167 only when the eccentricity approaches the unit value to one part in @xmath168 .
the kepler orbits are completely characterized by the energy @xmath8 , the angular momentum vector @xmath41 and the runge - lenz vector @xmath169 at the initial time @xmath170 .
these are six independent constants of the motion . in the lense - thirring picture
, the constants of the motion alone do not suffice for the full description of the orbit .
we complete the characterization by specifying the euler angles @xmath2 and @xmath11 of the periastron . their evolution over one period , under the influence of the spin ( disregarding radiation backreaction )
is given in ( [ dephi ] ) and ( [ depsi ] ) .
taking into account the gravitational radiation backreaction , the shifts ( [ depsi ] ) and ( [ dephi ] ) will vary from one period to another , as the constants @xmath9 and @xmath10 evolve .
one may interpret these as the radiation - induced changes in the parameters @xmath2 and @xmath11 : @xmath171 @xmath172 where @xmath173 and we have used that in zeroth order @xmath174 .
the total shift in the euler angle @xmath2 of the periastron point is given by the sum of ( [ eq : depsi ] ) and ( [ eq : dpsi ] ) , and likewise for @xmath11 .
we continue with some comments on the radiative losses in the constants of motion .
the terms independent of @xmath0 agree with the results of peters and mathews@xcite .
the terms linear in the spin @xmath0 in our averaged losses agree with ryan s corresponding results , when the parameters @xmath18 , @xmath19 and @xmath58 are converted to his parameters . in the present paper , the
energy and angular momentum losses have been directly computed .
this is to be contrasted with ryan s method where the backreaction forces have been used .
the perfect agreement achieved underscores the reliability of these computations .
when @xmath133 , the averaged losses in @xmath9 are equal to the averaged losses in @xmath10 up to first order in @xmath0 .
hence the equatorial plane of the orbits is stable with respect to radiation losses .
as we have shown in the previous section , the inclination of the orbit can not change significantly in the adiabatic era .
hence we conclude that it is insufficient to restrict the computation of signal templates from a black hole test particle system to equatorial carter orbits . in the limiting case of circular orbits , @xmath175
, there is no distinguished point on the ellipse .
the dependence of @xmath176 on @xmath2 vanishes .
as expected from the axial symmetry , the radiative losses , for arbitrary excentricity , do not depend on the initial angle @xmath11 . in the present paper
, we have developed a toolchest for the computation of radiation losses , which include a new parametrization of the orbit , new angular variables and the application of the residue theorem for obtaining time averages .
currently we are employing these procedures with the inclusion of finite mass effects.@xcite
this work has been supported by otka no .
t17176 and d23744 grants .
the algebraic packages reduce and maplev have been used for checking our computations .
introduce three angle variables @xmath177 and @xmath178 , where @xmath179 is the angle of rotation about @xmath80 , bringing the position of the particle to the node ; @xmath81 is the angle subtended by @xmath80 and @xmath7 and @xmath178 is the angle of precession of @xmath80 about @xmath7 .
the angle @xmath179 is the _ argument of the latitude _
, @xmath81 the _ inclination _ and @xmath178 the _ longitude of the node _
unlike @xmath58 , the angle @xmath81 is not constant .
let us express @xmath81 in terms of @xmath58 and @xmath70 .
computing the @xmath6 component of @xmath80 in two different ways we get the two sides of the equation @xmath180 with the time derivatives of polar angles ( [ polar ] ) after linearization in @xmath0 , the previous equation gives here we are using that the primed angles are equal , to zeroth order in @xmath0 , with the corresponding unprimed ones . taking the time derivative of @xmath81 we get @xmath182 with the new time - dependent euler rotations the coordinates of the particle become @xmath183 and the time derivatives of the polar angles are @xmath184 @xmath185 inserting these equations in ( [ polar ] ) , we get the changes of the primed euler angles in terms of the unprimed ones @xmath186 @xmath187 integrating these equations , we get the new angle variables and their shifts over one period @xmath188 note that these changes in the primed variables are equal to the shifts ( [ dephi ] ) and ( [ depsi ] ) of the unprimed variables . 99 b.carter , phys . rev . *
174 * , 1559 ( 1968 ) k.s.thorne , private communication a.ori , phys
* d55 * , 3444 ( 1997 ) l.kidder , c.will and a.wiseman , phys .
rev . * d47 * , 4183 ( 1993 ) b.m.barker and r.f.oconnell , gen.rel.gravitation * 11 * , 149 ( 1979 )
the computations of @xcite have been extended to the 5/2pn approximation by l.kidder , phys .
* d52 * , 821 ( 1995 ) m.shibata , phys .
rev . * d50 * , 6297 ( 1994 ) f.ryan , phys
* d52 * , r3159 ( 1995 ) f.ryan , phys .
rev . * d53 * , 3064 ( 1996 ) h. thirring and s. lense , phys
. zeitschr .
* 19 * , 156 ( 1918 ) , english translation : gen.rel.gravitation , * 16 * 727 ( 1984 ) l.d.landau and e.m.lifsitz , the classical theory of fields , pergamon , oxford , 1975 l. .
gergely and z.perjs , in _ proceedings of the virgo conference , eds .
i.ciufolini and f.fidecaro , world scientific , singapore , 1997 _ l.blanchet and t. damour , ann.inst.henri poincar , * a50 * , 377 ( 1989 ) t. damour and b. iyer , ann.inst.henri poincar , * 54 * , 115 ( 1991 ) l..gergely , z.perjs and m.vasth , in preparation .
the parametrization of the kepler orbit by the angle @xmath16 is close in spirit to ptolemy s idea of the _ equant_. both the equant and the orbit parameter are means for setting up a rule for determining the angular velocity of the orbiting body .
peters and s.mathews , phys .
131 * , 435 ( 1963 ) p.c .
peters , phys .
rev . * 136 * , b1224 ( 1964 ) | the gravitational radiation backreaction effects are considered in the lense - thirring approximation .
new methods for parameterizing the orbit and for averaging the instantaneous radiative losses are developed . to first order in the spin @xmath0 of the black hole , both in the absence and in the presence of gravitational radiation , a complete description of the test - particle orbit
is given .
this is achieved by two improvements over the existing descriptions .
first , by introducing new angle variables with a straightforward geometrical meaning .
second , by finding a new parametrization of a generic orbit , which assures that the integration over a radial period can be done in an especially simple way , by applying the residue theorem .
the instantaneous gravitational radiation losses of the system are computed using the formulation of _ blanchet , damour and iyer(1989)_. all losses are given both in terms of the dynamical constants of motion and the properly defined orbital elements @xmath1 and @xmath2 .
the radiative losses of the constants characterizing the lense - thirring motion , when suitably converted , are in agreement with earlier results of _ kidder , will and wiseman(1993 ) , ryan(1996 ) and shibata(1994)_. in addition , the radiative losses of two slowly changing orbital elements @xmath3 are given in order to complete the characterization of the orbit .
# 1@xmath4#1 |
it is now clear that chemical noise can sometimes play a significant role in the functioning of biochemical reaction networks @xcite .
such noise is implicated in the spontaneous flipping of genetic switches @xcite , and has been shown to have an adverse effect on the functioning of a synthetic chemical oscillator @xcite . in some circumstances
it has also been argued that noise can have a beneficial effect , for example in stochastic focusing @xcite .
chemical noise in reaction networks can be characterised by the root mean square ( rms ) deviation of number of molecules from the mean .
noise is important if the rms deviation is a significant fraction of the mean .
this is typically associated with systems where the mean number of molecules is rather small .
such situations are characteristic of gene regulatory networks where the concentrations of regulatory proteins can be small enough for there to be only a few molecules in the cell volume .
for example , for the lac operon in _
e. coli _ , the lac repressor is active at concentrations where there are only 1020 molecules present in the cell volume @xcite .
the mean and the rms deviation are ` point statistics ' , and have frequently been used to characterise the stochastic properties of biochemical reaction networks . in the present paper ,
we focus on the noise power spectra , which are a more refined characterisation of the stochastic properties .
we also focus on linear reaction networks for which the noise power spectra and point statistics can be calculated exactly . in the next section ,
we develop a general theory for computing the noise power spectra in linear reaction networks .
we relegate technical details to the appendix . in the following sections we apply the general theory to increasingly elaborate networks which model basic processes of biochemical interest , such as gene expression and signal detection .
we define a _ linear reaction network _ to be a network of chemical reactions which does not involve bimolecular or higher order reactions .
for such a network , the chemical rate equations can be written as @xmath0 where @xmath1 is the number of molecules of species @xmath2 , @xmath3 is a matrix of rate coefficients , and @xmath4 are source terms .
is a set of linear ordinary differential equations , which explains the origin of the phrase ` linear reaction network ' .
appendix [ sec : appkb ] indicates how @xmath3 and @xmath4 are related to the stoichiometry matrix and reaction rates which specify the network .
let @xmath5 be the instaneous number of molecules of species @xmath2 in the system .
define the mean value of @xmath1 in steady state to be @xmath6 , and the mean value out of steady state to be @xmath7 . by taking moments of the chemical master equation ( see appendix ) , one can prove that @xmath8 therefore , the chemical rate equations are _ exact _ for a linear reaction network , provided we work with the mean quantities @xmath7 .
in steady state , the mean values therefore solve @xmath9 this is a system of linear simultaneous equations which can be readily solved .
we now define the deviation away from the steady state mean values to be @xmath10 and introduce the steady - state variance - covariance matrix @xmath11 the diagonal elements of this are the variances @xmath12 the rms deviation mentioned in the introduction is @xmath13 .
it is often the case that @xmath14 and in fact some authors define the ` noise strength ' to be @xmath15 although we will not explicitly use this in the present paper .
this scaling means that the relative rms deviation @xmath16 which indicates that the relative important of noise decreases inversely with the square root of the number of molecules involved .
it is the basic reason why noise is important for systems where the number of molecules is small .
as many authors have noticed @xcite , the variance - covariance matrix at steady state can also be obtained by taking moments of the chemical master equation .
the details are unimportant for our arguments , and are relegated to appendix [ sec : appkb ] . for the remainder of the main text , we assume that the means and variance - covariance matrix have been computed .
we will say that these comprise the _ point statistics _ of the network .
to refine the description of the steady state beyond the point statistics , we introduce the set of correlation functions @xmath17 in steady state , the value chosen for the time origin is unimportant ( therefore the average could be regarded as a time average over starting times , instead of an ensemble average ) .
the correlation functions have the properties @xmath18 c_{ij}(t ) \to 0\quad\text{as}\quad t\to\infty,\\[6pt ] c_{ij}(t ) = c_{ji}(-t ) \end{array}\ ] ] ( the first of these is not always true ; see section [ sec : fast ] ) .
the last property shows that the autocorrelation functions are time - symmetric , @xmath19 , but it does not follow that the cross - correlation functions are time - symmetric , as is illustrated by the example in section [ sec : gem ] . closely related to the autocorrelation functions are the noise power spectra , which are the central topic of study in the present paper .
these are defined to be @xmath20 where @xmath21 is the fourier transform of @xmath22 .
we shall use the second form in eq . to compute the noise power spectra .
usually the noise power spectra obey the sum rule ( for an exception , see section [ sec : fast ] ) , @xmath23 now we turn to the computation of the correlation functions and the noise power spectra . in appendix
[ sec : appc ] we prove that the correlation functions satisfy @xmath24 if we subtract eq . from eq . , we find that the mean deviation out of steady state obeys @xmath25 this shows that the correlation functions decay in exactly the same way as deviations away from steady state , so that eq .
is an example of a regression theorem @xcite .
is to be solved with the initial conditions @xmath26 .
a convenient approach is by the use of the laplace transform , which allows the initial conditions to be automatically incorporated . taking the laplace transform of eq .
shows that @xmath27 where @xmath28 eq .
is a set of linear simultaneous equations which can readily be solved for @xmath29 .
the power spectra now follow almost for free . comparing eq . with eq .
shows that @xmath30 in other words @xmath31 is twice the real part of the analytic continuation of @xmath32 to @xmath33 .
eqs . and
are the key results of this section .
they indicate how exact results for the power spectra can be obtained for an arbitrary linear reaction network by purely algebraic methods . in appendix
[ sec : appl ] we prove that the same results can be obtained from the chemical langevin equations described by gillespie @xcite .
it is interesting to note that the power spectra are simpler to obtain than the correlation functions themselves , which require an inverse fourier or laplace transform .
to undertake such a step requires in general that a polynomial be factorised , which is not always possible .
in fact the route to the power spectra , via eqs , and in the appendix , just involves the solution of linear simultaneous equations . , @xmath34 , @xmath35 , and @xmath36 .
these correspond to cro in a recent model of phage lambda @xcite .
parameters for the implicit mrna model of eq .
are @xmath37 , @xmath38 , @xmath39 .
these are chosen so that the two models have matched point statistics , @xmath40 and @xmath41 , and matching values of @xmath42.[fig : fig1 ] ]
we first consider a very simple model of gene expression @xmath43 which represents a birth - death process in which @xmath44 copies of a are generated each time the first ( birth ) reaction fires .
this represents the ` burstiness ' of gene expression in prokaryotes @xcite .
the point statistics are @xmath45 the chemical rate equation is @xmath46 so the autocorrelation function obeys @xmath47 .
the solution , and its laplace transform , are @xmath48 the corresponding power spectrum is @xmath49 now we turn to a more realistic representation of gene expression which explicitly includes mrna , namely @xmath50 the first two reactions represent generation and degradation of m which corresponds to mrna .
the second two reactions represent generation of the protein a from m and the degradation of a. this is a model which has been analysed by other workers @xcite , and we recover previously known results for the point statistics . the mean values in
steady state are @xmath51 in this , @xmath52 is the mean number of copies of a produced in the lifetime of one mrna and can be taken to represent the ` burstiness ' of gene expression . comparing the last of eq . with the first of eq .
, it appears that @xmath53 in the present model corresponds to @xmath44 in the previous implicit mrna model . as we shall see though
, this interpretation does not carry through to the other statistical properties . the steady state variances and covariance
are ( see appendix [ sec : appkb ] ) @xmath54 \displaystyle s_{m\!a}=\frac{\rho{\langle{n_m}\rangle}}{\gamma+\lambda } = \frac{\gamma{\langle{n_a}\rangle}}{\gamma+\lambda}.\label{eq : mavcm } \end{array}\ ] ] in this case , comparing eq . with eq .
, we see that the correspondence @xmath55 fails for @xmath56 .
now we turn to the computation of the correlation functions and the power spectra .
the chemical rate equations are @xmath57 the correlation functions obey the homogeneous form of these equations . taking the laplace transform as indicated in the previous section , we find @xmath58 s\tilde c_{am } - s_{am } = - \lambda\tilde c_{am},\\[6pt ] s\tilde c_{m\!a } - s_{m\!a } = \rho \tilde c_{m\!m } - \gamma \tilde c_{m\!a},\\[6pt ] s\tilde c_{aa } - \sigma_a^2 = \rho \tilde c_{am } - \gamma \tilde c_{aa}. \end{array}\ ] ] because @xmath59 , there are four different laplace - transformed correlation functions ( note though that @xmath60 ) .
the solutions are @xmath61 \displaystyle \tilde c_{m\!a}=\frac{s_{m\!a}}{s+\gamma}+ \frac{\rho\sigma_m^2}{(s+\lambda)(s+\gamma)},\\[12pt ] \displaystyle \tilde c_{aa}=\frac{\sigma_a^2}{s+\gamma}+ \frac{\rho s_{am}}{(s+\lambda)(s+\gamma)}. \end{array}\ ] ] in this particular case , the inverse laplace transforms can easily be found by the method of partial fractions .
we give only the result for the cross - correlation function , which exhibits the interesting time asymmetry mentioned above , @xmath62 ( using @xmath63 for the second line ) .
finally , the power spectra are obtained .
the power spectrum for m is @xmath64 .
this might have been anticipated as a special case of the previous model with @xmath65 , since as far as m is concerned it is undergoing a simple birth - death process . the power spectrum for a is @xmath66 with a suitable choice of @xmath44 and rate cofficients @xmath67 and @xmath68 , the point statistics of the previous model in eq .
could be matched to the present model in eq . .
however , it is in general impossible to match the power spectra since the @xmath69 dependence is different .
[ fig : fig1 ] shows the power spectra for the two models with matched point statistics .
the presence of two correlation times in the explicit mrna model compared to a single relaxation time for the implicit mrna model is clearly seen .
we have gone through the calculations in some detail for these two model , but for the subsequent calculations we will only give the final results . a very simple model of signal detection is the reaction @xmath70 where a is the input signal and b is the output signal .
. has a chemical rate equation @xmath71 we have included an additional noise term @xmath72 in this , so it is actually a chemical langevin equation . taking the fourier transform yields @xmath73 the mean square modulus is then @xmath74 we have assumed that @xmath72 is uncorrelated with @xmath75 .
the theory of chemical langevin equations @xcite ( see also appendix [ sec : appl ] ) shows that @xmath76 , and for @xmath77 one has @xmath78 , thus finally @xmath79 this result has quite a striking interpretation .
the total noise in the output signal is made up of an _ extrinsic _ contribution ( first term ) plus an _ intrinsic _ contribution ( second term ) .
moreover , the extrinsic noise is equal the input signal noise multiplied by a low - pass filter function .
this is analogous to the behaviour of a passive rc circuit element @xcite .
this result , in one form or another , was derived by paulsson @xcite , and by shibata and fujimoto @xcite .
we refer to it as the ` spectral addition rule ' .
it is a potentially important result because it indicates that there is a trade - off between signal amplification by the detection motif and signal contamination by added intrinsic noise .
our exact results for linear reaction networks enable us to test the spectral addition rule .
we find that its validity is limited by the assumption that the intrinsic noise @xmath72 is uncorrelated with the input signal @xmath75 .
in particular , if the detection motif consumes molecules of the input signal , a correlation can arise which spoils the spectral addition rule . to demonstrate this
, we now consider the effect of adjoining eq . onto the gene expression models in the previous section .
we present results for the explicit mrna model , but similar conclusions can be drawn for the implicit mrna model too @xcite .
we therefore solve @xmath80 we find that @xmath81 and @xmath82 } { ( \omega^2+\lambda^2)(\omega^2+(\gamma+\nu)^2)(\omega^2+\mu^2)}\ ] ] with @xmath83 .
it is not hard to demonstrate that the spectral addition rule _ fails _ for these expressions . as an aside
, one can show @xmath84 . since the @xmath85 reaction in eq .
can be regarded as post - translational modification step , this shows that post - translational modification can reduce the noise associated with gene expression .
the reason is that the post - translational modification reaction smoothes the ` burstiness ' of gene expression by acting as a low - pass filter . to appreciate the influence of coupling the detection reaction to the input signal noise
, we now change the detection scheme so that the signal molecules a are _ not _ consumed , by replacing eq . with @xmath86
as far as b is concerned , it is important to note that the _ same _ chemical langevin equation eq
. holds for this detection scheme as for the previous scheme . for this variant
we find @xmath87 is as given in eq . and
@xmath88 } { ( \omega^2+\gamma^2)(\omega^2+\lambda^2)(\omega^2+\mu^2 ) } \label{eq : pbmab}\ ] ] one can check that in this case @xmath87 and @xmath89 do obey the spectral addition rule .
whilst we have only demonstrated the failure of the spectral addition rule for one particular case , the result is suggestive of the general conclusion that the spectral addition rule will only hold for detection schemes which do not consume input signal molecules .
these conclusions can also be reached by analysing the chemical langevin equations for the whole network @xcite .
there is a growing literature on the adiabatic elimination of fast reactions for stochastic chemical kinetics @xcite .
for example kepler and elston @xcite , and shibata @xcite , have a formalism that permits a systematic approach to the problem , involving the identification of the fast and slow variables in the system . in a similar example , bundschuh _
et al _ present simulation results which support the general strategy of elimination of fast variables in terms of slow variables @xcite . here
we examine the procedure for elimination of a fast equilibration reaction in the context of the _ exact _
results for a linear reaction network .
our results shed light on the way in which fast equilibration reactions can be systematically eliminated from a reaction network , whilst preserving the noise attributes . to make a concrete example
, we suppose that the signal detection motif of the previous section consists of a fast binding - unbinding step , followed by a slower detection step .
we replace @xmath85 by @xmath90 where a@xmath91 represents the bound state .
we suppose that the substrate which binds a is in excess , so that our reaction network is still a linear network .
the reaction scheme in its totality is now @xmath92 \text{a}{\overset{\gamma}{\to } } , & { } \quad \text{a}{\underset{k_b}{\overset{k_{\!f}}{\rightleftharpoons}}}\text{a}^*{\overset{\nu}{\to}}\text{b}{\overset{\mu}{\to}}. \end{array } \label{eq : maasb}\ ] ] the noise statistics for this network can be solved but the expressions are rather lengthy .
we therefore focus on the limit of a fast binding - unbinding reaction .
the equilibrium constant @xmath93 for the binding - unbinding reaction is defined via @xmath94 to take the limit of fast equilibration , we keep @xmath93 finite and allow @xmath95 .
the results are as follows .
let us first consider the final product species b. we find @xmath89 is given by eq .
but with ` renormalised ' reaction rates @xmath96 the same is true of the variance @xmath97 .
we can rationalise this as follows .
the chemical rate equations which involve the species in the fast equilibration reaction ( a and a@xmath91 ) are @xmath98 dn_{a^*}/dt = k_{\!f}n_a-(\nu+k_b)n_{a^*},\\[6pt ] dn_b / dt = \nu n_{a^*}-\mu n_b . \end{array}\ ] ] we now make a linear transformation from @xmath75 and @xmath99 to new variables @xmath100 and
@xmath101 : @xmath102 n_y = k n_a - n_{a^*}. \end{array } \label{eq : lintran}\ ] ] these are chosen because @xmath100 is a _
variable which is conserved by the equilibration reaction , and @xmath101 is a _ fast _ variable which vanishes in steady state . in terms of these concentration variables ,
the chemical rate equations become @xmath103 \displaystyle dn_y / dt + ( k+1 ) k_b n_y = k\rho n_m-\frac{\gamma-\nu}{k+1}\,n_x,\\[12pt ] \displaystyle dn_b / dt = \frac{k\nu}{k+1}\,n_x-\frac{\nu}{k+1}\,n_y-\mu n_b . \end{array } \label{eq : cxyb}\ ] ] the second of these shows that @xmath101 relaxes at a rate @xmath104 to a constant @xmath105 , for large @xmath106 .
this formalises the separation of timescales between @xmath101 and all the other concentration variables . in the large @xmath106 limit
therefore , @xmath107 and eqs .
become @xmath108 dn_b / dt = \nu_r n_x - \mu n_b . \end{array}\ ] ] with the reaction rates given by eq . .
these are ( a subset of ) the chemical rate equations for the following reaction network @xmath109 this is the same as the previously discussed scheme in eq . , with the replacement @xmath110 .
the power spectra follow from eqs . accordingly .
the important point to note is that this scheme not only has the correct chemical rate equations but also has the correct noise power spectra .
now let us turn to the power spectra for a and a@xmath91 in the original network .
we find that @xmath111 again these results are easy to rationalise , since inverting eqs . and setting @xmath112 shows that @xmath113 ( the proportionality factors become squared when computing the noise power spectra ) . for the variances , @xmath56 and @xmath114 , we find a different story though .
in fact , in the limit @xmath95 , we find a _ breakdown _ of the sum rule in eq . .
the reason is that @xmath115 for @xmath116 and @xmath117 .
the sum rule always holds for the left hand side of eq . , but
the limiting power spectra are defined in terms of the right hand side of eq . .
( as an aside , we find that the sum rule is satisfied for @xmath118 , which is equal to @xmath119 , confirming that the effect is confined to the individual species in the fast equilibration reaction . ) to handle this kind of situation , we define a ` sum rule deficit ' @xmath120 for the present situation , we find that @xmath121 interestingly , the same result is found for _ all _ situations involving a fast equilibration reaction which we have examined . the origin of the sum rule deficit lies in the noise generated by the fast equilibration reaction .
formally , at any finite @xmath106 we can define the contribution of the equilibration reaction to the total noise of species @xmath2 to be @xmath122 for @xmath116 and a@xmath91 , one has that @xmath123 as @xmath95 , at any finite value of @xmath69 , but that @xmath124 .
the reason is that , whilst @xmath125 vanishes as @xmath126 , it extends over a frequency range @xmath127 , so the total integrated contribution does not vanish in the limit @xmath95 . in words
, the equilibration reaction contributes high - frequency noise which vanishes at any particular finite frequency in the limit of infinitely fast equilibration , but makes a net non - zero contribution to the total integrated noise .
the universal magnitude of the sum rule deficit can be understood by analysing the simplest of all equilibration reactions , namely @xmath128 solving for the point statistics in this system one has @xmath129 and , from eqs . in the appendix , @xmath130 -k s_{aa^*}+\sigma_{a^*}^2 + k\sigma_a^2-s_{aa^*}-2{\langle{n_{a^*}}\rangle}=0\\[6pt ] k s_{aa^*}-\sigma_{a^*}^2+{\langle{n_{a^*}}\rangle}=0
\end{array } \label{eq : aasvcm}\ ] ]
these equations are linearly dependent and can not by themselves be solved for the elements of the variance - covariance matrix .
the reason is that the slow variable @xmath131 is exactly conserved by the reaction scheme in eq . .
however this implies that @xmath132 and so @xmath133 .
these are consistent with eqs . , which can now be solved to get @xmath134 this explains the universal value found for the sum rule deficit above . and @xmath135 .
the lines are for @xmath136 , 1.0 , @xmath137 , and @xmath95 ( dashed line).[fig : fig2 ] ] to complete the discussion , we turn finally to the correlation functions where the analogue of eq .
is @xmath138 for @xmath116 and @xmath117 .
the left hand side is always equal to @xmath139 .
one can show that the difference between the two sides gives an alternative expression for the sum rule deficit , @xmath140 this can be made clear by another example .
consider @xmath141 the steady state point statistics , which hold for arbitrary @xmath106 , are @xmath142 , @xmath143 , @xmath144 , @xmath145 .
we solve for the correlation functions using the methods described in section [ sec : gen ] .
similar results are obtained for both a and @xmath117 so we report results only for species a. the laplace - transformed autocorrelation function is @xmath146 writing @xmath147 allows us to perform the inverse laplace transform by the method of partial fractions . the decay rates @xmath148 are given by @xmath149 the net result is @xmath150 with @xmath151 for any finite value of @xmath106 , it is clear from this that @xmath152 and there is no sum rule deficit .
now let us take the limit @xmath95 with @xmath153 being held constant . to leading order we have @xmath154 \displaystyle a\to\frac{k}{k+1 } , & \displaystyle b\to\frac{1}{k+1 } , \end{array}\ ] ] and thus @xmath155.\ ] ] apart from the amplitude , this correlation function is characteristic of @xmath156 with @xmath157 .
this is entirely in accord with rewriting the original scheme in terms of the relevant slow variable .
the analysis shows that in the limit @xmath95 , we have @xmath158 .
thus there is a non - zero sum rule deficit @xmath159 where we have used @xmath160 .
again , the universal value for the sum rule deficit is recovered .
our analysis shows what happens to the correlation functions in the limit of fast equilibration ( similar results are found for @xmath161 ) . at any finite @xmath106
there are two decay rates in the correlation function . in the limit @xmath95 one of these decay rates
diverges whilst the other saturates to a finite value .
the diverging decay rate carries with it a finite amplitude and it is this that gives rise to the sum rule deficit .
the phenomenology is illustrated in fig .
[ fig : fig2 ] which shows typical results for @xmath162 at several increasing values of the equilibration rate .
the results in this section are supportive of the general strategy of elimination of the fast variables in terms of the appropriate slow variables .
to be specific , the strategy is to rewrite the chemical rate equations in terms of fast and slow variables , eliminate the fast variables , and re - interpret the reduced rate equations in terms of a reduced reaction network , possibly involving new chemical species ( in the present example , x replaces a and a@xmath91 ) .
our study has provided examples where one can rigorously prove that this strategy is successful .
firstly , let us make some general remarks about the use of noise power spectra to characterise the stochastic properties of chemical reaction networks . as we have shown , the noise power spectra are straightforward to calculate once one has the point statistics ( the mean values and the variance - covariance matrix ) , for a linear network , or for a linearisation of a non - linear network .
in fact , the power spectra are _ easier _ to calculate than the autocorrelation functions , which in general require the factorisation of a polynomial , unless one is satisfied with stopping at the laplace - transformed autocorrelation functions . because the power spectra are functions of frequency , they are a more refined measure than the point statistics , and can be used for instance as a sensitive test of whether two reaction networks can be mapped onto one another .
let us comment briefly on the situation for experiments and simulations .
both face similar difficulties in measuring the power spectra .
a signal of sufficiently long duration needs to be captured at sufficiently fine resolution in order to be able to estimate the power spectrum .
event - driven simulations though , such as those based on the gillespie algorithm @xcite , have an advantage since in principle the exact moments when a signal changes its value are known .
secondly , we discuss the biological relevance of our results .
our results indicate that the way noise is transmitted through a signal detection motif may be more complicated than previously thought @xcite . in particular , if the detection motif consumes the input signal molecules , a correlation can be set up between the intrinsic noise generated by the detection motif , and the input signal noise . we have examined this in the context of a post - translational modification reaction attached to the output of a gene expression module .
our analysis shows incidentally that post - translational modification can ameliorate the noise of gene expression , by smoothing out the ` burstiness ' of the translation step .
we also examined the effect of a fast equilibration reaction interposed in the network .
we find that , in the limit of infinitely fast equilibration , such a network can be exactly mapped onto a reduced or coarse - grained network through the use of suitably chosen slow variables .
this result supports the use of slow variables as a general strategy for model - order reduction by elimination of fast reactions .
this work was supported by the amsterdam centre for computational science ( accs ) .
the work is part of the research programme of the `` stichting voor fundamenteel onderzoek der materie ( fom ) '' , which is financially supported by the `` nederlandse organisatie voor wetenschappelijk onderzoek ( nwo ) '' .
pbw acknowledges the hospitality of trey ideker and the california institute for telecommunications and information technology ( calit2 ) at the university of california , san diego , where part of this work was written up .
technical details and proofs are relegated to this appendix . for further reading
we recommend gardiner @xcite , van kampen @xcite , gillespie @xcite , and swain @xcite .
we first describe the chemical master equation and present a useful moment relation . as a preliminary step
we need to define some basic quantities .
the stoichiometry matrix is a non - square matrix @xmath163 equal to the change in species @xmath2 due to the firing of reaction @xmath164 in the network .
we shall use roman indices to label chemical species and the greek symbol @xmath164 to label reactions . in terms of @xmath163 ,
the chemical rate equations can be written @xmath165 where @xmath166 is the flux through reaction channel @xmath164 .
the quantity @xmath166 is usually dependent on the current values of @xmath1 and is referred to by gillespie as the ` propensity function ' .
the propensity functions for a linear reaction network will be described shortly . in these terms
, the probability @xmath167 of being in the state characterised by @xmath1 molecules of species @xmath2 obeys @xmath168 if a reaction channel is impossible , for instance where @xmath169 , we set @xmath170 .
eq . is the chemical master equation .
first we need to specify in more detail the propensity functions for a linear reaction network .
we distinguish between zeroth- and first - order reactions in the network .
the former are pure generating reactions of the form ` @xmath173 ' .
the latter are monomolecular reactions and comprise the rest of the network ( in a linear reaction network , all reactions are either zeroth- or first - order ) .
the set of zeroth - order reactions is denoted by @xmath174 and the set of first - order reactions by @xmath175 .
the propensity function for a zeroth - order reaction is @xmath176 where @xmath177 is the reaction rate .
for the first order reactions , we introduce an indicator matrix @xmath178 , such that @xmath179 if the reaction @xmath164 involves species @xmath2 as the reactant , and @xmath180 otherwise .
armed with this , the propensity function for a first - order reaction is @xmath181 inserting eqs . and into eq
. obtains eq . in the main text , with @xmath182 the first moment of the chemical master equation
is found by setting @xmath183 in eq . .
it is easy to show that this gives eq . in the main text , with the above definitions of @xmath4 and @xmath3 .
the second moment of the chemical master equation is found by setting @xmath184 in eq . .
this gives a closed set of equations for the elements of the variance - covariance matrix .
in steady state , these equations can be reduced to @xmath185 where @xmath186 eq .
is a set of linear simultaneous equations for the elements @xmath187 of the variance - covariance matrix . as an example of the application of this machinery ,
consider the model for gene expression in eq .
in the main text .
there are two species , m and a , and four reactions .
the stoichiometry matrix is @xmath188 the reaction rates are @xmath189 the indicator matrix is @xmath190 in these , the reactions to the left of the vertical line correspond to @xmath191 and the reactions to the right of the line correspond to @xmath192 ( the indicator matrix is only defined for the latter reactions ) . from these one computes
@xmath193 obviously this could have been written down by inspection .
less trivially , one also finds @xmath194 \displaystyle & = \bigl(\begin{array}{cc}2\lambda{\langle{n_m}\rangle } & 0\\ 0 & 2\gamma{\langle{n_a}\rangle}\end{array}\bigr ) .
\end{array}\ ] ] eqs .
simplify to @xmath195 \rho\sigma_m^2-\gamma s_{m\!a}-\lambda s_{m\!a } = 0,\\[6pt ] 2\rho s_{m\!a}-2\gamma\sigma_a^2 + 2\gamma{\langle{n_a}\rangle}=0 .
\end{array}\ ] ] there are only three equations since eq .
is symmetric in @xmath2 and @xmath196 .
these are solved to get eqs . in the main text .
the proof of the regression theorem eq . in the main text is straightforward and parallels the development in gardiner @xcite .
we start by writing out an explicit expression for the correlation function @xmath197 { } \hspace{8em}{}\times p_s(n_{i,0})\;p(n_{i , t}|n_{i,0 } ; t ) . \end{array } \label{eq : app : ceq}\ ] ] in this , @xmath198 is the steady state probability distribution for the starting point @xmath199 , and @xmath200 is the conditional probability distribution for @xmath201 at time @xmath202 , given that the system started with @xmath199 at @xmath203 .
these probability distributions could in principle be found by solving the chemical master equation .
we now define two kinds of averages @xmath204 { \langle{f(n_i)}\rangle}_0= \int d\{n_{i,0}\}\;f(n_{i,0})\;p_s(n_{i,0 } ) .
\end{array}\ ] ] in words , the first is the average value of a function of @xmath205 at time @xmath202 , given that the system started with @xmath206 at @xmath203 .
the second is the average value of a function of @xmath199 in steady state conditions . in terms of these , eq .
can be written as @xmath207 this result is completely general @xcite . for a linear reaction network ,
taking the first moment of the chemical master equation proves an analogue to eq .
, namely @xmath208 combining this with eq
. demonstrates that @xmath209 obeys eq . in the main text .
we prove that the power spectra obtained from the chemical langevin equations for a linear reaction network are equivalent to the exact results obtained from the chemical master equation . as the first step
, gillespie has shown that the chemical langevin equations for a general network are @xcite @xmath210 in this , @xmath211 are independent unit gaussian white noise functions , one for each reaction channel . applying this to a linear reaction network obtains @xmath212 with @xmath213
it follows that the @xmath214 are gaussian white noise functions with the following statistics @xmath215 the correlation matrix @xmath216 is _ identical _ to the matrix that features in the computation of the variance - covariance matrix , given by eq . above , which can be interpreted in this context as a fluctuation - dissipation theorem .
working in terms of the deviation from steady state , and taking the fourier transform of the chemical langevin equations , obtains @xmath217 we conclude @xmath218 where @xmath219 is the inverse of @xmath220 and @xmath221 is the adjoint .
our task is to prove that eq .
is equivalent to eqs . and in the main text .
the problem is made easier if we rewrite everything in abstract matrix notation .
the result we have just obtained can be written as @xmath222 with @xmath223 ( @xmath224 is the transpose of @xmath225 ) .
the diagonal elements of @xmath226 are the power spectra from the chemical langevin equation route .
we now similarly rewrite the results from the analysis of the chemical master equation .
eqs . , and can be written respectively as @xmath227 i\omega { { \mathbf c}}-{{\mathbf s}}={{\mathbf k}}\cdot{{\mathbf c } } , \quad\text{or}\quad{{\mathbf c } } = { { \mathbf b}}\cdot{{\mathbf s}},\\[6pt ] { { \mathbf p}}_2={{\mathbf c}}+{{\mathbf c}}^\dag = { { \mathbf b}}\cdot{{\mathbf s } } + { { \mathbf s}}\cdot{{\mathbf b}}^\dag \end{array } \label{eq : app : p2}\ ] ] ( note that @xmath228 is real and symmetric ) .
the diagonal elements of @xmath229 are the power spectra from the chemical master equation route .
eliminate @xmath230 between the first of eqs . and eq . to get @xmath231
it follows from eq .
that @xmath232 on substituting these into eq .
, there is a cancellation of terms , and one finds @xmath233 comparison with the last of eqs .
shows that @xmath234 .
this contains our desired result , and proves that the power spectra obtained by the chemical langevin equation route are the same as those obtained by the ( exact ) chemical master equation route .
actually , we have a slightly stronger result , since it follows that @xmath235 , for any linear combination @xmath236 .
swain has presented an analysis of the chemical langevin equations for , effectively , a general linear reaction network @xcite .
he obtains general results for the correlation functions and the variance - covariance matrix in terms of the eigenvalues of @xmath3 .
the present section can be read as proving that the variance - covariance matrix can be found by a simpler route , by solving eq . in terms of the noise correlation matrix . for the correlation functions
though , our route involves an inverse laplace transform which in general requires that a polynomial be factorised , and is equivalent to solving the secular equation for @xmath3 .
so our methods do not provide any additional simplification compared to swain for the correlation functions .
as we have emphasised in the main text though , our results demonstrate that the power spectra are much easier to calculate than the correlation functions . | we present a simple method for determining the exact noise power spectra in linear chemical reaction networks .
we apply the method to networks which are representative of biochemical processes such as gene expression and signal detection .
our results clarify how noise is transmitted by signal detection motifs , and indicate how to coarse - grain networks by the elimination of fast reactions . |
the modeling of equilibrium figures is one major goal in astrophysics . in the case of compact objects ,
relativistic effects become relevant and it is necessary to treat them using general relativity . exact solutions in this field
are rare , and one is generally required to resort to approximation or numerical methods . a detailed treatment of these issues can be found in meinel et al .
( 2008 ) , friedman & stergioulas ( 2013 ) , where the latter includes important discussions of stability .
although it is the numerical approach , with its many possibilities for taking into account the properties of matter , that provides the most accurate means of modeling real astrophysical objects , it is essential that one pursues analytic methods in order to gain deeper insight into the nature of the theory of relativity and how it differs from newtonian theory .
one of the most fruitful avenues of research has always been the treatment of homogeneous matter , where it is precisely the extreme simplification that has permitted great headway . here
we consider one such object , the post - newtonian ( pn ) dedekind ellipsoid , in some depth , since we believe that doing so provides new insights by correcting previous errors and by setting out along a path that may provide new answers concerning the nature of stationary solutions within relativity
. in newtonian gravity , closed form solutions are known for various ellipsoidal figures of equilibrium including the maclaurin spheroids ( rigidly rotating , axially symmetric , and stationary ) , the tri - axial jacobi ellipsoids ( rigidly rotating , non - axially symmetric , and time independent in a rigidly rotating frame ) , the tri - axial dedekind ellipsoids ( non - axially symmetric and stationary ) , and the tri - axial riemann ellipsoids ( in general : non - axially symmetric and time independent in a rigidly rotating frame ) . for a summary of their properties see @xcite .
the fluids described by these solutions are homogeneous and have a velocity field , which is linear in cartesian coordinates .
moreover , they allow for the ratio of two half axes of the ellipsoid to be adjusted arbitrarily giving rise to one parameter families of solutions . for tri - axial ellipsoids ,
the third semi - axis is uniquely determined by choosing this ratio . the jacobi sequence and the dedekind sequence branch off from the maclaurin sequence thereby allowing an axially symmetric and rigidly rotating limit . in the search for analogous figures of equilibrium in general relativity ,
there is still much work to do .
however , for all of the aforementioned newtonian families of figures of equilibrium , a post - newtonian ( pn ) approximation was found in the sequence of papers @xcite
. hereafter we call these articles papers i and , if we refer only to the last one , paper ii .
the maclaurin sequence was also studied in @xcite . in the latter paper
, an algorithm was given that allows one to obtain the pn approximation to the maclaurin ellipsoids to arbitrary order .
since the newtonian jacobi ellipsoid has a time dependent quadrupole moment , the energy loss due to gravitational wave emission can be estimated .
this was done in @xcite .
thus , they can not describe figures of equilibrium in general relativity , assuming that the newtonian limit exists .
in fact , it was shown that the non - radiating final state is the maclaurin ellipsoid at the bifurcation point assuming that the jacobi ellipsoids evolve along the jacobi sequence .
similarly , the riemannian ellipsoids also lose energy due to gravitational wave emission .
the irrotational riemann ellipsoids were investigated to 1 pn order in @xcite .
however , the newtonian dedekind ellipsoids are stationary .
although they are non - axially symmetric they keep their form due to internal motion each fluid element moves along an ellipse in a plane perpendicular to the total angular momentum of the configuration . thus , they are a good starting point to investigate the question if _ stationary but not static relativistic stars are necessarily axially symmetric _ , cf
if dissipative effects are not neglected , it was shown in @xcite that this is always true , but in case of perfect fluids it is still an open question .
before we start deriving pn dedekind ellipsoids , we should state which properties should be satisfied by such a sequence . obviously , they should yield the newtonian dedekind ellipsoids in an appropriate limit , which is guaranteed if they are used as starting point of a pn approximation .
moreover , the pn dedekind ellipsoids should generalize as many properties of the newtonian dedekind ellipsoids as possible .
this includes the reflection symmetry with respect to the coordinate planes and they should approach the 1 pn maclaurin spheroids close to the bifurcation point .
the first property is respected in paper ii but not the second . in @xcite
, we showed that both requirements can be satisfied with a generalization of the ansatz for the 1 pn dedekind ellipsoids of paper ii . in the present paper ,
we study the entire sequence of these figures of equilibrium and discuss their properties .
all 1 pn sequences of figures of equilibrium studied in papers i admit singularities at certain axis ratios .
this implies that in a neighborhood of these points , the 1 pn approximation is not applicable any longer .
interestingly , these singularities along the pn maclaurin and pn jacobi sequences appear at axis ratios , where the newtonian sequence has special properties . in @xcite
, it was shown that the singularity along the 1 pn maclaurin sequence coincides with the first bifurcation point along the newtonian maclaurin sequence of a sequence of axially symmetric , stationary , rigidly rotating and homogeneous figures of equilibrium . a conjecture stating that all the bifurcation points of such sequences are reflected by a singularity in the pn approximation , cf .
@xcite , was proven in @xcite .
there it was pointed out that , if the bifurcation points are ordered appropriately , then there appears a singularity in the @xmath0th pn order for the @xmath0th bifurcation point . for the pn jacobi sequence ,
see @xcite , the singularity at the 1 pn order is related to the onset of a fourth - harmonic neutral mode of deformation of the newtonian jacobi ellipsoids ; also here the singularities in the pn approximation of jacobi ellipsoids is intimately tied to physical properties of the newtonian sequence .
however , the axis ratio , where the singularity obtained for the 1 pn dedekind ellipsoids appears in paper ii , could not be identified with a special point along the newtonian sequence by the authors of that paper . with our generalization of the ansatz for the dedekind ellipsoids , see @xcite , we are able to show that this singularity can be removed completely , suggesting that it is only due to an ansatz , which is too restrictive .
surprisingly , chandrasekhar & elbert already considered such a generalization in footnote 2 of paper ii , but discarded it because it was not helpful in curing the singularity in their opinion .
the paper is organized as follows .
we will first discuss the newtonian solution . on the one hand , this is done to fix our notation , but will also enable us to motivate certain limits and properties for the 1 pn generalized dedekind ellipsoids .
we will also determine the exterior solution in an explicit form using ellipsoidal harmonics . in section [ sec:1 pnapproach ] ,
we give the field equations for a 1 pn self - gravitating perfect fluid solution .
these are solved for the newtonian dedekind ellipsoids as starting point .
subsequently , we discuss the properties of this 1 parameter family in detail , in particular the singularities in parameter space . some explicit formulae , lengthy calculations , and figures are moved to the appendices for readability .
the newtonian dedekind ellipsoids describe rotating and tri - axial ellipsoids with a homogeneous mass density .
they are stationary in an inertial frame and are solutions to the coupled poisson and euler equations : @xmath1 where @xmath2 is the newtonian gravitational potential , @xmath3 the homogeneous mass density ( i.e. , a constant ) and @xmath4 the newtonian velocity of the particles .
the surface of vanishing pressure , i.e. , the surface of the configuration , is that of a tri - axial ellipsoid : @xmath5 we have chosen @xmath6 without loss of generality .
the three axes @xmath7 have to satisfy the relation @xmath8 the index symbols @xmath9 and @xmath10 , @xmath11 , see @xcite , are defined by @xmath12 it is also convenient to define @xmath13 these index symbols satisfy various identities that allow one to express any @xmath9 and @xmath10 as a linear combination of @xmath14 and @xmath15 , see @xcite .
we introduce the dimensionless axis ratios @xmath16 and @xmath17 .
the index symbols are homogeneous functions in @xmath18 : @xmath19 the dimensionless index symbols @xmath20 and @xmath21 will be denoted by @xmath22 and @xmath23 , respectively .
thus , equation in dimensionless form is given by @xmath24 .
it can be solved numerically and determines @xmath25 as a function of @xmath26 independently of @xmath18 ( see @xcite for a table of values for @xmath25 given @xmath26 ) .
hence , the free parameters of this solution are @xmath27 $ ] , @xmath18 and the mass density @xmath3 or , alternatively , the total mass of the configuration @xmath28 .
we further introduce a dimensionless constant @xmath29 and a constant @xmath30 via @xmath31 the solution of equations with a surface given by equations and reads in the interior , cf .
@xcite , @xmath32 several limits are possible in the parameter space of this family of solutions . focusing on the axis ratio @xmath26 , there are the limits @xmath33 and @xmath34 . in the latter case
, we have @xmath35 , too . in the first case , the dedekind ellipsoid approaches the maclaurin ellipsoid with @xmath36 marking the well - known bifurcation point along the maclaurin sequence . at this point , both the jacobi and the dedekind ellipsoids branch off . in this limit
, the free parameters @xmath18 and the mass density @xmath3 can be prescribed as an arbitrary function of @xmath26 leading to several qualitatively different possibilities depending on the behavior of these functions .
for instance , if @xmath37 , which implies @xmath38 as well as @xmath39 , and if moreover @xmath40 , then the maclaurin ellipsoids contract to a point particle with mass @xmath41 . in this limit
, the velocity fields vanish as well .
if on the other hand , @xmath42 , which implies @xmath43 , and @xmath3 approaches some value @xmath44 the entire space is filled with a rigidly rotating perfect fluid .
of course , this solution becomes unphysical for radii , where the fluid elements have an orbital velocity greater than the velocity of light . in the limit , where both , @xmath18 and @xmath3 , approach some finite and positive values @xmath45 and @xmath46 , a maclaurin ellipsoid is obtained . here
as well , @xmath45 is restricted by the physical requirement of subluminal motion .
let us turn our attention to @xmath34 . in order to discuss this limit , we shall need to use the following expansions and @xmath47 we have @xmath48 , if @xmath49 .
if the dependent variable is clear from context we drop the index of the landau symbol . ] @xmath50 the expansions of the other index symbols can be inferred from the recursion relations given in @xcite , cf . the comment after equation . in the most interesting case , where @xmath51 and @xmath52 with @xmath53 and @xmath54 implying @xmath38 and @xmath39 , the ellipsoids degenerate to a rod .
then the velocity field of equation always diverges logarithmically , which follows from equations .
in fact , two anti - parallel , non - interacting streams in the @xmath55-direction with infinite velocity emerge such that the solution is static .
the same holds true for the case @xmath56 . in these cases ,
the global solution is not admissible for a pn approximation sufficiently close to the limit @xmath57 . however , if @xmath58 is approached sufficiently fast , the limit describes a newtonian solution that can be interpreted as two anti - parallel streams of massless particles with a finite velocity .
hence , a pn approximation might be possible .
we present the details of this in appendix [ sec : limit_rod ] . if @xmath59 and @xmath54 , we obtain again a point mass .
if the limit of @xmath60 is sufficiently small for @xmath34 , then the velocity is subluminal during the limiting process .
we describe here the formalism with which the exterior solution of the 1 pn equilibrium figures is obtained in closed form in section [ sec:1 pnapproach ] . as a practical example , we apply the algorithm to the newtonian dedekind ellipsoids .
we introduce ellipsoidal coordinates @xmath61 with @xmath62 with @xmath63 and @xmath64 , see e.g. @xcite : @xmath65 these coordinates cover the octant @xmath66 , which is sufficient since the problem is reflection - symmetric with respect to the surfaces @xmath67 .
we assume that the pn configuration has this symmetry as well .
the surface of the ellipsoid is characterized by @xmath68 . using these coordinates ,
a separation of variables in the poisson equation with a density @xmath69 the support of which is an ellipsoid is possible .
the solution @xmath70 of @xmath71 will be of the form @xmath72 where the functions @xmath73 are the lam functions of the first kind .
their definition and the first few members of this complete set of functions can be found in @xcite .
the function @xmath74 will be obtained as a solution of the inhomogeneous lam equation @xmath75=\tilde g_m^n\left(\lambda^1\right)$ ] , where the lam operator @xmath76 is given by @xmath77 the characteristic values @xmath78 of the lam functions are also defined in @xcite .
the @xmath79 are the expansion coefficients of the density @xmath3 with respect to the ellipsoidal harmonics : @xmath80 since @xmath69 vanishes outside of the ellipsoid , the equation becomes homogeneous and the sole solution with the correct asymptotics is given by @xmath81 where @xmath82 denotes the lam functions of the second kind , see again @xcite .
these follow from the lam functions of the first kind via @xmath83 to the orders , which appear in the present paper , the @xmath84 can be given explicitly in closed form in terms of elliptic functions .
in general , we have to solve the inhomogeneous lam equation .
however , we can rely for all potentials that we have to calculate on a result by @xcite . there
the interior solution of the poisson equation with a density that is polynomial in cartesian coordinates inside an ellipsoid is given .
it is also established that this solution can be connected to an exterior solution , which vanishes at infinity , such that the solution is continuously differentiable everywhere .
thus , we obtain @xmath85 for @xmath86 simply by a coordinate transformation of ferrers interior solution from cartesian to ellipsoidal coordinates .
since the interior solution is polynomial in cartesian coordinates , the expansion of this in ellipsoidal surface harmonics @xmath87 terminates at finite order and we can simply read the @xmath74 off .
hence , @xmath88 can afterwards be obtained by @xmath89 and the potential @xmath2 is completely determined .
we illustrate this method for the dedekind ellipsoids , which are of the type considered by @xcite .
the interior solution can be written after a transformation to ellipsoidal coordinates as @xmath90 @xmath91 this implies that the exterior solution has the form @xmath92 where the constants @xmath88 are obtained from equations and .
in fact , this procedure allows us to obtain the exterior solution for mass densities of the form @xmath93 in closed form .
this enables us to determine the 1 pn metric in the exterior region in closed form for the dedekind ellipsoids as well as the jacobi ellipsoids in exactly the same way .
higher order ellipsoidal harmonics will be necessary in this scheme .
however , the calculations are tedious and will not be presented in detail here .
we will only repeat the form of the higher moments in the interior of the ellipsoid , which are the starting point for the straightforward calculations , in appendix [ sec : higher_moments ] .
in @xcite , a set of field equations was discussed whose solutions describe perfect fluids dynamically to first order in @xmath94 in general relativity . in a subsequent series of papers @xcite , solutions to these equations
were constructed using different newtonian configuration as a starting point namely maclaurin ellipsoids , jacobi ellipsoids and dedekind ellipsoids .
although these are all equilibrium figures , the field equations in @xcite allow for non - stationary solutions , too .
thus , one always has to determine and to solve the equations belonging to the dynamical aspects of the fluid .
in contrast , the projection formalism described in @xcite and the field equations derived therein implement the stationarity from the beginning and can be used only in the description of equilibrium figures . in our case , where we are primarily interested in stationary solutions the latter equations are more advantageous especially if one goes to higher pn orders .
however , to the 1 pn order both approaches yield the same result in the case of stationarity ; discrepancies will become apparent only at higher orders . for equations describing higher order pn corrections ,
see , e.g. , @xcite .
we use the same expansion of the metric as in @xcite : @xmath95 where greek indices run from @xmath96 to @xmath97 , @xmath98 denotes the kronecker delta , @xmath99 .
@xmath2 is the newtonian gravitational potential , which we assume here to be that of a dedekind ellipsoid as discussed in section [ sec : newtonian solution ] .
the contribution @xmath100 is defined momentarily .
the shape of the pn configuration is no longer that of the newtonian ellipsoid @xmath101 ( cf
. equation ) and we denote it by @xmath102 let us introduce a potential @xmath103 that is a solution of the poisson equation for this perturbed ellipsoid , i.e. , @xmath104 with @xmath105 for @xmath106 .
the symbol @xmath100 in equation is then defined by @xmath107 note that we use the expansion parameter @xmath108 to retain the compatibility with papers i and ii .
however , one could transform the results easily to a more physical expansion parameter , e.g. , @xmath109 .
the pressure and the velocity field are also expanded @xmath110 where @xmath111 is the newtonian velocity is defined as in paper ii and does not refer to the spatial components of the four - velocity @xmath112 , but is instead defined as @xmath113 . ] and @xmath114 is the newtonian pressure , cf . equations .
we assume that the homogeneous mass density does not change to any pn order .
the 1 pn equations of a self - gravitating perfect fluid read , see @xcite , [ eq:1 pn_equations ] @xmath115 where @xmath116 is the square of the newtonian velocity field . since the spatial part of the metric is easily obtained and already incorporated in equation the sole equations that remain to be solved in order to obtain the metric to 1 pn are the first two and the one which determines @xmath100 .
the other equations determine the 1 pn corrections to the pressure , velocity field and the surface .
for the 1 pn correction of the surface we choose the following ansatz [ eq : ansatz ] @xmath117 and for the velocity @xmath118 in paper ii , both , @xmath119 and @xmath120 , were not considered . in @xcite , we showed that these linear contributions can be used to allow a rigidly rotating axisymmetric limit of the 1 pn dedekind ellipsoids coinciding with the 1 pn maclaurin ellipsoids .
moreover , it seems natural to allow a 1 pn contribution to the newtonian constant @xmath30 analogous to the angular velocity in the case of rigid rotation , see papers i. furthermore , we introduced the constant @xmath121 and @xmath122 compared to paper ii .
the ansatz for the surface is the same as in paper i , i.e. , it originates from a lagrangian displacement of all fluid elements such that the coordinate volume remains constant .
the lagrangian displacement reads , cf .
equation ( 41 ) in paper ii , @xmath123 as was pointed out in @xcite , it is more physical to fix parameters with an immediate physical interpretation like the mass instead the coordinate volume .
however , by following paper ii , it is easier to compare our results with theirs .
as it was discussed in @xcite and in paper ii , an arbitrary contribution to the coordinate volume can be achieved by an additional displacement of the form @xmath124 , whose divergence does not vanish .
alternatively , one can arrive at the same result by considering a different underlying newtonian configuration , i.e. by substituting in our final results @xmath18 by @xmath125 while keeping @xmath26 and @xmath25 fixed .
note that higher order polynomials could also be allowed in the ansatz , which would eventually lead to a homogeneous system for their coefficients .
the form of the solution we use here is the minimal one needed to satisfy the inhomogeneous equations and get a configuration that is dedekind - like in the sense discussed in section [ sec : introduction ] . in this section
, we will repeat the interior solution for the metric components to an extent that is necessary to construct the exterior solution as well . since the changes in the ansatz do not affect equations and their solutions are the same as obtained in paper ii .
the equations are of the type treated in @xcite , namely they have a polynomial density due to the form of the newtonian solution .
thus , they can be expressed in terms of the higher moments of the homogeneous density @xmath126 , cf .
appendix [ sec : higher_moments ] . with these , the solutions to equations and read @xmath127,\\ u_1&=-\frac{a_1}{a_2 } \omega d_2,\quad u_2=\frac{a_2}{a_1 } \omega d_1,\\ u_3&=0 . \end{split}\end{aligned}\ ] ] a similar approach can be taken for @xmath100 .
it can be written as @xmath128 cf .
equation ( 58 ) in @xcite , i.e. , as the sum of derivatives of poisson integrals . as such @xmath100 is not continuously differentiable across the surface , which is a problem inherent to this coordinate system .
using surface adapted cartesian coordinates @xmath129 the metric components are continuously differentiable .
but we do not use them here but rather follow paper ii .
the densities in the poisson integrals are polynomial for the lagrangian displacement @xmath130 .
thus , the algorithm described in section [ sec : exterior_solution ] is applicable . taking the derivatives afterwards yields @xmath131 note that only six of the ten third order moments @xmath132 and only the diagonal terms of the second order moments @xmath133 are necessary , cf
. equation . in order to fix the constants @xmath134
, we solve the bianchi identity in equation in the next section .
although the changes in the ansatz will not change the calculations fundamentally , we will describe it in more detail and give also intermediate solutions and analytic expressions .
the main reason is that we were unable to reproduce the numerical data given in paper ii for the case @xmath135 ( see the discussion in @xcite for further details ) .
thus , it might prove helpful for rectifying this discrepancy or at the very least make our calculations repeatable .
the formulae , which are too lengthy for the text here , can be found in appendix [ appendix : explicit_results ] .
inserting our ansatz in the integrability condition for the gradient of the pressure , i.e. equation , and using the polynomial structure for a comparison of coefficients yields the following solution : @xmath136 thus , the constant @xmath137 is already determined completely and independently of the parameters @xmath138 . repeating the same for the continuity equation and using the results gives further constraints on the constants @xmath139 a further simplification is achieved by requiring the necessary condition that the normal component of the velocity vanishes at the surface up to 1 pn .
this gives @xmath140\right.\\ & \left.-6\bar{a}_2 ^ 2\left(\bar{b}_{112}+\bar{b}_{122}\right)\right.\\ & \left.- 2\left(2 \bar{a}_2 ^ 2 + 1\right)\left(\bar{a}_2 ^ 2\bar{b}_{123}+ \bar{b}_{23}\right)\bigg)\right.\\ & \left.-\frac{\left(w_1+w_2\right ) \bar{a}_2
\left(5 \bar{a}_2 ^ 2 + 1\right)}{2 \left(\bar{a}_2 ^ 2 - 1\right)}-\frac{\left(5 \bar{a}_2 ^ 2 + 1\right ) \bar{\omega } ^3}{2 } \right],\end{aligned}\ ] ] @xmath141,\\ \hat q_1&=\bar{\omega } \big(\frac{1 } { \bar{a}_2\left(\bar{a}_2 ^ 2 - 1\right)}\left[3 ( { \bar{a}}_2 ^ 2 + 1)(s_2-s_1)\right.\\ & \left.-(5\bar{a}_2 ^ 2 - 3)s_3 + 2\bar{a}_2 ^ 2s_4\right]- 6\bar{a}_2\left(\bar{b}_{112}+\bar{b}_{122}\right)\\ & - 2\frac { \left ( \bar{a}_2 ^ 2 + 2\right)}{{\bar{a}}_2}\left(\bar{a}_2 ^ 2\bar{b}_{123}+
\bar{b}_{23}\right)\big)\\ & -\frac{3\left(w_1+w_2\right ) \left ( \bar{a}_2 ^ 2 + 1\right)}{2 \left(\bar{a}_2 ^ 2 - 1\right)}-\frac{\left ( \bar{a}_2 ^ 2 + 5\right ) \bar{\omega } ^3}{2 \bar{a}_2}. \end{split}\end{aligned}\ ] ] up to now , all coefficients entering the 1 pn corrections to the velocity field can be given in terms of @xmath138 and @xmath134 .
to determine the surface coefficients @xmath134 , we have to impose that the pressure vanishes at the surface up to first pn order , which leads to a linear system of equations : @xmath142 we give the analytic and lengthy expressions of the coefficient matrix @xmath143 and the inhomogeneity @xmath144 in appendix [ appendix : surface_condition ] .
the equations for @xmath145 ensure that the pressure at the surface is constant . having solved those , the pn contribution to the central pressure @xmath146 is obtained using the equation for @xmath147 such that the pressure vanishes at the surface.the parameters @xmath148 and @xmath149 are plotted along the 1 pn dedekind sequence in figure [ fig : parameters ] in appendix [ appendix : solution_plotted ] using a parameterization described in section [ sec : singularities ] .
the singularity which was discovered in paper ii has its origin in a vanishing determinant of the coefficient matrix @xmath143 .
how this singularity can be removed is discussed in section [ sec : singularities ] .
a solution @xmath134 of equation depends on the @xmath138 .
hence , all constants entering our ansatz but @xmath137 are obtained in terms of the @xmath138 . the only requirement for the choice of the @xmath138 is that we have a dedekind - like configuration in the sense of section [ sec : introduction ] and that the resulting surface is still closed .
the latter is just a reformulation of the fact that the 1 pn corrections must be small compared to the newtonian quantities though offering here an explicit and necessary criterion . in this section ,
we discuss some properties of the family of solutions described in sections [ sec : metric_functions ] and [ sec : matter_functions ] .
the @xmath138 can be chosen independently for each @xmath26 along the family so that two free functions @xmath150 remain .
we assume here continuous functions @xmath138 in order to ensure a continuous family of 1 pn dedekind ellipsoids . in section [ sec :
singularities ] , we determine the conditions for the @xmath138 imposed by the requirement that the 1 pn dedekind ellipsoids have similar properties as in the newtonian case .
we treated one of these requirements in @xcite , which ensures that the 1 pn dedekind ellipsoids are axially symmetric and rigidly rotating in the limit @xmath33 coinciding with the 1 pn maclaurin ellipsoids .
this is achieved if @xmath151 holds in the limit .
analytic expressions can be found in that paper .
other restrictions do not ensue from this property .
let us first characterize the 1 pn corrections by two physical parameters of the dedekind ellipsoids the mass and the angular momentum .
according to the definition in @xcite , the 1 pn perturbation of the conserved mass @xmath152 is @xmath153 this is independent of the choice of @xmath138 .
the angular momentum of the newtonian solution as well as the 1 pn dedekind ellipsoids points in the @xmath154-direction .
the first evaluates to @xmath155 whereas the latter is rather lengthy and is shown in appendix [ appendix : angular_momentum ] . here
, we only present the plot , cf .
figure [ fig : mass_angular_momentum ] .
@xmath156 is linear in the @xmath138 as are all other quantities , which we investigate .
thus , we depict for all constants , say @xmath156 , the coefficients in front of the @xmath138 , i.e. , @xmath157 .
we use a solid line for @xmath158 , a dashed line for @xmath159 and a dot - dashed line for @xmath160 .
the part @xmath158 equals the respective constants in paper ii , though as discussed in length in @xcite the numerical values do not agree . + the singularity at @xmath161 , cf .
figure [ fig : angularmomentum ] , is the one already discovered in paper ii .
the importance of singularities in 1 pn approximations of equilibrium figures and the issues with the 1 pn dedekind sequence of that paper was discussed in section [ sec : introduction ] .
since we are able to introduce singularities in our 1 pn dedekind solutions at arbitrary points @xmath26 via @xmath162 , it is obvious that not all singularities are necessarily at physically distinguished points . that we can use the @xmath162 to remove the singularity at @xmath163 for _ all _ physical quantities
is shown in this section .
evaluating the determinant of the coefficient matrix @xmath143 , cf .
equation and appendix [ appendix : surface_condition ] , numerically shows that it changes sign at @xmath164 .
one of the eigenvalues , say @xmath165 , of the coefficient matrix @xmath166 vanishes there . multiplying equation with the eigenvector @xmath167 to the eigenvalue @xmath165 of the transposed matrix yields the condition @xmath168 in paper ii ,
whose results are obtained by setting @xmath135 , equation could not be satisfied . hence
, a singularity is inevitable there in lieu of our approach .
we have a more general inhomogeneity in equation , cf .
the @xmath169 .
choosing at @xmath170 @xmath171 equation is identically satisfied and the remaining system of four equations in equations is regular . to ensure that this holds in the limit @xmath172 as well
, higher orders have to be taken into account .
however , as one can readily check numerically , the equation @xmath173 is well - defined in the limit @xmath172 if @xmath174 .
thus , this equation can be used instead of , for instance , @xmath175 .
the resulting coefficient matrix is now regular and the solution finite .
let us introduce instead of @xmath119 a new parameter defined by @xmath176 , where @xmath177 and @xmath120 are arbitrary .
then , the surface coefficients at @xmath170 evaluate to @xmath178 since these parameters enter the velocity field linearly and their coefficients in equations are well - defined at @xmath164 , the entire solution is regular .
only equation is obtained as an extra condition for @xmath179 . as an example
, the angular momentum reads under this constraint @xmath180 in contrast to figure [ fig : angularmomentum ] , this is finite .
equation is not a good parametrization if @xmath34 or @xmath33 . for the axially symmetric limit , a good parametrization close to @xmath181
was discussed in @xcite and the limit @xmath34 is treated in detail in appendix [ sec : limit_rod ] .
these two limits yield additional constraints for @xmath150 in the neighborhood of the respective points that can be satisfied simultaneously .
which shapes can we expect for the 1 pn dedekind ellipsoids if no additional constraints are given ?
the gravitomagnetic effect , i.e. , parallel matter streams repel each other , is already included in a 1 pn approximation .
thus , one should find that the 1 pn dedekind ellipsoid is elongated in the @xmath154-direction compared to the newtonian figure , since all the newtonian streams are along ellipses in the same direction in parallel planes .
this qualitative argument is corroborated by the maclaurin ellipsoids , which are also elongated in this direction close to the bifurcation point , see @xcite .
moreover , the matter streams in the @xmath182-direction for @xmath183 are all parallel and anti - parallel to those in @xmath184 .
the latter are on average at a larger distance .
thus , the repulsive effect should prevail and in this direction too we have an elongation . in the @xmath55-direction , a similar argument holds .
which effect is stronger , the repulsion in the @xmath182-direction or in the @xmath55-direction , depends on the choice of @xmath138 .
however , this conclusion disregards the pressure entirely and can only provide a rough idea for the resulting 1 pn shape .
more importantly , the lagrangian displacement yields a vanishing 1 pn contribution to the coordinate volume and , thus , it does not allow for an elongation of the ellipsoid in all directions .
in fact , we always observe elongations in the @xmath154-direction , and deformation in the other directions can be adjusted with different choices of the @xmath138 .
the three qualitatively different case are depicted in figure [ fig : surface ] in appendix [ appendix : solution_plotted ] . to explicate the inner motion of the 1 pn dedekind ellipsoids , we integrated the velocity field of the fluid numerically and discuss here the trajectories of a generic fluid element . in the newtonian dedekind ellipsoid , all fluid elements move along ellipses that are in planes with constant @xmath154 .
the time of revolution coincides for all fluid elements . in the 1 pn approximation to these figures of equilibrium ,
the trajectories are distorted ellipses that are still closed and the motion is not any longer confined to planes with constant @xmath154 .
the latter follows directly from our ansatz in equation .
the velocity in the @xmath154-direction vanishes at the coordinate planes so that no fluid element moves from the upper half of the 1 pn ellipsoid ( @xmath185 ) to the lower ( @xmath186 ) and vice versa .
it also changes sign when the particle crosses the other coordinate planes .
moreover , the motion in the @xmath154-direction is periodic with half the time of revolution as period .
furthermore , the time of revolution of the fluid elements depends now on the starting point .
the trajectories for particles at the surface , which stay there during their motion , are exemplary and can be inferred partially from figure [ fig : surface ] .
we gratefully acknowledge discussions with m. ansorg , j. bik , j. friedman and r. meinel .
ng was financially supported by the grants no .
gauk 22708 and gar 205/09/h033 and dp by the deutsche forschungsgemeinschaft as part of the project `` gravitational wave astronomy '' ( sfb / tr7b1 ) .
moreover , ng gratefully acknowledges support from the dfg within the research training group 1620 `` models of gravity '' .
25 natexlab#1#1 , h. , & shibata , m. 1996 , , 54 , 4944 , h. , shibata , m. , & futamase , t. 1996 , , 96 , 81 bardeen , j. m. 1971 , , 167 , 425 byerly , w. e. 1893 , an elementary treatise on fourier series and spherical , cylindrical and ellipsoidal harmonics with application to problems in mathematical physics ( boston : ginn and company ) , s. 1965 , , 142 , 1513 . 1965 , , 142 , 1488 . 1967 , , 147 , 334 . 1967 , , 148 , 621 . 1970 , , 161 , 571 . 1971 , , 167 , 455 . 1971 , , 167 , 447 . 1987 , ellipsoidal figures of efquilibrium ( new york : dover ) chandrasekhar , s. , & elbert , d. d. 1974 , , 192 , 731 , erratum : 1978 , 220 , 303 chandrasekhar , s. , & nutku , y. 1969 , , 158 , 55 ferrers , n. m. 1875 , quarterly journal of pure and applied mathematics , 14 , 1 friedman , j. , & stergioulas , n. 2013 , rotating relativistic stars ( cambridge : cambridge univ . press ) geroch , r. 1971 , , 12 , 918 griffiths , j. b. , & podolsky , j. 2009 , exact space - times in einstein s general relativity ( cambridge : cambridge university press ) , n. , & petroff , d. 2010 , , 722 , 1207 , y. 1970 , theories of equilibrium figures of a rotating homogeneous fluid mass .
( washington d.c .
: nasa special publication ) israel , w. 1977 , , 15 , 935 , l. 1976 , , 208 , 873 lindblom , l. 1992 , , 340 , 353 meinel , r. , ansorg , m. , kleinwchter , a. , neugebauer , g. , & petroff , d. 2008 , relativistic figures of equilibrium ( cambridge : cambridge university press ) petroff , d. 2003 , phd thesis , friedrich - schiller - universitt , jena . 2003 , , 68 , 104029 , k. , asada , h. , & shibata , m. 1998 , , 100 , 703
the interior solutions for the higher moments of the homogeneous mass distribution were obtained in @xcite . they are given in @xcite and we repeat them here : @xmath187
we present the results that are too lengthy for the main text . in section [ sec : mass_and_angular_momentum ] , we discussed the 1 pn contribution to the angular momentum .
we gave the analytic expression for the newtonian angular momentum in equation and the plot for the 1 pn contribution , cf .
figure [ fig : angularmomentum ] . here
, we provide the analytic expression for @xmath156 , too : @xmath188 our numerical results and the numerical results in paper ii do not coincide as discussed in detail in @xcite .
we believe that there is a problem in the numerical evaluation of the right hand side of equation in paper ii .
however , this could not be explicitly seen since those expressions were not given .
we provide the lengthy analytical expression for equation , which we obtained and used in all numerical considerations in our text .
to shorten the results , we introduce a third kind of index symbols @xmath189 with the same meaning of @xmath190 and @xmath191 as in equation .
the coefficient matrix @xmath192 reads [ eq : coeff_mat ] m_00&= + m_01&=3 |a_3 ^ 2 |b_33-|b_13 + m_02&=3 |a_3 ^ 2 |b_33-|a_2 ^ 2 |b_23 + m_03&= ( |a_2 ^ 2 - 1 ) ( |c_112 - 2 |a_3 ^ 2 |c_1123+|a_3 ^ 4 |c_11233 ) @xmath193 the inhomogeneities are given by [ eq : inhom ] b_0&=- ( -2 |c_13 ( |a_1 - 2 |a_2 ^ 2 |b_12)-|a_3 ^ 2 ( -2 |a_1 ( |c_133+|c_233)+|a_3 ( 3 ( |b_1+|b_2)+|c_33)+ .. + & .. 2 |b_12 ( |a_2 ^ 2 ( 2 |c_133-|c_233)+3 |c_233))-2 |c_23 ( |a_1+(|a_2 ^ 2 - 3 ) & .|a_3 ^ 4 |a_3 ( 6 ( |c_133+|c_233)-8 |a_3 + 9 |c_333)+8 |a_^2 ) + b_1&=-(|a_1 ( 2 |a_2 ^ 2 |c_123 - 2 |a_3 ^ 2 |c_133)+|b_12 (-8 ( |a_2 ^ 2 - 1 ) & .. 8 |a_2 ^ 4 |b_223 - 4 |a_2 ^ 2 |c_223 + 16 |a_3 ^ 2 |a_2 ^ 2 |a_23 + 4 |a_3 ^ 2 ( |a_3 ^ 2-|a_2 ^ 2 ) |b_23)+ .
+ & .|a_3 ( 3 |a_2 ^ 2 |a_3 ^ 2 |c_123 - 3 |a_3 ^ 4 |c_133 - 6 |a_2 ^ 4 |a_3 ^ 2 |b_223 + 3 & .. 10 ( |a_2 ^ 2-|a_3 ^ 2 ) |a_3 ^ 6 |b_2333 + 10 |a_3 ^ 6 |b_333 - 5 |a_3 ^ 4 b_2&=- ( 2 |a_1 ( |a_2 ^ 2 ( 4 |b_12+|a_3 ^ 2 ( |c_1223+|c_1233))-|a_2 ^ 4 & ... 2 |a_3 ^ 2 ( |b_123 + 2 |a_223)+2 |b_123 + 7 |b_222 + 4 |b_23)+|a_2 ^ 6 ( |b_122 + 4 |b_123 + 2 |a_222)+|a_2 ^ 2 ( 2 |a_3 ^ 2 ( 2(|b_123+|b_13)- .... + & .... 5 |b_223)+|a_3 ^ 4 |b_133 + 2 |b_23)+|a_3 ^ 2 ( -4 ( |b_123+|b_13)+|a_3 ^ 2 & ... |a_2 ^ 2 |c_2223+|a_3 ^ 2 |c_2233)-4 ( |a_2 ^ 2 + 1 ) |b_12)+20 |a_2 ^ 4 & .. |a_3 ^ 2 ( -3 ( |c_1233+|c_2233)+20 |a_3 ^ 2 |b_2333 - 5 |c_2333)-12 |a_2 ^ 4 b_3&=-(-2 |a_1 ( 2 |b_113-|c_113 + 4 |a_2 ^ 2 |b_12+|a_3 ^ 2 ( |c_133 + 3 |a_3))+4 & .|b_12 ( |a_3 ^ 2 ( -4 |b_123 - 8 |b_13 + 4 |c_233 + 3 |a_3)-4 |a_2 ^ 2 ( -2 & ... |a_3 ^ 4 |b_133))+|a_3 ^ 2 |a_3 ( -3 ( |a_3 ^ 2 + 2 ) |b_113+|a_3 ^ 2 ( -(3 & .... 5 |c_1333 + 3 |c_233 + 5 |b_33))+5 |b_13)-2 |a_2 ^ 2 |a_2 ( 3|b_12+|a_3 ^ 2(|b_123 + |c_233))-4 |b_12 ^ 2 + 6 |a_3 ^ 4 |a_3 ^ 2+(w_2-w_1 ) |a_2 b_4&=- ( 2 ( |a_2 ^ 2-|a_3 ^ 2 ) |a_1 ( 2 |b_1123-|c_1123+|a_3 ^ 2 |c_1233)+4 & .... |a_3 ^ 2 |c_1233 + 2 |b_13)+3 ( |c_1223 + 2 |b_23))+|a_2 ^ 4 ( 3 |c_1123 + 6 & .. |a_3 ^ 2 ( ( |a_3 ^ 2 - 1 ) |c_1233 + 2 |b_23))+|a_3 ^ 2 |a_3 ( |a_3 ^ 2 ( 3 ( |c_1123+|c_1223)-6 |b_1123-|a_3 ^ 2 ( 3 ( |c_1233+|c_2233)+10 & .... 5 |c_2333)+5 |c_1233 + 10 |a_3 ^ 4 ( |b_1333+|b_2333))+|a_2 ^ 2 ( -3 ( |c_1123+|c_1223)+6 |b_1123+|a_3 ^ 2 ( 3(|c_1233 + |c_2233)- .... + & .... 10 |a_3 ^ 2 |b_2333 + 5 |c_2333)-5 |c_1233)-6 ( |a_3 ^ 2 - 1 ) |a_2 ^ 4 .. + & .. ( |a_3 ^ 2-|a_2 ^ 2 ) |c_1223+|a_3 ^ 2 ( |a_2 ^ 2-|a_3 ^ 2 ) |c_2233)+ ) + b_5&=-6 ( 2 ( |a_3 ^ 2 - 1 ) |a_1 ( -5 |b_1113+|b_113-|a_3 ^ 2 |c_1133)-24 |a_2 ^ 2 & . 4 |b_12 ( |a_2 ^ 2 ( 4 ( |a_3 ^ 2 - 1 ) |b_1113 + 6 |b_112 - 2 |a_3 ^ 2 |b_113 + 6 & .. 2 |a_3 ^ 2 ( |b_123 + 2 |c_123)+2 |a_2 ^ 4 |b_123 + 4 |b_23+|a_3 ^ 4 |b_233)+ & .. |a_3 ^ 2 ( 3 ( |c_1133+|c_1233)+5 |c_1333)+5 |c_1133 + 20 |a_3 ^ 4 |b_1333)+2 ( |a_3 ^ 2 - 1 ) |a_2 ^ 2 |a_2 ( |c_1123-|a_3 ^ 2 |c_1233)+ . + & .4 ( 2 |a_2 ^ 2 + 3 ) |b_12 ^ 2 + )
in @xcite , we took a close look at the axisymmetric limit @xmath194 of the pn dedekind ellipsoids .
it turns out that the pn maclaurin spheroids emerge in the limit but only if the pn velocity field is generalized as in equation .
here we consider the opposite limit of a rod along the @xmath195-axis , i.e. , @xmath196 , cf .
section [ sec : limiting_cases ] .
we begin by deriving conditions that arise from the behavior of the metric functions in this limit , where it will be important to treat @xmath3 and @xmath18 as functions of @xmath26 .
we then examine the surface and conclude that the only acceptable solution is a member of the weyl class , i.e. , an axially symmetric and static spacetime . the matter content collapses to a singularity along the axis and the limiting spacetimes contain the levi - civita spacetime and the curzon - chazy particle in the special cases in which the rod has infinite length ( @xmath42 ) or zero length ( @xmath197 ) , see e.g. @xcite . in order to include the levi - civita metric in the subsequent derivations
, we shall divide up the ellipsoid into slices defined by @xmath198 and with a thickness @xmath199 , which we denote by @xmath200 .
let us start by looking at @xmath201 , which is determined by the poisson equation .
the inhomogeneity is a sum of ( newtonian ) kinetic , inner and potential energy densities .
beginning with the kinetic energy contained in @xmath202 , we find @xmath203 it turns out that the ratio of the inner to the kinetic energy tends to zero in the limit .
furthermore , the potential energy is proportional to the kinetic one .
thus , it suffices to derive the form of the kinetic energy density in the limit .
we choose it to be a line energy density as required for a well - defined equation .
this then implies @xmath204 whereby @xmath205 is an arbitrary constant and the logarithmic term comes from the expansion of @xmath206 .
this yields a source for the potential @xmath201 of the form @xmath207 where @xmath208 and @xmath209 denote the dirac delta distribution and the heaviside step function , respectively . in cylindrical coordinates ( @xmath210 and @xmath211 )
the solution to equation reads @xmath212 now consider the potentials @xmath213 .
we remind the reader that @xmath214 vanishes for all @xmath26 .
we prove that @xmath215 is also zero in the exterior in the limit by looking at its multipoles .
the inhomogeneity in equation is proportional to the newtonian linear momentum density in the @xmath55-direction .
its integral over a slice @xmath216 vanishes because of the antisymmetry of @xmath217 .
the integrals over the halves of the slice with @xmath218 and @xmath219 are given to leading order in @xmath25 by @xmath220 which evidently tend to zero for @xmath221 .
an arbitrary multipole moment for , e.g. the density @xmath222 , again to leading order , is then bounded by @xmath223 which all tend to zero by virtue of the preceding equation .
this proves that @xmath224 vanishes for @xmath34 in the exterior .
this holds via corresponding arguments for @xmath225 , too .
hence , the time - like killing vector is hypersurface orthogonal and the spacetime is static in the limit . it remains to analyze @xmath100 and the pn surface . with the mass density given in equation
one can show that @xmath2 tends to zero in the exterior analogously to @xmath213 .
it , thus , suffices to consider the solution to the poisson equation for @xmath103 from equation .
the 1 pn surface is still defined by the condition of vanishing pressure at the surface . a necessary condition for the inhomogeneity in equation to have a well - defined limit in a distributional sense is that the mass @xmath226 contained in a slice @xmath202 is well - defined to 1 pn order : @xmath227 where @xmath228 denotes the 1 pn surface in cylindrical coordinates . using our standard notation for pn terms ,
the 1 pn order of the relevant term of the integrand is @xmath229 the integral over the newtonian contribution vanishes because of equation .
the second term can be written out explicitly using equation and an expansion in terms of @xmath25 .
the expansions of @xmath230 and @xmath134 will be denoted by @xmath231 the expression diverges like @xmath232 for @xmath35 in general .
hence , @xmath226 becomes singular in this limit as well , see equation .
these diverging terms vanish only if we have : @xmath233 however , the @xmath234 must also be consistent with equation to leading order , which provides the further conditions @xmath235 together with equations these equations imply that @xmath236 although these conditions are necessary they are not sufficient .
the next to leading order in @xmath25 in @xmath226 diverges in general logarithmically . to choose the parameters @xmath237
so that these terms vanish , we solve first the system of equations originating from equation , which reads to this order @xmath238 this can easily be solved and the solution , which depends on @xmath237 , can be inserted in @xmath226 .
now , the vanishing of the diverging terms requires the following behavior of the free parameters @xmath237 : @xmath239 with the free constants @xmath240 . the constants @xmath241 and @xmath242 in equation govern the length of the rod @xmath243 in the limit @xmath35 , which still diverges in general .
the choice @xmath244 ensures a finite length , which reads then @xmath245 with an argument like in equation , one can show that all multipole moments converge with the choices and and that @xmath100 is well - defined ; the inhomogeneity in equation tends to a line mass density , which is a polynomial in @xmath55 to order 4 : @xmath246 this is already of order @xmath247 .
hence , it is sufficient to take the newtonian length of the rod into account . the function @xmath100 can easily be calculated for the line density and is given by @xmath248,\\ n_{\pm}=&((a_1 \pm x_1)^2 + \rho^2)^{\frac 12 } \left ( 42 a_1 ^ 3 \mp 146 a_1 ^ 2 x_1 - 130 a_1 x_1 ^ 2 \pm 250 x_1 ^ 3 + 45 a_1 \rho^2 \mp 275 x_1 \rho^2\right ) .
\end{split}\end{aligned}\ ] ] with @xmath201 , @xmath100 , @xmath213 , and @xmath2 all metric functions are determined .
the metric describes an axially symmetric and static vacuum and , therefore , belongs to the weyl class .
the matter region degenerates to a rod along the @xmath195-axis and is described by a singularity .
the singular behavior can be read off the kretschmann scalar and is given to leading order in @xmath249 by @xmath250 the class of weyl metrics includes such prominent members as the lanczos metric and the curzon - chazy metric ( for an overview of these two metrics , see , e.g. @xcite ) .
these metrics result also in some special limits of the 1 pn dedekind ellipsoids .
first the limit @xmath39 is carried out as described above and afterward the following limits are taken : @xmath251 in the case of the lanczos metric and @xmath37 in case of the curzon - chazy particle . in the former case , the line densities and become @xmath252 which are constant along the entire @xmath55-axis .
the resulting spacetime is cylindrically symmetric .
the leading order in @xmath108 of the mass parameter is vanishing , such that it can be interpreted as an effective gravitational mass per unit length , cf .
@xcite . in the case
@xmath37 , the density vanishes and the density tends to the point density @xmath253 with @xmath254 is the total newtonian energy concentrated in this point .
this coincides with a 1 pn approximation to the curzon - chazy solution with a parameter @xmath255 with a vanishing leading order in @xmath108 .
as some analytical expressions are lengthy they were at some places suppressed in the article . in figure
[ fig : parameters ] , the main parameters describing the solution are plotted .
we use for all parameters , say , @xmath256 the usual splitting @xmath257 with the parameterization from section [ sec : singularities ] .
subsequently , the coefficients @xmath258 are depicted .
we use solid lines for @xmath147 , dashed lines for @xmath259 , and dot - dashed lines for @xmath260 .
this parameterization has the advantage that the singularity is already removed .
the results reflecting the parameterization in paper ii are recovered if we set @xmath261 and @xmath262 , which implies @xmath263 , cf.equation . in figure
[ fig : surface ] , three qualitatively different cases , cf .
section [ sec : gravitomagnetic ] , of the surfaces of the 1 pn dedekind ellipsoids are shown . | we derive a family of post - newtonian ( pn ) dedekind ellipsoids to first order .
they describe non - axially symmetric , homogeneous , and rotating figures of equilibrium .
the sequence of the newtonian dedekind ellipsoids allows for an axially symmetric limit in which a uniformly rotating maclaurin spheroid is recovered . however , the approach taken by @xcite to find the pn dedekind ellipsoids excludes such a limit . in @xcite
, we considered an extension to their work that permits a limit of 1 pn maclaurin ellipsoids . here
we further detail the sequence and demonstrate that a choice of parameters exists with which the singularity formerly found in @xcite along the sequence of pn dedekind ellipsoids is removed . |
s a major breakthrough in coding theory , polar codes , introduced by arikan in @xcite , can asymptotically achieve the channel capacity of binary symmetric channels using a successive cancellation ( sc ) decoder with a complexity of @xmath1 , where @xmath2 , is the code block length .
later , some improved algorithms of sc are described in @xcite @xcite @xcite . yet still
the performance of these decoders is inferior to that of the maximum likelihood ( ml ) decoder , or at least can not be proven to achieve the ml performance .
an ml decoder of polar codes is implemented by viterbi algorithm in @xcite ; but the decoding complexity grows exponentially with the code length . inspired by the sphere decoding ( sd ) for space - time block codes @xcite @xcite , an sd algorithm of polar codes is proposed in @xcite to perform ml decoding with a cubic complexity .
the sphere decoder configures an initial target radius , and the candidates in the sphere are enumerated .
whenever an @xmath3-length path corresponding to a smaller radius is found , this path is buffered as a candidate codeword , and based on which , the target radius is updated .
after that , the search is restarted in the new sphere with a smaller radius . until no paths can be found in the sphere with a smaller radius than the target
, the decoder outputs the buffered path as the decoding codeword .
the drawback of this naive decoding is that some paths which have already been spanned in the previous sphere may be respanned .
furthermore , when searching in the sphere , the decoder enumerates and expands the candidate path only based on the path length , namely , if one path has a longer length , it will be expanded with a higher priority .
this length - first search is not a good strategy , because the search rule only involves a simple information of the path length and neglects the information provided by the received sequence .
stack sequential algorithm , as stated in @xcite , is functionally equivalent to the sphere decoding . in this paper ,
a stack sd ( ssd ) algorithm of polar codes is applied , where the paths are stored in a stack with a descending order of the path metrics . in ssd , each possible path is visited at most once .
inspired by massey s work about the optimum frame synchronization @xcite , we derive an optimum path metric based on the ml rule .
further , a much simpler metric is also provided as an approximation of the proposed metric in the high signal - to - noise ratio ( snr ) regimes . with the help of the proposed path metrics , the ssd can efficiently enumerate the candidate paths and narrow the search range so as to remarkably reduce the computational complexity yet still achieve the ml performance .
the remainder of this paper is organized as follows .
describes the process of the polar coding and the signal model over the additive white gaussian noise ( awgn ) channel . derives the optimum and suboptimum path metric based on the ml rule and presents the details of ssd algorithm .
gives the performance and complexity analysis of the proposed ssd algorithm based on the simulation results . finally , concludes the paper .
given an @xmath4 polar code , the @xmath3-length code block ( codeword ) @xmath5 can be generated by @xmath6 where @xmath7 is a source - scrambling vector generated by the @xmath8 bit - reversal permutation matrix @xmath9 defined in @xcite , and the vector @xmath10 is the source block . as the @xmath11-th kronecker power of @xmath12 $ ]
, the matrix @xmath13 has a lower triangular structure , where the @xmath14-th element @xmath15 is taken the value in @xmath16 .
hence , the components @xmath17 can be written as @xmath18 the @xmath3 polarized subchannels are corresponding to the rows of the matrix @xmath13 , and the reliabilities of which are calculated using bhattacharyya parameters @xcite .
the entries in the source - scrambling vector @xmath19 with indices corresponding to the more reliable @xmath20 rows of the matrix @xmath21 are assigned information bits ; the other @xmath22 entries are related to the frozen bits which can be set to zeros when the channels are symmetric . the received sequence @xmath23 can be written as @xmath24 where @xmath25 is the energy of the transmitted signal , and @xmath26 is the transmitted signal vector , and @xmath27 is an @xmath3-length vector of i.i.d . awgns .
the components @xmath28 are statistically independent gaussian random variables with @xmath29 mean and variance @xmath30 where @xmath31 is the one - sided noise spectral density . and the elements @xmath32 are the bpsk modulated signals . the ml estimation of the transmitted bits can be obtained by minimizing the square euclidean distance ( sed ) towards the received sequence , that is , @xmath33 where @xmath34 is an all - one vector of length @xmath3 . after performing a bit - reversal permutation on @xmath35 , an estimation of the source block
is obtained , i.e. , @xmath36 where @xmath37 denotes the matrix inverse .
in fact , @xmath38 .
in this section , we first give a brief description of the code tree and provide an ssd with conventional enumeration strategy .
then , the optimum path metric is derived based on the ml rule and its approximations in high / low snr regimes are also provided .
finally , the proposed ssd algorithm with an efficient enumeration is described in detail . for the sd algorithm ,
the ml estimation @xmath35 can be obtained by enumerating all the transmitted signal vectors @xmath39 within a sphere of radius @xmath40 that is centered at @xmath41 . in @xcite ,
the sd algorithm is described as a series of depth - first searches on the code tree , and a strategy using parallel subtree searches is suggested to reduce the complexity .
but for each iteration of both methods , after updating the radius of sphere , the search process should be restarted from the preceding node .
thus , for a direct implementation of sd , a certain path in the code tree may be revisited many times . to avoid these revisitings , a stack implementation of sd
is applied in this paper . for finite bpsk signals ,
the sphere decoder can be regarded as a bounded search in a code tree . due to the lower triangular form of @xmath42
, the source - scrambling vector @xmath19 can be represented as a path in a tree of depth @xmath3 , where the values of bit @xmath43 at level @xmath44 correspond to branches at depth @xmath45 .
for example , fig.[fig_example ] gives a code tree with four levels .
inspired by @xcite , a stack is used in ssd to store the candidate paths .
the paths with solid branches are the candidates in the stack and sorted by path metric in descending order , on the other hand , those dash branches are visited and discarded ( since their euclidean distances to the sphere center are larger than the target ) . under the ssd algorithm ,
a possible candidate path is visited at most once .
let @xmath46 denote the binary labeling of an @xmath47-length path from the root to a certain branch at the @xmath44-th level , and we use @xmath48 to denote the corresponding transmitted signal vector .
given the received sequence @xmath41 , the corresponding sed between the path @xmath49 and the related segment of received sequence @xmath50 can be expressed as that in @xcite , @xmath51 if the branch @xmath43 is corresponding to a frozen bit and takes an incorrect value , the associated sed is @xmath52 .
the sphere decoder utilizes this sed as a measurement to judge whether the vector @xmath53 is inside the sphere . particularly , when @xmath54 , it is also used to update the radius of sphere .
for the conventional enumeration , since the sd is described as a series of depth - first searches , the length of the candidate path is taken for path sorting in ssd , @xmath55 initially , a null path @xmath56 which is corresponding to the root node is loaded in the stack and its metric is @xmath57 .
the search goes along the longest path in the stack .
whenever a candidate path @xmath49 is found to have a measurement @xmath58 that is larger than the target radius @xmath40 , it will be dropped directly .
when the path on the top of the stack achieves length @xmath3 , i.e. , @xmath54 , the @xmath3-length bit sequence corresponding to this path is buffered as a candidate vector , and the target radius is updated as @xmath59 .
this procedure goes on until the stack becomes empty , the buffered vector is then output as the decoding result .
in fact , this conventional enumeration is to select the candidate path in the stack one - by - one and check that whether the candidate is inside the sphere .
since the paths in the stack are sorted by the simple path metric @xmath60 , the sphere decoder lacks the valuable information to direct the search procedure .
given a path @xmath49 and its corresponding vector @xmath61 , the complement vector @xmath62 can take a possible value from a random vector @xmath63 , where each @xmath64 is either @xmath65 or @xmath66 with an equal probability of @xmath67 .
the optimum path metric for the enumeration is to maximize the likelihood probability @xmath68 since @xmath69 for all @xmath70 , it is equivalent to maximize @xmath71 since @xmath27 is a vector of i.i.d .
awgns , we have @xmath72 @xmath73 substituting this formula into ( [ metric_l2 ] ) and neglecting all the items independent of @xmath61 or @xmath74 , it is equivalent to maximize @xmath75 remember @xmath76 takes value in @xmath77 , the above summation can be expressed as @xmath78 the logarithmic form of ( [ metric_l3e ] ) is @xmath79 given a received sequence @xmath41 , the following summation @xmath80 is a constant . after subtracting ( [ const ] ) from @xmath81
, we obtain a path metric that is equivalent to the optimal one in ( [ metric_l1 ] ) , i.e. , @xmath82 where @xmath83 this ml path metric includes two terms : the first summation involves the correlation between the vector @xmath61 and the received sequence @xmath50 ; the second term , @xmath84 , is a correction term which further consists of two parts : the summation represents a kind of energy correction related to the received sequence @xmath50 ; the second part is corresponding to the depth of the path @xmath49 . compared with massey s metric @xcite , this metric has the same item for the first summation , but the correction term of which includes an additional part to manifest the length of the path .
similar to the works in @xcite , two approximations of the optimum path metric can also be derived in the cases of very high and very low snr .
when @xmath85 , the argument of the @xmath86 in ( [ m1_metric ] ) is much greater than @xmath87 so that the function @xmath88 can be approximated as @xmath89
. therefore , after dropping the constant @xmath90 , the high snr approximation of the optimum path metric in ( [ m1_metric ] ) can be expressed as @xmath91 which is much simpler than the metric @xmath92 . for practical implementation , after receiving @xmath41 , the values of @xmath93 for all @xmath94 can be pre - calculated and stored in a @xmath95-sized table .
thus , for each node visiting during the rest decoding procedure , no extra calculation is needed compared with the ssd using the conventional metric @xmath60 .
the similar method can also be applied for the calculation of metric @xmath92 .
when @xmath96 , the function @xmath97 can be approximated by @xmath98 . using this approximation in ( [ m1_metric ] ) , we have @xmath99 but according to our simulation results ( which are not provided in this paper ) , this metric is not efficient in reducing the complexity .
so we will not discuss it below . the ssd algorithm based on the proposed path metric @xmath92 or @xmath100
can be summarized as follows : * step 1*. initialization : 1 ) set the target radius as @xmath101 ; 2 ) a null path @xmath102 with @xmath103 is pushed into the stack . * step 2*. popping : a path @xmath49 is popped from the top of the stack , if the path reaches the depth @xmath3 , i.e. , @xmath54 , record the estimation vector @xmath104 and update the search radius @xmath105 , then go to * step 4*. * step 3*. expanding : the current path @xmath49 is extended to two new paths , i.e. , @xmath106 and @xmath107 . for each path , calculate the sed @xmath108 by ( [ msed_metric ] ) .
if the optimum path metric is used , calculate the path metric @xmath109 by ( [ m1_metric ] ) .
otherwise , if the high snr approximation is used , calculated the path metric @xmath110 by ( [ m2_metric ] ) . * step 4*. pushing : the paths with @xmath111 are pushed back into the stack , the others are simply dropped .
* step 5*. sorting : the paths in the stack are re - sorted from top to bottom in descending order according to the value of @xmath92 or @xmath100 . * step 6*. stopping : if the stack is empty , then stop the decoding process and output @xmath112 ; otherwise , go to . to efficiently implement an ssd decoder , the same technique in @xcite can be utilized to avoid the redundant copy operations and memory occupation .
compared with the conventional sd , the main difference of ssd algorithm is the sorting operations in * step 5*. the candidate path is enumerated from the most probable one by applying the proposed path metric @xmath92 or its approximation @xmath100 , therefore the complexity of ssd algorithm can be efficiently reduced .
in this section , the ssd is applied to polar codes and reed - muller ( rm ) codes with short code blocks . the performance and complexity of the ssd with the conventional metric and the proposed metrics are evaluated via simulations with bpsk modulation over awgn channels .
fig.[fig_perform ] presents the block - error - rate ( bler ) performances of ( 64 , 57 ) polar and rm codes under ssd with different path metrics .
we can see that the ssd algorithms with all the three metrics , the conventional metric ( @xmath60 in ( [ m0_metric ] ) ) , the ml rule ( @xmath92 in ( [ m1_metric ] ) ) or its high snr approximation ( @xmath100 in ( [ m2_metric ] ) ) , can achieve the same performance .
further , rm codes are shown to have better performances than polar codes under the ml decoding .
the computational complexities are shown in fig.[fig_complexity ] .
the average complexities are evaluated by counting the node visitings in the code tree .
compared with the ssd using the conventional metric @xmath60 , significant complexity reductions can be obtained by using the proposed path metrics : about @xmath0 times in the entire simulated snr regime . although the metric @xmath100 is an approximation of the metric @xmath92 in the case of high snr , the ssd algorithm with this metric can achieve almost the same complexity reduction in the medium to high bit snr regime , that is , @xmath113 .
interestingly , even in the low snr regime , only a slight increasing of complexity under the ssd algorithm with the metric @xmath100 is observed .
a stack - based sphere decoding algorithm with efficient enumeration is proposed to achieve the performance of ml decoding for the polar codes . by introducing the optimum path metric in the sorting operations ,
remarkable complexity reduction can be obtained under the ssd algorithm . moreover ,
a high snr approximation of the optimum metric is provided , which has a very simple form and is more suitable for practical applications . | sphere decoding ( sd ) of polar codes is an efficient method to achieve the error performance of maximum likelihood ( ml ) decoding .
but the complexity of the conventional sphere decoder is still high , where the candidates in a target sphere are enumerated and the radius is decreased gradually until no available candidate is in the sphere . in order to reduce the complexity of sd , a stack sd ( ssd ) algorithm with an efficient enumeration is proposed in this paper .
based on a novel path metric , ssd can effectively narrow the search range when enumerating the candidates within a sphere .
the proposed metric follows an exact ml rule and takes the full usage of the whole received sequence .
furthermore , another very simple metric is provided as an approximation of the ml metric in the high signal - to - noise ratio regime . for short polar codes ,
simulation results over the additive white gaussian noise channels show that the complexity of ssd based on the proposed metrics is up to @xmath0 times lower than that of the conventional sd .
polar codes , successive cancellation decoding , sphere decoding , maximum likelihood rule . |
energy and electronic transport plays a central role in a wide range of chemical and biological systems .
it is the fundamental mechanism for transporting the energy of an absorbed photon to a reaction center in light harvesting systems and for initiating a wide range of photo - induced chemical processes , including vision , dna mutation , and pigmentation .
the seminal model for calculating electron transfer rates was developed by marcus in the 1950s@xcite .
@xmath0 where @xmath1 is energy required to reorganize the environment following the transfer of an electron from donor to acceptor .
and @xmath2 is the driving force for the reaction , as illustrated in fig .
[ marcus ] .
if we assume that the nuclear motions about the equilibrium configurations of the donor and acceptor species is harmonic , the chemical reactions resulting from energy or charge transfer events can be understood in terms of intersecting diabatic potentials as sketched .
the upper and lower curves are the adiabatic potential energy surfaces describing the nuclear dynamics resulting from an energy or charge transfer event , taking the geometry of the donor state as the origin . ) .
energies are given in ev and the collective nuclear displacement is dimensionless . ]
as the transfer occurs by crossing an energy barrier , the transfer rate can be expected to be in the arrhenius form @xmath3 with @xmath4 as the activation energy . using @xmath5 we can relate the activation energy to both the reorganization energy and driving force , @xmath6 .
one of the most profound predictions of the theory is that as the driving force increases , the transfer rate reaches a maximum and further increases in the driving force lead to lower reaction rates , termed the inverted regime .
the existence of the inverted region was demonstrated unequivocally by miller _
@xcite in an elegant series of experiments that systematically tuned the driving force , reorganization energy , and diabatic coupling by careful chemical modification of the donor and acceptor . a number of years ago , our group developed a time - convolutionless master equation approach for computing state - to - state rates in which the coupling between states depends upon the nuclear coordinates@xcite .
this approach incorporates a fully quantum - mechanical treatment of both the nuclear and electronic degrees of freedom and recovers the well - known marcus expression in the semiclassical limit . the model is parameterized by the vibrational normal mode frequencies , and the electronic energies and energy derivatives at a reference configuration .
the approach has been used by our group to compute state - to - state transition rates in semi - empirical models for organic semiconducting light - emitting diode and photovoltaics @xcite .
we recently made a significant breakthrough in using this approach by tying it it to a fully _ ab initio _ quantum chemical approach for determining the diabatic states and electron / phonon coupling terms allowing unprecedented accuracy and utility for computing state - to - state electronic transition rates .
our methodology consists of two distinct components .
the first is the use of a diabatization schemes for determining donor and acceptor states in a molecular unit .
the other is a projection scheme which enables us to analyze the contribution of vibrations in reactions .
similar decomposition schemes have been presented by burghardt @xcite and the approach used here builds upon the method given in ref . .
we recently benchmarked this approach against both the experimental rates and recent theoretical rates presented by subotnik _
_ @xcite and successfully applied the approach to compute state - to - state transition rates in series of pt bridged donor - acceptor systems recently studied by weinstein s group .
we review here these latter results along with the details of our methods .
we consider a generic model for @xmath7 electronic states coupled linearly to a phonon bath . taking the electronic ground state of the system as a reference and assuming that the electronic states are coupled linearly to a common set of modes , we arrive at a generic form for the hamiltonian , here written for two coupled electronic states : @xmath8 here
, the first term contains the electronic energies , @xmath9 and @xmath10 computed at a reference geometry typically that of the donor or acceptor state .
the second term represents the linearized coupling between the electronic and nuclear degrees of freedom given in terms of the mass - weighted normal coordinates @xmath11 .
the diagonal terms give the adiabatic displacement forces between the reference geometry and the two states .
if we choose one of the states as the reference state , then either @xmath12 or @xmath13 will vanish .
the remaining two terms correspond to the harmonic motions of the nuclear normal modes , given here in mass - weighted normal coordinates . in the normal mode basis ,
the hessian matrix , @xmath14 , is diagonal with elements corresponding to the normal mode frequencies , @xmath15 . we now separate eq .
[ ham1 ] into diagonal and off - diagonal terms @xmath16 and perform a polaron transform using the unitary transformation @xcite .
@xmath17 under which the transformed hamiltonian is written in terms of the diagonal elements @xmath18 with the renormalized electronic energies , @xmath19 and off - diagonal terms , @xmath20 in the transformed ( or dressed ) picture the electronic transition from state @xmath21 to @xmath22 is accompanied by the excitations of all the normal modes . at this point
it is useful to connect the various terms in the phonon - dressed hamiltonian with specific physical parameters .
first , the reorganization energy is given by @xmath23 where the @xmath24 are the huang - rhys factors for each phonon mode .
these are related to the franck - condon factor describing the overlap between the @xmath25 vibronic state in one electronic state with the @xmath26 vibronic state in the other .
likewise , the energy difference between the renormalized energy gaps is related to the driving force of the state - to - state transition , @xmath27 transforming to the interaction representation and performing a trace over the phonons gives the spectral density in terms of the autocorrelation of the electron - phonon coupling operators .
using the explicit form of the electron - phonon coupling operators , one can arrive at a compact expression for the autocorrelation function of the electron / phonon coupling @xmath28 here , @xmath29 is the electron - phonon coupling term in the heisenberg representation and @xmath30 denotes a thermal average over the vibrational degrees of freedom .
the remaining terms are constructed from the normal mode frequencies @xmath31 and electron / nuclear couplings @xmath32 viz .
@xmath33 finally , @xmath34 is the bose population of vibrational normal mode @xmath35 , @xmath36 the spectral density and golden - rule rate can then be obtained by fourier transform @xmath37 and @xmath38 the formalism presented above requires both diagonal ( @xmath39 ) and off - diagonal ( @xmath40 ) derivative couplings between adiabatic states . until recently , these couplings were difficult to impossible to compute for molecular systems using standard quantum chemical means .
we next discuss how we have used the edmiston - ruedenberg ( er ) localization scheme @xcite to estimate the couplings .
we also present how one can construct a reduced set of harmonic modes that fully capture the electron / nuclear coupling .
a workaround is to transform to a diabatic representation , whereby the hamiltonian is written as @xmath41 fig .
[ adiadiab ] shows a sketch of the adiabatic and diabatic potentials for a model two level system . while the adiabatic representation is precisely defined in terms of electronic eigenstates , the diabatic representation offers several advantages
first , the sharp derivative couplings that depend upon the nuclear velocity in the adiabatic representation are transformed to smoother diabatic couplings , @xmath42 that only depend upon the nuclear positions .
second , potential energy surfaces are smoother and the avoided crossing is eliminated . a number of diabatization approaches have been developed and the reader is referred to ref . for a general review .
the problem now is how to obtain the transformation matrix @xmath43 while a number of methods are available a straightforward approach is to eliminate derivative coupling mathematically by requireing @xmath44 however , this is computationally very expensive especially for
complex molecular systems , and exact solutions generally do not exist .
@xcite an alternative approach is to use physical intuition rather than a purely mathematical constraint to define the diabatic states .
the edmiston - ruedenberg ( er ) diabatization scheme is such a scheme and is based on the idea that the diabatic states can be obtained by maximizing the total electron repulsion between localized states , @xmath45 when the adiabatic ( and diabatic ) energy minima are far enough away from the crossing points and the mixing angles between the diabatic and adiabatic states is small , we can use the gradients of the adiabatic potentials to approximate the diabatic potentials .
thus , if we perform calculations at the optimized geometry of the final acceptor state ( _ i .
e. _ about @xmath46 in fig .
[ marcus ] ) , we can write the hamiltonian as @xmath47 where @xmath48 is the harmonic oscillator hamiltonian for the vibrational normal modes . the linear assumption amounts to performing a series expansion of the full , multi - dimensional coupling term and keeping only the lowest order terms
. systematic improvement can be made by including higher - order ( e.g. quadratic ) off - diagonal couplings .
however , this would involve a substantial increase in the complexity of the theory .
the linear assumption is reasonable so long as the mixing angle is small .
@xcite we obtain the diabatic couplings @xmath49 and the mixing angle @xmath50 via er localization and transform the electronic hamiltonian from the adiabatic basis to the diabatic basis _ viz .
_ @xmath51 the diabatic coupling is then given by @xmath52 we then diagonalize the electronic part and transform the electron / nuclear coupling back into the adiabatic basis . in doing so
, we obtain the hamiltonian in the form given in eq .
[ ham1 ] @xmath53 alternatively , one can use the generalized mulliken - hush model ( gmh ) @xcite which works well for linear systems but does not generalize easily to systems with more than two charge centres .
within gmh , the diabatic mixing is given by @xmath54 where @xmath55 is the vertical excitation energy , @xmath56 and @xmath57 are the dipole moments of corresponding adiabatic states , and @xmath58 is the transition dipole moment between two states .
er localization can be seen as a extension of gmh which overcomes some drawbacks of gmh @xcite .
both er and gmh require convergence of the initial and final reference states and have be used to compute the coupling terms required for the tclme approach given above.@xcite while the marcus expression is elegant in its simplicity in requiring three parameters that can be obtained experimentally , it masks a wealth of detail that underlie the quantum transition .
considerable insight into the state - to - state dynamics can be revealed by examining the nuclear motions driving and coupling the electronic states .
our approach is based on earlier work by our group @xcite and burghardt _ et al .
central to the theory is that there exists a collective nuclear displacement coordinate that connects the initial geometry of the donor to the final geometry of the acceptor .
however , until this work a general systematic approach for determining such motions did not exist . generally speaking
, this collective coordinate involves all nuclear degrees of freedom .
however , the form of the electronic hamiltonian in eq .
[ ham1 ] suggests that there exists a subset of motions that are specific modes that capture the majority of the electronic / nuclear coupling and give a dominant contribution to the collective reaction coordinate . within the linearized approximation for the electronic / nuclear coupling , we can write a force tensor @xmath59 where @xmath60 is the electronic / nuclear coupling term in eq .
if we consider each unique element @xmath61 to be linearly independent , but non - orthogonal force vectors , one can develop a projection operator scheme to to parse the @xmath62-dimensional linear vector space spanned by the mass - weighted normal mode vectors into two subspaces : one spanned by three vectors describing the coupling between the electronic states and the other spanned by the remaining @xmath63 dimensional space spanned by motions that do not couple the electronic states .
this subspace can be generated by defining a projection operator @xmath64 in which the summation is limited to linearly independent vectors .
here @xmath65 , @xmath66 is outer product , and @xmath67 is unitary operator .
this @xmath68 matrix projects out all normal modes that are directly coupled to the electronic degrees of freedom and its complement @xmath69 projects out all modes not directly coupled . by diagonalizing the matrix @xmath70 we obtain a transformation , @xmath71 , between the normal coordinates and a new set of orthogonal coordinates .
both @xmath72 and @xmath73 are @xmath68 matrices . however , for a two - state system , the former will have exactly @xmath74 non - trivial eigenvalues , @xmath75 , with corresponding eigenvectors , @xmath76 , whereas the latter will have exactly @xmath77 non - trivial eigenvalues , @xmath78 , and corresponding eigenvectors , @xmath79 .
this the full @xmath68 transformation is formed by joining the non - trivial vectors from the two respective subspaces @xmath80 .
the transformed electron - phonon coupling constants are given by projecting the couplings in the normal mode basis on to the new basis .
@xmath81 by examining the types of molecular motions that compose the @xmath82 subspace , we can gain a deeper understanding of the specific classes of internal motion that are directly involved with the electron transfer process .
in addition , we can gain a computational advantage since presumably this reduced set of modes give the dominant contribution to the electron - phonon coupling and autocorrelation function given as the kernel in eq .
[ gr - expression ] .
it is crucial to notice that the vectors given in eq .
[ eq : locaiham ] are _ not linearly independent_. consequently , special care must be taken to generate the reduced sub - space . to facilitate this
, we develop an iterative lanczos approach , taking the normalized vector @xmath83 as a starting point . as
above , we initialize each step indexed by @xmath84 , by defining a projection operator @xmath85 and its complement @xmath86 . for the @xmath84-th mode .
we also construct @xmath87 as the total projection operator for all @xmath88 modes .
we then project the hessian matrix @xmath89 into each subspace _ viz .
_ @xmath90 and diagonalize each to obtain eigenvalues and eigenvectors @xmath91 and @xmath92 respectively . as above , @xmath93 and @xmath94 are @xmath68 matrices .
the first set will have a single non - trivial eigenvalue and the second set will have @xmath95 non - trivial eigenvalues . as above
we collect the non - trivial eigenvectors associated with each to form the orthogonal transformation matrix @xmath96 and again transform the full hessian @xmath89 into this new vector space to form the @xmath68 matrix @xmath97 . at each step in the iteration ,
the transformed hessian , @xmath97 is in the form of a @xmath98 tri - diagonal submatrix in the upper - left part of the matrix and a diagonal submatrix in the lower - right .
for example , after @xmath99 iterations , the hessian matrix takes the form : @xmath100 \label{omega - prime}\end{aligned}\ ] ] we note that only the @xmath84-th mode is coupled the @xmath95 remaining modes .
since all of the transformations are orthogonal , diagonalizing @xmath97 at any point returns the original hessian matrix . to continue iterating , we take the @xmath84-th row of @xmath97 and zero the first @xmath84 elements @xmath101 this is the coupling between the upper tridiagonal block and the lower diagonal block .
we thus obtain a new vector @xmath102 which is then reintroduced into the iteration scheme . for the first iteration , @xmath103 is parallel to the bare electron - phonon coupling vector @xmath104 and the associated frequency is @xmath105 .
the subsequent iterations introduce corrections to this via phonon - phonon coupling mediated via the electronic couplings .
for example , for the @xmath99 iteration , we would determine the active vector space in terms of the upper - left 3@xmath106 block of the matrix in eq .
[ omega - prime ] .
@xmath107\end{aligned}\ ] ] diagonalizing @xmath108 returns a set of frequencies and associated eigenvectors which are then used to compute the electron - phonon couplings in this reduced active space .
after @xmath109 iterations , @xmath97 is a fully tridiagonal matrix and diagonalizing this returns the original normal mode basis . at any point along the way , we can terminate the iteration and obtain a reduced set of couplings .
since the lanczos approach uses the power method for finding the largest eigenvector of a matrix , it converges first upon the vector with the largest electron / nuclear coupling which we refer to as the `` primary mode '' .
subsequent iterations produce reduced modes with progressively weaker electron / nuclear couplings and the entire process can be terminated after a few iterations .
after @xmath84-steps , the final electron - phonon couplings are then obtained by projecting the original set of couplings ( in the normal mode basis ) into the final vector space . for small systems , we find that accurate rates can be obtained with as few as 2 - 3 modes are sufficient to converge the autocorrelation function in eq .
the weinstein group at university of sheffield reported recently upon a series of donor - bridge - acceptor ( dba ) molecular triads whose electron transfer ( et ) pathways can be radically changed - even completely closed - by infrared light excitation of specific intramolecular vibrations @xcite .
the triads consist of a phenothiazine - based ( ptz ) donor linked to a naphthalene - monoimide ( nap ) acceptor via a pt - acetylide bridging unit.@xcite the structures of the triads are given in fig .
all three systems undergo a similar sequence of processes after following uv excitation : electron transfer from the pt - acetylide center to the nap acceptor , resulting in a charge - transfer state , @xmath110 , which due to strong spin - orbit coupling efficiently populates triplet charge - transfer state , ct .
further electron transfer leads to a fully charge - separated state ( css ) @xmath111 with the electron and hole localized on the acceptor and donor units respectively .
the charge transfer state can also undergo charge recombination to form a localized triplet exciton on the nap unit ( @xmath112nap ) , or the ground state .
both css and @xmath112nap decay to the singlet ground state on the nanoseconds and sub - millisecond time scales , respectively .
we also show in fig .
[ fig2]b the triplet energy along a linear interpolation coordinate connecting the @xmath112nap minimum energy geometry to the ct minimum energy geometry . between the two is a significant energy barrier reflecting the relative rotation of the nap and the ptz groups about the cc - pt - cc axis . the uv pump - ir push experiments performed on these triads showed that ir - excitation of bridge vibrations after the initial uv pump radically changes the relative yields of the intermediate states .
subsequent excitation of the -cc - pt - cc- localized vibrations by a timed ir pulse in the ct state of ptz - complex * 2 * at 1 ps after the uv pump decreases the yield of the css state , whilst increasing that of the @xmath112nap state .
ir - excitation in the course of electron transfer has caused a 100% decrease in the css yield in * 1 * , approximately 50% effect in * 2 * , and no effect in * 3*. this demonstration of control over excited state dynamics strongly suggests that the acetylide stretching modes are significantly involved in the electron / nuclear coupling in these systems and play central roles in the electron - transfer process .
the transferred charge can undergo either further separation to form the full charge - separated state ( css ) , or recombine to form a localized excitation , @xmath112nap .
both eventually decay to ground state .
_ showed that if a judiciously chosen ir pump is applied to excite the c@xmath113c bond after the initial uv excitation , the yield of intermediate states can be radically changed .
for example , when a ir pump with frequency = 1,940 @xmath114 is applied to excite the c@xmath113c in ptz - ch@xmath115-pt - nap , 1 ps after the uv pump , the yield of the electron transfer state decrease from 32% to 15% , while that of the @xmath112nap increases from 29% to 46% .
the most striking observation is that when a 1,908 @xmath114 ir pulse is applied to ptz - ch@xmath115-pt - nap 2 ps after uv excitation , the ct @xmath116 css step is completely switched off . @xcite .
quantum chemical analysis indicates that the electron - transfer rate is largely influenced by chemical modification of the ptz donor . from ptz - ch@xmath117-pt - nap to ptz - pt - nap , to ome - ptz - pt - nap , the donor strength increases , which increases the energy gap between ct and css states .
the driving force ( @xmath118 ) for the ct @xmath116 css transfer also increases from 0.2 ev in ptz - ch@xmath117-pt - nap , to 0.4 ev in ptz - pt - nap , to 0.6 ev in ome - ptz - pt - nap .
large @xmath118 accelerates the ct decay and hence decreases the lifetime of ct state .
comparing ptz - pt - nap and ptz - ch@xmath117-pt - nap , ct transfer to both charge separation and recombination slows down by a factor of about 5 ( the lifetime of ct increases from 3.3 to 14 ps and css from 190 ps to 1 ns ) . by appending methoxy groups to the ptz
, the donor strength is increased , and reaction is accelerated . as the result , the lifetime of ct in ome - ptz - pt - nap is further reduced to 1 ps .
proposed that the effect of infrared control is caused by the fact that the distance between ct energy minimum and the intersection of ct and css potential energy surfaces is small .
for all three molecules , two pess intersect where c@xmath113c bond is slightly longer than the equilibrium length .
when c@xmath113c bond gets excited , it elongates and helps molecules to pass the intersection .
if the energy gap between intersection and equilibrium geometry is much larger than c@xmath113c vibrational energy , the dynamics is barely affected ; if the energy gap is small , the vibrational excitation can radically change the dynamics .
@xcite .
we focus our attention on the ptz system and anticipate that the other systems in this study will exhibit similar behaviour due to the overall similarity of the various donor groups.@xcite for purposes of facilitating the calculations , the molecular structures are simplified such that the p(bu)@xmath119 moieties and octyl chain of the nap group were truncated to -ph@xmath119 and a single methyl group , respectively . in all quantum chemical calculations , we used the sdd pseudo - potential for pt and 6 - 31g(d , p ) for the other atoms .
we also used the polarizable continuum model ( pcm ) to account for the dichloromethane solventas used in ref .. the transition dipole moments and electron / hole distributions surfaces were calculated using the multiwfn ( v3.3.8 ) program.@xcite an energy level diagram based upon our calculations is sketched in fig .
[ ene_diagram]a together with the corresponding electron / hole distribution plots . [ cols="^,^,^,^,^ " , ] to compute the rates using the tclme expression ( eq .
[ gr - expression ] ) , we begin by computing the electron / nuclear correlation function and compare its convergence with respect to the number of lanczos modes . recall that the lanczos modes are determined by an iterative ranking algorithm that identifies superpositions of normal modes that optimize the electron / phonon coupling .
[ corrt1t3 ] gives a summary of these numerical tests in which we compute @xmath120 vs. time with an increasing number of lanczos modes . in all cases
, we compare to the `` exact '' result in which all nuclear modes were used .
the top two figures ( fig .
[ corrt1t3]a and b ) use the @xmath121nap as the reference geometry . in these cases ,
convergence of @xmath120 with respect to the number of modes proved to be problematic for both transitions considered .
correspondingly , the rates computed using this geometry also compare poorly against the observed experimental rates , although are an order of magnitude closer than marcus rates .
we speculate that this may signal a break - down in the condon approximation which insures separability between nuclear and electronic degrees of freedom . in table
[ ratesummary ] , we summarize both the experimental and computed state to state rates for the ptz system .
given the complexity and size of the system , overall the numerical rates computed using the exact tclme approach are in quantitative agreement with the experimental rates , particularly for those using the ct geometry as a reference point ( cf fig .
[ corrt1t3]c and d ) .
we note that fewer projected modes ( 30 - 50 ) are needed to converge the correlation function out to the first 50 fs when using the ct - geometry . furthermore , while marcus rate for the ct@xmath122 css transition agrees with the exact tclme result , it misses the ct @xmath123nap experimental rate by 4 orders of magnitude whereas the tclme rate is in much better agreement with the experimental rate .
if we compare the exact tclme rate , which uses the full set of normal modes in constructing the @xmath120 correlation function , to the rate computed used _
only _ the plm ( tclme+plm ) , for both the ct@xmath122 css and ct @xmath123nap rates , the single mode approximation is within 86% of the exact result .
this indicates that while multiple vibrational normal modes contribute to the electronic coupling , the linear combination identified by the projection algorithm carries the vast majority of the electron / phonon coupling .
this is consistent with our previous study of triplet energy transfer in small donor - bridge - acceptor systems .
@xcite as discussed earlier , our ranking algorithm allows us to rapidly determine the vibrational motions that optimize the electron / nuclear couplings .
in addition to providing an accurate way to compute rate constants , they provide additional insight into actual dynamics . here
, we shall focus upon the transitions originating from the ct geometry .
generally speaking , the highest ranked mode , termed the `` primary lanczos mode '' ( plm ) , captures much of the short - time dynamics of the transitions . in fig .
[ corrt1t3 ] , we show the electronic coupling correlation functions computed using different numbers of projected modes for all four transitions . for the ct @xmath116 @xmath112nap transition , the primary mode resembles the exact initial dynamics for the first 10 fs and roughly 10 or so modes are sufficient to converge the correlation function out to times longer than the correlation time . in table
[ ratesummary ] we see that for the ct geometry , the primary mode approximation is sufficient to obtain accurate rate constants .
on the contrary , it takes considerably more modes modes to recover the full correlation function for transition originating from the @xmath112nap geometry .
[ projt1](a - d ) shows the projection of the primary mode identified for each transition onto the normal vibrational modes of the originating state , _
i. e. _ , the primary modes calculated at ct geometry are projected onto the normal modes of the ct state , and those at the @xmath112nap geometry are projected onto the normal modes of the @xmath112nap state . in all four cases , the primary mode is dominated by symmetric and anti - symmetric contributions from the c@xmath113c displacements . while both transitions involve acetylene bond - stretching motions , the ct@xmath122css transition involves only the _ symmetric _ combination , whereas the ct@xmath124nap involves both the _ symmetric _ and _ anti - asymmetric _ combination .
it is tempting to conclude from this that that the secondary ir push used in the experiments preferentially excites the anti - symmetric mode and thus selectively enhances the ct@xmath124nap transition .
in fact , the computed ir oscillator strength of the anti - symmetric mode is an order of magnitude greater than the symmetric mode .
similarly , from experiment , the anti - symmetric normal mode extinction coefficient is 3 times larger than that for the symmetric normal mode .
however , the time - scale for the ir excitation is sufficient long enough that _ both _
symmetric and anti - symmetric cc modes are expected to be equally populated by the ir push pulse . in the ct@xmath124nap transition ,
both types of acetylene stretching motions ( symmetric and anti - symmetric ) contribute more or less equally to the electronic coupling while in the ct@xmath122css transition , _ only _ the symmetric acetylene motion carries the majority of the coupling .
this mechanism can be rationalized by the way the vibrational populations enter into our expression for electron / nuclear coupling correlation function in eq .
[ corrf ] . in principle , the expression was derived assuming a thermal population of the vibrational modes . however , if we assume that the role of the ir pulse is to excite the @xmath125 stretching modes by one vibrational quantum , then the value of @xmath126 appearing in eq .
[ corrf ] for those modes should be increased to @xmath127 .
consequently , driving these modes with the ir pulse _ increases _ the total electronic coupling , consistent with the experimental observation that ir excitation following formation of the ct states accelerates the ct@xmath124nap transition relative to the ct@xmath122css transition .
we here a review of our work in developing new tools for analysing electronic transitions in complex molecular systems .
central to our work is the notion that one can systematically identify a subset of vibrational modes that capture the majority of the electronic coupling to the nuclear motions .
these primary modes capture the short - time dynamics with sufficient accuracy for computing the salient correlation and response functions necessary for evaluating the golden - rule rates for state - to - state transitions . while not a central theme to this review , our time - convolutionless master equation method can be used for computing multi - state transitions and in cases where the state to state rates are time - dependent .
@xcite we believe that the key to understanding and ultimately controlling electron transfer pathways in a complex molecular species is through vibronic coupling .
the approach we have delineated in this article offers a systematic way to deduce a subset of nuclear motions that are most responsible for driving electronic transitions . when paired with the tclme approach for computing the state - to - state transitions
, we can obtain rate constants that are in quantitative agreement with experimental rates and probe deeper into the dynamics to understand which specific types of nuclear motions are involved in a given transition .
the algorithm illustrated here in the example of photo - induced charge - transfer should be of considerable utility for understanding of a multitude of light - induced reactions where several electronic states are involved in ultrafast transformations .
xy thanks tian lu for help with multiwfn .
the work at the university of houston was funded in part by the national science foundation ( che-1362006 , mri-1531814 ) and the robert a. welch foundation ( e-1337 ) .
we thank the weinstein group at the university of sheffield for sharing their experimental results and many detailed conversations regarding the ptz - pt - nap triad .
30ifxundefined [ 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 * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ , , vol .
( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) | we present a novel _ ab initio _
approach for computing intramolecular charge and energy transfer rates based upon a projection operator scheme that parses out specific internal nuclear motions that accompany the electronic transition .
our approach concentrates the coupling between the electronic and nuclear degrees of freedom into a small number of reduced harmonic modes that can be written as linear combinations of the vibrational normal modes of the molecular system about a given electronic minima . using a time - convolutionless master - equation approach , parameterized by accurate quantum - chemical methods , we benchmark the approach against experimental results and predictions from marcus theory for triplet energy transfer for a series of donor - bridge - acceptor systems .
we find that using only a single reduced mode termed the `` primary '' mode , one obtains an accurate evaluation of the golden - rule rate constant and insight into the nuclear motions responsible for coupling the initial and final electronic states .
we demonstrate the utility of the approach by computing the inelastic electronic transition rates in a model donor - bridge - acceptor complex that has been experimentally shown that its exciton transfer pathway can be radically modified by mode - specific infrared excitation of its vibrational mode . |
one of the important nonlinear processes affecting pulse propagation in massive wavelength division multiplexing ( wdm ) optical fiber communication systems is due to inter - pulse raman - induced crosstalk @xcite . in this process , which takes place during collisions between pulses from different frequency channels ( interchannel collisions ) , energy is transferred from high - frequency pulses to low - frequency pulses .
it is known that the magnitude of the raman - induced energy exchange in a single interchannel collision is independent of the frequency difference between the channels .
consequently , the magnitude of the cumulative energy shifts for a given pulse grows with the square of the number of channels , a result that is valid for linear transmission @xcite , conventional soliton transmission @xcite , and strongly dispersion - managed ( dm ) soliton transmission @xcite .
therefore , in a 100-channel system , for example , the raman crosstalk effects can be larger by a factor of @xmath1 compared with a two - channel system operating at the same bit rate per channel .
early studies of raman crosstalk in wdm transmission focused on the dependence of the energy shifts on the total number of channels @xcite , as well as on the impact of energy depletion @xcite and group velocity dispersion @xcite on the dynamics . later on attention turned to the combined effects of raman crosstalk and bit - pattern randomness in the on - off - keying ( ook ) transmission scheme , and it was found that the probability density function ( pdf ) of the pulse amplitudes is lognormal @xcite .
this finding means that the @xmath2th normalized moments of the amplitude grow exponentially with both propagation distance and @xmath3 .
furthermore , in several studies of conventional soliton transmission it was found that the dynamics of the frequency shift is strongly coupled to amplitude dynamics , and as a result , the @xmath2th normalized moments of the raman - induced self and cross frequency shifts also grow exponentially with propagation distance and @xmath3 ( see refs .
this intermittent dynamic behavior has important practical consequences , by leading to relatively high bit - error - rate ( ber ) values at intermediate and large propagation distances @xcite .
raman crosstalk effects were also recently investigated in hut - skipped amplified wdm transmission @xcite , in cable television overlay passive optical networks @xcite , in optical code - division multiple - access transmission systems @xcite , and in conjunction with four - wave - mixing @xcite .
one of the ways to overcome the detrimental effects of raman crosstalk on massive wdm transmission is by replacing the ook scheme by alternative encoding schemes , which are expected to be less susceptible to these effects .
the differential phase shift keying ( dpsk ) scheme , in which all time slots are occupied and the information is encoded in the phase difference between adjacent pulses , is one of the promising encoding methods that have attracted much interest in recent years @xcite . in dpsk transmission the raman - induced amplitude dynamics becomes ( approximately ) deterministic , and important questions arise regarding the character of this dynamics .
one major question concerns the possibility to achieve a stable equilibrium state for the amplitudes in all channels .
the study reported in ref .
@xcite demonstrated that this is not possible in unamplified optical fiber lines .
however , later experiments showed that the situation is quite different in amplified wdm transmission @xcite .
more specifically , it was found that the introduction of amplification into the system significantly reduces the raman - induced energy shifts . in the present paper
we suggest a dynamical explanation for this important experimental observation .
moreover , we demonstrate the robustness of dpsk transmission against inter - pulse raman crosstalk effects by showing that equilibrium states with non - zero amplitudes in all channels do exist , and by proving the stability of the equilibrium states . in the present study we consider optical solitons as an example for the pulses carrying the information for the following reasons .
first , as mentioned above , the raman - induced energy exchange in pulse collisions is similar in linear transmission , conventional soliton transmission , and strongly dm soliton transmission .
second , propagation of conventional solitons through an optical fiber is described by the nonlinear schrdinger ( nls ) equation , which is an integrable model @xcite . due to the integrability of the model , and to the fact that optical solitons are stable stationary solutions of the nls equation
, the derivation of the model for the raman - induced amplitude dynamics can be done in a rigorous manner .
third , conventional optical solitons have traditionally been considered as excellent candidates for information transmission in high - speed optical fiber lines and in all - optical networks @xcite .
furthermore , state - of - the - art transmission experiments already use all - raman distributed amplification @xcite , which is the most suitable amplification scheme for conventional solitons . since we consider transmission systems where the pulses in each frequency channel are well - separated , intrachannel four - wave - mixing ( fwm ) effects are negligible . in addition , we assume that loss is compensated by distributed raman amplification . it is a well - known fact in soliton theory that in the absence of loss interchannel fwm products completely vanish after the collisions , see ref .
@xcite for theory and refs .
@xcite for numerical simulations with the nls equation with and without delayed raman response .
we remark that even in systems employing lumped amplifiers ( and non - overlapping pulses ) , interchannel fwm can be significantly reduced by dispersion - management .
this is true both in the soliton regime @xcite and in the linear regime @xcite .
in contrast , the energy shift in a single collision between two strongly dm solitons is given by an expression with exactly the same form as the expression obtained for conventional solitons @xcite , i.e. , strong dispersion - management does not reduce raman crosstalk . in deriving the model for raman - induced amplitude dynamics
we fully take into account pulse walk - off .
in addition , we assume that the pulse sequences in all frequency channels are deterministic and that the sequences are either infinitely long or are subject to periodic temporal boundary conditions .
the first setup approximates long - haul transmission , while the second one corresponds to a closed fiber loop experiment .
we also assume that the constant net gain / loss in each channel is determined by the difference between distributed amplifier gain and fiber loss .
notice that in this feature our model is fundamentally different from the model derived in ref @xcite .
indeed , since in the latter model all channels experience net loss , it does not support an equilibrium state with non - zero amplitudes in all channels .
in contrast , in our model the net gain / loss of some channels is positive while for other channels it is negative , which is the underlying reason for the existence of equilibrium states with non - zero amplitudes in all channels . our model for the raman crosstalk dynamics in an @xmath0-channel system
consists of a system of @xmath0 coupled nonlinear ordinary differential equations ( odes ) for the amplitudes in different channels .
the system of coupled odes can be described in the jargon of population dynamics theory as an @xmath0-dimensional predator - prey model @xcite . after obtaining the model
we look for equilibrium states with non - zero amplitudes in all channels and establish their stability with respect to deviations of the initial amplitudes from the equilibrium values .
we also investigate the dynamic behavior induced by such deviations . in actual optical fiber transmission systems pulse dynamics
can be influenced by physical processes other than raman crosstalk .
it is therefore important to understand the manner in which these additional processes perturb the raman - induced amplitude dynamics . in the present study
we investigate the effects of three perturbations due to cross phase modulation ( xpm ) , raman self frequency shift ( sfs ) , and raman cross frequency shift ( xfs ) . for each of these physical processes
we construct the perturbed model describing amplitude dynamics in an @xmath0-channel system and investigate the existence of equilibrium states with non - zero amplitudes in all channels .
furthermore , we study the stability of the equilibrium state and the typical dynamic behavior of the amplitudes in two - channel systems in the presence of each of the three perturbations .
the rest of the paper is organized as follows . in section [ derivation ]
we derive the unperturbed model for raman - induced amplitude dynamics in wdm systems with @xmath4 channels , and in section [ stability ] we find the equilibrium states of the model and investigate their stability . in section [ simulations ]
we validate the predictions of section [ stability ] by numerical simulations .
the perturbed models with xpm , raman sfs , and raman xfs , are studied in sections [ xpm ] , [ sfs ] , and [ xfs ] , respectively , for two - channel systems .
section [ conclusions ] is reserved for conclusions . in appendix [ appendb ]
we obtain the perturbed models with xpm , raman sfs , and raman xfs in wdm transmission with @xmath4 channels .
propagation of short pulses of light through an optical fiber in the presence of delayed raman response is described by the following perturbed nls equation @xcite : @xmath5 where @xmath6 is proportional to the envelope of the electric field , @xmath7 is propagation distance and @xmath8 is time in the retarded reference frame .
the term @xmath9 represents the first order approximation for the fiber s delayed raman response @xcite and @xmath10 is the raman coefficient @xcite . when @xmath11 , the single - soliton solution of eq .
( [ cfs1 ] ) in a frequency channel @xmath12 is given by @xmath13 where @xmath14 , @xmath15 , and @xmath16 and @xmath17 are the soliton amplitude , phase and position , respectively .
let us describe the main raman - induced effects on a single collision between a soliton from the @xmath18th frequency channel and a soliton from the @xmath19th frequency channel .
we assume that @xmath20 , which is the typical situation in many wdm transmission systems even for adjacent channels @xcite . under this assumption
one can show that the most important effect of delayed raman response on the collision is an @xmath21 change in the soliton amplitude @xcite @xmath22 where the coupling constants @xmath23 depend on the specific approximation for the raman gain curve .
notice that due to the inclusion of the @xmath23 factors , eq .
( [ crosstalk3 ] ) does not rely on the triangular approximation for the raman gain curve .
if we adopt the triangular approximation , we obtain that @xmath24 for any @xmath18 and @xmath19 .
the effects of the collision in order @xmath25 will be described in section [ perturbed ] , where we obtain the perturbed models for amplitude dynamics . since @xmath20 ,
effects of order @xmath26 and higher will be neglected .
in addition , third order dispersion and self - steepening are neglected since the collision - induced effects of these conservative perturbations on the soliton amplitude and frequency are of higher order in both the parameter @xmath27 characterizing the perturbative process and @xmath28 ( see , e.g. , refs .
@xcite ) . consider now wdm transmission systems with @xmath4 channels and frequency difference @xmath29 between adjacent channels .
our model , which takes into account pulse walk - off , is based on the following assumptions .
( 1 ) the soliton sequences in all channels are deterministic in the sense that all time slots are occupied and each soliton is located at the center of a time slot of width @xmath30 . furthermore ,
when considering amplitude dynamics , the amplitudes are equal for all pulses from the same frequency channel , but are not necessarily equal for pulses from different channels .
this setup corresponds , for example , to return - to - zero ( rz ) transmission with differential - phase - shift - keying @xcite .
( 2 ) the sequences are either ( a ) infinitely long , or ( b ) subject to periodic temporal boundary conditions .
notice that setup ( a ) is an approximation for long - haul transmission systems , while setup ( b ) is an approximation for closed fiber - loop experiments .
( 3 ) the gain / loss in each channel is determined by the difference between distributed amplifier gain and fiber loss .
in particular , for some channels this difference can be slightly positive , resulting in small net gain , while for other channels this difference can be slightly negative , resulting in small net loss .
we emphasize that in this feature our model is essentially different from the models that are usually considered in studies of raman amplification schemes , where it is assumed that _ all _ pumps experience net loss @xcite . to obtain the dynamic equation for the amplitude of the @xmath18th - channel solitons we note that the distance traveled by these solitons while passing two successive solitons in the @xmath31 or @xmath32 channels is @xmath33 .
we denote by @xmath34 the location of the @xmath35th collision of a given @xmath18th - channel soliton with solitons in the @xmath32 or @xmath31 channel . using eq .
( [ crosstalk3 ] ) and summing over all collisions occurring within the interval @xmath36 $ ] , where @xmath37 , we obtain @xmath38 the constant @xmath39 on the right hand side of eq .
( [ crosstalk4 ] ) is the net gain / loss coefficient for the @xmath18th channel , which is assumed to be independent of @xmath7 . due to the periodicity of the pulse sequences the same equation is satisfied by the amplitudes of all @xmath18th - channel solitons .
moreover , the same equations with different @xmath18 values , where @xmath40 , describe the dynamics of the soliton amplitudes in all channels . going to the continuum limit we obtain @xmath41 , \label{crosstalk5}\end{aligned}\ ] ] where @xmath42 and @xmath40 .
the system ( [ crosstalk5 ] ) gives a complete description of the dynamics of the @xmath43 s .
notice that apart from the important fact that some gain / loss coefficients can be positive while others can be negative , the system ( [ crosstalk5 ] ) is similar to the one obtained in ref .
@xcite for the raman - induced amplitude dynamics of continuous waves in unamplified wdm transmission .
thus , our model also describes the raman - induced crosstalk dynamics of continuous waves in amplified wdm transmission systems .
in optical fiber communication systems it is usually desired to achieve a steady state in which the pulse amplitudes in all channels are equal and constant ( independent of @xmath7 ) @xcite .
we therefore look for an equilibrium state of the system ( [ crosstalk5 ] ) in the form @xmath44 for @xmath45 . setting the right hand sides of ( [ crosstalk5 ] ) equal to zero we arrive at @xmath46 therefore
, the gain required to maintain an equilibrium state with equal amplitudes is _ not _ flat with respect to frequency . instead
, the net gain / loss coefficient of high - frequency channels should be positive , while that of low - frequency channels should be negative .
in other words , high - frequency channels should be overamplified , whereas low - frequency channels should be underamplified compared with the reference ( @xmath47 ) channel . substituting eq .
( [ crosstalk6 ] ) into eq .
( [ crosstalk5 ] ) we arrive at a slightly simpler form of the model , which is convenient for analysis and numerical simulations , @xmath48 equation ( [ crosstalk7 ] ) can be described in population dynamics terminology as a predator - prey system with @xmath4 species @xcite .
the equilibrium states of eq .
( [ crosstalk7 ] ) with non - zero amplitudes are determined by @xmath49 the trivial solution of eq .
( [ crosstalk7a ] ) , that is , the solution with @xmath50 for @xmath51 , corresponds to the equilibrium state of eq .
( [ crosstalk7 ] ) with equal non - zero amplitudes .
however , due to the anti - symmetry of @xmath52 with respect to an interchange of @xmath18 and @xmath19 , eq .
( [ crosstalk7a ] ) has infinitely many non - trivial solutions , and these correspond to equilibrium states of eq .
( [ crosstalk7 ] ) with unequal non - zero amplitudes @xcite . in the case where the raman gain curve is described by the triangular approximation it is straightforward to show that the non - trivial equilibrium states of eq .
( [ crosstalk7 ] ) are determined by the following two equations : @xmath53 therefore , in this case the equilibrium states lie on a @xmath54-dimensional plane .
thus , for a 3-channel system , for example , the equilibrium states lie on the line segment @xmath55 , where @xmath56 .
we now prove stability of all equilibrium states of eq .
( [ crosstalk7 ] ) with non - zero amplitudes , @xmath57 , by constructing a lyapunov function for this equation . for this purpose
we consider the function @xmath58 , \label{crosstalk8}\end{aligned}\ ] ] where @xmath59 . taking the derivative of @xmath60 along trajectories of the system ( [ crosstalk7 ] )
we obtain @xmath61 from eq .
( [ crosstalk7a ] ) it follows that @xmath62 . using this relation together with eq .
( [ crosstalk9 ] ) and the anti - symmetry of @xmath52 we arrive at @xmath63 for @xmath64 .
it is straightforward to show that each term @xmath65 $ ] on the right hand side of eq .
( [ crosstalk8 ] ) satisfies @xmath66 for any @xmath67 , where @xmath68 only at @xmath69 .
therefore , @xmath70 for any vector @xmath71 for which @xmath67 for @xmath45 , where equality holds only at equilibrium points . combining this result with the result @xmath72 along trajectories of the system
we conclude that @xmath73 is a lyapunov function of the system ( [ crosstalk7 ] ) and therefore the equilibrium states @xmath74 are stable @xcite .
we note that since @xmath72 rather than @xmath75 this stability means that the values of @xmath76 are bounded for any @xmath7 but do not tend to @xmath77 for large propagation distance @xmath7 .
thus , the typical dynamics of the amplitudes @xmath76 for initial conditions that are off the equilibrium point is oscillatory .
notice that the stability is independent of the exact specification of the @xmath23 values , and therefore the equilibrium states are stable irrespective of the specific details of the approximation for the raman gain curve .
notice also that eq .
( [ crosstalk8 ] ) actually provides @xmath78 independent conserved quantities for the model ( [ crosstalk7 ] ) , where @xmath79 is the dimension of the solution space of eq .
( [ crosstalk7a ] ) .
for example , if the raman gain is described by the triangular approximation , @xmath80 , and there are @xmath81 conserved quantities in total . in order to check the predictions of the previous subsection we solve eq .
( [ crosstalk7 ] ) numerically by employing a fourth - order runge - kutta method . as a concrete example for the physical parameter values we consider a 101-channel system operating at 40 gbits / s per channel with frequency spacing @xmath82 ghz and dimensionless time slot width @xmath83 . with this choice
the pulse width is 5 ps , @xmath84 , @xmath85 , and @xmath86 .
these parameter values are typical for several state - of - the - art massive multichannel transmission experiments , see ref . @xcite and references therein .
taking @xmath87 and @xmath88 we obtain @xmath89 mw for the soliton peak power .
the dimensionless final propagation distance is taken as @xmath90 corresponding to @xmath91 km , but the main features of the dynamics can already be observed at considerably shorter distances .
we also choose @xmath92 , so that the trivial equilibrium state is @xmath93 , for all @xmath18 . to illustrate the impact of raman crosstalk on _ massive _ wdm transmission we focus attention on amplitude dynamics of solitons from faraway channels .
thus , we start by considering a two - channel system consisting of the reference channel ( @xmath47 ) and the highest - frequency channel @xmath94 . employing eq .
( [ crosstalk7 ] ) to this two - channel system while adopting the triangular approximation for the raman gain curve we obtain @xmath95 the solution of eq .
( [ simu1 ] ) for @xmath96 and @xmath97 is @xmath98- \ln\left[\eta_{0}(z)\right]=\kappa , \label{simu1a}\end{aligned}\ ] ] where the constant @xmath99 is determined by the values of @xmath100 and @xmath101 .
the @xmath7-dependence of the amplitudes @xmath102 and @xmath103 is shown in fig .
[ fig1 ] ( a ) for the initial condition @xmath104 and @xmath105 .
it is clearly seen that the amplitudes oscillate about their equilibrium value @xmath92 .
the dimensionless oscillation period is @xmath106 , corresponding to 2090 km @xcite .
similar oscillatory behavior is observed for other choices of the initial amplitudes .
figure [ fig1 ] ( b ) shows the phase portrait for the system @xmath107 .
since all trajectories are closed curves we conclude that the dynamics is indeed periodic and that the equilibrium state is stable .
next we consider a three - channel system consisting of the channels @xmath47 and @xmath108 . within the raman triangular approximation the amplitude dynamics
is described by @xmath109 figure [ fig3 ] ( a ) shows the phase portrait .
the solid line is the line segment of stable equilibrium points @xmath110 , where @xmath111 .
since all trajectories are closed orbits centered about the equilibrium line we conclude that the equilibrium states are indeed stable and that the amplitudes exhibit periodic oscillations .
this oscillatory dynamics is illustrated in fig .
[ fig3 ] ( b ) for the initial condition @xmath112 , @xmath105 , @xmath113 . as another example for possible dynamic scenarios exhibited by the unperturbed model ( [ crosstalk7 ] ) we study a four - channel system consisting of the channels @xmath114 and @xmath115 . as we demonstrate below
, a four - channel setup represents the simplest case where deviations from the triangular approximation for the raman gain curve lead to the emergence of new dynamical features .
we consider the following sets of values for the coupling constants @xmath23 : @xmath116 and @xmath117 , where @xmath118 for set ( 1 ) and @xmath119 for set ( 2 ) .
thus , set ( 1 ) corresponds to the triangular approximation , whereas set ( 2 ) represents a 10@xmath120 deviation from the triangular approximation for @xmath121 . for these values of the coupling constants
the dynamics of the amplitude is described by @xmath122 , \nonumber \\ & & \frac{d \eta_{16}}{dz}= 32c\eta_{16}\left(2+\eta_{48}-\eta_{-16}-2\eta_{-48}\right ) , \nonumber \\ & & \frac{d \eta_{-16}}{dz}= 32c\eta_{-16}\left(-2 + 2\eta_{48}+\eta_{16}-\eta_{-48}\right ) , \nonumber \\ & & \frac{d \eta_{-48}}{dz}= 32c\eta_{-48}\left[-(6 + 3p)+3(1+p)\eta_{48}+2\eta_{16}+\eta_{-16}\right ] .
\label{simu7}\end{aligned}\ ] ] we solve the system ( [ simu7 ] ) numerically with @xmath118 and @xmath119 and with the initial condition @xmath123 , @xmath124 , @xmath125 , and @xmath126 .
the results of our numerical simulations are presented in figs .
[ fig5 ] and [ fig6 ] . as can be seen , in both cases the amplitudes exhibit oscillatory dynamics .
however , in the triangular approximation case ( @xmath118 ) there is only a single period for the oscillations , whereas in the non - triangular approximation case ( @xmath119 ) two very different oscillation periods are observed .
the first period is not significantly different from the oscillation period for @xmath118 , while the second period is much longer than the first one ( see fig .
[ fig6 ] ) .
in fact , in order to verify that the amplitude dynamics is indeed periodic for @xmath119 one has to carry out the simulation up to a final propagation distance of @xmath127 .
notice that our choice of 40 gbits / s per channel systems is mainly for reasons of convenience , since with this value we can illustrate the oscillatory nature of the raman - induced amplitude dynamics already with two and three channels . since current soliton - based systems work at 10 gbits / s per channel it is useful to examine one example for the amplitude dynamics in such systems .
we therefore consider a 361-channel system operating at 10 gbits / s per channel with frequency spacing @xmath128 ghz and dimensionless time slot width @xmath83 .
for these parameter values the pulse width is 20 ps , @xmath84 , @xmath129 , and @xmath130 .
taking @xmath131 and @xmath132 we obtain @xmath133 mw for the soliton peak power .
the dimensionless final propagation distance is taken as @xmath134 corresponding to @xmath91 km , but the main dynamical features can already be observed at shorter distances .
as before we choose @xmath92 , so that the trivial equilibrium state is @xmath93 , for all @xmath18 .
to illustrate the dynamics we consider a seven - channel system consisting of the channels @xmath47 , @xmath135 , @xmath136 , and @xmath137 and adopt the triangular approximation for the raman gain curve .
the corresponding system of equations for amplitude evolution is solved numerically with the initial condition @xmath138 , @xmath139 , @xmath140 @xmath141 , @xmath142 , @xmath143 , and @xmath144 .
the results are shown in fig .
[ 7channels ] .
one can see that pulse amplitudes in all channels exhibit oscillatory behavior similar to the one observed in figures [ fig1 ] and [ fig3 ] for the 40 gbits / s per channel system .
thus , equilibrium states with non - zero amplitudes in all channels are stable in both 10 and 40 gbits / s per channel systems , in agreement with the predictions in section [ stability ] .
we now turn to study the effects of perturbations on the model described by eq .
( [ crosstalk7 ] ) .
we focus attention on the effects of cross phase modulation ( xpm ) , raman self frequency shift ( sfs ) , and raman cross frequency shift ( xfs ) .
it is well - known that each of these processes by itself does not change the soliton amplitude .
thus , in the absence of raman crosstalk the soliton amplitudes are constant and there is no amplitude dynamics to be considered .
as we show below xpm , raman sfs , and raman xfs affect amplitude dynamics by leading to @xmath7-dependence of the collision rate . for these reasons
, the three processes can be considered as perturbations to the model ( [ crosstalk7 ] ) . for each of the three processes we develop a perturbed model for amplitude dynamics and analyze the stability of the equilibrium state and the dynamic behavior in the new model .
here we choose to concentrate on a two - channel system consisting of channels @xmath47 and @xmath145 since such treatment is sufficient in order to uncover the main changes in the dynamics compared with the unperturbed model . without loss of generality
we assume @xmath146 for the coupling constant .
the general forms of the perturbed models in wdm systems with @xmath4 channels are obtained in appendix [ appendb ] . the xpm - induced position shift experienced by a soliton in the @xmath18th channel as a result of a single collision with a @xmath19th - channel soliton
is given by ( see , e.g. , refs .
@xcite ) @xmath147 our goal is to obtain a perturbed model for the dynamics of the amplitudes @xmath102 and @xmath148 in the two - channel system , which takes into account the xpm - induced position shift .
it is clear that the main effect of the position shift ( [ crosstalk26 ] ) is to lead to a change in the inter - collision distance @xmath149 , that is , @xmath149 is @xmath7-dependent in the perturbed model . in order to find this @xmath7-dependence
we first write down the equation for the location of the @xmath35th collision of the soliton from the zero time slot in the reference channel .
since this collision is with the soliton from the @xmath150 time slot in channel 1 , the collision distance @xmath34 is determined by @xmath151 where the first subscript in @xmath152 stands for the channel and the second subscript represents the time slot . taking into account the difference in group velocities , @xmath153 , and the position shifts experienced by the solitons in the collision we find @xmath154 solving eq .
( [ crosstalk28 ] ) for @xmath149 we obtain @xmath155 + \frac{2}{\beta_{1}^{3 } } \left[\eta_{0}(z_{l-1})+\eta_{1}(z_{l-1})\right ] .
\label{crosstalk29}\end{aligned}\ ] ] from the definition of @xmath156 and the periodicity of the pulse sequences it follows that @xmath157 .
therefore , eq . ( [ crosstalk29 ] ) can be rewritten as @xmath158\right\ } , \label{crosstalk30}\end{aligned}\ ] ] where @xmath159 .
notice that @xmath160 due to the fact that the xpm - induced position shifts are positive for the reference channel solitons and negative for the solitons in channel 1 .
the equations for the amplitudes @xmath102 and @xmath148 at @xmath161 are similar in form to eq .
( [ crosstalk4 ] ) . substituting relation ( [ crosstalk30 ] ) into the equations for @xmath162 and @xmath163 and going to the continuum limit we obtain @xmath164 , \nonumber \\ & & \frac{d \eta_{0}}{dz}= \eta_{0}\left[g_{0}+\frac{c\eta_{1 } } { 1+b\left(\eta_{0}+\eta_{1}\right)}\right ] .
\label{crosstalk31}\end{aligned}\ ] ] equation ( [ crosstalk31 ] ) represents the perturbed model that takes into account xpm effects .
notice that for the typical values @xmath83 and
@xmath165 , @xmath166 , i.e. , @xmath167 . thus , for such values the xpm perturbation can be considered as a weak perturbation .
we look for equilibrium states of eq .
( [ crosstalk31 ] ) that are of the form @xmath168 .
this requirement yields the expressions @xmath169 for the gain / loss coefficients .
thus , an important consequence of the xpm - induced position shifts is a change in the values of the gain / loss coefficients that are required for maintaining an equilibrium state with equal non - zero amplitudes .
taking into account the modified expressions for @xmath170 and @xmath171 , we can rewrite eq .
( [ crosstalk31 ] ) as @xmath172 , \nonumber \\ & & \frac{d \eta_{0}}{dz}= c\eta_{0}\left[-\frac{\eta}{1 + 2b\eta } + \frac{\eta_{1 } } { 1+b\left(\eta_{0}+\eta_{1}\right)}\right ] .
\label{crosstalk32}\end{aligned}\ ] ] to prove stability of the equilibrium point @xmath173 we look for a lyapunov function of the system ( [ crosstalk32 ] ) in the form @xmath174 .
\label{crosstalk33}\end{aligned}\ ] ] taking the derivative along trajectories of the system ( [ crosstalk32 ] ) we arrive at @xmath175 replacing @xmath176 and @xmath177 with the right hand sides of eq .
( [ crosstalk32 ] ) , we find @xmath178 -c\eta\left[\frac{\eta}{1 + 2b\eta}-\frac{\eta_{0 } } { 1+b\left(\eta_{0}+\eta_{1}\right)}\right ] \nonumber \\ & & -\frac{c\eta(\eta_{0}-\eta_{1})}{1+b\left(\eta_{0}+\eta_{1}\right)}=0 .
\label{crosstalk34a}\end{aligned}\ ] ] hence , @xmath73 is constant along trajectories of eq .
( [ crosstalk32 ] ) .
in addition , it is straightforward to show that @xmath73 attains its minimum at @xmath173 . since @xmath179
, it follows that @xmath180 for any @xmath181 such that @xmath96 and @xmath97 . combining these results
we conclude that @xmath73 is a lyapunov function for eq .
( [ crosstalk32 ] ) and that the equilibrium state @xmath173 is a center @xcite .
consequently , @xmath173 is a stable equilibrium point and deviations of the initial amplitude values from @xmath182 lead to oscillatory dynamics of @xmath183 and @xmath184 .
this means that the xpm perturbation does not change the stability properties of the equilibrium state . to illustrate these conclusions we numerically solve eq .
( [ crosstalk32 ] ) by use of a fourth - order runge - kutta scheme . for concreteness
we consider a 2-channel transmission system operating at 160 gbits / s per channel with time slot width @xmath83 . in this system
the pulse width is 1.25 ps and @xmath185 .
notice that wdm transmission at 160 gbits / s per channel has received much attention in recent years both theoretically @xcite and experimentally @xcite .
taking @xmath186 and @xmath88 we obtain @xmath187 mw for the soliton peak power .
the dimensionless final propagation distance is taken as @xmath188 corresponding to propagation over @xmath189 km , but the main dynamical features are observed already at much shorter distances . without loss of generality we choose @xmath92 , so that the equilibrium state is @xmath190 .
figure [ fig7 ] shows the @xmath7-dependence of @xmath148 and @xmath102 with the initial condition @xmath191 and @xmath105 for @xmath192 ( a ) and @xmath193 ( b ) .
the latter values of @xmath29 correspond to interchannel frequency spacing of @xmath194 ghz and @xmath195 ghz , respectively . in both cases the amplitudes exhibit oscillations about the equilibrium value in accordance with the prediction of the analytic calculations .
similar oscillatory dynamics is observed for other initial conditions and other values of @xmath29 .
furthermore , as can be seen from fig .
[ fig8 ] , all solutions curves in the phase plane are closed .
thus , our numerical simulations validate the predictions about stability of the equilibrium state in the presence of xpm .
the raman - induced sfs experienced by the @xmath18th - channel solitons is given by @xcite @xmath196 where we explicitly take into account the coupling of the frequency shift to the amplitude .
consider the effects of the raman sfs on amplitude dynamics in a two - channel system consisting of channels 0 and 1 .
since in general @xmath197 , the raman - induced sfs can lead to @xmath7-dependence of the frequency difference @xmath198 , which in turn would lead to @xmath7-dependence of the inter - collision distance @xmath199 .
therefore , the impact of the raman sfs on amplitude dynamics in a two - channel system can be taken into account by replacing the constant frequency difference @xmath29 by the @xmath7-dependent frequency difference @xmath200 in the equations for @xmath102 and @xmath148 , and by using relation ( [ crosstalk35 ] ) to obtain the dynamic equation for @xmath201 .
this calculation yields the following system : @xmath202 looking for equilibrium states of the system ( [ crosstalk36 ] ) in the form @xmath203 we obtain @xmath204 for the gain / loss coefficients . comparing this result with the result obtained in section [ unperturbed ] for the unperturbed model we see that the raman - induced sfs does not change the values of the gain / loss coefficients that are required for maintaining an equilibrium state with equal amplitudes . using the values @xmath204 we can rewrite the system ( [ crosstalk36 ] ) in a simpler form , @xmath205 we study the stability of the equilibrium state @xmath203 of ( [ crosstalk37 ] ) by linear stability analysis and by numerical simulations .
linear stability analysis predicts a bifurcation at @xmath206 . for @xmath207
the jacobian matrix of ( [ crosstalk37 ] ) has two purely imaginary eigenvalues and one zero eigenvalue , while for @xmath208 all three eigenvalues are real . in the latter case
one eigenvalue is positive , another is negative , and the third one is zero . for @xmath209
all three eigenvalues are equal to zero . based on this analysis one might suspect that the equilibrium state @xmath203 becomes unstable for @xmath208 .
however , since @xmath203 remains a non - hyperbolic equilibrium point , linear stability calculations might fail @xcite and one has to resort to numerical simulations to study stability .
we perform numerical simulations with eq .
( [ crosstalk37 ] ) for the same two - channel system that was considered in subsection [ xpm ] .
choosing @xmath92 , the equilibrium state is @xmath210 and the bifurcation value predicted by linear stability analysis is @xmath211 .
figure [ fig9 ] shows the @xmath7-dependence of the amplitudes and frequency difference for @xmath192 ( a ) and @xmath212 ( b ) .
the initial conditions are @xmath213 , @xmath214 , and @xmath215 in ( a ) , and @xmath191 , @xmath105 , and @xmath216 in ( b ) .
one observes that in ( a ) @xmath184 , @xmath183 , and @xmath200 tend away from their equilibrium values @xmath217 , while in ( b ) @xmath184 , @xmath183 , and @xmath200 oscillate about their equilibrium values @xmath218 .
similar dynamic behavior is obtained for other initial conditions in the neighborhood of @xmath217 or @xmath218 .
based on these observations we conclude that the equilibrium state @xmath217 is unstable , whereas @xmath218 is stable , in agreement with linear stability analysis .
this conclusion is further supported by the corresponding phase portraits that are shown in fig .
[ fig10 ] . for the system with @xmath192
the trajectories starting in the vicinity of @xmath217 tend away from @xmath217 .
in contrast , for the system with @xmath212 trajectories in the vicinity of @xmath218 are closed curves centered about @xmath218 . to further investigate the stability properties of a generic equilibrium point @xmath203 we perform detailed analysis of numerical simulations data for @xmath182 and @xmath29 values close to the bifurcation line predicted by linear stability calculations : @xmath219 .
figure [ fig11 ] is the bifurcation diagram obtained by this analysis .
it is seen that the bifurcation line obtained by numerical solution of eq .
( [ crosstalk37 ] ) closely agrees with the line predicted by linear stability computations despite of the fact that @xmath203 is a non - hyperbolic equilibrium state .
notice that the points in the @xmath220 plane that are above the bifurcation line are stable , while those that are below the bifurcation line are unstable .
thus , transmission in the two - channel system described in subsection [ xpm ] becomes unstable for frequency spacing values smaller than @xmath221 .
the second effect of delayed raman response on a single collision between a @xmath18th - channel soliton and a @xmath19th - channel soliton is an @xmath222 frequency shift @xcite .
this raman - induced xfs is given by @xcite @xmath223 we now consider the impact of this frequency shift on amplitude dynamics in a two - channel system
. since in general the amplitudes @xmath102 and @xmath148 vary with @xmath7 , the frequency shifts might lead to @xmath7-dependence of the frequency difference @xmath201 . in order to construct the perturbed model describing the effects of the raman xfs we first obtain an equation for the dynamics of @xmath201 . using eq .
( [ crosstalk40 ] ) we find @xmath224 and @xmath225 .
combining these relations with the definition of @xmath201 we arrive at @xmath226 in addition , we replace the constant frequency difference @xmath29 by the @xmath7-dependent frequency difference @xmath200 in the equations for @xmath102 and @xmath148 to obtain @xmath227 equations ( [ crosstalk41 ] ) and ( [ crosstalk42 ] ) represent the perturbed model for amplitude dynamics in the presence of the raman xfs .
we look for equilibrium states of the model in the form @xmath203 and find @xmath204 for the gain / loss coefficients .
therefore , the raman xfs does not modify the values of the gain / loss coefficients required for maintaining an equilibrium state with equal amplitudes .
substituting @xmath204 into eq .
( [ crosstalk42 ] ) we obtain the following simpler form of the model : @xmath228 it is possible to show that @xmath229 is a conserved quantity for the system ( [ crosstalk43 ] ) .
however , this does not guarantee the stability of the equilibrium point .
linear stability analysis predicts bifurcation at @xmath230 . for @xmath207
the jacobian matrix of ( [ crosstalk43 ] ) has two purely imaginary eigenvalues and one zero eigenvalue , while for @xmath208 all three eigenvalues are real . in the latter case
one eigenvalue is positive , another is negative , and the third one is zero . for @xmath209
all three eigenvalues are equal to zero . based on this analysis one
suspects that the equilibrium state @xmath203 becomes unstable for @xmath208 . since @xmath203 is a non - hyperbolic equilibrium point linear stability calculations might lead to erroneous conclusions @xcite .
we therefore investigate the stability of @xmath203 by numerical simulations with eq .
( [ crosstalk43 ] ) . for concreteness
we consider the two - channel system described in subsection [ xpm ] .
choosing @xmath92 , the equilibrium state is @xmath210 and the bifurcation value predicted by linear stability analysis is @xmath231 .
figure [ fig12 ] shows the dynamics of the amplitudes and frequency difference for @xmath232 ( a ) and @xmath212 ( b ) .
the initial conditions are @xmath213 , @xmath214 , and @xmath233 in ( a ) , and @xmath234 , @xmath105 , and @xmath235 in ( b ) .
it is observed that in ( a ) @xmath184 , @xmath183 , and @xmath200 tend away from their equilibrium values @xmath236 , while in ( b ) @xmath184 , @xmath183 , and @xmath200 oscillate about the equilibrium values @xmath218 .
additional numerical simulations with other initial conditions in the vicinity of @xmath236 or @xmath218 show the same dynamical behavior .
we therefore conclude that the equilibrium state @xmath236 is unstable , while @xmath218 is stable , in agreement with linear stability analysis .
this conclusion is also supported by the corresponding phase portraits . to check the stability of a generic equilibrium state @xmath203
we carefully analyze results of numerical simulations with eq .
( [ crosstalk43 ] ) for different @xmath182 and @xmath29 values .
we pay special attention to the region in the @xmath182-@xmath29 plane that is in the close neighborhood of the bifurcation line @xmath237 , predicted by linear stability computations .
figure [ fig13 ] shows the bifurcation diagram that is obtained by our analysis .
it is seen that the bifurcation line obtained by numerical simulations is in good agreement with the line predicted by linear stability calculations , despite of the non - hyperbolic character of @xmath203 .
notice that equilibrium points lying below the line @xmath238 are unstable , while those lying above it are stable .
as a practical consequence we note that transmission in the two - channel system described in subsection [ xpm ] is unstable for frequency spacing values smaller than @xmath239 .
we studied the deterministic effects of inter - pulse raman crosstalk in amplified wdm optical fiber transmission systems .
we considered conventional optical solitons as an example for the pulses carrying the information and assumed that the pulse sequences in all frequency channels are deterministic and that the sequences are either infinitely long or are subject to periodic temporal boundary conditions .
the first setup approximates return - to - zero ( rz ) differential - phase - shift - keyed ( dpsk ) long - haul transmission , while the second one corresponds to rz dpsk closed fiber loop experiments .
we assumed in addition that the constant gain / loss in each frequency channel is determined by the difference between distributed amplifier gain and fiber loss . under these assumptions we showed that the dynamics of pulse amplitudes in an @xmath0-channel transmission line is described by a system of @xmath0 coupled nonlinear ordinary differential equations ( odes ) , having the form of an @xmath0-dimensional predator - prey model .
we calculated the gain / loss coefficients required for maintaining an equilibrium state with equal non - zero amplitudes in all channels , and showed that high - frequency channels should be overamplified , while low - frequency channels should be underamplified compared with the middle ( reference ) channel .
this means that the net gain / loss profile should _ not _ be flat with respect to the frequency . with these values of the gain / loss coefficients we proved stability of equilibrium states with non - zero amplitudes in all channels by constructing lyapunov functions for the system of odes .
the stability was found to be independent of the exact details of the approximation for the raman gain curve .
furthermore , since the lyapunov functions are conserved quantities for the system , typical dynamics of the amplitudes for initial conditions that are off the equilibrium points is oscillatory . in an actual optical fiber line raman
crosstalk is not the only process impacting pulse dynamics .
it is therefore important to understand the manner in which other physical processes perturb the raman - induced amplitude dynamics described above . in this study we concentrated on the effects of three perturbations due to cross phase modulation ( xpm ) , raman self frequency shift ( sfs ) and raman cross frequency shift ( xfs ) . for each of these physical processes we constructed the corresponding perturbed model for an @xmath0-channel system and studied the dynamics in a two - channel system . for xpm - perturbed two - channel transmission we found that the gain / loss coefficients required for maintaining an equilibrium state with equal non - zero amplitudes are smaller compared with the values in the unperturbed case .
this is explained by noting that the xpm - induced position shifts tend to increase the inter - collision distance and thus to decrease the rate of collisions .
in contrast , the stability of the equilibrium state with equal non - zero amplitudes is not changed by xpm , i.e. , the equilibrium state remains a center . for two - channel systems perturbed by raman sfs or raman xfs we found that the values of the gain / loss coefficients required for maintaining the equilibrium state are the same as in the unperturbed case
however , the stability properties of the equilibrium state change as the frequency difference between the channels is decreased or increased , i.e. , the system undergoes a bifurcation .
the bifurcation curves are given by @xmath206 for the perturbed model with raman sfs and @xmath230 for the perturbed model with raman xfs .
in both models , for a fixed value of @xmath182 , two - channel transmission with @xmath207 is stable , while two - channel transmission with @xmath208 is unstable .
we therefore conclude that the raman - induced interplay between amplitude dynamics and frequency dynamics sets a bound on the smallest frequency spacing for stable transmission . in summary
, our study provides a quantitative explanation to the stability of wdm dpsk transmission against raman crosstalk effects .
this stability was demonstrated in experiments in a closed fiber loop @xcite .
the stable behavior of raman - induced amplitude dynamics in dpsk transmission is very different from the intermittent dynamic behavior exhibited by pulse parameters in on - off - keyed ( ook ) transmission due to the interplay between raman crosstalk and bit - pattern randomness @xcite .
this different dynamic behavior is an important advantage of dpsk transmission over ook transmission .
in this appendix we derive the perturbed models for raman - induced amplitude dynamics in the presence of xpm and raman sfs and xfs for wdm transmission lines with @xmath4 channels .
the rate of collisions of a soliton from the @xmath18th channel with solitons from the the @xmath19th channel in the _ unperturbed _ model is @xmath240 where @xmath33 is constant .
thus , eq . ( [ crosstalk4 ] ) for the change of the soliton amplitude in the interval @xmath241 $ ] in the _ unperturbed _ model can be written as @xmath242 in the _ perturbed _ transmission system , eq .
( [ appendb1 ] ) is replaced by @xmath243 where the @xmath7-dependent inter - collision distance @xmath244 is affected by the xpm - induced position shifts .
replacing @xmath245 by @xmath246 in eq .
( [ appendb2 ] ) while employing relation ( [ appendb3 ] ) we arrive at @xmath247 in order to find an expression for @xmath248 we write down an equation for the location of the collision of the soliton from the zeroth time slot in the @xmath18th channel with the soliton from the @xmath249 time slot in the @xmath19th channel : @xmath250 .
taking into account the different group velocities and summing over all xpm - induced position shifts during the collisions we arrive at the following generalization of eq .
( [ crosstalk28 ] ) : @xmath251 where @xmath252 is the kronecker delta function .
solution of eq .
( [ appendb5 ] ) for @xmath253 yields @xmath254\right\}. \label{appendb6}\end{aligned}\ ] ] taking the continuum limit in eqs .
( [ appendb4 ] ) and ( [ appendb6 ] ) we obtain @xmath255 . \label{appendb7}\end{aligned}\ ] ] equation ( [ appendb7 ] ) together with eq .
( [ appendb6 ] ) ( with @xmath156 replaced by @xmath7 ) represent the perturbed model for raman - induced amplitude dynamics in the presence of xpm in wdm transmission systems with @xmath4 channels .
consider the perturbed model with the raman - induced sfs .
the change in the amplitude of a @xmath18th - channel soliton within the interval @xmath241 $ ] is still given by eq .
( [ appendb4 ] ) .
however , now the @xmath7-dependent inter - collision distance @xmath244 is given by @xmath256 where @xmath257 .
the dynamics of @xmath258 is governed by @xmath259 substituting relation ( [ appendb11 ] ) into eq .
( [ appendb4 ] ) and going to the continuum limit we obtain @xmath260 .
\label{appendb13}\end{aligned}\ ] ] equations ( [ appendb12 ] ) and ( [ appendb13 ] ) with @xmath261 describe the raman - induced amplitude dynamics in the presence of raman sfs in transmission systems with @xmath4 channels .
it is straightforward to show that the gain / loss coefficients @xmath39 that are required for maintaining an equilibrium state with equal non - zero amplitudes in all channels are given by eq .
( [ crosstalk6 ] ) , that is , the values of these coefficients are not modified by the raman sfs . turning to the perturbed model with raman xfs
we observe that amplitude dynamics is described by eq .
( [ appendb13 ] ) . to obtain the dynamic equation for the frequency difference @xmath258
we first compute the change in the frequency experienced by a soliton in the @xmath18th channel within the interval @xmath241 $ ] . employing eq .
( [ crosstalk40 ] ) and summing over all collisions we arrive at @xmath262 the continuum limit of eq .
( [ appendb16 ] ) is @xmath263 and therefore the dynamics of @xmath264 is governed by @xmath265 .
\label{appendb18}\end{aligned}\ ] ] thus , the perturbed model for raman - induced amplitude dynamics in the presence of raman xfs is given by eqs .
( [ appendb13 ] ) and ( [ appendb18 ] ) , where @xmath261 .
a simple calculation shows that the values of the gain / loss coefficients required to maintain an equilibrium state with equal non - zero amplitudes in all frequency channels are the same as in the unperturbed model .
g. p. agrawal , _ nonlinear fiber optics _ ( academic , san diego , ca , 2001 ) .
r. h. stolen , proc .
ieee * 68 * , 1232 ( 1980 ) .
a. r. chraplyvy , electron . lett . * 20 * , 58 ( 1984 ) .
f. forghieri , r. w. tkach , and a. r. chraplyvy , ieee photon .
* 7 * , 101 ( 1995 ) .
d. n. christodoulides and r. b. jander , ieee photon .
. lett . * 8 * , 1722 ( 1996 ) .
p . ho , j. lightwave technol .
* 18 * , 915 ( 2000 ) .
s. chi and s. wen , opt
* 14 * , 1216 ( 1989 ) .
b. a. malomed , phys .
a * 44 * , 1412 ( 1991 ) .
s. kumar , opt . lett . * 23 * , 1450 ( 1998 ) .
a. peleg , opt . lett . * 29 * , 1980 ( 2004 ) .
t. i. lakoba and d. j. kaup , opt
* 24 * , 808 ( 1999 ) .
d. cotter and a. m. hill , electron . lett .
* 20 * , 185 ( 1984 ) .
a. r. sarkar , m. n. islam and m.g .
mostafa , opt .
quantum electron .
* 39 * , 659 ( 2007 ) .
m. muktoyuk and s. kumar , ieee photon .
technol . lett .
* 15 * , 1222 ( 2003 ) .
y. chung and a. peleg , nonlinearity * 18 * , 1555 ( 2005 ) . t. yamamoto and s. norimatsu , j. lightwave technol .
* 21 * , 2229 ( 2003 ) .
a. peleg , phys .
a * 360 * , 533 ( 2007 ) . y. chung and a. peleg , phys
. rev . a * 77 * , 063835 ( 2008 ) .
a. peleg , phys .
a * 373 * , 2734 ( 2009 ) .
x. zhou and m. birk , j. lightwave technol .
* 24 * , 1218 ( 2006 ) .
d. mazroa , s. zsigmond , and t. cinkler , photon .
network commun .
* 18 * , 77 ( 2009 ) .
b. colella , f. effenberger , c. shimer , and f. tian , `` raman crosstalk control in passive optical networks '' , in _ proc .
fiber opt .
_ , anaheim , ca , 2006 , paper nwd6 .
h. kim , s. b. jun , and y. c. chung , ieee photon .
* 19 * , 695 ( 2007 ) .
z. a. t. al - qazwini , m. k. abdullah , and m. b. mokhtar , opt
. eng . * 48 * , 015001 ( 2009 ) .
f. vanholsbeeck , s. coen , p. emplit , m. haelterman and t. sylvestre , opt . commun . * 250 * , 191 ( 2005 ) . c. xu , x. liu , and x. wei , ieee j. quantum electron . * 10 * , 281 ( 2004 ) .
a. h. gnauck and p. j. winzer , j. lightwave technol .
* 23 * , 115 ( 2005 ) .
v. j. mazurczyk , g. shaulov , and e. a. golovchenko , ieee photon .
* 12 * , 1573 ( 2000 ) .
h. kim , j. h. lee , and h. ji , opt .
express * 16 * , 20687 ( 2008 ) .
v. e. zakharov and a. b. shabat , sov .
jetp * 34 * , 62 ( 1972 ) .
m. n. islam , ed .
, _ raman amplifiers for telecommunications 1 : physical principles _
( springer , new york , 2004 ) . c. headley and g. p. agrawal , eds . , _ raman amplification in fiber optical communication systems _
( elsevier , san diego , ca , 2005 ) .
l. f. mollenauer , a. grant , x. liu , x. wei , c. xie , and i. kang , opt . lett . * 28 * , 2043 ( 2003 ) .
d. f. grosz , a. agarwal , s. banerjee , d. n. maywar , a. p. kng , j. lightwave technol .
* 22 * , 423 ( 2004 ) . c. rasmussen , t. fjelde , j. bennike , f. liu , s. dey , b. mikkelsen , p. mamyshev , p. serbe , p. van der wagt , y. akasaka , d. harris , d. gapontsev , v. ivshin , and p. reeves - hall , j. lightwave technol . * 22 * , 203 ( 2004 ) .
z. b. xu , k. rottwitt , c. peucheret , and p. jeppesen , ieee photon .
* 16 * , 329 ( 2004 ) .
j. d. ania - castan , t. j. ellingham , r. ibbotson , x. chen , l. zhang , and s. k. turitsyn , phys .
96 * , 023902 ( 2006 ) .
l. f. mollenauer and p. v. mamyshev , ieee j. quantum electron .
* 34 * , 2089 ( 1998 ) .
a. h. gnauck , r. w. tkach , a. r. chraplyvy , and t. li , j. lightwave technol .
* 26 * , 1032 ( 2008 ) .
v. volterra , `` variations and fluctuations of the number of individuals in animal species living together '' , translated in _ animal ecology _ , r. n. chapman , ed .
, ( mcgraw - hill , 1931 , new york ) .
r. h. stolen , j. p. gordon , w. j. tomlinson , and h. a. haus , j. opt . soc .
b * 6 * , 1159 ( 1989 ) .
the dimensionless @xmath7 in eq .
( [ cfs1 ] ) is @xmath266 , where @xmath267 is the actual position , @xmath268 is the soliton width , and @xmath269 is the second order dispersion coefficient .
the dimensionless retarded time is @xmath270 , where @xmath271 is the retarded time .
the spectral width is @xmath272 and the frequency difference is @xmath273 .
@xmath274 , where @xmath275 is proportional to the electric field and @xmath276 is the peak power .
the dimensionless second order dispersion coefficient is @xmath277 , where @xmath278 is the kerr nonlinearity coefficient .
the coefficient @xmath10 is given by @xmath279 , where @xmath268 is in picoseconds .
b. a. malomed , phys .
a * 43 * , 3114 ( 1991 ) .
a. peleg , m. chertkov , and i. gabitov , phys .
e * 68 * , 026605 ( 2003 ) .
a. peleg , m. chertkov , and i. gabitov , j. opt .
b * 21 * , 18 ( 2004 ) .
it is possible to show that the same model describes the drift part of the amplitude dynamics in rz ook transmission systems .
@xcite for a detailed derivation .
note that in this case the coefficient @xmath280 in eq .
( [ crosstalk5 ] ) is given by @xmath281 , where @xmath282 is the average fraction of occupied time slots .
for transmission with an even number of channels eq .
( [ crosstalk7a ] ) does not necessarily possess solutions other than the trivial solution .
however , when the raman gain is described by the triangular approximation eq .
( [ crosstalk7a ] ) has infinitely many solutions even for a @xmath81-channel system .
these solutions are given by expressions similar to the ones in eq .
( [ crosstalk7b ] ) .
the proof of stability of the corresponding equilibrium states is similar to the proof for transmission with an odd number of channels .
m. w. hirsch and s. smale , _ differential equations , dynamical systems , and linear algebra _
( academic press , new york , 1974 ) .
s. wiggins , _ introduction to applied nonlinear dynamical systems and chaos _ ( springer , new york , 2003 ) .
it is possible to show that @xmath283 is inversely proportional to @xmath10 .
b. cuenot , ieee photon .
* 15 * , 864 ( 2003 ) .
a. h. gnauck , g. raybon , p. g. bernasconi , j. leuthold , c. r. doerr , and l. w. stulz , ieee photon .
technol . lett .
* 15 * , 1618 ( 2003 ) .
m. daikoku , t. miyazaki , i. morita , h. tanaka , f. kubota , and m. suzuki , ieee photon .
. lett . * 18 * , 391 ( 2006 ) .
a. f. abas , a. hidayat , d. sandel , b. milivojevic , and r. noe , opt . fiber technol .
* 13 * , 46 ( 2007 ) .
f. m. mitschke and l. f. mollenauer , opt .
* 11 * , 659 ( 1986 ) .
j. p. gordon , opt . lett . * 11 * , 662 ( 1986 ) .
y. kodama and a. hasegawa , ieee j. quantum electron .
* 23 * , 510 ( 1987 ) . c. headley iii and g. p. agrawal ,
b * 13 * , 2170 ( 1996 ) .
1 . amplitude dynamics for the two - channel system described by eq .
( [ simu1 ] ) .
( a ) the @xmath7-dependence of @xmath103 ( solid curve ) and @xmath102 ( dotted curve ) with the initial condition @xmath104 and @xmath105 .
( b ) the corresponding phase portrait .
2 . amplitude dynamics for the three - channel system described by eq .
( [ simu2 ] ) .
( a ) the phase portrait .
( 1,1,1 ) is the second equilibrium point from the right .
( b ) the @xmath7-dependence of @xmath103 ( solid curve ) , @xmath102 ( dotted curve ) , and @xmath284 ( dashed curve ) , with the initial condition @xmath112 , @xmath105 , @xmath113 .
3 . the @xmath7-dependence of pulse amplitudes for the four - channel system described by eq .
( [ simu7 ] ) for @xmath118 ( a ) , and @xmath119 ( b ) .
the initial condition is @xmath123 , @xmath124 , @xmath125 , and @xmath126 .
the solid , dashed , dashed - dotted , and dotted curves represent @xmath285 , @xmath286 , @xmath287 , and @xmath288 , respectively . fig .
4 . the @xmath7-dependence of the amplitudes of pulses in the @xmath289 channel for @xmath118 ( a ) , and @xmath119 ( b ) .
the initial condition is the same as in fig [ fig5 ] and the final propagation distance is @xmath127 .
5 . the @xmath7-dependence of pulse amplitudes for a seven - channel system operating at 10 gbits / s per channel .
the initial amplitudes are @xmath138 , @xmath139 , @xmath140 @xmath141 , @xmath142 , @xmath143 , and @xmath144 .
the solid , dashed , dotted , and dashed - dotted curves in ( a ) represent @xmath290 , @xmath291 , @xmath292 , and @xmath293 , respectively .
the solid , dashed , and dotted curves in ( b ) correspond to @xmath294 , @xmath183 , and @xmath295 .
the @xmath7-dependence of pulse amplitudes for the xpm - perturbed two - channel system described by eq .
( [ crosstalk32 ] ) with @xmath192 ( a ) and @xmath193 ( b ) .
the initial condition is @xmath191 and @xmath105 .
the solid and dotted lines represent @xmath184 , and @xmath183 , respectively .
the @xmath7-dependence of soliton amplitudes and frequency difference for a two - channel system perturbed by the raman sfs [ eq .
( [ crosstalk37 ] ) ] .
( a ) the dynamics with @xmath192 and initial condition @xmath213 , @xmath214 , and @xmath215 .
( b ) the dynamics with @xmath212 and initial condition @xmath191 , @xmath105 , @xmath216 .
the solid , dashed and dotted lines represent @xmath184 , @xmath183 , and @xmath200 , respectively .
the bifurcation diagram for the two - channel system perturbed by the raman sfs .
the squares correspond to the bifurcation values obtained by numerical solution of eq .
( [ crosstalk37 ] ) , while the solid line is a fit of the form @xmath296 for the numerical data .
the dotted line represents the prediction of linear stability analysis : @xmath219 .
the @xmath7-dependence of soliton amplitudes and frequency difference for a two - channel system perturbed by the raman xfs [ eq .
( [ crosstalk43 ] ) ] .
( a ) the dynamics with @xmath232 and initial condition @xmath213 , @xmath214 , and @xmath233 .
( b ) the dynamics with @xmath212 and initial condition @xmath234 , @xmath105 , @xmath235 .
the solid , dashed and dotted lines represent @xmath184 , @xmath183 , and @xmath200 , respectively .
the bifurcation diagram for the two - channel system perturbed by the raman xfs .
the squares correspond to the bifurcation values obtained by numerical solution of eq .
( [ crosstalk43 ] ) , while the solid line is a fit of the form @xmath297 for the numerical data .
the dotted line is the prediction of linear stability analysis : @xmath237 . | we study the deterministic effects of raman - induced crosstalk in amplified wavelength division multiplexing ( wdm ) optical fiber transmission lines .
we show that the dynamics of pulse amplitudes in an @xmath0-channel transmission system is described by an @xmath0-dimensional predator - prey model .
we find the equilibrium states with non - zero amplitudes and prove their stability by obtaining the lyapunov function .
the stability is independent of the exact details of the approximation for the raman gain curve .
furthermore , we investigate the impact of cross phase modulation and raman self and cross frequency shifts on the dynamics and establish the stability of the equilibrium state with respect to these perturbations .
our results provide a quantitative explanation for the robustness of differential - phase - shift - keyed wdm transmission against raman crosstalk effects . |
in this paper we briefly discuss the usage of multigrid ( mg ) iteration schemes to solve the non - lte problem of the 2nd kind as defined by @xcite .
the era of extensive development of mg methods started in 1970 s by the work of @xcite .
several steps in using mg methods applied to radiative transfer were made by @xcite , @xcite , and @xcite .
these authors showed that this technique leads to a great improvement of the convergence rate .
this paper demonstrates how to apply these methods to a more general solution of polarized radiative transfer with realistic multilevel atomic models and complicated structure of overlapping lines .
the effects of depolarizing collisions is taken into account .
for the description of the atomic state , we adopt the density - matrix formalism and the representation in the basis of irreducible tensorial operators ( e.g. , * ? ? ?
* ; * ? ? ?
the elements of atomic density matrix have the usual form @xmath2 , where @xmath3 is the energy level of total angular momentum @xmath4 , and ( @xmath5 ) are the multipolar components of the level ( @xmath6 , @xmath7 ) . in stationary regime ,
the density - matrix elements are solutions of the local statistical equilibrium equations , @xmath8 the structure of the @xmath9-matrix has been extensively discussed by @xcite , @xcite , and @xcite .
we assume that this matrix has the form @xmath10 , where @xmath11 is the matrix of radiative rates , and @xmath12 is the matrix of collisional rates ( impact approximation ) . the radiative transfer equation for the set of four stokes parameters @xmath13 has the usual form @xmath14 where @xmath15 is the emission vector of the local sources , @xmath16 is the absorption matrix , and @xmath17 is the matrix of stimulated emission .
all these quantities are dependent on radiation frequency , @xmath18 , position vector , @xmath19 , and direction of propagation determined by the unit vector @xmath20 .
finally , @xmath21 is the parametrization of the radiation path along the @xmath20 direction .
most of the existing non - lte solvers use the methods based on @xmath22-operator splitting similar to the one of @xcite . depending on organization , these schemes are numerically equivalent to the jacobi or gauss - seidel smoothing procedures ( for details , see * ? ? ? * ; * ? ? ?
these smoothing procedures do reduce high frequencies of the solution fast , but poor convergence is achieved for low frequencies .
( with `` high frequencies '' we mean those which are comparable to the spatial frequency of grid points approximating the continuous scale . ) the principles of mg schemes are based on the idea of using coarse grids to reduce the low frequencies , and fine grids to smooth their high - frequency components .
it can be showed that such a process may lead to the optimal cpu time demands of @xmath1 , @xmath0 being the number of points per decade of optical scale . for comparison ,
the jacobi and gauss - seidel methods scales approximately as @xmath23 .
we have applied the non - linear version of the _ standard multigrid _ scheme based on the coarse grid correction ( cgc ) technique ( for details , see * ? ? ?
* ; * ? ? ?
cgc is the process of correction of the fine - grid approximation of the solution using the solutions on the coarse grids .
schematically , it can be described in the following way : the defect ( or residuum ) of the fine - grid approximation is computed by several calls of the sweeping procedure ( jacobi , etc . ) ; then both defect and the initial guess of the solution are restricted to the coarse grid , and a new solution on the coarse grid is obtained using these data .
this coarse - grid solution is interpolated to the fine grid , and the density - matrix components are corrected .
this process can be repeated recursively for every grid in order of increasing grid steps .
this recursive process leads to the so - called v , w , or more complicated diagrams , depending on the way in which the recursion is implemented @xcite . in our code
we use the parabolic short - characteristics technique of @xcite , and as a smoothing algorithm we use the jacobi iteration , similar to the one of @xcite , with modifications to include the effects of line overlapping .
for demonstrating the possibilities of this technique , we chose the model of a semi - infinite , plane - parallel , atmosphere .
the atmosphere is isothermal ( @xmath24k ) , with constant volume density of neutral hydrogen , @xmath25 , thermal electrons , @xmath26 , and protons , @xmath27 .
no magnetic field is taken into account .
the scattering of radiation is supposed to fulfill the conditions of complete frequency redistribution , and the @xmath28 and @xmath29 stokes parameters vanish due to the symmetries of the problem .
we adopted a hydrogen atomic model with 9 fine - structure energy levels ( @xmath30 to @xmath31 ) and all the non - vanishing multipole components @xmath32 of density matrix .
all the coherence elements ( @xmath33 ) are identically zero due to the model s symmetries . the coherences between different energy levels have been neglected due to the small natural widths of the levels in comparison to their separations , and due to the selection rules for dipole radiative transitions . collisional rates with thermal electrons and protons for fine structure transitions , @xmath34 , were computed in part using the semiclassical theory of @xcite , for transitions within the same shell , and in part using the data of collisional cross - sections from the atomic and molecular data information system ( amdis ; ` http://www-amdis.iaea.org/ ` ) , for transitions between different shells .
the number of logarithmically spaced nodes in the fine grid is @xmath35 and 5 grids were used in total .
the number of nodes in the `` @xmath36 '' grid is equal to @xmath37 .
the initial guess for the atomic density - matrix elements is given by the lte populations determined in the unpolarized case .
of atomic populations , @xmath38 , in the mg case ( solid line ) is compared to the jacobi solution ( dashed line ) .
the maximum relative error @xmath39 ( with respect to the fully converged solution ) for mg case ( dotted line ) and jacobi case ( dot - dashed line ) is showed as well .
the norm used is @xmath40 ( see * ? ? ?
* ; * ? ? ?
the graph shows the effect of 11 v - cycles with 2 pre- and 15 post - smoothing jacobi iterations.,width=457 ]
figure [ fig1 ] shows a comparison between the convergence rates of the jacobi and the mg methods . the effect of coarse - grid solutions is reflected by the dramatic evolution of the maximum relative change of the populations .
the coarser the grid the shorter the evaluation time , and the higher the rate of approaching to the truncation error of the grid . the maximum relative error dominated by the long - period components
is strongly reduced by the recursive cgc during the v - cycles .
the time saving of the mg method in this particular model is about a factor 4 compared to the jacobi method .
it must be noticed that the efficiency of mg increases as fast as the efficiency of the smoothing procedure .
the most important benefit from mg s is the asymptotical @xmath1 behavior , which designates the method for use in solutions of problems that necessitate strong refinements .
moreover the presented 1d geometry is the slowest case as pointed out by @xcite and @xcite .
bommier , v. , & sahal - brchot , s. 1978 , , 69 , 57 bommier , v. 1980 , , 87 , 109 brandt , a. 1977 , math.comp .
, 31 , 333 fabiani bendicho , p. , trujillo bueno , j. , & auer , l. 1997 , , 324 , 161 fano , u. 1957 , rev.mod.phys . , 29 , 74 hackbush , w. 1985 , multi - grid methods and applications ( berlin : springer ) kunasz , p. , & auer , l.h .
1988 , jqsrt , 39 , 67 .
landi deglinnocenti , e. 1987 , in numerical radiative transfer , ed .
w. kalkofen ( cambridge : cambridge univ . ) , 265 manso sainz , r. , & trujillo bueno , j. 2003 , in asp conf .
307 , solar polarization 3 , ed .
j. trujillo bueno & snchez almeida ( san francisco : asp ) , 251 paletou , f. , & lger , l. 2005 ` ( astro - ph/0507021 ) ` rybicki , g. b. , & hummer , d. g. 1991 , , 245 , 171 rybicki , g. b. , & hummer , d. g. 1992 , , 262 , 209 sahal - brchot , s. 1977 , 213 , 887 - 899 sahal - brchot , s. , vogt , e. , thoraval , s. , & diedhiou , i. 1996 , , 309 , 317 steiner , o. 1991 , , 242 , 290 vth , h. m. 1994 , , 284 , 319 | a new iterative method for non - lte multilevel polarized radiative transfer in hydrogen lines is presented .
iterative methods ( such as the jacobi method ) tend to damp out high - frequency components of the error fast , but converges poorly due to slow reduction of low - frequency components .
the idea is to use a set of differently coarsed grids to reduce both the short- and long - period errors .
this leads to the so - called multigrid ( mg ) methods . for the grid of @xmath0 spatial points , the number of iterations required to solve a non - lte transfer problem is of the order of @xmath1
this fact could be of great importance for problems with fine structure and for multi - dimensional models .
the efficiency of the so - called standard mg iteration in comparison to jacobi iteration is shown .
the formalism of density matrix is applied to the demonstrative example of 1d , semi - infinite , non - magnetic , 3-principal level hydrogen atmospheric model .
the effect of depolarizing collisions with thermal electrons is taken into account as well as general treatment of overlapping profiles . |
black hole accretion flow can reach a temperature of about @xmath1k if no cooling processes are included . here
, @xmath2 is the gravitational radius of a black hole , @xmath3 is the boltzmann constant , @xmath4 is the gravitational constant , @xmath5 is the mass of the black hole , @xmath6 is the mass of the proton and @xmath7 is the velocity of light . in a realistic flow ,
the temperature is smaller .
for instance , in a shakura - sunyaev ( 1973 ) keplerian disk , the temperature becomes close to a few times @xmath8k . in a two - temperature transonic flow
, the ions may remain hot ( @xmath9k ) while the electrons may be cooler ( ( @xmath10k ) depending on accretion rate of the keplerian and sub - keplerian components ( colpi , maraschi & treves , 1984 ; chakrabarti & titarchuk , 1995 ) .
since temperature inside a star is much less compared to the ion temperature mentioned above , considerable nucleosynthesis is expected inside an accretion disk .
it is true that the free - fall timescale @xmath11 s is very small compared to the stellar age and density of gas inside an accretion disk , @xmath12gm @xmath13 is also very small compared to the stellar density but the reaction cross - sections are often very temperature sensitive ( @xmath14 ) and the higher temperature could compensate for the lower density . considerable work has been done in the past on nucleosynthesis inside a disk taken into consideration .
paczyski and jaroszyski ( 1978 ) have studied the importance of nucleosynthesis at high temperature thin disks made up of helium and carbon . in these disks
, electron degeneracy pressure supports gravity .
taam & fryxell ( 1985 ) studied thin accretion disks of hydrogen - rich matter and showed that under certain condition , thermal instability of a keplerian disk may totally disappear .
they consider very high density disks where degenerate pressure in the mid - plane supports the disk structure .
chakrabarti , jin & arnett ( 1987 ) and jin , arnett & chakrabarti ( 1989 ) studied nucleosynthesis in thick accretion disks of realistic density and temperature .
these disks have very little radial velocity and a significant amount of nucleosynthesis was possible only if viscosity parameter of shakura - sunyaev ( 1973 ) is very low @xmath15 .
they showed that in order that _ thin _ disks remain stable against convection , only @xmath0 reactions are allowed in these disks in the radiation pressure supported regions .
however , for hot thick disks , even cno cycle would be allowed without destabilizing the disk .
these results were later repeated by independent groups ( arai & hashimoto , 1992 ; hashimoto et al .
more recently , chakrabarti & mukhopadhyay ( 1999 ) and mukhopadhyay & chakrabarti ( 2000 ) revisited the problem with a self - consistent accretion disk solution which includes heating and cooling .
they find that nucleosynthesis in some range of the parameter space could be high enough to have an energy release or absorption by nuclear reaction comparable to the viscous heating .
these studies are finding attention in recent years since one could detect lines from various isotopes from jets and companions and can , in principle , have some estimate of the conditions around a black hole . however , work mentioned above was purely numerical in nature and was mainly due to repeated use of two types of analysis : first , the steady state thermodynamic condition of a disk is determined by using fourth order runge - kutta method _ without _ nucleosynthesis taken into account . using temperature and
density distributions thus obtained at each point , nucleosynthesis code is run which mainly involves solving 255 linear equations simultaneously , one equation for each isotope . for each infinitesimal advancement of matter toward the black hole ,
nuclear composition was updated and energy release or absorption was computed locally .
after one complete run , the energy release is incorporated into the hydrodynamic equations and the integration scheme using the runge - kutta method is carried out to have updated thermodynamic conditions .
this process is continued till convergence . in the present paper ,
our approach is completely different .
we attack the problem analytically and estimate effects of nucleosynthesis by using sonic point analysis . this way the problem becomes fully self - consistent to begin with and no iterative step is needed .
presence of a saddle type sonic point is a necessary condition for a steady solution , but it is not sufficient . the flow must connect a black hole horizon with infinity .
of course , in order to have the problem under control , we chose only a few reactions which are likely to be very important , namely , the proton - proton and photo - dissociation of deuterium and helium .
since we are interested in stability , we considered only those reactions ( both exothermic and endothermic ) which change specific energy ( either way ) significantly .
we , however , neglect viscous heating or possible radiative transfer effects . in any successful model of the black hole accretion , a source of the soft photon and a hot , comptonising cloud must be present . in our model
, this comptonising cloud is the inner part of the disk itself . for concreteness , we concentrate on this model of black hole accretion , namely , one in which the flow close to a black hole is transonic ( this is a necessary condition for any model ) , but far away , it is sub - divided into keplerian and sub - keplerian components . according to the properties of the transonic disks ( see , chakrabarti & titarchuk , 1995 ; and references therein ) higher viscosity flow settles into the equatorial plane , while quasi - spherical flows with lower viscosity surround the keplerian disk .
soft photons emitted from the keplerian component are reprocessed by the electrons at the centrifugal pressure supported boundary layer and hard x - rays are emitted .
recent observational results strongly support the presence of comptonising clouds close to a black hole ( mirabel & rodriguez , 1999 ; homan et al . , 2000 ) which have properties very similar to our centrifugal pressure supported regions in the inner accretion disks .
we had to choose a particular model in order to compute the optical depth of matter through which soft photons are comptonized . since at high temperature ,
photo - dissociation is important one has to compute number of hard photons present in this inner disk .
ordinarily , photo - dissociation at a high temperature is taken for granted , but in the present situation , there are paucity of very hot photons since only few of them are emitted from the keplerian component and even fewer are intercepted by the sub - keplerian flow located ahead of the keplerian disk .
what is more , optical depth may be so low that most of the time these photons do not interact adequately with deuterium and helium . because of this
, we compute photon number density self - consistently from the emitted spectrum of a sub - keplerian flow . in the next section , we present the basic hydrodynamic equations which govern the steady state flow and perform the sonic point analysis .
we show that effect of the nuclear term is significant in changing topological properties of the flow . in 3 ,
we show that number of sonic points is larger in presence of nucleosynthesis .
we present a complete global solution and show how it is modified in presence of nucleosynthesis .
finally , in 4 , we draw our conclusions .
in what follows , we use the unit of distance ( x ) to be @xmath2 , unit of velocity to be @xmath7 and the unit of mass to be @xmath5 , where @xmath4 , @xmath5 and @xmath7 are the gravitational constant , mass of the black hole and the velocity of light respectively .
we use paczyski & wiita ( 1980 ) potential @xmath16 to describe the flow around a schwarzschild black hole .
the basic dimensionless hydrodynamic equations which govern the infalling matter in the steady state are given by : \(a ) the radial momentum equation : @xmath17 \(b ) the continuity equation : @xmath18 \(c ) the entropy equation : @xmath19 we do not have an azimuthal momentum equation , since we assumed an inviscid flow where angular momentum is conserved .
the sound speed is defined by , @xmath20 , where , @xmath21 and @xmath22 are the pressure and the density respectively and @xmath23 is the adiabatic index . also , @xmath24 is the conserved specific angular momentum ( in units of @xmath25 ) of the infalling matter , @xmath26 is the specific angular momentum in a keplerian disk , @xmath27 , @xmath28 is the vertically integrated density , @xmath29 is the half thickness of the disk [ @xmath30 , @xmath31 is the polytropic index , @xmath32 is the height averaged heat generation / absorption due to the nuclear effect of the disk . in order to isolate the effect of nucleosynthesis alone
, we neglect all the other heating / cooling terms
. the factor @xmath32 may be positive or negative depending on whether net nuclear energy is locally exothermic or endothermic . in future
, we plan to include viscous heating and radiative cooling terms .
the energy release rate for @xmath0 reaction is given by , @xmath33 where , @xmath34 is the number density of protons .
the energy absorption rate due to @xmath35 and @xmath36 dissociation is , @xmath37 where , the number density of the @xmath38th species ( proton , deuterium or helium for this case ) is @xmath39 .
@xmath40 is the avogadro number , @xmath41 is the mass abundance and @xmath42 is the mass number of the @xmath38th element and @xmath43 is the q - value of the corresponding reaction .
the reaction rates have been taken from standard literature ( e. g. , clayton , 1983 ) .
for simplicity , and in order that this work may be generalized even when a large number of nuclear species are considered , we can express this as , @xmath44 where , @xmath45 is a seven parameter family of functions ( thielemann , 1980 ) . for different reactions , the constant coefficients @xmath46s will be different . in our case , @xmath32 is the dimensionless height integrated nuclear energy generation rate @xmath47 . using eqs .
( 1b-1c ) , we eliminate all @xmath48 and @xmath49 terms and get , @xmath50- \frac{a^2(5x-3)}{x(x-1)}}{\frac{2a^2}{\vartheta}-(\gamma+1)\vartheta}. \eqno{(5)}\ ] ] at the sonic point , numerator and denominator both simultaneously should be zero , so , @xmath51 and @xmath52- \frac{a^2(5x-3)}{x(x-1)}=f(a,\vartheta)=0 .
\eqno{(7)}\ ] ] from ( 6 ) , we can get the mach number at the sonic point as , @xmath53 from ( 7 ) and ( 8) we can obtain the sound speed at the sonic point .
the energy and entropy at the sonic points are expressed as , @xmath54 @xmath55 unlike the case of a flow with conserved energy ( chakrabarti , 1989 ; hereafter c89 ) one can not compute everything about the flow using only two parameters , i.e. , energy and specific angular momentum . in the present context of nuclear energy generation ,
one requires to supply the accretion rate as well which in turn determines the density of the flow .
this is explicitly needed in eq .
( 7 ) above . during the infall of matter , rate of change of mass fraction for proton
, deuterium and helium can be written as , @xmath56 @xmath57 @xmath58 during the runge - kutta integration , these linear differential equations are also solved simultaneously to obtain their mass fractions at each point .
before we present results of our global analysis , we like to spend some time on how the photon number density is computed inside the transonic flow . usual treatment of photo - dissociation in the literature assumes the presence of a black body radiation in the flow .
however , advective flow close to a black hole is likely to be optically thin , and they would intercept soft photons from keplerian disks located farther out .
these soft photons are comptonized and in reality , the photo - dissociation would be done by these comptonized photons and the synchrotron photons rather than black body photons .
we need to compute the number of hard photons above @xmath59 mev in order to check if this is much larger compared to the deuterium nuclei .
if one considers a pure black body spectrum then the photon number density can be written as , @xmath60 where , @xmath61 .
but the spectrum of photons is not purely blackbody in nature . to calculate the number density of photon in our system we use results of chakrabarti ( 1997 , hereafter c97 ) .
we concentrate on fig . 4 of c97 where variation of photon spectrum other than black body is discussed .
4a of that paper , spectral variation is shown with energy for different values of keplerian accretion rate ( in units of eddington rate ) @xmath62 . for a complete dissociation of deuterium ,
at least @xmath59 mev energy is needed ( although with the energy of slightly less than @xmath59 mev dissociation may start . ) .
we first concentrate on those cases of @xmath62 where curves are extended at least to the region greater than @xmath59 mev .
each of these curves has one peak due to multi - colour black body and the other is a hump due to comptonization .
we concentrate on the hump at higher frequency to compute the number of hard photons . for the dissociation of a helium nuclei one
requires a higher energy ( @xmath63 mev ) . in fig .
4a of c97 , except for @xmath64 where there is a large peak due to the blackbody emission , other cases have significant amount of hard photon .
along with these photons , we include hard photons due to synchrotron radiation as obtained from a standard distribution @xmath65 normalized in a way such that its number is the same as that of the comptonized photons at the ` knee ' where blackbody distribution meets the comptonized power - law component .
this procedure provides a realistic estimate of hard photon . in any case ,
our main aim is to see if the resulting photon numbers are very large compared to the number of deuterium or helium nuclei .
so the exact number is not important . following the treatment of wagoner ( 1969 ) , we completely turn off the dissociation process if the number of hard photons at a given radius is fewer than about a thousand times that of the deuterium .
the same consideration applies for the dissociation of helium nuclei as well .
here we discuss a few results assuming matter falls towards a black hole of mass @xmath66 .
we choose the polytropic index @xmath67 to be @xmath68 .
we assume that the accretion flow consists of both keplerian and sub - keplerian matter ( chakrabarti & titarchuk , 1995 ) .
we assume that these two rates are @xmath62 and @xmath69 respectively and @xmath70 for concreteness .
we first compute the electron and proton temperatures in the flow by using two temperature hydrodynamic equations numerically ( chakrabarti & titarchuk , 1995 ) for different @xmath71 and @xmath72 rates and use the resulting ( cooler ) temperatures to compute nucleosynthesis in our analytical code .
carrying out of nucleosynthesis work in presence of comptonization would be otherwise very much prohibitive .
we verify a posteriori from the solution with nucleosynthesis that such an approximation is valid . in all the three cases listed below we choose @xmath73 and outer edge of the flow at @xmath74 .
a flow without nucleosynthesis has a smooth solution passing through the sonic point at @xmath75 with these parameters provided it is launched at the outer edge at @xmath76 with a radial velocity @xmath77 and sound speed @xmath78 . here
we choose @xmath79 , @xmath80 . in this case , the photon number density is @xmath81/cc .
we first check the global behaviour of the infalling matter , such as the position and nature of the sonic points if nucleosynthesis is turned on .
following c89 , we plot in fig .
1a , energy @xmath82 at the sonic point as a function of the entropy accretion rate @xmath83 at these points for a fixed specific angular momentum ( @xmath73 ) of the sub - keplerian and keplerian mixture .
dashed curve indicates the behaviour when nucleosynthesis is not included .
this behavior is similar to that observed in c89 .
the solid curve indicates the variation of energy with entropy accretion rate at the sonic point when nucleosynthesis in the flow is included .
the dashed curve is shifted vertically by @xmath84 in order to see the effects of nucleosynthesis prominently .
if the ratio of the number density of photon and the nuclear species to be dissociated in the disk is less than @xmath85 the dissociation is turned off because of the deficiency of photons ( wagoner , 1969 ) . for the present set of parameters
, it is seen that till @xmath86 nucleosynthesis does not affect the sonic points . in the range @xmath87 ,
the outer sonic points disappear as the solid curve shows a distinct gap in that range where nucleosynthesis is prominent .
subsequently , a new branch is originated .
as the inner sonic point is not affected , we can conclude that in these parameters the stability of the disk is not affected .
matter can pass through either the inner sonic point ( i ) and or through the outer sonic point ( o ) if the location of the latter point is far enough .
flows with sonic points located at the intermediate distances are unstable . in fig .
1b , we show the variation of the mach number as a function of radial distance .
the inner sonic point is chosen at @xmath88 .
the velocity of the matter at the outer edge of the disk ( assumed at @xmath89 ) is chosen to be @xmath90 as this allows a smooth solution through the inner sonic point when nuclear energy is not included .
the long - dashed curve is the resulting solution .
the sound speed at the launching point is @xmath91 .
all quantities are expressed in geometric units .
when nucleosynthesis is included , in order that the flow passes through the same inner sonic point smoothly , we had to raise the sound speed to be @xmath92 .
the solid curve indicates this solution .
if the flow with nucleosynthesis is given the same @xmath91 as that of the smooth solution without nucleosynthesis ( long dashed curve ) , the resulting solution becomes totally sub - sonic and is unphysical ( dash - dotted curve ) in the context of black hole accretion . for comparison , we present two other solutions ( short dashed and dotted ) which begin with different sound speed , and are also found to deviate from the true transonic solution ( solid curve ) .
2 shows the variation of the energy release due to nucleosynthesis as a function of the logarithmic radial distance of the flow ( solid curve ) .
the energy release is basically due to @xmath0 reaction as deuterium burning took place instantaneously after launching the flow .
note that the release is at the most @xmath93 ergs / gm / sec which the specific energy of the flow with an inner sonic point could be around @xmath94ergs / gm .
thus the nuclear cooling time scale is @xmath95s which is large compared to , infall or dynamical time scales @xmath96s .
thus , energetically nuclear energy release is not significant . here , we choose a comparatively hotter flow .
we use @xmath97 , @xmath98 photon number density is calculated as @xmath99/cc .
the global behaviour and particular solutions are shown in fig .
3(a - b ) .
they have similar properties as those of case i. in fig .
3a , we clearly see the disappearance of the outer sonic points between @xmath100 and @xmath101 .
the deviation from a ` non - nuclear ' solution is higher , and one requires to launch the flow with a higher speed of sound : @xmath102 . in this case
, the gas has to be heated at a much higher temperature in order that it may pass through the same sonic point . in that sense
the flow is prone to become more unstable than that of case i. fig .
2 ( dashed curve ) shows energy release for this case .
a hotter gas releases somewhat higher energy . in this case , we choose a cooler flow so that deuterium is dissociated closer to the black hole .
we use @xmath103 , @xmath104 and specific angular momentum of the mixture of keplerian and sub - keplerian gas is @xmath73 as before .
since accretion rate is large , there are virtually no comptonized hard photons . however , there are synchrotron photons which dissociate deuterium .
these synchrotron photons are computed from a power - law intensity distribution with a slope of @xmath105 joining with blackbody spectrum at roughly the same place where comptonized power - law joins .
2 shows the total energy release ( dotted upper curve ) as well as energy absorption due to deuterium in this case ( dotted , marked by @xmath106 ) .
as matter comes closer to the black hole , some deuterium is produced by the @xmath0 reaction which is also instantaneously destroyed .
this raises energy absorption closer to a black hole .
note also that only absolute value of @xmath106 is drawn here .
4 gives the variation of the abundance of deuterium , proton and neutron with logarithmic radial distance . in this case , deuterium is totally destroyed at around @xmath107 schwarzschild radii causing a rise in the protons and neutrons .
we studied analytically the effects of nucleosynthesis on black hole accretion flows of constant angular momentum .
we took only the basic reactions such as @xmath0 reaction producing deuterium and destruction of deuterium and helium due to photo - dissociation since ( a ) it is easier to handle only a few reactions analytically and ( b ) these reactions primarily control the energetics of the flow .
since we were dealing with an optically thin flow , the radiation could not be treated as a thermal blackbody and the photon number density had to be calculated from the actual thermal comptonized spectrum and from synchrotron emission from coronal electrons .
photons with energy higher than the threshold value were used to destroy deuterium and helium .
though , the energy released or absorbed were found to be not very significant compared to the rest mass of the flow it has a strong influence on the stability properties at the transonic region .
for instance , we find that when nucleosynthesis in included , the sound speed of the injected matter must be higher by at least @xmath108 to @xmath109 percent in order that the flow passes through inner sonic point .
in other words , the temperature of the flow at the injection point must be @xmath110 to @xmath111 percent higher .
if a physical process ( such as magnetic heating ) could not be found to enable this , the flow will not be transonic and would be unstable ( see , figs . 1b and 3b ) .
so far , in the literature , study of accretion flows around black holes had been done using two - step numerical procedure pure hydrodynamics is followed by pure nucleosynthesis and results of nucleosynthesis is fed back into hydrodynamics .
this procedure was repeated for the whole disk till a reasonable convergence occurs . in the present paper , we combined hydrodynamics and nucleosynthesis for the first time and studied analytically how nucleosynthesis influences topology of the flow . in some region of the parameter space , the outer sonic point is found to be absent .
how would a flow behave in this region ?
it is conjectured that matter would try to pass through ` some ' outer sonic point , becomes supersonic and subsequently pass through the inner sonic point .
since the outer sonic point is not present , the searching procedure to the true sonic point should give rise to an unstable or oscillating solution . only a _ time dependent
_ numerical simulation can tell if this conjecture is correct .
this will be attempted in the future .
arai , k. & hashimoto , m. 1992 , astron .
268 , 131 chakrabarti , s.k .
1989 , apj , 347 , 365 chakrabarti , s.k .
1997 , apj , 484 , 313 chakrabarti , s.k . ,
jin , l. & arnett , w.d .
, 1987 , 313 , 674 chakrabarti , s.k .
& mukhopadhyay , b. 1999 , astron .
344 , 105 chakrabarti , s.k .
& titarchuk , l.g .
1995 , apj , 455 , 623 clayton , d. 1983 , principles of stellar evolution and nucleosynthesis ( new york : mcgraw - hill ) colpi , m. , maraschi , l. & treves , a. 1984 , apj , 280 , 319 hashimoto , m. , eriguchi , y. , arai , k. & mller 1993 , astron .
268 , 131 homan et al .
2000 , apj ( in press ) jin , l. , arnett w.d . &
chakrabarti , s.k .
1989 , apj , 336 , 572 mirabel i.f . , rodriguez , l.f . 1999 ,
ara&a 37 , 409 mukhopadhyay , b. & chakrabarti , s.k .
2000 , astron . astrophys . 353 , 1029 paczyski , b. & jaroszyski , 1978 , acta astr . 28 , 111 paczyski , b. & wiita , p.j .
1980 , a & a , 88 , 23 shakura , n.i . and sunyaev , r.a .
, 1973 _ astron .
astrophys . _
* 24 * 337 taam , r. e. & fryxell , b. a. , 1985 , apj , 294 , 303 thielemann , f.k .
1980 , ph.d . thesis ( mpi , garching ) wagoner , r. v. , 1969 , apjs , 18 , 247 : variation of the energy with entropy accretion rate as the sonic points are changed . i and @xmath100 represent the inner and outer saddle type sonic points and the unmarked curves represent the unphysical ` o ' type sonic point .
dashed curve , shifted by @xmath84 along energy axis , shows the sonic point behavior when nucleosynthesis is turned off . :
variation of the mach number of the flow with logarithmic radial distance .
matter injected at @xmath112 with radial velocity @xmath113 moves in with injected sound speed @xmath114 to pass through the inner sonic point when nucleosynthesis is turned off ( long - dashed curve ) . when nucleosynthesis is turned on ,
sound speed must be modified to @xmath115 in order that the flow passes through the inner sonic point ( solid curve ) .
dash - dotted curve is the ( unphysical ) solution produced when nucleosynthesis is included but the flow had the same initial parameters as that of the long - dashed curve .
short - dashed and dotted curves are for some other initial parameters to indicate non - transonicity .
see text for details .
: variation of net energy release @xmath32 ( in ergs / gm / sec ) as matter is accreted on a black hole .
solid and dashed curves are drawn for cases i and ii and the dotted curves are drawn for the case iii discussed in the text .
@xmath106 is the energy absorption when deuterium is dissociated . :
same as in fig .
1b , except for a hotter flow . in this case , when nucleosynthesis is turned on , the sound speed must be modified to @xmath116 in order that the flow passes through the inner sonic point ( solid curve ) .
see text for details . | we study the effect of nuclear reaction on a thin , axisymmetric , differentially rotating , inviscid , steady accretion flow around a black hole from an analytical point of view .
we find that for most of the reasonable disk parameters , when @xmath0-reaction , dissociation of deuterium and helium are taken into account , the _ transonic _ region of the disk continues to have the inner sonic point and if the temperature of the flow at the injection point could be raised ( by say , some heating processes ) the flow would to pass through this inner sonic point .
otherwise , the flow may be unstable .
we use the sonic point analysis to study the solution . in the rest of the disk parameters
the inner sonic point is absent altogether and the flow will definitely be unstable . psfig.tex |
the major expected source of gravitational waves for the advanced ligo @xcite class of gravitational wave detectors are compact binary systems .
the long inspiral signals from such binaries will be detected by matched filtering with theoretical templates , which allows signal - to - noise to be built up over many orbits @xcite .
gravitational waves ( gws ) are also emitted as broad band bursts when compact objects undergo close passages , either during single parabolic or hyperbolic encounters , or during repeated eccentric encounters @xcite .
such events occur rarely , but are more likely in dense stellar environments , such as globular clusters or galactic nuclear clusters .
the brief duration of such bursts do not allow a large integrated buildup of signal to noise , and they may be difficult to detect without some electromagnetic trigger .
recently , @xcite showed that during binary inspiral of neutron stars ( nss ) resonant tidal excitation of the interface mode a natural mode of a neutron star peaked at the crust - core boundary could result in an isotropic resonant shattering flare , and that these were consistent with short gamma - ray burst ( sgrb ) precursors observed seconds before some sgrbs @xcite .
coincident timing of such precursor flares and the gw inspiral signal can be used to provide strong constraints on the ns equation of state @xcite . in this paper
we show that resonant shattering flares can also occur during close passages with other compact objects , such as another ns or a black hole ( bh ) , and that such flares could serve as electromagnetic counterparts to gravitational wave bursts , allowing triggered searches for these bursts .
tidal energy transfer during close encounters can be determined in a newtonian approximation through the procedure outlined in @xcite .
while a fully relativistic formulation would be preferable , the newtonian formulation is sufficiently accurate for periapse distance much larger than the neutron star radius , relativistic effects would only slightly modify the frequencies and increase the strength of the interaction .
in general the energy transfer rate to a star is given by = d^3 x u where the fluid velocity @xmath0 is the time derivative of the lagrangian displacement @xmath1 and @xmath2 is the gravitational potential . to examine the response of a neutron star with mass @xmath3 it is convenient to decompose the potential due to a star with mass @xmath4 into spherical harmonics @xmath5 , u(r , , ) = _ l=0^ _ m =- l^l u_lm y_lm^ * ( , ) + u_lm = w_lm e^im(t ) where @xmath6 is the true anomaly of the system , @xmath7 denotes the complex conjugate , @xmath8 is the distance between the stars , and @xmath9 is the co - moving coordinate system centered with @xmath3 .
note that for the purposes of mode excitation we are only concerned with the tidal ( @xmath10 ) component of the potential .
assuming the normalization for the spherical harmonics given in @xcite , the constants @xmath11 are w_lm & = & ( -)^(l+m)/2 ^1/2 + & & ^-1 where @xmath12 is defined to be zero when @xmath13 is a non - integer .
the energy transferred to a particular mode ( assuming no non - linear effects , such as crust fracture ) during a periapse passage for a parabolic or highly eccentric encounter can be estimated by eq ( 40 ) of @xcite e_nlm = 2 ^ 2 ( ) ^2 ( ) ^2l + 2 |q_nl|^2 |k_nlm|^2 where @xmath14 is the overlap integral for the ns displacement eigenmode @xmath15 , with radial mode number @xmath16 , q_nl d^3 x _ nlm^ * [ r^2 y_l , m ( , ) ] , [ eq : q ] and k_nlm = 2 ^ 3/2 i_lm(_nlm ) i_lm(_nlm ) _ 0^(1 + x^2)^-ldx .
here @xmath17^{1/2 }
( r_{\rm min}/r_1)^{3/2}$ ] and @xmath18 are the keplerian frequency at the neutron star surface and the mode frequency respectively , both scaled by the keplerian frequency at periapse
. modes with frequency much higher than periapse keplerian frequency ( @xmath19 ) can not be strongly excited .
in contrast if the the periapse distance is too small , the stars may collide , or a tidal disruption may occur . we can calculate the energy transfer to the interface mode @xmath20 , by utilizing the i - mode frequencies and overlap integrals for various equations of state given by @xcite and summarized in table 1 . if @xmath21 , the mode energy required for the crust to reach the breaking strain , then a shattering flare can occur .
@xmath22 for various equations of state are shown as a function of periapse distance @xmath23 for parabolic encounters between @xmath24 neutron star ( figure 1 ) and between a @xmath25 black hole and a @xmath26 neutron star ( figure 2 ) .
, to the mode energy required to reach the breaking strain , @xmath27 , for a parabolic or highly eccentric @xmath25 bh- @xmath26 ns encounter as a function of periapse distance ( @xmath23 ) , for various equations of state .
below a critical periapse distance the mode energy exceeds the breaking energy ( @xmath28 ) and a resonant shattering flare could occur . the upper axis shows the keplerian orbital frequency at the periapse , @xmath29^{1/2}/2\pi$ ] . below : the sky and observer inclination averaged signal to noise ratio ( snr ) for a single advanced ligo gravitational wave detector for a gravitational wave burst at 50 mpc from a parabolic 10 @xmath30 - 1.4 @xmath30 close encounter with periapse @xmath23 . ] .resonant mode properties for the @xmath31 i - mode for various equations of state from @xcite .
@xmath32 is the radius of a @xmath26 ns , @xmath33 is the i - mode frequency , @xmath34 is the overlap integral for the i - mode and the tidal field , while @xmath27 is the mode energy required to reach the breaking strain in the crust . [
cols="<,^,^,^,>",options="header " , ]
the process that produces a resonant shattering flare is outlined in figure [ fig : flow ] . during a close encounter ( or at orbital resonance )
energy is extracted from the kinetic energy of the orbit , through resonant tidal coupling .
the interface mode is excited strongly , which drives the mode to an amplitude at which the breaking strain of the crust is exceeded . the crust fractures , depositing @xmath35 erg of seismic energy into the crust , where @xmath36 is the breaking strain @xcite , @xmath37 is the shear modulus , and @xmath38 cm is roughly the thickness of the crust .
these broad spectrum seismic waves are peaked at characteristic frequency @xmath39hz , where @xmath40 is the density of the crust .
low frequency seismic waves can not couple efficiently to the magnetic field @xcite , and the energy builds up in the neutron star crust as the interface mode is driven further , and more fractures occur .
this seismic energy builds until the crust reaches the elastic limit @xmath41erg , when it shatters , scattering the mode and seismic energy to high - frequency oscillations which can then couple to the magnetic field .
strong perturbations of the magnetic field result in strong transverse electric fields , which can accelerate particles to high energy , sparking a pair - photon fireball .
the luminosities of resonant shattering flares are expected to be up to @xmath42 erg / s @xcite , if the precursor flare timescales are assumed .
@xcite found precursors occurring in 3 out of 49 sgrs analyzed , implying that not every binary merger should result in a detectable shattering flare .
we note that the extraction of seismic energy from the crust by the magnetic field is limited by the strength of the magnetic field at the surface of the neutron star .
the maximum luminosity that can be extracted from the crust by the magnetic field can be estimated by l_max
~10 ^ 47 erg s^-1 ( v / c ) ( b_surf/10 ^ 13g)^2 ( r/10 ^ 10 cm)^2 , where @xmath43 is the maximum velocity of the perturbation to the field line , @xmath44 is the neutron star radius and @xmath45 is the local surface field strength , which can be significantly higher than the large scale dipole field .
thus only shattering flares from neutron stars with sufficiently strong surface fields can be detected .
, to the mode energy required to reach the breaking strain , @xmath27 , for a parabolic or highly eccentric @xmath26 ns- @xmath26 ns encounter as a function of periapse distance ( @xmath23 ) , for various equations of state .
below a critical periapse distance the mode energy exceeds the breaking energy ( @xmath28 ) and a resonant shattering flare could occur .
the upper axis shows the keplerian orbital frequency at the periapse , @xmath46^{1/2}/2\pi$ ] . below :
the sky and observer inclination averaged signal to noise ratio ( snr ) for a single advanced ligo gravitational wave detector for a gravitational wave burst at 50 mpc from a parabolic 1.4 @xmath30 - 1.4 @xmath30 close encounter with periapse @xmath23 . ]
to calculate the expected gravitational wave signal to noise ratio due to parabolic encounters we follow the procedure outlined in @xcite .
the strain due to a gravitational wave burst due to a parabolic encounter is given by @xcite h(f ) = where @xmath47 is the redshift , @xmath48 is the luminosity distance , and @xmath49 is the total gravitational wave energy emitted by encounter per unit frequency , which is given for a parabolic ( @xmath50 ) encounter in the non - relativistic limit by equation ( 46 ) from @xcite .
the signal to noise ratio ( snr ) for a sky and orientation averaged signal on a single detector is given by @xcite = where @xmath51 is the spectral noise density for a given detector .
in figures 1 and 2 the snr for is shown for the ns - ns and bh - ns encounters assuming a single encounter at 50 mpc ( @xmath52 ) , for advanced ligo , with spectral noise density given by @xcite .
blind detection ( snr @xmath53 coincident at each detector , see e.g. * ? ? ?
* ) of a single gravitational wave burst from a neutron star close encounter would be extremely challenging at reasonable distances , with fairly low snr even for close passages , in particular for ns - ns encounters . using x - ray or gamma - ray detections of resonant shattering flares as electromagnetic counterparts ,
triggered gw searches could be performed , significantly lowering the snr threshold for gw burst detection @xcite .
networks of detectors can also be used to enhance burst detection , through coincident and coherent methods @xcite .
@xcite also show that repeated gw bursts from eccentric captures can be combined with the final chirp to boost the integrated snr by roughly an order of magnitude , and would optimistically allow detection of bursts from bh - ns eccentric captures out to @xmath54 mpc , and ns - ns encounters to @xmath55 mpc .
the pattern of these repeated bursts can be modelled for given orbital parameters .
resonant shattering flares can be seen significantly father than the gw bursts .
if they occur for a given system , they will happen for sufficiently close passages which are also those that contribute the largest component of the gw burst signal .
if repeated flares are seen , these could also be used to characterize the orbit and target a burst search to accumulate snr over multiple passages . however
, significant changes to the current gravitational wave templates may be necessary to detect eccentric captures and mergers @xcite .
close encounters of neutron stars with other compact objects are much more likely to occur in dense stellar environments , such as globular clusters and galactic nuclei . while it is beyond the scope of this paper to perform an extremely detailed evaluation of the event rates for close encounters of compact objects , we will briefly discuss the event rates for such encounters in both these environments , and provide updated estimates for some of the rates in the literature .
@xcite calculated the parabolic encounter rate for compact objects in globular clusters using simplified globular cluster models , predicting a rate of @xmath56 detection per year for advanced ligo in optimistic scenarios .
however their detection rates are dominated by rare distant events involving close encounters of @xmath57 black holes . @xcite
examine various dynamical pathways to sgrbs in globular clusters , including binary interactions and tidal capture .
they calculate a high rate of close encounters for two neutron stars , @xmath58 yr@xmath59 gpc@xmath60 using as a calibration the estimate of @xmath61 nss in the collapsed core of m15 , from the fokker - planck calculations of @xcite .
this would require an extremely high neutron star retention fraction .
subsequent more careful calculations by @xcite have determined the number of neutron stars in the core of m15 to be closer to @xmath62 , consistent with @xmath63 retention fraction estimates from pulsar kick velocities @xcite .
this reduces the estimates of @xcite to @xmath64 yr@xmath59 gpc@xmath60 .
the i - mode grows quickly until the maximum strain at the base of the crust exceeds the breaking strain at mode energy @xmath65 erg .
a fracture occurs , releasing @xmath66 erg of low frequency seismic energy ( c ) per fracture , however the mode continues to be driven by the resonance .
as more fractures occur , more energy is deposited into seismic energy in the crust .
when the total seismic energy in the ns crust exceeds the elastic limit of the curst @xmath67 erg , the crust shatters , scattering the mode energy and elastic energy to high frequency oscillations ( d ) .
high frequency oscillations can couple strongly to the magnetic field @xcite by strongly vibrating their footprints ( e ) .
strong perturbations of the magnetic field at the neutron star surface drive strong electric fields , which can accelerate charged particles , triggering pair production and a relativistic fireball with luminosity @xmath68 erg / s .
[ fig : flow ] ] while there are many globular clusters per galaxy , high kick velocities at ns birth significantly lower their retention fraction .
the deeper potentials of galactic nuclei may provide dense stellar environments where the ns retention fraction is higher and close encounters are more likely to occur .
@xcite and @xcite both provide estimates of @xmath69 black hole close passages per year within a few gpc detectable by advanced ligo , with gw - detectable ns - bh encounters estimated to be @xmath70 of this .
however , as we discuss in the appendix below , they scale by a factor @xmath71 , representing the contribution due to the variance of the nuclear cluster density across galaxies .
they take this factor to be @xmath72 , but we find @xmath71 is more correctly evaluated to be significantly lower , even with the most optimistic assumptions . in appendix
a we have re - evaluated the rates for single - single eccentric captures of compact objects in nuclear star clusters containing massive central black holes , assuming a simplified isothermal density distribution , as in @xcite .
we find that the @xmath25 bh - bh eccentric capture encounter rate is @xmath73 yr@xmath59 gpc@xmath60 , significantly lower than the previously estimated values .
however a more top - heavy imf @xcite along with enhanced segregation and spatial flattening for heavier bhs may help to increase this value . the ns - ns and ns - bh ( 10 @xmath30 ) rates can be estimated in a similar fashion for an isothermal @xmath74 density distribution to be @xmath75 yr@xmath59 gpc@xmath60 , and @xmath76 yr@xmath59 gpc@xmath60 with the range mainly due to uncertainty in the imf and mass loss models for ns progenitors @xcite .
however , there is reason to suspect that the slope of the density distribution is somewhat flattened due to interaction with segregated bhs ( see e.g. * ? ? ? * and references therein ) . taking a neutron star density distribution @xmath77 as a lower bound for our rates we find in this case @xmath78
. note that in evaluating the above rates , we have taken very optimistic assumptions about systematic vs intrinsic variation of the relevant observations , as in @xcite .
possibly more realistic estimates for this intrinsic scatter reduce these rates by a factor of @xmath79 . in the high density , high relative - velocity region near the center of nuclear star clusters
, there may be a significant rate of hyperbolic passages where the periapse is sufficiently close to trigger a shattering flare , but insufficient to result in eccentric capture by gw emission . in appendix b we have calculated the rate of encounters that result in a shattering flare during the first close passage , and find this to be higher than the eccentric capture rate for the fiducial model used .
we find , for the optimistic assumptions about intrinsic variation used above , and assuming our canonical isothermal model @xmath80
yr@xmath59 gpc@xmath60 . for a more flattened density profile @xmath77
we have @xmath81 yr@xmath59 . taking our less generous estimates for the intrinsic scatter across galaxies
significantly reduces these rates by a factor of @xmath82 . in the above discussion
we have primarily considered single - single interactions in determining the event rates in dense clusters .
binary - single or binary - binary interactions have larger cross sections and could increase the rates for such events ( see e.g. * ? ? ?
recently @xcite and @xcite demonstrated that kozai - lidov type interactions can drive the inner binaries of hierarchical triple systems towards extreme eccentricity , with collisions occurring when the periapse distance is driven below the stellar radius .
they claim that the rates for such kozai - oscillation driven collisions between white dwarfs in field triple star systems can be comparable to the sn type ia rate .
similar interactions could potentially drive close encounters of nss or bhs in triple systems , however , the periapse for shattering flares is two orders of magnitude smaller than those considered by @xcite , and nss and bhs are substantially more rare than white dwarfs , particularly outside of dense clusters . within clusters
such triple systems would need to reach these extreme eccentricities quickly , before other encounters ionize away the softer less - bound outer companion .
we have calculated the energy transfer to the interface mode through tidal interaction for neutron stars during close encounters with other compact objects , and shown that resonant shattering flares can occur during parabolic or eccentric encounters if the periapse is sufficiently close and the local surface field of the ns is sufficiently high .
such flares are similar to resonant shattering flares during binary inspirals , and should have luminosity @xmath42 erg s@xmath59 @xcite .
broad band gravitational wave bursts are also generated by such encounters .
while they are rare , there is intense interest in such gravitational wave burst events which are detectable by the next generation of gravitational wave detectors .
gw bursts with high snr are also those for which shattering flares may occur and act as an electromagnetic counterpart to trigger burst searches .
highly eccentric captures of nss in dense stellar environments are expected to result in repeated gw bursts @xcite , but may also lead to repeated shattering flares at each periapse passage .
we have also reviewed and updated the rates for compact object encounters presented in the literature for dense stellar environments , and find that these rates should be revised significantly downwards . while these estimated rates for close encounters involving neutron stars within the horizon of advanced ligo are low , shattering flares during from such encounters can be detected significantly farther , and may occur at a rates of up to @xmath80 yr@xmath59 gpc@xmath60 , subject to large model uncertainties .
more conservative assumptions substantially lower this rate .
this research was supported by generous funding from the lorne trottier chair in astrophysics and cosmology as well as the canadian institute for advanced research , and brought to you by the letter @xmath71 .
i would like to thank bence kocsis , ryan oleary , and cole miller for detailed and enlightening discussions involving the rates , as well as samaya nissanke , tony piro , enrico ramirez - ruiz , kostas gourgouliatos , andrew cumming , sterl phinney , and gil holder for insightful conversation .
here we will carefully estimate the eccentric capture rate of compact objects in nuclear star clusters .
we begin by following the general procedure outlined in the appendix c of @xcite , and calculate the eccentric capture rate for a single fiducial galaxy .
the cross section for eccentric capture is basically the cross section for which the energy emitted by gw ( or lost due to tidal interactions ) during a close encounter exceeds the kinetic energy of the objects at infinity .
the maximum periapse for capture is given by @xcite as r_p , max = ^2/7 = 190 km ( ) ^2/7 ( ) ( ) ^-4/7 where @xmath83 is the symmetric mass ratio , @xmath84 is the total mass of the two objects , and @xmath85 is the relative speed of the two objects at infinity .
we only consider gravitational wave capture here since , for compact objects , the tidal capture cross section is much smaller than the gravitational wave capture cross section .
we can then calculate the cross section , using the standard formula for gravitational focussing _ cs & = & r_p , max^2 + & = & 1.3 10 ^ 23 cm^2 ( /0.25)^2/7 ( m_tot/20 m_)^2 ( v_rel/84 km s^-1)^-18/7 where we ve scaled this to fiducial values that will be used later . the rate of eccentric captures for a single galaxy with central supermassive black hole of mass @xmath86 and velocity dispersion @xmath87 is then _
gal(m_s ) _ r_min^r_i dr 4r^2 n_1(r , m_s ) n_2(r , m_s ) _ cs v_rel where @xmath88 and @xmath89 are the number densities of each type of object as a function of radius , @xmath90 is the radius of influence of the black hole and @xmath91 is approximately the radius inside which there is only a single object .
we take the density of objects to be the same as in an isothermal distribution , and the velocity distribution to be maxwellian at each radius with relative speed @xmath92 , which captures the behavior well near the central black hole , where the contribution to the rate is the largest .
the number densities are then given by n_i(r ) = ( ) ^-2 where @xmath93 defines the dynamical radius inside which twice the mass of the central black hole is contained such that @xmath94 and @xmath95 are the number of objects and total mass fraction of type @xmath96 within @xmath93 respectively .
we take these scalings of @xmath97 and @xmath98 for simplicity to highlight the sources of uncertainty and provide a rate estimate for the simplest case . for
how different and more realistic scalings may alter the basic single galaxy rate , see detailed discussion in @xcite and appendix c of @xcite .
it is also convenient to define the geometric means of the number @xmath99 and mass fraction @xmath100 , such that ^2 n_1 n_2 = , ^2 _ 1 _ 2 , as well as the fiducial scalings @xmath101 km s@xmath59 ) , @xmath102 , and @xmath103 . approximating @xmath104 we find _
gal(m_s ) & & _ r_min^r_i dr 4r^2 ( ) ^-4 _ cs v_rel + & & 1.3 10 ^ 30 cm^3 s^-1 _ 0.25 ^ 2/7 m_20 ^ 2 _ 84 ^ -11/7 ( ) ^-11/14 _
^-1^r_i / r_dyn dx x^-2 + 11/14 where we have averaged over the maxwellian distribution , @xmath105 , and changed integration variables to @xmath106 .
assuming that @xmath107 , we have _
gal(m_s ) = 1.210 ^ -10 yr^-1 ( r_i / r_dyn)^31/14 _ 0.25 ^ -23/28 _ 2.5 ^ 31/14 m_20 ^ -3/14 m_4e6 ^ -11/14 _ 84 ^ 31/7 [ 1galrate ] where @xmath108 is scaled to the milky way , and @xmath109 as in @xcite . applying the @xmath110 relation ,
@xmath111@xcite we finally have _
gal(m_s ) = 1.210 ^ -10 yr^-1 ( r_i / r_dyn)^31/14 _ 0.25 ^ -23/28 _ 2.5 ^ 31/14 m_20 ^ -3/14 m_4e6 ^ 9/28 thus far , this agrees relatively well with @xcite and @xcite . from the scatter in the inferred nuclear star cluster relaxation time @xmath112 for the low @xmath113 galaxies given in figure 1 of @xcite , @xcite and @xcite claim that the variance of the central number density scales the average rate per galaxy by a factor @xmath114 , increasing their total rate substantially . here , in calculating the average rate over many galaxies , we will carefully consider the effect of variation in both the @xmath110 relation and the scaling of the central density implied by @xcite related to the parameter @xmath115 and show that such a substantial increase in the inferred rate is not warranted . * variation in the @xmath110 relation : * we begin by taking the generous assumption that the intrinsic scatter in the @xmath110 relation is @xmath116 dex in @xmath86 @xcite .
we take the distribution of @xmath86 for fixed @xmath113 to be log - normal such that @xmath117 where @xmath118 is a random variable with log - normal probability distribution with geometric mean @xmath119 and scale factor @xmath120 .
our single galaxy rate ( [ 1galrate ] ) has scaling such that for a fixed smbh mass bin @xmath121 . using the properties of log - normal random variables the scaling to the average rate over this distribution
is given as _
m = 2.25 . where @xmath122 denotes averaging of @xmath123 over the distribution .
* variation in @xmath115 : * the ratio of the radius of influence , @xmath124 , to the dynamical radius , @xmath93 , determines the relative density of the nuclear cluster , and varies for different n - body models from @xmath125 .
@xcite . for simplicity
we will take this ratio to be a log - normal distributed random variable independent of @xmath118 .
figure 1 of @xcite showed an estimate of the relaxation time at the dynamical radius as a function of the velocity dispersion @xmath113 for nuclei of early - type galaxies in the acs virgo cluster survey @xcite . while the majority of the scatter in this distribution is for low - luminosity unresolved nuclear star clusters , an indication that much of this scatter may be due observational uncertainty , for the purposes of this discussion we will assume , as in @xcite , that this scatter is intrinsic .
we will again assume , for simplicity , that the relaxation time @xmath112 for fixed @xmath113 is distributed log - normally , with a generous estimate of 1.5 dex standard deviation for the scatter in @xmath126 , such that scale factor @xmath127 . the nuclear relaxation time is given by t_r(r_dyn ) = @xcite , where @xmath128 is the coulomb logarithm . for fixed @xmath113 we can rewrite the @xmath112 in terms of the random variables @xmath118 and @xmath115 .
t_r(r_dyn ) ( ) ^-3 ~_disp^5 c_m^2 ( ) ^-3 . this gives , for independent log - normal random variables @xmath118 and @xmath115 , _
t_r^2 = ( 2_m)^2 + ( 3 _ dyn)^2 where @xmath129 is the standard deviation of @xmath130 .
this then gives the scaling due to variation in @xmath115 to the average rate of _ dyn = = 6.1 .
* final rate : * the central black hole mass function can be estimated by ( m_s ) 0.0077 mpc^-3 ( ) ^+ 1 [ massfn ] @xcite where @xmath131 , @xmath132 and @xmath133 , assuming the local hubble constant @xmath134 km s@xmath59 mpc@xmath59 .
integrating our mass dependence @xmath135 over the mass function , we can obtain the effective density n_gal , eff = _
m_s , min^m_s , max m_4e6 ^ 9/28 ( m_s ) with which to multiply our single galaxy rate evaluated at @xmath136 . the mass function ( [ massfn ] )
is only constructed to be valid between @xmath137 , however we expect significant contribution from smaller galaxies .
the halo mass function increases for lower mass , however for dwarf galaxies the stellar mass ( and therefore smbh mass ) to halo mass ratio drops significantly and the expected number density at that mass should also fall .
if we integrate down to only @xmath138 this gives us an effective density @xmath139 mpc@xmath60 , while extending this mass function down to a cutoff of @xmath140 yields @xmath141 mpc@xmath60 . with this in mind
we take the fiducial value of the effective density to be @xmath142 mpc@xmath60 . scaling to the milky way where @xcite assume a fiducial value of @xmath143 , we have our final rate of eccentric captures in galactic nuclei , ^(gn , ec)_tot & = & d^3 n_gal , eff ( m_4e6 = 1 ) + & & 0.6 yr^-1 ( ) ( ) ( ) ^31/14 ( ) ^-2 _ 0.25 ^ -23/28 _ 2.5 ^ 31/14 m_20 ^ -3/14 n_gal , 5 d_2gpc^3 , [ finalrate ] within @xmath144 gpc , where we take 2 gpc for the fiducial value as it is roughly the advanced ligo horizon distance for @xmath25 bh - bh eccentric captures . here
we ve also included an additional factor @xmath145 which may increase the density for nuclear clusters where significant flattening has occurred ( b. kocsis , private communication ) . assuming the generous fiducial values for @xmath146 and @xmath147 , and no significant flattening , the rate for eccentric capture of @xmath25 bh - bh encounters is _ bh - bh^(gn , ec ) 0.02 yr^-1 gpc^-3 .
less generous estimates for the intrinsic scatter , @xmath148 @xcite , @xmath149 give @xmath150 and @xmath151 , which reduces the above rate by a factor of @xmath79 .
we are now ( finally ) ready to estimate the rates for ns eccentric captures in galactic nuclei .
there is large uncertainty in the ns production rate in nuclear star clusters , mostly due to two factors .
first , the imf is unknown and could range from the standard salpeter imf , to an extremely top - heavy imf ( e.g. * ? ? ?
second , there is great uncertainty in the effects of mass - loss for deteriming the fraction of stars with @xmath152 that will become neutron stars , and the fraction that will collapse to form black holes ( see e.g. * ? ? ? * for discussion ) . with these considerations
, we will assume that between @xmath63 of the stars in a nuclear star cluster will become neutron stars .
the low end of this range is for a standard imf , with low mass loss and metallicity , while the high end roughly corresponds to a top - heavy imf with more mass loss in the ns progenitors . assuming an average stellar mass in nuclear star clusters of @xmath153 , this gives @xmath154 .
neutron stars have only had time to segregate in galaxies for which @xmath155 km s@xmath59 @xcite .
unfortunately these low mass galaxies also have escape velocities @xmath156 , and thus ns kick velocities of @xmath157 km s@xmath59 will significantly reduce the retention fraction after ns formation , much like in globular clusters .
so we will assume the scaling above with no significant enhancements due to segregation or flattening .
substituting @xmath158 , @xmath159 and @xmath160 into ( [ finalrate ] ) and continuing to use the very generous assumptions above for @xmath161 and @xmath162 we get a rate for eccentric capture of ^(gn , ec)_ns - ns 0.04 - 6 yr^-1 gpc^-3 . for @xmath25
bh - @xmath26 ns encounters we take @xmath163 , @xmath164 , @xmath165 , @xmath166 , and get a rate ^(gn , ec)_bh - ns 0.05 - 0.6 yr^-1 gpc^-3
. however , in systems where the bh s dominate the core , a cusp of more massive objects tends to flatten out the distribution of lighter objects , compared to an isothermal density profile .
@xcite find that the distribution can approach a power law index of 1.5 for the lighter objects in their fokker - planck calculation .
repeating the above calculations for @xmath167 as a lower bound , we find ^(gn , ec)_ns - ns 0.003 - 0.3 yr^-1 gpc^-3 , noting that the rate in a single galaxy is no longer dominated by the contribution due to the innermost objects .
eccentric captures , described above , provide multiple close passages , however , the cross section for eccentric capture only exceeds that for shattering flares from the first passage below @xmath168 km s@xmath59 . near the center of nuclear star clusters where the density and relative velocity dispersion are the highest the rate for hyperbolic passages that result in shattering flares , but are not bound through gw emission can be significant . in this appendix
we calculate this rate for shattering flares during first passage .
we begin by assuming that that the speed at infinite separation is small compared to the speed at periapse @xmath169 .
the maximum periapse distance for shattering is then only a function of the stellar masses , and the equation of state r_p , s = r_p , s ( , m_tot , eos ) , which can be determined through setting @xmath170 as in @xmath171 of the main text above .
we can then calculate the single galaxy rate as for eccentric captures , giving ^(gn , sf)_gal & & _ r_min^r_i 4 r^2 ( ) ^-4 dr + & & 1.3 10 ^ -10 yr^-1 ( ) ( ) ^5/2 ( ) ^5/2 _ 0.25 ^
-5/4 m_2.8 ^ -3/2 c_m^5/4 m_4e6 ^ 3/4 where we ve used the @xmath110 relation @xmath172 and the fact that @xmath173 over a maxwellian distribution . to calculate the effective density we take n^(sf)_gal , eff = _
m_s , min^m_s , max m_4e6 ^ 3/4 ( m_s ) 0.1 mpc^-3 which gives us our rate for shattering flare encounters of ^(gn , sf)_tot & & 0.5 yr^-1 _ m _ dyn ( ) ( ) ^5/2 ( ) ^5/2 ( ) _ 0.25 ^ -5/4 m_2.8 ^ -3/2d_2gpc^3 .
similar to the case for eccentric captures , less generous estimates for the intrinsic scatter significantly reduce these rates .
taking @xmath148 and @xmath149 , we have @xmath177 and @xmath178 , which reduces the above rates by a factor of @xmath82 . | we demonstrate that resonant shattering flares can occur during close passages of neutron stars in eccentric or hyperbolic encounters .
we provide updated estimates for the rate of close encounters of compact objects in dense stellar environments , which we find are substantially lower than given in previous works .
while such occurrences are rare , we show that shattering flares can provide a strong electromagnetic counterpart to the gravitational wave bursts expected from such encounters , allowing triggered searches for these events to occur . |
despite decades of study our knowledge of the fundamental resonance spectrum of the nucleon is still incomplete . many resonance properties are uncertain and some resonances predicted to exist are yet to be observed .
most predictions of the resonance spectrum are based on phenomenological models , such as the constituent quark model or the di - quark model , but lattice predictions directly from quantum chromodynamics ( qcd ) are fast developing . experimental differentiation between these theoretical approaches based on the presently established resonance spectrum
is , however , inconclusive due to the many `` missing '' and poorly established resonances . insufficient accuracy and quantity of observables
obtained results in too many ambiguities and the famous case of the missing resonances in baryon spectroscopy @xcite .
have they not yet been observed or are they simply not there ? a promising tool to learn about the resonance spectrum is the photoproduction of pions from nucleon targets , as many resonances are expected to couple to the pion decay channel and polarised real photons , with a well - understood electromagnetic interaction , provide a powerful probe . in such pseudo - scalar meson production reactions
the measurement of differential cross section , all three single - polarisation ( beam , target , recoil nucleon ) and a total of four well chosen double - polarisation observables has the potential to give the first model independent extraction of the production amplitudes .
this will greatly enhance our capabilities to determine the nucleon excitation spectrum in a nearly model - independent way .
. this effort is being carried out at various electron - beam facilities worldwide .
one of the three single - polarisation observables , and a crucial one to constrain pwa s , is the beam asymmetry , @xmath0 , from linearly polarised photons : @xmath4 where @xmath5 is the linear beam polarisation and @xmath6 is the angle between the reaction plane and the photon polarisation plane .
data from the neutron is crucial in determining the resonance isospin and achieving a reliable extraction of the electromagnetic couplings of the excited states .
the world dataset on the neutron , however , is extremely sparse , consisting of only three fixed - angle experiments limited to the 1.70 - 2.05 gev range in energy and 35@xmath7 - 90@xmath7 in scattering angle . @xcite@xmath3@xcite@xmath3@xcite . a new , extensive photoproduction experiment using a polarised photon beam and
a liquid deuterium target has been carried out at jefferson laboratory , virginia , usa , march - june 2007 .
the experiment ran with both a circularly and a linearly polarised beam .
the linear polarisation plane was flipped frequently between two orthogonal orientations , parallel and perpendicular to the lab floor .
six photon energy settings were employed between 1.1 and 2.3 gev ( produced via coherent bremsstrahlung from an electron beam at 3.3 - 5.2 gev ) .
a single charged particle trigger was used , yielding a total of @xmath8 events recorded .
we present very preliminary measurements of @xmath0 from the analysis of the @xmath9 channel .
jefferson laboratory is home to a 1.4 km race - track electron accelerator operating at energies up to 6 gev ( an upgrade to 12 gev is currently underway ) .
200 @xmath10a of continuous current is split and delivered simultaneously to three experimental halls .
our experiment was conducted in hall b , where the electron beam is passed through a diamond crystal emitting linearly polarised photons via the process of coherent bremsstrahlung .
the scattered electrons are momentum analysed in the hall b photon tagging facility@xcite , which serves to `` tag '' the photons through a coincidence timing measurement between the deflected electrons and the event time in the detector .
the target cell is positioned in the centre of clas .
clas @xcite is a formidable multi - layer onion of detectors ( scintillators , drift chambers and calorimeters ) providing nearly full coverage in the azimuthal angle and from 8@xmath11 to 140@xmath11 in scattering angle ( lab frame ) which , combined with a toroidal magnetic field , offer excellent sensitivity to charged particles .
the reaction of interest can be considered as quasi - free with a spectator proton : @xmath12 the following data cuts are then applied to obtain a first selection of the @xmath9 channel : select two - particle events with particle masses of the @xmath13 and _ p _ in the final state .
cut on `` missing mass '' of the recoiling system to be consistent with the mass of a spectator proton .
select low `` missing momentum '' in the spectator system - below 0.15 gev / c .
cut on _ p _ and @xmath13 being back - to - back in centre - of - mass system ( cms ) .
the exact photon was identified through timing coincidence of the reaction in clas with the tagger .
this is made possible by the bunched nature of the beam , which arrives at 2 ns intervals .
as can be seen from ( [ asym_eq ] ) , the beam asymmetry is extracted from a @xmath14 fit to the @xmath6-distribution .
@xmath6 is , as before , the angle between the reaction plane and the photon polarisation plane . to make the following simplification possible
, we need to keep the plane from which we measure @xmath6 constant , so we choose to always measure from the horizontal polarisation plane , which is also the horizontal plane of clas in the lab frame . in order to reduce systematics ,
the polarisation plane was rotated between two orthogonal directions during the experiment , with @xmath6 being always measured from the horizontal plane of clas , simplifying the expression thus ( fig .
[ f5 ] ) : @xmath15 @xmath16 @xmath17 where @xmath18 is the differential cross - section from ( [ asym_eq ] ) using a parallel beam polarisation and @xmath19 , consequently , a perpendicular polarisation .
we make the assumption that @xmath20 .
calibrations are still underway , but a very preliminary evaluation of @xmath21 0.4@xmath22 of the data indicates a tiny statistical uncertainty ( fig .
it is our goal to reduce systematics , which are currently very large , to @xmath21 5@xmath22 .
4.7 cm 4.7 cm
the determination of the beam asymmetry , @xmath0 , from a linearly polarised photon beam incident on a neutron ( in a liquid deuterium target ) is underway .
early analysis shows that the data quality is good , statistical uncertainty is tiny and a sizeable asymmetry can be seen at backward angles .
a full analysis is to follow soon in scattering angle range 20@xmath11 - 145@xmath11 ( in cms ) and entire photon energy range 1.1 - 2.3 gev .
the measurement promises to greatly expand the sparse world data - set on the neutron and aid in constraining the determination of the reaction amplitudes in pion photoproduction thus providing strong constraints on the nucleon excitation spectrum .
9 v. burkert , t. s. h. lee , _ int .
j. of mod .
* 13 * , 1035 ( 2004 ) .
i. barker , a. donnachie , j. storrow , _ nucl .
b _ * 95 * , 347 ( 1975 ) . w .-
chiang , f. tabakin , _ phys .
c _ * 55 * , 2054 ( 1997 ) .
j. alspector _
_ , _ phys .
_ * 28 * , 1403 ( 1972 )
. l. abrahamian _
. j. nucl .
phys . _ * 32 * , 69 ( 1980 ) .
f. adamyan _
_ , _ j. phys
. g _ * 15 * , 1797 ( 1989 ) . d.
sober _ et al .
_ , _ nucl .
instr . and meth . a _ * 440 * , 263 ( 2000 ) . b. mecking
_ , _ nucl . instr . and meth .
a _ * 503 * , 513 ( 2003 ) . | we present a preliminary analysis of the photon beam asymmetry observable ( @xmath0 ) from the photoproduction reaction channel @xmath1 .
this new data was obtained using the near-@xmath2 cebaf large acceptance spectrometer ( clas ) at jefferson laboratory , usa , employing a linearly polarised photon beam with an energy range 1.1 - 2.3 gev .
the measurement will provide new data to address the poorly established neutron excitation spectrum and will greatly expand the sparse world data - set both in energy and angle . |
many chemical and biological systems exhibit excitability . in small ( zero - dimensional ) geometry
they show threshold behaviour , i.e. small perturbations immediately decay , whereas sufficiently large perturbations decay only after a large excursion .
this behaviour is crucial for the electrical activation of cardiac tissue or the propagation of nerve pulses where activation should only be possible after a sufficiently large stimulus .
moreover , the decay to the rest state allows for the medium to be repeatedly activated - also crucial for the physiological functioning of the heart and the nervous system .
one - dimensional excitable media support travelling pulses , or rather , periodic wave trains ranging in wavelength @xmath0 from the localized limit @xmath1 to a minimal value @xmath2 below which propagation fails .
pulses and wave trains are best - known from nerve propagation along axons . in two dimensions one typically observes spiral waves .
spirals have been observed for example in the auto - catalytic belousov - zhabotinsky reaction @xcite , in the aggregation of the slime mold dictyostelium discoideum @xcite and in cardiac tissue @xcite . for certain system parameters the propagation of isolated pulses and wave trains may fail ( see for example @xcite ) .
the analytical tools employed to describe these phenomena range from kinematic theory @xcite , asymptotic perturbation theory @xcite to dynamical systems approaches @xcite .
numerical observations reveal that independent of the detailed structure of a particular excitable media model , the bifurcation behaviour of excitable media is generic .
for example , a saddle - node bifurcation is generic for single pulses and for wave trains .
however , there exists no general theory which accounts for all bifurcations which may appear . in this paper
we develop a normal form for excitable media which is build around the observation that the propagation failure of a one - dimensional wave train is mediated by the interaction of a pulse with the inhibitor of the preceding pulse .
+ in section [ sec-1d ] we briefly review some basic properties of excitable media and illustrate them with a specific example . in section [ sec - nf ]
we introduce the normal form .
the properties and the bifurcation scenarios of this normal form are investigated in section [ sec - bt ] . in section [ sec - num1 ]
we show how the parameters of the normal form can be determined from numerical simulations of the excitable medium and then compare the predictions of the normal form with actual numerical simulations of a partial differential equation .
the paper concludes with a discussion in section [ sec - disc ] .
most theoretical investigations of excitable media are based on coupled reaction - diffusion models .
we follow this tradition and investigate a two - component excitable medium with an activator @xmath3 and a non - diffusive inhibitor @xmath4 described by @xmath5 this is a reparametrized version of a model introduced by barkley @xcite .
note that the diffusion constant @xmath6 is not a relevant parameter as it can be scaled out by rescaling length .
although the normal form which we will introduce in section [ sec - nf ] is independent of the particular model used , we illustrate some basic properties of excitable media using the particular model ( [ barkley ] ) . later in section [ sec - num1 ]
we show correspondence of the predictions of our normal form with numerical simulations of the model ( [ barkley ] ) .
our choice of model is motivated by the fact that this model incorporates the ingredients of an excitable system in a compact and lucid way .
thus , for @xmath7 the rest state @xmath8 is linearly stable with decay rates @xmath9 along the activator direction and @xmath10 along the inhibitor direction .
perturbing @xmath3 above the threshold @xmath11 ( in 0d ) will lead to growth of @xmath3 . in the absence of @xmath4
the activator would saturate at @xmath12 leading to a bistable system . a positive inhibitor growth factor @xmath13 and @xmath14 forces the activator to decay back to @xmath15 .
finally also the inhibitor with the refractory time constant @xmath16 will decay back to @xmath17 .
for @xmath18 the system is in zero - dimensional systems no longer excitable but instead bistable . in order to study pulse propagation in one - dimensional excitable media
it is useful to first consider the case of constant @xmath4 .
the resulting bistable model is exactly solvable @xcite and the pulse velocity is @xmath19 $ ] .
hence , excitability requires that @xmath11 is below the stall value @xmath20 .
the quantity @xmath21 characterizes the strength of excitability and @xmath22 coincides with the solitary pulse velocity for @xmath23 .
clearly , for @xmath24 and not too large @xmath25 , pulse propagation fails for @xmath13 larger than some @xmath26 .
the critical growth factor @xmath26 marks the onset of a saddle - node bifurcation @xcite .
the saddle - node can be intuitively understood when we consider the activator pulse as a heat source , not unlike a fire - front in a bushfire . due to the inhibitor
the width of the pulse decreases with increasing @xmath13 .
hence , the heat contained within the pulse decreases . at a critical width , or critical @xmath13 ,
the heat contained within the pulse is too small to ignite / excite the medium in front of the pulse .
[ fig - pulses ] ( a ) ( b ) even if a given set of equation parameters allows for propagation of a single isolated pulse , the system may not necessarily support a pulse in a periodic box of finite length or a wave train consisting of several of such pulses : if the distance @xmath0 between two consecutive pulses of the train becomes too small , the pulses run into the refractory tail of the preceding pulse ( see figure [ fig - pulses ] ) , and may consecutively decay .
hence , propagation failure for periodic wave trains is controlled by the decay of the inhibitor , and propagation is only possible when the inter - pulse distance @xmath0 is larger than a critical wavelength @xmath2 .
note that @xmath2 diverges for @xmath27 when the decay rate of the inhibitor @xmath28 vanishes .
the critical wavelength @xmath2 is a lower bound for the wavelength for the existence of periodic wave trains .
one can also think of keeping @xmath0 fixed and , as before , vary @xmath13 .
then the saddle - node @xmath29 is a monotonically increasing function . in the next section we present a normal form which incorporates the saddle - node bifurcation , and
moreover predicts other types of bifurcations .
it is well known that an isolated pulse undergoes a saddle - node bifurcation above a certain threshold value of the refractoriness @xmath26 .
the corresponding generic normal form for such a saddle - node bifurcation is given by @xmath30 where @xmath31 is , for example , the amplitude or velocity of a pulse with the corresponding values at the saddle - node bifurcation subtracted .
the bifurcation parameter @xmath32 measures the distance from the bifurcation point and is proportional to @xmath33
. the normal form ( [ sn0 ] ) accurately describes the behaviour of isolated solitary pulses in one - dimensional excitable media close to criticality .
we use this normal form for a saddle - node bifurcation for an isolated pulse ( [ sn0 ] ) as a seed to construct a normal form for general travelling in excitable media incorporating finite wavelength effects
. in particular we look at a single pulse on a ring with finite length , i.e. in a one - dimensional box with periodic boundary conditions , and at wave trains with a finite wavelength .
both cases are depicted in figure [ fig - pulses ] . in that case
the saddle - node ( [ sn0 ] ) will be disturbed and will depend on the length of the periodic box in the case of a single pulse ( figure [ fig - pulses]a ) or on the wavelength of the wave train ( figure [ fig - pulses]b ) .
the interaction of the pulse ( or a member of a wave train ) with the preceding pulse ( or more accurate with its inhibitor ; see figure [ fig - pulses ] ) modifies as discussed above the bifurcation behaviour .
we may therefore extend the saddle - node normal form to @xmath34 where @xmath35 describes the inhibitor of the preceding pulse who is temporally displaced by @xmath36 where @xmath0 is the wavelength of the pulse train , i.e. the distance between two consecutive pulses , and @xmath37 is its uniform velocity .
we neglect here a possible temporal dependency of @xmath38 .
note that in the case of a single pulse in a ring , @xmath35 describes the inhibitor of the single pulse which had been created by the pulse at the time of the last revolution around the ring and @xmath0 is simply the length of the periodic box .
+ we assume an exponential decay ( in space and time ) of the inhibitor of well separated pulses .
this is the case for the system ( [ barkley ] ) .
we may write @xmath39 where the function @xmath40 and the decay - rate @xmath41 depend on the particular model chosen ; for example for the model ( [ barkley ] ) we have @xmath42 . in the limiting case of isolated pulses
we note that @xmath43 and @xmath44 , and hence we retrieve the unperturbed saddle - node bifurcation ( [ sn0 ] ) .
the ansatz for @xmath35 is a simplification where we ignore the cumulative effect of the inhibitor .
( note that the equation for the inhibitor @xmath4 in ( [ barkley ] ) can be solved directly and involves an integral over @xmath3 , i.e. involves history. ) the unknown function @xmath45 can be taylor - expanded around the saddle - node @xmath46 .
we summarize and arrive at the following normal form @xmath47 where @xmath48 with @xmath42 for the model equation ( [ barkley ] ) .
pulses have a nonzero width @xmath49 which implies that the temporal delay @xmath36 has to be modified to @xmath50 .
this equation already produces qualitatively all the results we will present in the subsequent sections .
however , much better quantitative agreement is achieved by taking into account that a variation in the amplitude implies a change in velocities and henceforth a change of the effective inhibition .
if we allow for a single pulse to have a temporarily varying pulse amplitude or , in the case of a wave train consisting of distinct members , if we allow for different amplitudes of individual members of the wave train , we have to take into account that the propagation behaviour is amplitude dependent : larger pulses have larger velocities .
hence , a pulse @xmath51 which is larger than its predecessor @xmath52 runs further into the inhibitor - populated space created by its predecessor .
if @xmath53 the finite wavelength induced shift of the bifurcation is stronger compared to the case of equal amplitudes .
conversely , if @xmath54 the finite wavelength induced shift of the bifurcation is weaker compared to the case of equal amplitudes .
this effect is stronger the larger the difference of the two amplitudes @xmath55 .
the inclusion of the amplitude differences affects the bifurcation behaviour depending continuously on the difference @xmath55 .
we thus add a term @xmath56 with @xmath57 into the wavelength dependant inhibitor term in ( [ nf ] ) , and arrive at @xmath58 or after relabeling of @xmath59 @xmath60 it is this equation which we propose as a normal form to study bifurcations of one - dimensional wave trains .
before we show how to determine the parameters of the normal form , we will describe its properties with a main emphasis on bifurcations .
besides the well - known saddle - node bifurcation we identify a symmetry preserving hopf bifurcation and a symmetry breaking spatially inhomogeneous pitchfork bifurcation .
numerical integration of partial differential equation models of excitable media such as ( [ barkley ] ) confirm these bifurcation scenarios of the normal form ( [ nf ] ) .
although some of these bifurcations have been previously observed in numerical simulations , up to now there did not exist a unified framework to study these bifurcations .
the normal form is able to identify these bifurcations as being generic for excitable media , rather than as being particular to certain models of excitable media .
this is the main achievement of our present work .
numerical simulations of excitable media show that the bifurcations of a _ single _ propagating pulse in a ring ( as in figure [ fig - pulses]a ) are different from the bifurcations of a wave train consisting of several distinct pulses ( as in figure [ fig - pulses]b ) .
we first look at a single propagating pulse before in section [ sec - pf ] we look at the interaction of different pulses in a wave train . equation ( [ nf ] ) has the following stationary solutions @xmath61 \;.\end{aligned}\ ] ] the upper solution branch is stable whereas the lower one is unstable .
the two solutions coalesce in a saddle - node bifurcation with @xmath62 since @xmath63 is small we have @xmath64 .
this indicates that the saddle - node of a periodic wave train occurs at smaller values of the bifurcation parameter @xmath32 than for the isolated pulse , and the bifurcation is shifted to the left with respect to the isolated pulse ( see figure [ fig - hopfsketch ] ) .
this is a well known fact which we numerically verified .
note that the limiting case of an isolated pulses with @xmath65 implies @xmath66 and hence , @xmath67 .
the saddle - node of the isolated pulse with @xmath68 at @xmath69 described by ( [ sn0 ] ) is recovered .
[ fig - hopfpde ] besides this stationary instability the normal form ( [ nf ] ) also allows for a non - stationary bifurcation which we investigate in the next section .
the stability of the homogeneous solution @xmath70 with respect to small perturbations of the form @xmath71 can be studied by linearizing the normal form around @xmath70 .
we obtain @xmath72 besides the stationary saddle - node bifurcation ( [ sn_wt ] ) at @xmath73 ( cf .
( [ sn_wt ] ) ) , a hopf bifurcation @xmath74 is possible with @xmath75 in anticipation of the study of wave trains consisting of several distinct pulses we call this hopf bifurcation of a single pulse in a ring a _ symmetry preserving hopf bifurcation_. from ( [ homhopfa ] ) we infer that a hopf bifurcation is only possible provided @xmath76 , i.e. if the coupling is strong enough and the pulse feels the presence of the inhibitor of the preceding pulse sufficiently strong . since @xmath77 the symmetry preserving hopf bifurcation sets in before the saddle - node bifurcation , independent of the value of @xmath78 .
moreover , the hopf bifurcation branches off the upper stable branch of the homogeneous stationary solutions ( [ statio ] ) . in figure [ fig - hopfsketch ]
we show a schematic bifurcation diagram with the saddle - node bifurcation and the subcritical hopf bifurcation for a single pulse in a ring .
+ in numerical simulations of the barkley model and also the fitzhugh - nagumo equations @xcite we could verify this scenario for a single pulse in a ring .
a hopf bifurcation had been previously observed numerically @xcite for the barkley model @xcite and in @xcite for the modified barkley model ( [ barkley ] ) .
hopf bifurcations have also been reported to occur in several other models of excitable media . in @xcite
a hopf bifurcation was found in the @xmath79-variable beeler - reuter model @xcite , and in @xcite in the @xmath80-variable noble model @xcite and in the @xmath81-variable karma model @xcite .
we show here that hopf bifurcations are generic for travelling waves in excitable media .
+ in numerical simulations of model ( [ barkley ] ) the hopf bifurcation was found to be subcritical .
typical temporal behaviour of the maximal amplitude of the activator for model ( [ barkley ] ) close to the bifurcation is shown in figure [ fig - hopfpde ] .
the inset shows the maximal amplitude slightly above the bifurcation point for about @xmath82 periods .
we counted more than @xmath83 periods before stability was visibly lost and the maximal amplitude collapses to zero .
we note that this may have easily lead to the wrong conclusion that the bifurcation is in fact supercritical rather than subcritical .
[ fig - hopfsketch ] the saddle - node and the symmetry preserving hopf bifurcation coalesce in a co - dimension-@xmath81 bogdanov - takens bifurcation for @xmath84 . at the bogdanov - takens point
we have @xmath85 .
the hopf bifurcation and the saddle - node bifurcation have been suggested before to be an unfolding of a bogdanov - takens point in @xcite and later in @xcite .
the normal form provides a framework to study this unfolding .
we were able to numerically verify the condition @xmath86 derived from ( [ homhopfa ] ) by simulating the full partial differential equation ( [ barkley ] ) .
the parameters @xmath78 and @xmath38 will be determined further down in sections [ sec - num1 ] .
we have also numerically simulated the fitzhugh - nagumo @xcite equations to check that this bifurcation is not particular to our chosen model ( [ barkley ] ) .
+ the bogdanov - takens point is apparent in our normal form ( [ nf ] ) and can be derived from it .
close to the saddle - node and the hopf bifurcation when @xmath87 , the dynamics exhibits critical slowing down .
we may therefore expand @xmath88 .
the normal form ( [ nf ] ) becomes at the bogdanov - takens point @xmath89 where @xmath90 and @xmath91 satisfies the stationary version of the normal form ( [ nf ] ) and is given by ( [ statio ] ) .
the linear part of ( [ bt ] ) exhibits the correct eigenvalue structure of a bogdanov - takens point .
the bifurcation parameters , @xmath92 and @xmath93 , measure the distance from the hopf bifurcation and the distance from the saddle - node bifurcation , respectively .
numerical simulations of systems such as ( [ barkley ] ) reveal that a group of several pulses in a ring do not undergo a symmetry preserving hopf bifurcation on increasing the refractoriness @xmath13 , but instead develop a symmetry breaking , spatially inhomogeneous instability whereby every second pulse dies . in figure [ fig - pdepitch ]
we show an example of such an inhomogeneous instability .
spatially inhomogeneous bifurcations have been observed before for periodically paced excitable media @xcite .
here we show that this bifurcation is generic for wave trains in excitable media and does not require external pacing .
+ [ fig - pdepitch ] the above mentioned inhomogeneous alternating instability is contained in our normal form . to investigate spatial instabilities we need to distinguish between consecutive pulses .
we may rewrite the normal form as @xmath94 where the subscript @xmath95 numbers the pulses in a wave train which interact with their nearest neighbours .
linearizing around the homogeneous solution @xmath96 according to @xmath97 yields as a condition for stationary instabilities ( ie .
@xmath73 ) @xmath98 hence stationary instabilities are possible for @xmath99 and for @xmath100 . in the homogeneous case @xmath99 ,
the instability is yet again the spatially homogeneous saddle - node ( [ sn_wt ] ) . for @xmath100
this is a new type of instability , and we have @xmath101 at the bifurcation point .
this spatially inhomogeneous bifurcation will be identified further down as a subcritical pitchfork bifurcation . the criterion @xmath100 for the inhomogeneous bifurcation is corroborated by numerical simulations where every _ second _ pulse dies within a wave train ( see figure [ fig - pdepitch ] ) .
note that in traveling wave coordinates of a partial differential equation model for excitable media , this instability would correspond to a subcritical period - doubling bifurcation .
+ in order to study this @xmath100-bifurcation within our normal form we need to consider two populations of pulses ( @xmath102 ) , @xmath31 and @xmath103 , which interact via their inhibitors with each other . we extend our normal form for the case @xmath100 to @xmath104 the system ( [ nf2 ] ) for wave trains supports two types of stationary solutions ; firstly the homogeneous solution ( [ statio ] ) , @xmath105 , which may undergo a saddle - node bifurcation described by ( [ sn_wt ] ) .
there exists another stationary solution , an alternating mode , with @xmath106 associated with this solution is a pitchfork bifurcation at @xmath107 when @xmath108 comparing ( [ sn_wt ] ) with ( [ ai_pfa ] ) shows that the pitchfork bifurcation sets in before the saddle - node bifurcation .
the upper branch of the homogeneous solution @xmath109 given by ( [ statio ] ) at the pitchfork bifurcation point @xmath110 coincides with ( [ ai_pfb ] ) .
hence the pitchfork bifurcation branches off the upper branch of the homogeneous solution .
it is readily seen that the pitchfork bifurcation is subcritical because there are no solutions @xmath111 possible for @xmath112 .
we now look at the stability of the homogeneous solution @xmath113 .
we study perturbations @xmath114 and @xmath115 .
linearization yields as a condition for nontrivial solutions @xmath116 and @xmath117 .
@xmath118 the upper sign refers to an antisymmetric mode @xmath119 whereas the lower sign refers to a symmetric mode @xmath120 .
stationary bifurcations occur at @xmath73 .
the symmetric mode then coincides with the saddle - node bifurcation ( [ sn_wt ] ) whereas the antisymmetric mode terminates at the pitchfork bifurcation ( [ ai_pfb ] ) . + non - stationary hopf bifurcations are possible if @xmath74 .
we then have @xmath121 and @xmath122 one has physical solutions with a single - valued positive @xmath123 only for the symmetric case ( the lower signs ) which reproduces our results ( [ homhopfa ] ) and ( [ homhopfb ] ) for the symmetry preserving hopf bifurcation .
for @xmath124 the hopf bifurcation moves towards the saddle - node ( [ sn_wt ] ) and coalesces with it at @xmath85 in a bogdanov - takens point as described in section [ sec - bt ] .
for @xmath125 the limiting value of @xmath126 is @xmath127 which coincides with the pitchfork bifurcation @xmath128 in another codimension-2 bifurcation . at this bifurcation
the hopf bifurcation has a period @xmath129 which corresponds exactly to the inhomogeneous pitchfork bifurcation with @xmath100 whereby every second pulse dies . for values @xmath130
the hopf bifurcation always comes after the pitchfork bifurcation which has been numerically verified with simulations of the full system ( [ barkley ] ) .
+ [ fig - pitchsketch ] this allows us to sketch the full bifurcation scenario for a wave train in a periodic ring as depicted in figure [ fig - pitchsketch ] .
in this section we determine the parameters of the normal form ( [ nf ] ) from numerical simulations of the full partial differential equations ( [ barkley ] ) .
we determine the free parameters @xmath49 , @xmath32 , @xmath131 , @xmath132 , @xmath133 and @xmath134 .
we are then in the position to test how well the normal form ( [ nf ] ) reproduces the solution behaviour of the full partial differential equation ( [ barkley ] ) .
we use here as equation parameters for ( [ barkley ] ) @xmath135 , @xmath136 and @xmath137 .
+ the parameter @xmath49 which modifies the delay time @xmath38 due to the finite width of a pulse is easily determined as a typical width of a pulse in the parameter region of interest .
we find @xmath138 .
we note that there is some ambiguity in the determination of @xmath49 , and one may as easily justify @xmath139 $ ] .
+ the parameters @xmath32 and @xmath131 can be determined by studying the isolated pulse with @xmath140 .
the normal form reduces to @xmath141 we have @xmath142 with @xmath26 being the critical @xmath13 at the saddle - node .
solutions of ( [ para1 ] ) are obtained by quadrature @xmath143 for small deviations from the saddle - node @xmath46 this solution may be expanded to obtain @xmath144 which obviously corresponds to the solution of equation ( [ para1 ] ) linearized around the saddle - node .
the solution ( [ para1b ] ) has its inflection point at the saddle - node @xmath46 where its slope is @xmath32 .
we can therefore determine @xmath32 by measuring @xmath145 at the inflection point for different values of @xmath146 .
this allows us to determine @xmath147 via @xmath148 . in figure [ fig - dtxeps ]
we show the results of @xmath145 versus @xmath146 .
the numerical results are obtained by letting a stable pulse which was created at some @xmath149 , decay in an environment with @xmath150 .
the relaxation then allows us to determine the slope at the saddle - node . using a least - square fit we obtain @xmath151 .
+ [ fig - dtxeps ] the parameter @xmath131 can now be determined by looking at the stationary problem @xmath152 .
the behaviour of the amplitude of the activator close to the saddle - node versus @xmath13 is depicted in figure [ fig - epsx ] .
it clearly demonstrates quadratic behaviour typical for saddle - nodes .
the normal form for the saddle - node of an isolated pulse ( [ sn0 ] ) yields @xmath153 which we can use upon using the above measured value of @xmath147 to obtain @xmath154 from a least square fit . + [ fig - epsx ] to determine the missing parameters @xmath134 , @xmath132 and @xmath133 we need to study a pulse in a periodic ring of finite length @xmath0 .
the self - interaction of the pulse with its own inhibitor modifies the saddle - node bifurcation as discussed in section [ sec - nf ] .
we will look here at the symmetry preserving hopf bifurcation . to study the hopf bifurcation we study the circulation of a single pulse in a periodic ring instead of a wave train consisting of more than one pulse . in the latter case the subcritical pitchfork bifurcation sets in before the symmetry preserving hopf bifurcation . combining the expressions for the angular frequency @xmath123 and the deviation @xmath155 of the pulse amplitude from the saddle - node at the bifurcation point , ( [ homhopfa ] ) and ( [ homhopfb ] )
, we can eliminate the so far undetermined parameter @xmath78 to determine @xmath133 .
we obtain @xmath156 in figure [ fig - xhhom ] we show a plot of numerically obtained values for the quantity @xmath157 by integrating the partial differential equation ( [ barkley ] ) , and the result of our normal form ( [ xhom ] ) . in the numerical simulations of ( [ barkley ] ) we measured the frequency of the subcritical hopf bifurcation @xmath123 and @xmath155 which is the difference between the amplitude at the symmetry preserving hopf bifurcation and the saddle - node value of the isolated pulse . expression ( [ xhom ] ) matches the numerical simulations well for @xmath158 .
+ [ fig - xhhom ] we can now determine the parameters @xmath134 and @xmath132 by looking at the shift of the bifurcation parameter @xmath32 and the shift of the critical amplitude @xmath31 at the saddle - node for finite wavelength @xmath0 with respect to the values at the saddle - node for an isolated pulse with infinite wavelength .
the saddle - node is shifted with respect to the isolated pulse with @xmath159 .
we express the shifted finite-@xmath0 saddle - node by @xmath160 where @xmath161 expresses the shift in the bifurcation parameter at the saddle - node and @xmath162 expresses the shifted value of the amplitude when compared to the isolated pulse .
the values for @xmath163 and @xmath164 can be measured by numerically solving ( [ barkley ] ) and determing the finite wavelength induced saddle - node .
we did so by expressing ( [ barkley ] ) in travelling wave coordinates and treating the problem as a boundary value problem .
the thereby obtained values for @xmath163 and @xmath164 have to be compared with the corresponding expressions of the normal form . in the normal form ( [ nf ] )
the finite wavelength induced shift is represented by the finite length correction @xmath165 neglecting @xmath166 which is justified for not too large deviations from the isolated pulse , we obtain @xmath167 where @xmath49 and @xmath133 had already been determined . note that the @xmath166-term , @xmath168 appears , of course , correctly in the shift of the bifurcation parameter @xmath169 in ( [ sn_wt ] ) .
although the inclusion of @xmath170 is not necessary for the existence of a hopf bifurcation ( see ( [ homhopfa],[homhopfb ] ) ) , it is significant to obtain good quantitative agreement . whereas for the stationary bifurcations the inclusion of @xmath133 is in effect a redefinition of @xmath78 , it is vital in the case of the hopf bifurcation because it allows for a decoupled dependence of the frequency @xmath123 and the amplitude @xmath171 on @xmath0 . in figure [ fig - shiftpara ]
we show a plot of the bifurcation parameter shift as a function of length @xmath0 .
agreement of numerically obtained values from a simulation of the full partial differential equation ( [ barkley ] ) with the expression derived from our normal form ( [ shifta ] ) , which implies @xmath172 is assured provided @xmath173 .
we recall that @xmath26 is the critical refractoriness at the saddle - node of an isolated pulse . in figure [ fig - shiftx ]
we show a plot of the pulse amplitude shift as a function of length @xmath0 .
agreement of numerically obtained values from a simulation of the full partial differential equation ( [ barkley ] ) with the expression derived from our normal form ( [ shiftb ] ) is given provided @xmath174 . combining these results we can solve for @xmath134 and @xmath132 and obtain @xmath175 and @xmath176 .
this finalizes the determination of the free parameters of the normal form ( [ nf ] ) .
+ [ fig - shiftpara ] [ fig - shiftx ] we are now in the position to check the validity of our normal form by testing whether the normal form ( [ nf ] ) with the above determined parameters is able to reproduce observations of the numerical simulation of the full partial differential equation ( [ barkley ] ) .
we already noted in section [ sec - bt ] qualitative agreement i.e. the correct bifurcation behaviour .
here we show quantitative agreement with the behaviour of ( [ barkley ] ) .
+ in particular we look at the symmetry preserving hopf bifurcation described in section [ sec - bt ] . in figure [ fig - xhh ] and
figure [ fig - om ] we show a comparison of the analytical results for the frequency and the amplitude at the hopf bifurcation , ( [ homhopfa ] ) and ( [ homhopfb ] ) , with results obtained from numerically integrating ( [ barkley ] ) .
the figures show good agreement .
note that the parameters @xmath134 and @xmath49 were not determined by fitting data representing the symmetry preserving hopf bifurcation and henceforth the two figures , figure [ fig - xhh ] and figure [ fig - om ] , are indeed predictions .
+ [ fig - xhh ] [ fig - om ] figure [ fig - xpf ] shows a comparison of numerical simulations of ( [ barkley ] ) and our analytical results ( [ ai_pfb ] ) for the spatially inhomogeneous pitchfork bifurcation . for the determination of the parameters of the normal form we have not used any fitting which involved results from the spatially inhomogeneous pitchfork bifurcation
the agreement in figure [ fig - xpf ] therefore demonstrates that the normal form can indeed be used to obtain quantitative agreement and make quantitative predictions .
[ fig - xpf ]
we have constructed a normal form for travelling waves in one - dimensional excitable media which takes the form of a delay differential equation .
the construction is based on the well - known observation that the interaction of a pulse with the inhibitor of the preceding pulse modifies the generic saddle - node bifurcation of an isolated pulse .
the normal form ( [ nf ] ) exhibits a rich bifurcation behaviour which we could verify by numerically simulating the partial differential equation ( [ barkley ] ) .
besides the well known saddle - node bifurcations for isolated pulses and for periodic wave trains the normal form also exhibits a symmetry preserving hopf bifurcation and a symmetry breaking , spatially inhomogeneous pitchfork bifurcation .
moreover , the normal form shows that the saddle - node and the hopf bifurcation are an unfolding of a bogdanov - takens point as previously suggested in @xcite .
the symmetry preserving hopf bifurcation is found to occur before the saddle - node bifurcation for a single pulse in a ring . for a wave train consisting of several pulses in a ring , the hopf- and the saddle - node bifurcations occur after the symmetry breaking pitchfork bifurcation in which every second pulse dies .
we could verify these scenarios in numerical simulations of the modified barkley - model ( [ barkley ] ) and the fitzhugh - nagumo equations @xcite .
these bifurcations have been observed before but had previously not been described within a unified framework of one normal form .
+ we were able to determine the parameters of the normal form from numerical simulations of the partial differential equation ( [ barkley ] ) .
using these numerically determined parameters we showed excellent agreement between the normal form and the full partial differential equation ( [ barkley ] ) .
we could quantitatively describe the symmetry preserving hopf bifurcation and the inhomogeneous pitchfork bifurcation .
moreover , we were able to quantify the bogdanov - takens bifurcation .
+ the symmetry preserving hopf bifurcation has been studied intensively before .
it was observed numerically for example for the barkley - model @xcite .
interest has risen in the hopf bifurcation in the context of cardiac dynamics because it leads to propagation failure of a single pulse on a ring .
it is believed to be related to a phenomenon in cardiac excitable media which goes under the name of _
alternans describe the scenario whereby action potential durations are alternating periodically between short and long periods .
the interest in alternans has risen as they are believed to trigger spiral wave breakup in cardiac tissue and ventricular fibrillation @xcite . besides numerical investigations of the barkley model @xcite , the modified barkley model @xcite , the beeler - reuter model @xcite , the noble - model @xcite and the karma - model @xcite , where a hopf bifurcation has been reported ,
there have been many theoretical attempts to quantify this bifurcation for a single - pulse on a ring . since
the pioneering work @xcite alternans have been related to a period - doubling bifurcation .
it was proposed that the bifurcation can be described by a one - dimensional return map relating the action potential duration ( @xmath177 ) to the previous recovery time , or diastolic interval ( @xmath178 ) , which is the time between the end of a pulse to the next excitation .
a period - doubling bifurcation was found if the slope of the so called restitution curve which relates the @xmath177 to the @xmath178 , exceeds one . a critical account on the predictive nature of the restitution curve for period - doubling bifurcations
is given in @xcite .
in @xcite the instability was analyzed by reducing the partial differential equation describing the excitable media to a discrete map via a reduction to a free - boundary problem . in @xcite
the hopf bifurcation could be described by means of a reduced set of ordinary - differential equations using a collective coordinate approach . in @xcite
the bifurcation was linked to an instability of a single integro - delay equation .
the condition for instability given by this approach states - as in some previous studies involving one - dimensional return maps - that the slope of the restitution curve needs to be greater than one .
however , as evidenced in experiments @xcite and in theoretical studies @xcite alternans do not necessarily occur when the slope of the restitution curve is greater than one .
in further studies it would be interesting to see how our criterion @xmath179 is related to that condition . note that @xmath49 is related to the @xmath177 and @xmath180 to the recovery time @xmath178 .
+ in the context of alternans the hopf bifurcation has been described as a supercritical bifurcation @xcite ( although their occurrence is related to wave break up @xcite ) .
further study will explore whether the subcritical character of the hopf bifurcation we find is model - dependant or indeed generic .
+ to go beyond the case of a single pulse circulating in a ring , periodically stimulated excitable media have been studied in the context of alternans @xcite . in @xcite one - dimensional maps
were developed to study the hopf bifurcation and the transition to conduction blocks . in @xcite
a nonlocal partial differential equation has been proposed to study spatiotemporal dynamics of alternans .
it would be interesting for further studies to see how the transition to conduction blocks explored in these paced cardiac excitable media @xcite can be described by the spatially inhomogeneous pitchfork bifurcation we found in our normal form .
the pitchfork bifurcation however does not require a fixed pacing site and does not require external pacing but rather is dynamically induced .
this may aid in investigating the formation of conduction blocks purely as a dynamical phenomenon of wave trains .
+ an interesting scenario in our normal is the coalescence of the spatially inhomogeneous pitchfork bifurcation with the hopf bifurcation when @xmath181 .
this condition implies @xmath182 .
then the hopf frequency is in resonance with the spatial instability in which every second pulse dies .
connections to alternans of this scenario are planned for further research .
+ ideally , one would like to deduce the normal form directly from a model for excitable media and determine its parameters without relying on numerical simulations of a particular excitable medium .
one initial path along that avenue could be to use the non - perturbative approach developed in @xcite to determine the parameters .
this method was developed to study critical wave propagation of single pulses and pulse trains in excitable media in one and two dimensions .
it was based on the observation that close to the bifurcation point the pulse shape is approximately a bell - shaped function .
numerical simulations show that this is the case for the barkley model ( [ barkley ] ) close to the saddle - node bifurcation . a test function approximation that optimises the two free parameters of a bell - shaped function , i.e. its amplitude and its width , allows us to find the actual bifurcation point , @xmath26 , and determine the pulse shape for close - to - critical pulses at excitabilities near @xmath26 .
this method has so far also been successfully applied to other non - excitable reaction - diffusion equations @xcite . to apply the method for our purpose
is planned for future work .
+ g.a.g gratefully acknowledges support by the australian research council , dp0452147 .
g.a.g . would like to thank bjrn sandstede for kindly helping with producing some of the numerical saddle - node values using auto .
initial parts of the simulations were done with the xdim interactive simulation package developed by p. coullet and m. monticelli . | we present a normal form for travelling waves in one - dimensional excitable media in form of a differential delay equation . the normal form is built around the well - known saddle - node bifurcation generically present in excitable media .
finite wavelength effects are captured by a delay .
the normal form describes the behaviour of single pulses in a periodic domain and also the richer behaviour of wave trains .
the normal form exhibits a symmetry preserving hopf bifurcation which may coalesce with the saddle - node in a bogdanov - takens point , and a symmetry breaking spatially inhomogeneous pitchfork bifurcation .
we verify the existence of these bifurcations in numerical simulations .
the parameters of the normal form are determined and its predictions are tested against numerical simulations of partial differential equation models of excitable media with good agreement .
* excitable media are often found in biological and chemical systems .
examples of excitable media include electrical waves in cardiac and nerval tissue @xcite , camp waves in slime mold aggregation @xcite and intracellular calcium waves @xcite .
excitable media support localized pulses and periodic wave trains . in 2 dimensions rotating vortices ( or spirals ) and in 3 dimensions scroll waves @xcite
are possible .
the critical behaviour of pulses , wave trains and spirals , i.e. propagation failure , is often associated with clinical situations .
the study of spiral waves is particularly important as they are believed to be responsible for pathological cardiac arrhythmias @xcite .
spiral waves may be created in the heart through inhomogeneities in the cardiac tissue .
some aspects of spiral wave break up can be studied by looking at a one - dimensional slice of a spiral i.e. at a one - dimensional wave train @xcite .
+ we investigate critical behaviour relating to one - dimensional wave propagation .
we develop a normal form which allows us to study the bifurcation behaviour of critical waves . in particular
, the normal form predicts a hopf bifurcation and a symmetry breaking pitchfork bifurcation .
the symmetry breaking pitchfork bifurcation can be numerically observed as an instability where every second pulse of a wave train dies .
this seems to be related to alternans @xcite , which are discussed in the context of cardiac electric pulse propagation . * |
the seyfert 1 galaxy mcg-6 - 30 - 15 is the primary `` disk - line '' target , in which a relativistically broadened iron k - emission line is suggested to originate in the innermost region of the accretion disk around the black hole ( e.g. , tanaka et al .
mcg-6 - 30 - 15 is also known to exhibit characteristic x - ray spectral variations ; inoue & matsumoto ( 2001 , 2003 ) , fabian et al .
( 2002 ) and matsumoto et al .
( 2003 ) reported significantly small variability in the iron line energy band of mcg 6 - 30 - 15 .
it is found that the energy - dependent root mean square ( rms ) variability is suppressed around the iron line energy , particularly in longer timescales .
inoue and matsumoto ( 2001 , 2003 ) proposed that the absorbed spectrum due to photo - ionized warm absorbers mimic shape of the strongly red - shifted iron line , and variation of the warm absorbers may explain the apparently small variability in the iron line energy band . on the other hand , miniutti and fabian ( 2004 ) proposed that the suppressed variability in the iron line band may be explained by the general relativistic light - bending effects , which take place in the very vicinity of the black hole .
miniutti et al .
( 2007 ) claim that the suzaku spectrum of mcg-6 - 30 - 15 is in fact consistent with the light - bending model .
niedwiecki & ycki ( 2008 ) and niedwiecki & miyakawa ( 2010 ) independently re - examined the light - bending model , and concluded that it is possible to explain the suppressed variability in the iron energy band assuming a particular configuration and movement of the illuminating source .
ycki et al .
( 2010 ) , on the other hand , claim that the light - bending model is not able to explain the suzaku spectrum , if a broader energy range is adopted than used by miniutti et al . ( 2007 ) .
the primary question in mcg-6 - 30 - 15 is whether the iron line is truely broad and its little variability is a consequence of the general relativistic light bending effects or not .
the critical point to address the problem is to model the underlying continuum spectrum correctly , since the broad iron line parameters are dependent on the choice of the continuum spectral model .
in fact , on one hand , miller , turner , and reeves ( 2008 ) claim that the `` disk - line '' is not required , when the continuum spectrum is modeled with multiple warm absorbers . on the other hand ,
reynolds et al .
( 2009 ) argue that the absorption - dominated model over - predicts the 6.4 kev fluorescent line emission , and the disk - line model is more physically reasonable .
miller et al . ( 2009 ) counterargue that the calculation in reynolds et al . ( 2009 ) neglects opacity at the 6.4 kev line energy , and conclude that variation in the partial - covering fraction may dominate the observed x - ray spectral variability .
however , the partial covering scenario by miller et al . ( 2009 ) requires a number of free parameters to model the observed spectrum .
therefore , in order to examine the partial covering scenario , reducing the number of parameters , if possible , and constructing a concrete physical picture to explain the spectral variation is certainly intriguing . recently , using the suzaku data taken in january 2006 , miyakawa et al .
( 2009 ) ( hereafter paper i ) found a clear correlation between the intensity in the 6 10 kev band and the spectral ratio of 0.5 3.0 kev / 6 10 kev in a time - scale of @xmath9 sec .
essentially the same spectral correlation was confirmed for longer timescales up to @xmath814 years in the rxte data ( miyakawa 2010 ) .
paper i adopted a spectral model composed of a direct power - law component , its reflection component by a neutral matter ( with the solid - angle @xmath10 ) , two warm absorbers with different ionization states .
the observed spectral variation requires change of the apparent slope of the direct component , whereas the shape and intensity of the reflection component being invariable .
a mildly broad iron emission line was required at @xmath11 kev with an intrinsic width of 0.29 kev ( 1 @xmath12 ) and the equivalent - width 100 ev ; such an extremely broadened `` disk line '' as is claimed by miniutti et al .
( 2007 ) was not required . in this paper ,
following paper i , we attempt to comprehensively understand spectral variability of mcg-6 - 30 - 15 using suzaku and chandra archival data .
our goal is to find a spectral model which can naturally explain the observed spectral variation of mcg-6 - 30 - 15 with a minimum set of free - parameters .
if we are successful to construct such a reasonable spectral model , we will see if the relativistically distorted iron emission line is truely present or not in the energy spectrum .
we believe this is an important step forward to fully understand x - ray spectral properties of mcg-6 - 30 - 15 , seyfert galaxies , and black holes .
we use the same suzaku data used in paper i , taken in 2006 from january 9 to 14th ( 143 ksec exposure ) , from 23 to 26 ( 99 ksec ) and from 27 to 30 ( 97 ksec ) .
we use the xis data in 1 10 kev , and pin data in 10 40 kev . in the present paper , we are interested in spectral variations above 1 kev , so we did not use the data below 1 kev .
all of the spectral fits were made with xspec v11 ( arnaud 1996 ) . in the following ,
the xspec model names used in the spectral analysis are explicitly given . a constant factor to adjust normalizations between xis and pin
is fixed at 1:1.086 ( ishida et al .
2007 ) . for more details about suzaku data reduction ,
see paper i. the chandra satellite observed mcg-6 - 30 - 15 several times . in this paper , we use the high energy transmission grating spectrometer ( hetgs ) data taken between 2004 may 19 and 27 , when the source was observed four times resulting in a good exposure time of 522 ksec . for data reduction we used the ciao 4.0 software package ; `` tgextract '' is used to produce pha2 spectral files from the level2 event file .
hetgs is composed of high energy grating ( heg ) and medium energy grating ( meg ) .
the @xmath13 and @xmath14 orders of the heg were combined , and so were the @xmath13 and @xmath14 orders of the meg .
the heg and meg spectra were fitted separately .
errors quoted in this paper are at statistical 90@xmath15 confident level .
in paper i , spectral variation of mcg-6 - 30 - 15 was studied assuming the `` two - component '' spectral model given as , @xmath16 where @xmath17 is the intrinsic power - law spectrum with a high - energy cut - off ( fixed at 160 kev ; guainazzi et al .
1999 ) , @xmath18 is normalization factor of the directly observed component , @xmath19 is normalization factor of the reflected component , @xmath20 is reflection albedo by the cold optically thick matter ( `` pexrav '' in xspec : magdziarz @xmath21 zdziarski 1995 ) , @xmath22 and @xmath23 represent attenuation by a high- and a low - ionized warm absorbers , respectively , and @xmath24 is an iron @xmath25 emission line .
effect of interstellar absorption is always taken into account in the spectral fitting ( `` phabs '' ; balucinska - church @xmath21 mccammon 1992 ) , but its term is omitted for simplicity in the above expression and hereafter .
amount of the disk reflection is measured with @xmath26 , solid angle of the reflector seen from the central source .
ratio of the normalization of the reflection component to that of the cutoff power - law component corresponds to @xmath26/2@xmath27 , assuming isotropic emission of the central source .
we used xstar version 2.1kn8 ( kallman et al . 2004 ) to model the warm absorbers .
the temperature , pressure and density of the warm absorbers are assumed to be @xmath28 k , 0.03 dyne @xmath29 and @xmath30 , respectively .
the incident photo - ionizing spectrum is assumed to be a power - law with the index 2.0 , and the solar abundance by greeves , noels and sauval ( 1996 ) is adopted for the ionized material .
we made a grid model by running xstar for different values of @xmath31 and @xmath32 ; the log @xmath31 values are from 0.1 to 5 ( erg cm s@xmath33 ) and the @xmath32 values are from @xmath34 to @xmath35 ( @xmath2 ) .
the number of steps for log @xmath31 and @xmath32 are both 20 , thus our grid model has @xmath36 grid - points .
the redshift of a warm absorber and the source are fixed at 0.001 and 0.00775 , respectively ( young et al .
2005 , fisher et al .
1995 ) . in paper
i , we found that the spectral variation at various timescales is successfully explained by variations of two parameters , normalization factor of the directly observed component ( @xmath18 ) and apparent slope of the direct component , such that the slope gets steeper for greater normalizations .
the apparent slope change is described by either change of the index of the power - law component ( @xmath17 ) , ionization state of the low - ionized warm absorber , or column density of the low - ionized warm absorber ( @xmath23 ) .
below , we scrutinize the spectral model in paper i carefully , and introduce a more elaborated model . using the chandra hetgs , young et al .
( 2005 ) resolved a narrow emission line at 6.4 kev in mcg-6 - 30 - 15 , which we confirmed via re - analysis of the archival data .
the equivalent width of the line is measured as @xmath37 ev , which requires a cold reflector with a solid angle of @xmath26/2@xmath27 @xmath8 0.3 viewed from the central x - ray source ( e.g. , george @xmath21 fabian 1991 ) . on the other hand , the best - fit values of @xmath18 and @xmath19 obtained in paper
i indicates that @xmath38 .
consequently , in order to account for the observed disk reflection spectrum of an amount of @xmath26/2@xmath27 @xmath8 1 including the outer disk reflection with @xmath26/2@xmath27 @xmath8 0.3 , we need another disk reflection component or a different spectral component which is similar to disk reflection , corresponding to an amount of @xmath26/2@xmath27 @xmath8 0.7 .
thus , we separate the @xmath19 term in ( [ eq1 ] ) into two terms .
one is the reflection continuum component from a cold reflector emitting a narrow iron @xmath39 line , solid angle of which is @xmath26/2@xmath27 @xmath8 0.3 .
the other is introduced to represent the remaining part of the @xmath19 term ; for this , we adopt a heavily absorbed component by warm absorber , since the absorbed component is similar to the reflection component in shape , and does not have to accompany the fluorescent iron emission line .
in fact , a partial covering model has been proposed to explain the continuum spectral shape and variation of mcg-6 - 30 - 15 ( e.g. , matsuoka et al . 1990 ; mckernan and yaqoob 1998 ; miller , turner and reeves 2008 , 2009 ) .
resultantly , we introduce a new spectral model expressed as @xmath40 where @xmath41 is normalization of the direct component , @xmath42 is that for the heavily absorbed component , @xmath43 stands for another warm absorber that is for the ionized optically thick partial absorber .
the normalization factor of the outer reflection component , @xmath44 is constant and related to the average of @xmath41 so that @xmath45 , and @xmath24 is a narrow iron emission line with a fixed equivalent width of @xmath46 ev .
hereafter , this model is referred as the `` three - component '' model .
we try the three - component model to the suzaku average spectrum in 1 40 kev .
we found the fit improves if we additionally put a weak iron neutral edge at 7.11 kev and an absorption line at 7.0 kev .
the former is considered to be due to a slight difference of the assumed @xmath26 from the exact value , and the latter due to deficiency of the h - like iron absorption line equivalent - width already included in the xstar model .
figure @xmath47 and table [ miyakawa_model_ave_suzaku ] show the fitting result for the suzaku average spectrum with the three - component model in 1 40 kev .
parameters of the high - ionized and low - ionized warm absorbers are similar to those in paper i. for the newly introduced absorption component ( @xmath43 ) , we obtain @xmath48 1.6 @xmath49 10@xmath50 @xmath2 and @xmath51 .
normalizations of the heavily absorbed component and the direct component are 0.60 and 1.44 ( @xmath52 photons s@xmath33@xmath2 at 1 kev ) , respectively , which suggests that the source is partially covered by the ionized thick matter by @xmath53 30 % .
we note that the absorbing matter is marginally compton thick , so that the compton scattering may not be negligible for precise spectral calculation ( e.g. , yaqoob 1997 ) .
this is not considered in the present paper , and will be a subject for future work .
cc + @xmath32 ( @xmath54 @xmath29 ) & 1.5@xmath55 + + @xmath32 ( @xmath56 @xmath29 ) & 2.4@xmath57 + log @xmath31 & [email protected] + + @xmath32 ( @xmath54 @xmath29 ) & [email protected] + log @xmath31 & 1.54@xmath59 + + 10@xmath60 ph / s/@xmath61 at 1 kev & [email protected] + + 10@xmath60 ph / s/@xmath61 at 1 kev & [email protected] + + @xmath32 ( @xmath35 @xmath29 ) & 1.6@xmath62 + log @xmath31 & 1.57@xmath63 + + photon index & 1.91@xmath64 + @xmath65 ( kev ) & 160(fixed ) + + & @xmath660.3 + + cosincl & 0.866(fixed ) + + line e ( kev ) & 6.35 ( fixed ) + sigma ( kev ) & 0.01 ( fixed ) + norm ( @xmath67ph / s/@xmath61 ) & [email protected] + ew ( ev ) & 27@xmath587 + edge e ( kev ) & 7.11 ( fixed ) + maxtau & [email protected] + line e ( kev ) & 7.0 ( fixed ) + sigma ( kev ) & 0.01(fixed ) + norm ( @xmath68ph / s/@xmath61 ) & @xmath692.4 + line e ( kev ) & [email protected] + sigma ( kev ) & 0.01(fixed ) + norm ( @xmath67ph / s/@xmath61 ) & @xmath700.5 + reduced chi - square ( d.o.f ) & 1.12 ( 203 ) + + [ miyakawa_model_ave_suzaku ] ( 90mm,90mm)figure1.eps cc + @xmath32 ( @xmath54 @xmath29 ) & 3.5@xmath71 + + e ( kev ) & [email protected] + index & @xmath72 ( fixed ) + @xmath73 ( @xmath74 ) & 220@xmath75 + @xmath76 ( @xmath74 ) & 400 ( fixed ) + inclination ( deg ) & 30 ( fixed ) + norm ( @xmath67 ) & 1.3@xmath77 + ew ( ev ) & 40@xmath5820 + + @xmath32 ( @xmath56 @xmath29 ) & 2.9@xmath78 + log @xmath31 & [email protected] + + @xmath32 ( @xmath54 @xmath29 ) & 3.7@xmath79 + log @xmath31 & 1.52@xmath80 + + 10@xmath60ph / s/@xmath61 at 1 kev & 1.44@xmath81 + + 10@xmath60ph / s/@xmath61 at 1 kev & 0.55@xmath82 + + @xmath32 ( @xmath35 @xmath29 ) & 1.7@xmath83 + log @xmath31 & 1.65@xmath84 + + photon index & 1.91@xmath64 + @xmath65 ( kev ) & 160(fixed ) + + & @xmath850.3 + + cosincl & 0.866(fixed ) + + line e ( kev ) & 6.35 ( fixed ) + sigma ( kev ) & 0.01 ( fixed ) + norm ( @xmath68ph / s/@xmath61 ) & 5.4@xmath86 + ew ( ev ) & 12@xmath87 + edge e ( kev ) & 7.11 ( fixed ) + maxtau & [email protected] + line e ( kev ) & 7.0 ( fixed ) + sigma ( kev ) & 0.01(fixed ) + norm ( @xmath68ph / s/@xmath61 ) & @xmath88 + line e ( kev ) & [email protected] + sigma ( kev ) & 0.01(fixed ) + norm ( @xmath67ph / s/@xmath61 ) & @xmath700.5 + reduced chi - square ( d.o.f ) & 1.07 ( 200 ) + + [ miyakawa_diskline_model_ave_suzaku ] ( 90mm,90mm)figure2.eps several authors claim presence of an extremely broad `` disk line '' in the energy spectrum of mcg-6 - 30 - 15 ( e.g. , miniutti et al .
therefore , we examine if presence of the disk line is reconciled with our three - component model .
we analyze the time - average spectrum of mcg-6 - 30 - 15 adding a disk line model from a fast - rotating kerr black hole ( laor 1991 ) .
thus , the model we adopt here has both the narrow line from very far from the black hole , and a putative broad line which is from the region very close to the black hole .
we try to fit the spectrum with laors disk line model with varying the inner disk radius , @xmath73 .
table [ miyakawa_diskline_model_ave_suzaku ] shows the best - fitting result for the suzaku average spectrum adding laor s model to the three - component model , where the fit is slightly improved with @xmath89 200 @xmath90 and the reduced chi - square is 1.07 ( @xmath91/d.o.f = 214/200 ) .
note that the additional line is only mildly broadened with @xmath89 200 @xmath90 , and the equivalent width of the broad disk line is just 40 @xmath58 20 ev .
figure @xmath92 shows the chi - squares of fitting to the average spectrum as a function of the inner disk radius . as seen from this figure ,
@xmath93 chi - squares become greater as @xmath73 gets less than @xmath94 200 @xmath90 , and it exceeds 2.71 when @xmath73 is less than 9 @xmath90 .
namely , our three - component model is only reconciled with a mildly broad disk line from the region @xmath95 with 90 % confidence .
the reason we do not require a strong , extremely distorted disk line is primarily because that the newly introduced heavily absorbed component in our three - component model ( the @xmath96 term in equation [ eq2 ] ) has an ionized iron edge feature which resembles the broad disk line shape ( figure [ miyakawa_model_fit_suzaku ] ) .
secondary reason is that the warm absorbers ( @xmath23 and @xmath22 in equation [ eq2 ] ) have spectral curvatures below @xmath85 kev without which the residual may look like a low - energy tail of the disk line .
consequently , we see that the disk line feature is dependent on the choice of the continuum spectral models , and an extremely distorted disk line is not required in our three - component model . next , we apply the three - component model to the eight `` intensity - sliced spectra '' ( the same spectral sets as used in paper i ) to investigate for spectral variations .
the method of creating the intensity - sliced energy spectra is as follows : ( 1 ) we created a light curve ( the average of xis0 , xis2 , and xis3 ) , with a bin - width of 128 s in the 0.2 12 kev band .
we found that the counting rate varies from @xmath97 to @xmath98 cts s@xmath33 .
( 2 ) we chose eight time - periods when the source intensity is in the ranges of 11.75 , 1.752.50 , 2.503.25 , 3.254.00 , 4.004.75 , 4.755.50 , 5.506.25 , 6.257.00 cts s@xmath33 .
these intensity ranges are chosen so that the exposure time for each intensity bin be approximately equal .
( 3 ) from the eight time - periods corresponding to the different source flux levels , we created eight intensity - sliced energy spectra ( the sum of xis0 , xis2 , and xis3 ) . here
, all the parameters are made the same for the eight spectra except the following three parameters ; the direct power - law normalization , @xmath41 , the absorbed power - law normalization , @xmath42 , and ionization degree of the low - ionized warm absorber , @xmath23 .
attenuation by a warm absorber is expressed as @xmath99 , where @xmath100 is energy and ionization - degree dependent cross - section and @xmath101 is the hydrogen column density of the low - ionized absorber .
variation of @xmath23 is due to either a variation of @xmath100 or @xmath101 , or both . here , we investigate two cases , one with a variation of the cross - section @xmath100 due to change of the ionization parameter @xmath31 , the other with a variation of @xmath102 .
eight spectra are fitted simultaneously , and the both cases equally exhibit acceptable fits with reduced chi - square values 1.09 and 0.95 ( d.o.f = 1135 ) , respectively .
thus , we could not tell if the ionization parameter or column density of the low - ionized warm absorber is more variable from spectral fitting alone .
when @xmath101 is fixed , the top and bottom panel of figure @xmath103 shows the relation between @xmath41 and @xmath104 and that between @xmath41 and @xmath42 , respectively .
similarly , when @xmath31 is fixed , the top and bottom panel of figure @xmath105 shows the relation between @xmath41 and @xmath101 and that between @xmath41 and @xmath42 , respectively . in either case ,
correlations between each two of the three parameters are obvious .
note that we see two independent correlations among three free parameters .
this indicates that there should be a single principle parameter which is primarily responsible for the observed spectral variations ( section [ sec : vpcmodel ] ) .
( 90mm,90mm)figure3a.eps ( 90mm,90mm)figure3b.eps ( 90mm,90mm)figure4a.eps ( 90mm,90mm)figure4b.eps we have seen that the three - component model is effective to describe the suzaku spectral variation , and that the ionization degree of the low - ionized warm absorber is variable with intensities .
next , we will see if the three - component model is valid to describe the chandra hetg spectra and their variations .
strength of chandra hetg compared to suzaku is that it can study variation of individual low - energy absorption lines corresponding to change of the ionization degree .
first , we fit the chandra / hetgs spectra with the three - component model in 1.0 7.5 kev . since the energy range is much smaller than that of suzaku , some parameters are not constrained and thus fixed to the suzaku best - fit values . in figure @xmath106 and table [ miyakawa_model_ave_chandra ]
, we show the fitting result for the average spectrum .
we see that the chandra hetg spectra are successfully modeled with the three - component model with similar parameters to those of suzaku .
next , we study possible ionization degree variations in the chandra / hetgs data by extracting the `` bright spectrum '' and `` faint spectrum '' as follows : 1 ) create a light curve with a time - bin - width of 128 sec , and calculate average counting rates .
2 ) create the `` bright spectrum '' from the period when the meg and heg count rates ( 0.410 kev ) are higher than the average , and the `` faint spectrum '' when the count rates are lower than the average .
we fit the bright / faint spectra in the 1.151.55 kev band with a single power - law and two negative gaussians for the mgxi ( he - like 1.34 kev ) and mgxii ( h - like 1.47 kev ) absorption lines .
we also fit the bright / faint spectra in the 1.752.10 kev band separately with a single power - law and two negative gaussians to investigate for the sixiii ( he - like 1.851.86 kev ) and sixiv ( h - like 2.00 kev ) lines . in figure @xmath107
, we show the fitting result for the bright / faint spectra in both energy bands .
we could fit the bright and faint spectra with the common photon index , while only the power - law normalization and the normalization of absorption lines are varied .
the best fit parameters are shown in table @xmath108 .
we notice that the mgxi absorption line equivalent width is larger in the faint state , while that of the mgxii line is larger in the bright state .
this is understood as due to change of the ionization state of the warm absorber , such that mgxi is more abundant in the faint state , while mgxii is more abundant in the bright state .
in fact , ion fraction of mgxi and that of mgxii become equal at @xmath109 ( e.g. , kallman @xmath21 bautista 2001 ) , that is about the best - fit value of the low - ionized warm absorber ( tables [ tab : chandra3comp ] ) . as for si absorption lines ,
the sixiii line equivalent width is oppositely larger for the bright spectrum , which is explained by the low - ionized warm absorber with @xmath110 , since ion fraction of sixiii increases when @xmath104 varies from @xmath111 to @xmath112 ( kallman @xmath21 bautista 2001 ) .
fraction of sixiv should increase more dramatically when @xmath104 increases accordingly ( kallman @xmath21 bautista 2001 ) , but the observed equivalent widths of the sixiv line is not much different .
this result is explained by the presence of the high - ionized warm absorber at @xmath113 , which produces a significant amount of sixiv but not sixiii .
consequently , observed variation of the mg and si absorption line equivalent - widths is primarily explained by change of the ionization degree of the low - ionized warm absorber according to the flux changes .
( 90mm,90mm)figure5.eps ( 90mm,90mm)figure6a.eps ( 90mm,90mm)figure6b.eps [ around_mg ] [ tab : chandra3comp ] cc + @xmath32 ( @xmath54 @xmath29 ) & 1.5(fixed ) + + @xmath32 ( @xmath56 @xmath29 ) & 2.4(fixed ) + log @xmath31 & [email protected] + redshift ( @xmath114 ) & 1.8@xmath115 + + @xmath32 ( @xmath54 @xmath29 ) & 3.7(fixed ) + log @xmath31 & [email protected] + redshift ( @xmath114 ) & [email protected] + + 10@xmath60 ph / s/@xmath61 at 1 kev & [email protected] + + 10@xmath60 ph / s/@xmath61 at 1 kev & @xmath1160.2 + + @xmath32 ( @xmath35 @xmath29 ) & 1.6(fixed ) + log @xmath31 & 1.57(fixed ) + redshift ( @xmath114 ) & 1(fixed ) + + photon index & [email protected] + @xmath65 ( kev ) & 160(fixed ) + + & @xmath660.3 + + cosincl & 0.866(fixed ) + + line e ( kev ) & 6.34 ( fixed ) + sigma ( kev ) & 0.018 ( fixed ) + norm ( @xmath67ph / s/@xmath61 ) & [email protected] + ew ( ev ) & 25@xmath588 + edge e ( kev ) & 7.11 ( fixed ) + maxtau & 0.13@xmath117 + line e ( kev ) & 6.98@xmath118 + sigma ( kev ) & 0.01(fixed ) + norm ( @xmath68 ph / s/@xmath61 ) & @xmath119 + reduced chi - square ( d.o.f ) & 0.92 ( 5099 ) + + [ miyakawa_model_ave_chandra ] lccc & bright & faint + + line e ( kev ) & + sigma ( 10@xmath120 kev ) & + norm ( 10@xmath121ph / s/@xmath61 ) & @xmath1220.3 & @xmath1230.2 + ew ( ev ) & @xmath1240.5 & @xmath1250.4 + + line e ( kev ) & + sigma ( 10@xmath120 kev ) & + norm ( 10@xmath121ph / s/@xmath61 ) & @xmath1230.2 & @xmath1260.2 + ew ( ev ) & @xmath1270.3 & @xmath1230.5 + photon index & + norm ( 10@xmath120ph / s/@xmath61 at 1 kev ) & [email protected] & [email protected] + reduced chi - square & + & bright & faint + + line e ( kev ) & + sigma ( 10@xmath128kev ) & + norm ( 10@xmath121 ph / s/@xmath61 ) & @xmath1230.2 & [email protected] + ew ( ev ) & @xmath1290.5 & @xmath1300.7 + + line e ( kev ) & + sigma ( 10@xmath128kev ) & + norm ( 10@xmath121 ph / s/@xmath61 ) & @xmath1310.4 & @xmath1320.4 + ew ( ev ) & @xmath1331.0 & @xmath1341.4 + photon index & + norm ( 10@xmath120 ph / s/@xmath61 at 1 kev ) & [email protected] & [email protected] + reduced chi - square & + [ table_chandra][table_chandra2 ] ccc & bright & faint + + @xmath32 ( @xmath54 @xmath29 ) & + + @xmath32 ( @xmath56 @xmath29 ) & + log @xmath31 & [email protected] & @xmath135 + redshift ( @xmath114 ) & + + @xmath32 ( @xmath54 @xmath29 ) & + log @xmath31 & [email protected] & [email protected] + redshift ( @xmath114 ) & + + 10@xmath60 ph / s/@xmath61 at 1 kev & [email protected]&@xmath136 + + 10@xmath60 ph / s/@xmath61 at 1 kev & + + @xmath32 ( @xmath35 @xmath29 ) & + log @xmath31 & + redshift ( @xmath114 ) & + + photon index & + @xmath65 ( kev ) & + + & + + cosincl & + + line e ( kev ) & + sigma ( kev ) & + norm ( @xmath68ph / s/@xmath61 ) & + ew ( ev ) & 14@xmath588 & @xmath137 + edge e ( kev ) & + maxtau & + line e ( kev ) & + sigma ( kev ) & + norm ( @xmath68 ph / s/@xmath61 ) & + reduced chi - square ( d.o.f ) & + + [ miyakawa_model_slice_chandra ] ccc & bright & faint + + @xmath32 ( @xmath54 @xmath29 ) & + + @xmath32 ( @xmath56 @xmath29 ) & + log @xmath31 & [email protected] & [email protected] + redshift ( @xmath114 ) & + + @xmath32 ( @xmath54 @xmath29 ) & [email protected] & [email protected] + log @xmath31 & + redshift ( @xmath114 ) & + + 10@xmath60 ph / s/@xmath61 at 1 kev & [email protected]&@xmath138 + + 10@xmath60 ph / s/@xmath61 at 1 kev & + + @xmath32 ( @xmath35 @xmath29 ) & + log @xmath31 & + redshift ( @xmath114 ) & + + photon index & + @xmath65 ( kev ) & + + & + + cosincl & + + line e ( kev ) & + sigma ( kev ) & + norm ( @xmath68ph / s/@xmath61 ) & + ew ( ev ) & 17@xmath588 & @xmath139 + edge e ( kev ) & + maxtau & + line e ( kev ) & + sigma ( kev ) & + norm ( @xmath68 ph / s/@xmath61 ) & + reduced chi - square ( d.o.f ) & + + [ miyakawa_model_slice_chandra_columnb ] we further simultaneously fit the bright / faint spectra in the 1.07.5 kev band with the three - component model ( table @xmath140 and [ miyakawa_model_slice_chandra_columnb ] ) by varying normalization of the direct component and ionization degree of the low - ionized warm absorber . in stead of allowing the normalization of the absorbed power - law component to be free , we fixed it to the value of a constant total normalization minus the direct power - law normalization , in order to confirm the anti - correlation between the direct component and absorbed component as observed in suzaku ( bottom panels in figures [ relation_for_the_slice_spectra ] and [ relation_for_the_slice_spectra_nh ] ) . in table
@xmath140 , we show results of allowing the ionization degrees of the low - ionized warm absorber to be free while the column density is fixed ; we see the parameter is significantly variable ( @xmath104 = [email protected] and [email protected] ) .
also , ionization degree of the high - ionized warm absorber is made a free parameter , but these values are hardly variable ( @xmath104 = [email protected] and [email protected] ) . because of the presence of constant high - ionized warm absorber , varying column density of the low - ionized warm absorber effectively changes the average ionization states .
therefore , in stead of varying ionization degree of the low - ionized warm absorber , the bright and faint spectral variation can be equally explained by varying its column density .
this result is shown in table [ miyakawa_model_slice_chandra_columnb ] . in summary ,
chandra / hetgs data confirm that the spectral variation is described by change of only two spectral parameters in the framework of the three - component model ( equation [ eq2 ] ) , normalization of the direct component ( @xmath41 ) and either ionization degree or column - density of the low - ionized warm absorber(@xmath23 ) .
next , we create a time - sequence of the energy spectra from suzaku data , and apply the three - component model .
we have made energy spectra every 20 ksec , and performed spectral fits to all the sequential 20 ksec spectra with the three - component model , to investigate time variation of the spectral parameters .
we chose a 20 ksec interval , since variation amplitudes in the 110 kev band is maximum at around 2080 ksec ( inoue , miyakawa and ebisawa 2011 ) . in the spectral fits ,
@xmath41 , @xmath42 and @xmath23 are treated as time - variable in equation ( [ eq2 ] ) .
the fits are made for the three observation sequences separately ( obsid:700007010 , 700007020 and 700007030 ) .
we have seen that observed variation of @xmath23 is equally represented with change of the column density @xmath102 or ionization parameter , @xmath31 ( section [ sec : sliced ] ) .
hereafter , we vary @xmath102 while @xmath31 is assumed to be invariable throughout the observation .
in fact , as we shall see below , variation of @xmath102 is more physically reasonable than that of @xmath31 .
figure [ direct_absorbed_20ksec ] shows a correlation between @xmath41 ( direct power - law normalization ) and @xmath42 ( absorbed power - law normalization ) , where an inverse correlation is seen between them , as well as in the bottom panels in figure [ relation_for_the_slice_spectra ] and [ relation_for_the_slice_spectra_nh ] .
the correlation coefficient between @xmath41 and @xmath42 is @xmath141 .
if we fit with a linear relation , @xmath142 , we obtain @xmath143 , @xmath144 and @xmath145 ( d.o.f.=35 ) . on the other hand ,
fits with a constant @xmath41 ( @xmath146 with @xmath147 for d.o.f.=36 ) or @xmath42 ( @xmath148 for d.o.f.=36 ) are not acceptable .
we found that slope of the anti - correlation between @xmath41 and @xmath42 is not very far from @xmath14 , which suggests that sum of the two parameters may be rather constant .
therefore , we introduce the `` total normalization '' , @xmath149 , as , @xmath150 similarly , we introduce the `` covering fraction '' , @xmath151 , as @xmath152 ( 90mm,90mm)figure7.eps ( 90mm,90mm)figure8a.eps ( 90mm,90mm)figure8b.eps ( 90mm,90mm)figure8c.eps ( 90mm,90mm)figure9a.eps ( 90mm,90mm)figure9b.eps ( 90mm,90mm)figure9c.eps obviously , @xmath153 thus , instead of @xmath154 , we can equivalently use @xmath155 to describe the observed spectral variation . in figure
[ correlation_20ksec ] , we show correlations among these three parameters .
cccccc & & & + @xmath156 & @xmath157 & cc@xmath158 & @xmath159 & @xmath160 & @xmath161 + @xmath149 & @xmath151 & @xmath162 & @xmath163 & @xmath164 & @xmath165 + @xmath149 & @xmath101 & @xmath166 & @xmath163 & n.a . &
@xmath167 + @xmath151 & @xmath102 & @xmath168 & @xmath169 & n.a . & @xmath170 + + [ tab : correlation ] we study variations of and correlations among the three parameters , @xmath155 , by calculating the correlation coefficients between the two , as well as fitting the correlation with a constant parameter or a linear - function .
results are summarized in table [ tab : correlation ] .
consequently , we can see a good `` orthogonality '' between @xmath149 and @xmath151 , and @xmath149 and @xmath102 ( i.e. , the fitting parameter @xmath171 is almost zero , where @xmath172 or @xmath173 ) .
namely , the total power - law normalization @xmath149 is not significantly variable , while the covering fraction , @xmath151 , and column - density of the low - ionized absorber , @xmath101 , are more variable .
furthermore , @xmath151 and @xmath102 indicate a good correlation with the correlation - coefficient 0.64 .
therefore , spectral variation of mcg-6 - 30 - 15 at a timescale of 20 ksec is primarily described by a single parameter , either @xmath151 or @xmath101 .
we remind , however , there is a hint of weak anti - correlation between @xmath149 and @xmath151
. further sophistication of the model may explain this correlation ( section [ sec : blocker ] ) .
figure [ 20ksec_par_var ] indicates another representation of the spectral parameter variations . as a function of time ,
the three parameters , @xmath174 , respectively normalized by their average values , are indicated .
the @xmath175 values around the average ( 1.0 ) are 2.4 , 2.9 and 10.3 for @xmath176 respectively ( d.o.f.=36 ) .
we can see @xmath149 is the least variable , while @xmath151 and @xmath101 are in sync and more significantly variable . in the following ,
we will discuss physical meaning of these spectral variations .
[ phenomenological_inte_section ] we have seen that suzaku and chandra energy spectra of mcg6 - 30 - 15 are successfully described by the three - component model introduced with equation ( [ eq2 ] ) .
we found @xmath42 , normalization of the heavily absorbed component , is anti - correlated with normalization of the direct component , @xmath41 ( bottom panels of figures [ relation_for_the_slice_spectra ] and [ relation_for_the_slice_spectra_nh ] , and figure [ direct_absorbed_20ksec ] ) . the spectral variation indicates that the `` total normalization '' , @xmath149 , ( equation [ eq3 ] ) is not so variable , while the covering fraction , @xmath151 , ( equation [ eq4 ] ) and column - density of the low - ionized warm absorber are in sync and more significantly variable ( figure [ correlation_20ksec ] and [ 20ksec_par_var ] )
. how should we interpret this characteristic spectral variation ?
since @xmath177 and @xmath178 represent amounts of the non - absorbed and absorbed fluxes emitted from the central x - ray source , respectively , the anti - correlation is naturally understood that a rather constant x - ray source is partially obscured by absorbing matters with variable covering fraction , @xmath151 .
the `` difference variation function analysis '' introduced by inoue , miyakawa and ebisawa ( 2011 ) also confirms that the ( @xmath149 , @xmath151 ) parameter - set is more orthogonal and fundamental than the ( @xmath41 , @xmath42 ) parameter - set .
correlation between the low - ionized absorber @xmath101 and @xmath151 ( bottom panel of figure [ correlation_20ksec ] ) is understood as follows : the low - ionized absorber represented by @xmath179 is optically thin ( table [ miyakawa_model_ave_suzaku ] ) , so that @xmath180 on the other hand , the same absorption effect can be expressed by a partial covering model with a fixed column - density as , @xmath181 @xmath182 namely , @xmath183 that is seen in the bottom panel of figure [ correlation_20ksec ] .
as we have seen , the observed spectral variation of mcg-6 - 30 - 15 is primarily explained by variation of the partial covering fraction , @xmath151 .
below , we designate our model as the `` variable partial covering model '' . in the framework of the variable partial covering model , the suzaku intensity - sliced spectra in 1 40 kev
is explained mostly by change of the covering fraction . using the total normalization @xmath149 and the covering fraction @xmath151 ,
our three - component model ( [ eq2 ] ) is rewritten as follows : @xmath184 in equations ( [ eq9 ] ) and ( [ eq6 ] ) , only variable parameters are @xmath151 and @xmath149 . in figure [ covering_fraction_change ] and table [ table1_slice ] , we show an example of the simultaneous fit of the suzaku intensity - sliced spectra with variable @xmath151 ( the same dataset as used in section [ sec : sliced ] ) . the eight intensity - sliced spectra are successfully fitted with the constant normalization , @xmath185 photons s@xmath33 @xmath2 at 1 kev ( except the brightest one where @xmath149 is 1.5 times higher ) , while @xmath151 is variable from 0.63 ( dimmest ) to null ( brightest ) .
the well - known small variability in the iron energy band ( section 1 ) is also nicely explained in the framework of the variable partial covering model . in order to see energy dependence of the variation at a specific timescale
, we took the following method to calculate the `` difference variation function '' , @xmath186 ( inoue , miyakawa and ebisawa 2011 ) : ( 1 ) we choose a time - interval of interest , @xmath187 , to investigate for spectral variations ; here , we take @xmath18840 ksec . ( 2 )
energy spectra are calculated for every @xmath189 , namely , 20 ksec .
( 3 ) for every two contiguous energy spectra , we recognize the `` brighter spectrum '' and the `` fainter spectrum '' , @xmath190 and @xmath191 , respectively .
( 4 ) we average all the brighter spectra and all the fainter spectra .
( 5 ) thus , for a given time - period @xmath187 , we create a single `` averaged bright spectrum '' @xmath192 and a single `` averaged faint spectrum '' @xmath193 .
( 6 ) we calculate difference of the averaged bright spectrum and the averaged faint spectrum relative to the average , as @xmath194 in figure [ cov_fac_simulation ] , we show the difference variation function thus calculated at a timescale of 40 ksec . together
, we show a model prediction of the difference variation function calculated in the framework of the variable partial covering model ( eq . 29 in inoue , miyakawa and ebisawa 2011 ) , where the covering fraction @xmath151 is assumed to be the major variable parameter , while the total normalization @xmath149 adds minor fluctuation .
agreement of the observation and model calculation is obvious .
next , we estimate physical parameters of the warm absorbers in equation ( [ eq6 ] ) , @xmath22 , @xmath23 and @xmath43 , which are high - ionized warm absorber , thin low - ionized warm absorber and thick partial absorber , respectively .
we provide suffixes , @xmath195 and @xmath196 to distinguish column densities and ionization parameters of these warm absorbers . in the following ,
@xmath197 is distance from the central x - ray source to the absorber , @xmath198 is a representative thickness of the absorbing region along the line of sight , and @xmath199 is density .
( 90mm,90mm)figure10.eps ( 90mm,90mm)figure11.eps ccccccccc state & slice 1 & slice 2 & slice 3 & slice 4 & slice 5 & slice 6 & slice 7 & slice 8 + + @xmath32 ( @xmath54 @xmath29 ) & + + @xmath32 ( @xmath56 @xmath29 ) & + log @xmath31 & + + @xmath151 & 0.632 & 0.356 & 0.269 & 0.195 & 0.181 & 0.166 & 0.122 & @xmath200 + + @xmath201 ( @xmath202 ) & + @xmath32 & + log @xmath31 & + + 10@xmath60ph / s/@xmath61 at 1 kev & & 2.71 + + & + + & + + @xmath32 ( @xmath35 @xmath29 ) & + log @xmath31 & + + photon index & + @xmath65 ( kev ) & + + pexrav k ( 10@xmath120 ) & + + cosincl & + + line e ( kev ) & + sigma ( kev ) & + norm ( @xmath67ph / s/@xmath61 ) & + edge e ( kev ) & + maxtau & + line e ( kev ) & + sigma ( kev ) & + norm ( @xmath68ph / s/@xmath61 ) & + line e ( kev ) & + sigma ( kev ) & + norm ( @xmath67ph / s/@xmath61 ) & + reduced chisq ( d.o.f ) & + + [ table1_slice ] first , in general , we note the following relations hold : @xmath203 @xmath204 for the luminosity , we take the representative value @xmath205 erg / s , which is valid for order estimation , since the flux variation is much less than an order of magnitude . from model
fitting ( table [ table1_slice ] ) , @xmath206 @xmath207 10@xmath208 erg cm / s , @xmath209 @xmath207 10@xmath210 @xmath2 .
hence , @xmath211 our analysis has indicated that the high - ionized warm absorber is constant , while partial covering fraction is variable .
this suggests the variable partial absorbers be embedded in the largely extended , static high - ionized warm absorber .
hence , we assume @xmath212 @xmath8 @xmath213 , and take @xmath214 10@xmath215 cm and @xmath216 10@xmath217 cm@xmath120 for typical parameters of the high - ionized warm absorbers . from model
fitting ( table [ table1_slice ] ) , @xmath219 10@xmath220 erg cm / s , @xmath221 10@xmath222 @xmath2 .
hence , @xmath223 here , we may further constrain the parameters of the partial absorber from variation timescale of the partially absorbed component . in our variable partial covering model , typical variation timescale ( @xmath224 sec ; figure [ 20ksec_par_var ] )
corresponds to the crossing time of the x - ray absorbing clouds in front of the x - ray source .
we will call these clouds as the `` low - ionized clouds '' , hereafter . assuming that size of the x - ray source and that of a low - ionized cloud have similar dimensions , the crossing timescale is expressed as @xmath225 , where @xmath226 is the velocity of a low - ionized cloud . hence , @xmath227 where @xmath226 is normalized to a typical velocity of the broad line region ( blr ) clouds . combining @xmath228 and @xmath229
, we estimate parameters of the partial absorbers as follows : @xmath230 @xmath231 note that a low - ionized cloud may have internal ionization structures , which we will see below . in our variable partial covering model
, the low - ionized warm absorber ( @xmath23 ) and the partial absorber ( @xmath43 ) are closely associated . presumably , they are different parts of the low - ionized clouds .
we point out that presence of such internal structures in the blr clouds is also suggested in mrk 766 ( risaliti et al .
2011 ) . in our model
, the low - ionized warm absorber has a fixed column - density @xmath201 with a variable partial covering factor ( equation [ eq9 ] ) ; we simply write this column density as @xmath101 below . from model
fitting ( table [ table1_slice ] ) , @xmath232 @xmath2 and @xmath233 10@xmath234 erg cm / s .
if we take the distance of the low - ionized clouds @xmath235 , @xmath236 @xmath237 note that values of @xmath238 and @xmath239 are between those of the high - ionized absorber ( @xmath240 cm@xmath120 and @xmath241 ) and the partial absorber ( @xmath242 @xmath2 and @xmath243 ) , respectively .
the thickness @xmath244 is much smaller than that of the partial absorber @xmath245 cm .
therefore , the low - ionized absorber is considered to be in the boundary layer between the low - ionized cloud and the high - ionized warm absorber .
we may call this putative layer as the `` cloud envelope '' .
meanwhile , it is known that there takes place thermal instability in the range of @xmath246 erg cm / s ( reynolds @xmath21 fabian 1995 ) .
therefore , it is reasonable that the cloud envelope continuously connects the partial absorber in the low - ionized cloud and the surrounding constant high - ionized warm absorber .
we notice the thomson optical depth ( @xmath247 ) of the partial absorber is about unity . presumably , this is not just a coincidence , but suggests that the low - ionized cloud has internal density gradients extending up to more than unit thomson optical depth , such that attenuation with continuously different optical depths is approximated by unit thomson optical depth .
thus , the low - ionized cloud may well have central , cold thomson thick cores ( @xmath248 @xmath2 ) , which are completely opaque to incoming x - rays .
we may not see these `` x - ray blockers '' directly , but if we take into account the effect of the x - ray blockers , the weak anti - correlation between the total normalization @xmath149 and the covering fraction @xmath151 ( top - panels of figure [ correlation_20ksec ] ) might be explained . [ cols="<,^,^,^,^,^,^,^",options="header " , ] finally , we present a physical picture surrounding the x - ray source to explain the observations . in table [ tab : absorbers ] , we summarize parameters of the absorbers in our model .
basic points of our variable partial covering model are the following : 1 . observed x - ray flux and spectral variability is primarily caused by random passages of the low - ionized clouds in front of the central x - ray source .
a typical size of each low - ionized cloud is @xmath245 cm and a typical velocity is @xmath249 cm / s ( equation [ sizeofcloud ] ) .
central x - ray source is not always fully blocked , nor fully exposed .
the covering fraction is variable between null to @xmath250 ( table [ table1_slice ] ) , and it fluctuates by an amplitude of @xmath251 % ( figure [ 20ksec_par_var ] ) . we may estimate the average covering fraction @xmath252 from intensity - sliced spectra as @xmath252 = @xmath253 , where @xmath254 and @xmath255 are exposure time for each spectrum and the total observation , respectively .
we thereby obtain @xmath256 0.3 .
the integrated covering fraction is considered to be about 0.1 for blr ( e.g. , blandford et al .
thus , our results suggest that the partial absorption takes place in the blr , and the low - ionized clouds corresponds to the blr blobs .
we may estimate number of the absorbing clouds in the field of view , @xmath257 , and size of the central x - ray source , @xmath258 , as follows .
assuming that fluctuation of the covering fraction , @xmath259 , is determined by statistical fluctuation of the number of clouds , @xmath260 , @xmath257 is estimated as @xmath261 .
the average covering fraction , @xmath262 , may be written as @xmath263 , thus , @xmath264 cm .
typically , blr blobs have a velocity of @xmath226 @xmath8 10@xmath265 cm / sec . assuming the kepler velocity has a similar value , @xmath266 500 @xmath267 , where @xmath267 is the schwarzschild radius .
identifying this with @xmath268 10@xmath269 cm ( equation [ eq : r2 ] ) , the central black hole mass of mcg-6 - 30 - 15 is @xmath270 .
thus , the central x - ray source ( @xmath271 cm ) is extended to @xmath272 around the black hole . besides variation of the blr coluds ,
the x - ray source is variable on a timescale down to @xmath8 1 ksec ( inoue , miyakawa and ebisawa 2011 ) , which may be compared with the free fall time , @xmath273 .
then , the intrinsic source variation arises at a few times the schwarzshild radius , around the innermost stable circular orbit .
we point out that solid angle of the low - ionized clouds seen from the x - ray source should be much smaller than 4@xmath27 , otherwise we will have to see strong fluorescent lines from the clouds . if a thick absorber with @xmath274 @xmath2 covers the power - law component completely ( @xmath275 ) ,
equivalent width of the iron line to the reflection component is @xmath8 1 kev . considering @xmath276 0.3 and the ratio of the reflection component to the direct component , 10 to 1 , the expected iron line equivalent width is @xmath8 30 ev . as mentioned in subsection 5.5.2 , we found the fit improves if a mildly broad disk - line is added , where the best - fit inner radius and the equivalent width are @xmath277 200 @xmath90 and @xmath278 ev , respectively ( table [ miyakawa_diskline_model_ave_suzaku ] ) .
these mildly broadened , weak iron emission lines may well be expected from fluorescence in the blr clouds .
figure @xmath279 gives a schematic view of internal structure of the low - ionized cloud and the variable partial covering model surrounding the black hole .
the low - ionized cloud ( blr cloud ) has internal ionization structure .
a putative blocker in the core would block the incoming x - rays completely .
the main body of the low - ionized cloud is the thick partial absorber surrounding the core .
still the outer cloud envelope corresponds to the thin low - ionized warm absorber .
central x - ray emission region around the black hole is mildly extended , and partially covered by these low - ionized clouds .
variation of the covering fraction explains most of the observed flux and spectral variations .
finally we comment that the idea of partial covering is not new for mcg-6 - 30 - 15 to explain its spectral shape and varaition ( e.g. , matsuoka et al .
1990 ; mckernan and yaqoob 1998 ; miller , turner and reeves 2008 , 2009 ) .
recently , similar partial covering models are proposed to explain spectral variations of ngc 3516 ( turner et al .
2008 ) , mrk 766 ( risaliti et al . 2011 ) and other seyfert galaxies ( turner & miller 2009 ) .
presumably , partial covering of rather constant x - ray source is a common mechanism in seyfert galaxies .
it is of interest to study spectral variations of seyfert galaxies systematically to see if the observed x - ray variablity is truely due to intrinsic luminosity variation or explained by partial covering of the constant source .
we have analyzed suzaku and chandra archival data of mcg-6 - 30 - 15 and constructed a new model to explain the observed spectral variation .
our main conclusions are summarized below : 1 .
we have shown that the observed energy spectra and spectral variation can be explained by the `` three - component model '' ( equation [ eq2 ] ) that includes ( 1 ) a direct power - law component , ( 2 ) a heavily absorbed power - law component by thich photoionized material , and ( 3 ) a cold disk reflection component far from the black hole .
the first two components are affected by two warm abosrbers having different ionization states .
the ionized iron k - edge of the heavily absorbed component and spectral curvature due to the warm absorbers explain most of the seemingly broad `` disk line '' spectral feature .
consequently , general relativistic interpretation of the broad iron emission line feature in mcg-6 - 30 - 15 is not confirmed .
the claim that mcg-6 - 30 - 15 is a kerr black hole with nearly extreme rotation assuming the relativistic line - broadening ( e.g. , miniutti et al .
2007 ) is questionable .
3 . we propose the `` variable partial covering model '' , in which a central , moderately extended x - ray source is partially covered by variable low - ionized clouds in the line of sight . in this model ,
observed x - ray spectral variation is primarily caused by change of the partial covering fraction of the central source .
these absorbing clouds presumably correspond to the fast - moving broad line region ( blr ) clouds .
the authors would like to thank profs .
p. ycki , y. terashima , y. ueda and k. makishima for fruitful discussion .
this research has made use of public suzaku data obtained through the data archives and transmission system ( darts ) , provided by the institute of space and astronautical science ( isas ) at japan aerospace exploration agency ( jaxa ) , public chandra data obtained through the high energy astrophysics science archive research center ( heasarc ) at nasa / goddard space flight center . for data reduction
, we used the software provided by the high energy astrophysics science archive research center ( heasarc ) at nasa / goddard space flight center and chandara x - ray center ( cxc ) at harvard - smithsoniann center for astrophysics .
k.e . acknowledges ms .
akiko sato for helping us illustrate our model and producing a beautiful artwork . | we propose a simple spectral model for the seyfert 1 galaxy mcg-6 - 30 - 15 that can explain most of the 1 40 kev spectral variation by change of the partial covering fraction , similar to the one proposed by miller et al .
( 2008 ) .
our spectral model is composed of three continuum components ; ( 1 ) a direct power - law component , ( 2 ) a heavily absorbed power - law component by mildly ionized intervening matter , and ( 3 ) a cold disk reflection component far from the black hole with moderate solid - angle ( @xmath0 ) accompanying a narrow fluorescent iron line .
the first two components are affected by the surrounding highly ionized thin absorber with @xmath1@xmath2 and @xmath3 .
the heavy absorber in the second component is fragmented into many clouds , each of which is composed of radial zones with different ionization states and column densities , the main body ( @xmath4@xmath2 , @xmath5 ) , the envelope ( @xmath6@xmath2 , @xmath7 ) and presumably a completely opaque core .
these parameters of the ionized absorbers , as well as the intrinsic spectral shape of the x - ray source , are unchanged at all .
the central x - ray source is moderately extended , and its luminosity is not significantly variable .
the observed flux and spectral variations are mostly explained by variation of the geometrical partial covering fraction of the central source from 0 ( uncovered ) to @xmath80.63 by the intervening ionized clouds in the line of sight .
the ionized iron k - edge of the heavily absorbed component explains most of the seemingly broad line - like feature , a well - known spectral characteristic of mcg-6 - 30 - 15 .
the direct component and the absorbed component anti - correlate , cancelling their variations each other , so that the fractional spectral variation becomes the minimum at the iron energy band ; another observational characteristic of mcg-6 - 30 - 15 is thus explained . |
quantum discord is the most well - known measure of non - classical correlations and is thought to differentiate quantum and classical systems @xcite .
it is defined as the minimum difference between two classically equivalent formulations of the mutual information between subsystems @xmath2 and @xmath3 , @xmath4 and @xmath5 , where @xmath6 is the shannon entropy when @xmath7 is described by a classical probability distribution , and is the von neumann entropy when @xmath7 describes a quantum system . in the quantum case
the conditional entropy depends on the measurement basis , @xmath8 , where @xmath9 is a complete set of orthonormal projectors on @xmath10 such that @xmath11 , @xmath12 is the probability of observing outcome @xmath13 on system @xmath2 , and @xmath14/p_k$ ] is the state of system @xmath3 conditional on the measurement of system @xmath2 returning measurement outcome @xmath13 .
the quantum discord , @xmath15 , thus reduces to @xmath16 where @xmath17 is the reduced density matrix of system @xmath2 . in order to calculate this quantity ,
full state knowledge is required in addition to a minimization over all possible projective measurements on subsystem @xmath2 . while there is no analytical expression for the quantum discord of a general state , the discord for various sets of two - qubit states
have been found @xcite .
in addition , the quantum discord for all two - qubit states has been reduced to the problem of solving a set of transcendental equations @xcite .
one of the results we report here is an expression for the qd in the final state of a two - qubit dqc1 algorithm that can be calculated with only the outcome of the dqc1 algorithm ( trace of the unitary ) and the initial polarization of the top register . motivated by the difficultly in computing the qd , dakic , vedral and brukner @xcite proposed the gqd , which is defined as @xmath18 where @xmath19 is the set of all zero discord states ( @xmath20 . )
these can be written as @xmath21 .
the quantity @xmath22 is the square of the hilbert - schmidt norm of hermitian operators .
note that the gqd , like the original quantum discord , is not symmetric in the subsystems , and @xmath23 is defined as @xmath24 where @xmath25 . shortly after the proposal of this measure of quantum correlations , luo and fu @xcite showed that equation is equivalent to the minimization @xmath26 where @xmath27 is a projective measurement on system @xmath2 , and @xmath28 .
the minimization can be performed analytically for arbitrary @xmath29-dimensional @xcite and @xmath30-dimensional @xcite states .
the resulting expressions are in terms of quantities that are not efficiently experimentally accessible .
however , a method for measuring a tight lower bound on the gqd has been recently proposed @xcite , requiring only a constant number of measurements on up to four copies of the state . in this report
we provide an analytical expression for the gqd of @xmath31-dimensional dqc1-states , and describe a dqc1-algorithm to efficiently estimate it . along @xmath32 . at the conclusion of the circuit ,
measurements of @xmath33 and @xmath34 result in the value of the real and imaginary components of the trace of the unitary , scaled by the value of the polarization of the first qubit . ]
the dqc1 model of computation is an example of a mixed state quantum information processor @xcite .
as shown in figure [ dqc1fig ] , this model has access to only one qubit with non - zero polarization , accompanied by a register of @xmath1 maximally mixed qubits .
the computation consists of two gates : a hadamard on the top , single qubit register , followed by an @xmath1-qubit unitary on the bottom register that is controlled by the state of the top qubit .
the final state of this simple algorithm is what we call the dqc1-state and can be written as @xmath35 expectation value measurements of the pauli matrices @xmath36 and @xmath37 yield the real and imaginary values of the trace of the unitary : @xmath38 and @xmath39 . since calculating the trace of a unitary
does not have a known efficient classical algorithm , a dqc1 computer is believed to be more powerful than its classical counterpart . despite its apparent improvement over classical methods
, dqc1 is known to have zero bipartite entanglement between the top and bottom registers and can only have a small amount of entanglement across any other bipartite splitting @xcite .
it can , however , have quantum correlations characterized by the quantum discord @xcite .
this has been confirmed in two experiments to date : measurement of the discord in a two - qubit optics setup using full state tomography @xcite , and a four - qubit nmr implementation that witnesses discord with a small number of experiments @xcite .
in order to find an expression for the geometric discord in a dqc1-state , let us write eqn . as @xmath40 the first term , the purity of the state of the total system , is invariant under unitary transformations
. therefore it can be calculated for the initial state , and only depends on the initial polarization of the top register : @xmath41 in order to calculate the other two terms , consider parameterizing the measurement on the top qubit as @xmath42 , where @xmath43 and @xmath44 ; @xmath45 and @xmath46 are real and @xmath47 $ ] ; and @xmath48 $ ] .
thus , the state after measurement is @xmath49 for the dqc1-state , the terms @xmath50 and @xmath51 are equivalent , and evaluate to @xmath52 where @xmath53 are the eigenphases of @xmath54 ( i.e. @xmath55 ) , and we have used @xmath56 .
thus , we are left with the task of minimizing @xmath57 over the measurement parameters @xmath58 $ ] and @xmath48 $ ] .
examining first and second order derivatives of @xmath59 with respect to @xmath45 and @xmath60 , we find the optimal measurement parameters to be @xmath61 which reduces the gqd to @xmath62,\ ] ] where @xmath63 , and can be evaluated with a dqc1-algorithm using back - to - back applications of the control-@xmath54 .
the depth of the circuit that evaluates @xmath64 is at most double the one that evaluates @xmath65 , and therefore has the same efficiency of evaluation as the standard dqc1 algorithm .
worth noting here that @xmath66 indicates that the optimal measurement is always in the transverse ( x - y ) plane of the bloch sphere , independent of the implemented unitary .
also note that eqn .
allows for the ability to experimentally evaluate the optimal measurement parameter @xmath67 using the same data collected for the evaluation of @xmath64 . the gqd does not depend on the eigenvectors of the unitary , but rather the distribution of eigenphases
this gives rise to classes of unitaries that generate the same gqd in a dqc1 circuit .
for instance , reproducing a result from ref .
@xcite , the set of unitaries that generate zero gqd ( and hence , zero qd ) will have @xmath68 , implying @xmath69 for some @xmath70 .
this is possible if and only if the eigenphases of @xmath54 are @xmath71 . that is to say @xmath72 , where @xmath2 is a binary observable ( @xmath73 ) . for a unitary drawn randomly according to the haar measure , its eigenphases
are randomly distributed over the unit circle @xcite .
thus , for large @xmath1 , @xmath64 approaches zero and @xmath74 approaches its maximum value of @xmath75 , i.e. the gqd of a dqc1-state decreases exponentially with the number of qubits .
this is in contrast to the average qd of a dqc1-state , which was shown in ref .
@xcite to be independent of the number of qubits for large @xmath1 .
we measure the quantum correlations , as quantified by the gqd , in a liquid - state nuclear magnetic resonance implementation of the dqc1 algorithm .
the qubits in nmr are ensembles of spin-1/2 nuclei . in this experiment
, we use the four carbon-13 nuclei in the molecule trans - crotonic acid ( molecular information and hamiltonian parameters can be found in ref . @xcite . )
the experiment is implemented in a 16.7 t magnetic field bruker avance spectrometer , where , for carbon-13 nuclei ( @xmath76 ) at room temperature , the initial polarization of the thermal state is @xmath77 .
the spins are manipulated with radio frequency pulses numerically generated using the grape algorithm @xcite .
the pulses are designed to have a fidelity of no less than 0.998 and are adjusted for non - linearities in the pulse generation and transmission by placing a pickup coil at the location of the sample and running a feedback loop to iteratively adjust the pulse shape for optimal transmission . , normalized by @xmath78 , as a function of the measurement parameters @xmath45 and @xmath60 , for @xmath79 the output of the four - qubit dqc1 algorithm described in the text , and ( ii ) the geometric distance at the optimal measurement axis @xmath80 .
the dashed lines indicate the parameters that correspond to the optimal measurement .
the experimental data point for the geometric discord of @xmath79 is shown on plot ( ii ) with error bars propagated from experimental uncertainties and the spectral fit . ] for the particular instance of the dqc1 circuit in this experiment , we choose to implement one of the unitary matrices used for approximating the jones polynomial for braids with four strands @xcite .
this is a problem that completely encapsulates the power of the dqc1 model and is of interest in many fields of physics and math .
the unitary has the form @xmath81 , where @xmath82 and @xmath83 .
the dqc1-state , after the application of the control-@xmath54 , is known to have nonzero discord as witnessed in a recent experiment @xcite .
using eqn . , we calculate the expected value of the gqd to be @xmath84 .
as shown in figure [ fig_results ] , this is achieved for an optimal measurement with @xmath85 and @xmath86 rad .
also shown in figure [ fig_results](ii ) is the experimental value of gqd in the dqc1-state , which is found to be @xmath87 by measuring the outcome of a dqc1 circuit with back - to - back applications of control-@xmath54 . for our experimental polarization ( @xmath88 ) ,
the gqd in the final state is @xmath89 .
in the case where the bottom register is a single qubit , we report an analytical expression for the qd of the dqc1-state , thereby contributing to the sets of two - qubit states for which the qd has been analytically solved . commencing from eqn . , a minimization must be performed over the conditional entropy term , @xmath90 , which , in the case of a two - qubit dqc1-state , reduces to @xmath91 where @xmath92 , @xmath93 , and @xmath94 and @xmath60 parameterize the measurement on the top qubit ( see the gqd case above ) .
the minimization with respect to the measurement parameters @xmath45 and @xmath60 results in an optimal measurement characterized by @xmath95 this allows us to write the analytical expression for the qd of a two - qubit dqc1-state as @xmath96 where @xmath97 is the binary entropy , and @xmath98 , which can be measured directly with a dqc1-algorithm .
hence , the quantum discord of a two - qubit dqc1-state is zero for @xmath99 or @xmath100 , and maximum for @xmath101 .
recall , from eqn .
, that the gqd is zero for @xmath102 and maximum for @xmath103 .
consider a general single - qubit unitary , expressed as @xmath104 , where @xmath105 $ ] is a rotation by @xmath106 about the unit vector in three dimensions @xmath107 , @xmath108 are the pauli matrices , and @xmath70 a global phase .
it follows from @xmath109 that @xmath110 and @xmath111 .
thus , the qd and gqd are simultaneously maximum for @xmath112 , and simultaneously zero for @xmath113 , independent of the rotation axis @xmath114
. examples of unitaries that can be written as @xmath115 rotations include bit - flip ( or not ) and hadamard .
in this article we provided analytical expressions for the quantum correlations present in a dqc1-state in terms of experimentally accessible quantities . the geometric quantum discord can be computed for @xmath31-dimensional dqc1-states by implementing a dqc1 algorithm with a second application of the controlled unitary .
the algorithm was experimentally demonstrated for a four - qubit liquid - state nmr implementation of the dqc1 model .
we also showed , that for the special case of a two - qubit system , the quantum discord can be calculated using outcome of the dqc1-circuit ( the trace of the unitary ) . for a typical dqc1-state , where the unitary is chosen uniformly at
random according to the haar measure , we found that the geometric measure of discord scales with @xmath116 , in contrast to the quantum discord , which is known to be independent of @xmath1 ( for large @xmath1 ) .
this suggests that the geometric quantum discord fails to completely quantify the correlations defined in the original entropic measure of quantum discord .
finally , the work we have presented here further supports the suspicion that the apparent speedup exhibited by ( the dynamics of ) dqc1 is not necessarily captured by the geometric measure of quantum discord at the conclusion of the algorithm .
32ifxundefined [ 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/physrevlett.88.017901 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop ( ) @noop * * , ( ) @noop ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.81.5672 [ * * , ( ) ] link:\doibase 10.1103/physreva.72.042316 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.050502 [ * * , ( ) ] , link:\doibase 10.1103/physrevlett.101.200501 [ * * , ( ) ] , @noop * * , ( ) @noop * * , ( ) @noop `` , '' ( ) , @noop `` , '' ( ) , @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1016/j.jmr.2004.11.004 [ * * , ( ) ] @noop * * , ( ) | we describe an efficient dqc1-algorithm to quantify the amount of geometric quantum discord present in the output state of a dqc1 computation .
dqc1 is a model of computation that utilizes separable states to solve a problem with no known efficient classical algorithm and is known to contain quantum correlations as measured by the discord . for the general case of a @xmath0-qubit dqc1-state we provide an analytical expression for the geometric quantum discord and find that its typical ( and maximum ) value decreases exponentially with @xmath1 .
this is in contrast to the standard quantum discord whose value for typical dqc1-states is known to be independent of @xmath1 .
we experimentally demonstrate the proposed algorithm on a four - qubit liquid - state nuclear magnetic resonance quantum information processor . in the special case of a two - qubit dqc1 model
, we also provide an expression for the quantum discord that only requires the outcome of the dqc1 algorithm . since its inception in 2001 , quantum discord ( qd ) @xcite and related measures
have been used to quantify the amount of non - classical correlations in a physical system .
in addition to quantifying the most well known quantum correlations of entanglement , they also capture the quantum correlations that exist in separable states .
discord measures are based on the premise that if a measurement on one part of a bipartite state disturbs the total state , there must be correlations stronger than what is found in the classical world .
quantum discord has been studied extensively over the past few years , from operational definitions @xcite , to witnesses @xcite and analytical expressions for specific sets of states @xcite .
however , it is generally difficult to calculate as it requires both full state knowledge and an optimization over all projective measurements of a subsystem . even in the simple case of two qubits ,
a general closed form expression does not exist .
the measure of geometric quantum discord ( gqd ) was introduced @xcite as a simple geometric measure of the distance from a given state to the closest classical ( zero discord ) state .
dqc1 is a model of mixed state quantum computation @xcite that contains limited entanglement @xcite yet is thought to outperform classical methods .
it does generate non - classical correlations as measured by the quantum discord @xcite , which , in hind sight , is not surprising since almost all quantum states have non - zero discord @xcite .
quantum discord has indeed been witnessed in the final states of dqc1 computations @xcite , even at very small values in highly mixed quantum states @xcite . to further explore the role of quantum correlations in the dqc1 model and beyond , their quantification becomes a necessary pursuit . in this work
, we derive an analytical expression for the geometric discord in the final state of a dqc1 computation of arbitrary dimension , in terms of quantities that can be efficiently estimated on a dqc1 computer .
we then demonstrate its experimental evaluation in a four - qubit implementation of a dqc1 algorithm .
additionally , we derive a simple expression for the quantum discord in the special case of a two - qubit dqc1-state . |
starting with the synthesis of graphene@xcite a new era of two dimensional ( 2d ) materials have emerged . among them
we find hexagonal boron nitride , molybdenum disulphide , dichalcogenides , layered oxides@xcite , silicene and more recently one layer of black phosphorous ( phosphorene)@xcite . these materials , in general , have diverse properties , but some of them are semiconductors , a dream of stacking them in a certain order has become a reality resulting in structures with desired properties@xcite .
it has been possible not only to stack them in certain order but also rotate layers as well to exploit certain properties that twisted structures posses@xcite .
for instance in twisted bilayer graphene a plethora of interesting angle dependent properties have been found @xcite .
single layer phosphorene is a direct band gap semiconductor ( 2 @xmath0 ) .
the addition of more layers reduce the value of the gap , allowing in principle to control its size which goes inverse with the layer number down to the value of the bulk ( 0.3 @xmath0)@xcite .
nonetheless the main optical absorption peak of monolayer phosphorene is in 1.2 @xmath0 due to the presence of excitons@xcite . on the other hand
it has been shown in ref.[@xcite ] that a displacement of some atoms in crystal phosphorene results in a honeycomb structure like graphene .
the authors studied its stability by first principle methods and found that it can be as stable as phosphorene ; this new 2d allotrope , named blue phosphorene , has an unit cell composed of two sublattices , like in honeycomb lattice but with the difference that the two atoms are not in the same horizontal plane , forming a periodically buckled topology .
the blue phosphorene is a semiconductor with an indirect band gap of @xmath1 2 @xmath0 . along this line
a theoretical work based on first principles calculation proposes the use of blue phosphorene nanoribbons@xcite to modulate the band gap , in this case the gap decreases for narrower nanoribbons .
nonetheless the system remains an indirect band gap semiconductor for either armchair or zigzag edges . another way to modulate the physical properties of
blue phosphorene is changing the stacking .
two blue phosphorene layers stacked in aa or ab fashion have also shown a decrease in the band gap compared with the blue phosphorene monolayer .
unfortunately the band gap is also indirect@xcite which is a real limitation to use this 2d allotrope in the solar cell industry due to the poor efficiency that characterizes indirect band gap semiconductors@xcite . in this article
we show that twisted bilayer blue phosphorene(tbbp ) is a direct band gap semiconductor at least in three of the four structures studied .
the relative rotation angle ( rra ) between layers modulates the value of the gap .
the band gap goes from 0.95 @xmath0 for a rra of 7.3@xmath2 to a value of 1.6 @xmath0 for an angle of 21.8@xmath2 . unlike stacked aa and ab bilayer blue phosphorene ,
tbbp has absorption peaks in the visible range of the absorption spectra , and additionally the presence of flat valence bands in tbbp structures could improve the optical absorption probability .
these findings makes tbbp a good candidate for a solar cell device , since an ideal solar cell device would be composed of a battery of stacked direct band gap semiconductors , each of them with a different value of the band gap and hopefully around 1.5 @xmath0 to take advantage of the solar spectrum@xcite .
( a ) two stacked ab blue phosphorene layers .
( b ) lateral view of the two layers in ab stacking .
( c ) twisted structure for a relative rotating angle of @xmath3 . ]
crystalline blue phosphorene is composed of an hexagonal lattice with a basis of two p atoms like the honeycomb structure of graphene .
the results of relaxation processes determined by _ ab initio _ methods shows that the two atoms are not in the same plane .
in fact , there is a vertical shift between them of 1.24 @xcite . on the other hand ,
one of the ways of stacking two layers is in the bernal or ab stacking . in bilayer
graphene it is the most stable configuration .
the unit cell of stacked ab blue phosphorene will have four atoms , two in each layer .
one of the atoms at the top layer will have directly below it an atom of the bottom layer whereas the second atom of the upper layer will be located exactly above the center of a hexagon of the layer beneath , as shown in fig [ xyz ] ( a ) .
we consider a system composed of two coupled blue phosphorene layers when there is a rra between them .
the unit cell of the rotated structure was constructed following a procedure exposed in several previous works@xcite .
starting from a stacked ab bilayer one rotates one layer from a point in the lattice to an equivalent location .
this can be done by taking a vector @xmath4 and rotating it around a line perpendicular to the layers passing through the origin to an equivalent position at @xmath5 , where @xmath6 and @xmath7 are the monolayer lattice vectors ; _
n_,_m _ are integers . this way the rotational angle and all the relevant geometric information of the bilayer twisted cell can be expressed in terms of @xmath8 and @xmath9 . in figure [ xyz](a
) we show the stacked ab blue phosphorene , in panel ( b ) a lateral view of the two layers where the buckled structure can be observed , and in panel ( c ) the twisted unit cell obtained doing a commensurate rotation .
in this case we selected a hollow site of the stacked ab bilayer as a center of rotation , which is also the origin of the twisted unit cell .
the hollow site is the point where an atom of the bottom layer have directly above a center of an hexagon of the top layer .
the calculations were performed within the density functional theory ( dft ) approach , as implemented in the siesta code@xcite .
the kohn - sham orbitals were expanded in a localized orbitals basis set ( double-@xmath10 , single polarized in the present work ) .
the twisted blue phosphorene bilayers were simulated considering cells with @xmath11 ( tbbp-28 ) , @xmath12 ( tbbp-76 ) , @xmath13 ( tbbp-148 ) and @xmath14 ( tbbp-244 ) atoms , corresponding to rra ( @xmath15 ) of @xmath16 , @xmath17 , @xmath18 , and @xmath19 respectively , and with a vacuum region of @xmath20 .
for the exchange - correlation potential we employed vdw - kbm approach@xcite , that includes the van der waals ( vdw ) interactions within the self - consistent process .
all structures were fully relaxed considering a force convergence tolerance of @xmath21 ev / per atom .
the brillouin zone ( bz ) was sampled by a @xmath22 monkhorst - pack mesh that ensure the total energy convergence .
the optical absorption spectrum is obtained through the imaginary part of the dielectric function , implemented in the siesta code using the dipolar approximation .
this approximation does not take into account the many body effects as gw corrections and the electron - hole interaction .
these calculations are prohibitive for such a big systems as tbbps . so that , the dipolar approach is certainly a good first order approximation to understand the physical properties of tbbps , in particular the optical absorption spectrum .
all optical spectra were calculated with a gaussian smearing of 0.06 @xmath0 . to ensure the convergence of the optical spectrum and the total density of states we employed a k - point grids of @xmath23 for aa and ab stacking , @xmath24 for rras of @xmath16 and @xmath17 , and @xmath25 for rras of @xmath18 and @xmath19 respectively .
we have calculated the geometrical structure , the band structure , the total density of states , and the optical absorption of monolayer bp , bilayers stacked aa and ab of bp and tbbp with different rra .
the main geometric parameters and the band gap for each system are shown in the table [ tabla ] . for the bp monolayer the results obtained by us for the lattice constant , and the band gap
are consistent with a previous siesta calculations@xcite . in the case of bp
stacked aa and ab the results are in agreement with a previous calculation@xcite , where pbe and dft - d2 corrections were employed to incorporate a vdw interaction . there , the authors obtained a distance of @xmath26 and @xmath27 for aa and ab stacking , respectively .
our results for the binding energy of all structures show that the aa staking is more favourable than the ab and tbbp with a binding energy per atom of @xmath28 ev . in the case of tbbp
, the rra does not make any perceptible variation in the binding energy per atom , the value obtained for all structures were @xmath29 ev / atom .
this results suggest that the tbbps are more stable than the ab staking regardless of the rra .
.geometric parameters , band gap , and binding energy of several blue phosphorene configurations .
@xmath30 - lattice constant , @xmath31 - maximal distance between layers and @xmath32 - binding energy by atom . [ cols="^,^,^,^,^,^",options="header " , ] in figure [ f1 ] we show the results of the dft calculations of the band structure for monolayer and bilayer staked aa and ab of blue phosphorene , the red arrow indicates the indirect band gap .
the three structures are indirect band gap semiconductors , the size of the gap varies from @xmath1 0.6 to 2 @xmath0 .
the density of states ( dos ) of the same three structures ( figure [ f2](a ) ) shows clearly the evolution of the size of the band gap .
the top valence band in the two bilayer systems is almost flat around the @xmath33 point of the bz , and the vhs below the fermi level can be associated with this valence band . a strong absorption of a material at a given energy is usually associated with the presence of vhs in the dos .
the absorption spectra in figure [ f2](b ) ) show strong absorptions peaks for energies around @xmath1 4 @xmath0 , but below this energy and above the absorption edges there is not any strong absorption peak whatsoever .
this fact is certainly similar to what is observed in other hexagonal flat structures as graphene which is transparent in the visible spectrum range . we were interested to see what happens when we rotate the layers . according to our previous experience
, many fascinating angle dependent properties may appear .
we have calculated the band structure of four twisted structure by dft , and the results are shown in the figure [ f3 ] .
it is possible to observe that the rotation angle changes the electronic properties of the system .
in fact , for certain values of the relative rotation angle , the bilayer blue phosphorene becomes a direct band gap semiconductor .
at least in three of the four structures the minimum distance between top valence band an lower conduction band lies at the same k - point .
however the top valence band is almost flat throughout the whole bz and , as long as the angle is smaller this effect is reinforced .
one might explain such a behaviour as a result of the shrinking of the bz .
the flat region around the @xmath33 point observed in stacked ab bilayer have become relatively larger in the reduced zone of the twisted structures . as a result of this band flatness
, optical absorption peaks will appear near the band gap value with amplitudes larger than those appearing in the stacked ab and aa structures .
= 21.8@xmath2 , ( b ) @xmath15=13.2@xmath2 , ( c ) @xmath15=9.4@xmath2 and ( d ) @xmath15=7.3@xmath2 .
the ( red vertical ) arrows show the band gap . ]
the value of the band gap changes with the angle .
likely , more results are needed in order to establish the full variation pattern but in accordance with the four cases investigated here it is possible to advance the decrease of the band gap as a function of the rotation angle . in table
[ tabla ] the values of the band gap for the angles studied are listed .
it has a value of 1.33 @xmath0 for the larger angle , 21.8@xmath2 and goes to 0.95 @xmath0 for a rra of 7.3@xmath2 .
the distance between layers in the relaxed structures diminishes slightly as the angle gets lower .
the latter is expected since our starting point was a stacked ab bilayer .
= 21.8@xmath2 , ( b ) @xmath15=13.2@xmath2 , ( c ) @xmath15=9.4@xmath2 and ( d ) @xmath15=7.3@xmath2 .
panel ( e ) is for the imaginary part of the dielectric of four twisted bilayer blue phosphorene structures , the insert in the panel ( e ) shows the imaginary part of the dielectric function in different incident photon energy range . ]
the dos of the twisted structures , figure [ f4](a ) , express in another format the facts we were mentioning ; that is , the slight dependence of the band gap with the angle and the presence of vhs just in the band edge of the valence bands .
the absorption spectra in figure [ f4](b ) reveal the presence of absorption peaks below 2 @xmath0 , the position of the peaks changes with the angle and follows in principle the same behaviour of the band gap .
the intensity of the peaks diminishes also for lower angles .
its is important to notice that we have not included many - electron effects in our calculations .
the presence of these flat bands which is in many cases related with the appearance of many body effects as well as the difference of the experimental optical band gap with the theoretical ( single electron calculations ) in its cousin bilayer black phosphorene lead us to suggest that further studies of these structure should be made considering many body effects .
however the task is complicated due to the large size of the unit cell .
we have studied by dft calculations the electronic structure of two coupled layers of blue phosphorene and found that a relative rotation between layers changes the electronic properties of the system .
bilayer stacked aa or ab blue phosphorene are indirect band gap semiconductor while some of the twisted structures are direct band gap semiconductor and moreover the size of the band gap depends on the rra .
the absorption spectra show peaks in the visible range of the spectra which makes twisted bilayer blue phosphorene another candidate for solar cell applications .
the dependence of the band gap with the angle suggests that stacking several of these structures together will increase the efficiency in solar cell devices as it will cover a wider range in energy of the solar spectrum as well as a larger interface .
the authors acknowledge financial support from the universidad de medelln research office through projects 746 and 747 and chilean fondecyt grant 11130129 .
dao and cad are grateful to the colombian agencies codi- universidad de antioquia ( estrategia de sostenibilidad de la universidad de antioquia and project : `` manipulacin de propiedades optoelectrnicas de nanoestructuras semiconductoras por ondas acsticas superficiales . '' ) , facultad de ciencias exactas y naturales - universidad de antioquia ( cad - exclusive dedication project 2015 - 2016 ) , and el patrimonio autnomo fondo nacional de financiamiento para la ciencia , la tecnologa y la innovacin , francisco jos de caldas .
10 url # 1`#1`urlprefixhref # 1#2#2 # 1#1 k. s. novoselov , d. jiang , f. schedin , t. j. booth , v. v. khotkevich , s. v. morozov , a. k. geim , two - dimensional atomic crystals , proceedings of the national academy of sciences of the united states of america 102 ( 30 ) ( 2005 ) 1045110453 .
k. kim , m. yankowitz , b. fallahazad , s. kang , h. c. p. movva , s. huang , s. larentis , c. m. corbet , t. taniguchi , k. watanabe , s. k. banerjee , b. j. leroy , e. tutuc , van der waals heterostructures with high accuracy rotational alignment , nano letters 16 ( 3 ) ( 2016 ) 19891995 .
i. brihuega , p. mallet , h. gonzlez - herrero , g. trambly de laissardire , m. m. ugeda , l. magaud , j. m. gmez - rodrguez , f. yndurin , j .- y .
veuillen , unraveling the intrinsic and robust nature of van hove singularities in twisted bilayer graphene by scanning tunneling microscopy and theoretical analysis , phys .
( 2012 ) 196802 .
w. yan , m. liu , r .- f .
dou , l. meng , l. feng , z .- d .
chu , y. zhang , z. liu , j .- c .
nie , l. he , angle - dependent van hove singularities in a slightly twisted graphene bilayer , phys .
( 2012 ) 126801 .
v. tran , r. soklaski , y. liang , l. yang , layer - controlled band gap and anisotropic excitons in few - layer black phosphorus , physical review b - condensed matter and materials physics 89 ( 23 ) ( 2014 ) 235319 .
j. xie , m. s. si , d. z. yang , z. y. zhang , d. s. xue , a theoretical study of blue phosphorene nanoribbons based on first - principles calculations , journal of applied physics 116 ( 7 ) ( 2014 ) 073704 .
http://arxiv.org/abs/1405.4407 [ ] .
t. zdanowicz , t. rodziewicz , m. zabkowska - waclawek , theoretical analysis of the optimum energy band gap of semiconductors for fabrication of solar cells for applications in higher latitudes locations , solar energy materials and solar cells 87 ( 14 ) ( 2005 ) 757 769 , international conference on physics , chemistry and engineering .
j. dai , x. c. zeng , bilayer phosphorene : effect of stacking order on bandgap and its potential applications in thin - film solar cells , the journal of physical chemistry letters 5 ( 7 ) ( 2014 ) 12891293 , pmid : 26274486 . j. m. soler , e. artacho , j. d. gale , a. garca , j. junquera , p. ordejn , d. snchez - portal , the siesta method for ab initio order- n materials simulation , journal of physics : condensed matter 14 ( 11 ) ( 2002 )
27452779 . http://arxiv.org/abs/0111138 [ ] . | we report that two rotated layers of blue phosphorene behave as a direct band gap semiconductor .
the optical spectrum shows absorption peaks in the visible region of the spectrum and in addition the energy of these peaks can be tuned with the rotational angle .
these findings makes twisted bilayer blue phosphorene a strong candidate as a solar cell or photodetection device .
our results are based on ab initio calculations of several rotated blue phosphorene layers .
blue phosphorene , bilayer , rotated , ab initio |
description logics ( dls ) @xcite are a family of logic - based knowledge representation formalisms , which can be used to represent the conceptual knowledge of an application domain in a structured and formally well - understood way .
they are employed in various application domains , such as natural language processing , configuration of technical systems , databases , and biomedical ontologies , but their most notable success so far is the adoption of the dl - based language owl @xcite as standard ontology language for the semantic web . in dls
, concepts are formally described by _
concept terms _ , i.e. , expressions that are built from concept names ( unary predicates ) and role names ( binary predicates ) using concept constructors .
the expressivity of a particular dl is determined by which concept constructors are available in it . from a semantic point of view
, concept names and concept terms represent sets of individuals , whereas roles represent binary relations between individuals .
for example , using the concept name @xmath1 , and the role name @xmath2 , the concept of _ women having a daughter _ can be represented by the concept term @xmath3 and the concept of _ women having only daughters _ by @xmath4 knowledge representation systems based on dls provide their users with various inference services that allow them to deduce implicit knowledge from the explicitly represented knowledge .
an important inference problem solved by dl systems is the subsumption problem : the subsumption algorithm allows one to determine subconcept - superconcept relationships .
for example , the concept term @xmath1 subsumes the concept term @xmath5 since all instances of the second term are also instances of the first term , i.e. , the second term is always interpreted as a subset of the first term . with the help of the subsumption algorithm ,
a newly introduced concept term can automatically be placed at the correct position in the hierarchy of the already existing concept terms .
two concept terms @xmath6 are _ equivalent _ ( @xmath7 ) if they subsume each other , i.e. , if they always represent the same set of individuals .
for example , the terms @xmath8 and @xmath9 are equivalent since the value restriction operator ( @xmath10 ) distributes over the conjunction operator ( @xmath11 ) .
if we replace the value restriction operator by the existential restriction operator ( @xmath12 ) , then this equivalence no longer holds .
however , for this operator , we still have the equivalence @xmath13 the equivalence test can , for example , be used to find out whether a concept term representing a particular notion has already been introduced , thus avoiding multiple introduction of the same concept into the concept hierarchy .
this inference capability is very important if the knowledge base containing the concept terms is very large , evolves during a long time period , and is extended and maintained by several knowledge engineers .
however , testing for equivalence of concepts is not always sufficient to find out whether , for a given concept term , there already exists another concept term in the knowledge base describing the same notion . on the one hand
, different knowledge engineers may use different names for concepts , like @xmath14 versus @xmath15 . on the other hand , they may model on different levels of granularity .
for example , assume that one knowledge engineer has defined the concept of _ men loving fast cars _ by the concept term @xmath16 a second knowledge engineer might represent this notion in a somewhat different way , e.g. , by using the concept term @xmath17 these two concept terms are not equivalent , but they are meant to represent the same concept .
the two terms can obviously be made equivalent by substituting the concept name @xmath18 in the first term by the concept term @xmath19 and the concept name @xmath20 in the second term by the concept term @xmath21 .
this leads us to _ unification of concept terms _ , i.e. , the question whether two concept terms can be made equivalent by applying an appropriate substitution , where a substitution replaces ( some of the ) concept names by concept terms .
of course , it is not necessarily the case that unifiable concept terms are meant to represent the same notion .
a unifiability test can , however , suggest to the knowledge engineer possible candidate terms .
a _ unifier _
( i.e. , a substitution whose application makes the two terms equivalent ) then proposes appropriate definitions for the concept names . in our example , we know that , if we define @xmath20 as @xmath21 and @xmath18 as @xmath19 , then the concept terms @xmath22 and @xmath23 are equivalent w.r.t .
these definitions .
unification in dls was first considered in @xcite for a dl called @xmath24 , which has the concept constructors _ conjunction _ ( @xmath11 ) , _ value restriction _ ( @xmath10 ) , and the _ top concept _ ( @xmath25 ) .
it was shown that unification in @xmath24 is decidable and exptime - complete , i.e. , given an @xmath24-unification problem , we can effectively decide whether it has a solution or not , but in the worst - case , any such decision procedure needs exponential time .
this result was extended in @xcite to a more expressive dl , which additionally has the role constructor _
transitive closure_. interestingly , the _ unification type _ of @xmath24 had been determined almost a decade earlier in @xcite . in fact , as shown in @xcite , unification in @xmath24 corresponds to unification modulo the equational theory of idempotent abelian monoids with several homomorphisms . in @xcite
it was shown that , already for a single homomorphism , unification modulo this theory has unification type zero , i.e. , there are unification problems for this theory that do not have a minimal complete set of unifiers .
in particular , such unification problems can not have a finite complete set of unifiers . in this paper , we consider unification in the dl @xmath26 .
the @xmath0-family consists of inexpressive dls whose main distinguishing feature is that they provide their users with _ existential restrictions _ ( @xmath12 ) rather than value restrictions ( @xmath10 ) as the main concept constructor involving roles .
the core language of this family is @xmath0 , which has the top concept , conjunction , and existential restrictions as concept constructors .
this family has recently drawn considerable attention since , on the one hand , the subsumption problem stays tractable ( i.e. , decidable in polynomial time ) in situations where @xmath24 , the corresponding dl with value restrictions , becomes intractable : subsumption between concept terms is tractable for both @xmath24 and @xmath0 @xcite , but allowing the use of concept definitions or even more expressive terminological formalisms makes @xmath24 intractable @xcite , whereas it leaves @xmath0 tractable @xcite . on the other hand , although of limited expressive power , @xmath0 is nevertheless used in applications , e.g. , to define biomedical ontologies .
for example , both the large medical ontology snomed ct and the gene ontology can be expressed in @xmath0 , and the same is true for large parts of the medical ontology galen @xcite .
the importance of @xmath0 can also be seen from the fact that the new owl2 standard contains a sub - profile owl2el , which is based on ( an extension of ) @xmath0 .
unification in @xmath0 has , to the best of our knowledge , not been investigated before , but matching ( where one side of the equation(s ) to be solved does not contain variables ) has been considered in @xcite . in particular ,
it was shown in @xcite that the decision problem , i.e. , the problem of deciding whether a given @xmath0-matching problem has a matcher or not , is np - complete .
interestingly , @xmath24 behaves better w.r.t .
matching than @xmath0 : for @xmath24 , the decision problem is tractable @xcite . in this paper , we show that , w.r.t .
the unification type , @xmath24 and @xmath0 behave the same : just as @xmath24 , the dl @xmath0 has unification type zero .
however , w.r.t .
the decision problem , @xmath0 behaves much better than @xmath24 : @xmath0-unification is np - complete , and thus has the same complexity as @xmath0-matching .
regarding unification in dls that are more expressive than @xmath0 and @xmath24 , one must look at the literature on unification in modal logics .
it is well - known that there is a close connection between modal logics and dls @xcite .
for example , the dl @xmath27 , which can be obtained by adding negation to @xmath0 or @xmath28 , corresponds to the basic ( multi-)modal logic @xmath29 .
decidability of unification in @xmath29 is a long - standing open problem .
recently , undecidability of unification in some extensions of @xmath29 ( for example , by the universal modality ) was shown in @xcite .
the undecidability results in @xcite also imply undecidability of unification in some expressive dls ( e.g. , @xmath30 @xcite ) .
the unification types of some modal ( and related ) logics have been determined by ghilardi ; for example in @xcite he shows that @xmath31 and @xmath32 have unification type finitary .
unification in sub - boolean modal logics ( i.e. , modal logics that are not closed under all boolean operations , such as the modal logics corresponding to @xmath0 and @xmath28 ) has , to the best of our knowledge , not been considered in the modal logic literature .
in addition to unification of concept terms as introduced until now , we will also consider unification w.r.t .
a so - called acyclic tbox in this article .
until now , we have only talked about concept terms , i.e. , complex descriptions of concepts that are built from concept and role names using the concept constructors of the given dl . in applications of dls , it is , of course , inconvenient to always use such complex descriptions when referring to concepts . for this reason ,
dls are usually also equipped with a terminological formalism . in its simplest form , this formalism allows to introduce abbreviations for concept terms .
for example , the two concept definitions @xmath33 introduce the abbreviation @xmath1 for the concept term @xmath34 and the abbreviation @xmath35 for the concept term @xmath36 . a finite set of such concept definitions is called an _ acyclic _
tbox if it is unambiguous ( i.e. , every concept name occurs at most once as left - hand side ) and acyclic ( i.e. , there are no cyclic dependencies between concept definitions ) .
these restrictions ensure that every defined concept ( i.e. , concept name occurring on the left - hand side of a definition ) has a unique expansion to a concept term that it abbreviates . inference problems like subsumption and unification can also be considered w.r.t . such acyclic tboxes .
as mentioned above , the complexity of the subsumption problem increases for the dl @xmath24 if acyclic tboxes are taken into account @xcite .
in contrast , for @xmath0 , the complexity of the subsumption problem stays polynomial in the presence of acyclic tboxes .
we show that , for unification in @xmath0 , adding acyclic tboxes is also harmless , i.e. , unification in @xmath0 _ w.r.t .
acyclic tboxes _ is also np - complete .
this article is structured as follows . in the next section ,
we define the dl @xmath0 and unification in @xmath0 more formally . in section [ subs ] , we recall the characterization of subsumption and equivalence in @xmath0 from @xcite , and in section [ type : zero ] we use this to show that unification in @xmath0 has type zero . in section [ decision : problem ] , we show that unification in @xmath0 is np - complete .
the unification algorithm establishing the complexity upper bound is a typical `` guess and then test '' np - algorithm , and thus it is unlikely that a direct implementation of this algorithm will perform well in practice . in section [ goal : oriented : sect ] , we introduce a more goal - oriented unification algorithm for @xmath0 , in which non - deterministic decisions are only made if they are triggered by `` unsolved parts '' of the unification problem . in section [ eq : th : sect ] , we point out that our results for @xmath0-unification imply that unification modulo the equational theory of semilattices with monotone operators @xcite is np - complete and of unification type zero . more information about description logics can be found in @xcite , and about unification theory in @xcite .
this article is an extended version of a paper @xcite published in the proceedings of the 20th international conference on rewriting techniques and applications ( rta09 ) .
in addition to giving more detailed proofs , we have added the goal - oriented unification algorithm ( section [ goal : oriented : sect ] ) and the treatment of unification modulo acyclic tboxes ( subsection [ unif : mod : tbox ] ) .
in this section , we first define the syntax and semantics of @xmath0-concept terms as well as the subsumption and the equivalence relation on these terms . then
, we introduce unification of @xmath0-concept terms , and finally extend this notion to unification modulo an acyclic tbox . starting with a set @xmath37 of concept names and a set @xmath38 of role names , _
_ are built using the following concept constructors : the nullary constructor _ top - concept _
( @xmath25 ) , the binary constructor _ conjunction _ ( @xmath39 ) , and for every role name @xmath40 , the unary constructor _ existential restriction _ ( @xmath12 ) .
the semantics of @xmath0 is defined in the usual way , using the notion of an interpretation @xmath41 , which consists of a nonempty domain @xmath42 and an interpretation function @xmath43 that assigns binary relations on @xmath42 to role names and subsets of @xmath42 to concept terms , as shown in the semantics column of table [ el - syn - sem ] . ' '' '' concept name & @xmath44 & @xmath45 + ' '' '' role name & @xmath46 & @xmath47 + ' '' '' top - concept & @xmath25 & @xmath48 + ' '' '' conjunction & @xmath49 & @xmath50 + ' '' '' existential restriction & @xmath12 & @xmath51 + ' '' '' subsumption & @xmath52 & @xmath53 + ' '' '' equivalence & @xmath54 & @xmath55 + the concept term @xmath56 _ is subsumed by _ the concept term @xmath57 ( written @xmath58 ) iff @xmath53 holds for all interpretations @xmath59 .
we say that @xmath56 _ is equivalent to _ @xmath57 ( written @xmath7 ) iff @xmath58 and @xmath60 , i.e. , iff @xmath55 holds for all interpretations @xmath59 .
the concept term @xmath56 _ is strictly subsumed by _ the concept term @xmath57 ( written @xmath61 ) iff @xmath58 and @xmath62 .
it is well - known that subsumption ( and thus also equivalence ) of @xmath26-concept terms can be decided in polynomial time @xcite . in order to define unification of concept terms ,
we first introduce the notion of a substitution operating on concept terms . to this purpose
, we partition the set of concepts names into a set @xmath63 of concept variables ( which may be replaced by substitutions ) and a set @xmath64 of concept constants ( which must not be replaced by substitutions ) .
intuitively , @xmath63 are the concept names that have possibly been given another name or been specified in more detail in another concept term describing the same notion .
the elements of @xmath64 are the ones of which it is assumed that the same name is used by all knowledge engineers ( e.g. , standardized names in a certain domain ) . a _ substitution _
@xmath65 is a mapping from @xmath63 into the set of all @xmath26-concept terms .
this mapping is extended to concept terms in the obvious way , i.e. , 1 .
@xmath66 for all @xmath67 , 2 .
@xmath68 , 3 .
@xmath69 , and 4 .
@xmath70 .
an _ @xmath26-unification problem _ is of the form @xmath71 , where @xmath72 are @xmath26-concept terms .
the substitution @xmath65 is a _ unifier _ ( or _ solution _ ) of @xmath73 iff @xmath74 for @xmath75 . in this case
, @xmath73 is called _ solvable _ or _ unifiable_. when we say that @xmath26-unification is _ decidable _ , then we mean that the following decision problem is decidable : given an @xmath26-unification problem @xmath73 , decide whether @xmath73 is solvable or not . accordingly
, we say that @xmath26-unification is _ np - complete _ if this decision problem is np - complete . in the following , we introduce some standard notions from unification theory @xcite , but formulated for the special case of @xmath26-unification rather than for an arbitrary equational theory .
unifiers can be compared using the instantiation preorder @xmath76 .
let @xmath73 be an @xmath26-unification problem , @xmath77 the set of variables occurring in @xmath73 , and @xmath78 two unifiers of this problem .
we define @xmath79 if @xmath80 , then we say that @xmath81 is an _ instance _ of @xmath65 .
let @xmath73 be an @xmath26-unification problem .
the set of substitutions @xmath82 is called a _
complete set of unifiers _ for @xmath73 iff it satisfies the following two properties : 1 .
every element of @xmath82 is a unifier of @xmath73 ; 2 .
if @xmath81 is a unifier of @xmath73 , then there exists a unifier @xmath83 such that @xmath80 .
the set @xmath82 is called a _ minimal complete set of unifiers _ for @xmath73 iff it additionally satisfies 1 .
if @xmath84 , then @xmath80 implies @xmath85 .
the unification type of a given unification problem is determined by the existence and cardinality of such a minimal complete set .
let @xmath73 be an @xmath26-unification problem .
this problem has type 1 .
_ unitary _ iff it has a minimal complete set of unifiers of cardinality @xmath86 ; 2 .
_ finitary _ iff it has a finite minimal complete set of unifiers ; 3 .
_ infinitary _ iff it has an infinite minimal complete set of unifiers ; 4 .
_ zero _ iff it does not have a minimal complete set of unifiers .
note that the set of all unifiers of a given @xmath26-unification problem is always a complete set of unifiers .
however , this set is usually infinite and redundant ( in the sense that some unifiers are instances of others ) . for a unitary or finitary @xmath26-unification problem ,
all unifiers can be represented by a finite complete set of unifiers , whereas for problems of type infinitary or zero this is no longer possible .
in fact , if a problem has a finite complete set of unifiers @xmath82 , then it also has a finite _ minimal _ complete set of unifiers , which can be obtained by iteratively removing redundant elements from @xmath82 .
for an infinite complete set of unifiers , this approach of removing redundant unifiers may be infinite , and the set reached in the limit need no longer be complete .
this is what happens for problems of type zero .
the difference between infinitary and type zero is that a unification problem of type zero can not even have a non - redundant complete set of unifiers , i.e. , every complete set of unifiers must contain different unifiers @xmath78 such that @xmath80 .
more information on unification type zero can be found in @xcite .
when we say that _ @xmath26 has unification type zero _ , we mean that there exists an @xmath26-unification problem that has type zero . before we can prove in section [ type : zero ] that this is indeed the case , we must have a closer look at equivalence in @xmath26 in section [ subs ] . but first , we consider unification modulo acyclic tboxes .
a _ concept definition _ is of the form @xmath87 where @xmath44 is a concept name and @xmath56 is a concept term . a
@xmath88 is a finite set of concept definitions such that no concept name occurs more than once on the left - hand side of a concept definition in @xmath88 .
the tbox @xmath88 is called _ acyclic _ if there are no cyclic dependencies between its concept definitions . to be more precise
, we say that the concept name @xmath44 _ directly depends on _ the concept name @xmath89 in a tbox @xmath88 if @xmath88 contains a concept definition @xmath87 and @xmath89 occurs in @xmath56 .
let _ depends on _ be the transitive closure of the relation _ directly depends on_. then @xmath88 contains a _ terminological cycle _ if there is a concept name @xmath44 that depends on itself .
otherwise , @xmath88 is called _
acyclic_. given a tbox @xmath88 , we call a concept name @xmath44 a _ defined concept _ if it occurs as the left - side of a concept definition @xmath87 in @xmath88 .
all other concept names are called _ primitive concepts_. the interpretation @xmath90 is a model of the tbox @xmath91 iff @xmath92 holds for all concept definitions @xmath87 in @xmath91 .
subsumption and equivalence w.r.t .
a tbox are defined as follows : @xmath93 ( @xmath94 ) iff @xmath53 ( @xmath55 ) holds for all models @xmath59 of @xmath95 .
subsumption and equivalence w.r.t .
an acyclic tbox can be reduced to subsumption and equivalence of concept terms ( without tbox ) by _ expanding _ the concept terms w.r.t .
the tbox : given a concept term @xmath56 , its expansion @xmath96 w.r.t .
the acyclic tbox @xmath88 is obtained by exhaustively replacing all defined concept names @xmath44 occurring on the left - hand side of concept definitions @xmath87 in @xmath88 by their defining concept terms @xmath56 . given concept terms @xmath6 , we have @xmath97 iff @xmath98 @xcite .
the same is true for equivalence , i.e. , @xmath99 iff @xmath100 .
this expansion process may , however , result in an exponential blow - up @xcite , and thus this reduction of subsumption and equivalence w.r.t .
an acyclic tbox to subsumption and equivalence without a tbox is not polynomial .
nevertheless , in @xmath26 , subsumption ( and thus also equivalence ) w.r.t .
acyclic tboxes can be decided in polynomial time @xcite . in our definition of unification
modulo acyclic tboxes , we assume that all defined concepts are concept constants .
in fact , defined concepts already have a definition in the given tbox , and thus it does not make sense to introduce new ones for them by unification . in this setting , a _ substitution _ @xmath65 is a mapping from @xmath63 into the set of all @xmath26-concept terms _ not containing any defined concepts_. the extension of @xmath65 to concept terms is defined as in the previous subsection , and its application to @xmath95 is defined as @xmath101 an _ @xmath26-unification problem modulo
an acyclic tbox _ is of the form @xmath102 , where @xmath72 are @xmath26-concept terms , and @xmath88 is an acyclic @xmath26-tbox .
the substitution @xmath65 is a _ unifier _ ( or _ solution _ ) of @xmath73 _ modulo @xmath88 _ iff @xmath103 for @xmath75 . in this case
, @xmath73 is called _ solvable modulo @xmath88 _ or _ unifiable modulo @xmath88_. coming back to our example from the introduction , assume that one knowledge engineer has written the concept definition @xmath104 to the tbox , whereas a second one has written the definition @xmath105 where all the concept names occurring on the left - hand side of these definitions are primitive concepts .
then the substitution that replaces @xmath18 by @xmath19 and @xmath20 by @xmath21 is a unifier of @xmath106 w.r.t .
the tbox @xmath88 consisting of these two definitions .
using expansion , we can reduce unification modulo an acyclic tbox to unification without a tbox .
in fact , the following lemma is an easy consequence of the fact that @xmath107 holds for all @xmath26-concept terms @xmath56 . [ expansion : lem ] the substitution @xmath65 is a unifier of @xmath108 modulo @xmath88 iff it is a unifier of @xmath109 .
since expansion can cause an exponential blow - up , this is not a polynomial reduction . in the remainder of this subsection
, we show that there actually exists a polynomial - time reduction of unification modulo an acyclic tbox to unification without a tbox .
we say that the @xmath0-unification problem @xmath73 is in _ dag - solved form _ if it can be written as @xmath110 , where @xmath111 are distinct concept variables such that , for all @xmath112 , @xmath113 does not occur in @xmath114 .
for @xmath115 , let @xmath116 be the substitution that maps @xmath113 to @xmath117 and leaves all other variables unchanged .
we define the substitution @xmath118 as @xmath119 for @xmath120 , and @xmath121 for all other variables @xmath122 .
the following is an instance of a well - known fact from unification theory @xcite .
[ jk : lemma ] let @xmath110 be an @xmath0-unification problem in dag - solved form .
then , the set @xmath123 is a complete set of unifiers for @xmath73 .
there is a close relationship between acyclic tboxes and unification problems in dag - solved form .
in fact , if @xmath88 is an acyclic tbox , then there is an enumeration @xmath124 of the defined concepts in @xmath88 such that @xmath125 and @xmath126 does not occur in @xmath127 .
consequently , the corresponding unification problem @xmath128 ( where @xmath129 are now viewed as concept variables ) is in dag - solved form .
in addition , it is easy to see that , for any @xmath0-concept term @xmath56 , we have @xmath130 .
the @xmath26-unification problem @xmath102 is solvable modulo the acyclic tbox @xmath88 iff @xmath131 is solvable .
are treated as concept constants in @xmath73 , and as concept variables in @xmath131 . ]
assume that @xmath81 is a unifier of @xmath102 modulo @xmath88 .
then it is a unifier of @xmath132 , by lemma [ expansion : lem ] .
since @xmath133 and @xmath134 , we have @xmath135 .
consequently , if we define the substitution @xmath136 by setting @xmath137 for all concept variables and defined concepts @xmath122 , then @xmath136 is a unifier of @xmath138 .
in addition , since @xmath139 is a unifier of @xmath140 , @xmath136 is also a unifier of @xmath131 .
conversely , assume that @xmath136 is a unifier of @xmath131 .
in particular , this implies that @xmath136 is a unifier of @xmath140 . by lemma [ jk : lemma ]
, @xmath141 is a complete set of unifiers for @xmath140 , and thus there is a substitution @xmath81 such that @xmath142 for all concept variables occurring in the unification problem @xmath131 . since @xmath133 and @xmath134
, this implies that @xmath81 is a unifier of @xmath109 , and thus of @xmath102 modulo @xmath88 , by lemma [ expansion : lem ] .
since the size of @xmath140 is basically the same as the size of @xmath88 , the size of @xmath143 is linear in the size of @xmath73 and @xmath88 .
thus , the above lemma provides us with a polynomial - time reduction of @xmath0-unification w.r.t .
acyclic tboxes to @xmath0-unification .
[ reduction : thm ] @xmath0-unification w.r.t .
acyclic tboxes can be reduced in polynomial time to @xmath0-unification .
in order to characterize equivalence of @xmath26-concept terms , the notion of a reduced @xmath26-concept term is introduced in @xcite .
a given @xmath26-concept term can be transformed into an equivalent reduced term by applying the following rules modulo associativity and commutativity of conjunction : @xmath144 a\sqcap a \rightarrow a & \mbox{for all concept names $ a\in n_{\mathit{con}}$}\\[.3em ] \exists\ , r.c\sqcap \exists\ , r.d \rightarrow \exists\ , r.c & \mbox{for all $ \mathcal { el}$-concept terms $ c , d$ with $ c\sqsubseteq d$ } \end{array}\ ] ] obviously , these rules are equivalence preserving .
we say that the @xmath26-concept term @xmath57 _ is reduced _ if none of the above rules is applicable to it ( modulo associativity and commutativity of @xmath11 ) , and that @xmath56 _ can be reduced to _
@xmath57 if @xmath57 can be obtained from @xmath56 by applying the above rules ( modulo associativity and commutativity of @xmath11 ) . the @xmath26-concept term @xmath57 is a _ reduced form _ of @xmath56 if @xmath56 can be reduced to @xmath57 and @xmath57 is reduced .
the following theorem is an easy consequence of theorem 6.3.1 on page 181 of @xcite .
[ cor1 ] let @xmath6 be @xmath26-concept terms , and @xmath145 reduced forms of @xmath6 , respectively . then @xmath7 iff @xmath146 is identical to @xmath147 up to associativity and commutativity of @xmath11 .
this theorem can also be used to derive a recursive characterization of subsumption in @xmath0 .
in fact , if @xmath58 , then @xmath148 , and thus @xmath56 and @xmath39 have the same reduced form .
thus , during reduction , all concept names and existential restrictions of @xmath57 must be `` eaten up '' by corresponding concept names and existential restrictions of @xmath56 .
[ cor2 ] let @xmath149 and @xmath150 , where @xmath151 are concept names .
then @xmath58 iff @xmath152 and for every @xmath153 , there exists an @xmath154 , such that @xmath155 and @xmath156 .
note that this corollary also covers the cases where some of the numbers @xmath157 are zero .
the empty conjunction should then be read as @xmath25 .
the following lemma , which is an immediate consequence of this corollary , will be used in our proof that @xmath26 has unification type zero .
[ lem1 ] if @xmath6 are reduced @xmath26-concept terms such that @xmath158 , then @xmath56 is either @xmath25 , or of the form @xmath159 where @xmath160 ; @xmath161 are reduced and pairwise incomparable w.r.t .
subsumption ; and @xmath162 .
conversely , if @xmath6 are @xmath26-concept terms such that @xmath159 and @xmath162 , then @xmath158 . the following lemma states several other obvious consequences of corollary [ cor2 ] . [ 1:new : lem ] 1 .
the existential restriction @xmath12 is reduced iff @xmath56 is reduced .
2 . let @xmath163 be concept names or existential restrictions .
then the conjunction @xmath163 is reduced iff @xmath164 are reduced and pairwise incomparable w.r.t .
3 . let @xmath165 and @xmath166 be conjunctions of @xmath0-concept terms .
if , for all @xmath167 , there exists @xmath168 , such that @xmath169 , then @xmath58 . if @xmath170 are concept names or existential restrictions , then the implication in the other direction also holds . in the proof of decidability of @xmath26-unification , we will make use of the fact that the inverse strict subsumption order is well - founded . [
well : founded1 ] there is no infinite sequence @xmath171 of @xmath26-concept terms such that @xmath172 .
we define the _ role depth _ of an @xmath26-concept term @xmath56 as the maximal nesting of existential restrictions in @xmath56 .
let @xmath173 be the role depth of @xmath174 .
since @xmath175 for @xmath176 , it is an easy consequence of corollary [ cor2 ] that the role depth of @xmath117 is bounded by @xmath173 , and that @xmath117 contains only concept and role names occurring in @xmath174 .
in addition , it is known that , for a given natural number @xmath173 and finite sets of concept names @xmath37 and role names @xmath38 , there are , up to equivalence , only finitely many @xmath26-concept terms built using concept names from @xmath177 and role names from @xmath178 and of a role depth bounded by @xmath173 @xcite .
consequently , there are indices @xmath179 such that @xmath180 .
this contradicts our assumption that @xmath181 .
to show that @xmath26 has unification type zero , we exhibit an @xmath26-unification problem that has this type . [
type : zero : thm ] let @xmath182 be variables .
the @xmath26-unification problem @xmath183 has unification type zero .
it is enough to show that any complete set of unifiers for this problem is redundant , i.e. , contains two different unifiers that are comparable w.r.t .
the instantiation preorder .
thus , let @xmath82 be a complete set of unifiers for @xmath73 .
first , note that @xmath82 must contain a unifier that maps @xmath122 to an @xmath26-concept term not equivalent to @xmath25 or @xmath184 .
in fact , consider a substitution @xmath136 such that @xmath185 and @xmath186 .
obviously , @xmath136 is a unifier of @xmath73 .
thus , @xmath82 must contain a unifier @xmath65 such that @xmath187 .
in particular , this means that there is a substitution @xmath188 such that @xmath189 .
obviously , @xmath190 would imply @xmath191 , and thus @xmath192 , which is , however , not the case .
similarly , @xmath193 would imply @xmath194 , and thus @xmath195 , which is also not the case .
thus , let @xmath83 be such that @xmath196 and @xmath197 .
without loss of generality , we assume that @xmath198 and @xmath199 are reduced .
since @xmath65 is a unifier of @xmath73 , we have @xmath158 .
consequently , lemma [ lem1 ] yields that @xmath56 is of the form @xmath159 where @xmath160 , @xmath161 are reduced and pairwise incomparable w.r.t .
subsumption , and @xmath162 .
we use @xmath65 to construct a new unifier @xmath200 as follows : @xmath201 where @xmath202 is a new variable ( i.e. , one not occurring in @xmath6 ) .
the second part of lemma [ lem1 ] implies that @xmath200 is indeed a unifier of @xmath73 .
next , we show that @xmath203 . to this purpose
, we consider the substitution @xmath188 that maps @xmath202 to @xmath204 , and does not change any of the other variables .
then we have @xmath205 and @xmath206 .
note that the second equivalence holds since we have @xmath207 .
since @xmath82 is complete , there exists a unifier @xmath208 such that @xmath209 .
transitivity of the relation @xmath76 thus yields @xmath210 .
since @xmath65 and @xmath81 both belong to @xmath82 , we have completed the proof of the theorem once we have shown that @xmath211 .
assume to the contrary that @xmath85 .
then we have @xmath212 , and thus there exists a substitution @xmath213 such that @xmath214 , i.e. , @xmath215 recall that the concept terms @xmath161 are reduced and pairwise incomparable w.r.t . subsumption .
in addition , since @xmath216 is reduced and not equivalent to @xmath184 , none of the concept terms @xmath161 can be equivalent to @xmath25 . finally , @xmath202 is a concept name that does not occur in @xmath161 .
all this implies that @xmath217 is reduced .
obviously , any reduced form for @xmath218 is a conjunction of at most @xmath219 existential restrictions .
thus , theorem [ cor1 ] shows that the above equivalence @xmath220 actually can not hold . to sum up
, we have shown that @xmath82 contains two distinct unifiers @xmath78 such that @xmath210 .
since @xmath82 was an arbitrary complete set of unifiers for @xmath73 , this shows that this unification problem can not have a minimal complete set of unifiers .
before we can describe our decision procedure for @xmath0-unification , we must introduce some notation .
an @xmath0-concept term is called an _ atom _ iff it is a concept name ( i.e. , concept constant or concept variable ) or an existential restriction @xmath221 . is _ not _ an atom . ] obviously , any @xmath0-concept term is ( equivalent to ) a conjunction of atoms , where the empty conjunction is @xmath25 .
the set @xmath222 of _ atoms of an @xmath0-concept term @xmath56 _ is defined inductively : if @xmath223 , then @xmath224 ; if @xmath56 is a concept name , then @xmath225 ; if @xmath226 then @xmath227 ; if @xmath228 , then @xmath229 .
concept names and existential restrictions @xmath221 where @xmath57 is a concept name or @xmath25 are called _
flat atoms_. an @xmath0-concept term is _ flat _ iff it is a conjunction of flat atoms ( where the empty conjunction is @xmath25 ) .
the @xmath0-unification problem @xmath73 is _ flat _ iff it consists of equations between flat @xmath0-concept terms . by introducing new concept variables and eliminating @xmath25 ,
any @xmath0-unification problem @xmath73 can be transformed in polynomial time into a flat @xmath0-unification problem @xmath230 such that @xmath73 is solvable iff @xmath230 is solvable .
thus , we may assume without loss of generality that our input @xmath0-unification problems are flat .
given a flat @xmath0-unification problem @xmath231 , we call the atoms of @xmath232 the _ atoms of @xmath73_. atoms of @xmath73 that are not variables ( i.e. , not elements of @xmath63 ) are called _ non - variable atoms of @xmath73_. the unifier @xmath65 of @xmath73 is called _ reduced _ iff , for all concept variables @xmath122 occurring in @xmath73 , the @xmath0-concept term @xmath233 is reduced .
it is _ ground _ iff , for all concept variables @xmath122 occurring in @xmath73 , the @xmath0-concept term @xmath233 does not contain variables .
obviously , @xmath73 is solvable iff it has a reduced ground unifier . given a ground unifier @xmath65 of @xmath73 , the _ atoms of @xmath65 _ are the atoms of all the concept terms @xmath233 , where @xmath122 ranges over all variables occurring in @xmath73 .
[ occurrence : remark ] in the following , we consider situations where all occurrences of a given reduced atom @xmath57 in a reduced concept term @xmath56 are replaced by a more general concept term , i.e. , by a concept term @xmath234 with @xmath235 .
however , when we say _ occurrence of @xmath57 in @xmath56 _ , we mean occurrence modulo equivalence ( @xmath236 ) rather than syntactic occurrence .
for example , if @xmath237 , @xmath238 , and @xmath239 , then the term obtained by replacing all occurrences of @xmath57 in @xmath56 by @xmath234 should be @xmath240 , and not @xmath241 .
since @xmath56 and @xmath57 are reduced , equivalence is actually the same as being identical up to associativity and commutativity of @xmath11 . in particular
, this means that any concept term that ( syntactically ) occurs in @xmath56 and is equivalent to the atom @xmath57 is also an atom , i.e. , only atoms can be replaced by @xmath234 . in order to make this meaning of occurrence explicit
we will call it _ occurrence modulo @xmath242 _ in the following
. we will write @xmath243 to express that the atoms @xmath244 and @xmath245 are identical up to associativity and commutativity of @xmath11 . obviously , @xmath243 implies @xmath246 .
[ monotone : greater : lem ] let @xmath247 be @xmath0-concept terms such that @xmath57 is a reduced atom , @xmath248 , and @xmath56 is reduced and contains at least one occurrence of @xmath57 modulo @xmath242 . if @xmath249 is obtained from @xmath56 by replacing all occurrences of @xmath57 by @xmath234 , then @xmath250 .
we prove the lemma by induction on the size of @xmath56 . if @xmath251 , then @xmath252 , and thus @xmath253 , which yields @xmath250 .
thus , assume that @xmath254 . in this case , @xmath56 can not be a concept name since it contains the atom @xmath57 . if @xmath255 , then @xmath57 occurs in @xmath204 modulo @xmath242 . by induction
, we can assume that @xmath256 , where @xmath257 is obtained from @xmath204 by replacing all occurrences of @xmath57 ( modulo @xmath242 ) by @xmath234 .
thus , we have @xmath258 by corollary [ cor2 ] .
finally , assume that @xmath259 for @xmath260 atoms @xmath161 . since @xmath56 is reduced , these atoms are incomparable w.r.t .
subsumption , and since the atom @xmath57 occurs in @xmath56 modulo @xmath242 we can assume without loss of generality that @xmath57 occurs in @xmath204 modulo @xmath242 .
let @xmath261 be respectively obtained from @xmath161 by replacing every occurrence of @xmath57 ( modulo @xmath242 ) by @xmath234 , and then reducing the concept term obtained this way . by induction , we have @xmath256 .
assume that @xmath262 .
since the concept constructors of @xmath0 are monotone w.r.t .
subsumption @xmath263 , we have @xmath264 , and thus @xmath262 means that @xmath265 .
consequently , @xmath259 and the reduced form of @xmath266 must be equal up to associativity and commutativity of @xmath11 .
if @xmath266 is not reduced , then its reduced form is actually a conjunction of @xmath267 atoms , which contradicts @xmath265 .
if @xmath266 is reduced , then @xmath256 implies that there is an @xmath268 such that @xmath269 .
however , then @xmath270 contradicts the fact that the atoms @xmath161 are incomparable w.r.t .
subsumption .
proposition [ well : founded1 ] says that the inverse strict subsumption order on concept terms is well - founded .
we use this fact to obtain a well - founded strict order @xmath271 on ground unifiers .
let @xmath78 be ground unifiers of @xmath73 .
we define 1 .
@xmath272 iff @xmath273 holds for all variables @xmath122 occurring in @xmath73 .
@xmath274 iff @xmath272 and @xmath275 , i.e. , iff @xmath273 holds for all variables @xmath122 occurring in @xmath73 , and @xmath276 holds for at least one variable @xmath122 occurring in @xmath73 .
if @xmath73 contains @xmath219 variables , then @xmath277 is the @xmath219-fold product of the order @xmath263 with itself .
since the strict part @xmath278 of the inverse subsumption order @xmath263 is well - founded by proposition [ well : founded1 ] , the strict part @xmath271 of @xmath277 is also well - founded @xcite .
the ground unifier @xmath65 of @xmath73 is called _ is - minimal _ iff there is no ground unifier @xmath81 of @xmath73 such that @xmath274 .
the following proposition is an easy consequence of the fact that @xmath271 is well - founded .
[ minimal : unif : prop ] let @xmath73 be an @xmath0-unification problem . then @xmath73 is solvable iff it has an is - minimal reduced ground unifier . in the following ,
we show that is - minimal reduced ground unifiers of flat @xmath0-unification problems satisfy properties that make it easy to check ( with an np - algorithm ) whether such a unifier exists or not .
[ atoms : equiv : lem ] let @xmath73 be a flat @xmath0-unification problem and @xmath279 an is - minimal reduced ground unifier of @xmath73 . if @xmath56 is an atom of @xmath279 , then there is a non - variable atom @xmath57 of @xmath73 such that @xmath280 . the main idea underlying
the proof of this crucial lemma is that an atom @xmath56 of a unifier @xmath65 that violates the condition of the lemma ( i.e. , that is not of the form @xmath280 for a non - variable atom @xmath57 of @xmath73 ) can be replaced by a concept term @xmath147 such that @xmath281 , which yields a unifier of @xmath73 that is smaller than @xmath65 w.r.t .
@xmath271 . before proving the lemma formally ,
let us illustrate this idea by two examples .
first , consider the unification problem @xmath282 the substitution @xmath283 is a unifier of @xmath284 that does not satisfy the condition of lemma [ atoms : equiv : lem ] .
in fact , @xmath89 is an atom of @xmath285 , but none of the non - variable atoms @xmath57 of @xmath284 ( which are @xmath44 , @xmath286 , and @xmath287 ) satisfy @xmath288 .
the unifier @xmath285 is not is - minimal since @xmath289 , which can be obtained from @xmath285 by replacing the offending atom @xmath89 with @xmath25 , is a unifier of @xmath284 that is smaller than @xmath285 w.r.t .
the unifier @xmath290 is is - minimal , and it clearly satisfies the condition of lemma [ atoms : equiv : lem ] .
second , consider the unification problem @xmath291 the substitution @xmath292 is a unifier of @xmath293 that does not satisfy the condition of lemma [ atoms : equiv : lem ] .
in fact , @xmath294 is an atom of @xmath295 , but none of the non - variable atoms @xmath57 of @xmath293 ( which are @xmath44 , @xmath89 , @xmath286 , and @xmath296 ) satisfy @xmath297 .
the unifier @xmath295 is not is - minimal since @xmath298 , which can be obtained from @xmath295 by replacing the offending atom @xmath294 with @xmath299 , is a unifier of @xmath293 that is smaller than @xmath295 w.r.t .
the unifier @xmath300 is is - minimal , and it clearly satisfies the condition of lemma [ atoms : equiv : lem ] .
assume that @xmath279 is an is - minimal reduced ground unifier of @xmath73 .
since @xmath279 is reduced , all atoms of @xmath279 are reduced . in particular
, this implies that @xmath56 is reduced , and since @xmath279 is ground , we know that @xmath56 is either a concept constant or an existential restriction
. _ first _ , assume that _
@xmath56 is of the form @xmath44 for a concept constant @xmath44 _ , but there is no non - variable atom @xmath57 of @xmath73 such that @xmath301 .
this simply means that @xmath44 does not appear in @xmath73 .
let @xmath302 be the substitution obtained from @xmath279 by replacing every occurrence of @xmath44 by @xmath25 . since equivalence in @xmath0 is preserved under replacing concept names by @xmath25 , and since @xmath44 does not appear in @xmath73 , it is easy to see that @xmath302 is also a unifier of @xmath73 .
however , since @xmath303 , this contradicts our assumption that @xmath279 is is - minimal .
_ second _ , assume that _
@xmath56 is an existential restriction of the form @xmath304 _ , but there is no non - variable atom @xmath57 of @xmath73 such that @xmath280 .
we assume that @xmath56 is maximal ( w.r.t .
subsumption ) with this property , i.e. , for every atom @xmath249 of @xmath279 with @xmath305 , there is a non - variable atom @xmath234 of @xmath73 such that @xmath306 .
let @xmath307 be all the non - variable atoms of @xmath73 with @xmath308 ( @xmath309 ) . by our assumptions on @xmath56 , we actually have @xmath310 and , by lemma [ lem1 ] , the atom @xmath311 is also an existential restriction @xmath312 ( @xmath309 ) .
we consider the conjunction @xmath313 which is @xmath25 in case @xmath314 .
given an @xmath0-concept term @xmath315 , the concept term @xmath316}}$ ] is obtained from @xmath315 by replacing every occurrence of @xmath56 ( modulo @xmath242 ) by @xmath147 .
the substitution @xmath317}}$ ] is obtained from @xmath279 by replacing every occurrence of @xmath56 ( modulo @xmath242 ) by @xmath147 , i.e. , @xmath317}}(x ) : = \gamma(x){^{[c/\widehat{d}]}}$ ] for all variables @xmath122 .
we will show in the following that @xmath317}}$ ] is a unifier of @xmath73 that is smaller than @xmath279 w.r.t .
this will then again contradict our assumption that @xmath279 is is - minimal .
@xmath318}}$ ] .
obviously , @xmath147 subsumes @xmath56 .
we claim that this subsumption relationship is actually strict .
in fact , if @xmath314 , then @xmath319 , and since @xmath56 is an atom , it is not equivalent to @xmath25 .
if @xmath320 , then @xmath321 would imply ( by corollary [ cor2 ] ) that there is an @xmath322 , with @xmath323 .
however , this would yield @xmath324 , which contradicts the fact that @xmath310 .
thus , we have shown that @xmath325 .
lemma [ monotone : greater : lem ] implies that @xmath303 . to complete the proof of lemma [ atoms : equiv : lem ] , it remains to show the next lemma .
@xmath317}}$ ] is a unifier of @xmath73 .
consider an equation in @xmath73 of the form @xmath326 where @xmath327 and @xmath328 are flat atoms , and define @xmath329 and @xmath330 .
we know that @xmath331 are conjunctions of atoms of the form @xmath332 and @xmath333 , where each conjunct @xmath334 is a reduced ground atom that is either an atom of @xmath279 or equal to @xmath335 for a non - variable atom @xmath336 of @xmath73 .
since @xmath279 is a unifier of @xmath73 , we have @xmath337 . 1 .
[ part : one ] since @xmath56 is an atom , we obviously have @xmath338}}= a_1{^{[c/\widehat{d}]}}\sqcap\ldots\sqcap a_\mu{^{[c/\widehat{d}]}}$ ] and @xmath339}}= b_1{^{[c/\widehat{d}]}}\sqcap\ldots\sqcap b_\nu{^{[c/\widehat{d}]}}$ ] .
now , we show that @xmath338}}= \gamma{^{[c/\widehat{d}]}}(l_1\sqcap\ldots\sqcap l_m)$ ] and @xmath339}}= \gamma{^{[c/\widehat{d}]}}(r_1\sqcap\ldots\sqcap r_n)$ ] .
we concentrate on proving the first identity since the second one can be shown analogously .
to show the first identity , it is enough to prove that @xmath340}}= \gamma{^{[c/\widehat{d}]}}(l_j)$ ] holds for all @xmath168 .
a. [ case : a ] if @xmath341 is a variable @xmath122 , then @xmath317}}(x ) = \gamma(x){^{[c/\widehat{d}]}}$ ] holds by the definition of @xmath317}}$ ] .
b. [ case : b ] if @xmath341 is a concept constant @xmath44 , then @xmath342}}= a$ ] since @xmath56 is an existential restriction .
thus , we have @xmath317}}(a ) = a = a{^{[c/\widehat{d}]}}= \gamma(a){^{[c/\widehat{d}]}}$ ] . c. [ case : c
] otherwise , @xmath341 is an existential restriction @xmath343 . by our assumption on @xmath56
, we have @xmath344 , and thus @xmath340}}= \exists\ , r_j.\left(\gamma(l_j){^{[c/\widehat{d}]}}\right)$ ] . in addition
, we have @xmath317}}(l_j ) = \exists\ , r_j.\gamma{^{[c/\widehat{d}]}}(l_j')$ ] .
thus , it is enough to show @xmath345}}= \gamma{^{[c/\widehat{d}]}}(l_j')$ ] . since @xmath341 is a flat atom , we know that @xmath346 is either a concept constant , the top - concept @xmath25 , or a concept variable . in the first to cases , we can show @xmath345}}= \gamma{^{[c/\widehat{d}]}}(l_j')$ ] as in ( [ case : b ] ) , and in the third case we can show this identity as in ( [ case : a ] ) .
2 . because of ( [ part : one ] ) , if we can prove that @xmath338}}\equiv r{^{[c/\widehat{d}]}}$ ] , then we have shown that @xmath317}}$ ] solves the equation @xmath326 .
+ without loss of generality , we concentrate on showing that @xmath338}}\sqsubseteq r{^{[c/\widehat{d}]}}$ ] . since @xmath338}}= a_1{^{[c/\widehat{d}]}}\sqcap\ldots\sqcap a_\mu{^{[c/\widehat{d}]}}$ ] and @xmath339}}= b_1{^{[c/\widehat{d}]}}\sqcap\ldots\sqcap b_\nu{^{[c/\widehat{d}]}}$ ] , it is thus sufficient to show that , for every @xmath347 , there exists a @xmath348 , such that @xmath349}}\sqsubseteq b_i{^{[c/\widehat{d}]}}$ ] ( see ( 3 ) of lemma [ 1:new : lem ] ) .
since @xmath350 and @xmath351 are atoms , we actually know that , for every @xmath347 , there exists a @xmath348 , such that @xmath352 .
thus , it is sufficient to show that @xmath352 implies @xmath349}}\sqsubseteq b_i{^{[c/\widehat{d}]}}$ ] .
this is an easy consequence of the next lemma since @xmath353 satisfy the conditions of this lemma .
[ hilf:2 ] let @xmath354 be reduced ground atoms such that @xmath89 is an atom of @xmath279 or of the form @xmath355 for a non - variable atom @xmath57 of @xmath73 . if @xmath356 , then @xmath342}}\sqsubseteq b{^{[c/\widehat{d}]}}$ ] .
we show @xmath342}}\sqsubseteq b{^{[c/\widehat{d}]}}$ ] by induction on the size of @xmath44 . 1 .
first , assume that @xmath357 , which implies that @xmath342}}= \widehat{d } = \gamma(d_1)\sqcap\ldots\sqcap\gamma(d_n)$ ] .
a. if @xmath89 is of the form @xmath358 for a non - variable atom @xmath57 of @xmath73 , then there is an @xmath359 , such that @xmath360 , which shows that @xmath342}}\sqsubseteq b$ ] . since @xmath361 and the constructors of @xmath0 are monotone w.r.t .
subsumption , we also have @xmath362}}$ ] , and thus @xmath342}}\sqsubseteq b{^{[c/\widehat{d}]}}$ ] .
b. assume that @xmath89 is an atom of @xmath279 .
if @xmath363 , then @xmath364}}= \widehat{d}$ ] , and thus @xmath342}}= b{^{[c/\widehat{d}]}}$ ] , which implies @xmath342}}\sqsubseteq b{^{[c/\widehat{d}]}}$ ] .
otherwise , since @xmath365 are reduced atoms , @xmath366 implies @xmath367 .
together with @xmath368 , this shows that @xmath369 .
thus , the maximality of @xmath56 implies that there is a non - variable atom @xmath57 of @xmath73 such that @xmath358 .
thus , we are actually in case ( a ) , which yields @xmath342}}\sqsubseteq b{^{[c/\widehat{d}]}}$ ] .
2 . now , assume that @xmath370 .
if there is no occurrence ( modulo @xmath242 ) of @xmath56 in @xmath44 , then we have @xmath342}}= a \sqsubseteq b \sqsubseteq b{^{[c/\widehat{d}]}}$ ] .
+ otherwise , @xmath44 is of the form @xmath371 and @xmath56 occurs in @xmath336 ( modulo @xmath242 ) .
obviously , @xmath356 then implies that @xmath89 is of the form @xmath372 with @xmath373 .
the concept terms @xmath374 are conjunctions of reduced ground atoms , i.e. , @xmath375 and @xmath376 where @xmath377 are reduced ground atoms . by corollary [ cor2 ] , for every @xmath378 , there exists @xmath379 such that @xmath380 . + in order to be able to assume , by induction , that @xmath380 implies @xmath381}}\sqsubseteq f_h{^{[c/\widehat{d}]}}$ ] , we must show that the conditions in the statement of the lemma hold for the concept terms @xmath382 , where @xmath383 plays the rle of @xmath44 and @xmath384 plays the rle of @xmath89 . since we already know that @xmath377 are reduced ground atoms , it is sufficient to show that each of the atoms @xmath385 is an atom of @xmath279 or of the form @xmath355 for a non - variable atom @xmath57 of @xmath73 .
we know that @xmath386 is an atom of @xmath279 or an instance ( w.r.t .
@xmath279 ) of a non - variable atom of @xmath73 . in the first case , the atoms @xmath385 are clearly also atoms of @xmath279 . in the second case , @xmath387 for a non - variable atom @xmath234 of @xmath73 . if @xmath234 is a ground atom , then @xmath385 are also ground atoms that are atoms of @xmath73 , and thus they are instances ( w.r.t .
@xmath279 ) of non - variable atoms of @xmath73 . otherwise , since @xmath73 is flat , @xmath234 is of the form @xmath388 for a variable @xmath122 and @xmath389 . in this case , @xmath385 are clearly atoms of @xmath279 .
+ thus , we can assume by induction : @xmath390}}\sqsubseteq f_h{^{[c/\widehat{d}]}}$}\ ] ] it remains to show that this implies @xmath342}}\sqsubseteq b{^{[c/\widehat{d}]}}$ ] .
a. if @xmath391 , then @xmath342}}= \exists\ , s.(e_1{^{[c/\widehat{d}]}}\sqcap\ldots\sqcap e_\kappa{^{[c/\widehat{d}]}})$ ] and @xmath364}}= \exists\ , s.(f_1{^{[c/\widehat{d}]}}\sqcap\ldots\sqcap f_\lambda{^{[c/\widehat{d}]}})$ ] , and thus property @xmath392 yields @xmath342}}\sqsubseteq b{^{[c/\widehat{d}]}}$ ] .
b. assume that @xmath393 . in this case , @xmath56 can not occur ( modulo @xmath242 ) in any of the concept terms @xmath394 , which implies that @xmath395}}\sqcap\ldots\sqcap f_\lambda{^{[c/\widehat{d}]}})$ ] .
since we have @xmath342}}= \exists\ , s.(e_1{^{[c/\widehat{d}]}}\sqcap\ldots\sqcap e_\kappa{^{[c/\widehat{d}]}})$ ] , property @xmath392 yields @xmath342}}\sqsubseteq b$ ] . since we also have @xmath362}}$ ] , this yields @xmath342}}\sqsubseteq b{^{[c/\widehat{d}]}}$ ] .
thus , we have shown in all cases that @xmath342}}\sqsubseteq b{^{[c/\widehat{d}]}}$ ] , which completes the proof of lemma [ hilf:2 ] .
overall , we have thus completed the proof of lemma [ atoms : equiv : lem ] .
the next proposition is an easy consequence of this lemma .
[ main : prop ] let @xmath73 be a flat @xmath0-unification problem and @xmath279 an is - minimal reduced ground unifier of @xmath73 . if @xmath122 is a concept variable occurring in @xmath73 , then @xmath396 or there are non - variable atoms @xmath170 ( @xmath160 ) of @xmath73 such that @xmath397 .
if @xmath398 , then it is a non - empty conjunction of atoms , i.e. , there are atoms @xmath399 ( @xmath160 ) such that @xmath400
. then @xmath399 are atoms of @xmath279 , and thus lemma [ atoms : equiv : lem ] yields non - variable atoms @xmath401 of @xmath73 such that @xmath402 for @xmath403 .
consequently , @xmath397 .
this proposition suggests the following non - deterministic algorithm for deciding solvability of a given flat @xmath0-unification problem .
[ algo:1 ] let @xmath73 be a flat @xmath0-unification problem . 1 .
for every variable @xmath122 occurring in @xmath73 , guess a finite , possibly empty , set @xmath404 of non - variable atoms of @xmath73 .
we say that the variable @xmath122 _ directly depends on _ the variable @xmath405 if @xmath405 occurs in an atom of @xmath404 .
let _ depends on _ be the transitive closure of _ directly depends on_. if there is a variable that depends on itself , then the algorithm returns `` fail . ''
otherwise , there exists a strict linear order @xmath406 on the variables occurring in @xmath73 such that @xmath407 if @xmath122 depends on @xmath405 .
we define the substitution @xmath65 along the linear order @xmath406 : 1 . if @xmath122 is the least variable w.r.t .
@xmath406 , then @xmath404 does not contain any variables .
we define @xmath233 to be the conjunction of the elements of @xmath404 , where the empty conjunction is @xmath25 .
2 . assume that @xmath408 is defined for all variables @xmath409 .
then @xmath404 only contains variables @xmath405 for which @xmath408 is already defined .
if @xmath404 is empty , then we define @xmath410 .
otherwise , let @xmath411 .
we define @xmath412 .
test whether the substitution @xmath65 computed in the previous step is a unifier of @xmath73 .
if this is the case , then return @xmath65 ; otherwise , return `` fail . ''
this algorithm is trivially _ sound _ since it only returns substitutions that are unifiers of @xmath73 .
in addition , it obviously always terminates .
thus , to show correctness of our algorithm , it is sufficient to show that it is complete .
if @xmath73 is solvable , then there is a way of guessing in step 1 subsets @xmath404 of the non - variable atoms of @xmath73 such that the _ depends on _ relation determined in step 2 is acyclic and the substitution @xmath65 computed in step 3 is a unifier of @xmath73 .
if @xmath73 is solvable , then it has an is - minimal reduced ground unifier @xmath279 . by proposition [ main : prop ] ,
for every variable @xmath122 occurring in @xmath73 we have @xmath396 or there are non - variable atoms @xmath170 ( @xmath160 ) of @xmath73 such that @xmath397 . if @xmath396 , then we define @xmath413 .
otherwise , we define @xmath414 .
we show that the relation _ depends on _ induced by these sets @xmath404 is acyclic , i.e. , there is no variable @xmath122 such that @xmath122 depends on itself .
if @xmath122 directly depends on @xmath405 , then @xmath405 occurs in an element of @xmath404 .
since @xmath404 consists of non - variable atoms of the flat unification problem @xmath73 , this means that there is a role name @xmath46 such that @xmath415 .
consequently , we have @xmath416 . thus ,
if @xmath122 depends on @xmath122 , then there are @xmath417 role names @xmath418 such that @xmath419 .
this is clearly not possible since @xmath420 can not be subsumed by an @xmath0-concept term whose role depth is larger than the role depth of @xmath420 . to show that the substitution @xmath65 induced by the sets @xmath404 is a unifier of @xmath73
, we prove that @xmath65 is equivalent to @xmath279 , i.e. , @xmath421 holds for all variables @xmath122 occurring in @xmath73 .
the substitution @xmath65 is defined along the linear order @xmath406 . if @xmath122 is the least variable w.r.t .
@xmath406 , then the elements of @xmath404 do not contain any variables .
if @xmath404 is empty , then @xmath422 .
otherwise , let @xmath411 . since the atoms @xmath311 do not contain variables , we have @xmath423 .
thus , the definitions of @xmath404 and of @xmath65 yield @xmath424 .
assume that @xmath425 holds for all variables @xmath409 .
if @xmath426 , then we have again @xmath422 . otherwise , let @xmath411 .
since the atoms @xmath311 contain only variables that are smaller than @xmath122 , we have @xmath427 by induction .
thus , the definitions of @xmath404 and of @xmath65 yield @xmath428 . note that our proof of completeness actually shows that , up to equivalence , the algorithm returns all is - minimal reduced ground unifiers of @xmath73 .
[ np : complete : thm ] @xmath0-unification is np - complete .
np - hardness follows from the fact that @xmath0-matching is np - complete @xcite . to show that the problem can be decided by a non - deterministic polynomial - time algorithm
, we analyze the complexity of our algorithm . obviously , guessing the sets @xmath404 ( step 1 ) can be done within np . computing
the _ depends on _ relation and checking it for acyclicity ( step 2 ) is clearly polynomial .
steps 3 and 4 are more problematic .
in fact , since a variable may occur in different atoms of @xmath73 , the substitution @xmath65 computed in step 3 may be of exponential size .
this is actually the same reason that makes a naive algorithm for syntactic unification compute an exponentially large most general unifier @xcite . as in the case of syntactic unification ,
the solution to this problem is basically structure sharing . instead of computing the substitution @xmath65 explicitly
, we view its definition as an acyclic tbox .
to be more precise , for every concept variable @xmath122 occurring in @xmath73 , the tbox @xmath429 contains the concept definition @xmath430 if @xmath426 and @xmath431 if @xmath411 ( @xmath160 ) . instead of computing @xmath65 in step 3
, we compute @xmath429 . because of the acyclicity test in step 2 , we know that @xmath429 is an acyclic tbox .
the size of @xmath429 is obviously polynomial in the size of @xmath73 , and thus this modified step 3 is polynomial .
it is easy to see that applying the substitution @xmath65 to a concept term @xmath56 is the same as expanding @xmath56 w.r.t .
the tbox @xmath429 , i.e. , @xmath432 .
this implies that , for every equation @xmath433 in @xmath73 , we have @xmath434 iff @xmath435 .
thus , testing in step 4 whether @xmath65 is a unifier of @xmath73 can be reduced to testing whether @xmath434 holds for every equation @xmath433 in @xmath73 . since subsumption ( and thus equivalence ) in @xmath0 w.r.t .
acyclic tboxes can be decided in polynomial time @xcite , this completes the proof of the theorem . in subsection
[ unif : mod : tbox ] , we have shown that there exists a polynomial - time reduction of unification modulo an acyclic tbox to unification without a tbox .
thus , theorem [ np : complete : thm ] also yields the exact complexity for @xmath0-unification w.r.t .
acyclic tboxes .
@xmath0-unification w.r.t .
acyclic tboxes is np - complete .
the problem is in np since theorem [ reduction : thm ] states that there is a polynomial - time reduction of @xmath0-unification w.r.t .
acyclic tboxes to @xmath0-unification , and we have just shown that @xmath0-unification is in np .
np - hardness for @xmath0-unification w.r.t .
acyclic tboxes follows from np - hardness of @xmath0-unification since @xmath0-unification can be viewed as the special case of @xmath0-unification w.r.t .
acyclic tboxes where the tbox is empty .
the np - algorithm introduced in the previous section is a typical `` guess and then test '' np - algorithm , and thus it is unlikely that a direct implementation of this algorithm will perform well in practice . here
, we introduce a more goal - oriented unification algorithm for @xmath0 , in which non - deterministic decisions are only made if they are triggered by `` unsolved parts '' of the unification problem . as in the previous section ,
we assume without loss of generality that our input unification problem @xmath436 is flat . for a given flat equation @xmath433 , the concept terms @xmath6
are thus conjunctions of flat atoms .
we will often view such an equation as consisting of four sets : the left - hand side @xmath56 is given by the set of variables occurring in the top - level conjunction of @xmath56 , together with the set of non - variable atoms occurring in this top - level conjunction ; the right - hand side @xmath57 is given by the set of variables occurring in the top - level conjunction of @xmath57 , together with the set of non - variable atoms occurring in this top - level conjunction . to be more precise ,
let @xmath437 denote the equation @xmath433 , where @xmath438 and @xmath439 for concept variables @xmath440 and non - variable atoms @xmath441 .
then we define @xmath442 obviously , the equation @xmath443 is uniquely determined ( up to associativity , commutativity , and idempotency of conjunction ) by the four sets @xmath444 . instead of viewing an equation @xmath437 as being given by a pair of concept terms , we can thus also view it as being given by these four sets . in the following
, it will often be convenient to employ this representation of equations .
if , with this point of view , we say that we _
add an atom _ to the set @xmath445 or @xmath446 , then this means , for the other point of view , that we conjoin this atom to the top - level conjunction of the left - hand side or right - hand side of the equation .
in addition , if we say that the equation @xmath437 _ contains _ the variable @xmath122 , then we mean that @xmath447 .
similarly , if we say that the left - hand side of @xmath437 contains @xmath122 , then we mean that @xmath448 , and if we say that the right - hand side of @xmath437 contains @xmath122 , then we mean that @xmath449 ) . inside
non - variable atoms @xmath450 are not taken into consideration here . ]
in addition to the unification problem itself , the algorithm also maintains , for every variable @xmath122 occurring in the input problem @xmath436 , a set @xmath404 of non - variable atoms of @xmath436 .
initially , all the sets @xmath404 are empty .
we call the set @xmath404 the current assignment for @xmath122 , and the collection of all these sets the
_ current assignment_. throughout the run of our goal - oriented algorithm , we will ensure that the current assignment is _ acyclic _ in the sense that no variable depends on itself w.r.t .
this assignment ( see ( 2 ) of algorithm [ algo:1 ] ) .
an acyclic assignment induces a substitution @xmath65 , as defined in ( 3 ) of algorithm [ algo:1 ] .
we call this substitution the _
current substitution_. initially , the current substitution maps all variables to @xmath25 .
the algorithm applies rules that can 1 .
change an equation of the unification problem by adding non - variable atoms of the input problem @xmath436 to one side of the equation ; 2 . introduce a new flat equation of the form @xmath451 , where @xmath365 are atoms of the input problem @xmath436 or @xmath25 ; 3 .
add non - variable atoms of the input problem @xmath436 to the sets @xmath404 .
another property that is maintained throughout the run of our algorithm is that all equations @xmath437 are _ expanded _ w.r.t . the current assignment in the following sense : for all variables
@xmath122 we have @xmath452 given a flat equation @xmath437 that contains the variable @xmath122 , the _ expansion _ of @xmath437 w.r.t .
the assignment @xmath404 for @xmath122 is defined as follows : if @xmath448 then all elements of @xmath404 are added to @xmath445 , and if @xmath449 then all elements of @xmath404 are added to @xmath446 .
the following lemma is an immediate consequence of the definition of expanded equations and of the construction of the current substitution .
[ expanded : lem ] if the equation @xmath433 is expanded w.r.t . the current assignment , then @xmath453 implies that the current substitution @xmath65 solves this equation , i.e. , @xmath454 .
we say that an equation @xmath437 is _ solved _ if @xmath455 . an atom @xmath456 is called _ solved in @xmath437 _ ; atoms @xmath457 that are not solved in @xmath437 are called _ unsolved in @xmath437_. obviously , an equation @xmath437 is solved iff all atoms @xmath458 are solved in @xmath437 . basically , in each step , the goal - oriented algorithm considers an unsolved equation and an unsolved atom in this equation , and tries to solve it . picking the unsolved equation and
the unsolved atom in it is do nt care non - deterministic , i.e. , there is no need to backtrack over such a choice .
once an unsolved equation and an unsolved atom in it was picked , do nt know non - determinism comes in since there may be several possibilities for how to solve this atom in the equation , some of which may lead to overall success whereas others wo nt . in some cases , however , a given equation uniquely determines the assignment for a certain variable @xmath122 . in this case
, we make this assignment and then label the variable @xmath122 as _
finished_. this has the effect that the set @xmath404 can no longer be extended .
initially , none of the variables occurring in the input unification problem is labeled as finished .
we say that the variable @xmath122 is _ unfinished _ if it is not labeled as finished .
[ goal : oriented : alg ] let @xmath436 be a flat @xmath0-unification problem . we define @xmath459 and @xmath413 for all variables @xmath122 occurring in @xmath436 .
none of these variables is labeled as finished .
as long as @xmath73 contains an unsolved equation , do the following : 1 . if the eager - assignment rule applies to some equation @xmath437 , then apply it to this equation ( see figure [ eager ] ) .
2 . otherwise , let @xmath437 be an unsolved equation and @xmath44 an unsolved atom in @xmath437 .
if neither of the rules decomposition ( see figure [ decomp ] ) and extension ( see figure [ extent ] ) applies to @xmath44 in @xmath437 , then return `` fail . ''
if one of these rules applies to @xmath44 in @xmath437 , then ( do nt know ) non - deterministically choose one of these rules and apply it .
once all equations of @xmath73 are solved , return the substitution @xmath65 that is induced by the current assignment .
the l - variant of the eager - assignment rule applies to the equation @xmath437 if there is an unfinished variable @xmath448 such that 1 .
all variables @xmath460 are finished ; 2 .
@xmath461 .
its application sets @xmath462 . 1 .
if this makes the current assignment cyclic , then return `` fail .
otherwise , label @xmath122 as finished and expand all equations containing @xmath122 w.r.t . the new assignment for @xmath122 . the _ eager - assignment rule _
is described in figure [ eager ] .
note that , after a non - failing application of this rule , the equation it was applied to is solved since the expansion of this equation w.r.t .
the new assignment for @xmath122 adds all elements of @xmath446 to @xmath445 . as an example , consider the equations @xmath463 and assume that @xmath464 and none of the three variables @xmath465 is finished .
an application of the eager - assignment rule to the first equation labels @xmath405 as finished , but does not change anything else .
the subsequent application of the eager - assignment rule to the second equation changes the assignment for @xmath202 to @xmath466 , labels @xmath202 as finished , and expands the second and the third equation w.r.t . the new assignment for @xmath202 .
thus , we now have the equations @xmath467 since @xmath468 are finished , the eager - assignment rule can now be applied to the third equation .
this changes the assignment for @xmath122 to @xmath469 , labels @xmath122 as finished , and adds @xmath184 to the left - hand side of the third equation .
now all equations are solved .
the current assignment induces a substitution @xmath65 with @xmath470 and @xmath471 , which is a unifier of the original set of equations .
the l - variant of the decomposition rule applies to the unsolved atom @xmath44 in the equation @xmath437 if 1 .
@xmath472 ; 2 .
@xmath44 is of the form @xmath473 ; 3 .
there is at least one atom of the form @xmath474 .
its application chooses ( do nt know ) non - deterministically an atom of the form @xmath474 and 1 .
adds @xmath12 to @xmath446 ; 2 . creates a new equation @xmath475 and expands it w.r.t .
the assignments of all variables contained in this equation , unless this equation has already been generated before .
if the equation has already been generated before , it is not generated again .
the _ decomposition rule _ is described in figure [ decomp ] .
this rule solves the unsolved atom @xmath473 by adding it to the other side . for this to be admissible
, one needs a more specific atom @xmath296 on that side , where the `` more specific '' is meant to hold after application of the unifier .
thus , to ensure that the unifier @xmath65 computed by the algorithm satisfies @xmath476 , the rule adds the new equation @xmath475 .
obviously , if the substitution @xmath65 solves this equation , then it satisfies @xmath477 , and thus @xmath476 . as an example , consider the equation @xmath478 and assume that @xmath426 and that @xmath122 is unfinished .
an application of the l - variant of the decomposition rule to this equation adds @xmath479 to the right - hand side of this equation , and thus solves it .
in addition , it generates the new equation @xmath480 , which is solved .
the current assignment induces a substitution @xmath65 with @xmath481 , which solves the original equation .
the _ extension rule _ is described in figure [ extent ] .
basically , this rule solves the unsolved atom @xmath44 by extending with this atom the assignment of an unfinished variable contained in the other side of the equation . as an example , consider the equation @xmath482 where @xmath44 is a concept constant , @xmath426 , and @xmath122 is unfinished .
an application of the extension rule to @xmath44 in this equation extends the assignment for @xmath122 to @xmath483 , and expands this equation by adding @xmath44 to the right - hand side .
the equation obtained this way is solved .
the substitution @xmath65 induced by the current assignment replaces @xmath122 by @xmath44 , and solves the original equation .
the l - variant of the extension rule applies to the unsolved atom @xmath44 of the equation @xmath437 if 1 .
@xmath472 ; 2 .
there is at least one unfinished variable @xmath484 its application chooses ( do nt know ) non - deterministically an unfinished variable @xmath484 and adds @xmath44 to @xmath404 . 1 .
if this makes the current assignment cyclic , then return `` fail .
otherwise , expand all equations containing @xmath122 w.r.t . the new + assignment for @xmath122 .
[ goal : oriented : thm ] algorithm [ goal : oriented : alg ] is an np - algorithm for testing solvability of flat @xmath0-unification problems .
first , we show that the algorithm is indeed an np - algorithm . for this
, we consider all _ runs _ of the algorithm , where for every ( do nt care ) non - deterministic choice exactly one alternative is taken . since a single rule application can obviously be realized in polynomial time , it is sufficient to show the following lemma .
every run of the algorithm terminates after a polynomial number of rule applications .
each application of the eager - assignment rule finishes an unfinished variable .
thus , since finished variables never become unfinished again , it can only be applied @xmath485 times , where @xmath485 is the number of variables occurring in the input unification problem @xmath436 .
this number is clearly linearly bounded by the size of @xmath436 .
every application of the decomposition rule or the extension rule turns an unsolved atom in an equation into a solved one , and a solved atom in an equation never becomes unsolved again in this equation . for a fixed equation , in the worst case
every atom of @xmath436 may become an unsolved atom of the equation that needs to be solved .
there is , however , only a linear number of atoms of @xmath436 .
each equation considered during the run of the algorithm is either descended from an original equation of @xmath436 , or from an equation of the form @xmath475 for atoms @xmath296 and @xmath12 of @xmath436 .
thus , the number of equations is also polynomially bounded by the size of @xmath436 .
overall , this shows that the decomposition rule and the extension rule can only be applied a polynomial number of times .
next , we show soundness of algorithm [ goal : oriented : alg ] .
we call a run of this algorithm _ non - failing _ if it terminates with a unification problem containing only solved equations .
let @xmath436 be a flat @xmath0-unification problem .
the substitution @xmath65 returned after a successful run of algorithm [ goal : oriented : alg ] on input @xmath436 is an @xmath0-unifier of @xmath436 .
first , note that the rules employed by algorithm [ goal : oriented : alg ] indeed preserve the two invariants mentioned before : 1 .
the current assignment is always acyclic ; 2 .
all equations are expanded .
in fact , whenever the current assignment is extended , the rules test acyclicity ( and return `` fail , '' if it is not satisfied ) .
in addition , they expand all equations w.r.t . the new assignment .
now , assume that the run of the algorithm has terminated with the @xmath0-unification problem @xmath486 , in which all equations are solved .
the first invariant ensures that the final assignment constructed by the run is acyclic , and thus indeed induces a substitution @xmath65 . because of the second invariant , lemma [ expanded : lem ] applies , and thus we know that @xmath65 is a solution of @xmath486 .
it remains to show that the substitution @xmath65 is also a solution of the input problem @xmath436 . to this purpose , we take all the equations that were considered during the run of the algorithm , i.e. , present in @xmath436 or in any of the other unification problems generated during the run .
let @xmath487 denote the set of these equations .
we define the relation @xmath488 on @xmath487 as follows : @xmath489 if @xmath437 was transformed into @xmath490 using one of the rules of algorithm [ goal : oriented : alg ] .
to be more precise , the eager - assignment rule transforms equations containing @xmath122 from the current unification problem @xmath73 by expanding them w.r.t . the new assignment for @xmath122 .
the same is true for the extension rule .
the decomposition rule transforms an equation @xmath437 containing the unsolved atom @xmath473 by adding this atom to the other side , which needs to contain an atom of the form @xmath296 . for this new equation @xmath490
, we have @xmath489 .
the decomposition rule may also generate a new equation @xmath491 of the form @xmath475 ( if this equation was not generated before ) .
however , we do not view this equation as a successor of @xmath437 w.r.t .
@xmath488 , i.e. , we do _ not _ have @xmath492 .
equations @xmath475 that are generated by an application of the decomposition rule are called _ d - equations_. equations that are elements of the input problem @xmath436 are called _ i - equations_. any equation @xmath490 that is not an i - equation or a d - equation has a unique predecessor w.r.t .
@xmath488 , i.e. , there is an equation @xmath493 such that @xmath489 . starting with the set @xmath494 we will now step by step extend @xmath495 by a predecessor of an equation in @xmath495 until no new predecessors can be added .
since @xmath487 is finite , this process terminates after a finite number of steps .
after termination we have @xmath496 , and thus in particular @xmath497 .
this is due to the fact that , for every element @xmath498 of @xmath487 , there are @xmath499 elements @xmath500 such that @xmath501 and @xmath502 .
thus , it is enough to show that the set @xmath495 satisfies the following _ invariant _ : @xmath503 since @xmath65 is a solution of @xmath486 , this invariant is initially satisfied . to prove that it is preserved under adding predecessors of equations in @xmath495 , we start with the equations of minimal role depth .
to be more precise , if the equation @xmath437 is of the form @xmath433 , we define the _ role depth of @xmath437 w.r.t .
@xmath65 _ to be the role depth for a definition . ] of the concept term @xmath504 .
the strict order @xmath271 on @xmath487 is defined as follows : @xmath505 iff the role depth of @xmath437 w.r.t .
@xmath65 is larger than the role depth of @xmath490 w.r.t .
we write @xmath506 if @xmath437 and @xmath490 have the same role depth w.r.t .
the following is an easy consequence of the definition of @xmath65 and of our rules : @xmath507 assume that we have already constructed a set @xmath495 such that the invariant @xmath392 is satisfied .
let @xmath490 be an equation in @xmath495 such that 1 .
there is an @xmath508 with @xmath489 ; 2 .
@xmath490 is of minimal role depth with this property , i.e. , if @xmath509 is such that @xmath510 and @xmath511 has a predecessor @xmath512 w.r.t .
@xmath488 , then @xmath513 .
if no such equation @xmath490 exists , then we are finished , and we have @xmath496 . otherwise , let @xmath490 be such an equation and @xmath437 its predecessor w.r.t . @xmath488 .
we add @xmath437 to @xmath495 . in order to show that the invariant @xmath392 is still satisfied
, we make a case distinction according to which rule was applied to @xmath437 to produce @xmath490 : 1 . _
eager - assignment .
_ by an application of this rule , the assignment for @xmath122 is modified from @xmath426 to @xmath514 , where @xmath515 are non - variable atoms .
in addition , @xmath122 is labeled as finished . since the assignment of a finished variable can not be changed anymore , we know that we also have @xmath514 in the final assignment , and thus @xmath516 .
the rule modifies equations as follows : all equations containing @xmath122 are expanded w.r.t . the assignment @xmath514 . since @xmath437 is transformed into @xmath490 using this rule , it must contain @xmath122 .
we assume for the sake of simplicity that @xmath122 is contained in the left - hand side of @xmath437 , but not in the right - hand side , i.e. , @xmath437 is of the form @xmath517 and the new equation @xmath518 obtained from @xmath437 is @xmath519 . since @xmath65 solves @xmath490 , we have @xmath520 , which shows that @xmath65 also solves @xmath437 .
_ without loss of generality , we consider the l - variant of this rule .
thus , the equation @xmath437 is of the form @xmath521 , and @xmath490 is obtained from @xmath437 by adding @xmath12 to the right - hand side , i.e. , @xmath490 is of the form @xmath522 . we know that @xmath65 solves @xmath490
thus , if we can show @xmath477 , then we have @xmath523 , which shows that @xmath65 solves @xmath437 .
+ consequently , it is sufficient to prove @xmath477 .
the decomposition rule also generates the equation @xmath475 and expands it w.r.t .
the assignments of all the variables contained in this equation , unless this equation has already been generated before . thus , either this application or a previous one of the decomposition rule has generated the equation @xmath475 , and then expanded it ( w.r.t . the current assignment at that time ) to an equation @xmath524 . since atoms are never removed from an assignment , the atoms present in the assignment at the time when the decomposition rule generated the equation @xmath475 are also present in the final assignment used to define the substitution @xmath65 .
thus , if we can show that @xmath65 solves @xmath524 , then we have also shown that @xmath65 solves @xmath475 , and thus satisfies @xmath477 .
+ since equations are never completely removed by our rules , but only modified , there is a sequence of equations @xmath525 such that @xmath502 .
property @xmath526 thus yields @xmath527 .
in addition , the role depth of @xmath475 w.r.t .
@xmath65 is the same as the role depth of @xmath524 w.r.t .
consequently , we have @xmath528 for all @xmath167 .
now , assume that @xmath529 .
then there is an @xmath530 such that @xmath531 , but @xmath532 .
this contradicts our assumption that @xmath490 is minimal .
thus , we have shown that @xmath533 , and this implies that @xmath65 solves @xmath524 .
+ overall , this finishes the proof that @xmath65 solves @xmath437 .
_ by an application of this rule , the assignment for @xmath122 is modified by adding a non - variable atom @xmath44 to it . since atoms are never removed from an assignment , we know that we also have @xmath534 in the final assignment , and thus @xmath535 .
the rule modifies equations as follows : all equations containing @xmath122 are expanded w.r.t . the new assignment for @xmath122 . since @xmath437 is transformed into @xmath490 using this rule , it must contain @xmath122 .
we assume for the sake of simplicity that @xmath122 is contained in the left - hand side of @xmath437 , but not in the right - hand side , i.e. , @xmath437 is of the form @xmath517 and the new equation @xmath490 obtained from @xmath437 is @xmath536 .
since @xmath65 solves @xmath490 , we have @xmath537 , which shows that @xmath65 also solves @xmath437 . to sum up
, we have shown that the invariant @xmath392 is still satisfied after adding @xmath437 to @xmath495 .
this completes the proof of soundness of our procedure .
it remains to show completeness of algorithm [ goal : oriented : alg ] .
thus , assume that the input unification problem @xmath436 is solvable .
proposition [ minimal : unif : prop ] tells us that @xmath436 then has an is - minimal reduced ground unifier @xmath279 , and proposition [ main : prop ] implies that , for every variable @xmath122 occurring in @xmath436 , there is a set @xmath538 of non - variable atoms of @xmath436 such that @xmath539 where , for a set of non - variable atoms @xmath540 of @xmath436 , the expression @xmath541 denotes the conjunction of the elements of @xmath540 ( where the empty conjunction is @xmath25 ) .
let @xmath436 be a flat @xmath0-unification problem , and assume that @xmath279 is an is - minimal reduced ground unifier of @xmath436 .
then there is a successful run of algorithm [ goal : oriented : alg ] on input @xmath436 that returns a unifier @xmath65 that is equivalent to @xmath279 , i.e. , satisfies @xmath421 for all variables @xmath122 occurring in @xmath436 .
the algorithm starts with @xmath459 and the initial assignment @xmath542 for all variables @xmath122 occurring in @xmath436 .
it then applies rules that change @xmath73 and the current assignment as long as the problem @xmath73 contains an unsolved equation .
we use @xmath279 to guide the ( do nt know ) non - deterministic choices to be made during the algorithm .
we show that this ensures that the run of the algorithm generated this way _ does not fail _ and that the following _ invariants _ are satisfied throughout this run : 1 .
@xmath279 is a unifier of @xmath73 ; 2 . for all atoms
@xmath543 there exists an atom @xmath544 such that @xmath545 ; 3 . for all finished variables
@xmath122 we have @xmath546 . before constructing a run that satisfies these invariants ,
let us point out two interesting consequences that they have : 1 .
_ the current assignment is always acyclic .
_ in fact , if @xmath122 directly depends on @xmath405 , then there is an atom @xmath547 that has the form @xmath548 for some role name @xmath46 .
invariant @xmath549 then implies that there is an @xmath544 such that @xmath550 .
thus , if @xmath122 depends on @xmath122 , then there are @xmath417 role names @xmath418 such that @xmath419 , which is impossible .
_ for each variable @xmath122 occurring in @xmath436 , we have @xmath551 , where @xmath65 is the current substitution induced by the current assignment .
_ this is again a consequence of invariant @xmath549 .
indeed , recall that the fact that the current assignment is acyclic implies that there is a strict linear order @xmath406 on the variables occurring in @xmath73 such that @xmath407 if @xmath122 depends on @xmath405 .
the current substitution @xmath65 is defined along this order .
we prove @xmath551 by induction on this order .
+ consider the least variable @xmath122 .
if @xmath426 , then @xmath481 , and thus @xmath551 is trivially satisfied .
otherwise , we know , for every @xmath547 , that it does not contain any variables , which implies that @xmath552 for some atom @xmath553 .
obviously , this yields @xmath554 .
+ now , assume that @xmath555 holds for all variables @xmath409 . since the concept constructors of @xmath0 are monotone w.r.t .
subsumption , this implies @xmath556 for all concept terms @xmath56 containing only variables smaller than @xmath122 . if @xmath404 is empty , then @xmath557 is trivially satisfied . otherwise , we know , for every @xmath547 , that it contains only variables smaller than @xmath122 .
this yields @xmath558 for some atom @xmath553 .
again , this implies @xmath554 .
since @xmath279 was assumed to be an is - minimal unifier of @xmath436 , the consequence @xmath559 implies that @xmath65 can only be a unifier of @xmath436 if @xmath65 is equivalent to @xmath279 .
if the run has terminated successfully , then the final substitution @xmath65 obtained by the run is a unifier of @xmath436 ( due to soundness ) .
thus , in this case the computed unifier @xmath65 is indeed equivalent to @xmath279 .
consequently , to prove the lemma , it is sufficient to _ construct a non - failing run of the algorithm that satisfies the above invariants . _
the invariants are initially satisfied since @xmath279 is a unifier of @xmath436 , the initial assignment for all variables @xmath122 occurring in @xmath436 is @xmath426 , and there are no finished variables .
now , assume that , by application of the rules of algorithm [ goal : oriented : alg ] , we have constructed a unification problem @xmath73 and a current assignment such that the invariants are satisfied . 1 . if all equations in @xmath73 are solved , then the run terminates successfully , and we are done .
2 . if there is an unsolved equation to which the _ eager - assignment rule
_ applies , then the algorithm picks such an equation @xmath437 and applies this rule to it . without loss of generality , we assume that the l - variant of the rule is applied .
the selected equation @xmath437 is of the form @xmath560 where @xmath561 are non - variable atoms , and @xmath562 are finished variables .
because the left - hand side of the equation does not contain any non - variable atoms , we know that @xmath563 ( since the algorithm keeps all equations expanded ) . since @xmath564 are finished , we thus have @xmath565 ( by invariant @xmath566 ) .
we also know that @xmath567 for all @xmath568 .
since the variables @xmath569 are finished , invariant @xmath566 implies that @xmath570 .
+ the new assignment for @xmath122 is @xmath571 , all equations containing @xmath122 are expanded w.r.t .
this assignment , and @xmath122 becomes a finished variable .
first , we show that @xmath566 is satisfied .
nothing has changed for the variables that were already finished before the application of the rule .
however , @xmath122 is now also finished .
thus , we must show that @xmath546 .
we know that @xmath279 solves the equation @xmath437 ( by @xmath572 ) .
this yields @xmath573 .
regarding @xmath549 , the only assignment that was changed is the one for @xmath122 .
since the new assignment for @xmath122 is @xmath571 , and we have already shown that @xmath574 , the invariant @xmath549 holds by corollary [ cor2 ] . note that this also implies that the new assignment is acyclic , and thus the application of the eager - assignment rule does not fail . finally , consider the invariant @xmath572 .
the rule application modifies equations containing @xmath122 by adding the atoms @xmath515 . since @xmath574 , an equation that was solved by @xmath279 before this modification ,
is also solved by @xmath279 after this modification . to sum up
, we have shown that the application of the eager - assignment rule does not fail and preserves the invariants .
if there is _ no _ unsolved equation to which the _ eager - assignment rule _ applies , then the algorithm picks an unsolved equation @xmath437 and an unsolved atom @xmath44 occurring in this equation .
we must show that we can apply either the decomposition or the extension rule to @xmath44 in @xmath437 such that the invariants stay satisfied . without loss of generality , we assume that the unsolved atom @xmath44 occurs on the left - hand side of the equation @xmath437 .
a. first , assume that _
@xmath44 is an existential restriction @xmath473_. the selected unsolved equation @xmath437 is thus of the form @xmath575 where @xmath576 and @xmath577 are ( variable or non - variable ) atoms and @xmath578 .
since @xmath279 solves this equation ( by invariant @xmath572 ) , corollary [ cor2 ] implies that there must be an @xmath167 , such that @xmath579 .
a. if @xmath580 is an existential restriction @xmath581 , then we have @xmath582 .
we apply the decomposition rule to @xmath44 and @xmath580 .
the application of this rule modifies the equation @xmath437 to an equation @xmath490 by adding the atom @xmath44 to the right - hand side .
in addition , it generates the equation @xmath475 and expands it w.r.t .
the assignments of all variables contained in this equation ( unless this equation has been generated before ) . after the application of this rule , the invariants @xmath549 and @xmath566 are still satisfied since the current assignments and the set of finished variables remain unchanged .
regarding invariant @xmath572 , since @xmath279 solves @xmath437 , it obviously also solves @xmath490 due to the fact that @xmath583 and @xmath580 is a conjunct on the right - hand side of @xmath437 .
in addition , @xmath582 implies that @xmath279 also solves the equation @xmath475 .
since invariant @xmath549 is satisfied , this implies that @xmath279 also solves the equation obtained from @xmath475 by expanding it w.r.t .
the assignments of all variables contained in it .
b. assume that there is no @xmath167 , such that @xmath580 is an existential restriction satisfying @xmath583 .
thus , if @xmath580 is such that @xmath583 , then we know that @xmath584 is a variable .
we want to apply the extension rule to @xmath44 and @xmath122 . to be able to do this , we must first show that @xmath122 is not a finished variable . + thus , assume that @xmath122 is finished , and let @xmath585 .
invariant @xmath566 yields @xmath586 , and thus there is a @xmath587 , such that @xmath588 .
since @xmath44 is an existential restriction , the non - variable atom @xmath589 must also be an existential restriction , and since the equation @xmath437 is expanded , @xmath590 occurs on the right - hand side of this equation .
this contradicts our assumption that there is no such existential restriction on the right - hand side .
thus , we have shown that @xmath122 is not finished , which means that we can apply the extension rule to @xmath44 and @xmath122 .
+ the application of this rule adds the atom @xmath44 to the assignment for @xmath122 , and it expands all equations containing @xmath122 w.r.t . this new assignment , i.e. , it adds @xmath44 to the left - hand side and/or right - hand side of an equation whenever @xmath122 is contained in this side . since we know that @xmath591 , it is easy to see that , if @xmath279 solves an equation before this expansion , it also solves it after the expansion .
thus invariant @xmath572 is satisfied .
invariant @xmath549 also remains satisfied .
in fact , if @xmath592 , then @xmath593 implies that there is a @xmath587 , such that @xmath594 .
the fact that @xmath549 is satisfied by the new assignment also implies that this new assignment is acyclic , and thus the application of the extension rule does not fail .
invariant @xmath566 is still satisfied since @xmath122 is not finished , and the assignments of variables different from @xmath122 were not changed .
b. second , assume that _
@xmath44 is a concept name_. the selected unsolved equation @xmath437 is thus of the form @xmath595 where @xmath576 and @xmath577 are ( variable or non - variable ) atoms , and @xmath596 .
since @xmath279 solves this equation ( by invariant @xmath572 ) , corollary [ cor2 ] implies that there must be an @xmath167 , such that @xmath597 . since @xmath596 , we know that @xmath584 is a variable .
we want to apply the extension rule to @xmath44 and @xmath122 . to be able to do this , we must first show that @xmath122 is not a finished variable .
+ thus , assume that @xmath122 is finished , and let @xmath585 .
invariant @xmath566 yields @xmath598 , and thus there is a @xmath587 , such that @xmath599 .
since @xmath44 is a concept name , the non - variable atom @xmath589 must actually be equal to @xmath44 , and since the equation @xmath437 is expanded , @xmath600 occurs on the right - hand side of this equation .
this contradicts our assumption that @xmath44 is an unsolved atom .
thus , we have shown that @xmath122 is not finished , which means that we can apply the extension rule to @xmath44 and @xmath122 .
the application of this rule adds the atom @xmath44 to the assignment for @xmath122 , and it expands all equations containing @xmath122 w.r.t .
this new assignment .
the proof that this rule application does not fail and preserves the invariants is identical to the one for the case where @xmath44 was an existential restriction . to sum up
, we have shown that algorithm [ goal : oriented : alg ] always terminates ( in non - deterministic polynomial time ) and that it is sound and complete .
this finishes the proof of theorem [ goal : oriented : thm ] .
unification problems and their types were originally not introduced for description logics , but for equational theories @xcite . in this section ,
we show that the above results for unification in @xmath0 can actually be viewed as results for an equational theory .
as shown in @xcite , the equivalence problem for @xmath0-concept terms corresponds to the word problem for the equational theory of semilattices with monotone operators . in order to define this theory
, we consider a signature @xmath601 consisting of a binary function symbol @xmath602 , a constant symbol @xmath86 , and finitely many unary function symbols @xmath603 .
terms can then be built using these symbols and additional variable symbols and free constant symbols .
the equational theory of _ semilattices with monotone operators _ is defined by the following identities : @xmath604 & \{f_i(x\wedge y)\wedge f_i(y ) = f_i(x\wedge y ) \mid 1\leq i\leq n\ } \end{array}\ ] ] a given @xmath0-concept term @xmath56 using only roles @xmath605 can be translated into a term @xmath606 over the signature @xmath601 by replacing each concept constant @xmath44 by a corresponding free constant @xmath607 , each concept variable @xmath122 by a corresponding variable @xmath608 , @xmath25 by @xmath86 , @xmath11 by @xmath602 , and @xmath609 by @xmath610 .
for example , the @xmath0-concept term @xmath611 is translated into @xmath612 .
conversely , any term over the signature @xmath601 can be translated back into an @xmath0-concept term .
let @xmath6 be @xmath0-concept term using only roles @xmath605
. then @xmath7 iff @xmath613 .
as an immediate consequence of this lemma , we have that unification in the dl @xmath0 corresponds to unification modulo the equational theory @xmath614 .
thus , theorem [ type : zero : thm ] implies that @xmath614 has unification type zero , and theorem [ np : complete : thm ] implies that @xmath614-unification is np - complete .
the equational theory @xmath614 of semilattices with monotone operators has unification type zero , and deciding solvability of an @xmath614-unification problem is an np - complete problem .
since the unification problem introduced in theorem [ type : zero : thm ] contains only one role @xmath46 , this is already true in the presence of a single monotone operator .
in this paper , we have shown that unification in the dl @xmath0 is of type zero and np - complete .
there are interesting differences between the behavior of @xmath0 and the closely related dl @xmath28 w.r.t .
unification and matching .
though the unification types coincide for these two dls , the complexities of the decision problems differ : @xmath28-unification is exptime - complete , and thus considerably harder than @xmath0-unification .
in contrast , @xmath28-matching is polynomial , and thus considerably easier than @xmath0-matching , which is np - complete .
in addition to showing the complexity upper bound for @xmath0-unification by a simple `` guess and then test '' np - algorithm , we have also developed a more goal - oriented np - algorithm that makes ( do nt know ) non - deterministic decisions ( i.e. , ones that require backtracking ) only if they are triggered by unsolved atoms in the unification problem . as future work , we will consider also unification of concept terms for other members of the @xmath0-family of dls @xcite .
in addition , we will investigate unification modulo more expressive terminological formalisms . on the practical side ,
we will optimize and implement the goal - oriented @xmath0-unification algorithm developed in section [ goal : oriented : sect ] .
we intend to test the usefulness of this algorithm for the purpose on finding redundancies in @xmath0-based ontologies by considering extensions of the medical ontology snomed ct .
for example , in @xcite , two different extensions of snomed ct by so - called post - coordinated concepts were considered .
the authors used an ( incomplete ) equivalence test to find out how large the overlap between the two extensions is ( i.e. , how many of the new concepts belonged to both extensions ) .
as pointed out in the introduction , the equivalence test can not deal with situations where different knowledge engineers use different names for concepts , or model on different levels of granularity .
we want to find out whether using unifiability rather than equivalence finds more cases of overlapping concepts . of course ,
in the case of unification one may also obtain false positives , i.e. , pairs of concepts that are unifiable , but are not meant to represent the same ( intuitive ) concept .
it is also important to find out how often this happens .
another problem to be dealt with in this application is the development of heuristics for choosing the pairs of concepts to be tested for unifiability and for deciding which concept names are turned into variables .
franz baader .
characterizations of unification type zero . in n.
dershowitz , editor , _ proceedings of the 3rd international conference on rewriting techniques and applications _ , volume 355 of _ lecture notes in computer science _ , pages 214 , chapel hill , north carolina , 1989 .
springer - verlag .
franz baader .
terminological cycles in a description logic with existential restrictions . in georg gottlob and toby
walsh , editors , _ proc . of the 18th int .
joint conf . on artificial intelligence ( ijcai 2003 ) _ , pages 325330 , acapulco , mexico , 2003 .
morgan kaufmann , los altos .
franz baader , sebastian brandt , and carsten lutz . pushing the @xmath0 envelope . in leslie pack kaelbling and alessandro saffiotti , editors , _ proc . of the 19th int .
joint conf . on artificial intelligence ( ijcai 2005 ) _ ,
pages 364369 , edinburgh ( uk ) , 2005 .
morgan kaufmann , los altos .
franz baader and ralf ksters .
matching in description logics with existential restrictions . in _ proc .
of the 7th int .
conf . on principles of knowledge representation and reasoning ( kr 2000 )
_ , pages 261272 , 2000 .
franz baader and ralf ksters .
unification in a description logic with transitive closure of roles . in robert nieuwenhuis and andrei
voronkov , editors , _ proc . of the 8th int .
conf . on logic for programming and automated reasoning ( lpar 2001 )
_ , volume 2250 of _ lecture notes in artificial intelligence _ , pages 217232 , havana , cuba , 2001 .
springer - verlag .
franz baader , ralf ksters , and ralf molitor .
computing least common subsumers in description logics with existential restrictions . in _ proc .
of the 16th int .
joint conf . on artificial intelligence
( ijcai99 ) _ , pages 96101 , 1999 .
franz baader and barbara morawska .
unification in the description logic @xmath0 . in ralf
treinen , editor , _ proc . of the 20th int .
on rewriting techniques and applications ( rta 2009 ) _ , volume 5595 of _ lecture notes in computer science _ , pages 350364 .
springer - verlag , 2009 .
sebastian brandt .
polynomial time reasoning in a description logic with existential restrictions , gci axioms , and
what else ? in ramon lpez de mntaras and lorenza saitta , editors , _ proc . of the 16th eur .
conf . on artificial intelligence ( ecai 2004 ) _ , pages 298302 , 2004 .
james r. campbell , alejandro lopez osornio , fernan de quiros , daniel luna , and guillermo reynoso .
semantic interoperability and snomed ct : a case study in clinical problem lists . in k.a .
kuhn , j.r .
warren , and t .- y .
leong , editors , _ proc . of the 12th world congress on health ( medical ) informatics ( medinfo 2007 ) _ , pages 24012402 .
ios press , 2007 .
jean - pierre jouannaud and claude kirchner . solving equations in abstract algebras : a rule - based survey of unification . in j .-
lassez and g. plotkin , editors , _ computational logic : essays in honor of a. robinson_. mit press , cambridge , ma , 1991 .
yevgeny kazakov and hans de nivelle .
subsumption of concepts in @xmath28 for ( cyclic ) terminologies with respect to descriptive semantics is pspace - complete . in _ proc . of the 2003 description logic workshop ( dl 2003)_. ceur electronic workshop proceedings , http://ceur-ws.org/vol-81/ , 2003 .
hector j. levesque and ron j. brachman . a fundamental tradeoff in knowledge representation and reasoning . in ron
j. brachman and hector j. levesque , editors , _ readings in knowledge representation _ , pages 4170 .
morgan kaufmann , los altos , 1985 .
alan rector and ian horrocks .
experience building a large , re - usable medical ontology using a description logic with transitivity and concept inclusions . in _ proceedings of the workshop on ontological engineering , aaai spring symposium ( aaai97 ) _ , stanford , ca , 1997 .
aaai press . | the description logic @xmath0 has recently drawn considerable attention since , on the one hand , important inference problems such as the subsumption problem are polynomial . on the other hand , @xmath0 is used to define large biomedical ontologies .
unification in description logics has been proposed as a novel inference service that can , for example , be used to detect redundancies in ontologies .
the main result of this paper is that unification in @xmath0 is decidable .
more precisely , @xmath0-unification is np - complete , and thus has the same complexity as @xmath0-matching .
we also show that , w.r.t .
the unification type , @xmath0 is less well - behaved : it is of type zero , which in particular implies that there are unification problems that have no finite complete set of unifiers . |
pioneering experiments on ultracold gases of atoms trapped in optical lattices allowed for a direct observation of quantum many - body phenomena , such as the quantum phase transition from mott phase to superfluid phase.@xcite optical lattices are realized by counterpropagating laser beams , which form a periodic potential.@xcite the bosonic particles located on the optical lattice gain kinetic energy when tunneling through the potential wells of neighboring sites of the periodic potential and they exhibit a repulsive interaction when a lattice site is occupied by more than one atom .
a condensate of ultracold atoms can be driven from superfluid phase to mott phase by gradually increasing the intensity of the laser beams .
the potential wells of the optical lattice are shallow for low laser - beam intensity .
thus the bosonic particles can overcome the barrier easily and are delocalized on the whole lattice .
however , for large intensity of the laser beams the potential wells are deep and there is little probability for the atoms to tunnel from one lattice site to another .
this physical behavior can be described by the bose - hubbard ( bh ) model @xcite provided the gas of ultracold atoms is cooled such that only the lowest bloch band of the periodic potential has to be taken into account.@xcite the ground state of the bh model is superfluid when the local on - site repulsion between the atoms is small in comparison to the nearest - neighbor hopping strength whereas it is a mott state for integer particle density and large on - site repulsion compared to the hopping strength . due to these characteristics of the bh model
the depth of the potential wells in optical lattices can be associated directly with the ratio of the on - site repulsion and the hopping strength .
ultracold atoms confined in optical lattices provide a very clean experimental realization of a strongly correlated many - body problem and the internal physical processes are well understood in comparison to conventional condensed - matter systems .
there is large experimental control over the system parameters , such as the particle number , lattice size , and depth of the potential wells .
furthermore the sites of the optical lattice can be addressed individually due to the mesoscopic scale of the lattice.@xcite the quantum phase transition from mott phase to superfluid phase has been first observed experimentally for ultracold rubidium atoms trapped in a three - dimensional optical lattice @xcite and subsequently as well in optical lattices of two dimensions.@xcite the corresponding theoretical model , the two - dimensional ( 2d ) bh model , has already been investigated to some detail in literature . the phase diagram , which describes the quantum phase transition from mott phase to superfluid phase , has been investigated thoroughly at the mean - field level ( possibly including gaussian - fluctuation corrections).@xcite more accurate results for the phase diagram from quantum monte carlo @xcite ( qmc ) simulations , variational approaches,@xcite and strong - coupling perturbation theory@xcite are also available .
the phase diagram for arbitrary integer fillings has been obtained recently using the so - called diagrammatic process chain approach.@xcite spectral functions of the two - dimensional bh model have been evaluated within a strong - coupling approach @xcite and a variational mean field approach.@xcite in the present paper we evaluate the border of the quantum phase transition from mott phase to superfluid phase for the first two mott lobes by means of the variational cluster approach ( vca),@xcite and show that this method provides quite accurately the boundaries of the mott phase , as compared with more demanding qmc simulations and perturbative expansions .
in addition , we study in detail the spectral functions of the two - dimensional bh model in both the first and the second mott lobe , which require computing the green s function in real frequency domain . we also present the densities of states and momentum distributions corresponding to the spectral functions . finally ,
as a technical point , we present an extension of the so - called @xmath0-matrix formalism , which has been originally proposed for fermionic ( anticommutator ) green s functions,@xcite to bosonic ( commutator ) green s functions.@xcite as we show below , this extension is nontrivial due to the nonunitary nature of the bogoliubov transformation for bosonic particles .
this paper is organized as follows . in sec .
[ sec : model ] the bh model is introduced .
section [ sec : method ] contains a short review on the variational cluster approach and the extension of the @xmath0-matrix formalism .
section [ sec : results ] is devoted to the spectral properties of the bh model in two dimensions . here
the phase diagram , spectral functions , densities of states , and momentum distributions are presented . finally , we summarize and conclude our findings in sec . [
sec : conclusion ] .
the ( grand - canonical ) hamiltonian of the bh model @xcite is given by @xmath1 where @xmath2 is the nearest - neighbor hopping strength , @xmath3 is the local on - site repulsion , and @xmath4 is the chemical potential . the angle brackets in the first part of the hamiltonian specify to sum over pairs of nearest neighbors ( each pair counted once ) . the operator @xmath5 creates a particle at lattice site @xmath6 whereas @xmath7 annihilates a particle at site @xmath6 .
the total particle number @xmath8 is conserved , since @xmath9 = 0 $ ] .
the particles of the bh model are of bosonic character and thus the commutation relation @xmath10 = \delta_{ij}$ ] is satisfied . the first term of the hamiltonian models the hopping of a particle from lattice site @xmath11 to lattice site @xmath6 .
the second part describes the local on - site repulsion , which remains zero when a lattice site is unoccupied or occupied by only one particle .
however , it increases proportional to @xmath3 for each additionally added particle .
we consider the on - site repulsion @xmath3 as unit of energy .
the third part of the hamiltonian is necessary to perform calculations in grand - canonical ensemble , where the chemical potential @xmath4 controls the total particle number of the system .
we use vca@xcite to evaluate the phase diagram and the spectral functions of the 2d bh model .
vca is a variational extension of the cluster perturbation theory @xcite and is based on the self - energy functional approach ( sfa ) which has been originally proposed for fermionic systems by m. potthoff.@xcite vca has been extended to bosonic systems as well.@xcite the sfa is based on the fact that dyson s equation for the exact green s function is recovered at the stationary point of the grand potential @xmath12 $ ] considered as a functional of the self - energy @xmath13
. thus @xmath13 corresponds , at the stationary point , to the real physical self - energy .
the self - energy functional @xmath12 $ ] can not be evaluated directly as it contains the legendre transform @xmath14 $ ] of the luttinger - ward functional.@xcite however , the functional @xmath14 $ ] just depends on the interaction term of the hamiltonian , i.e. , on the second term of eq.([eq : bhm ] ) , and is thus equivalent for all hamiltonians which share a common interaction part . due to this property @xmath14 $ ] can be eliminated from the expression of the self - energy functional @xmath12 $ ] . for this purpose
an exactly solvable , so - called `` reference , '' system @xmath15 is constructed , which must be defined on the same lattice and must have the same interaction part as the original system @xmath16 .
thus both the self - energy functional of the original system @xmath12 $ ] and the one of the reference system @xmath17 $ ] contain the same @xmath14 $ ] , which can be eliminated by comparison from the expressions of the two self - energy functionals .
this yields for bosonic systems @xcite @xmath18 & = \omega^\prime[{\mathsf{\sigma } } ] & & - { \mbox{tr}}\,\ln(-({\mathsf{g}}_0^{\prime\,-1}-{\mathsf{\sigma } } ) ) \nonumber \\ & & & + { \mbox{tr}}\,\ln(-({\mathsf{g}}_0^{-1}-{\mathsf{\sigma } } ) ) \ ; \mbox { , } \label{eq : num : om5}\end{aligned}\ ] ] where quantities with prime correspond to the reference system and @xmath19 is the free green s function .
the free green s function is defined as @xmath20 , where @xmath21 contains the hopping matrix and all other one - particle parameters of the hamiltonian except for the chemical potential @xmath4 , which is already treated separately in the definition .
the symbol @xmath22 denotes a summation over bosonic matsubara frequencies and a trace over site indices .
the self - energy functional @xmath12 $ ] given by eq.([eq : num : om5 ] ) is exact . in order to be able to evaluate the functional
, the search space of the self - energy @xmath13 has to be restricted,@xcite which consists in an approximation .
more precisely , the functional @xmath12 $ ] is evaluated for the subset of self - energies available to the reference system @xmath15 , see fig.[fig : num : vca ] .
practically , this is achieved by varying the single - particle parameters of the reference hamiltonian in order to find the stationary point of the grand potential .
thus the functional @xmath12 $ ] becomes a function of the set @xmath23 of single - particle parameters of @xmath24 @xmath25 leading to the stationary condition @xmath26 is restricted to self - energies @xmath27 which are accessible via the reference system @xmath15 .
( b ) lattice decomposition of a square lattice into @xmath28 site clusters .
, scaledwidth=48.0% ] in vca the reference system is given by the decomposition of the total system into identical clusters , see fig.[fig : num : vca ] . in order to implement the @xmath0-matrix approach
, we solve each cluster by means of the band lanczos method.@xcite the green s function of the total system is obtained via the relation @xmath29 which can be deduced from the dyson equation of the total system @xmath30 and the reference system @xmath31 .
the self - energy can be eliminated and it follows that @xmath32 the expression in parenthesis defines the matrix @xmath33 with eq.([eq : num : gvca ] ) the grand potential @xmath34 can be rewritten as @xmath35 the decomposition of the @xmath36-site lattice into clusters of @xmath37 sites can be described by a superlattice .
the original lattice is obtained by attaching a cluster to each site of the superlattice.@xcite a partial fourier transform from superlattice indices to wave vectors @xmath38 , which belong to the first brillouin zone of the superlattice , yields the total green s function @xmath39 due to the diagonality of @xmath40 in the superlattice indices its partial fourier transform does not depend on @xmath38 .
the matrices in eq.([eq : num : cpteqfourier ] ) are now defined in the space of cluster - site indices and are thus of size @xmath41 .
the @xmath36 wave vectors @xmath42 from the brillouin zone of the total lattice can be expressed as @xmath43 where @xmath44 belongs to both the reciprocal superlattice and the first brillouin zone of the total lattice.@xcite the frequency integration implicit in the expression for the grand potential , given in eq.([om ] ) , can be carried out analytically , yielding at zero temperature@xcite @xmath45 where @xmath46 and @xmath47 are the poles of the cluster green s function and total green s function , respectively .
the number of clusters @xmath48 is denoted as @xmath49 .
the poles @xmath46 of the cluster green s function can be readily obtained from the lanczos method , whereas the poles of the total green s function @xmath47 can be evaluated with the so - called @xmath0-matrix formalism , which was originally proposed for fermionic green s functions.@xcite here , we extend this formalism to the generic case , i.e. , we include bosonic green s functions .
as we will see , this extension is nontrivial , since it involves non - unitary transformations . for zero temperature
, the cluster green s function reads @xcite @xmath50 where @xmath51 is the ground state of the @xmath52 particle system , @xmath53 is its ( grand - canonical ) energy , and @xmath54 ( @xmath55 ) for bosonic ( fermionic ) green s functions .
the first term on the right - hand side of eq.([num : gf ] ) describes single - particle excitations from the @xmath52 particle ground state and can thus be referred to as particle term , whereas the second part corresponds to single - hole excitations and can be called hole term . inserting the identity @xmath56 into each part of eq.([num : gf ] ) ,
where @xmath57 are the eigenvectors of the reference hamiltonian with corresponding eigenvalues @xmath58 , yields the lehmann representation of the green s function @xmath59 which can be cast into the form @xmath60 in eq.([num : gfqmatrix ] ) , we have introduced the following notation : @xmath61 @xmath62 and @xmath63 where @xmath64 is the hilbert space of an @xmath65 particle system . with @xmath66 the cluster green s function
can be written in matrix notation @xmath67 with the help of this expression the vca green s function eq.([eq : num : cpteqfourier ] ) can be rewritten as @xmath68^{-1 } { \mathsf{s}}\,{\mathsf{q}}^\dagger \nonumber \\ & = { \mathsf{q } } \frac{1 } { { \mathsf{g}}^{\prime\,-1 } - { \mathsf{s}}\,{\mathsf{q}}^\dagger\,{\mathsf{v}}\,{\mathsf{q } } } { \mathsf{s}}\,{\mathsf{q}}^\dagger \;\mbox { , } \label{eqn : num : gvcarewritten}\end{aligned}\ ] ] where in the third step we expanded the fraction in a taylor series .
the matrix @xmath69 is diagonal and contains the poles of the cluster green s function @xmath40 , see eq.([eq : num : gsmall ] ) .
it can be written as @xmath70 with @xmath71 . plugging this into eq.([eqn : num : gvcarewritten ] )
yields @xmath72 we introduce the matrix @xmath73 .
this matrix can be diagonalized as @xmath74 , where @xmath75 is a diagonal matrix containing the eigenvalues of @xmath76 and @xmath77 is the matrix of the eigenvectors of @xmath76 .
the eigenvalue equation of the matrix @xmath76 can be rewritten as @xmath78 , where @xmath79 is the inverse of @xmath77 and not its transpose as @xmath76 is a non - symmetric matrix . from
that we obtain @xmath80 therefore ,
the poles of the total green s function @xmath81 in eq.([eqn : num : gqmatrixlambda ] ) are the eigenvalues of the matrix @xmath76 .
the matrices @xmath81 and @xmath82 are defined on the space of cluster - site indices .
thus @xmath81 and @xmath82 are of size @xmath41 and depend on the wave vector @xmath83 , see eq.([eq : num : cpteqfourier ] ) .
the matrix @xmath0 is of size @xmath84 , where @xmath85 is the dimension of the krylov space generated in the band lanczos method . due to the dependence of @xmath82 on @xmath83 the diagonalization of the matrix @xmath76 yields
@xmath85 eigenvalues @xmath86 , which are used in eq.([eqn : num : omintzero ] ) .
the diagonalization has to be repeated for all wave vectors @xmath83 . with that the grand potential @xmath34 can be evaluated .
the crucial point is that for bosonic green s functions , the entries of the diagonal matrix @xmath87 can be both @xmath88 as well as @xmath89 , see eq.([eq : num : sdef ] ) . therefore , the eigenvalue problem is not symmetric.@xcite the factorization of the total lattice into clusters breaks the translational symmetry of the lattice .
hence the total green s function would depend on two wave vectors @xmath42 and @xmath90 , which is certainly not correct for a periodic lattice .
this has to be circumvented by a periodization prescription that provides a total green s function @xmath91 depending only on one wave vector @xmath42 .
the periodization prescription proposed in ref .
( greens - function periodization ) reads as follows : @xmath92 where @xmath42 is a wave vector of the total lattice and @xmath93 refers to lattice sites @xmath94 of the cluster .
the wave vectors @xmath38 in eq.([eq : num : cpteqfourierfull ] ) can be replaced by the total wave vectors @xmath42 as they just differ by a reciprocal superlattice wave vector , see eq.([eqn : kdecomposition ] ) . with eqs.([eqn : num : gqmatrixlambda ] ) and ( [ eqn : m ] ) the periodized green s function can be rewritten in matrix notation @xmath95 where the vector @xmath96 and its adjoint @xmath97 contain @xmath37 plane waves @xmath98 there exists as well an alternative periodization prescription where the self - energy @xmath13 is periodized.@xcite this self - energy periodization should prevent spurious gaps , which arise in the spectral function .
however , at least for fermion systems , this procedure yields spurious metallic bands in the mott phase for arbitrarily large @xmath3 . since we do not observe any spurious gaps in the spectral function of the 2d bh model we use the periodization on the green s function defined in eq.([eq : num : cpteqfourierfull ] ) . with the wave - vector
resolved green s function of the total system @xmath99 we are able to calculate the single - particle spectral function @xmath100 the density of states @xmath101 and the momentum distribution @xmath102 the frequency integration can be evaluated directly by means of the @xmath0-matrix formalism , which yields a sum of the residues of the green s function , see eq.([eqn : greenfunctionmatrix ] ) , corresponding to negative poles @xmath103 , @xmath104
the bh model exhibits a quantum phase transition from a mott to a superfluid phase when the ratio between the hopping strength and the on - site repulsion @xmath105 is increased or when particles are added to or removed from the system .
the mott phase is characterized by an integer particle density , a gap in the spectral function and zero compressibility.@xcite , scaledwidth=48.0% ] the first two mott lobes of the 2d bh model obtained by means of vca are shown in fig.[fig : pd ] .
we used the chemical potential @xmath106 as variational parameter , which ensures a correct particle density of the total system.@xcite in contrast to the one - dimensional results @xcite the mott lobes of the 2d bh model are round shaped .
the gray shaded area in fig.[fig : pd ] presents the phase boundaries calculated within the process chain approach by n. teichmann _ et al . _ in refs . and , which are basically identical to the qmc results by b. capogrosso - sansone _ et al .
_ , see ref . .
the agreement is quite good for small hopping .
however , vca seems to overestimate the critical value of the hopping @xmath107 , which determines the tip of the mott lobe . for the critical hopping of the first mott lobe ,
we obtain approximately @xmath108 and for the second one @xmath109 .
latest process chain approach , @xcite qmc ( ref . ) and strong - coupling perturbation theory @xcite results yield @xmath110 and @xmath111 for the critical parameter of the first and second mott lobe , respectively .
the spectral functions @xmath112 and the densities of states @xmath113 for parameters of the first mott lobe are shown in fig.[fig : spectrallobe1 ] . , left column , and density of states @xmath113 , right column , in the first mott lobe for the parameters ( a ) @xmath114 , @xmath115 , ( b ) @xmath116 , @xmath117 and ( c ) @xmath118 , @xmath119 .
the captions of the subfigures refer to the marks in fig.[fig : pd].[fig : spectrallobe1],scaledwidth=48.0% ] the spectral function is displayed on the conventional path around the brillouin zone @xmath120 over @xmath121 to @xmath122 and back to @xmath123 , and we use an artificial imaginary - frequency broadening @xmath124 .
a peculiarity of bosonic systems is that the hole band of the spectral function has negative spectral weight whereas the particle band has positive spectral weight .
this follows from the definition of the bosonic green s function which has a negative sign in front of the hole term , see eq.([num : gf ] ) . in the figures we always plot the absolute value of the spectral function .
the local density of states is defined as a wave - vector summation of @xmath112 , see eq.([eq : spe : dos ] ) .
therefore we observe a negative peak in the density of states , which corresponds to the hole band of the spectral function . for bosonic green s functions
the density of states is not a probability distribution , as it contains negative values .
taking the absolute value would yield an all positive density of states , however , it would not be normed and is thus no probability distribution either . for small hopping , the gap in the spectral function is large and the bands are rather flat , i.e. , the width of the bands is small , see fig.[fig : spectrallobe1 ] .
the corresponding density of states contains two well - separated peaks . for increasing hopping , the gap of the spectral function is decreasing and the width of the bands is increasing .
pursuant to the spectral function , the peaks in the density of states become broader for increasing hopping .
the intensity of the two bands is almost constant for small hopping independent of the wave vector @xmath42 , whereas for large hopping a large intensity can be observed at @xmath125 .
the boundaries of the mott lobes correspond to the chemical potential of the state with one additional particle ( hole ) , which is obtained directly from the single - particle ( single - hole ) minimum excitation energy .
for this reason , we evaluate the phase diagram in fig .
[ fig : pd ] by taking the minimal gap of the spectral function for each @xmath105 , which always occurs at @xmath125 .
the spectral functions and densities of states in the second mott lobe corresponding to the marks iv and v in fig.[fig : pd ] are shown in fig.[fig : spectrallobe2 ] . , left column , and density of states @xmath113 , right column , in the second mott lobe for the parameters ( a ) @xmath126 , @xmath127 and ( b ) @xmath128 , @xmath129 .
the captions of the subfigures refer to the marks in fig.[fig : pd].[fig : spectrallobe2],scaledwidth=48.0% ] qualitatively they are very similar to the spectral functions and densities of states in the first mott lobe .
particularly , the intensity distribution of the bands seem to be strongly related . yet
the peaks of the density of states are larger due to the twice as large particle density within the second mott lobe and thus the absolute value of the spectral weight in the second mott lobe is larger than the one in the first mott lobe .
the momentum distribution @xmath130 corresponding to the spectral functions in the first and second mott lobe are shown in fig.[fig : nofk ] . in ( a ) the first mott lobe and ( b ) the second mott lobe .
the roman numerals in the legends refer to the parameters marked in fig.[fig : pd].[fig : nofk],scaledwidth=48.0% ] the particle density in the first mott lobe is one and thus @xmath130 is centered around one in fig.[fig : nofk ] . for the second mott lobe @xmath130 is centered around two , see fig.[fig : nofk ] .
the particle density @xmath130 is extremely flat for small hopping whereas it is peaked at @xmath125 for large hopping , which is already a precursor for the bose - einstein condensation where all particles condense in the @xmath125 state .
this behavior directly reflects the intensity distribution of the bands in the spectral function .
there is excellent quantitative agreement between our vca results for the momentum distribution and results obtained by means of qmc and a strong - coupling perturbation theory with scaling ansatz,@xcite see fig.[fig : nofkcomparison ] . obtained by means of vca and a strong - coupling perturbation theory with scaling ansatz.@xcite the data identified with the letters sc correspond to the strong - coupling results .
the relative deviations between vca results and strong - coupling results with scaling ansatz are shown in ( b ) .
the roman numerals in the legends refer to the parameters marked in fig.[fig : pd].[fig : nofkcomparison],scaledwidth=48.0% ] we compare the momentum distributions for the parameters i ( @xmath114 ) and ii ( @xmath116 ) , and observe that the relative deviations between our vca results and an approach obtained by combining strong - coupling perturbation theory with a scaling ansatz @xcite are almost zero for small hopping @xmath114 and less than one percent for medium hopping @xmath116 .
this latter methods is certainly more accurate than vca in the evaluation of the momentum distribution .
however , it should be mentioned that the information about the critical point ( critical exponents and critical hopping strength @xmath107 ) have to be inserted `` by hand , '' in order to optimize the results .
this information , in turn , must be extracted , e. g. , from a qmc calculation . on the other hand ,
our vca results are obtained directly without the need to introduce external parameters .
in the present paper , we presented and discussed results obtained within the variational cluster approach for the spectral properties of the two - dimensional bose - hubbard hamiltonian .
this is a minimal model to describe bosonic ultracold atoms confined in optical lattices,@xcite and it undergoes a quantum phase transition from a mott to a superfluid phase depending on the chemical potential @xmath4 , and the ratio between the hopping strength and the on - site repulsion @xmath105 . in particular , we determined the first two mott lobes of the phase diagram and found reasonable agreement with essentially exact results from qmc simulations and from the process chain approach . in particular , the variational cluster approach yields very good results for the phase boundaries apart from the region close to the lobe tip . here
, strong - coupling expansions and qmc calculations are , clearly , much more accurate .
yet it should be emphasized that the computational effort is considerably lower for vca than for qmc .
furthermore , we evaluated spectral functions in the first and second mott lobe . an important aspect of vca is that the green s function of the system is obtained directly in the real frequency domain , which allows for a direct calculation of the spectral function .
on the other hand , qmc quite generally provides correlation functions in imaginary time .
imaginary - time correlation functions have to be analytically continued to real frequencies , which is a very ill - conditioned problem , as the data contain statistical errors . in qmc
this analytical continuation is best carried out by means of the maximum entropy method .
a very accurate dispersion ( without spectral weight ) has been also obtained by a strong - coupling expansion.@xcite the intensity distribution of the spectral weight is similar for the spectral functions of both mott lobes , leading to an evenly distributed spectral weight for small hopping strengths and to a distribution sharply peaked at @xmath131 for large hopping strengths .
the latter indicates a precursor to the bose - einstein condensation occurring above a certain critical hopping .
we also evaluated the densities of states and momentum distributions corresponding to the calculated spectral functions .
we compared our vca results for the momentum distribution with strong - coupling perturbation - theory results , where a scaling ansatz has been used , and found excellent quantitative agreement .
finally , as a technical point , we extended the @xmath0-matrix formalism to deal with bosonic green s functions , which , in contrast to the fermionic case , produces a non - symmetric eigenvalue problem .
we are grateful to n. teichmann for providing the process chain approach data of the phase diagram used in fig.[fig : pd ] .
we thank j. k. freericks for sending us the self - consistently solved strong - coupling results with scaling ansatz shown in fig.[fig : nofkcomparison ] .
we made use of parts of the alps library ( ref . ) for the implementation of lattice geometries and for parameter parsing .
m.k . wants to thank p. pippan for fruitful discussions .
we acknowledge partial financial support from the austrian science fund ( fwf ) under the doctoral program `` numerical simulations in technical sciences ''
w1208-n18 ( m.k . and w.v.d.l . )
and under project no .
p18551-n16 ( e.a . ) . | spectral properties of the two - dimensional bose - hubbard model , which emulates ultracold gases of atoms confined in optical lattices , are investigated by means of the variational cluster approach .
the phase boundary of the quantum phase transition from mott phase to superfluid phase is calculated and compared to recent work . moreover the single - particle spectral functions in both the first and the second mott lobe are presented and the corresponding densities of states and momentum distributions are evaluated . a qualitatively similar intensity distribution of the spectral weight can be observed for spectral functions in the first and the second mott lobe . |
star - forming disc galaxies like the milky way must accrete @xmath0 1 @xmath2 of fresh gas each year ( see * ? ? ? * and references therein ) and have built their discs gradually over the last 10 gyr ( e.g. * ? ? ?
a central question is the origin of the accreted gas and how this gas reaches the thin disc whitin which the process of star formation takes place .
the virial - temperature corona , in which disc galaxies are embedded , is the only reservoir of baryons capable of sustaining an accretion rate of @xmath3 for a hubble time .
we present the results of a set of grid - based hydrodynamical simulations supporting the idea that the gas needed by the disc to form stars is drawn from this corona .
coronae of disc galaxies are similar in many respects to the hot atmospheres of giant elliptical galaxies and galaxy clusters , but with lower gas temperature and density ( e.g. * ? ? ?
* ; * ? ? ?
* ) . as the hot gas of these more massive systems ,
the coronal gas is unlikely to fragment into clouds via thermal instability @xcite , but it is expected to cool monolithically and feed the central black hole rather than produce an extended cold disc in which stars can form . however ,
if the gas needed to feed star formation has to be drawn from the corona , a mechanism that makes the hot gas accrete onto the disc must be at work .
there is abundant evidence that star formation in galaxies like the milky way powers a galactic fountain : ejection of gas from the mid - plane by supernova explosions @xcite . through the fountain a significant fraction ( from 10 to 25 % ) of the whole hicontent of the galaxy
is carried into its halo ( see * ? ? ?
* and references therein ) . in this work , supported by several lines of argument , we hypothesise that the transfer of gas from the corona to the star - forming disc is effected by the hiclouds ejected by the galactic fountain @xcite .
our hydrodynamical simulations suggest that the gas accretion proceeds through the following steps : ( i ) stripping of gas from fountain clouds by the corona as a result of the kelvin - helmholtz instability , ( ii ) mixing of the ( high metallicity ) stripped gas with a comparable amount of coronal gas in the turbulent wake of the clouds ; as a consequence of the mixing , the gas cooling time of the coronal gas is reduced to a value lower than the cloud s flight time , ( iii ) formation of knots of cold gas that accrete onto the disc in a dynamical time .
the stripped gas leads to condensation of coronal gas only if the mass - loss rate exceeds a critical value @xmath4 , determined by the physical properties of the cloud and the corona .
in addition , dimensional analysis suggests that the actual mass - loss rate must lie close to @xmath4 . in view of the proposed scenario and of these considerations
the aims of the simulations are two - fold : ( i ) to provide an estimate of the actual mass - loss rate @xmath5 for comparison with @xmath4 , ( ii ) to determine the critical ambient pressure ( dependent on metallicity ) above which the mass of cool gas increases with time through condensation in the wake . in figure [ nocool ] the evolution of the mass of gas at @xmath6 k ( accreted gas ) , for a simulation with physical conditions representative of galactic coron ,
is shown ( see caption for details ) .
in particular , in the left panel of the figure , the mass loss of a cloud is displayed and the agreement between the simulation and the analytical prediction is remarkable . from the right panel of figure [ nocool ] it has been possible to derive an estimate of the global accretion rate ( @xmath7 ) which is of the same order as that required to feed star formation . by comparing the figures in the two panels
it is apparent that the cooled mass of coronal gas is comparable to the mass lost by the cloud . in other words ,
when the cooling is switched on the evaporation of some cloud mass produces an accretion of roughly the same amount of coronal mass .
we have used hydrodynamical simulations to check whether the interaction between the galactic - fountain clouds , powered by a star - forming disc , and the virial - temperature corona , in which disc galaxies are embedded , could lead the coronal gas to cool promptly in the clouds wake and accrete onto the disc to feed star formation .
the results of our analysis , described in detail in @xcite , can be summarized as follows : * the interaction between the galactic fountain and the hot corona causes the fountain clouds to be stripped of some of their gas .
the simulations provide a reliable estimate of the mass - loss rate and confirm that its value lies close to @xmath4 , in agreement with analytical expectations ; * the condensation of coronal gas in the cloud s wake prevails over evaporation if the metallicity and/or pressure of the corona are high enough . in our simulations
this happens for likely parameters of the coron in disc galaxies , condensation is present also for densities as low as @xmath8 , provided that [ fe / h ] @xmath9 ; * the derived global accretion rate is of the same order as that required to feed star formation
. therefore the condensation of the hot corona seems a viable mechanism to sustain the star formation rate in star - forming galaxies .
we used cpu time assigned under the inaf - cineca agreement 2008 - 2010 .
f. marinacci gratefully acknowledges the support from the marco polo program by university of bologna and the hospitality of the rudolf peierls centre for theoretical physics . | star - forming disc galaxies such as the milky way need to accrete @xmath0 1 @xmath1 of gas each year to sustain their star formation .
this gas accretion is likely to come from the cooling of the hot corona , however it is still not clear how this process can take place .
we present simulations supporting the idea that this cooling and the subsequent accretion are caused by the passage of cold galactic - fountain clouds through the hot corona .
the kelvin - helmholtz instability strips gas from these clouds and the stripped gas causes coronal gas to condense in the cloud s wake . for likely parameters of the galactic corona and of typical fountain clouds we obtain a global accretion rate of the order of that required to feed the star formation . address = dipartimento di astronomia , universit di bologna , via ranzani 1 , 40127 bologna , italy .
address = rudolf peierls centre for theoretical physics , oxford university , keble road , oxford ox1 3np , uk .
address = dipartimento di astronomia , universit di bologna , via ranzani 1 , 40127 bologna , italy .
address = dipartimento di astronomia , universit di bologna , via ranzani 1 , 40127 bologna , italy .
address = dipartimento di astronomia , universit di bologna , via ranzani 1 , 40127 bologna , italy .
address = inaf - osservatorio astronomico di bologna , via ranzani 1 , 40127 , bologna , italy . |
if @xmath0 is a not everywhere defined operator on a locally convex vector space @xmath1 and @xmath2 is a lipschitz curve system on the riemann sphere then @xmath3 if @xmath4 and the resolvent operators of @xmath0 with poles from @xmath2 applied @xmath5 times to @xmath6 give a continuous function on @xmath7 . the slowly growing holomorphic functions on the complement of the lipschitz curve system @xmath2 form the algebra @xmath8 .
this algebra has a natural filtration : @xmath9 if @xmath10 is a superposition of rational functions with a pole in @xmath2 with multiplicity at most @xmath5 . in case
@xmath11 it means that @xmath12 where @xmath13 are finite borel measures on @xmath2 .
having such representations we may define the operator @xmath14 on @xmath15 if @xmath9 .
we do that by substituting @xmath0 into @xmath16 formally . in the special case above @xmath17 the objective of this paper
is to elaborate the notions used above and demonstrate the viability of this functional calculus in arithmetical sense . that includes the statements of well - definedness , multiplicativity , and locality , which hold under realistic conditions .
inspiration for this work came from three sources : one is the author s interest in taking square roots along special curves in locally convex algebras , cf .
@xcite ; another one is the classical theory of fractional powers of operators with radial spectral cuts , see komatsu @xcite , or the book of martnez carracedo and sanz alix @xcite ; and the third one is the very classical riesz functional calculus , cf .
dunford and schwartz @xcite , vii.3 .
we will see how these fit into our picture . for a comprehensive review of related matters
see the book of haase @xcite .
although this latter one concentrates on the sectorial case but it is exhaustive in that regard . in terms of the construction of the calculus outlined above , we can notice that the relationship between the function @xmath18 and the not everywhere defined operator @xmath14 is not immediate .
although we want @xmath14 to specialize to @xmath18 for scalars , we obtain @xmath14 through certain measures @xmath13 : @xmath19 this character also appears in the structure of the paper : sections 15 comprise a function theoretical discussion . in sections 1 and 2 we go through a short discussion of lipschitz curve systems and some elementary properties of integrals of cauchy type .
section 3 is about some finer properties .
the main purpose of this section is that we can formulate the statement of locality later in a slightly stronger form than it would be easier otherwise .
so despite its relative length one can skip this section without much consequence .
section 4 is a reminder to some elementary identities , which , in fact , will be generalized later , but they are included here for the sake of better readability .
section 5 is about normal forms of slowly growing holomorphic functions outside a lipschitz curve system @xmath2 .
practically , it says that internal distributions on @xmath2 ( whatever they are ) are in bijective correspondence to slowly growing holomorphic functions as boundary values , cf .
berenstein and gay @xcite , 3.6 .
the function theoretical part is relatively lengthy but it turns out that once we properly understand what we want to generalize the task itself is surprisingly easy . in sections 69 the functional calculus is actually discussed . in section 6
we restate the statements of section 4 in greater generality , this time applied to multivalued linear operators .
the key observation is that the resolvent identity remains valid , so we can generalize everything . in section 7
we define the function calculus and we discuss its basic properties . in section 8
we examine that how our functional calculus applies to fractional powers . in section 9
we mention some modifications which can be applied . in terms of conventions , in what follows `` measure '' means borel measure of bounded variation .
they are always considered on @xmath20 , @xmath21 , locally compact topological spaces unless noted otherwise .
we remark that such measures can be well approximated by discrete measures in weak sense , and integration is a question of sequential convergence . in the first part we consider complex measures and scalar integrals . but eventually , in the second part we also consider vector measures or rather vector integrals , depends on viewpoint .
one can take a look at e. g. @xcite in order to see how to develop a sufficient integration theory along such measures , but the reader is rather assumed to have working knowledge regarding vector integration .
i would like to thank markus haase for some useful comments .
let @xmath22\rightarrow\mathbb c$ ] be a simple continuous curve , @xmath23)$ ] .
we say that @xmath24 is a lipschitz curve if @xmath2 is locally the graph of a lipschitz function .
we say that @xmath24 is a short lipschitz curve if @xmath2 is actually the graph of a lipschitz function .
in the latter case we say that the vector @xmath25 is transversal in angle @xmath26 if for any @xmath27 , @xmath28 the angle between @xmath25 and the line through @xmath29 and @xmath30 is @xmath26 .
if we introduce a coordinate system @xmath31 such that its ordinate points into the direction of @xmath25 then we find that the curve has shape @xmath32 where @xmath33 . from the definition it is clear that being a lipschitz curve is independent of the parametrization . but bi - lipschitz parametrizations can always be chosen , ie . for which there exist constants @xmath34 such that @xmath35 .
for example , the arc - length parametrization is always like that .
it is not hard to see that a curve @xmath22\rightarrow\mathbb c$ ] is a lipschitz curve with a bi - lipschitz parametrization if its derivative exist in @xmath36 sense and its representative function @xmath37 can be chosen so that any @xmath38 $ ] has a neighborhood @xmath39 such that the set @xmath40 is contained in a bounded , closed convex set not containing @xmath41 . in what follows
we always consider lipschitz curves with bi - lipschitz parametrizations .
if @xmath42 $ ] then we define the set of curve directions at @xmath43 as @xmath44}}^{\overline{\hspace{5 mm } } } \frac{\gamma(t)-\gamma(t_0)}{|\gamma(t)-\gamma(t_0)|}}$ ] .
let @xmath23)$ ] .
going counterclockwise , the analytic boundary @xmath45 of @xmath46 in @xmath47 contains the endpoint @xmath48 , the left boundary @xmath49 , the endpoint @xmath50 , and the right boundary @xmath51 .
if @xmath52 is a subset of @xmath2 then we use the notation @xmath53 , and similarly with @xmath54 and @xmath55 .
in particular , if @xmath56 is open then the analytic boundary @xmath57 of @xmath58 in @xmath59 will be an open subset of the analytic boundary @xmath45 of @xmath46 in @xmath47 .
if @xmath60 is a continuous function then we can consider its maximal continuous extension @xmath61 to the analytic boundary of @xmath58 in @xmath59 such that the domain of the extension intersects the analytic boundary in an open subset . in that case
we define @xmath62 and @xmath63 such that @xmath64 , and @xmath65 , wherever @xmath61 makes sense .
[ lem : boundary ] suppose that @xmath59 is an open set , @xmath60 is holomorphic and it extends to the analytic boundary @xmath66 .
then , we claim , @xmath67 in distributional sense on @xmath59 , where the measure @xmath68 slightly prolonging @xmath24 , if necessary , we can forget about the endpoints .
furthermore , the question is local , and we can assume that @xmath24 is locally of shape @xmath69
. then we can define the functions @xmath70 locally . for such functions
the statement is a consequence of green s theorem , because the first derivatives extend to the analytic boundary continuously .
the general case follows from taking the limit @xmath71 .
we say that the simple curves @xmath72\rightarrow\mathbb c$ ] ( @xmath73 ) and the points @xmath74 ( @xmath75 ) form a curve system if the sets @xmath76)$ ] are pairwise disjoint except at endpoints , the @xmath74 are distinct and they do not lie in any of @xmath76)$ ] . then @xmath77)$ ] is the supporting set of the curve system .
the endpoints of the @xmath78 and the points @xmath74 are the nodes of the curve system , the elements of the @xmath79 are the regular points .
if the curves @xmath78 are lipschitz curves then we say that the curve system is a piecewise lipschitz curve system .
a curve @xmath22\rightarrow\mathbb c$ ] is a piecewise lipschitz curve if it can be cut into lipschitz pieces .
this happens if and only if @xmath80)$ ] is the supporting set of a piecewise lipschitz curve system .
we say that a curve system is a lipschitz curve system if it is piecewise lipschitz and @xmath76)\cap\gamma_j([a_j , b_j])\neq\emptyset$ ] implies that @xmath76)\cup\gamma_j([a_j , b_j])$ ] is the supporting set of a lipschitz curve , ie .
the components of the curve system join as pieces of lipschitz curves .
if it causes no confusion then we rather call the supporting set @xmath2 of a lipschitz curve system as the lipschitz curve system . from lemma [ lem : boundary
] one can immediately deduce [ lem : boundary2 ] suppose that @xmath2 is a lipschitz domain , @xmath59 is an open subset of @xmath47 , and @xmath10 is a holomorphic function on @xmath58 .
suppose that @xmath10 extends to the analytic boundary of @xmath58 in @xmath59 continuously .
then we claim , @xmath67 in distributional sense on @xmath59 , where in regular points the measure @xmath81 is locally given as @xmath82 such that @xmath83 is a local parametrization of @xmath2 , and @xmath81 extends to the additional , finitely many nodes of @xmath2 as the zero measure ( ie .
@xmath81 is a continuous measure ) .
the relationship of lemma [ lem : boundary ] and lemma [ lem : boundary2 ] is typical for what follows : if we have a result concerning lipschitz curves then it extends to lipschitz curve systems with appropriate modifications . using local charts
lipschitz curves and curve systems can be defined on any @xmath84 @xmath85-manifold @xmath86 .
if @xmath86 is oriented then left and right transversal directions , etc . can also be defined . for us
only the case @xmath87 , the riemann sphere , will be of importance .
an open subset @xmath88 is a lipschitz domain if there is a lipschitz curve system @xmath89 such that @xmath59 is a connected component of @xmath90 .
we can consider the analytic boundary @xmath91 of @xmath59 in @xmath92 .
the analytic boundary @xmath91 and the ordinary boundary @xmath93 has the same number of components , but the analytic boundary is richer , there is a surjective continuous map @xmath94 .
topologically , each component of @xmath91 is either a point or a circle .
in fact , for each connected component @xmath95 of @xmath91 we can find an oriented boundary curve @xmath96_{a\sim b}\rightarrow k_i$ ] , such that a. ) @xmath97_{a\sim b}$ ] is a possibly degenerate closed interval @xmath97 $ ] with the endpoints identified , b. ) the underlying curve @xmath98 is a piecewise lipschitz curve , although not necessarily simple , c. ) @xmath59 `` lies on the left '' of @xmath99 , in particular in case of a bounded @xmath59 it goes counterclockwise .
these oriented boundary curves are unique up to bi - lipschitz reparametrizations .
let @xmath22\rightarrow\mathbb c$ ] be a lipschitz curve .
we define the associated logarithm function as @xmath100 in fact , @xmath101 , where @xmath102 is the unique branch of logarithm such that the function vanishes at @xmath103 . regarding local behaviour , for @xmath104 , @xmath105 we see that @xmath106 where @xmath107 is an appropriate branch of @xmath108 , and @xmath109 is @xmath110 or @xmath111 if @xmath16 on the left or right side of @xmath24 , respectively .
+ + considering it seems reasonable to extend the logarithm function to @xmath112 by taking @xmath113 .
indeed , we do so but we do not really consider @xmath114 to be in the domain of @xmath115 but we use @xmath116 just as a special expression when a formula calls for it . notice that @xmath116 can be understood as the cauchy principal value of the defining integral of @xmath115 .
[ po : summary ] more generally , consider a finite borel measure @xmath117 on @xmath97 $ ] , and set @xmath118 then the function @xmath18 has the following properties : a. ) @xmath18 is holomorphic on @xmath46 , @xmath18 vanishes at @xmath103 .
b. ) @xmath119 .
c. ) @xmath18 is a function of class @xmath120 ( @xmath121 ) .
d. ) @xmath18 can be understood in distributional sense , and then @xmath122 in particular , if such functions @xmath18 are given then the measure @xmath117 can be recovered .
e. ) if @xmath123 $ ] and @xmath124 stays in a closed sector not containing any curve direction of @xmath24 at @xmath114 then @xmath125 f. ) suppose that @xmath126\cap(u , v)$ ] .
if @xmath10 continuously extends to the analytic boundary @xmath127 then @xmath128 where @xmath129 in particular , @xmath130 is continuous .
g. ) assume that @xmath131 $ ] and @xmath132 where @xmath130 is hlder - continuous on @xmath133 .
then @xmath10 extends to @xmath134 continuously . in particular , ( )
holds on @xmath133 .
generalizing the case of the logarithm function here it is reasonable to set @xmath135 one can see that @xmath136 is the cauchy principal value of @xmath137 .
again , this is a reasonable notation , but we do not consider @xmath83 to be in the domain of @xmath18 .
h. ) assume that @xmath138 $ ] and @xmath139 where @xmath130 is hlder - continuous on @xmath140 .
then @xmath141 where @xmath142 extends continuously to @xmath143
. similar statement holds around @xmath144 .
i. ) more generally , if @xmath130 is a piecewise hlder continuous function , then the picture is pretty much the same except @xmath18 may have logarithmic singularities around the discontinuities of @xmath130 .
\(a ) and ( b ) are immediate from the definition .
( c ) : on @xmath145 the function @xmath18 is given as the convolution of the function @xmath146 and the compactly supported finite measure @xmath147 .
the function @xmath146 is of class @xmath120 ( @xmath121 ) , hence we see that @xmath18 is a function of class @xmath120 ( @xmath121 ) .
( d ) : this convolution can also be understood in distributional sense .
we know that the function @xmath148 is a fundamental solution of the @xmath149 operator ; from which @xmath150 follows .
( e ) follows from a simple estimate .
( f ) follows from lemma [ lem : boundary ] and point ( d ) .
( g ) : we can assume that @xmath24 is a short lipschitz curve with the direction @xmath124 transversal , and @xmath130 is hlder - continuous on @xmath97 $ ] .
then for @xmath151 , @xmath152 , @xmath153 we see that @xmath154 from that the local behaviour is immediate .
( h ) : if @xmath155 then @xmath18 extends continuously to the analytic boundary of @xmath90 for @xmath156 .
one sees this immediately by slightly prolonging @xmath24 at @xmath157 and extending @xmath130 as identically zero .
otherwise , we can write @xmath158 , so @xmath18 can be written as @xmath159 plus a function continuously extending to the analytic boundary for @xmath156 .
( i ) follows by cutting @xmath24 into pieces .
a convenient lemma in the other direction is : suppose that @xmath160)$ ] , and @xmath18 is a function on @xmath90 such that a. ) @xmath10 is holomorphic on @xmath46 , and @xmath10 vanishes at @xmath103 ; b. ) @xmath10 extends to a distribution on @xmath47 ; c. ) if @xmath105 then there are functions @xmath62 and @xmath63 coming from continuous extensions of @xmath10 to the analytic boundary on the left and right side of @xmath24 respectively , and @xmath161 d. ) for z=@xmath162 and @xmath163 there exists a direction @xmath124 which is not a curve direction at @xmath16 and @xmath164 then , we claim , @xmath165 let temporarily @xmath166 denote the right side of .
if @xmath166 is such a function then due to [ po : summary].f the distribution @xmath167 is a distribution supported at the endpoints of @xmath24 . then consider @xmath168 .
we see that @xmath169 is a rational function with singular support at the endpoints of @xmath24 at most , and one which vanishes at @xmath103 . also , @xmath149 vanishes on @xmath170 .
then , from liouville s theorem we see that @xmath171 .
due to [ po : summary].e and condition ( d ) here we see that @xmath172 has actually no singular support , consequently @xmath173 .
[ po : logref1 ] as an immediate consequence of this lemma we may deduce : a. ) for @xmath174 @xmath175 b. ) for @xmath176 @xmath177 [ po : logref2 ] in the special case when @xmath97=[-1,1]$ ] and @xmath178 it yields a. ) for @xmath174 @xmath179 b. ) for @xmath176 @xmath180 one can make similar statements concerning lipschitz curve systems . there
the role of endpoints will be played by the nodes of the curve system .
* a. * first , we prove some estimates .
suppose that @xmath181 is a measure on the set @xmath182 and @xmath183 $ ] is a measurable function .
then we define the straightening of @xmath10 along @xmath117 as the function @xmath184 $ ] given by @xmath185 simply speaking , the straightening operation rearranges the values of @xmath10 into a monotone decreasing function meanwhile it takes into respect that `` how much a given value occurs '' according to @xmath117 . if @xmath186 , then the value @xmath187 is discarded . using step - functions
it is easy to prove [ lem : pmp ] let @xmath188 be a measure , and @xmath189 be functions as above .
a. ) then @xmath190 b. ) if the pairing is monotone in the sense that for all @xmath191 and @xmath192 @xmath193 implies @xmath194 then we have equality .
we introduce some terminology , so we can formulate some technical statements about lipschitz curves more easily . if @xmath22\rightarrow\mathbb c$ ] is a short lipschitz curve then we define the sets @xmath195 as the set of global curve directions at @xmath162 and the set of forward directions of @xmath24 , respectively
. we can consider the closed cone of curve directions @xmath196 and the cone of forward directions @xmath197 .
let us denote the semi - angle of the cone @xmath197 by @xmath198 .
this is always smaller than @xmath199 for such a short lipschitz curve .
one can see that @xmath200}$ ] is the set of curve directions at @xmath162 .
for @xmath201 we use the notation @xmath202 . [
lem : estkey ] suppose that @xmath24 is a short lipschitz curve as above , @xmath203 .
suppose that the ( minimal ) angle between the direction of @xmath25 and those of @xmath204 is @xmath205 .
a. ) let @xmath206 , @xmath207 .
let @xmath208 be the angle between the direction of @xmath25 and the axis of @xmath197 .
let @xmath209 . notice that @xmath210 .
let us set @xmath211 for @xmath212 $ ] and @xmath213 .
let @xmath214 be @xmath215 times the lebesgue measure restricted to @xmath216
. then @xmath217 .
b. ) we claim that @xmath218 c. ) if the arc - length of @xmath24 is @xmath219 then @xmath220 the latter two inequalities are monotone in @xmath198 and @xmath205 , so they apply if the semi - angle of @xmath197 is at most @xmath198 , and the angle between the direction of @xmath25 and those of @xmath204 is _ at least _ @xmath205 .
a. ) we can make the assumptions that @xmath221 , the axis of @xmath197 is the positive @xmath222-axis , @xmath223 , and @xmath24 is parametrized by its @xmath222 coordinate .
+ one can see from the picture that @xmath224 and @xmath225 . from this
the statement is immediate .
b. ) @xmath226 } |f(t)|^2\,|d\gamma(t)|\leq \int_{[0,\infty ) } |f(x)|^2 \frac{dx}{\cos\alpha}$ ] , and the latter term can be computed exactly . if @xmath227 then the latter integral is composed of two pieces and integration yields @xmath228 if @xmath229 then there is one piece and it yields @xmath230 , which can be estimated alike .
c. ) let @xmath231 .
. then one can see that @xmath233 if @xmath234 .
consequently , @xmath235 we estimate the last term .
doing that we can forget about the condition @xmath236 .
then the integral is maximal when @xmath237 is concentrated around @xmath238 , but the density relative to the lebesgue measure is at most @xmath215 .
hence we can estimate as @xmath239 using @xmath240 and @xmath241 we can continue by @xmath242 [ lem : double ] suppose that @xmath24 is a short lipschitz curve and @xmath25 is a transversal vector in angle @xmath26 .
let @xmath243 , @xmath42 $ ] .
set @xmath244 , @xmath245 .
then , we claim , @xmath246 } \frac1{|\gamma(t)-s|}\frac1{|\gamma(t)-r(s)|}|d\gamma(t)|\leq \frac{2\pi(c^{1/2}+c^{-1/2})}{|r(s)-s|\sin^2\theta}.\ ] ] we cut the interval @xmath97 $ ] into pieces @xmath247 $ ] and @xmath248 $ ] . for each piece
( with possible change of orientation ) point ( b ) of the previous lemma applies with the choice @xmath249 , @xmath250 .
adding up the estimates on the pieces and applying @xmath251 we obtain @xmath246 } \frac1{|\gamma(t)-s|^2}|d\gamma(t)|\leq\frac{2\pi}{\varepsilon\sin^2\theta};\qquad \int_{[a , b ] } \frac1{|\gamma(t)-r(s)|^2}|d\gamma(t)|\leq\frac{2\pi}{c\varepsilon\sin^2\theta}.\ ] ] applying the cauchy inequality we obtain the estimate @xmath252 . substituting @xmath253 the statement follows .
[ po : star0 ] now we generalize the statements above to lipschitz curve systems . if @xmath11 is a lipschitz curve system , @xmath254 then we can imagine the constituent curves of @xmath2 as follows : we have short lipschitz curves @xmath255\rightarrow\mathbb c$ ] ( @xmath73 ) such that @xmath256 , and the sets @xmath257 are pairwise disjoint .
we say that these curves form a star system around @xmath258 .
apart from these there may be other constituents of @xmath2 but these do not affect the local structure around @xmath258 .
we often consider star systems as above for the purposes of local analysis .
we generally denote the maximal cone semi - angle of the @xmath259 by @xmath198 , and we set @xmath260 .
[ lem : estkey3 ] suppose that we have a star system around @xmath258 as above .
assume that @xmath261 is an open cone from @xmath262 .
also assume that the angle between any vector in @xmath261 and any vector in @xmath263 is @xmath264 .
( see figure a. ) then , we claim , for any @xmath265 @xmath266 this is immediate from lemma [ lem : estkey].b and the cauchy inequality
. @xmath267 [ po : star ] consider a star system as earlier .
let @xmath268 , so that we can formulate the statement in simpler terms .
we can assume that the @xmath78 are listed such that sets @xmath257 come in counterclockwise direction .
suppose that @xmath269 are unit vectors such that @xmath270 points between @xmath271 and @xmath272 , and @xmath273 points between @xmath271 and @xmath274 .
in other words @xmath275 and @xmath276 close @xmath277 together from left and right .
we can assume that the angle between @xmath278 and @xmath275 is @xmath279 , the angle between @xmath275 and @xmath280 is @xmath281 , the angle between @xmath280 and @xmath276 is @xmath282 , the angle between @xmath276 and @xmath283 is @xmath279 .
we also assume that @xmath270 and @xmath273 are in angle @xmath284 from the axis of @xmath259 .
we define the `` sector '' @xmath285 as the open set between the curve @xmath277 and @xmath275 , and the `` sector '' @xmath54 as the open set between the curve @xmath277 and @xmath276 .
these sets are not completely well - defined , but they are quite well - defined in a sufficiently small neighborhood of @xmath258 , ie . locally .
so , we just use @xmath285 and @xmath54 without further ado , although what follows works only locally .
( see figure b. ) now , we can define a sort of reflection map @xmath286 as follows : take a point @xmath287 , then draw a line perpendicular to the axis of @xmath288 .
suppose that this line intersects @xmath277 at @xmath289 .
then we set @xmath290 , where @xmath187 is a positive constant .
one can show that if @xmath291 then the map indeed goes @xmath292 .
the very crude estimate @xmath293 also holds .
[ lem : estkey4 ] @xmath294 where @xmath295 from lemma [ lem : double ] we can estimate the term @xmath296 , from lemma [ lem : estkey3 ] we can estimate the terms @xmath297 directly .
regarding to the situation above we say that @xmath285 is a sector siding @xmath277 from the right , and @xmath54 is a sector siding @xmath277 from the left .
one can notice that if @xmath285 is siding @xmath277 then so do the sectors obtained by closing @xmath270 to @xmath298 , and similarly for @xmath54 .
[ lem : log1 ] suppose that @xmath24 is a short lipschitz curve .
let @xmath299 be arbitrary , @xmath300 .
let @xmath301 , @xmath207 , here @xmath302)$ ] .
suppose that @xmath219 is the arc - length of the curve .
then , we claim , @xmath303 we can assume that @xmath24 is in arc - length parametrization , @xmath97=[-\ell/2,\ell/2]$ ] , the axis of @xmath197 is the @xmath222 axis , @xmath304 is monotone increasing .
let @xmath305 if @xmath306 , @xmath307 if @xmath308 , @xmath309 if @xmath310
. then @xmath311 . in particular
, it is enough to consider @xmath312 instead of @xmath313 on @xmath314 $ ] .
the straightening of @xmath315 is maximal when @xmath316 , and it yields the formula in the statement .
[ lem : log2 ] suppose that @xmath24 is a short lipschitz curve .
suppose that @xmath219 is the arc - length of @xmath24 .
let @xmath317 , and @xmath318 be arbitrary .
then @xmath319 @xmath320 holds , where @xmath321 \}$ ] .
if @xmath322 is a subset of @xmath323 then @xmath324 @xmath325_{u=0}^\ell\leq n!\ , \ell\left(1+\log^+\frac2{\ell\cos\alpha } \right)^p .
\label{eq : logprod}\ ] ] using @xmath326 we can decompose our original integral , and from we obtain the estimate in the statement .
[ lem : log3 ] suppose that @xmath11 is a piecewise lipschitz curve system , @xmath174 .
then , we claim , there exist a number @xmath327 such that for any lipschitz curve @xmath24 lying in @xmath2 and @xmath317 , @xmath300 arbitrary the inequality @xmath328 @xmath329 holds , where @xmath321 \}$ ] , and @xmath219 is the arc - length of @xmath330 .
this follows from the previous lemma by cutting @xmath2 to small pieces .
one can similarly estimate @xmath331 where @xmath332 , @xmath333 and many other functions .
the concrete examples above are also easy @xmath334-estimates .
* b. * applications to integrals of cauchy type .
[ lem : plus1 ] consider the situation of [ po : summary ] .
j. ) assume that @xmath131 $ ] and @xmath132 where @xmath130 is continuous on @xmath133 .
then we claim : if @xmath10 extends to @xmath335 continuously then it also extends to @xmath336 continuously and vice versa .
in particular , @xmath337 holds on @xmath133 .
k. ) suppose that on @xmath138 $ ] and @xmath338 , where @xmath339 is continuous .
if @xmath124 stays in a closed sector not containing any curve direction of @xmath24 at @xmath162 then @xmath340 j. ) due to the local compactness of @xmath341))^+$ ] it is enough to show that the infinitesimal oscillation of @xmath10 near @xmath342 ( @xmath105 ) is @xmath41 .
suppose that the direction @xmath25 is transversal to @xmath24 in angle @xmath26 in a neighborhood of @xmath114 .
suppose that @xmath343 $ ] is an even smaller neighborhood of @xmath344 such that @xmath345}|m(s)-m(t)|<\delta$ ] .
we see that @xmath346 @xmath347\setminus[a_2,b_2]}\left(\frac1{\gamma(t)-\gamma(s)-u\xi } - \frac1{\gamma(t)-\gamma(s)+u\xi}\right)\left(m(t)-m(s)\right ) d\gamma(t)\ ] ] @xmath348}\left ( \frac1{\gamma(t)-\gamma(s)-u\xi}-\frac1{\gamma(t)-\gamma(s)+u\xi}\right ) \left(m(t)-m(s)\right ) d\gamma(t).\ ] ] after the equality , as a function of @xmath349 , every term extends to @xmath350)^+$ ] continuously , except the last one , but which , according to lemma [ lem : double ] , we can estimate by @xmath351 this shows that the infinitesimal oscillation of @xmath10 near @xmath342 is at most @xmath352
. the @xmath353 can be arbitrarily small , so we see that the infinitesimal oscillation of @xmath10 near @xmath342 is actually @xmath41 .
k. ) suppose that @xmath354\subset[a , b]$ ] is so small that @xmath355}$ ] is short and the semi - angle of @xmath356}$ ] is @xmath198 .
let @xmath357\subset[a , b]$ ] be even smaller so that @xmath358}|\gamma(t)-\gamma(a)| < \delta$ ] .
@xmath359\setminus[a , b_2 ] } \left(m(t)-m(a)\right)\frac{d\gamma(t)}{\gamma(t)-(\gamma(a)+\theta)}\ ] ] @xmath360 } \left(m(t)-m(a)\right)\frac{d\gamma(t)}{\gamma(t)-(\gamma(a)+\eta)}.\ ] ] dividing by @xmath361 we see that in the right side of the equation the first term has the required asymptotic behaviour , the second term goes to zero as we divide a bounded quantity by a logarithmic one . according to lemma [ lem : estkey].c , the @xmath362 of the absolute value of the first term divided by @xmath363 is less or equal to @xmath364 .
choosing @xmath353 well , this latter term can be arbitrarily small , so the statement follows .
let us consider a star system @xmath2 as in point [ po : star0 ] .
assume that @xmath365\rightarrow\mathbb c$ ] are continuous functions .
for @xmath366 let us define @xmath367 [ lem : plus2 ] suppose that @xmath261 is an open cone from @xmath262 .
assume the angle between any two directions from @xmath261 and @xmath368 is @xmath279 .
suppose that @xmath369 , and @xmath18 is bounded on the open segment @xmath370 connecting @xmath258 and @xmath29 . then , we claim , a. ) @xmath371 b. ) @xmath18 extends continuously to the analytic boundary of @xmath372 .
a. ) let @xmath373 . according to proposition [ lem : plus1].k @xmath374 , but which has to vanish according to the boundedness of @xmath10 in the expression .
b. ) the statement is already known except for around @xmath258 .
we can assume that @xmath375 .
we can assume that @xmath261 is so small that the angle between any two vectors in @xmath261 is @xmath376 .
let @xmath377 be a linear map sending @xmath261 into @xmath378 .
choosing it well , we can assume that there is a positive number @xmath187 such that for @xmath379 we have @xmath380 , @xmath381 .
let us chose @xmath382 such that @xmath383 and @xmath384 } |m_i(t)-m_i(a_1)|<\delta.\ ] ] now , for @xmath379 @xmath385 @xmath386 } ( m_i(t)-m_i(a_i ) ) \left(\frac1{\gamma_i(t)-z}-\frac1{\gamma_i(t)-l(z)}\right ) d\gamma_i(t)+\ ] ] @xmath387 } ( m_i(t)-m_i(a_i ) ) \left(\frac1{\gamma_i(t)-z}-\frac1{\gamma_i(t)-l(z)}\right ) d\gamma_i(t).\ ] ] then , on the right side of the equation every term is continuous on @xmath261 except possibly the last one . according to lemma [ lem : estkey3 ] , the absolute value of the last term @xmath388 } ( m_i(t)-m_i(a_i ) ) \frac{l(z)-z}{(\gamma_i(t)-z)(\gamma_i(t)-l(z ) ) } d\gamma_i(t)\right|\leq\ ] ] @xmath389 in particular , the infinitesimal oscillation of @xmath10 on @xmath261 around @xmath258 is less or equal than twice the value above .
now , @xmath353 can be arbitrarily small , so we know that @xmath10 extends continuously .
[ lem : plus3 ] suppose that the sectors @xmath285 and @xmath54 are siding @xmath277 from left and right respectively .
if @xmath10 extends to the analytic boundary of @xmath54 continuously then it also extends to the analytic boundary of @xmath285 continuously .
it is similar to the previous ones , but we have to use lemma [ lem : estkey4 ] .
here we collect some identities of rational functions on @xmath47 . [
lem : rid ] suppose that @xmath390 are three different points in @xmath47 .
then a. ) @xmath391 b. ) for @xmath174 , @xmath392 c. ) for @xmath393 @xmath394 now assume that @xmath22\rightarrow\mathbb c$ ] is a lipschitz curve , @xmath23)$ ] .
drawing on the basic integral formula we can get finer statements . in what follows @xmath395 .
[ lem : rif ] a. ) suppose that @xmath396 $ ] and @xmath397 .
then @xmath398 b. ) suppose that @xmath396 $ ] and @xmath399 modulo @xmath400 as @xmath401 .
then @xmath402 c. ) suppose that @xmath403 and @xmath393 .
then @xmath404 next , we are going to examine superpositions of the identities before .
[ po : meascont ] suppose that @xmath405 and @xmath182 are @xmath20 , @xmath21 , locally compact topological spaces , and we assign to each point @xmath406 a measure @xmath407 on @xmath405 .
we say that this assignment is continuous if a. ) the variation map @xmath408 is bounded and continuous ; b. ) the map @xmath409 is weakly continuous on @xmath182 .
+ in this case , if @xmath117 is a measure on @xmath182 then we may form the parametrized measure @xmath410 on @xmath411 , most readily to be defined as a radon measure . integrating in the variable @xmath344
we obtain the superposition measure @xmath412 in fact , both the parametrized and the superposition measures can also be taken if @xmath413 is not necessarily continuous but @xmath182 is a disjoint union of the borel sets @xmath414 such that on each of them @xmath413 is continuous , and the @xmath415 are @xmath20 , @xmath21 , locally compact spaces themselves . now , if @xmath416 is a bounded continuous , locally convex vector space valued function then @xmath417 we will consider superpositions only in very special cases , where the general arguments above can also be carried out by more direct means .
we introduce the measure @xmath418 on @xmath97 $ ] .
we define the function @xmath419 such that @xmath420}(x)$ ] if @xmath421 , @xmath422 if @xmath423 , and @xmath424}(x)$ ] if @xmath425
. then we set @xmath426 with this notation @xmath427 the map @xmath428 is continuous in the sense of [ po : meascont ] .
[ lem : addsuper ] suppose that @xmath429 is finite measure on @xmath97 $ ] .
suppose that @xmath42 $ ] , @xmath174 .
then @xmath430 where @xmath431 more explicitly , @xmath432 where @xmath433 \displaystyle{-\int^b_{s = t}\mu(s ) } & \qquad\text{if } \,t > t_0\,.\end{cases}}$ ] similarly to the additive case , let us define the measure @xmath434 on @xmath97 $ ] as follows : if @xmath435 then we set @xmath436}(t)\frac{(n_1+n_2 + 1)!}{n_1!n_2 ! } \frac{(\gamma(t)-\gamma(a_1))^{n_2}(\gamma(b_1)-\gamma(t))^{n_1}}{(\gamma(b_1)-\gamma(a_1))^{n_1+n_2 + 1 } } \,d\gamma(t);\ ] ] if @xmath437 then we set @xmath438 , the one - point measure at @xmath439 ; and if @xmath440 then we define @xmath441
. then @xmath442 the map @xmath443 is continuous in the sense of [ po : meascont ] on the sets @xmath444\times[a , b]\,:\ , x < y \}$ ] , @xmath445\times[a , b]\,:\ , x = y \}$ ] , and @xmath446\times[a , b]\,:\ , x > y \}$ ] , separately .
indeed , only the boundedness property is nontrivial , but from the bi - lipschitz property we obtain @xmath447 .
[ lem : prodsuper ] suppose that @xmath448 , @xmath449 are finite measures on @xmath97 $ ] , @xmath393 .
then @xmath450 holds with @xmath451\times[a , b]}\xi_{s_1,s_2}^{(\gamma , n_1,n_2)}(t)\mu_1(s_1)\mu_2(s_2),\ ] ] ie . a superposition with respect to the product measure @xmath452 .
the next step is to pass from lipschitz curves to lipschitz curve systems . in these cases , however , the normalization processes are not entirely canonical .
we start with the additive case .
suppose that @xmath11 is a connected lipschitz curve system , @xmath453 .
then we can construct a choice function @xmath454 , such that a. ) @xmath455 is a finite set of possibly degenerate lipschitz curves @xmath72\rightarrow\mathbb c$ ] @xmath456 lying in @xmath2 ; b. ) if @xmath457 then @xmath458)$ ] ; c. ) @xmath459 is a borel function .
we define the measure @xmath460 as @xmath461 if @xmath457 . the map @xmath462 is such that we can consider superposition in the sense of [ po : meascont ] : [ lem : addsuper2 ] suppose that @xmath11 is a connected lipschitz curve system , @xmath463 , @xmath174 , and the map @xmath464 is constructed according to a suitable choice function @xmath459 .
suppose that @xmath117 is a measure on @xmath2 .
then @xmath465 where @xmath466 if @xmath459 is piecewise constant on @xmath2 then @xmath467 piecewise , where @xmath468 is the natural variation measure oriented in some direction and @xmath469 is an integral function of @xmath470 in the same direction
. we can deal with the multiplicative case as follows : suppose that @xmath11 is a lipschitz curve system .
then we can construct a choice function @xmath471 , such that a. ) @xmath455 is a finite set containing the empty set @xmath472 and possibly degenerate lipschitz curves @xmath72\rightarrow\mathbb c$ ] @xmath456 lying in @xmath2 ; b@xmath473 . ) if @xmath474 then @xmath475)$ ] ; b@xmath476 . )
there is constant @xmath318 such that @xmath477 implies @xmath478 ; c. ) @xmath459 is a borel function .
( the most natural case is when @xmath477 if and only if @xmath479 and @xmath480 are in different connected components . )
we define the measures @xmath481 @xmath482 such that @xmath483 then the maps @xmath484 are such that we can consider superposition in the sense of [ po : meascont ] : [ lem : prodsuper2 ] suppose that @xmath11 is a lipschitz curve system , @xmath448 and @xmath449 are measures on it .
suppose that the map @xmath484 is obtained according to a choice function @xmath459 as above .
then @xmath485 where @xmath486 the following statements deal with global conformal transformations .
[ lem : confchange ] let @xmath487 be a conformal map , @xmath488 @xmath489 a. ) if @xmath490 , @xmath491 , @xmath492 then @xmath493 b. ) suppose that @xmath11 is a lipschitz curve system , @xmath117 is a measure on @xmath2 .
assume that @xmath494 .
suppose that @xmath395 and @xmath492 .
then @xmath495 for the purpose of the following statements let @xmath56 be a simply connected lipschitz domain with oriented boundary curve @xmath496 with underlying curve @xmath22_{a\sim b}\rightarrow\mathbb c$ ] .
[ lem : ratambig ] for any polynomial @xmath497 with degree @xmath498 @xmath499 [ lem : vancirc ] assume that @xmath500 is a holomorphic function which extends to the analytic boundary of @xmath59 continuously .
suppose that @xmath501_{a\sim b}\rightarrow\mathbb c$ ] gives the boundary values in the sense that at @xmath502 the extended value is @xmath503 .
suppose that @xmath174 , @xmath504 .
then @xmath505 [ lem : weakcirc ] suppose that @xmath397 , @xmath504 , @xmath42/\{a\sim b\}$ ] .
a. ) if @xmath506 then @xmath507 where @xmath508 is a branch such that it is continuous on @xmath509/\{a\sim b\})\setminus\{t_0\}$ ] .
b. ) if @xmath117 is a measure @xmath59 and @xmath504 then @xmath510 assumed that i. ) for a fixed @xmath258 the function @xmath511 is a branch such that it is continuous except at @xmath512 , ii . ) @xmath513 is lebesgue integrable . such choice functions @xmath514 can always be found .
let @xmath515 be the riemann sphere .
suppose that @xmath89 is a closed set .
we define @xmath8 as the set of functions @xmath10 such that a. ) @xmath10 is holomorphic on @xmath90 ; b. ) @xmath10 is slowly growing around @xmath2 , ie .
there exist @xmath174 , and @xmath516 such that @xmath517 , where @xmath518 is the distance function according to a riemannian metric on @xmath92 .
the growth condition says that @xmath10 grows around @xmath2 polynomially with exponent at most @xmath5 .
( if @xmath11 then this local condition can be checked using the ordinary euclidean distance function @xmath519 . ) it is easy to see that @xmath8 is an algebra , the algebra of slowly growing holomorphic functions on @xmath90
. we will be interested primarily in the very special case when @xmath2 is a lipschitz curve system .
* a*. first we examine the case of lipschitz curves on @xmath47 : let @xmath22\rightarrow\mathbb c$ ] be a lipschitz curve , @xmath23)$ ] .
a function @xmath520 has a power series expansion around @xmath103 : @xmath521 we define the operation of global primitive function @xmath86 if @xmath522 . then @xmath523 is the primitive function which vanishes at @xmath103 : @xmath524 it is easy to see that @xmath525 is still in @xmath526 .
it is often useful to write the expansion around infinity in terms of @xmath527 , so @xmath528 we can notice that @xmath529 and @xmath530 but the other terms do vary .
nevertheless , @xmath531 can always be expressed from @xmath532 .
they encode the same information , the asymptotic expansion modulo @xmath533 .
suppose that @xmath174 , @xmath117 is a measure on @xmath97 $ ] , such that @xmath534 in this case we say that @xmath18 is written in a simple @xmath535-normal form with base point @xmath43 .
a priori we do not ask that @xmath536 should be understood in the sense of the previous point , but it is easy to see that they must be the same .
it is reasonable to say that @xmath18 allows a simple @xmath537-normal form if @xmath538 , independent of base points . more generally ,
if @xmath539 are measures on @xmath97 $ ] and @xmath540 then we say that @xmath18 is written in an @xmath535-normal form . on the other hand
, we may also notice that if a function is of shape like above then it is a slowly growing holomorphic function , in fact , of exponent @xmath541 .
a simple @xmath542-normal formal is certainly an @xmath542-normal form . on the other hand ,
if @xmath18 allows an @xmath542-normal form then it allows a simple @xmath542-normal form with any base point in @xmath80)$ ] .
indeed , the measure @xmath543 can be gathered into a constant @xmath544 , then by induction , using lemma [ lem : addsuper ] , we can push the measures up to the @xmath535-th level , except the singular term in the base point .
in fact , this gives a well - defined reduction process , whose validity is based solely on lemma [ lem : addsuper ] .
[ prop : normal1 ] every @xmath545)}$ ] allows a simple @xmath542-normal form for certain @xmath5 .
if @xmath5 and the base point are fixed then the simple normal form is unique .
first we prove unicity .
the @xmath546 are determined by the power series expansion of @xmath10 around infinity .
then it is enough to show that @xmath547 determines @xmath117 .
indeed , its @xmath5-th primitive function is @xmath548 , from which @xmath117 can be recovered according to the @xmath149-formula , [ po : summary].d .
existence is as follows : if @xmath18 is of exponent @xmath549 then locally its @xmath5-th primitive function extends lipschitz continuously to the analytic boundary . however , there are obstructions to have a global primitive function . for that reason we take the global @xmath5-th primitive function @xmath550 of @xmath551 .
we see this primitive function locally lipschitz continuously extends to the analytic boundary except at @xmath552 .
let @xmath553 .
this is a locally lipschitz function except at @xmath43 .
if we choose another @xmath554 and proceed with that then we see that the difference @xmath555 in the @xmath5-th primitive functions is @xmath556 .
now , @xmath130 @xmath557 can be recovered from @xmath166 @xmath558 using the @xmath149 operator , and which yields @xmath559 . from
that we can see that the discontinuity of @xmath339 is not at all too bad , @xmath339 is a piecewise lipschitz function .
let @xmath560 .
we see that @xmath561 extends holomorphically to @xmath562 except at @xmath43 , and it is vanishing at infinity in all orders . from this
we deduce that @xmath563 .
* b. * next , we assume that @xmath11 is a lipschitz curve system .
suppose that @xmath564 @xmath565 are the connected components of @xmath2 .
let @xmath566 be chosen .
if with appropriate @xmath5 , constants @xmath567 , and a measure @xmath117 on @xmath2 @xmath568 then we say that @xmath18 is written in a simple @xmath535-normal form .
similarly , if @xmath539 are measures on @xmath2 and @xmath569 then we say that @xmath18 is written in an @xmath535-normal form . for lipschitz curve systems things go in similar manner : ( @xmath5)-normal forms can be reduced to simple ( @xmath5)-normal forms by lemma [ lem : addsuper2 ] , every element @xmath10 of @xmath8 allows a normal form , etc .
there is , however , one important difference : it is not true that ( the @xmath570 chosen ) the simple normal form is unique .
first , the constant terms @xmath571 can be accumulated into a single constant term . apart from that ,
the higher @xmath572 are fixed , because they are obstructions to taking the @xmath573-th primitive functions around @xmath574 .
( in fact the @xmath574 can be nested so that one encloses the other one but we can go in the inside - out direction . ) but the @xmath117 can also be a source of ambiguity . assume that @xmath575 and @xmath2 contains a cycle @xmath576 .
assume that @xmath497 is a polynomial of order @xmath577 .
then @xmath578 clearly , a term @xmath579 can be added the measure - term @xmath117 without consequences .
let @xmath580 be the number of independent cycles in @xmath2 .
then due to the reasons above we have already found a @xmath581 dimensional family of simple normal forms .
what is true , however , is that by this we have already recovered all simple normal forms for @xmath10 . from the discussion above
we see that for each @xmath520 we can define number @xmath582 which is the smallest natural number @xmath5 such that @xmath10 allows a ( simple ) @xmath542-normal form .
it is immediate that @xmath583 gives an additive filtration of @xmath526 , in particular @xmath584 with equality if @xmath585 .
what is not so obvious is that it also gives a multiplicative filtration in the sense that @xmath586 this , however , follows from lemma [ lem : prodsuper2 ] .
* c. * extension to @xmath92 .
[ po : conf ] suppose that @xmath487 is a conformal map , @xmath11 is a lipschitz curve system , and @xmath494 .
suppose that the function @xmath587 has an @xmath542-normal form as above .
then lemma [ lem : confchange].b can be applied and we find that @xmath588 with appropriate measures @xmath589 .
substituting @xmath590 we can immediately see that @xmath591 in other terms , the new coefficient measures belong to an @xmath542-normal form of the function @xmath592 with respect to the curve system @xmath593 .
the argument above , of course , works in a reversed way : if we know that @xmath592 allows a ( simple ) @xmath542-normal form with respect to @xmath593 then we may consider a generalized ( simple ) @xmath542-form of @xmath10 with respect to @xmath2 : assume that @xmath89 is lipschitz curve system and @xmath594 .
then we can define appropriate normal forms as follows : suppose that @xmath487 is a conformal map such that @xmath494 .
then @xmath595 , and it allows , a normal form @xmath596 then by defining the measure @xmath597 we find that @xmath598 so , we see that generalized ( simple ) normal forms can be defined similarly to ( simple ) normal forms but @xmath599 substituted by @xmath600 .
the importance of point [ po : conf ] beside providing inspiration is that it immediately shows that @xmath601 above does not depend on @xmath602 as long as @xmath494 .
consequently , the filtration @xmath583 extends to general @xmath8 , and @xmath603 is conformal invariant in the sense that @xmath604 with @xmath602 arbitrary .
the concrete form of @xmath602 is not really important here . in order just to indicate what kind of statements we may obtain this way we state
suppose that @xmath22\rightarrow\overline{\mathbb c}$ ] is a lipschitz curve , @xmath23)$ ] , @xmath520 .
suppose that @xmath605 , @xmath42 $ ] .
then for @xmath606 , we claim , there are unique coefficients @xmath607 and a measure @xmath608 on @xmath97 $ ] such that @xmath609 where @xmath610 and @xmath611 ( ie . a generalized simple normal form ) .
analogous statements hold for general lipschitz curve systems @xmath2 .
indeed , the case @xmath612 is what we have studied before .
the case @xmath613 follows from taking @xmath614 .
* d. * preparation for the locality argument .
[ po : prep ] in what follows we give a recipe for manipulating certain measures .
although it involves choices the method itself is quite well - defined : suppose that @xmath11 is a lipschitz curve system , @xmath615 , where @xmath616 , @xmath617 are open , and @xmath617 contains the infinite component of @xmath618 .
assume that @xmath594 such that @xmath619 .
now suppose that @xmath620 are measures on @xmath47 such that they are supported on @xmath2 and @xmath621 suppose that the connected components of @xmath622 are @xmath623 .
our first step among the manipulations is that we cut up the @xmath624 with some lipschitz curves such that it yields simply connected lipschitz domains @xmath625 .
we can choose the cutting curves such that they are pairwise disjoint simple curves connecting regular points of @xmath2 .
we do that for all @xmath573 such that @xmath626 .
let @xmath627 , where @xmath628 and @xmath629 is the supporting set of the cuts .
the curve system @xmath630 has nodes of two type : there are the old nodes coming from @xmath2 and there are the new nodes coming from the cuts . in the second step , we push the measures @xmath631 to the boundary and we sum them up .
let @xmath632 where the latter two terms are obtained according to lemma [ lem : weakcirc].b .
at this point , the supporting set of the measures is @xmath630 .
we know that @xmath633 for @xmath634 . in the third step , using appropriate additive reduction we have @xmath635 where the @xmath570 are from different connected components of @xmath630 .
due to residue behaviour the @xmath636 vanish , so @xmath637 because of the nature of additive reduction and the properties of @xmath638 we know that locally @xmath639 , where @xmath640 is the variation measure of @xmath630 with piecewise constant orientation , and @xmath641 is a piecewise continuous function .
we can assume that the discontinuities happen only at the nodes of @xmath630 .
[ lem : prepared ] consider @xmath642 which is @xmath643 times an @xmath5-th primitive function of @xmath644 .
we claim that @xmath602 extends to the analytic boundary of @xmath645 .
if the @xmath646 is the oriented boundary curve of @xmath625 , @xmath647 are the corresponding extended values , and @xmath648 are the underlying curves of @xmath646 then @xmath649 first , we claim that @xmath650 extends continuously to the analytic boundary of @xmath651 .
indeed , if @xmath344 is a regular point of @xmath630 , then @xmath344 bounds @xmath617 , on which side @xmath650 is a polynomial , so by proposition [ lem : plus1 ] @xmath650 extends to the other side continuously , too .
if @xmath344 is an old node of @xmath630 then the argument is similar , except we have to use proposition [ lem : plus3 ] . at new nodes ,
we claim the behaviour is not far from regular points . indeed , let us slightly relocate the cut , then we obtain @xmath652 and @xmath653 .
then @xmath344 is a regular point of @xmath652 and the difference @xmath654 is locally polynomial .
this indeed proves the existence of the continuous extension .
now consider @xmath655 this is equal to @xmath650 but on @xmath617 where it vanishes .
nevertheless @xmath654 is a polynomial of degree @xmath656 on the components of @xmath617 . in particular , taking @xmath5-th derivatives and dividing by @xmath643 we obtain .
the fourth step of our manipulation process is to replace the normal form by the normal form .
notice that this replacement can be done by an additive reduction process , no matter how we argued for the existence of the normal form .
a multivalued linear operator @xmath0 on a vector space @xmath1 is simply a linear subspace of @xmath657 . if @xmath0 is a multivalued linear operator and @xmath658 , then we may form the multivalued linear operator @xmath659 this yields an associative @xmath660-action on multivalued linear operators . in simple cases
we still use the colloquial notation @xmath661 , @xmath662 , @xmath663 , @xmath664 , @xmath665 , etc .
we have to keep in mind that the definition is a matrix action , so @xmath666 is to be imagined as at least three consecutive matrix actions ( @xmath667 , @xmath668 , @xmath669 ) executed but which , naturally , also simplifies according to the associativity of the action . the only possible source of confusion is when @xmath670 because then we have to keep in mind that we started with @xmath0 .
so , in that case it is better to write @xmath671 . if @xmath672 then we have a restriction of one of @xmath673 , @xmath674 or @xmath675 .
if @xmath676 then we have bijective correspondence of multivalued linear operators .
if @xmath677 is the corresponding conformal map on the riemann sphere , then we use the short notation @xmath678 .
we say that the multivalued linear operator @xmath0 is an ( not necessarily everywhere defined ) operator , if @xmath679 implies @xmath680 , ie .
its values are well - defined . in that case
we define @xmath681 as @xmath682 where @xmath683 .
[ lem : quasi1 ] a. ) if @xmath684 , @xmath685 then @xmath686 makes sense if and only if @xmath687 makes sense , and in that case @xmath688 .
b. ) suppose that @xmath689 make sense , @xmath684
. then @xmath690 makes sense if and only if @xmath691 makes sense , and in that case @xmath692 .
c. ) if @xmath693 and @xmath694 make sense then @xmath695 and @xmath696 make sense , moreover @xmath697 , @xmath698 .
@xmath699 makes sense , @xmath684 , @xmath685 then @xmath693 makes sense .
[ lem : quasi2 ] suppose that @xmath700 , @xmath701 , @xmath702 .
if i. ) @xmath703 and @xmath704 make sense , or ii . ) @xmath705 makes sense , @xmath706 , + then @xmath707 including the statement that every expression in the equation makes sense .
set @xmath708 .
then @xmath709 with @xmath710 , @xmath711 .
what we have to prove is that @xmath712 including that every term is well - defined .
i. ) we know that @xmath693 and @xmath713 exist .
in particular , @xmath714 also exists . using lemma [ lem : quasi1].c
we see that @xmath715 and @xmath695 also exist , and they are equal to @xmath6
. then lemma [ lem : quasi1].b implies that the equality above holds .
ii . ) in that case @xmath716 .
then , according to lemma [ lem : quasi1].d @xmath693 exists .
by this we have reduced the problem to case ( i ) .
[ lem : quasi3 ] suppose that @xmath717 , @xmath28 .
if i. ) @xmath718 and @xmath719 make sense , or ii . ) @xmath720 makes sense + then the resolvent identity holds : @xmath721 @xmath722 including the statement that all expressions in the identity are well - defined . in short terms :
the resolvent identity holds if either side makes sense .
let @xmath723 , @xmath724 . then @xmath725 in particular , @xmath693 and @xmath714 exist , hence @xmath695 and @xmath715 exist and they are equal to @xmath6 .
we have to prove that @xmath726 ie . that @xmath727 i. ) the equality follows from lemma [ lem : quasi1].b .
ii . ) in that case @xmath728 and the statement is reduced to part ( i ) by lemma [ lem : quasi1].d .
the resolvent identity is still valid , which is of crucial importance .
it makes possible to generalize the identities concerning rational functions .
if @xmath729 , @xmath730 , @xmath0 is a multivalued linear operator on @xmath1 , @xmath4 then we use the notation @xmath731 ie .
@xmath732 is @xmath664 iterated @xmath5 times on @xmath6 multiplied by the constant @xmath187 .
now , using the resolvent identity above one can easily prove the full analogue of lemma [ lem : rid ] : [ lem : qrid ] suppose that @xmath717 , @xmath28 .
then b. ) for @xmath174 @xmath733 if either side makes sense .
c. ) for @xmath393 @xmath734 if either side makes sense . assume from now on that @xmath1 is a locally convex vector space .
suppose that @xmath89 and @xmath0 is a multivalued linear operator .
for @xmath174 we define @xmath735 as the set of elements @xmath6 such that for all @xmath736 the functions @xmath737 and @xmath738 are well - defined and continuous . notice that @xmath739 is simply the domain of @xmath0 if it is an operator . in order to make this notion more friendly ,
we give another characterization : [ lem : res1 ] a. ) if @xmath740 then @xmath741 if and only if for all @xmath736 the functions @xmath742 are well - defined and continuous .
b. ) if @xmath613 , @xmath743 then @xmath741 if and only if for all @xmath736 the functions @xmath744 are well - defined and continuous . one might think that @xmath745 is characterized by a global property with respect to @xmath2 .
this is not the case , it is a local property . using the resolvent identity
it is not hard to prove [ lem : res2 ] we have @xmath741 if and only if for any @xmath254 there is an open set @xmath506 such that @xmath746 .
now we can state the generalization of lemma [ lem : rif ] .
for the sake of simplicity we write @xmath747 instead of @xmath748 .
this expression can also be understood either formally or as a vector measure which is the product of the scalar measure @xmath749 and a continuous vector valued function .
[ lem : qrif ] now assume that @xmath22\rightarrow\mathbb c$ ] is a lipschitz curve , @xmath23)$ ] .
a. ) suppose that @xmath396 $ ] , @xmath397 , @xmath750 .
then @xmath751 b. ) under similar assumption , suppose that @xmath399 modulo @xmath400 as @xmath401 .
then @xmath752 c. ) suppose that @xmath403 , @xmath393 , @xmath753 .
then @xmath754 part ( a ) follows from lemma [ lem : qrid].a . part ( b ) and ( c ) follow from part ( a ) .
when we consider superpositions as in lemma [ lem : addsuper ] and [ lem : addsuper2 ] it is essential that the integrals in question exist .
for this reason we assume that the locally convex vector space @xmath1 is sequentially complete .
proceeding with the generalization of lemma [ lem : addsuper2 ] : [ qlem : addsuper2 ] suppose that @xmath11 is a connected lipschitz curve system , @xmath463 , @xmath174 , and the map @xmath464 is constructed according to a suitable choice function @xmath459 .
suppose that @xmath117 is a measure on @xmath2 , @xmath755 , @xmath1 is a sequentially closed locally convex vector space .
then @xmath756 where @xmath466 it follows from the fubini integral formulae .
when it comes to multiplication as in lemma [ lem : prodsuper ] and [ lem : prodsuper2 ] , in order to generalize them we have to apply for the sequential closedness of @xmath0 .
the reason is that closedness allows to interchange integration and operator actions : [ lem : closein ] suppose that @xmath117 is a measure on @xmath182 , @xmath757 is a continuous @xmath1-valued function on @xmath182 , and @xmath0 is an operator on @xmath1 such that @xmath758 is also continuous on @xmath182 .
assume that @xmath1 is sequentially complete , @xmath0 is sequentially closed
. then @xmath759 including the statement that both sides make sense .
it is a trivial observation that if @xmath658 , @xmath760 and @xmath0 is sequentially closed then @xmath761 is sequentially closed , too . using that we can state [ qlem : prodsuper2 ]
suppose that @xmath11 is a lipschitz curve system , @xmath448 and @xmath449 are measures on it .
suppose that the map @xmath484 is obtained according to a choice function @xmath459 as earlier .
assume that @xmath762 , @xmath1 is a sequentially closed locally convex vector space , @xmath0 is a sequentially closed multivalued linear operator .
then @xmath763 where @xmath486 it follows from the previous lemma and the fubini integral formulae .
the generalization of the rest of section 4 is straightforward .
[ qlem : confchange ] let @xmath487 be a conformal map , @xmath488 @xmath489 a. ) suppose that @xmath490 .
then @xmath764 b. ) suppose that @xmath11 is a lipschitz curve system , @xmath117 is a measure on it .
assume that @xmath494 .
suppose that @xmath3 . then @xmath765 for the purpose of the following statements let @xmath56 be a simply connected lipschitz domain with oriented boundary curve @xmath496 with underlying curve @xmath22_{a\sim b}\rightarrow\mathbb c$ ] .
[ qlem : ratambig ] for any polynomial @xmath497 with degree @xmath498 , @xmath750 @xmath766 cutting @xmath24 to pieces , it follows from lemma [ lem : qrif].b .
[ qlem : vancirc ] assume that @xmath500 is a holomorphic function which extends to the analytic boundary of @xmath59 continuously .
suppose that @xmath501_{a\sim b}\rightarrow\mathbb c$ ] gives the boundary values in the sense that at @xmath502 the extended value is @xmath503 .
suppose that @xmath174 , @xmath767 .
then @xmath768 if @xmath769 is a similar curve inside @xmath59 then the statement is a consequence of green s theorem , or rather the fundamental integral formula behind it . due to the fact that @xmath59 is a lipschitz domain
, we can well approximate @xmath24 by such @xmath769 , and by continuity the statement follows .
[ qlem : weakcirc ] a. ) if @xmath397 , @xmath506 , @xmath42/\{a\sim b\}$ ] , @xmath770 then @xmath771 where @xmath508 is a branch such that it is continuous on @xmath509/\{a\sim b\})\setminus\{t_0\}$ ] .
b. ) if @xmath117 is a measure on @xmath59 , @xmath397 , @xmath506 , @xmath42/\{a\sim b\}$ ] , @xmath770 , then @xmath772 as long as i. ) for a fixed @xmath258 the @xmath511 is a branch such that it is continuous except at @xmath512 , ii . ) @xmath513 is lebesgue integrable .
a. ) by the the standard encircling argument and partial integration @xmath773 b. ) this is just the superposition statement of part ( a ) .
let us fix a lipschitz curve system @xmath89 , a sequentially complete locally convex space space @xmath1 , and a multivalued linear operator @xmath0 on @xmath1 .
assume now that @xmath9 .
then we define the operator @xmath14 on @xmath774 as follows : consider any generalized normal form @xmath775 where @xmath487 is a conformal map such that @xmath776 .
then , for @xmath777 we define @xmath778 ie . the evaluation of the normal form for @xmath0 on @xmath6 .
we have to show that @xmath779 above does not depend on the particular choice of the generalized normal form . if @xmath9 , @xmath3 then a. ) @xmath779 is well - defined .
b. ) @xmath779 is linear in @xmath10 and @xmath6 .
c. ) if @xmath602 is a conformal map , then @xmath780 d. ) if @xmath0 is sequentially closed then @xmath14 defines linear maps @xmath781 part ( a ) : first , we consider the case when @xmath782 can be taken
. then we have to show that @xmath779 does not depend on the particular choice of the normal form belonging to @xmath10 . we know that normal forms can be reduced to simple normal forms with fixed poles using lemma [ lem : addsuper2 ] , while according to theorem [ qlem : addsuper2 ] the evaluation for @xmath0 does not change .
then the simple normal forms differ only by some polynomial cycles whose evaluation vanishes according to theorem [ qlem : ratambig ] . in the second step
, one must show that using different @xmath602 the evaluation does not change .
indeed , theorem [ qlem : confchange].b shows that while obtaining a generalized normal form according to point [ po : conf ] the evaluation of @xmath779 does not change .
the general case then follows from a formal change of variable @xmath783 .
parts ( b ) and ( c ) follow from the definition immediately ; ( d ) follows from lemma [ lem : closein ] .
next we prove the multiplicative property , from which one can develop the spectral mapping theorem , etc .
suppose that @xmath0 is a sequentially closed multivalued linear operator on the sequentially complete locally convex space space @xmath1 , and @xmath784 , @xmath785 .
then @xmath786 , and for @xmath787 @xmath788 after a conformal change of variable we can assume that @xmath11 .
lemma [ lem : prodsuper2 ] and theorem [ qlem : prodsuper2 ] show that a certain normal form of the product function @xmath789 gives the evaluation @xmath790 .
suppose that @xmath2 is a lipschitz curve system , @xmath791 , @xmath792 vanishes on @xmath617 , and @xmath793 .
then @xmath794 after a conformal change of variable we can assume that @xmath11 .
let us follow the evaluation @xmath779 during the preparation argument of [ po : prep ] . according to theorem [ qlem : addsuper2 ] and theorem [ qlem
: weakcirc].b we see that the evaluation does not change
. using functional theoretic arguments , in lemma [ lem : prepared ] , we saw that the prepared normal form can be assumed to be of the very special form .
but then , according to theorem [ qlem : vancirc ] , the evaluation @xmath779 vanishes . the case of the classical riesz functional calculus ( in case of lipschitz curves ) can be considered as the special case when @xmath2 is a closed simple curve bounding the bounded open set @xmath59 which contains the spectrum of @xmath0 , and @xmath795 where @xmath315 is holomorphic in a neighborhood of @xmath796 .
the locality property is the counterpart of the statement that one can relocate the integrating paths .
if we want to consider @xmath797 then the natural idea is take the function @xmath798)}$ ] where @xmath102 is an appropriate branch of logarithm and @xmath799\rightarrow\overline{\mathbb c}$ ] is a lipschitz curve starting at @xmath800 and ending at @xmath801 .
this prescription is unique except that there is an ambiguity @xmath802 in the logarithm function . regarding the natural choice we may proceed as
follow : suppose that @xmath799\rightarrow\overline{\mathbb c}$ ] is a lipschitz curve and @xmath803\rightarrow\mathbb c\mathsf z+\mathbb c$ ] is a formal linear function such that @xmath804 , where @xmath805 , @xmath806 , @xmath807 is a scalar multiple of @xmath808 except at @xmath809 and it is a scalar multiple of @xmath810 except at @xmath811 ( in particular , @xmath812 ) .
we assume that the scalar multipliers are locally nonvanishing lipschitz curves in technical sense .
then , we can define in formal sense @xmath813 if we substitute @xmath814)$ ] into @xmath815 then we obtain a concrete representative @xmath816 . also , simplifying the right side one can easily extract a normal form . for @xmath817 we may define @xmath818 as the average of @xmath819 and @xmath820 . according to the situation above , for @xmath821 we may define powers as @xmath822 for @xmath814)$ ] with @xmath823 and for @xmath817 with @xmath824 , respectively .
a very natural case is when @xmath825 .
this yields the function @xmath115 and @xmath826)}\ ] ] ( cf .
[ po : logref1 ] ) .
the other very natural case is when we consider the straight segment connecting @xmath827 and @xmath815 : @xmath828 for @xmath829 $ ] .
the associated logarithm function ( in the reverse direction ) is @xmath830 ie .
the classical logarithm function cut along @xmath831 .
the associated powers are @xmath832 ( one can pass from @xmath833 by substituting @xmath834 , @xmath835 , cf .
[ po : logref2 ] . )
a. ) if @xmath174 , @xmath836)}(a)$ ] , @xmath837 then @xmath838 can be iterated on @xmath6 arbitrarily many times and @xmath839 b. ) let @xmath174 , and @xmath840 $ ] or @xmath841 . if @xmath842)}(a)$ ] , @xmath843 then the function @xmath844 is well - defined and continuous .
moreover , if @xmath845 then @xmath846 c. ) it yields @xmath847 as long as @xmath848 . in particular ,
if @xmath849 then the @xmath850 form a local group on @xmath851)}(a)$ ] with infinitesimal generator @xmath838 . for @xmath397 @xmath852 part ( a ) follows from multiplicativity .
regarding ( b ) : for @xmath176 @xmath853 if @xmath176 then the convergence of the normal form measures implies the statement .
let @xmath854 . if @xmath855 , @xmath856 ( say @xmath857 ) then @xmath858 @xmath859 @xmath860 proceeding backwards we can make sure of convergence . in the last step we collapse the terms belonging to the same sum of powers .
that proves the statement for @xmath861 . for @xmath862
$ ] we can argue similarly .
part ( c ) follows from the power series expansion .
[ po : powext ] one can make a similar statement for @xmath863 here for , @xmath176 @xmath864 assume now that @xmath865 .
one can notice that the expression above make sense if @xmath6 is such that for some @xmath866 @xmath867 is bounded .
we denote the set of such elements @xmath6 by @xmath868 .
[ lem : powlim ] if @xmath869 then the function @xmath870 is continuous , and for @xmath871 @xmath872 the choice @xmath873 yields the same expression as on right side of .
hence , it is entirely reasonable to define @xmath874 for @xmath869 as in .
what happened is that we used convergence improvement by singling out a particular normal form , but which idea was supported by a continuity property . using lemma [ lem : powlim ] above
, one can easily extend the semigroup property , etc .
the formula above already applies to positive operators in hilbert spaces : if @xmath0 is positive then @xmath875 , which shows that @xmath876 for @xmath877 .
now , is a classical formula ready to relate to other ones usual in the theory of fractional powers .
there are many ways to improve the holomorphic calculus above .
a. ) if one examines the proofs before , one can see that instead of lipschitz curve systems it is sufficient to consider piecewise lipschitz curve systems where at nodes the sets of curve directions regarding the different pieces are disjoint .
b. ) when we consider logarithms and exponentials as in the previous section we can take ( possibly closed ) piecewise lipschitz curves .
for example , lemma [ lem : log3 ] shows that logarithmic singularities are rather nice , and by deleting smaller and smaller neighborhoods of the nodes taking limits allow us , by continuity , to pass from disjoint unions of lipschitz curves to piecewise lipschitz curves .
c. ) instead of allowing all possible normal forms one can restrict to simple normal forms with , say , @xmath36 measure terms . in this case
the condition @xmath3 can be relaxed , principally the iterated resolvent does not need to be defined at the diagonal of @xmath7 .
mm c. a. berenstein , r. gay : _ complex variables . an introduction .
_ graduate texts in mathematics 125 .
springer - verlag , new york , berlin , heidelberg , 1991 .
n. dunford , j. t. schwartz : _ linear operators , part i : general theory .
_ interscience publishers , new york , 1967 .
m. haase : _ the functional calculus for sectorial operators .
_ operator theory : advances and applications 169 .
birkhuser verlag , basel , 2006 .
h. komatsu : fractional powers of operators .
_ pacific j. math .
_ * 19 * ( 1966 ) , 285346 .
lakos : notes on lebesgue integration .
arxiv : math.fa/0506185 .
lakos : elementary spectral calculations in rings .
arxiv : math.ra/0610921 . c. martnez carracedo , m. sanz alix : _ the theory of fractional powers of operators . _ north - holland mathematics studies , 187 .
north - holland , amsterdam , 2001 . | we develop a holomorphic functional calculus for ( multivalued linear ) operators on locally convex vector spaces .
this includes the case of fractional powers along lipschitz curves . |
recently , slow anomalous diffusion , which is defined by the sublinear dependence of the mean square displacement ( msd ) on time , has been found in various phenomena including lipid granules diffusing in living fission yeast cells @xcite , colloidal particles diffusing on sticky surfaces @xcite and in networks of entangled actin filaments @xcite , mrna molecules diffusing in _ e. coli _
@xcite , chromosomal loci diffusing in bacteria @xcite , telomeres diffusing in nuclei of eukaryote cells @xcite , and proteins diffusing in dextran solutions @xcite . to understand these phenomena , two types of slow diffusion models have been extensively studied so far : ( 1 ) continuous - time random walks ( ctrws ) @xcite and ( 2 ) the generalized langevin equation ( gle ) and fractional brownian motions ( fbms ) @xcite . for the ctrw model ,
if the probability density function ( pdf ) of the waiting times between jumps of the particle is a power law @xmath0 with @xmath1 , the ensemble - averaged msd ( eamsd ) shows slow diffusion @xcite : @xmath2 '' to stand for an asymptotic relation , i.e. , @xmath3 means @xmath4 for some asymptotic limit such as @xmath5 .
on the other hand , we use `` @xmath6 '' to stand for a stricter asymptotic relation : @xmath7 means @xmath8 for some asymptotic limit .
furthermore , we use `` @xmath9 '' when @xmath10 is an approximation of @xmath11 , and `` @xmath12 '' when @xmath10 is proportional to @xmath11 , that is , they are not necessarily asymptotic relations . ] . typical origins of such power - law waiting times are complex energy landscapes @xcite and diffusion in inner degrees of freedom @xcite . similarly , eamsd of gle shows slow diffusion if the memory kernel @xmath13 decays algebraically @xmath14 with @xmath1 @xcite .
a possible origin of this non - markovian memory effect is the viscoelasticity of the medium @xcite . a primary difference between these two stochastic processes ctrws and gle is in their ergodic properties .
even if the memory kernel of a gle is given by a power function as stated above , the gle satisfies the ergodic property , that is , time - averaged quantities coincide with ensemble - averaged quantities @xcite .
in contrast , this equivalence between the two averages does not hold in the ctrws with power - law waiting times .
however , such ctrws exhibit an extended form of ergodicity time - averaged quantities become random variables following a distribution function @xcite .
this distributional ergodicity is called weak ergodicity breaking @xcite or infinite ergodicity @xcite .
in addition to the above mentioned examples , ctrws are frequently used in various fields of science @xcite . however , there are cases in which some finite - size effects should be considered in order to compare the model with experimental data
. in particular , cutoffs at the tail of the power - law waiting time distribution @xcite and spatial confinement effects @xcite arise in many systems .
the typical origins of waiting time cutoffs are physical limits with respect to energy and spatial extensions ( e.g. , @xcite ) .
theoretical analysis of the ctrws with such cutoffs is difficult , because it is necessary to investigate the transient behavior .
however , this difficulty is avoidable by using the tempered stable distribution ( tsd ) @xmath15 @xcite .
the tsd has exponentially smooth cutoffs , and is a modified version of the truncated stable distribution , which has a sharp cutoff at the tail @xcite .
furthermore , the tsd has the property called infinite divisibility @xcite , which allows a rigorous analysis even for transient behavior .
the tsd is a special model for the cutoffs , but it shows typical behavior of cutoff distributions such as a slow convergence to the gaussian distribution @xcite .
furthermore , the generalized fractional fokker - planck equation ( gffpe ) , which we will use in this paper to study eamsd , has been derived for ctrws with tempered stable waiting times @xcite . the ergodic properties of the system without confinement have also been clarified in @xcite , where the existence of a clear transition from weak ergodicity breaking ( an irreproducible regime ) to ordinary ergodicity ( a reproducible regime ) was shown analytically for the case of a non - equilibrium initial ensemble ( see sec . [
sec : tsd ] for a precise definition of the non - equilibrium and equilibrium ensembles ) .
another important finite - size effect is spatial confinements .
for example , to understand the transport phenomena in cells @xcite , confinements due to cell membranes should be considered .
confinement effects in ctrws with power - law waiting times were studied numerically in @xcite and analytically in @xcite ; it was found that the time - averaged msd ( tamsd ) shows a crossover from normal diffusion at short timescales to anomalous slow diffusion at longer timescales .
they also reported numerically the weak ergodicity breaking in tamsd .
recently , the ctrw model with the two finite - size effects ( i.e. , waiting time cutoffs and spatial confinements ) has been used as a model for the transport of lipid granules in living fission yeast cells @xcite , in which confinement effects are caused by a hookean force exerted by optical tweezers .
the model clearly explains the experimental results such as the crossover in tamsd and weak ergodicity breaking .
these studies mainly used numerical simulations , but a detailed theoretical analysis has not yet been reported . also , they studied only non - equilibrium ensemble , and thus the dependences of the ergodic properties on initial ensembles are still unknown . in this paper , we present theoretical results for ctrws with two finite - size effects : the effects of cutoffs in the waiting time distribution and the confinement effects .
in particular , we focus on tamsd as an observable , and study the crossover from normal to anomalous diffusion in the tamsd and ergodic properties in terms of the scatter of diffusion constant for the tamsd .
the tamsd , which is often used in single - molecule tracking experiments @xcite , is defined as @xmath16 where @xmath17 is the position of the particle at time @xmath18 , @xmath19 is the total measurement time , and @xmath20 is the time interval .
hereinafter we assume that @xmath21 . here
we define a generalized diffusion constant @xmath22 as @xmath23 . in some experiments
, it has been reported that @xmath22 behaves like a random variable depending on each time series @xcite .
therefore , the scatter in tamsd or @xmath22 has been used to check the consistency of the model with experimental data @xcite . in this paper
, we use the tsds @xcite as waiting time distributions of ctrws . for the tsd , it is possible to explicitly write the convoluted waiting distributions of any order [ eq .
( [ e.prob.na ] ) ] .
moreover , we use the numerical method for the tsd presented in @xcite and appendix [ sec : num.method ] , and study the initial and boundary value problem of the gffpe to understand the confinement effect .
the rest of this paper is organized as follows . in sec .
[ sec : tsd ] , we introduce the tsds .
then , in sec . [ sec : ea.tamsd ]
, we show the crossover from normal to anomalous diffusion in tamsd by using the gffpe . in sec .
[ sec : stat.tamsd ] , ergodic properties of tamsd are studied using renewal theoretic analysis . secs .
[ sec : conclusion ] and [ sec : discussion ] are devoted to conclusion and discussion . in the appendices ,
we summarize some technical matters , including a derivation of gffpe from ctrws as well as the ergodic properties of general observables .
in this paper , we consider ctrws confined in a one - dimensional lattice with unit lattice constant : @xmath24 .
jumps are permitted only to the nearest neighbor sites without preferences , although this can be generalized to jump length distributions with zero mean and finite variances .
it is also assumed that the particle is reflected if it goes beyond the permitted region ( for example , if the particle jumps into the site @xmath25 , it is pushed back to @xmath26 ) .
this system is a continuous - time version of the discrete - time random walks ( dtrws ) with reflecting barriers @xcite .
furthermore , successive waiting times of the particle @xmath27 between jumps are assumed to be mutually independent and follow the tsd @xmath28 .
the laplace transform of the tsd , @xmath29 , is given by @xmath30 \right ) , \label{e.laplace.of.ptl}\ ] ] where @xmath31 is the stable index , @xmath32 is a scale factor , and @xmath33 is a parameter characterizing the smooth cutoff . note that when @xmath34 , this is the laplace transform of the one - sided stable distributions .
equivalently , the characteristic function of the tsd , @xmath35 , is given by @xmath36 \right),\ ] ] this characteristic function is a special case of the tsd given in @xcite .
this is because @xmath37 takes only positive values , and we need only one - sided distributions . more precise definition of the tsd and the derivations of the above equations are presented in appendix [ sec : app.tsd ] .
the tsd in real space is also derived explicitly as follows ( for a derivation , see appendix [ sec : app.tsd ] ) : @xmath38 where @xmath39 is the gamma function .
when @xmath40 , @xmath41 is the one - sided @xmath42-stable distribution with a power - law tail : @xmath43 as @xmath44 @xcite .
therefore , the tsd [ eq . ( [ e.prob.1b ] ) ] is the one - sided stable distribution multiplied by the exponential function @xmath45 : @xmath46 .
thus , @xmath15 behaves as @xmath47 when @xmath44 .
moreover , the @xmath48-times convoluted pdf @xmath49 , which is the pdf of the sum of the successive waiting times @xmath50 , is expressed by using @xmath51 : @xmath52 therefore , the @xmath48-times convoluted pdf @xmath53 is also explicitly derived from eqs .
( [ e.prob.1b ] ) and ( [ e.prob.na ] ) .
using eq .
( [ e.prob.na ] ) , we derive transient properties of ctrws , including various crossovers , in the following sections .
even though we use the tsd @xmath15 as a waiting time distribution , we should further specify the first waiting time @xmath54 , or equivalently , the initial ensemble ( here , the first waiting time @xmath54 is the time interval between the start of the measurement and the first jump .
we always assume that measurements start at @xmath55 . ) .
in this paper , we consider two kinds of initial ensembles .
the first one is a typical and most frequently used non - equilibrium ensemble , for which the first waiting times @xmath54 of the particles are chosen from @xmath56 .
the second one is the equilibrium ensemble , for which the first waiting times @xmath54 are chosen from the equilibrium waiting time distribution @xmath57 . here
, @xmath58 can be defined by its laplace transformation @xcite : @xmath59 where @xmath60 is the mean waiting time for @xmath15 .
note that the second and subsequent waiting times , @xmath61 , are chosen from @xmath15 for both ensembles .
numerical methods to generate random variables following @xmath15 and @xmath57 are summarized in appendix [ sec : num.method ] .
@xmath57 can be expressed analytically as follows @xmath62 where @xmath63 is defined by @xmath64 .
see appendix [ sec : app.tsd ] for a derivation of eq .
( [ e.prob.1b.eq ] ) .
the mean waiting time for @xmath58 is given by @xmath65 , which is much longer than @xmath66 if @xmath67 is small .
the confinement effects on ctrws with power - law waiting times were investigated in @xcite by using the fractional fokker - planck equation ( ffpe ) . here instead of the ffpe , we use gffpe , which was derived in @xcite , to incorporate the smooth cutoff into the waiting time distribution and to study the ensemble averages of the tamsd .
there are three timescales in the present model : ( 1 ) time interval @xmath20 , ( 2 ) total measurement time @xmath19 , and ( 3 ) timescale of the cutoff @xmath68 .
let us define the laplace variables @xmath69 and @xmath70 conjugate to @xmath20 and @xmath19 , respectively .
since we assume that @xmath21 ( @xmath71 ) , it is sufficient to consider the following three cases : @xmath72 \mathrm{(b ) } & \delta \ll 1/\lambda \ll t & ~~\longleftrightarrow~~ & s \ll \lambda \ll u , \\[.2 cm ] \mathrm{(c ) } & 1/\lambda \ll \delta \ll t & ~~\longleftrightarrow~~ & s \ll u \ll \lambda .
\end{array}\end{aligned}\ ] ] here it is expected that the standard random walk behavior arises in case ( c ) .
thus , we only study the cases ( a ) and ( b ) in this paper . in this subsection ,
we rewrite the ensemble averages of tamsd @xmath73 and @xmath74 by using the eamsd @xmath75 ) and ( [ e.confine.ea-tamsd.laplace.eq ] ) ] . here and in the followings , we use the bracket @xmath76 for the average over the non - equilibrium initial ensemble , while @xmath77 for the average over the equilibrium initial ensemble .
also , we assume that initial position @xmath78 is uniformly distributed on the lattice @xmath79 for both ensembles .
first , let @xmath80 be the waiting time distribution of a renewal process with a finite mean , @xmath81 .
moreover , we define @xmath82 as the pdf of the forward recurrence time @xmath37 @xcite , i.e. , @xmath82 is the waiting time distribution at time @xmath18 . here , note that @xmath18 is not necessarily a renewal time .
particularly , @xmath83 for the non - equilibrium initial ensemble , since we assume that we start measurements at @xmath55 , whereas @xmath84 in general .
by contrast , for the equilibrium ensemble , @xmath85 where @xmath86 is the pdf of the forward recurrence time at time @xmath18 for the equilibrium ensemble , and @xmath87 is defined through its laplace transformation @xmath88 [ see eq .
( [ e.laplace.of.ptl.eq ] ) ] .
this relation [ eq .
( [ e.time_translation_inv ] ) ] is obvious because of the time - translation invariance of the equilibrium state , and a proof is given in appendix [ sec : app.time.trans ] .
let us begin with the non - equilibrium ensemble . using the pdf @xmath89
, we can express the ensemble average of the displacement during @xmath90 $ ] as follows : @xmath91 ^ 2 \right\rangle \notag\\[0.1 cm ] & & \quad= \sum_{l = \pm 1}^{}\frac { 1}{2 } \int_0^{\delta } d\tau w_{\mathrm{e}}(\tau ; t ' ) \left\langle \left|x(t'+\delta ) - x(t'+\tau ) + l\right|^2 \right\rangle \notag\\[0.1 cm ] \label{e.confine.frec_time } & & \quad= \int_0^{\delta } d\tau w_{\mathrm{e}}(\tau ; t ' ) \left[\left\langle \left|x(\delta - \tau ) - x(0)\right|^2 \right\rangle + 1\right].~~\end{aligned}\ ] ] this is an exact relation , while an approximated version of this equation is already presented in ref.@xcite .
we also show a more detailed derivation in appendix [ sec : app.eq10 ] . from eqs .
( [ e.tamsd.1 ] ) and ( [ e.confine.frec_time ] ) , we have @xmath92 & & \quad= % % \int_0^{t-\delta } dt ' % % \int_0^{\delta } d\tau \frac { w_{\mathrm{e}}(\tau ; t')}{t - \delta } \langle x^2(\delta - \tau ) \rangle % % \nonumber\\[0.2 cm ] % % & = & \int_0^{\delta } d\tau \bar{w}_{\mathrm{e}}(\tau ; t ) \left [ \left\langle \left|x(\delta - \tau ) - x(0)\right|^2 \right\rangle + 1 \right ] , \qquad \label{e.confine.ea - tamsd}\end{aligned}\ ] ] where @xmath93 is defined as @xmath94 from eq .
( [ e.confine.ea-tamsd ] ) , the ensemble average of tamsd is given by the convolution of @xmath93 and the eamsd @xmath95 .
in this and the following subsections , we study these two factors using laplace transformations . here
we further rewrite eqs .
( [ e.confine.ea-tamsd ] ) and ( [ e.confine.w_bar ] ) using laplace transformations .
first , the laplace transform of eq .
( [ e.confine.w_bar ] ) with respect to @xmath37 gives @xmath96 where we have defined the laplace transformations of @xmath97 and @xmath93 as @xmath98 and @xmath99 , respectively .
furthermore , the laplace transform of eq .
( [ e.confine.w_bar.laplace ] ) with respect to @xmath19 gives @xmath100(s ) % % \nonumber\\[0.2 cm ] % % & = & % % e^{- s\delta } % % \int_s^{\infty } ds ' % % \mathcal{l } % % \left [ % % \int_0^{t } dt ' \tilde{w}_{\mathrm{e}}(u ; t ' ) % % \right](s ' ) % % \nonumber\\[0.2 cm ] \label{e.confine.w_bar.laplace.2 } = e^{- s\delta } \int_s^{\infty } ds ' \frac { \breve{w}_{\mathrm{e}}(u ; s')}{s'},\ ] ] where the double laplace transformations @xmath101 and @xmath102 are defined as @xmath103 and @xmath104 , respectively .
in addition , by taking the laplace transformation of eq .
( [ e.confine.ea-tamsd ] ) with respect to @xmath20 , we obtain @xmath105 ( u , t ) \notag\\[0.1 cm ] \label{e.confine.ea - tamsd.laplace } & & \quad= \tilde{\bar{w}}_{\mathrm{e}}(u ; t ) \left\ { \,\mathcal{l } \left [ \left\langle \left|x(\delta ) - x(0)\right|^2 \right\rangle \right ] ( u ) + \frac { 1}{u } \right\}.\end{aligned}\ ] ] for the non - equilibrium ensemble , the following relation between the waiting time distribution @xmath106 and forward recurrence time distribution @xmath107 is well known @xcite : @xmath108 see appendix [ sec : app.time.trans ] for a derivation . if we choose tsd [ eq .
( [ e.prob.1b ] ) ] for the waiting time distribution @xmath80 , we have @xmath109 where we have used eq .
( [ e.laplace.of.ptl ] ) and @xmath110 . using the eq .
( [ e.confine.w_e.tlf ] ) and @xmath111 , the integral on the rhs of eq .
( [ e.confine.w_bar.laplace.2 ] ) can be approximated as @xmath112 \displaystyle \frac { \lambda^{1-\alpha } u^{\alpha-1}}{\alpha s } , & ~\mathrm{for}~~s
\ll \lambda \end{array } \right.\end{aligned}\ ] ] then , the inverse laplace transformation of eq .
( [ e.confine.w_bar.laplace.2 ] ) with respect to @xmath70 gives @xmath113 \displaystyle \frac { \lambda^{1-\alpha } u^{\alpha-1}}{\alpha } , & ~\mathrm{for}~~1/\lambda \ll t \end{array } \right.\end{aligned}\ ] ] where we have used @xmath114 [ note that @xmath115 because of eq .
( [ e.confine.timescales.2 ] ) ] . for the case of the equilibrium ensemble , eq .
( [ e.confine.frec_time ] ) should be replaced by @xmath116 ^ 2 \right\rangle_{\mathrm{eq } } \notag\\[0.1 cm ] \label{e.confine.frec_time.eq } & & \qquad= \int_0^{\delta } d\tau w^{\mathrm{eq}}(\tau ) \left [ \left\langle \left|x(\delta - \tau ) -
x(0 ) \right|^2\right\rangle + 1 \right].~~\end{aligned}\ ] ] note that the ensemble average in the right - hand side ( rhs ) is taken over the non - equilibrium ensemble .
since rhs of the above equation is independent of @xmath18 , we have the following equation through the calculation similar to that in the non - equilibrium case : @xmath117 ( u ) \notag\\[0.1 cm ] \label{e.confine.ea - tamsd.laplace.eq } & & \qquad= \tilde{w}^{\mathrm{eq}}(u ) \left\ { \mathcal{l } \left [ \left\langle \left|x(\delta ) - x(0 ) \right|^2\right\rangle \right ] ( u ) + \frac { 1}{u } \right\}.\quad\end{aligned}\ ] ] moreover , the left - hand - side ( lhs ) of eq .
( [ e.confine.frec_time.eq ] ) with @xmath55 is just the eamsd with respect to the equilibrium initial ensemble .
thus , using eq .
( [ e.tamsd.1 ] ) we have @xmath118 ^ 2 \right\rangle_{\mathrm{eq } } = \langle \overline{(\delta x)^2 } ( \delta ) \rangle_{\mathrm{eq } } \label{e.confine.ergode.eq}\ ] ] this is a manifestation of ergodicity for the equilibrium ensemble . note that eq .
( [ e.confine.ergode.eq ] ) is valid in general , if @xmath119 is finite ( and thus @xmath87 exists ) . if we choose tsd [ eqs .
( [ e.prob.1b ] ) and ( [ e.laplace.of.ptl ] ) ] for the waiting time distribution @xmath80 , we have the following relation for @xmath120 @xcite : @xmath121 } } { \left\langle \tau \right\rangle u } \simeq \frac { 1}{\alpha } \left ( \frac { \lambda}{u}\right)^{1 - \alpha},\ ] ] where we used the assumption @xmath122 [ eq . ( [ e.confine.timescales.2 ] ) ] and @xmath123 . finally , we obtain the following estimate for the laplace transformation of eamsd : @xmath138 ( u ) \simeq \begin{cases } % % \frac { 2 l^2}{\pi^2 u ( \delta_c u ) ^{\alpha } } = \dfrac { 1}{cu^{1+\alpha } } , & \mathrm{for}\ , \delta_c u \gg 1 \\[0.35 cm ] \dfrac { l^2}{6u } , & \mathrm{for}\ , \delta_c u \ll 1 , \end{cases}\end{aligned}\ ] ] where we used the rhs of eq .
( [ e.confine.eamsd.laplace.2 ] ) for the case of @xmath139 , while we rewrote the rhs of eq .
( [ e.confine.eamsd.laplace.2 ] ) by using zeta functions ( see appendix [ sec : appe - riemann ] ) for the case of @xmath140 as @xmath141 % % \bigg\ } % % \nonumber\\[0.2 cm ] % % & = & % % \frac { 16l^2}{\pi^{4 } u } % % \sum_{\begin{subarray}{c}n=1 \\ n : \mathrm{odd}\end{subarray } } ^{\infty } % % \frac { 1 } { ( \delta_c u ) ^{\alpha } + n^{2}}\frac { 1}{n^{2 } } % % \nonumber\\[0.2 cm ] % % \label{e.confine.eamsd.laplace.3 } % % & = & \frac { 16l^2}{\pi^{4 } \delta_c^{\alpha } u^{1+\alpha } } \bigg [ \frac { \pi^{2}}{8 } - \sum_{\begin{subarray}{c}n=1 \\ n : \mathrm{odd}\end{subarray } } ^{\infty } \frac { 1 } { ( \delta_c u ) ^{\alpha } + n^{2 } } \bigg].\ ] ] then , the summation term can be neglected since @xmath140 . from eq .
( [ e.confine.eamsd.laplace.4 ] ) , we obtain the eamsd for non - equilibrium ensemble as follows : @xmath142 \dfrac { l^2}{6 } , & \mathrm{for}\ , \delta \gg \delta_c .
\end{cases}\end{aligned}\ ] ] as shown in the next subsection , the tamsd for the non - equilibrium ensemble behaves differently .
namely , the ergodicity is broken even at @xmath143 .
in contrast , the ergodicity is satisfied for the equilibrium ensemble at @xmath144 . in this subsection
, we derive asymptotic behavior of the ensemble - averaged tamsd using the results from the preceding subsections .
( color online ) ( a ) tamsd @xmath145 vs. time interval @xmath20 in log log form for the non - equilibrium ensemble ( a ) : @xmath146 .
the total measurement time @xmath19 is set as @xmath147 , and the cutoff parameter @xmath67 as @xmath148 .
other parameters are set as @xmath149 , and @xmath150 .
tamsd is calculated for 8 different realizations of trajectories and different symbols correspond to different realizations .
( b ) the ensemble average of tamsd in log log form ( circles ) .
the lines are the theoretical predictions given by eq .
( [ e.confine.ea-tamsd.final.1.real ] ) .
note that no adjustable parameters were used to obtain these theoretical lines .
, width=309 ] ( color online ) ( a ) tamsd @xmath145 vs. time interval @xmath20 in log log form for the non - equilibrium ensemble ( a ) : @xmath146 .
the total measurement time @xmath19 is set as @xmath147 , and the cutoff parameter @xmath67 as @xmath148 .
other parameters are set as @xmath149 , and @xmath150 .
tamsd is calculated for 8 different realizations of trajectories and different symbols correspond to different realizations .
( b ) the ensemble average of tamsd in log log form ( circles ) .
the lines are the theoretical predictions given by eq .
( [ e.confine.ea-tamsd.final.1.real ] ) .
note that no adjustable parameters were used to obtain these theoretical lines . , width=309 ] first , we start with the non - equilibrium ensemble for @xmath152 . from eqs .
( [ e.confine.ea-tamsd.laplace ] ) , ( [ e.confine.bar_w_e ] ) , and ( [ e.confine.eamsd.laplace.4 ] ) , we have leading terms in @xmath153 and @xmath154 as follows : @xmath155 ( u , t ) \notag\\[0.1 cm ] \label{e.confine.ea - tamsd.final.1 } & \qquad= \begin{cases } \dfrac { u^{- 2}}{c\gamma ( 1+\alpha ) t^{1-\alpha } } , & \mathrm{for}~ \delta_c
u \gg 1 \\[0.4 cm ] \dfrac { l^{2}}{6 } \dfrac { u^{\alpha-2}}{\gamma ( 1+\alpha ) t^{1 -\alpha } } , & \mathrm{for}~ \delta_c u \ll 1 . \end{cases}\end{aligned}\ ] ] the inverse laplace transformation with respect to @xmath69 gives @xmath156 \displaystyle \frac { l^{2}\delta^{1 -\alpha } } { 6\gamma(1+\alpha ) \gamma ( 2 - \alpha ) t^{1 -\alpha } } , & \mathrm{for}~ \delta_c \ll \delta .
\end{array } \right.\end{aligned}\ ] ] thus , the ensemble - averaged tamsd shows normal diffusion at a short timescale ( @xmath157 ) and anomalous slow diffusion at a longer timescale ( @xmath158 ) .
as expected , these results perfectly coincide with those of the previous studies [ eqs .
( 10 ) and ( 11 ) in @xcite ] .
also , in this regime , the tamsd depends on the measurement time @xmath19 .
that is , the diffusion becomes slower , as the measurement time increases .
we call this behavior _ aging _ in this article . in fig .
[ f.confine.msd.1](a ) , tamsds for 8 different trajectories are shown .
although these tamsds show similar scaling behavior , the diffusion constant @xmath22 of each tamsd @xmath159 seems randomly distributed .
this behavior is analyzed in the next section . in fig .
[ f.confine.msd.1 ] ( b ) , the ensemble - averaged tamsd @xmath160 is displayed .
the solid and dashed lines are the theoretical predictions given by eq .
( [ e.confine.ea-tamsd.final.1.real ] ) .
( color online ) ( a ) tamsd @xmath145 vs. time interval @xmath20 in log log form for the non - equilibrium ensemble ( b ) : @xmath161 .
the total measurement time @xmath19 is set as @xmath162 , and the cutoff parameter @xmath67 as @xmath148 .
the other parameters are the same as in fig .
[ f.confine.msd.1 ] .
tamsd is calculated for 8 different realizations of trajectories and different symbols correspond to different realizations .
( b ) the ensemble average of tamsd in log log form ( circles ) .
the lines are theoretical predictions given by eq .
( [ e.confine.ea-tamsd.final.2.real ] ) .
no adjustable parameters were used to obtain these theoretical lines .
, width=309 ] ( color online ) ( a ) tamsd @xmath145 vs. time interval @xmath20 in log log form for the non - equilibrium ensemble ( b ) : @xmath161 .
the total measurement time @xmath19 is set as @xmath162 , and the cutoff parameter @xmath67 as @xmath148 .
the other parameters are the same as in fig .
[ f.confine.msd.1 ] .
tamsd is calculated for 8 different realizations of trajectories and different symbols correspond to different realizations .
( b ) the ensemble average of tamsd in log log form ( circles ) .
the lines are theoretical predictions given by eq .
( [ e.confine.ea-tamsd.final.2.real ] ) .
no adjustable parameters were used to obtain these theoretical lines .
, width=309 ] next , we study the case of @xmath164 . from eqs .
( [ e.confine.ea-tamsd.laplace ] ) , ( [ e.confine.bar_w_e ] ) , and ( [ e.confine.eamsd.laplace.4 ] ) , we have @xmath165 ( u , t ) = \left\ { \begin{array}{ll } \displaystyle \frac { 1}{c\lambda^{\alpha-1}\alpha } \frac { 1}{u^{2 } } , & \mathrm{for}~ \delta_c u \gg 1 \\[0.5 cm ] \displaystyle \frac { 1}{\lambda^{\alpha-1}\alpha } \,\frac { l^2}{6u^{2-\alpha } } , & \mathrm{for}~ \delta_c u \ll 1 . \end{array } \right.\end{aligned}\ ] ] then , taking the inverse laplace transformation , we obtain @xmath166 \displaystyle \frac { 1 } { \lambda^{\alpha-1 } \alpha } \frac { l^2\delta^{1 -\alpha}}{6\gamma ( 2 - \alpha ) } , & \mathrm{for}~ \delta_c \ll \delta . \end{array } \right.\end{aligned}\ ] ] thus , a crossover from the normal to anomalous diffusion similar to that in case ( a ) can be observed even in the long measurement time limit @xmath143 , whereas the aging behavior@xmath19-dependence of the diffusion constant vanishes .
note also that even in this limit , @xmath144 , the ( ensemble - averaged ) tamsd does not coincide with the eamsd [ eq . ( [ e.confine.eamsd.5 ] ) ] .
[ f.confine.msd.2](a ) shows tamsds for 8 different trajectories . in this case , the scatter of tamsds , observed in fig .
[ f.confine.msd.1](a ) , is diminished .
the ensemble - averaged tamsd is also shown in fig .
[ f.confine.msd.2](b ) by circles , where the theoretical predictions given by eq .
( [ e.confine.ea-tamsd.final.2.real ] ) are shown by solid and dashed lines . for the case of the equilibrium initial ensemble
, we have the following relation from eqs .
( [ e.confine.ea-tamsd.laplace.eq ] ) , ( [ e.confine.w_e.tlf.eq ] ) , and ( [ e.confine.eamsd.laplace.4 ] ) : @xmath167 ( u ) \simeq \begin{cases } \dfrac { 1}{c \lambda^{\alpha-1 } \alpha } \dfrac { 1}{u^2 } , & \text{for}~ \delta_c u \gg 1 \\[0.5 cm ] \dfrac { 1}{\lambda^{\alpha-1}\alpha } \,\dfrac { l^2}{6u^{2-\alpha } } , & \text{for}~ \delta_c u \ll 1 . \end{cases}\end{aligned}\ ] ] through the inverse transformation , we have @xmath168 \dfrac { 1}{\lambda^{\alpha-1}\alpha } \dfrac { l^2 \delta^{1-\alpha}}{6\gamma(2-\alpha ) } , & \text{for}~ \delta \gg \delta_c . \end{cases}\end{aligned}\ ] ] the equation ( [ e.confine.ea-tamsd.eq ] ) is equivalent to eq .
( [ e.confine.ea-tamsd.final.2.real ] ) , the tamsd for the non - equilibrium ensemble [ case ( b ) ] .
however , eq . ( [ e.confine.ea-tamsd.eq ] ) is valid for arbitrary measurement times @xmath19 , while eq .
( [ e.confine.ea-tamsd.final.2.real ] ) is valid only for long measurement times ( @xmath163 ) .
in addition , the aging behavior is absent in the equilibrium case . in fig .
[ f.confine.msd.3](a ) and ( b ) , tamsds for 8 different trajectories are shown for a short and long measurement times @xmath19 , respectively . surprisingly , the scatter of tamsd is even broader than that in the non - equilibrium case [ fig .
[ f.confine.msd.1](a ) ] at short measurement times @xmath19 as shown in fig .
[ f.confine.msd.3](a ) . in summary , for the non - equilibrium ensemble , the scatter of tamsd appears with the aging behavior [ the non - equilibrium ensemble ( case a ) ] , whereas for the equilibrium ensemble the scatter appears without aging .
these scatters are also studied quantitatively through theoretical analysis in the next section .
the theoretical predictions given by eq .
( [ e.confine.ea-tamsd.eq ] ) are also shown by dotted and dashed lines in fig .
[ f.confine.msd.3](b ) . furthermore , from eq .
( [ e.confine.ergode.eq ] ) , the eamsd for the equilibrium initial ensemble @xmath169 ^ 2 \right\rangle_{\mathrm{eq}}$ ] is also given by the rhs of eq .
( [ e.confine.ea-tamsd.eq ] ) . in fig .
[ f.confine.msd.3](b ) , a numerically obtained eamsd @xmath169 ^ 2 \right\rangle_{\mathrm{eq}}$ ] is displayed by a solid curve , which is consistent with the theory [ eq .
( [ e.confine.ea-tamsd.eq ] ) : the dotted and dashed lines in fig .
[ f.confine.msd.3](b ) ] .
thus , the eamsd for the equilibrium ensemble @xmath170 ^ 2 \right\rangle_{\mathrm{eq}}$ ] [ eq .
( [ e.confine.ea-tamsd.eq ] ) ] is different from the eamsd for the non - equilibrium ensemble @xmath171 ^ 2 \right\rangle$ ] [ eq .
( [ e.confine.eamsd.5 ] ) ] .
( color online ) ( a ) tamsd @xmath145 vs. time interval @xmath20 in log log form for the equilibrium ensemble with @xmath146 .
the total measurement time @xmath19 is set as @xmath147 , and the cutoff parameter @xmath67 as @xmath148 .
the other parameters are the same as in fig .
[ f.confine.msd.1 ] .
tamsd is calculated for 8 different realizations of trajectories and different symbols correspond to different realizations .
( b ) the same as the figure ( a ) except that @xmath162 ( @xmath172 ) .
the symbols are tamsds for 8 different trajectories and the thick solid curve is the eamsd @xmath173 ^ 2 \right\rangle_{\mathrm{eq}}$ ] .
the dotted and dashed lines are theoretical predictions given by eq .
( [ e.confine.ea-tamsd.eq ] ) .
note that no adjustable parameters were used to obtain these theoretical lines .
, width=309 ] ( color online ) ( a ) tamsd @xmath145 vs. time interval @xmath20 in log log form for the equilibrium ensemble with @xmath146 .
the total measurement time @xmath19 is set as @xmath147 , and the cutoff parameter @xmath67 as @xmath148 .
the other parameters are the same as in fig .
[ f.confine.msd.1 ] .
tamsd is calculated for 8 different realizations of trajectories and different symbols correspond to different realizations .
( b ) the same as the figure ( a ) except that @xmath162 ( @xmath172 ) .
the symbols are tamsds for 8 different trajectories and the thick solid curve is the eamsd @xmath173 ^ 2 \right\rangle_{\mathrm{eq}}$ ] .
the dotted and dashed lines are theoretical predictions given by eq .
( [ e.confine.ea-tamsd.eq ] ) .
note that no adjustable parameters were used to obtain these theoretical lines .
, width=309 ]
the statistical property of tamsd is dominated by the property of the number of jumps @xmath174 until time @xmath19 ( see sec . [
sec : stat.tamsd ] ) .
therefore , we first study @xmath174 in secs .
[ s.real-space-analysis ] and [ s.laplace-space-analysis ] on the basis of the analysis presented in @xcite . in the sec .
[ sec : tsd ] , we define @xmath175 as the time when the @xmath48-th jump occurs for a trajectory @xmath176 : @xmath177 . from this definition
, we have the following relation : @xmath178 & = & \mathrm{prob } ( t_n > t ) = \mathrm{prob } \left ( \sum_{k=1}^{n } \tau_k > t \right ) , \label{e.recipro.a}\end{aligned}\ ] ] where @xmath179 is the probability and @xmath27 are the waiting times between jumps .
( color online ) ( a ) pdf of the diffusion coefficient @xmath189 for the non - equilibrium tamsd @xmath190 for small @xmath20 .
each pdf is normalized so that its mean value equals unity .
@xmath189 is calculated from tamsd by least - square fitting over the interval @xmath191 .
the results for two different values of measurement times are presented : @xmath192 ( circles ) and @xmath193 ( squares ) .
the other parameters are set as @xmath194 , and @xmath150 .
the lines correspond to the theoretical predictions given by eq .
( [ e.tml ] ) .
no adjustable parameters were used to obtain these curves .
( b ) pdf of the diffusion coefficient @xmath189 for the equilibrium tamsd @xmath195 for small @xmath20 .
the parameter values are the same as in the figure ( a ) .
the lines are to guide the eye.,width=309 ] ( color online ) ( a ) pdf of the diffusion coefficient @xmath189 for the non - equilibrium tamsd @xmath190 for small @xmath20 .
each pdf is normalized so that its mean value equals unity .
@xmath189 is calculated from tamsd by least - square fitting over the interval @xmath191 .
the results for two different values of measurement times are presented : @xmath192 ( circles ) and @xmath193 ( squares ) .
the other parameters are set as @xmath194 , and @xmath150 .
the lines correspond to the theoretical predictions given by eq .
( [ e.tml ] ) .
no adjustable parameters were used to obtain these curves .
( b ) pdf of the diffusion coefficient @xmath189 for the equilibrium tamsd @xmath195 for small @xmath20 .
the parameter values are the same as in the figure ( a ) .
the lines are to guide the eye.,width=309 ] as shown below [ sec .
[ sec : stat.tamsd ] ] , tamsd behaves as @xmath196 for small @xmath20 .
thus , the diffusion constant @xmath189 , defined as @xmath197 , shows the same statistical property as that of @xmath174 .
[ f.confine.ml ] shows the pdf of @xmath189 for two different measurement times @xmath19 .
note that the pdf narrows as @xmath19 increases .
the analytical results given in eq .
( [ e.tml ] ) are also depicted by solid and dashed lines in the figure .
the equilibrium case is also analyzed in a similar way . in this case , eq .
( [ e.recipro.b ] ) should be replaced by @xmath198 where @xmath199 means a convolution .
in contrast to the non - equilibrium case , however , it seems difficult to obtain a simple expression of the pdf of @xmath174 for the equilibrium case . as shown in fig .
[ f.confine.ml ] , a qualitative difference appears for short measurement time regime , @xmath200 .
in fact , there is a peak at @xmath201 for the equilibrium initial ensemble , which is due to the mean waiting time of @xmath58 , @xmath202 ( @xmath203 ) , is much longer than that of @xmath15 , @xmath66 ( @xmath204 ) .
namely , there are many trajectories that are trapped more than the measurement time @xmath205 @xcite . next , to clarify the ergodic properties of the system , we study the relative standard deviation ( rsd ) @xmath206 , where @xmath207 is the cumulant . the quantity @xmath208 is a measure of ergodicity @xcite .
in fact , if @xmath209 , time averages of an observable give the same value independent of the trajectory ; however , if @xmath210 , they are different from one trajectory to another .
this quantity @xmath208 was also used in some molecular dynamics simulations to characterize the ergodicity breaking and non - gaussian fluctuations of lipid motions in cell membranes @xcite , and to determine the longest relaxation time in entangled polymers @xcite .
let us begin with the non - equilibrium ensemble .
in order to derive an analytical expression of @xmath208 , we start with the laplace transformation of eq .
( [ e.recipro.b ] ) : @xmath211 & = & \frac { 1 - e^{-nc[(\lambda + s)^{\alpha } - \lambda^{\alpha}]}}{s},\end{aligned}\ ] ] where @xmath212 is the laplace transformation of @xmath213 in terms of @xmath19 .
next , let us define a function @xmath214 as @xmath215 .
note that @xmath214 is the laplace transformation ( with respect to @xmath19 ) of the pdf of @xmath174 .
furthermore , we perform a discrete laplace transformation of @xmath214 with respect to @xmath48 as follows : @xmath216 \right ) } { 1- \exp \left ( - \nu - c[(\lambda + s)^{\alpha } - \lambda^{\alpha } ] \right)},\end{aligned}\ ] ] where we define @xmath217 as @xmath218 . using the assumption @xmath219 , we have @xmath220 } % % { \nu + c[(\lambda + s)^{\alpha } - \lambda^{\alpha } ] } % % \nonumber\\[.2 cm ] % % & = & % % \frac { 1}{s } % % \times % % \frac % % { 1 } % % { 1 + \frac { \nu}{c[(\lambda + s)^{\alpha } - \lambda^{\alpha } ] } } % % \nonumber\\[.2 cm ] % % & = & = \frac { 1}{s } \sum_{k=0}^{\infty } \left(- \frac { \nu}{c}\right)^k \left[(\lambda + s)^{\alpha } - \lambda^{\alpha}\right]^{-k}.\end{aligned}\ ] ] [ [ first - moments ] ] first moments + + + + + + + + + + + + + in addition , @xmath222 is usually called the renewal function in the renewal theory , and eqs .
( [ e.laplace.1order.a ] ) and ( [ e.laplace.1order.b ] ) can also be derived by the renewal equation @xcite . in contrast , higher order moments of @xmath174 can not be derived from the renewal equation , and we have to use eq .
( [ e.laplace.3 ] ) .
[ [ second - moments ] ] second moments + + + + + + + + + + + + + + similarly , the laplace transform @xmath227(s)$ ] of the second moment @xmath228 is given by @xmath229(s ) % % & \simeq & % % \frac { 2}{c^{2}s}~ % % \frac { 1}{\left [ ( \lambda + s)^{\alpha } - \lambda^{\alpha } \right]^{2 } } % % \nonumber\\[.2 cm ] & \simeq & \left\ { \begin{array}{ll } \displaystyle \frac { 2}{c^{2}s^{2\alpha+1 } } , & \hspace*{.1 cm } s \gg \lambda \\[.3 cm ] \displaystyle \frac { 2}{c^{2}\lambda^{2\alpha-2 } \alpha^{2 } s^{3 } } \left [ 1 + ( 1 - \alpha ) \frac { s}{\lambda}\right ] , & \hspace*{.1 cm } s \ll \lambda .
\end{array } \right .
\nonumber\\[.0cm]\end{aligned}\ ] ] then , the inverse transform is given by @xmath230 \displaystyle \frac { 2}{c^{2 } \lambda^{2\alpha-2 } \alpha^{2 } } \left [ \frac { t^{2}}{2 } + \frac { 1-\alpha}{\lambda}t\right ] , & \hspace*{.3 cm } t \gg 1/\lambda .
\end{array } \right.\end{aligned}\ ] ] [ [ relative - standard - deviation ] ] relative standard deviation + + + + + + + + + + + + + + + + + + + + + + + + + + + using eqs .
( [ e.laplace.1order.b ] ) and ( [ e.laplace.2order.b ] ) , we obtain asymptotics of the rsd @xmath231 : @xmath232 \displaystyle \sqrt { \frac { 1 - \alpha}{\lambda t } } & \hspace*{.3 cm } t \gg 1/\lambda .
\end{array } \right.\end{aligned}\ ] ] we define a crossover time @xmath233 between the two regimes , @xmath151 and @xmath234 , as the intersection of the two functions in eqs .
( [ e.relative.sd ] ) ; therefore , we have @xmath235 as shown in fig .
[ f.confine.rsd ] , the rsd remains almost constant before the crossover time @xmath233 , and starts decaying rapidly after the crossover . in the figure ,
the rsd for the exponential waiting - time distribution , which has the same mean waiting time @xmath236 as that of the tsd with @xmath237 , is also shown by pluses .
it is clear that the rsd for the exponential distribution ( pluses ) decays much more rapidly than that for the tsd ( triangles ) .
( color online ) ( a ) rsd @xmath238 vs. total measurement time @xmath19 for the non - equilibrium initial ensemble .
@xmath189 is calculated from the tamsd @xmath145 by a least - square fitting over the interval @xmath239 .
three different values of @xmath67 are used : @xmath240 ( circles ) , @xmath241 ( squares ) , and @xmath242 ( triangles ) .
in addition , @xmath42 , @xmath32 , and @xmath26 are set as @xmath243 , @xmath244 , and @xmath245 , respectively .
the lines correspond to the theoretical predictions given by eq .
( [ e.relative.sd ] ) ; the solid line is the result for short timescales @xmath246 , and the dashed lines are for long timescales @xmath247 .
the intersections of the solid and dashed lines correspond to the crossover times @xmath233 given by eq .
( [ e.tc ] ) .
( b ) rsd @xmath248 vs. total measurement time @xmath19 for the equilibrium initial ensemble .
the parameter values are the same as those in the figure ( a ) , except that the results only for two different values of @xmath67 are plotted for clarity : @xmath249 ( squares ) , and @xmath242 ( triangles ) .
the lines correspond to the theoretical predictions given by eq .
( [ e.relative.sd.eq ] ) .
( a ) and ( b ) the plus signs in both figures are the rsd for the case in which the waiting time distribution is given by the exponential distribution @xmath250 with the same mean waiting time as that of the tsd with @xmath237 ( triangles ) : @xmath251 .
the dot dashed line is a theoretical prediction for the exponential distribution : @xmath252 .
note that the scales of vertical axes of two figures are different . for comparison ,
the theoretical prediction of rsd for the non - equilibrium ensemble at short timescales is depicted by a dotted line in the figure ( b).,width=309 ] ( color online ) ( a ) rsd @xmath238 vs. total measurement time @xmath19 for the non - equilibrium initial ensemble .
@xmath189 is calculated from the tamsd @xmath145 by a least - square fitting over the interval @xmath239 .
three different values of @xmath67 are used : @xmath240 ( circles ) , @xmath241 ( squares ) , and @xmath242 ( triangles ) .
in addition , @xmath42 , @xmath32 , and @xmath26 are set as @xmath243 , @xmath244 , and @xmath245 , respectively .
the lines correspond to the theoretical predictions given by eq .
( [ e.relative.sd ] ) ; the solid line is the result for short timescales @xmath246 , and the dashed lines are for long timescales @xmath247 .
the intersections of the solid and dashed lines correspond to the crossover times @xmath233 given by eq .
( [ e.tc ] ) .
( b ) rsd @xmath248 vs. total measurement time @xmath19 for the equilibrium initial ensemble .
the parameter values are the same as those in the figure ( a ) , except that the results only for two different values of @xmath67 are plotted for clarity : @xmath249 ( squares ) , and @xmath242 ( triangles ) .
the lines correspond to the theoretical predictions given by eq .
( [ e.relative.sd.eq ] ) .
( a ) and ( b ) the plus signs in both figures are the rsd for the case in which the waiting time distribution is given by the exponential distribution @xmath250 with the same mean waiting time as that of the tsd with @xmath237 ( triangles ) : @xmath251 .
the dot dashed line is a theoretical prediction for the exponential distribution : @xmath252 .
note that the scales of vertical axes of two figures are different .
for comparison , the theoretical prediction of rsd for the non - equilibrium ensemble at short timescales is depicted by a dotted line in the figure ( b).,width=309 ] the calculation for the equilibrium case is almost parallel to that for the non - equilibrium case except that we should use eq .
( [ e.recipro.b.eq ] ) instead of eq .
( [ e.recipro.b ] ) .
thus , we only show the final results .
first , the generating function @xmath217 is given by @xmath253^{-k+1}.\ ] ] using this function , we obtain @xmath254 @xmath255 and the rsd for the equilibrium initial ensemble @xmath256 \displaystyle \sqrt { \frac { 1 - \alpha}{\lambda t } } & \hspace*{.3 cm } t \gg 1/\lambda .
\end{array } \right.\end{aligned}\ ] ] although the rsd slowly decays at the short measurement timescales , the relative fluctuations are even larger than the non - equilibrium case as shown in fig .
[ f.confine.rsd ] . in this section ,
we discuss the statistical properties of tamsd for one - dimensional ctrws , although the analysis for free ctrws in the followings can be generalized for higher dimensional systems @xcite .
properties applicable to a general class of observables are summarized in appendix [ sec : hopf ] . in the absence of the confinement effect @xcite ( i.e. , without reflecting boundaries )
, the relations @xmath272 and @xmath273 for @xmath274 hold because of the mutual independence of @xmath275 , @xmath276 and @xmath277 .
in addition , since @xmath275 is independent of the initial ensemble , the same relations can be obtain also for the equilibrium ensemble : @xmath278 and @xmath279 . then , @xmath280 [ eq . ( [ e.hopf.hk ] ) ] satisfies the law of large numbers [ eq .([e.app.hopf.2 ] ) ] ; therefore , we have @xmath281 for arbitrary @xmath20 . note that this result is independent of the choice of the initial ensembles .
therefore , tamsd increases in proportion to @xmath20 , and has the same statistical property as @xmath174 . in particular , the diffusion constant @xmath189 is given by @xmath282 furthermore , the pdf of @xmath189 is given by eq .
( [ e.tml ] ) , and the rsd of @xmath189 shows the asymptotics such as the ones given in eq .
( [ e.relative.sd ] ) .
although a similar property holds for the system with reflecting boundaries , the calculation becomes more complicated , because @xmath283 are not mutually independent .
here , we present only an outline of the proof .
the main tool is the spectral decomposition of the @xmath48-time transition matrix @xmath284 of the dtrws with reflecting boundaries @xcite : @xmath285 where @xmath26 is the total number of sites .
the ket @xmath286 is the @xmath26-dimensional eigenvector of the transition matrix defined by @xmath287 for @xmath288 , and @xmath289 for @xmath290 . here
, @xmath291 is a site index ( 1 @xmath292 ) .
also , the bra @xmath293 is defined by the transpose of @xmath286 .
the real numbers @xmath294 are the associated eigenvalues : @xmath295 for @xmath288 and @xmath296 for @xmath297 .
thus , the first term ( @xmath288 ) in eq .
( [ e.decay.corr.tmatrix ] ) corresponds to the eigenmode with the unit eigenvalue ( the non - decaying mode ) , which is a uniform state , and the other terms decay exponentially fast to zero .
each element of the transition matrix , i.e. , @xmath298 , is the transition probability from the site @xmath299 to the site @xmath291 during @xmath48 jumps .
an important point is that each element of the matrix tends to @xmath300 exponentially fast as @xmath48 increases .
it follows that correlation functions for the dtrws decay exponentially , too . by using this fact ,
we can show ( after some lengthy calculations ) that the correlation functions for the ctrws @xmath301 , which can be expressed with correlation functions for the dtrws such as @xmath302 , also decay exponentially . as a result
, the law of large numbers holds for the @xmath280 even for the confined system , and we obtain @xmath303 the function @xmath304 can be derived by using the transition matrix [ eq .
( [ e.decay.corr.tmatrix ] ) ] with a suitable hydrodynamic limit , however an easier way here is to utilize the results in sec . [
sec : ea.tamsd ] .
for example , from eqs .
( [ e.confine.ea-tamsd.final.1.real ] ) and ( [ e.laplace.1order.b ] ) [ or eqs .
( [ e.confine.ea-tamsd.final.2.real ] ) and ( [ e.laplace.1order.b ] ) ] we obtain @xmath305 \dfrac { n_t}{t } \dfrac { cl^2\delta^{1 - \alpha}}{6\gamma ( 2 - \alpha ) } , & \text{for } \quad \delta \gg \delta_c , \end{cases}\end{aligned}\ ] ] for the non - equilibrium initial ensemble .
similarly , from eqs .
( [ e.confine.ea-tamsd.eq ] ) and ( [ e.mean.nt.eq ] ) , we have eq . ( [ e.tamsd.final ] ) also for the equilibrium ensemble .
thus , eq .
( [ e.tamsd.final ] ) is valid for both non - equilibrium and equilibrium ensembles .
thus , all the differences between the two ensembles are included in the statistical properties of @xmath174 in eq .
( [ e.tamsd.final ] ) .
see , for example , eqs .
( [ e.laplace.1order.b ] ) and ( [ e.mean.nt.eq ] ) .
up to now , two characteristic properties have been reported for confined ctrws @xcite : ( i ) there is a crossover from normal to anomalous diffusion , and ( ii ) tamsds are distributed depending on trajectories ( i.e. , weak ergodicity breaking ) .
these results are for the non - equilibrium ensemble [ there is no equilibrium ensemble for ctrws with power - law waiting times with stable index @xmath31 , because the system never reaches an equilibrium state . ] . in this paper
, in addition to this confinement effect , we incorporated a cutoff into the waiting time distribution .
for the case of the non - equilibrium ensemble , we analytically determined that the property ( i ) persists even in the long measurement time limit @xmath144 ( this fact was first found numerically in @xcite ) .
in contrast , the property ( ii ) , that is , the distributional behavior of tamsds , appears for short measurement times @xmath19 , whereas ergodicity holds for longer measurement times .
the important point is that as compared to common distributions such as exponential distribution , it takes a very long time to observe ergodic behavior for the case where the waiting time distribution is given by the tsd .
in addition , this transition from weak ergodicity breaking to ergodicity is a transition from an irreproducible regime to a reproducible regime .
furthermore , as shown in eq .
( [ e.tc ] ) , the crossover time @xmath233 between weak ergodicity breaking and ergodicity is proportional to @xmath68 [ also , the crossover time between anomalous and normal diffusion in eamsd for free ctrws is proportional to @xmath68 ; see the text below eq .
( [ e.laplace.1order.b ] ) ] .
because the mean waiting time @xmath66 is given by @xmath306 , we have @xmath307 thus , the crossover time @xmath233 is not proportional to the mean waiting time @xmath66 ; in fact , @xmath233 can be much longer than @xmath66 for @xmath42 close to 1 .
these facts may be important for estimating the crossover time in experiments @xcite .
in contrast to the ctrws with power - law waiting times with stable index @xmath31 , the ctrws with cutoff waiting times have an equilibrium state .
it is surprising that the crossover from normal to anomalous diffusion [ property ( i ) ] , and a scatter in tamsd [ property ( ii ) ] exist even for the case of the equilibrium initial ensemble .
the scatter in tamsd is even broader than the case of the non - equilibrium ensemble .
the main difference from the non - equilibrium case is that there is no aging for the equilibrium ensemble .
another difference between the two ensembles is that the rsd decays algebraically even in the short time regime for the equilibrium case , whereas it does not decay for the non - equilibrium case .
these properties for the equilibrium ensemble might fit well with some experimental results @xcite ( see the next section ) .
the important point is that the absence of aging in experimental data does not necessarily exclude the possibility of the ctrws .
we also presented a numerical method to generate the equilibrium ensemble ( appendix [ sec : num.method ] ) .
first , we compare the model studied in this paper ( in particular the equilibrium ctrws ) with the experimental data presented in @xcite . in @xcite , they studied lipid granules in a harmonic potential , and used confined non - equilibrium ctrws as a theoretical model .
they presented the following quantitative agreements between the experimental data and the model : ( 1 ) the crossover from normal to anomalous diffusion in tamsd and ( 2 ) the scatter distribution of the tamsd .
they set the total measurement time @xmath308 sec for these data , which seems to correspond to the case ( b ) in our classification [ eq .
( [ e.confine.timescales.2 ] ) : namely , a time regime after aging ] .
see the fig . 6 in the supplementary material of @xcite . however , there seems to be one disagreement between the experimental data and the non - equilibrium ctrws : i.e. , ( 3 ) the aging property .
in fact , the non - equilibrium ctrws show aging in short measurement times [ time regime ( a ) in eq .
( [ e.confine.timescales.2 ] ) ] , whereas the experimental data do not ( see the fig . 6 in the supplementary material ) . on the other hand , the equilibrium ctrws reproduce all three properties . in the time regime ( b )
( [ e.confine.timescales.2 ] ) ] , the equilibrium ctrws show quantitatively the same behavior as the non - equilibrium ctrws with respect to the properties ( 1 ) and ( 2 ) .
thus , the equilibrium ctrws will also reproduce these experimental observations successfully .
in addition , the equilibrium ctrws show no aging even in the time regime ( a ) , and thus the model agrees the experimental data with respect to the property ( 3 ) , too .
second , although we show that time averages for a class of observables including tamsd follow the ml distribution ( appendix [ sec : hopf ] ) , different distributions can arise for different classes of observables @xcite . as an example , we show in appendix [ sec : hopf ] a case in which long time averages follow the pdf called the bilateral ml distribution @xcite . third , there are qualitatively different types of random walk models in random environments . recently
, anomalous properties have been found in both experiments @xcite and molecular dynamics simulations @xcite .
the ctrws , which are a model of random walks in random environments , are often used as a model of these anomalous properties . however , there are similar random walk models in random environments with qualitatively different statistical behavior : for example , barrier models @xcite , random force models @xcite and the reptation model for entangled polymers @xcite . to the best of the authors knowledge , the ergodic properties , such as the behavior of rsd , of these systems are still unclear .
finally , besides random walks in random environments , there are still several different mechanisms for anomalous behavior , e.g. , gle , fbms , and diffusion on fractal structures @xcite .
therefore , it is important to develop techniques for time series analysis to elucidate which mechanism is the actual cause of the anomalous properties observed in various experiments and molecular dynamical simulations @xcite .
it is also important to investigate systems in which these mechanisms are combined @xcite , and those in which nonlinear dynamics are coupled with these anomalous mechanisms @xcite .
we thank s. shinkai for drawing our attention to ref.@xcite , and t. uneyama for fruitful discussions concerning possible connections between ctrws and entangled polymer systems .
in addition , we are grateful to an anonymous referee for his helpful comments on eq .
( [ e.confine.frec_time ] ) and appendix [ sec : app.eq10 ] .
this study is partially supported by a grant - in - aid for young scientists ( b ) ( 22740262 ) .
here we briefly introduce the one - sided tsd . in general , the infinitely divisible distributions @xmath309 is defined as follows : if @xmath310 is the characteristic function ( fourier transform ) of @xmath309 , then there exists a characteristic function @xmath311 such that @xmath312 .
note that this is not trivial , because the function @xmath313 is not necessarily a characteristic function of a pdf ( i.e. , a non - negative function ) .
now , we derive eq .
( [ e.prob.1b ] ) . from eqs .
( [ e.prob.1a ] ) and ( [ e.cfunc3 ] ) , we have @xmath330 where we have changed the integration path from the real to the imaginary axis by using the cauchy s integral theorem .
then , we have @xmath331 finally , we obtain eq .
( [ e.prob.1b ] ) by taylor expansion of the exponential function @xmath332 and the integral representation of the gamma function @xmath333 .
the equilibrium waiting time distribution ( [ e.prob.1b.eq ] ) can be derived by using eq .
( [ e.app.rv.equil ] ) as follows : @xmath334
let @xmath335 be a random variable following the tsd [ eq .
( [ e.prob.1b ] ) ] .
we briefly review the method to generate @xmath335 numerically according to @xcite .
first , a random variable @xmath336 , which follows the one - sided stable distributions , can be generated by the following equation @xcite : @xmath337^{(1-\alpha)/\alpha},\end{aligned}\ ] ] where @xmath338 is a uniform noise in @xmath339 $ ] , and @xmath340 is an exponential noise with mean 1 .
to generate @xmath335 , note that tsd [ eq . ( [ e.prob.1b ] ) ] is just a stable distribution multiplied by the exponential factor @xmath341 .
therefore , we first generate @xmath336 , then we accept it with a probability @xmath342 . if @xmath336 is rejected , it will be regenerated according to eq .
( [ e.rv_levy ] ) until it is accepted .
thereafter , the random variable that is finally accepted follows tsd @xcite . similarly , to generate equilibrium noise @xmath343
, first we create an auxiliary random variable @xmath344 by accepting @xmath345 with a probability @xmath346 .
then , @xmath343 is obtained by @xmath347 , where @xmath348 is a uniform noise in @xmath349 $ ] .
this can be proved as follows .
first , let us consider a time axis @xmath350 covered by many non - overlapping time intervals .
each interval is obtained from @xmath351 ( the intervals are assumed to be mutually independent ) . then , choose a time @xmath19 randomly from the time axis .
now , the random variable @xmath344 is defined by the length of the time interval which includes the randomly chosen time @xmath19 .
in addition , let @xmath352 be the pdf of the variable @xmath344 .
then , @xmath352 is given by @xmath353 because the probability with which an interval is chosen is proportional to the length of that interval @xmath37 .
thus , the random variable @xmath344 can be generated from the stable noise @xmath336 by accepting it with the probability @xmath354 ( the factor @xmath355 is added to make the maximum probability be unity for computational efficiency . ) .
moreover , by the laplace transform of the above equation , we have @xmath356 next , let us define a new variable @xmath357 , where @xmath348 is the uniform noise in @xmath349 $ ] .
then , we have @xmath358 by differentiating the both side , we obtain @xmath359 where @xmath360 is the pdf of @xmath361 .
the laplace transform of the above equation leads to @xmath362 inserting eq .
( [ e.app.rv.auxiliary ] ) into the above equation , we obtain @xmath363/\left\langle \tau \right\rangle s$ ] , which is the laplace transform of the equilibrium waiting time pdf [ eq .
( [ e.laplace.of.ptl.eq ] ) ] .
therefore , @xmath361 follows the equilibrium waiting time distribution @xmath364 .
this method to generate the equilibrium waiting time might be applicable to other probability densities with finite moments .
let @xmath80 be a waiting time distribution for a renewal process , and @xmath87 be the associated equilibrium waiting time distribution . here
, we show the relation @xmath365 [ eq .
( [ e.time_translation_inv ] ) ] by the method presented in @xcite .
first , let us define @xmath366 where @xmath367 if the inside of the bracket is satisfied , 0 otherwise .
in addition , @xmath368 are renewal times defined by @xmath369 , where @xmath370 are successive waiting times .
the laplace transform with respect to @xmath37 and @xmath18 gives @xmath371 & = \begin{cases } \dfrac { \tilde{w}^{\mathrm{eq } } ( u ) - \tilde{w}^{\mathrm{eq } } ( s)}{s - u } , & \text{for } \,\ , n=0 \\[0.1 cm ] \tilde{w}^{\mathrm{eq}}(s)\tilde{w}^{n-1 } ( s ) \dfrac { \tilde{w } ( u ) - \tilde{w } ( s ) } { s - u } , & \text{for } \,\
, n\geq 1 .
\end{cases}\end{aligned}\ ] ] because @xmath372 , we have @xmath373 here , if we replace @xmath374 with @xmath375 , we obtain the forward recurrence time distribution for the non - equilibrium ensemble [ eq . ( [ e.confine.w_e ] ) ] : @xmath376 it follows that @xmath377 by inserting eq .
( [ e.app.inv.w.eq ] ) into eq .
( [ e.app.inv.w_e.eq ] ) , we find that @xmath378 .
the inverse laplace transform gives the relation we want .
in this appendix , we present a detailed derivation of eq .
( [ e.confine.frec_time ] ) , which we owe to an anonymous referee . here , we consider the non - equilibrium initial ensemble , i.e. , @xmath379 is the renewal time ( the particle jumps at @xmath379 ) .
first , we write the lhs of eq .
( [ e.confine.frec_time ] ) as @xmath380 ^ 2 \right\rangle = \int_{-\infty}^{\infty}dx x^2 l(x;t ' , t'+\delta),\ ] ] where @xmath381 is the pdf of the displacement in the interval @xmath382 $ ] .
furthermore , @xmath381 can be rewritten as @xmath383 where @xmath384 is the probability of having @xmath48 jumps in the interval @xmath382 $ ] , @xmath385 is the pdf of single jump length , and @xmath386 is the @xmath48-times convoluted pdf of @xmath385 with @xmath387 .
note that we set @xmath388/2 $ ] in our model .
moreover , the following recursion relation is the most essential ingredient : @xmath389 & = \int_{0}^{\delta}d\tau w_e(\tau;t ' ) p ( n-1 ; 0 , \delta-\tau)\end{aligned}\ ] ] for @xmath390 . note that we can replace @xmath391 with @xmath392 ( and @xmath393 with @xmath394 ) , because both @xmath391 and @xmath392 are the renewal times . putting these equations together ,
we obtain @xmath395 ^ 2 \right\rangle\notag\\[0.2 cm ] & = \int_{0}^{\delta}d\tau w_e(\tau;t ' ) \sum_{n=0}^{\infty } p
( n ; 0 , \delta-\tau ) \int_{-\infty}^{\infty } dx x^2 l^{\ast ( n+1 ) } ( x).\end{aligned}\ ] ] now , if we put @xmath388/2 $ ] into the convolution of the above equation , @xmath396(x)$ ] , we have eq .
( [ e.confine.frec_time ] ) .
next , we assume that the pdf @xmath409 can be separable , i.e. , @xmath410 .
we further assume that @xmath411 in the hydrodynamic limit @xmath412 , where @xmath413 is the mean squared displacement of a single jump , i.e. , @xmath414
. for one - dimensional ctrws with jumps only to the nearest neighbor sites , we have @xmath415 .
moreover , we use the tsd , @xmath416 , as the waiting time distribution @xmath106 : @xmath417 % % \right ) % % \nonumber \\[0.2 cm ] & \simeq & 1 - c \left [ ( \lambda + u)^{\alpha } - \lambda^{\alpha } \right],\end{aligned}\ ] ] where @xmath418 is assumed . under these assumptions ,
( [ e.appc.fourier-laplace.3 ] ) can be rewritten as follows : @xmath419 } \tilde{p } ( k , u).\qquad\end{aligned}\ ] ] the inverse fourier and laplace transformations with respect to space and time lead to @xmath420 where @xmath132 is a constant given by @xmath421 , and @xmath422 is an operator defined by @xmath423 with function @xmath424 , which is defined by its laplace transform : @xmath425 as @xmath426 , eq .
( [ e.appc.ffpe.3 ] ) leads to the usual ffpe @xcite .
thus , the above equations ( [ e.appc.ffpe.3]) ( [ e.appc.ffpe.5 ] ) are called generalized ffpe @xcite . note also that gffpe is the equation for non - equilibrium initial ensemble .
in contrast , in order to describe the system started with the equilibrium ensemble , the multi - point fractal diffusion equation would be necessary @xcite .
the initial - boundary value problem of gffpe [ eq . ( [ e.appc.ffpe.3 ] ) ] under the boundary condition @xmath427 and initial condition @xmath428 can be solved by a standard method for the diffusion equation @xcite .
in fact , by assuming the separability @xmath429 , we obtain @xmath430 where @xmath431 is a constant .
@xmath432 satisfying the boundary condition [ eq .
( [ e.appd.mix.2 ] ) ] is given by @xmath433 and @xmath434 ( @xmath435 ) .
similarly , @xmath436 is derived in its laplace form : @xmath437 then , we have the solution in the following form : @xmath438 .
finally , @xmath439 in eq .
( [ e.appd.separable.t.laplace.2 ] ) can be determined from the initial condition [ eq .
( [ e.appd.mix.3 ] ) ] as @xmath440 finally , we obtain @xmath441 note that @xmath442 in the above equation is just the transition probability @xmath443 used in sec . [
sec : ea.tamsd ] .
first , we define the functions @xmath444 and @xmath445 as @xmath446 by using the integral representation of the gamma function , @xmath447 it is easy to obtain the following formula : @xmath448 % % & = & % % \sum_{n=1}^{\infty } % % \frac { 1}{\gamma(\beta ) } % % \int_{0}^{\infty } dv v^{\beta-1 } e^{-2nv } % % \\[0.2 cm ] % % & = & % % \frac { 1}{2^{\beta } } % % \sum_{n=1}^{\infty } % % \frac { 1}{\gamma(\beta ) } % % \int_{0}^{\infty } dv v^{\beta-1 } e^{-nv } \frac { \zeta(\beta)}{2^{\beta}},\end{aligned}\ ] ] where @xmath449 is the riemann zeta function . from eq .
( [ e.appe.even.zeta.int ] ) , we have @xmath450 in particular , we have @xmath451 and @xmath452 .
in this appendix , we summarize the ergodic properties of general observables . as an example
, we derive the spatial distribution of free ctrws .
first , we consider a renewal process with renewal times @xmath453 ( see sec . [
sec : tsd ] ) .
we also define observables that take nonzero values only at the renewal times @xmath454 : @xmath455 where @xmath456 are random variables with the same mean value @xmath457 ( @xmath458 ) .
@xmath456 are assumed to be independent of @xmath174 , but not necessarily mutually independent .
we also assume the ergodicity with respect to the operational time , i.e. , the number of renewals @xmath48 : @xmath459 this relation is just the law of large numbers . a sufficient condition for eq .
( [ e.app.hopf.2 ] ) is that the correlation function @xmath460 decays faster than @xmath461 @xcite . from eq .
( [ e.app.hopf.1 ] ) , the time average of the function @xmath271 is given by @xmath462 note that the value of the rhs of eq .
( [ e.app.hopf.3 ] ) depends on the trajectories of the renewal process in general . for example , if @xmath463 and the pdf of the renewal time @xmath37 is give by a power law @xmath464 with @xmath465 , the rhs of the above equation is equivalent to @xmath466 , which follows the ml distribution [ eq .
( [ e.tml ] ) with @xmath34 ] as @xmath144 .
first , we consider the case in which @xmath468 . in this case , we obtain the following equation from eq .
( [ e.app.hopf.2 ] ) : @xmath469 thus , eq .
( [ e.app.hopf.3 ] ) can be rewritten as @xmath470 similarly , if another observable @xmath471 defined by @xmath472 satisfies the same conditions as @xmath271 , we have @xmath473 note that the rhs is not a random variable . equation ( [ e.app.hopf.f.2 ] ) is a stochastic version of hopf s ergodic theorem for dynamical systems @xcite .
next , we consider the case @xmath474 .
we assume that the correlation function @xmath475 decays faster than @xmath476 ; then , from the central limit theorem , we have @xmath477 where @xmath478 is a random variable following the gaussian distribution with mean 0 and variance @xmath479 @xcite . the variance @xmath479 is given by @xmath480 @xcite . from eqs .
( [ e.app.hopf.3 ] ) and ( [ e.app.hopf.6 ] ) , we obtain @xmath481 if we define a random variable @xmath482 as @xmath483 , its characteristic function @xmath484 is given by @xmath485 where @xmath486 is the laplace transform of the pdf @xmath487 [ eq . ( [ e.tml ] ) ] .
although , in general , it is difficult to obtain the pdf of @xmath482 , we can derive it explicitly for @xmath34 . in this case , since @xmath488 , which is the laplace transform of the ml distribution , we obtain @xmath489 to derive this equation , let us introduce an auxiliary variable @xmath490 in eq .
( [ e.cumulative.1 ] ) as @xmath491 .
then double laplace transformations with respect to @xmath490 and @xmath182 give @xmath492 , where @xmath70 is the laplace variable conjugate to @xmath490 .
finally , taking the inverse laplace transform with respect to @xmath490 and setting @xmath493 , we obtain eq .
( [ e.app.hopf.9 ] ) . here
note that @xmath494 is the integral of the pdf @xmath495 . by using eqs .
( [ e.app.hopf.8 ] ) and ( [ e.app.hopf.9 ] ) , we obtain the characteristic function for @xmath40 explicitly : @xmath496 note that the pdf of @xmath482 is symmetric with respect to the @xmath497-axis from its definition and that the even order moments of @xmath482 are equivalent to those of the ml distribution with index @xmath498 and scale factor @xmath499 [ eq . ( [ e.app.hopf.9 ] ) ] .
this means that the pdf of @xmath482 , @xmath500 , is a symmetric extension of the ml distribution ; therefore we obtain @xmath501 for @xmath502 .
the pdf @xmath500 is called the bilateral mittag leffler distribution @xcite .
the simplest example of this pdf is the spatial distribution of one - dimensional free ctrws .
let the random variable @xmath503 be the displacement of the @xmath504-th jump ( e.g. , @xmath505 for ctrws with jumps only to the nearest neighbor sites ) .
if the jumps are symmetric , then @xmath506 .
thus , the time average [ eq . ( [ e.app.hopf.7 ] ) ] becomes a rescaled spatial position and the pdf [ eq .
( [ e.app.hopf.11 ] ) ] is the spatial distribution of the free ctrws .
note that this pdf is equivalent to the one derived from the ffpe [ eq . ( 46 ) in @xcite ] .
91ifxundefined [ 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/physrevlett.106.048103 [ * * , ( ) ] link:\doibase 10.1039/c0cp01879a [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.228102 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.92.178101 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.104.238102 [ * * , ( ) ] link:\doibase 10.1103/physreve.82.011913 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.018102 [ * * , ( ) ] link:\doibase 10.1103/physreve.83.041919 [ * * , ( ) ] \doibase doi : 10.1529/biophysj.104.051078 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.180602 [ * * , ( ) ] link:\doibase 10.1103/physreve.82.011129 [ * * , ( ) ] link:\doibase 10.1103/physrevb.12.2455 [ * * , ( ) ] link:\doibase 10.1016/0370 - 1573(90)90099-n [ * * , ( ) ] link:\doibase 10.1529/biophysj.106.092619 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physreve.83.031926 [ * * , ( ) ] \doibase doi : 10.1016/0378 - 4371(86)90060 - 9 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1016/s0378 - 4371(02)01361 - 4 [ * * , ( ) ] link:\doibase 10.1103/physreve.73.016111 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.74.1250 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.85.1774 [ * * , ( ) ] link:\doibase 10.1103/physreve.79.011112 [ * * , ( ) ] link:\doibase 10.1103/physreve.81.021103 [ * * , ( ) ] link:\doibase 10.1103/physreve.80.046125 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physreve.83.062101 [ * * , ( ) ] \doibase 10.1051/jp1:1992238 [ * * , ( ) ] link:\doibase 10.1103/physreve.82.016209 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , , ) link:\doibase 10.1103/physreve.82.030102 [ * * , ( ) ] link:\doibase 10.1063/1.857415 [ * * , ( ) ] link:\doibase 10.1038/nphys1760 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physreve.82.011117 [ * * , ( ) ] link:\doibase 10.1103/physreve.84.021137 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.73.2946 [ * * , ( ) ] link:\doibase 10.1103/physreve.52.1197 [ * * , ( ) ] \doibase doi : 10.1016/s0375 - 9601(00)00059 - 1 [ * * , ( ) ] link:\doibase 10.1103/physreve.76.041105 [ * * , ( ) ] link:\doibase 10.1103/physreve.79.031120 [ * * , ( ) ] link:\doibase 10.1016/j.physleta.2009.05.027 [ * * , ( ) ] link:\doibase 10.1103/physreve.78.051106 [ * * , ( ) ] link:\doibase 10.1103/physreve.80.011109 [ * * , ( ) ] link:\doibase 10.1073/pnas.0508366103 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.97.048302 [ * * , ( ) ] \doibase doi : 10.1016/s0006 - 3495(93)81253 - 0 [ * * , ( ) ] @noop * * , ( ) @noop _ _ , ed .
( , ) chap .
@noop _ _ ( , ) @noop _ _ ( , , ) @noop _ _ , ed . ,
( , ) chap
. @noop _ _ ( , , ) @noop * * , ( ) link:\doibase 10.1063/1.1559676 [ * * , ( ) ] link:\doibase 10.1063/1.1763136 [ * * , ( ) ] @noop * * , ( ) @noop ( ) link:\doibase 10.1103/physrevlett.107.178103 [ * * , ( ) ] link:\doibase 10.1063/1.4752768 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.99.210601 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1529/biophysj.108.139691 [ * * , ( ) ] link:\doibase 10.1103/physreve.83.021902 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.53.175 [ * * , ( ) ] link:\doibase 10.1103/physreve.69.061903 [ * * , ( ) ] link:\doibase 10.1529/biophysj.104.050765 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1080/00018730110116353 [ * * , ( ) ] link:\doibase 10.1103/physreve.84.051138 [ * * , ( ) ] @noop * * , ( ) @noop _ _ , ed . ,
( , ) http://www.jstor.org/stable/2285309 [ * * , ( ) ]
link:\doibase 10.1016/0167 - 7152(95)00113 - 1 [ * * , ( ) ] @noop _ _ ( , ) link:\doibase 10.1063/1.1704269 [ * * , ( ) ] link:\doibase 10.1016/s0370 - 1573(00)00070 - 3 [ * * , ( ) ] http://stacks.iop.org/0295-5075/77/i=1/a=10002 [ * * , ( ) ] http://stacks.iop.org/1742-5468/2007/i=08/a=p08001 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop ( ) link:\doibase 10.1073/pnas.1003693107 [ * * , ( ) ] @noop _ _ ( , , ) | the effects of spatial confinements and smooth cutoffs of the waiting time distribution in continuous - time random walks ( ctrws ) are studied analytically .
we also investigate dependences of ergodic properties on initial ensembles ( i.e. , distributions of the first waiting time ) . here
, we consider two ensembles : the equilibrium and a typical non - equilibrium ensembles . for both ensembles
, it is shown that the time - averaged mean square displacement ( tamsd ) exhibits a crossover from normal to anomalous diffusion due to the spacial confinement and this crossover does not vanish even in the long measurement time limit . moreover , for the non - equilibrium ensemble , we show that the probability density function of the diffusion constant of tamsd follows the transient mittag - leffler distribution , and that scatter in the tamsd shows a clear transition from weak ergodicity breaking ( an irreproducible regime ) to ordinary ergodic behavior ( a reproducible regime ) as the measurement time increases .
this convergence to ordinary ergodicity requires a long measurement time compared to common distributions such as the exponential distribution ; in other words , the weak ergodicity breaking persists for a long time . in addition , it is shown that , besides the tamsd , a class of observables also exhibits this slow convergence to ergodicity .
we also point out that , even though the system with the equilibrium initial ensemble shows no aging , its behavior is quite similar to that for the non - equilibrium ensemble . |
hypernuclei provide a unique laboratory for the investigation of hadronic many - body systems with strangeness -1 and for the study of new aspects of the strong and weak interactions in nuclei . because direct measurements of hyperon - nucleon ( @xmath5 ) scattering at low energies are prohibited by the short hyperon lifetime ,
hypernuclear spectra are the only way to study this interaction .
thus , a unique opportunity to obtain this information is provided by observing the fine structure of hypernuclei caused by the specific spin - dependence of the hyperon - nucleon interaction .
such characteristics are realized in practice only in @xmath2 hypernuclei and are hardly seen in other hypernuclei .
thus the spectroscopy of @xmath2 hypernuclei has a unique value in strangeness nuclear physics . in the past ,
hypernuclear spectroscopy has been carried out with limited resolution only by means of hadronic reactions , such as the strangeness exchange and associated production reactions , that use meson beams and in which a neutron is converted into a @xmath2 @xcite .
more recently , @xmath6-ray spectroscopy has been used to measure hypernuclear transition energies . here
, a few - kev energy resolution has been obtained , which has allowed precise level assignments and the measurement of doublet spacings @xcite .
however , the method is limited to the bound region below particle emission thresholds and to bound levels reached following particle emission
. the experimental knowledge can be enhanced using the @xmath7 electroproduction reaction .
this reaction is characterized by a large momentum transfer to the hypernucleus ( @xmath8 250 mev / c ) and strong spin - flip contributions , even at zero @xmath9 production angles @xcite , resulting in the excitation of both natural- and unnatural - parity states @xcite . in the @xmath7 reaction a proton
is converted into a @xmath2 hyperon so that one can produce and study new hypernuclei , not accessible in the standard reactions .
together with experiments in hall c @xcite , the e94 - 107 experiment in hall a at jefferson lab @xcite started a systematic study of high resolution hypernuclear spectroscopy in the @xmath10-shell region of nuclei , such as the hypernuclei produced in electroproduction on @xmath0be , @xmath11c , and @xmath12o targets .
results on and have been published @xcite .
the results for , which was long ago suggested as a good candidate for electroproduction studies @xcite because of the predicted large splitting of the ground - state and second - excited - state doublets are presented in this paper .
as in the previous experiments @xcite , the data are interpreted using shell - model calculations that include both @xmath2 and @xmath13 hyperons in @xmath14 states coupled to @xmath15-shell core wave functions optimized to fit a wide range of @xmath15-shell properties @xcite .
the @xmath7 reaction is described with distorted - wave impulse approximation ( dwia ) calculations @xcite that use the saclay - lyon ( sla ) model @xcite for the elementary @xmath16 reaction .
the sla model was successfully applied in the analysis of electroproduction experiments on @xmath11c and @xmath12o targets @xcite , which suggests that this model provides a reasonable prediction for the elementary cross section at very small @xmath9 production angles and at the center of mass energy of this experiment .
li core states are shown on the left along with the spectroscopic factors for proton removal from @xmath0be .
all excitation energies are in kev . on the right ,
the factors giving the relative population of levels in purely non - spin - flip ( @xmath17 ) and purely spin - flip ( @xmath18 ) production reactions on @xmath0be are given.,width=328 ] in a shell - model approach , one can define five @xmath19 two - body matrix elements for a hypernucleus with an @xmath20-level @xmath2 coupled to a @xmath15-shell nuclear core . these can be put into a one - to - one correspondence with the parameters ( radial integrals ) @xmath21 , @xmath22 , @xmath23 , @xmath24 , and @xmath25 associated with the average central , spin - spin , @xmath2-spin - orbit , nucleon - spin - orbit , and tensor components of the in - medium ( effective ) @xmath26 interaction @xcite , given by @xmath27 where @xmath28 denotes the spin and @xmath29 the angular momentum .
the constant @xmath21 simply contributes @xmath30 to the binding energy of every @xmath31 configuration and therefore does not affect the spectrum , only the overall binding energy @xmath32 .
the value of @xmath33 mev used is very close to the value that reproduces the experimental @xmath32 value ( table 2 of ref .
@xcite ) . in the weak - coupling limit ( quite good because the @xmath26 interaction for @xmath34 is a spatial monopole ) , only operators that depend on the spin of the @xmath2 particle ( @xmath22 , @xmath23 , @xmath25 ) contribute to doublet spacings while @xmath24 contributes to the spacing between doublets .
we use values ( in mev ) @xmath35 that fit the spectrum of the five bound levels of determined from and experiments .
the main parameters used for the corresponding @xmath26-@xmath36 interaction are @xmath37 and @xmath38 mev , making a total of six @xmath5 parameters that affect the spectrum . the calculated spectrum for is shown in fig . [ fig : lli9 ] ,
together with the core states for @xmath39li ( the first four are known and the other two are taken from the @xmath15-shell calculation ) , while table [ tab : contributions ] shows the contributions to the various level spacings for the three lowest doublets .
the contributions listed in table [ tab : contributions ] do not add up to exactly the spacings in fig .
[ fig : lli9 ] because small contributions from configuration mixing are not included ( see the caption to table [ tab : contributions ] ) . the spectroscopic factors ( @xmath40 with @xmath41 ) for proton removal from @xmath0be control the population of states via electroproduction .
the structure factors on the right of the figure for pure non - spin - flip and spin - flip transitions @xcite are normalized such that in the weak - coupling limit ( @xmath5 interactions turned off ) the @xmath17 and @xmath18 values for a doublet each sum to @xmath40 for the core state .
lrrrrr @xmath42 & @xmath43 & @xmath22 & @xmath23 & @xmath24 & t + + @xmath44 & 116 & 531 & @xmath45 & @xmath45 & @xmath46 + + @xmath47 & 79 & @xmath48 & 13 & 11 & 91 + + @xmath49 & 90 & 494 & @xmath50 & @xmath51 & @xmath52 + + @xmath53 & 63 & 441 & @xmath54 & 56 & @xmath55 + + @xmath56 & -6 & 8 & @xmath57 & @xmath58 & @xmath59 + lrrrrrr @xmath60 & @xcite & @xcite & @xcite & fit4 & ( 6 - 16 ) & ( 8 - 16 ) + + @xmath61 & 1.00 & 1.03 & 0.78 & 1.00 & 1.00 & 0.95 + + @xmath62 & 0.42 & 0.39 & 0.47 & 0.45 & 0.40 & 0.20 + + @xmath63 & 0.33 & 0.30 & 0.51 & 0.36 & 0.35 & 0.33 + + @xmath64 & & & & 0.04 & 0.06 & 0.24 + the @xmath40 values for @xmath65li in fig.[fig : lli9 ] are in good agreement with the values from @xmath66 studies @xcite ( see table [ tab : c2s ] ) . from a @xmath67 study @xcite ,
larger relative values were extracted for the excited @xmath68 and @xmath69 states of @xmath39li .
the @xmath40 values for the interaction used in the present work are listed in fig . [
fig : lli9 ] .
the values for all the @xmath15-shell interactions derived in connection with these hypernuclear studies are similar and in agreement with those for the cohen and kurath ( 6 - 16)2bme interaction @xcite .
the values for the other two cohen and kurath interactions put more strength in the second @xmath68 state than in the first ( as noted in ref .
@xcite for the ( 8 - 16)pot interaction ) .
the reason for this is that the @xmath62 states are rather purely @xmath70 , @xmath71 , rather than with strongly mixed @xmath72 and @xmath71 components , as happens for the other interactions .
strength for the @xmath64 state would be immediately noticeable in electroproduction because the states based on the @xmath64 state lie close to the neutron threshold at 3.73 mev and should therefore be narrow . the states in the first - excited doublet
are predicted to be nearly degenerate , in part because of the contributions from @xmath2-@xmath13 coupling ( see table [ tab : contributions ] ) .
in addition , the @xmath73 state contains a 3.5% admixture of a @xmath2 coupled to the @xmath39li ground state which lowers the @xmath73 state by another @xmath74 kev .
thus , if the shell - model predictions are reasonable , five peaks should , in principle , be resolved in below the particle - decay threshold by an electroproduction experiment with good energy resolution .
hall a at jlab is well suited to perform @xmath7 experiments . scattered electrons
are detected in the high resolution spectrometer ( hrs ) electron arm while coincident kaons are detected in the hrs hadron arm @xcite .
the disadvantage of smaller electromagnetic cross sections is compensated for by the high current and high duty cycle properties of the beam . throughout the experiment ,
the same equipment has been used in very similar kinematical conditions on c , be , and h@xmath75o targets .
the use of a pair of septum magnets permitted particle detection at very forward angles @xcite and a ring imaging cherenkov ( rich ) detector @xcite has been used in the hadron arm to provide an unambiguous identification of kaons when combined with the standard particle identification apparatus of hall a , based on aerogel cherenkov detectors @xcite . in the present experiment a 92.5 mg/@xmath76 solid @xmath77be target with a beam current of @xmath78100 @xmath79a
was used at a beam energy of 3775 mev . both hrss
were physically positioned at an angle of 12.5@xmath80 , but the pair of septum magnets yielded an effective angle for both the scattered electron and the hadron detection of @xmath81 .
[ be1 ] shows the observed binding - energy spectrum of .
the broad peak centered at a small positive binding energy corresponds to the states in fig .
[ fig : lli9 ] corresponding to the lowest three states of @xmath39li .
the rise in cross section starting at 0 mev corresponds to states with the @xmath2 in a @xmath15 orbit and , because these states are unbound , the states are broad and no structure is observed . as in ref .
@xcite , the background was determined from the binding energy spectrum obtained with a coincidence time shifted with respect to the coincidence time between secondary electrons and produced kaons and was rather flat for values of binding energy ranging from 15 mev to 0 mev .
its value was calculated as the average of the counts in the range 9.95 mev @xmath82 binding energy @xmath82 18.35 mev . for the calculation of the absolute cross section
, we computed the following quantities : detector efficiencies , detector dead time , detector phase space , kaon survival in hrs , integrated luminosity .
the calculation of efficiencies for the standard hrs package are well established and implemented in the hall a analysis software .
therefore , those procedures were used for that purpose . for the rich and aerogel cherenkov detectors
, we used one detector to determine the efficiency of the other one in the following way : we selected a pure sample of kaons by means of aerogel detectors and we measured the fraction of those kaons detected by the rich and vice versa .
the detector dead time was measured by the hall a data acquisition system .
the detector phase space was calculated using the simc code @xcite .
kaon survival is calculated considering the average path length inside the hrs arm .
the integrated luminosity was calculated by means of beam current monitor devices .
then , the absolute cross section @xmath83 was computed according to @xmath84 where @xmath85 is the event number in the experiment , @xmath86 is the global detector efficiency , @xmath87 is 1-detector dead time , @xmath88 is the detector phase space , @xmath89 is the kaon survival in the hrs , and @xmath90 is the integrated luminosity .
[ simulbe ] shows the background - subtracted experimental binding - energy spectrum , together with monte carlo simulations@xcite ( red curve ) and the same simulations with the radiative effects turned off ( blue curve ) .
the error bars in the data are statistical .
the simulations used the five peak positions and widths listed as configuration @xmath91 in table [ peak_configurations ] .
the red curve fits the experimental data well with a corresponding @xmath92/ndf value of 36.69/35 .
several other peak configurations , with different numbers , heights , positions and widths of the peaks , have been found to reproduce the red curve .
all of those are also expected to generate the same spectrum ( the blue curve of fig .
[ simulbe ] ) when radiative corrections are turned off , since radiative corrections are independent of the assumptions regarding the number and type of the peaks that build up the experimental spectrum . in practice , the simulated data do not overlap perfectly with the experimental data , which produces small systematic errors on the radiatively corrected spectrum .
the unfolding for radiative corrections has been done bin - by - bin .
the content of each bin of the radiatively corrected spectrum was obtained by multiplying the corresponding bin of the experimental spectrum by the ratio of the blue and red curves of fig .
[ simulbe ] for that bin . in order to avoid possible removals of background enhancements or to artificially null the spectrum in the regions where the blue curve is zero , the ratio between the blue and red curves of fig .
[ simulbe ] was performed after summing the background to each of them .
the background value was then subtracted from the result of the product of the ratio with the corresponding bin .
the result is shown in fig .
[ databe ] which presents the radiatively unfolded experimental data ( points with statistical errors ) compared to a theoretical prediction ( thin green line ) .
the band at the bottom of the histogram represents the systematic errors in the radiative unfolding .
the theoretical histogram was obtained using the procedure described in section [ sec : theory ] assuming an energy resolution of 730 kev ( fwhm ) .
once radiative corrections have been applied , the binding - energy spectrum resolution is small enough to clearly show a three - peak structure in the spectrum . a more detailed description of the procedure employed to determine the radiatively unfolded spectrum of fig . [ databe ]
is given in appendix [ appendix a ] .
[ cols="^,^,^,^,^,^,^,^ " , ] @xmath93 means that the spectrum with the radiative effects subtracted ( @xmath94 ) does not depend on the peak configurations @xmath91 , @xmath95 , @xmath96 as long as all the configurations considered @xmath97 , @xmath98 , @xmath96 fit the experimental spectrum with no radiative effects applied .
it has to be noted that only two assumptions were made in deriving the conclusion quoted above * the single intervals @xmath99 $ ] are so small that @xmath100 defined by eq .
( [ eq:10 ] ) is constant in it . *
the matrix @xmath101 is invertible .
this is usually the case considering that usually @xmath102 and @xmath103 if @xmath104 and @xmath105 to determine the spectrum with the radiative effects `` turned off '' ( blue curve of fig .
[ simulbe ] ) the monte carlo simc was employed @xcite .
the red curve of fig .
[ single_peaks ] shows a single peak of the excitation energy spectrum as predicted by simc for the reaction @xmath0be@xmath7 when using the e94 - 107 experimental apparatus ( position and amplitude of the peak are arbitrary ) .
the blue curve of fig . [ single_peaks ] shows the same peak when the radiative effects in the monte carlo simc are `` turned off '' .
several peak configurations , made up by a number of peaks like the one of fig . [ single_peaks ] red curve , fit the experimental excitation energy spectrum after being normalized to it .
table [ peak_configurations ] quotes five of them .
their corresponding excitation energy spectra ( normalized to the experimental data ) are shown in fig [ monte_carlo_spectra with_radiative_effects ] .
for each configuration , table [ peak_configurations ] quotes the position and relative amplitude of the peaks ( here and in the following the amplitude of a peak is defined as the integral of the peak over the whole energy spectrum ) and the value of the @xmath92 test , @xmath106 where @xmath107 = @xmath91 , @xmath95 , @xmath6 , @xmath108 , or @xmath109 , @xmath110 is the number of counts in the @xmath111 interval of the experimental excitation energy spectrum , and @xmath112 is the number of counts in the same interval as predicted by the normalized peak configuration @xmath107 .
the @xmath92 tests were performed in the interval -1.515 mev @xmath113 excitation energy @xmath113 3.735 mev , corresponding to 35 degrees of freedom . because of the properties of the subtraction of radiative effects from spectra quoted above , all the peak configurations @xmath91 , @xmath95 , @xmath6 , @xmath108 , and @xmath109 produce the same `` radiatively corrected '' spectrum .
the spectra of fig . [ monte_carlo_spectra without_radiative_effects ]
are obtained from fig .
[ monte_carlo_spectra with_radiative_effects ] by turning off the radiative effects , that is replacing the `` fig . [ single_peaks ]
red curve - like '' peaks with `` fig .
[ single_peaks ] blue curve - like '' peaks , without changing positions and amplitudes of the peaks .
all plots of fig .
[ monte_carlo_spectra without_radiative_effects ] are quite equal , as confirmed by the @xmath92 test , @xmath114 with @xmath115 and @xmath116 = @xmath91 , @xmath95 , @xmath6 , @xmath108 , or @xmath109 . in the worst case ( @xmath117 and @xmath118 ) eq . ( [ eq:17 ] ) yielded a value of 28.387 with 40 degrees of freedom . in fig .
[ histogram_comparisons](a ) , fig .
[ monte_carlo_spectra without_radiative_effects](a ) ( dashed line ) and fig . [ monte_carlo_spectra without_radiative_effects](c ) ( continuous line ) are shown together . because the configurations @xmath91 and @xmath6 produce the two most different `` radiatively corrected '' simc results , the difference between the two curves plotted in fig .
[ histogram_comparisons](a ) was chosen as the systematic error due to the ambiguity of the peak structure underlying the energy spectrum . as shown in fig .
[ histogram_comparisons](b ) this error is small compared to the statistical error .
the method to obtain radiative corrected spectra described in this appendix was used , in the analysis of the @xmath119li spectrum , because of the difficulties in establishing the peak structure underlying the experimental spectrum .
it is relatively new and it could be worthwhile hence to make some considerations about its reliability .
the method relies on eq .
( [ eq:14 ] ) that is mathematically correct .
the uncertainties on the radiative corrected spectrum @xmath120 derived by eq .
( [ eq:14 ] ) originate obviously from the uncertainties on the experimental spectrum @xmath121 and on the function @xmath122 ( `` detector function '' in the following ) that provides @xmath120 once @xmath121 has been measured .
if @xmath121 and @xmath122 were exempt from errors @xmath120 would be `` perfect '' . to understand the effects on the reconstructed radiative spectrum of the uncertainties on the measured spectrum and on the detector function
it could be worthwhile to look at the results of the method to derive a neutron energy spectrum from the proton recoil energy measurement ( see for example @xcite ) .
this method consists in determining a neutron energy spectrum measuring the energies of the protons generated by scatterings of the neutrons in a radiator and is formally similar to the one described in this appendix to derive radiative corrected spectra from the experimental ones .
formally , the connection between the neutron energy spectrum and the proton recoil energy spectrum can be expressed by a formula like eq .
( 1 ) quoted in ref .
@xcite that can be concisely expressed as : @xmath123 here @xmath124 is the number of neutrons in the neutron spectrum with an energy included in the interval @xmath125 , @xmath126 is the number of protons in the experimental spectrum with an energy @xmath127 , and @xmath128 is the `` detector function '' . defining @xmath129 and @xmath130 the arrays whose elements are @xmath124 and @xmath126 respectively ( @xmath131 and @xmath132 covering the whole neutron and proton spectra ) , eq . ( [ eq : recprot1 ] ) transforms into : @xmath133 eq .
( [ eq : recprot2 ] ) is formally equivalent to eq .
( [ eq:14 ] ) ( in eq .
( [ eq : recprot2 ] ) @xmath128 is a diagonal matrix ) .
however , in eq .
( [ eq : recprot2 ] ) the knowledge on @xmath130 is ( sometimes greatly ) affected by the uncertainties on the measured recoil proton energy .
these uncertainties mean that the determination of the number of protons @xmath134 whose real energy is included in the @xmath111 interval of the proton energy spectrum @xmath130 and hence the proton energy spectrum itself are affected by ( sometimes not negligible ) uncertainties too .
several factors affect the proton energy measurement : detector calibration , background subtraction , and , above all , proton energy losses in the detector elements ( including their entrance windows ) and in the air between them . to correct for proton energy losses ,
the proton energies are shifted by the estimated average energy loss over possible proton paths , or , sometimes , in low energy regions , by unfolding techniques .
the proton energy losses set usually the low limit of the reconstructed neutron energy spectrum .
the detector function @xmath128 in eq .
( [ eq : recprot2 ] ) is affected by uncertainties too .
it includes the detector efficiency that depends on the geometry ( and on the connected problem of the determination of the scattering angle of the detected proton ) and on the differential n - p scattering cross section which are both sources of systematic errors .
the differential n - p scattering cross section is often obtained by parameterizations . despite these problems , the method to derive a neutron energy spectrum from the proton recoil energy measurement provides usually satisfactorily results . in ref @xcite ,
the method was applied to determine the spectra of nearly monoenergetic neutrons form the reaction @xmath135li(p , n)@xmath135be measured for eight incident proton energies .
the situation was here complicated by the fact that , together with the neutrons generated in the reaction under study , which corresponded to a well - defined peak in the neutron energy spectrum , the experiment detected neutrons by other reactions , as the three - body breakup process @xmath135li(p , n@xmath136he)@xmath91 , that generated a long tail in the low energy region of the neutron energy spectrum . despite that , the reproduction of the peaks of the eight neutron energy spectra was excellent , while the neutron counts in the tails of these spectra was somehow bigger than the corresponding parts of the spectra obtained with a time of flight ( tof ) detector for three of the eight incident proton energies .
the authors decided to rely on the tof detector results for the tails of these three spectra because the tof detector was free from the problems concerning the effects of the proton energy loss by reactions in the detectors and becuase it extended to lower energies than the method based on the proton recoil energy measurement . in ref .
@xcite the situation was improved with respect to ref .
@xcite because the detector function @xmath128 was simply equal , for all the proton energies , to @xmath137 , with @xmath138 , the proton scattering angle , measured by two silicon strip detectors for the most energetic protons and , less precisely , through the coordinates of the conversion point of the neutron inside the converter and the coordinates of the silicon detector closer to the converter for the protons whose energies were not big enough to make them reach the other silicon strip detector .
the use of a segmented converter decreased the uncertainties on the proton energies due to energy losses inside the converter itself . as a consequence ,
the minimum neutron energy detectable was lower than the one of ref .
the double differential neutron yield for the reaction @xmath139c(d , n ) at 40 mev was obtained this way .
the good successes obtained in determining neutron spectra from the measurements of the proton recoil energies and the understanding of the effects that could make this method less effective ensure that we can rely on the method described in this appendix to obtain , in the experiments performed at jlab , radiative corrected spectra from experimental binding energy spectra of hypernuclei . in fact , in these experiments , the hypernuclei are generated in point - like targets and the experimental binding energy spectra are obtained by the measurements of the momenta and the scattering coordinates of the secondary electrons and of the produced kaons in the reaction @xmath140 .
these momenta and scattering coordinates are determined very precisely by magnetic spectrometers ( at the level of @xmath141 for the momenta measured by hrs arms ) .
the energy losses of the secondary electrons and of the produced kaons inside the target and along the path to the detectors are small .
moreover , thanks to the excellent particle identification apparatus employed , the experimental spectrum of the experiment e94 - 107 was pratically background free , the only small background coming from kaons from accidental coincidences . the uncertanties on @xmath121 of eq .
( [ eq:14 ] ) are hence much smaller than the uncertanties on @xmath130 of eq .
( [ eq : recprot2 ] ) . besides , the uncertainties on the detector function @xmath122 in eq .
( [ eq:14 ] ) are smaller than the uncertainties on the corresponding function @xmath128 in eq .
( [ eq : recprot2 ] ) because of the simpler geometries involved , the smaller uncertainties on the scattering angles of the detected particles and of the qed cross sections involved in the function @xmath122 better known than the neutron - proton cross sections involved in the function @xmath128 .
in the hall a experimental setup , scattered electrons and produced kaons of the reactions @xmath77be@xmath7 and @xmath11c@xmath7 were detected by the high resolution spectrometer ( hrs ) electron arm and by the hrs hadron arm , respectively , while the primary electrons were provided by the cebaf accelerator . the cebaf accelerator electron beam energy and the central momenta and angles of the hrs electron and hadron arms were set according to the kinematics of the reactions and are taken as constant for the course of the experiment ( their variations being of the order of @xmath142 for the cebaf electron beam energy and the central momenta of the hrs electron and hadron arms , and practically zero for the spectrometer central angles ) .
however , the actual values of the cebaf accelerator electron beam energy and of the central momenta and angles of the hrs electron and hadron arms , although constant , differ by unknown amounts from the nominal set values , and are referred to as `` kinematical uncertainties '' .
although small ( the experimental uncertainties on the cebaf accelerator electron beam energy and on the spectrometer central momenta being of the order of @xmath141 - @xmath143 and those on the spectrometer central angles of the order of @xmath144 ) , these kinematical uncertainties cause a global shift in the binding - energy spectrum that hence has to be calibrated . in fact
, the binding energy is expressed as @xmath145 with @xmath146 and @xmath147 where @xmath148 , @xmath149 , @xmath150 , @xmath151 , @xmath152 , and @xmath153 are the energies and the momenta of the primary electron , of the scattered electron and of the produced kaon respectively , @xmath154 is the @xmath2 mass , @xmath155 is the target mass , and @xmath156 is the mass of the residual nucleus , that is of the nucleus with a-1 nucleons and z-1 protons ( where a is the number of nucleons and z is the atomic number of the target , respectively ) . the change in the binding - energy spectrum @xmath157 caused by the kinematical uncertainties can be expressed by the formula @xmath158 where @xmath148 is the cebaf electron beam energy , @xmath159 and @xmath160 are the central momenta of hrs electron arm and hrs hadron arm , @xmath161 and @xmath162 are the central angles ( defined as the angles between the central axes and the cebaf beam line ) of hrs electron arm and hrs hadron arm , and @xmath163 , @xmath164 , @xmath165 , @xmath166 and @xmath167 are the kinematical uncertainties , that is the differences between the nominal and the actual values of the cebaf electron beam energy and of the central momenta and the central angles of hrs electron and hrs hadron arms .
@xmath149 , @xmath150 , @xmath151 , @xmath152 , and @xmath153 in eq .
( [ eq:18 ] ) - ( [ eq:20 ] ) are functions of @xmath148 , @xmath159 , @xmath160 , @xmath161 and @xmath162 , as well as ( with the exception of @xmath149 ) of the scattering variables .
for example , the components normal to the beam line in the horizontal plane of @xmath151 and @xmath153 respectively ( called here @xmath168 and @xmath169 , respectively ) are given by the expressions @xmath170 and @xmath171 , while @xmath172 and @xmath173 , where @xmath174 ( @xmath175 ) is the percentage difference between the momentum of the scattered electron ( produced kaon ) and the central momentum of the hrs electron arm ( hrs hadron arm ) , @xmath176 ( @xmath177 ) is the angle between the electron ( produced kaon ) direction in the horizontal plane , and @xmath161 ( @xmath162 ) , @xmath178 ( @xmath179 ) are the angles between the electron ( produced kaon ) direction in the vertical plane and the cebaf beam line , and @xmath180 is the kaon mass .
for the kinematics of the experiment e94 - 107 ( see the nominal values of @xmath148 , @xmath150 , and @xmath152 at the end of this appendix ) or if @xmath183 , the change in the binding - energy spectrum @xmath157 caused by the kinematical uncertainties is @xmath184 where @xmath185 and @xmath186 the term @xmath187 does not depend on the target mass and causes a global shift of the binding - energy spectrum without changing the peak shapes and relative positions , while the term @xmath25 is equal to zero only when @xmath188 .
when this condition is not fulfilled , the main effect of @xmath25 on the binding - energy spectrum is to broaden the peaks because , in this case , depending on scattering variables through @xmath189 , it produces non - zero , scattering variable dependent values of @xmath157 and , as a consequence , an unphysical dependence on the scattering variables of the binding energy that broadens the peaks corresponding to the bound states .
because of the presence of the coefficient @xmath190 , the effect of @xmath25 on the position of the peaks is negligibly small .
the actual values of @xmath148 , @xmath159 , @xmath160 , @xmath161 , and @xmath162 are hence those for which @xmath191 ( which places a peak at its known position in the binding energy spectrum and minimizes its fwhm ) . to determine the @xmath2 binding energy of the hypernucleus produced in the reaction @xmath77be@xmath7 , we took advantage of the fact that the experiment used the same set - up employed for the study of the hypernucleus .
the values of @xmath148 , @xmath159 , @xmath160 , @xmath161 , and @xmath162 were thus the same in the productions of both and and were determined by positioning the ground state at its known position of @xmath192 mev @xcite in the binding - energy spectrum and minimizing its fwhm .
when minimizing the ground - state fwhm , it has to be taken into account that this ground state is actually a doublet whose energy splitting , assumed to be equal to the ground - state energy splitting , is 161.6 @xmath4 0.2 kev @xcite .
minimizing the ground - state fwhm implies hence some sort of distortion because it artificially narrows the positions of the peaks making up the doublet .
however , the ground - state doublet energy splitting is small enough with respect to the energy resolution of the experiment to make the approximation of assuming the ground state as a single peak still valid .
no attempt to minimize the fwhm was performed on the other peaks of the binding - energy spectrum .
another possibile source of distortion comes from the term @xmath25 in eq .
( [ eq:24 ] ) , which , although small because of the presence of the factor @xmath193 in it ( see eq .
( [ eq:26 ] ) ) , can potentially change the positions of the excited states with respect to each other and with respect to the ground state . during all the process of minimization of the kinematical uncertainties , the positions of the peaks of the energy spectrum as resulted by a fitting procedure were checked to ensure that the relative peak positions did not change within the errors ( the error of a position peak being defined as the standard deviation resulting by the fitting procedure ) .
it has to be stressed that the term @xmath187 in the expression of @xmath157 given by eq .
( [ eq:24 ] ) dominates because of the presence of the factor @xmath193 in the term @xmath25 ( see eq .
( [ eq:26 ] ) ) .
the positioning of the ground state at its known value in the binding spectrum was thus mainly performed choosing a set of values @xmath163 , @xmath164 , and @xmath165 that produced a value of @xmath187 equal to the difference between the measured and the expected position of the ground state .
the minimization of the ground state peak fwhm produces only second order effects on the position of the peaks in the binding - energy spectrum and was peformed mainly to choose the right set of values @xmath163 , @xmath164 , and @xmath165 among the @xmath194 sets of values that produced the desired value of @xmath187 .
things are much different for the reaction @xmath16 , where the target mass is small and the minimization of the fwhm of the peak ( which is a single peak ) plays a role as important as the positioning of the peak in the binding - energy spectrum to its zero value .
the procedure described above resulted in the set of values @xmath195 mev , @xmath196 mev , @xmath197 mev , @xmath198 , and @xmath199 . replacing with these values the nominal ones @xmath200 mev , @xmath201 mev , @xmath202 mev , @xmath203 , and @xmath204 ) in the @xmath205 expression , a value of @xmath206 ( stat . )
mev was obtained for the ground state .
the statistical error of @xmath207 mev is the error in the position of the first peak in the four - gaussian fit of the binding - energy spectrum ( see fig .
[ databe ] and table [ tab : results ] ) .
when evaluating the systematic error of the binding energy , one has to consider that , as quoted above , the term @xmath187 in @xmath157 dominates the effects of the term @xmath25 .
when neglecting the energy loss in the target , this means that if the single values of @xmath163 , @xmath164 , and @xmath165 were wrong , the binding energy would not be affected significantly as long as the sum @xmath187 ( and hence the position of the ground state ) is correctly reproduced .
the major source of systematic error is hence due to the energy loss in the target .
the difference between the shifts of the ground - state position in the and binding - energy spectra due to the energy loss in the targets of @xmath11c and @xmath0be , respectively , was evaluated to be equal to @xmath208 kev through the use of the monte carlo code simc @xcite .
this value was added in quadrature to the error of @xmath209 kev quoted for the ground - state binding energy @xcite to give a @xmath210 kev systematic error on the binding energy .
we want to remember f. cusanno , who enthusiastically and greatly contributed to this paper before his untimely death .
we acknowledge the jefferson lab physics and accelerator division staff for the outstanding efforts that made this work possible . this material is based upon work supported by the department of energy , office of energy research , under contract de - ac05 - 06or23177 , by the italian istituto nazionale di fisica nucleare , by the grant agency of the czech republic under grant no .
p203/12/2126 , by the french cea and cnrs / in2p3 , and by the u.s . national science foundation . | background : : in the absence of accurate data on the free two - body hyperon - nucleon interaction , the spectra of hypernuclei can provide information on the details of the effective hyperon - nucleon interaction .
purpose : : to obtain a high - resolution spectrum for the @xmath0be@xmath1 reaction .
method : : electroproduction of the hypernucleus has been studied for the first time with sub - mev energy resolution in hall a at jefferson lab on a @xmath0be target . in order to increase the counting rate and to provide unambiguous kaon identification , two superconducting septum magnets and a ring imaging cherenkov detector ( rich )
were added to the hall a standard equipment .
results : : the cross section to low - lying states of is concentrated within 3 mev of the ground state and can be fitted with four peaks .
the positions of the doublets agree with theory while a disagreement could exist with respect to the relative strengths of the peaks in the doublets . a @xmath2 separation energy , @xmath3 , of 8.36 @xmath4 0.08 ( stat . ) @xmath4 0.08 ( syst . ) mev was measured , in agreement with an earlier experiment . , , |
a self - map @xmath0 of a metric space @xmath1 is _ non - expansive _ if @xmath2 .
a general problem consists in studying the asymptotic behaviour of the orbits of @xmath0 .
this is motivated in particular by the celebrated theorem of denjoy and wolff @xcite on the iteration of holomorphic self - maps of the unit disk ( these maps are non - expansive in poincare s metric ) .
other motivations arise from the cases of non - expansive self - maps of banach spaces and of self - maps of cones that are non - expansive in hilbert s , thompson s or riemannian metric , which have received a considerable attention , due in particular to applications in game theory @xcite , discrete event systems @xcite , quadratic optimal control or filtering @xcite , and non - linear perron - frobenius theory @xcite .
several denjoy - wolff type results , either in the setting of metric spaces , or concerning the special case of cones have appeared , see in particular @xcite .
the reader may consult the monograph @xcite for an overview of the field . in this paper , we consider the linear _ escape rate _
@xmath3 the latter always exists , by a standard subadditive argument .
our main result is the following maximin characterisation of the escape rate .
[ th - minimax ] let @xmath0 be a non - expansive self - map of a complete metrically star shaped hemi - metric space @xmath1 .
then , @xmath4 where the maximum , which is attained , is taken over the set of martin functions @xmath5 of @xmath1 .
if in addition the hemi - metric @xmath6 is bounded from below ( in particular , if @xmath6 is a metric ) , and if @xmath7 , then any function attaining the maximum is a horofunction .
let us explain the terminology .
_ hemi - metrics _ are analogous to metrics , but @xmath8 and @xmath9 may differ .
their definition is a variation of that of _ weak metrics _ in @xcite .
the notion of _ metrically star shaped _
space is the object of definition [ defistarshaped ] below .
it requires a distinguished family of geodesics with a common origin to satisfy a convexity inequality .
hence , it is a mild form of busemann s classical non - positive curvature condition @xcite .
_ martin functions _ and _ horofunctions _ are special lipschitz functions of constant @xmath10 which arise when compactifying the hemi - metric space @xmath1 .
the set of horofunctions ( the _ horoboundary _ ) provides an abstract boundary of @xmath11 .
see [ subsec - horo ] for details .
theorem [ th - minimax ] is inspired by the classical collatz - wielandt characterisation of the perron root @xmath12 of a @xmath13 non - negative matrix @xmath14 .
the latter shows that @xmath15 here , @xmath16 denotes the set of real non - negative numbers , and so @xmath17 is the standard positive cone .
nussbaum showed in @xcite that the latter formula holds more generally when the map @xmath18 is replaced by a non - linear continuous order - preserving @xmath0 of the standard positive cone .
when @xmath0 preserves the interior of this cone , it is non - expansive in the ( reverse ) _ funk hemi - metric _
@xcite defined on this interior by @xmath19 thus , @xmath20 the collatz - wielandt formula and its non - linear extension in @xcite will be shown to be special instances of the maximin formula .
theorem [ th - minimax ] is also motivated by the case of a banach space @xmath11 .
then , one may consider the limit @xmath21 for instance , in the setting of zero - sum games , @xmath0 is the dynamic programming or _ shapley operator _ @xcite ( which acts on a banach space of continuous functions equipped with the sup - norm ) , and the latter expression represents the limit of the mean payoff per time unit as a function of the initial position , when the horizon of the game tends to infinity , see [ appli - shapley ] for more background .
a theorem of pazy @xcite shows that the limit does exist if @xmath11 is a hilbert space ; further improvements , under suitable strict convexity assumptions were done by reich@xcite and then by kohlberg and neyman @xcite : the limit exists in the weak ( resp .
strong ) topology if @xmath11 is reflexive and strictly convex ( resp .
if the norm of the dual space @xmath22 is frchet differentiable ) . without strict convexity , the limit may not exist , even when @xmath11 is finite dimensional , however , a general result of kohlberg and neyman @xcite ( corollary [ linear ] below ) shows that there is always a linear form @xmath23 of norm @xmath10 such that @xmath24 holds for all @xmath25 , where @xmath26 is precisely the term at the left - hand side of .
the theorem of kohlberg and neyman may also be thought of as a special case of theorem [ th - minimax ] .
indeed , any horofunction @xmath5 attaining the maximum in satisfies @xmath27 for all @xmath25 .
we shall see that the kohlberg - neyman theorem follows readily from this result .
we shall also recover as corollaries a generalisation of a result of gaubert and gunawardena @xcite ( corollary [ linearcompact ] ) concerning shapley operators , as well as a refinement of a result of gunawardena and walsh @xcite valid in symmetric cones , corollary [ cor - symm ] below .
theorem [ th - minimax ] turns out to be related to a denjoy - wolff type theorem of karlsson @xcite , theorem [ thmkarlsson ] below .
karlsson showed that an inequality similar to holds .
he used a general subadditivity argument , which does not require any non - positive curvature condition .
the statement of @xcite assumes the metric space to be proper , i.e. , closed balls to be compact , but this assumption can be relaxed by defining horofunctions with respect to the topology of pointwise convergence as we do here .
however , there are essential discrepancies between and the result of @xcite : the term @xmath28 is replaced there by the linear escape rate @xmath29 , and the horofunction @xmath5 _ depends _ on the choice of the point @xmath30 , whereas it does not in .
we shall see that without a non - positive curvature condition , the formula may not hold ( we give an example in which @xmath31 and @xmath5 necessarily depends on @xmath30 ) .
hence , the maximin theorem [ th - minimax ] holds only under more restrictive circumstances than the denjoy - wolff type result of karlsson . by comparison ,
the interest of the more special theorem [ th - minimax ] lies in its strong duality nature : it allows one to certify that @xmath32 , or dually , that @xmath33 , by exhibiting an element @xmath34 such that @xmath35 , or dually a martin function @xmath5 such that @xmath36 .
it should be noted in this respect that when @xmath11 is a banach space with a polyhedral norm , the horoboundary admits a simple effective description , see @xcite .
we say that @xmath37 is a _ hemi - metric _ on a set @xmath11 if the two following conditions are satisfied for all @xmath38 : 1 .
@xmath39 2 .
@xmath40 if and only if @xmath41 .
we then say that @xmath42 is a hemi - metric space .
notice that a hemi - metric is generally not a metric , since it is neither symmetric nor non - negative .
this definition is closely related to the one of a _ weak metric _ in @xcite , in which @xmath43 is required to be non - negative .
the last condition of the definition corresponds to the _ weak separation _ condition in the latter reference .
we allow @xmath43 to take negative values in order to deal with order - preserving positively homogeneous self - maps of cones ( section [ sec - cones ] ) .
to any hemi - metric , one can canonically associate a metric by the following lemma . for any hemi - metric @xmath43 ,
the function @xmath44 is a metric on @xmath11 . verifying that @xmath6 is symmetric and satisfies the triangular inequality is easy .
the positivity of @xmath6 comes from the fact that for all @xmath30 and @xmath45 , @xmath46 . finally , if @xmath47 then both @xmath48 and @xmath49 are non - positive so by the same argument they are both null , thus @xmath41 .
in the sequel , @xmath11 is equipped with the topology induced by the metric @xmath6 .
we shall say that @xmath42 is _ complete _ when the associated metric space @xmath1 is complete .
a geodesic joining a point @xmath50 to a point @xmath34 is a map @xmath51\rightarrow x$ ] such that @xmath52 , @xmath53 , and such that for all @xmath54 , @xmath55 [ defistarshaped ] we say that @xmath42 is _ metrically star - shaped _ with centre @xmath56 if there exists a family of geodesics @xmath57 , such that @xmath58 joins the centre @xmath56 to the point @xmath45 , and such that the following inequality is satisfied for every @xmath59 and @xmath60 $ ] : @xmath61 if the hemi - metric @xmath43 is not a metric , we also require that for any @xmath45 , the quantity @xmath62 tends to 0 as @xmath63 goes to 1 .
the condition is a form of _ metric convexity _ @xcite . in particular
any busemann space @xcite is metrically star - shaped , but our definition is less demanding since we only require the inequality to be satisfied for one specific choice of geodesics .
any banach space @xmath64 is metrically star - shaped with respect to any centre @xmath56 : it suffices to take the straight lines as geodesics , i.e. , for any choice of centre @xmath56 , the choice of @xmath65 yields a metrically star shaped space .
notice that some banach spaces are not busemann spaces @xcite .
the next example relies on the following notion .
[ def - heminorm ] a map @xmath66 from a banach space @xmath64 to @xmath67 is a _ hemi - norm _ if it can be written as @xmath68 where @xmath69 is a bounded subset of the dual space @xmath70 , and @xmath71 a hemi - norm is always lipschitz .
it is easily verified that if @xmath66 is a hemi - norm , then @xmath72 is a hemi - metric .
note that condition holds if and only if the orthogonal set of @xmath69 is reduced to the zero vector . as in the case of banach spaces , @xmath42 is metrically star shaped : it suffices to choose the straight lines as geodesics .
we shall say that the hemi - norm is _ compatible _ with the banach space @xmath64 if @xmath73 .
then , the metric @xmath6 obtained from @xmath43 is the one associated to the norm of @xmath11 , and in particular , the hemi - metric space @xmath42 is complete .
the norm of the banach space @xmath11 is a special case of compatible hemi - norm , in which @xmath74 where @xmath75 is the dual unit ball of @xmath70 .
then , by bauer maximum principle ( * ? ? ?
* theorem 7.69 ) , the supremum is attained at some extreme point of @xmath75 , so the set @xmath69 arising in the definition of the hemi - norm ( see equation [ e - hemi - norm ] ) can be taken to be either @xmath76 or the set of extreme points of @xmath76 .
[ exampletop ] a useful example of hemi - norm is the following .
let @xmath77 denote a compact topological space , let @xmath78 denote the banach space of continuous functions from @xmath77 to @xmath79 , equipped with the sup - norm .
consider @xmath80 , where @xmath81 denotes the `` top '' operator which gives the maximum of a function , i.e. , @xmath82 the metric @xmath6 obtained from @xmath43 is the sup - norm .
we can take for @xmath69 the set of evaluations functions @xmath83 with @xmath84 .
another class of examples concern symmetric cones , which include cones of positive semidefinite matrices .
we shall discuss them further in section [ sec - cones ] .
in this section we consider a metrically star - shaped space @xmath42 with centre @xmath56 , as well as a map @xmath85 that is non - expansive with respect to the hemi - metric @xmath43 , meaning that for all @xmath86 , @xmath87 to each non - expansive mapping @xmath0 , we associate the two following quantities : @xmath88 @xmath89 thus , the number @xmath29 measures the _ linear escape rate _ of the orbits of @xmath0 .
it is well defined .
indeed , for any @xmath50 , since @xmath0 is non - expansive , the sequence @xmath90 is _ subadditive _ , meaning that @xmath91 for all @xmath92 , and so , a classical argument ( see ( * ? ? ?
* lemma 1.18 ) ) shows that the limit @xmath93 does exist and is equal to @xmath94 .
moreover , since @xmath95 the limit is independent of the choice of @xmath50 .
[ inegalitefacile ] the following inequality is satisfied for any non - expansive mapping @xmath0 : @xmath96 observe that for any @xmath50 and @xmath97 , @xmath98 and take the infimum on both @xmath99 and @xmath30 .
the horofunction boundary of a metric space was defined by gromov @xcite .
see also @xcite , ( * ? ? ?
ii ) and @xcite . in fact
, the same construction can be performed with a hemi - metric .
the details can be found in @xcite , in which a horofunction - like boundary is defined for discrete optimal control problems , by analogy with the martin compactification arising in the theory of markov processes . see also @xcite .
let us fix an arbitrary point @xmath100 ( the _ basepoint _ ) .
we define a map @xmath101 from @xmath11 to the set functions from @xmath11 to @xmath67 by associating to any @xmath50 the following function @xmath102 : @xmath103(y):=\delta({\bar{x}},x)-\delta(y , x).\ ] ] using the triangular inequality , we get that , for all @xmath104 , @xmath105(y){\leqslant}\delta({\bar{x}},y)\ ] ] hence , the set @xmath106 can be identified to a subset of the product space @xmath107\subset{\mathbb{r}}^x$ ] . by tychonoff
s theorem , the later space is compact for the product topology .
hence , the closure of @xmath108 in the topology of pointwise convergence ( which is the same as the product topology ) is compact .
we denote by @xmath109 this closure .
the elements of the _ boundary _ @xmath110 are called _ horofunctions_. we will call _ martin _ functions the elements of @xmath109 ( the _ martin space _ ) .
note that the choice of the basepoint is irrelevant ( changing the basepoint translates all the martin functions by the same constant ) .
the term of horofunction is sometimes used in a more restrictive sense , to denote what is also called a _
busemann function _ , which is a horofunction obtained as the limit of a family of functions @xmath111 taken along an `` infinite geodesic '' @xmath112 , see @xcite .
note also the sign in the definition of @xmath102 ( which implies , in the banach space case , that the map @xmath102 is concave ) .
this sign is chosen consistently with potential theory ( the opposite of the distance is the analogue of the martin kernel ) . in the metric geometry literature ,
the opposite choice is most often made .
[ rk - technical ] the horofunctions are often defined as the closure of @xmath108 in the topology of uniform convergence on bounded sets ; indeed , the injection @xmath113 is continuous for this topology , and this injection is an embedding if @xmath11 is a complete geodesic space ( * ? ?
it can be checked that every map @xmath102 is lipschitz of constant @xmath10 with respect to the metric @xmath6 , and so , by the ascoli - arzela theorem , the closure of @xmath108 in this sense is compact as soon as @xmath11 is proper ( meaning that every closed ball is compact ) . in the present work
, we do not require @xmath11 to be proper , but define rather @xmath114 as the closure of @xmath108 in the topology of pointwise convergence , so that @xmath114 is always compact . note however that the injection @xmath113 may _ not _ be an embedding from @xmath11 to @xmath114 ( the topology on @xmath114 is too weak in general , so the inverse of the map @xmath113 may not be continuous ) .
[ remark - topo - martin ] the topology of the martin space is metrisable as soon as @xmath11 is a countable union of compact sets , see for instance ( * ? ? ?
* remark 7.10 ) .
then , every horofunction is the limit of a sequence of functions @xmath115 , where @xmath116 is a sequence of elements of @xmath11 .
we shall derive theorem [ th - minimax ] from the following result .
[ maintheorem ] let @xmath42 be a complete metrically star - shaped hemi - metric space , and let @xmath85 be non - expansive .
then there exists a martin function @xmath117 such that for all @xmath50 , @xmath118 moreover , if @xmath43 is bounded from below ( in particular , if @xmath43 is a metric ) and if @xmath119 , then @xmath5 is necessarily a horofunction .
let @xmath56 and @xmath120 as in definition [ defistarshaped ] ; for any @xmath121 denote by @xmath122 the function @xmath123 and recall that @xmath124 is @xmath125-contracting by definition . by completeness , we can thus define for any @xmath125 the point @xmath126 as the only solution of the fixed point equation @xmath127 then for any @xmath50 , @xmath128 since the space @xmath109 is compact , the family of functions @xmath129 admits a limit point @xmath117 as @xmath125 tends to @xmath10 . passing to the limit in the previous inequality , and using the additional assumption in definition [ defistarshaped ]
, we deduce that @xmath130 .
assume now by contradiction that @xmath43 is bounded from below , that @xmath119 , and that @xmath131 for some @xmath132 .
then , we deduce from @xmath133 and @xmath134 that @xmath135 the right - hand side of this expression tends to @xmath136 as @xmath99 tends to infinity , but the left - hand side of this expression is bounded above independently of @xmath99 , since @xmath137 is bounded from below .
this is impossible , hence , @xmath138 is a horofunction .
( more generally , the same argument shows that for any net of functions @xmath139 converging to @xmath5 , the net @xmath140 can not have a bounded subnet . )
we obtain as immediate consequences the following corollaries : [ cor - pumping ] let @xmath141 and @xmath0 be as in theorem [ maintheorem ] .
then , there exists a martin function @xmath117 such that for all @xmath50 , @xmath142 moreover , if @xmath43 is bounded from below and if @xmath119 , then @xmath5 is necessarily a horofunction .
[ coregalite ] if the assumptions of theorem [ maintheorem ] are satisfied , then @xmath143 .
we deduce from that @xmath144 which yields @xmath145 .
then , the result follows by lemma [ inegalitefacile ] .
theorem [ th - minimax ] follows readily by combining theorem [ maintheorem ] and corollary [ coregalite ] .
we now apply theorem [ maintheorem ] to non - expansive self - maps of a banach space @xmath11 , and more generally , to those maps that are non - expansive in a _
compatible hemi - norm _ , as defined in section [ subsec - hemi ] .
[ seminormtheorem ] let @xmath64 be a banach space , and let @xmath146 be a compatible hemi - norm , where @xmath69 is a bounded subset of the dual space @xmath70 .
let @xmath0 be non - expansive for @xmath43 .
then , for every @xmath50 , there exists a linear form @xmath147 in the weak - star closure of @xmath69 such that for every @xmath148 , @xmath149 let @xmath150 denote the closure of @xmath69 in the weak - star topology . observe that , for all @xmath151 , @xmath152 since @xmath69 is bounded , @xmath150 is weak - star compact , and so , the map @xmath153 attains its maximum on @xmath150 .
fix now @xmath30 and @xmath45 in @xmath11 .
it follows from the previous discussion that for all @xmath132 , @xmath154 for some @xmath155 , independent of @xmath156 , and so @xmath157 in other words , @xmath147 is in the super - differential of @xmath158 at point @xmath30 .
let @xmath5 be the martin function that appears in corollary [ cor - pumping ] ; @xmath159 for some net @xmath160 .
by the previous observation , for every @xmath161 we can find an element @xmath162 such that for every @xmath148 , @xmath163 taking the limit along some subnet , we can find a linear form @xmath147 in @xmath150 such that for every @xmath148 , @xmath164 since by corollary [ cor - pumping ] @xmath165 , the theorem is established .
we now recover the following result .
[ linear ] let @xmath64 be a banach space , let @xmath85 be non - expansive and assume that @xmath119 .
then for any @xmath50 , there exists a continuous linear form @xmath147 of norm one , such that @xmath166 for all @xmath25 .
moreover , @xmath147 can be taken in the weak - star closure of the set of extreme points of the dual unit ball .
we apply theorem [ seminormtheorem ] , taking @xmath69 to be the set of extreme points of the dual unit ball .
this gives a @xmath147 in the weak - star closure of @xmath69 satisfying ( [ e - kn ] ) .
in particular @xmath167
. moreover , @xmath168 , and since @xmath169 , @xmath170 , which shows that @xmath171 .
actually , the assumption that @xmath119 can be dispensed with in the finite dimension case : [ linearfinite ] let @xmath64 be a finite dimensional banach space , and let @xmath85 be non - expansive . then for any @xmath50 , there exists a linear form @xmath147 of norm one , in the closure of the set of extreme points of the unit ball , such that @xmath172 for all @xmath25 .
the proof is similar to the one of corollary [ linear ] .
we still take @xmath69 as the set of extreme points of the dual unit ball , which belongs to the dual unit sphere , and we conclude since the unit sphere of a finite dimensional banach space is closed . the case in which @xmath173 , in finite dimension , has been studied in particular by lins in @xcite , who proved that either @xmath0 has a fixed point or there exists a linear form @xmath174 of norm @xmath10 such that @xmath175 for all @xmath50 .
this result is established by showing first that in the latter case , there is a horofunction @xmath5 such that @xmath176 as @xmath99 tends to @xmath177 ( modulo a change of sign convention for horofunctions ) .
as noted in remark 3.1 of @xcite , the construction of the linear form @xmath174 from the horofunction @xmath5 relies essentially on a sub - differentiability argument .
an argument of the same nature is used here to derive corollary [ linearfinite ] from theorem [ maintheorem ] .
note however that corollary [ linearfinite ] and the existence of the horofunction in @xcite are independent results ( none of them can be recovered from the other one ) . in this section
@xmath77 is a compact topological space , and @xmath78 is the banach space of continuous functions from @xmath77 to @xmath67 , endowed with the sup norm
. we will be interested in functions @xmath85 that are _ order - preserving _ , meaning that @xmath178 where @xmath179 denotes the canonical partial order on functions .
we shall say that @xmath0 _ commutes with the addition of a constant _ if @xmath180 where @xmath181 denotes the function @xmath182 . as in example [ exampletop ]
, we set @xmath183 and consider @xmath80 . recall that
@xmath43 is a hemi - metric compatible with the sup norm and that @xmath42 is metrically star - shaped .
it is easy to see that @xmath0 is order - preserving and commutes with the addition of a constant if and only if @xmath184 for all @xmath30 and @xmath45 , meaning that @xmath0 is non - expansive for @xmath43 ( see in particular @xcite , or @xcite ) . for this special hemi - metric @xmath43 ,
we denote by @xmath185 and @xmath186 the quantities defined in ( [ defrhobar ] ) and ( [ deftimecycle ] ) .
theorem [ maintheorem ] and corollary [ coregalite ] imply that @xmath187 for all @xmath30 in @xmath11 .
we next derive from theorem [ seminormtheorem ] the following generalisation of a result established in @xcite when @xmath77 is finite .
[ linearcompact ] let @xmath77 be a compact topological space , and let @xmath78 be the banach space of continuous functions from @xmath77 to @xmath67 , endowed with the sup norm .
suppose that @xmath85 is order - preserving and commutes with the addition of a constant .
then for every @xmath50 , there exists @xmath188 such that , @xmath189 for all @xmath25 .
we apply theorem [ seminormtheorem ] to the hemi - norm @xmath190 , taking for @xmath69 the set of evaluation functions @xmath191 , @xmath192 . since @xmath77 is compact , the set @xmath69 is weak - star closed , so that the linear form @xmath147 in theorem [ seminormtheorem ] can be written as @xmath193 for some @xmath194 .
order - preserving maps @xmath0 commuting with the addition of a constant arise as _ shapley operators _ in game theory .
suppose now that @xmath77 , @xmath195 and @xmath196 are three compact metric spaces , that @xmath197 is a continuous function from @xmath198 to @xmath67 , and that @xmath199 is continuous , where @xmath200 , the set of probabilities on @xmath77 , is endowed with the weak topology .
then the shapley operator of the two - player , zero - sum stochastic game with state space @xmath77 , action sets @xmath195 and @xmath196 , payoff @xmath197 , and transition probability @xmath201 , is defined as @xmath202 \mu(da)\nu(db ) \right\}\\ & = \inf_{\nu\in \delta(b ) } \sup_{\mu\in \delta(a ) } \left\ { \int_{a\times b } \left[g(a , b,\omega)+\int_\omega q(dw'|a , b,\omega ) x(\omega')\right ] \mu(da)\nu(db ) \right\ } \enspace.\end{aligned}\ ] ] the map @xmath0 is order - preserving and commutes with the addition of a constant .
it is shown in @xcite that it preserves @xmath203 .
then , @xmath204 represents the average reward per time unit if both players play optimally in the @xmath99-stage game and if the starting state is @xmath205 .
thus , corollary [ linearcompact ] shows that there is an initial state @xmath206 which is the best for the maximising player in the long term .
in particular @xmath207 converges as @xmath99 goes to infinity . to every result concerning
the hemi - norm @xmath81 corresponds a dual result concerning the hemi - norm @xmath208 .
in particular , @xmath209 moreover , there exists @xmath210 satisfying @xmath211 for all @xmath25 .
this is readily deduced by applying corollary [ linearcompact ] to the map @xmath212 .
theorem [ maintheorem ] should be compared with the following result .
( recall that a metric space is _ proper _ if every closed ball is compact . )
[ thmkarlsson ] if @xmath0 is a non - expansive self - map of a proper metric space @xmath42 , then , for all @xmath50 , there exists a horofunction @xmath5 ( depending on @xmath30 ) , such that @xmath213 holds for all @xmath214 .
note first two differences in the statements of theorem [ maintheorem ] and of karlsson s theorem : we require a metric convexity assumption whereas karlsson does not ; karlsson makes the general assumption that the metric space is proper whereas we do not .
actually , the compactness issue appears to be a mere technicality : it can be checked that the proof of theorem [ thmkarlsson ] remains valid even if @xmath11 is not proper if one defines horofunctions with respect to the topology of pointwise convergence as we do here ( see remark [ rk - technical ] ) .
a key difference however lies in the conclusions of both results . in theorem
[ maintheorem ] , horofunctions are _ independent _ of the choice of @xmath30 , whereas in karlsson s result , the horofunction does depend on @xmath30 .
note also that inequalities involving @xmath28 as in theorem [ maintheorem ] and corollary [ cor - pumping ] are stronger than their counterparts involving @xmath29 since @xmath215 .
this is illustrated in example [ examplenonconvexemetrique ] below , in which we consider a non - star - shaped space for which theorem [ thmkarlsson ] applies but not theorem [ maintheorem ] .
this example shows that the assumptions of theorem [ maintheorem ] can not be weakened : it is primordial that the space @xmath11 is metrically star - shaped .
hence , theorem [ maintheorem ] and karlsson s result are incomparable : the latter holds under more general circumstances , but the former yields a stronger conclusion .
[ examplenonconvexemetrique ] let @xmath216 be the torus with its canonical metric @xmath217 ; and let @xmath218 , with the metric @xmath219 let @xmath220 , and consider the mapping @xmath221 , @xmath222 which is non - expansive for the metric @xmath6 .
it is straightforward to check that @xmath223 but @xmath224 , so inequation can not be satisfied ( otherwise one would have @xmath225 by corollary [ coregalite ] ) .
in fact we prove that it is not even true that there exists a function @xmath117 such that for all @xmath30 , @xmath226 fix a basepoint @xmath227 in @xmath11 .
since the metric on @xmath218 is the sum of the two metrics @xmath228 and @xmath229 , we can write any martin function @xmath5 of @xmath11 as the sum @xmath230 where @xmath231 is a martin function of @xmath228 and @xmath232 is a martin function of @xmath229 , relatively to the basepoints @xmath233 and @xmath234 , respectively ( see prop . 9.1 in @xcite ) .
the only martin functions on @xmath235 are @xmath236 with @xmath237 .
the martin functions of @xmath229 are either of the form @xmath238 , with @xmath239 , or the affine functions @xmath240 and @xmath241 .
we search a martin function @xmath5 such that , for any @xmath242 , @xmath243 i.e. , @xmath244 any martin function @xmath245 is bounded .
if @xmath232 is of the form @xmath246 or @xmath247 , then , for a fixed @xmath248 , the left - hand side of ( 3.7 ) remains bounded from above as @xmath249 , whereas the right - hand side of ( 3.7 ) tends to infinity .
the only possibility is thus @xmath250 . in that case , becomes @xmath251 since the sequence @xmath252 is dense in @xmath235 , and since @xmath245 is continuous , we have that @xmath253 .
this is true for any @xmath254 , so @xmath245 is constant .
but there are no constant martin function of @xmath255 .
karlsson s result shows however that for any choice of @xmath242 , there exists a horofunction @xmath5 ( depending on @xmath248 ) such that is satisfied .
in this section , we recall or establish preliminary results concerning cones .
we consider a ( closed , convex and pointed ) cone @xmath256 in @xmath257 with non - empty interior @xmath258 .
we equip @xmath256 with the partial order defined by @xmath259 if @xmath260 . for all @xmath261 ,
we set @xmath262 since @xmath256 is closed , as soon as the two latter sets are non - empty , their respective infimum and supremum are attained ; in particular , @xmath263 and @xmath264 are finite and positive .
the ( reverse ) _ funk hemi - metric _ on the interior of @xmath256 is defined by @xmath265 more generally , we shall use the notation @xmath266 as soon as @xmath261 are such that @xmath267 for some @xmath268 .
the terminology `` funk metric '' is borrowed from @xcite ; it refers to @xcite .
the map @xmath269 is easily checked to be a hemi - metric on @xmath270 . indeed : @xmath48 is finite since @xmath30 is in the interior of @xmath256 ; the triangular inequality and the fact that @xmath271 for any @xmath50 are easy to verify since @xmath256 is pointed ; the fact that @xmath272 follows from the fact that @xmath256 is closed and pointed .
the term `` reverse '' is by opposition to the hemi - metric @xmath273 which has a different horoboundary @xcite .
the quantities @xmath263 and @xmath264 can be expressed in terms of the extreme rays of the _ dual cone _
the latter is the set of continuous linear forms taking non - negative values on @xmath256 . for each cone @xmath275
, @xmath276 denotes the set of _ representatives _ of the extreme rays of @xmath275 ( i.e. , the non - zero vectors belonging to these extreme rays ) .
[ lem - funk ] let @xmath277 be such that @xmath267 and @xmath278 for some @xmath279 . then , @xmath280 moreover , when @xmath281 , the latter suprema are attained , and the condition that @xmath282 can be replaced by @xmath283 . when @xmath284 , the latter infima are attained .
let now recall that if @xmath290 is in the interior of @xmath256 , and if @xmath291 , then , @xmath292 can not vanish .
indeed , we can find a ball @xmath196 centred at @xmath293 , such that @xmath294 for all @xmath295 .
if @xmath296 , then , @xmath297 holds for all @xmath295 , and so @xmath298 .
choose now a vector @xmath299 , and define @xmath300 .
observe that @xmath301 is compact and convex , and that each extreme ray of @xmath256 can be identified to the unique extreme point of @xmath301 which generates this ray .
let @xmath302 denote the value of the last supremum in , so that the inequality @xmath303 holds for all extreme points @xmath174 of @xmath301 .
since every element @xmath174 of @xmath301 is a barycentre of extreme points of @xmath301 , we deduce that the same inequality holds for all points @xmath174 of @xmath301 , which shows that @xmath302 coincides with the first supremum in . finally , when @xmath281 , we consider the map @xmath304 , @xmath305 .
since @xmath306 is linear , it attains its maximum at an extreme point of the compact convex set @xmath307 , which implies that both suprema in are attained .
similarly , when @xmath284 , we note that every @xmath308 maximising @xmath309 attains the first infimum in , and so , the second infimum is attained at an extreme point of @xmath310 .
let @xmath256 be a closed convex pointed cone of non - empty interior in a finite dimensional banach space , with basepoint @xmath233 .
then , any martin function @xmath5 for the ( reverse ) @xmath311 hemi - metric corresponds to a vector @xmath312 , @xmath313 and @xmath5 is a horofunction if and only if @xmath314 .
we shall also be interested in the following variant of the funk hemi - metric , considered by papadopoulos and troyanov @xcite : @xmath315 the map @xmath316 is easily seen to be a hemi - metric .
the following simple application of the arguments of @xcite characterises the boundary of the hemi - metric @xmath316 .
[ prop - walsh - funkp ] let @xmath256 be a closed convex pointed cone of non - empty interior in a finite dimensional banach space .
then , the martin space for the @xmath316 hemi - metric consists of the functions , with @xmath312 , together with the functions @xmath317 with @xmath318 , and an element is a horofunction if and only if it is of the form , with @xmath312 , or of the form with @xmath319 .
any function @xmath5 of the martin space is the pointwise limit of a sequence of functions @xmath320 , where @xmath321 is a sequence of elements of @xmath270 . if @xmath322 as @xmath323 tends to infinity , then , for all @xmath281 , @xmath324 also tends to infinity , and so @xmath325 for @xmath323 large enough
then , the result of @xcite shows that @xmath326 is of the form , for some @xmath312 .
it remains to examine the case in which @xmath327 contains a bounded subsequence .
then , possibly after replacing @xmath327 by a subsequence , we may assume that @xmath327 converges to a point @xmath318 .
using the fact established in @xcite that for @xmath281 , @xmath266 is continuous in the second argument @xmath328 , we deduce that @xmath329 converges to the function . finally , the characterisation of the horofunctions is straightforward .
the metric @xmath330 associated to the funk hemi - metric , @xmath331 is called the _ thompson
metric_. we recall that for this metric , the map @xmath332 with @xmath333 and @xmath334 is a ( straight line ) geodesic joining @xmath56 to @xmath45 , see @xcite .
however , as shown in @xcite , this choice of geodesics does not satisfy busemann s non - positive curvature condition ( even when @xmath335 ) . actually , it is shown in @xcite that only a weaker inequality is valid .
the latter inequality is optimal ( it is straightforward to find explicit counter examples to the convexity inequality ) .
however , when @xmath256 is the cone @xmath336 of @xmath13 positive semi - definite matrices , there is a different choice of geodesics between any pair of matrices @xmath337 in the interior of @xmath256 , involving geometric means : @xmath338 see in particular @xcite .
the latter are indeed known to be geodesics in thompson s and hilbert s metric .
note that when @xmath339 is the identity matrix , this geodesic seen with logarithmic glasses is a straight line : @xmath340 .
the general form of the geodesic is obtained by considering the linear transformation @xmath341 , which is an automorphism of the cone of positive definite matrices , isometric both in thompson s and hilbert s metric , and sending @xmath339 to the identity .
this construction actually makes sense more generally when @xmath256 is a _ symmetric cone _
@xcite , meaning that @xmath256 is self - dual ( @xmath342 ) , and that the group of automorphisms leaving invariant the interior of @xmath256 acts transitively on this interior .
symmetric cones include in particular the lorentz cones , the cones of real or complex positive semidefinite matrices , and cartesian products of such cones .
it is shown in @xcite that the metric convexity property of definition [ defistarshaped ] holds for any symmetric cone , for the latter choice of geodesics , for any choice of the centre , both for the thompson and hilbert metric .
a similar observation was made by lawson and lim @xcite .
actually , the arguments in @xcite apply as well to the funk and @xmath316 hemi - metrics .
hence , the present results apply to those self - maps of the interior of a symmetric cone that are non - expansive in any of the previously mentioned hemi - metrics .
[ rk - funk - symmetric ] when @xmath343 is the cone of positive semi - definite matrices , when @xmath281 , and @xmath344 , @xmath263 coincides with the maximal eigenvalue of the symmetric matrix @xmath345 .
moreover , denoting by @xmath346 the ( rank one ) orthogonal projector on any eigenline corresponding to this eigenvalue , one can check that the linear form @xmath347 attains the supremum in .
when @xmath348 , with @xmath349 for some @xmath268 , a maximising linear form @xmath174 can be constructed in the same way , replacing @xmath350 by the square root of the moore - penrose inverse of @xmath30 .
a dual argument allows one to construct a linear form attaining the infimum in .
note also that these constructions can be easily adapted to the case of a symmetric cone @xmath256 .
[ rk - psd ] in the case of the cone of symmetric positive definite matrices , the thompson , hilbert , funk , and @xmath316 hemi - metrics all have the following form @xmath351 where @xmath352 is a symmetric function @xmath353 ( by _ symmetric _ , we mean that @xmath352 is invariant by permutation of its variables ) , @xmath354 denotes the sequence of eigenvalues of a matrix @xmath14 , counted with multiplicities , so that @xmath355 denote the eigenvalues of the matrix @xmath356 ( which are real and positive since @xmath195 and @xmath196 are positive definite ) .
the previously listed hemi - metrics are obtained by taking for @xmath352 the following maps , @xmath353 , @xmath357 respectively .
we may consider more generally any symmetric hemi - norm @xmath352 .
then , the corresponding map @xmath358 can be checked to be a hemi - metric , which is of finsler type @xcite . indeed , measuring the length of a vector @xmath359 in the tangent space of @xmath336 at point @xmath11 by the hemi - norm @xmath360 , we get @xmath361 the infimum being taken over the set of curves @xmath362 from @xmath363 $ ] to the interior of the cone , such that @xmath364 and @xmath365 .
we recover as special cases of this construction the invariant riemannian metric @xcite ( @xmath352 is the euclidean norm ) , the log @xmath66-schatten metric @xcite ( @xmath352 is the @xmath366 norm ) , as well as metrics arising from symmetric gauge functions @xcite ( the latter are symmetric hemi - norms that are invariant by a change of sign of each variable , and take positive values except at the origin ) .
for any choice of symmetric hemi - metric @xmath352 , the map yields a family of geodesics for the hemi - metric @xmath358 joining a centre @xmath339 to any matrix @xmath359 , for which @xmath367 is a metrically star shaped space .
this can be proved along the lines of @xcite , where the same property is established in the special case in which @xmath352 is a symmetric gauge function .
the reader may also consult @xcite , in which the significance of several matrix inequalities in terms of non - positive curvature is brought to light .
we shall now consider specifically the case of order - preserving and positively homogeneous maps .
a self - map @xmath0 of @xmath256 is _ order - preserving _ if @xmath368 .
it is _ positively homogeneous _ ( of degree one ) if @xmath0 commutes with the product by a positive constant .
these notions make sense more generally if @xmath0 is only defined on a subset of @xmath256 invariant by any such product .
it is readily checked that a self - map @xmath0 of the interior of @xmath256 is order - preserving and positively homogeneous if and only if it is non - expansive in the funk metric .
[ cor - symm ] let @xmath0 be an order - preserving and positively homogeneous self - map of the interior of a symmetric cone @xmath256 , and let @xmath369 then , for all @xmath281 , @xmath370 and there exists an extreme ray of @xmath256 , such that for any element @xmath346 in this ray , @xmath371 the map @xmath0 is non - expansive in the funk metric . by corollary [ coregalite ] , @xmath225 .
let @xmath5 be the martin function appearing in theorem [ maintheorem ] .
then , there exists a sequence of elements of @xmath270 , @xmath321 such that @xmath373 converges to @xmath5 .
let us now fix @xmath374 .
since @xmath342 , any element @xmath174 of an extreme ray of @xmath274 can be written as @xmath375 where @xmath346 is an element of an extreme ray of @xmath256 .
then , it follows from lemma [ lem - funk ] that for all @xmath323 , there exists an element @xmath376 such that @xmath377 a ray of @xmath256 is known to be extreme if and only if it contains a primitive idempotent of the associated jordan algebra ( * ? ? ?
iv.3.2 ) . moreover ,
the set of these primitive idempotents is compact ( exercise 5 of chap . 4 in @xcite ) .
since the previous property is invariant by a scaling of @xmath378 , we may assume that each @xmath378 is a primitive idempotent , and so , possibly after extracting a subsequence , we may assume that @xmath378 converges to a primitive idempotent @xmath346 , which therefore belongs to an extreme ray of @xmath256 . for all @xmath379
, we get @xmath380 which concludes the proof .
we next point out a variant of corollary [ cor - symm ] concerning order - preserving and sub - homogeneous self - maps of the interior of a symmetric cone .
recall that _ sub - homogeneous _ means that for all @xmath281 and @xmath381 , @xmath382 .
order - preserving and sub - homogeneous maps are easily seen to be characterised by the following property @xmath383 [ cor - symm2 ] let @xmath0 be an order - preserving and sub - homogeneous self - map of the interior of a symmetric cone @xmath256 , and let @xmath384 then , for all @xmath281 , @xmath385 and , if @xmath119 , there exists an extreme ray of @xmath256 , such that for any element @xmath346 in this ray , @xmath371 since the hemi - metric @xmath316 is bounded from below , the observation made at the end of the proof of theorem [ maintheorem ] shows that if @xmath327 is any sequence of points of @xmath270 such that @xmath373 converges to the martin function @xmath5 in this theorem , then , the sequence @xmath327 must be such that @xmath322 .
then , it follows from the characterisation of the boundary of the @xmath316 metric ( proof of proposition [ prop - walsh - funkp ] ) that @xmath5 , which can not be of the form , is necessarily of the form .
therefore , we conclude as in the proof of corollary [ cor - symm ] .
corollary [ cor - symm ] should be compared with a related result of gunawardena and walsh ( * ? ? ?
* th . 1 ) , in which the map @xmath0 is only required to be non - expansive in thompson s metric .
then , it is shown that all the accumulation points of the sequence @xmath386 belong to the same face of a ball which respect to the norm in the tangent space of @xmath256 at point @xmath30 . if @xmath0 is an order preserving and positively sub - homogeneous self - map of the interior of a ( closed , convex and pointed ) cone @xmath256 in a finite dimensional vector space , we define the _ radial extension _
@xmath387 of @xmath0 to the closed cone @xmath256 by @xmath388 where @xmath156 is any point in the interior of @xmath256 .
one verifies that this definition is independent of the choice of @xmath156 ; equivalently , we may define @xmath389 as @xmath390 .
the order preserving and positively ( sub-)homogeneous character of @xmath0 carries over to @xmath387 .
it is shown in @xcite that @xmath0 does not necessarily have a continuous extension to @xmath256 , but that it always does if @xmath256 is polyhedral .
[ lem - ourcw ] let @xmath0 be an order - preserving and positively sub - homogeneous self - map of the interior of a ( convex closed and pointed ) cone @xmath256 in a finite dimensional vector space , and consider the martin function @xmath5 in associated to a vector @xmath312 .
then , for all @xmath391 , @xmath392 since , @xmath393 , the left - hand side condition in holds if and only if @xmath394 for @xmath395 and @xmath396 , define @xmath397 , where @xmath156 is a given element in @xmath270 . then , @xmath398 .
taking @xmath399 in , and letting @xmath400 tend to zero , we get @xmath401 and so , the condition at the right - hand side of holds .
conversely , fix @xmath281 , and choose @xmath402 such that @xmath403 .
we have @xmath404 cancelling @xmath402 , we arrive at .
observe that when @xmath0 is positively sub - homogeneous , the following _ recession map _
@xmath405 is well defined .
of course , @xmath406 if @xmath0 is positively homogeneous .
we now arrive at the following corollary of theorem [ th - minimax ] , which extends the collatz - wielandt formula .
[ th - ourcw ] let @xmath0 be an order - preserving and positively sub - homogeneous self - map of the interior of a symmetric cone in a finite dimensional vector space , and assume that @xmath407 then , @xmath408 moreover , when @xmath0 is positively homogeneous , the same conclusion remains valid even when the condition @xmath7 does not hold .
we apply theorem [ maintheorem ] to the space @xmath258 equipped with the hemi - metric @xmath409 .
let us denote by @xmath186 the escape rate of @xmath0 in this hemi - metric , so that @xmath410 since @xmath411 , the hemi - metrics @xmath311 and @xmath316 coincide as soon as @xmath316 or @xmath311 is positive .
hence , @xmath7 implies that @xmath412 and that the term @xmath413 in can be replaced by @xmath414 .
together with , this gives the first equality in .
we observed in the proof of corollary [ cor - symm2 ] that when the escape rate in the @xmath316 hemi - metric is positive , the corresponding horofunction @xmath5 in theorem [ maintheorem ] is necessarily of the form .
since the limit coincides with the infimum in the definition of the recession function ( see ) , the right - hand side condition in is equivalent to @xmath415 then , using lemma [ lem - ourcw ] , we rewrite the maximin characterisation in theorem [ maintheorem ] as @xmath416 moreover , it follows from remark [ rk - funk - symmetric ] that the latter infimum is attained .
finally , when @xmath0 is positively homogeneous , we equip the space @xmath270 with the hemi - metric @xmath311 instead of @xmath316 .
then , the first equality in follows readily from theorem [ maintheorem ] and ( the assumption that @xmath7 is not needed any more ) , and the proof of the second inequality is unchanged . when @xmath256 is the standard positive cone , and @xmath0 is a _ continuous _ order preserving and positively homogeneous self - map of the _ closed _
cone @xmath256 , nussbaum @xcite showed that @xmath418 which is more accurate than .
moreover , some of the results of @xcite have been extended in @xcite to more general ( normal , possibly infinite dimensional ) cones .
however , corollary [ th - ourcw ] is applicable when @xmath0 is only defined on the _ interior _ of the cone and does not extend continuously to its boundary ( this situation does occur in practice , see the discussion in @xcite of the matrix harmonic mean ) .
[ ex - riccati ] we finally illustrate the previous denjoy - wolff type results with an example originating from quadratic optimal control .
let @xmath419 be @xmath13 positive semidefinite matrices , and let @xmath14 be a @xmath13 invertible matrix .
consider the following discrete time riccati operator , which is a self - map of the interior of the cone @xmath336 of @xmath13 positive semidefinite symmetric matrices , @xmath420 where @xmath421 denotes the transpose of @xmath14 .
we refer the reader to @xcite for more background on these maps . in particular , @xmath0 is order - preserving , and , as shown by liverani and wojtkowski @xcite , @xmath0 is non - expansive in the thompson metric .
we noted above that the latter property , together with the preservation of order , implies that @xmath0 is positively sub - homogeneous , and so , corollaries [ cor - symm2 ] and [ th - ourcw ] apply to the map @xmath0 .
if the matrices @xmath195 and @xmath196 are positive definite , then @xmath0 is known to be a strict contraction in thompson metric @xcite .
it follows that @xmath0 has a unique fixed point in @xmath367 to which every orbit converges .
in particular , the escape rate @xmath29 is zero .
we next examine a more interesting case in which @xmath196 is of rank one , leading to a positive escape rate .
assume that @xmath422 , write @xmath423 for some ( column ) vector @xmath424 , and assume that @xmath14 is of the form @xmath425 , with @xmath426 , where @xmath427 is the identity matrix .
we claim that @xmath428 . to see this ,
let @xmath290 denote a non - zero vector orthogonal to @xmath424 , and consider @xmath429 .
then , it is easily checked that the radial extension of @xmath0 satisfies @xmath430 for all @xmath396 , and so , @xmath431 . using
, we deduce that @xmath432 . to show that the equality holds
, we observe that @xmath433 .
hence , for all @xmath434 , @xmath435 , where @xmath436 denotes the maximal eigenvalue of @xmath195 , which by , implies that @xmath437 . letting @xmath63 tend to infinity , we arrive at @xmath438 , and so @xmath428 .
a possible choice of the horofunction @xmath5 appearing in theorem [ maintheorem ] is @xmath439 where the basepoint is the identity matrix @xmath427 .
then , the linear form @xmath440 constructed in the proof of corollary [ cor - symm2 ] can be checked to be @xmath441 ( or any scalar multiple of this form ) . for @xmath422 ,
the cone of positive semidefinite matrices @xmath336 has the shape of a lorentz ( ice - cream ) cone .
then , the superlevel set @xmath442 is nothing but the interior of the translated lorentz cone @xmath443 .
this is illustrated in figure [ fig - denjoy ] .
applied to a discrete riccati operator on the cone of positive definite matrices @xmath444 ( example [ ex - riccati ] ) .
the horoballs ( superlevel sets ) of the horofunction @xmath5 are nested cones ( in green ) intersecting the cone @xmath445 ( in blue ) along the ray generated by the matrix @xmath446 .
the map @xmath0 sends every horoball @xmath447 to the horoball @xmath448 .
three such horoballs are represented , with apices @xmath446 , @xmath449 , @xmath450 , corresponding to @xmath451 , respectively . ]
this work originates from discussions of brian lins and the first author at the aim workshop on non - negative matrix theory in palo alto in 2008 , where they jointly obtained an early version of theorem [ maintheorem ] concerning the special banach space case .
the first author thanks brian lins for these discussions , and also for more recent observations .
m. gromov .
hyperbolic manifolds , groups and actions . in _ riemann surfaces and related topics : proceedings of the 1978 stony brook conference ( state univ .
new york , stony brook , n.y . , 1978 )
_ , volume 97 of _ ann . of math .
_ , pages 183213 , princeton , n.j . , 1981 .
princeton univ . press .
a. neyman .
stochastic games and nonexpansive maps . in _
stochastic games and applications ( stony brook , ny , 1999 ) _ , volume 570 of _ nato sci .
ser . c math .
_ , pages 397415 .
kluwer acad .
publ . , dordrecht , 2003 . | we establish a maximin characterisation of the linear escape rate of the orbits of a non - expansive mapping on a complete ( hemi-)metric space , under a mild form of busemann s non - positive curvature condition ( we require a distinguished family of geodesics with a common origin to satisfy a convexity inequality ) .
this characterisation , which involves horofunctions , generalises the collatz - wielandt characterisation of the spectral radius of a non - negative matrix .
it yields as corollaries a theorem of kohlberg and neyman ( 1981 ) , concerning non - expansive maps in banach spaces , a variant of a denjoy - wolff type theorem of karlsson ( 2001 ) , together with a refinement of a theorem of gunawardena and walsh ( 2003 ) , concerning order - preserving positively homogeneous self - maps of symmetric cones .
an application to zero - sum stochastic games is also given . |
current knowledge of the ckm matrix element @xmath3 is derived from the mesonic decays @xmath4 or @xmath5 .
experimental knowledge of the @xmath6 semileptonic decay can lead to an independent estimate of @xmath3 if the effect of the strong interaction in the decay are understood , _
e.g. _ , via lattice qcd .
a first lattice study of the baryonic semileptonic decay was performed by the ukqcd collaboration @xcite .
we report our initial results for the dominant form factors of this decay .
the semileptonic decay @xmath7 can be parametrized in terms of six form factors , @xmath8 and @xmath9 , for i = 1 , 2 , 3 .
@xmath10u^{(r)}_{q'}(v ' ) .
\label{velform}\end{aligned}\ ] ] here @xmath11 is the weak current and @xmath12 are polarisation states of the baryons . since both @xmath6 and @xmath13 are hadrons containing a single heavy quark , heavy quark effective theory ( hqet ) is applicable @xcite .
hence the matrix element is taken between baryons of a given velocity , and the form factors are functions of the scalar @xmath14 . to leading order in hqet , the combinations @xmath15 and @xmath16 involving the dominant form factors @xmath17 and @xmath16 can be written in terms of a single function , called the ( baryonic ) isgur - wise function , @xmath18 .
this function is normalised at zero - recoil , @xmath19 .
the simulations are performed on the asqtad quenched lattices at @xmath20 generated by the milc collaboration @xcite . these are @xmath21 improved @xmath22 64 lattices , with @xmath23 = 1.33 gev , as determined from @xmath24 .
we use three light quark masses near the strange quark mass , @xmath25 .
two heavy quark @xmath26 values , 0.104 and 0.114 bracket the charm quark , and other two , 0.064 and 0.077 bracket the bottom .
we use the clover action for the valence quarks , with a tadpole improved clover coefficient .
the value for the tadpole improvement factor @xmath27 is taken from the landau gauge fixed mean link .
fermilab formalism is used for the heavy quarks .
results are presented for 300 lattices for two - point functions , and 237 lattices for three - point functions .
the dispersion relation is shown in fig . [ disp_fig ] .
the fitted energy values agree very well with the expectation from the lattice dispersion relation .
the chiral extrapolations for a fixed heavy quark mass are shown in fig .
[ chiral_fig ] .
the baryon kinetic mass @xmath28 is estimated as @xmath29 , where @xmath30 and @xmath31 are the baryon and heavy quark kinetic(rest ) masses respectively .
we use a linear fit for these extrapolations . in fig .
[ baryon_mass ] , we have shown the chirally extrapolated baryon mass as a function of the heavy quark mass , along with the coresponding meson masses taken from the milc collaboration .
our values for @xmath32 and @xmath33 are 5.626(36)gev and 2.300(27)gev .
different form factors contribute to different matrix elements in the three - point function . for @xmath34 , the dominant contribution to
three - point functions comes from the vector form factors and for @xmath35 , axial - vector form factor @xmath16 gives the dominant contribution .
we present results for the isgur - wise function from vector as well as axial - vector data .
@xmath6 is created at time 0 and @xmath13 is annihilated at time @xmath36 in lattice units .
the time at which the current acts is varied , and we study three - point function as a function of this time @xmath37 . for the results presented here ,
the intial baryon is at rest and the final baryon is moving with different velocities giving different values for @xmath38 . on the lattice , one
is restricted to region near @xmath39 as data starts getting noisy for high momenta . in this region ,
@xmath40 can be approximated by 1 .
then for an initial baryon of mass @xmath41 decaying to a final baryon of mass @xmath42 moving with a momentum @xmath43 , if we consider the sum of the co - efficients of @xmath44 and @xmath45 , for large @xmath46 and @xmath47 , the three - point expression simplifies to where @xmath54 and @xmath55 are known from the two - point functions .
we fit this to a form @xmath56 and consider the ratio @xmath57}{a[(m',\vec{0 } ) \rightarrow ( m,\vec{0 } ) ] } & & \hspace{-0.5cm}= \big ( \frac{f_1(\omega)+f_2(\omega)+f_3(\omega)}{f_1(1)+f_2(1)+f_3(1 ) } \big)\nonumber\\ & & \hspace{-0.9cm}\cdot\big ( \frac{z'_s(|\vec{q}|)}{z'_s(0)}\big ) \cdot \big ( \frac{e+m}{2e}\big).\end{aligned}\ ] ] first factor on the rhs is(1/m ) corrections , which are further multiplied by ( 1-@xmath38 ) . for the small momenta accessible on the lattice
, @xmath58 is a very good approximation to the baryonic isgur - wise function .
the suffix @xmath59 emphasizes that the infinite mass limit has not been taken . ]
the isgur - wise function @xmath60 .
the second and third factors are known from the two - point functions .
the third factor may be approximated by 1 to 0.5 per cent accuracy .
the second factor differs from 1 by upto 10% over our range of @xmath61 .
the ratio is independent of the renormalization constant @xmath62 because we have the same heavy quark transition in both numerator and denominator .
our results for the isgur - wise function from the vector current are shown in fig .
[ vec_iw ] .
the isgur - wise function obtainted from the axial - vector current ( @xmath35 case ) is shown in fig .
[ ax_iw ] .
the isgur - wise function seems to be quite insensitive to the heavy quark mass .
we have also studied the light quark mass dependence of the isgur - wise function .
the isgur - wise function is expected to fall slower for smaller light quark masses , by a heuristic argument .
we do see such a trend in this preliminary study , but it is very far from clear with the statistical errors we have .
this is shown in fig .
[ light ] . | we present results from a lattice study of the semileptonic decay @xmath0 .
we use @xmath1 improved quenched lattices of the milc collaboration , with lattice spacing @xmath2 fm . for the valence quarks ,
the tadpole - improved clover action is used , with the fermilab method employed for the heavy quarks .
form factors are extracted from the vector as well as the axial - vector part of the current . |
the spectra of quasars show a @xmath0forest `` of absorption lines blueward of the ly-@xmath1 emission line ( lynds 1971 , sargent et al . 1980 , weymann , carswell , & smith 1981 ) .
observational and theoretical work in recent years has shown that most of this absorption can be attributed to neutral hydrogen in galaxies and large - scale structure along the line of sight ( cen et al .
1994 , lanzetta et al .
1995,1996 , zhang et al .
1995 , hernquist et al .
1996 , miralda - escud et al .
1996 , bi & davidsen 1997 , chen et al .
1998 , theuns et al .
1998 , ortiz - gil et al . 1999 ,
impey , petry , & flint 1999 , dav et al . 1999 ,
bryan et al .
1999 ) . in aggregate ,
qso spectra show an increasing line density with increasing redshift such that @xmath19 ( sargent et al .
1980 , weymann , carswell , & smith 1981 , young et al .
1982 , murdoch et al .
1986 , lu , wolfe , & turnshek , 1991 , bechtold 1994 , kim et al .
but the line density within an individual quasar spectrum decreases with proximity to the ly-@xmath1 emission line ( weymann , carswell , & smith 1981 , murdoch et al .
this is generally thought to be due to enhanced ionization of neutral hydrogen in the vicinity of the quasar due to ionizing photons from the quasar itself .
this @xmath0proximity effect '' can be used to measure the mean intensity of the uv background , denoted @xmath3 ( carswell et al .
1987 , bajtlik , duncan , & ostriker 1988 , hereafter bdo ) .
@xmath3 has been measured at @xmath20 by a variety of authors ( bdo , lu , wolfe , & turnshek 1991 , giallongo et al . 1993,1996 , bechtold 1994 , williger et al .
1994 , cristiani et al .
1995 , fernndez - soto et al .
1995 , lu et al .
1996 , savaglio et al . 1997 ,
cooke et al .
1997 , scott et al .
the results , summarized in paper ii of this series ( scott et al .
2000b ) , are in general agreement with the predictions of models of the uv background which integrate the contribution from known population of quasars and include reprocessing effects in an inhomogeneous intergalactic medium ( haardt & madau 1996 , hereafter hm96 , fardal et al . 1998 ) . in paper ii
, the mean intensity of the ionizing background was found to be @xmath21 ergs s@xmath5 @xmath6 hz@xmath5 sr@xmath5 at @xmath22 .
the decline of the quasar space density from @xmath17 to the present is expected to drive a corresponding decline in the intensity of the uv background .
kulkarni & fall ( 1993 , hereafter kf93 ) measured @xmath23 ergs s@xmath5 @xmath6 hz@xmath5 sr@xmath5 at @xmath24 from a subset of the now complete hst quasar absorption line key project sample presented by bahcall et al .
( 1993 ) .
much of this previous work has relied upon the technique for measuring @xmath3 outlined by bdo .
this technique requires the entire sample of absorption lines to be binned according to the ratio of the quasar flux at the physical position of the absorber to the background flux .
this is done for several initial guesses of the background intensity ; and the value that gives the lowest @xmath25 between the binned data and the ionization model is chosen as the best fit @xmath3 . however , this is not the optimal technique to use at low redshift where absorption line densities are low .
kf93 developed a maximum likelihood technique to address this issue and used it in their measurement of @xmath3 at @xmath24 .
however , their measurement was based upon a sample of only 13 qsos and less than 100 lines , and has correspondingly large error bars . in addition
, the value these authors find is lower than the predictions of the models of haardt & madau ( 1996 ) , though consistent within the uncertainties , as shown in figure 13 of paper ii and in figure [ fig : lowzcomp ] of this paper . given the importance of the value of the hi ionization rate to the hydrodynamical evolution of the low redshift universe , performing this measurement with a much larger line sample is worthwhile .
the low redshift hydrodynamic simulations of theuns et al .
( 1998 ) and dav et al .
( 1999 ) indicate that the evolution of the ionizing background is the primary driver behind the change of character of the ly-@xmath1 forest from high redshift to low redshift , specifically , the break in the number distribution of ly-@xmath1 lines at @xmath26 ( morris et al .
1991 , bahcall et al .
1991 , weymann et al .
the growth of structure pulling gas from low density regions into high density regions also contributes to this and other attributes of the evolution of the ly-@xmath1 forest .
shull et al .
( 1999 ) estimate the local ionizing background including contributions to the background from starburst galaxies as well as seyferts and qsos .
their models include a treatment of the opacity of the low redshift ly-@xmath1 forest using information drawn from recent observational work ( weymann et al .
1998 , penton et al .
they find that starbursts and agn could contribute approximately equally to the ionizing background at low redshift , each @xmath27 ergs s@xmath5 @xmath6 hz@xmath5 sr@xmath5 .
the full hst / fos archival data set is presented in paper iii and can also be found at or , so we describe the data used in this paper only briefly in [ sec - data ] . in [ sec - zsys ] and [ sec - flux ] , we discuss our treatment of two parameters of each sample object which are integral to the proximity effect analysis , systemic redshifts and lyman limit fluxes .
we outline the proximity effect analysis in [ sec - analysis ] and we present our results in [ sec - disc ] .
we discuss the recovery of @xmath3 from simulated ly-@xmath1 forest spectra in
[ sec - sims ] .
a comparison of the results from radio loud and radio quiet qsos is given in
[ sec - rl ] .
the maximum likelihood solutions for @xmath3 found when allowing for an equivalent width threshold that varies across each sample spectrum are discussed in [ sec - varthr ] .
solutions for the hi ionization rate are given in
[ sec - gam ] .
the effect of a non - zero @xmath28 on our calculations is discussed in [ sec - omegal ] , and the effect of the ionizing background on the ly-@xmath1 forest line density is discussed in [ sec - dndz ] .
we provide comparisons with previous observational work on the low redshift uv background in [ sec - prevres ] and with models of this background in [ sec - models ] .
a discussion of possible systematic effects on this analysis is given in
[ sec - systematics ] .
we conclude with a summary of the results in [ sec - summary ] .
the reduction of the fos data is described in paper iii .
table [ tab - z ] lists the objects used in the proximity effect analysis along with the object s redshift and classification in the nasa extragalactic database . for the reasons outlined in scott et al .
( 2000b ) we have removed from the full fos sample of paper iii the spectra of quasars known to be lensed , as well as those that show damped ly-@xmath1 absorption , associated absorption , or broad intrinsic absorption . for our primary proximity effect sample
, we also remove objects classified as blazars ( bl lacs and optically violent variables ) on the grounds that their continua are highly variable .
however , we also perform the proximity effect analysis with associated absorbers , damped ly-@xmath1 absorbers , and blazars included in order to determine if they affect the results obtained . as discussed in paper iii , objects observed only in the period before the costar upgrade to the hst optics and with the a-1 fos aperture are particularly subject to irregular line spread functions . we have omitted those data from this analysis as well .
the distributions in redshift of the qsos and absorption lines used in this paper are shown in figure [ fig - samphist ] .
qso redshifts based on the ly-@xmath1 emission line have been shown to be blueshifted from the systemic redshift based on narrow emission lines by up to @xmath29 km s@xmath5 .
generally , the forbidden oiii doublet at 4959 , 5007 is taken to be the most reliable indicator of the qso systemic redshift ; though other lines such as mg ii @xmath302796,2803 and balmer lines have been shown to trace the systemic redshift as well , with some spread .
( zheng & sulentic 1990 , tytler & fan 1992 , laor et al . 1994,1995 , corbin & boroson 1996 ) however , the results of nishihara et al .
( 1997 ) , m@xmath31intosh et al .
( 1999 ) , and scott et al .
( 2000b ) indicate that in fact h@xmath32 may not reflect the systemic redshift of high redshift qsos .
spectra of the emission lines h@xmath32 , [ oiii]@xmath335007 , or mg ii were obtained for several objects in our total proximity effect sample .
the observations were carried out on the nights of 19 december 1995 , 14 january 1996 , 20 and 21 april 1996 , 12 and 13 december 1996 , and 2 february 1997 .
these observations are summarized in table [ tab - zhst ] .
the 19 december 1995 and 13 december 1996 observations were made using the 1.5 meter tillinghast telescope at the fred lawrence whipple observatory using the fast spectrograph ( fabricant et al .
1998 ) and a thinned loral 512x2688 ccd chip ( gain = 1.06 , read noise = 7.9 e@xmath34 ) binned by a factor of 4 in the cross - dispersion direction .
observations were made using a 300 lines mm@xmath5 grating blazed at 4750 and a 3@xmath35 slit . these spectra cover a wavelength range of 36607540 .
this is listed as set - up ( 1 ) in table [ tab - zhst ] .
the january , april , and december 10 and 12 , 1996 observations were made using the steward observatory bok 90 inch telescope using the boller and chivens spectrograph with a 600 l mm@xmath5 grating blazed at 6681 in the first order , a 1.5@xmath35 slit , and a 1200 x 800 ccd array with a gain of 2.2 e@xmath34 adu@xmath5 and a read noise of 7.7 e@xmath34 , binned 1x1 . for the january 1996 observations ,
the data were obtained with one of two grating tilts , one resulting in wavelength coverages of 36005825 and 68709140 . for the april 1996 data ,
the wavelength ranges were 41406370 and 52807550 .
two grating tilts were also used for the december 1996 data , giving wavelength coverages of 45006700 and 56107860 .
the spectrum of one object , 0827 + 2421 , was obtained on 15 february 1997 at the multiple mirror telescope with the blue channel spectrograph , a 2 @xmath35 slit , the 3k x 1k ccd array , and the 800 l mm@xmath5 grating blazed at 4050 with spectral coverage of 43656665 .
the spectra are shown in figure [ fig - zspec ] and the lines used for redshift measurements are labeled .
taking a simple cursor measurement of each line centroid , we find a mean [ oiii]-balmer line @xmath36v of -30 @xmath37 1010 km s@xmath5 for 31 objects and a mean [ oiii]-mg ii @xmath36v of 58 @xmath37 576 km s@xmath5 for 31 objects .
the mean blueshift of the ly-@xmath1 emission line with respect to [ oiii ] is 289 @xmath37 727 km s@xmath5 based on 51 measurements .
the redshifts measured for each object in our sample are shown in table [ tab - z ] ; and the results are shown in figure [ fig - zhist ] .
gaussian fits to the lines give similar results .
we therefore treat both balmer lines and mg ii in addition to [ oiii ] as good systemic redshift indicators for these low redshift objects . in the case of a qso for which we have only a ly-@xmath1 emission line measurement of the redshift
, we add 300 km s@xmath5 to this value to estimate its systemic redshift .
our method for estimating lyman limit fluxes for each qso is the same as that described in paper ii . for objects with spectral coverage between the ly-@xmath1 and civ emission lines ,
we extrapolate the flux from 1450 in the quasar s rest frame to 912 using @xmath38 and a spectral index @xmath1 measured primarily from the spectral region between the ly-@xmath1 and c iv emission lines .
figure [ fig - fnus ] shows the fos spectra for which these fits were made along with the power law fits themselves . in some cases , @xmath1 is poorly constrained from these fits , especially if there was little spectral coverage redward of ly-@xmath1 emission in the data . if another measurement of the spectral index was available in the literature for these objects , we used it ; otherwise , we used our measurement . table [ tab - flux ] lists the lyman limit flux for each object in this proximity effect sample and either a ) the flux at 1450 , or some other appropriate wavelength free of emission features , measured from the fos data , or b ) a directly measured lyman limit flux and the reference . if available from the extracted archive data , red spectra and the fits to them are presented for objects which were observed only with pre - costar fos and a-1 aperture , though these data were not subsequently used for any ly-@xmath1 forest studies . see table 4 of paper iii .
in figure [ fig - zl ] , we show qso lyman limit luminosities versus emission redshift for this hst / fos sample combined with the high redshift objects presented in papers i and ii . only at the lowest redshifts
is there any trend of luminosity with redshift .
the distribution of ly-@xmath1 lines in redshift and equivalent width is given by : @xmath39 the distribution in redshift and hi column density , n , is : @xmath40 the parameter @xmath41 is the redshift distribution parameter .
the quantities @xmath42 in equ .
[ eq : dndzdw ] and @xmath32 in equ .
[ eq : dndzdnh1 ] are the line rest equivalenth width and column density distribution parameters , respectively .
the quantities @xmath43 and @xmath44 are normalizations .
the bdo method for measuring @xmath3 consists of binning all lines in the sample in the parameter @xmath45 , the ratio of qso to background lyman limit flux density at the physical location of the absorber : @xmath46 for various values of @xmath3 .
the value of @xmath3 that results in the lowest @xmath25 between the binned data and the ionization model , @xmath47^{-(\beta-1 ) } , \label{eq : dndx}\ ] ] is considered to be the optimal value .
this ionization model follows from the assumption that the column densities of lines are modified by the presence of the qso according to @xmath48 where @xmath49 is the column density a given line would have in the absence of the qso .
the 1@xmath50 errors are found from @xmath51 for 7 degrees of freedom ( press et al .
1992 ) .
the value of @xmath45 for each line in a given sample depends not only upon the value of @xmath3 assumed , but also on the cosmological model , as @xmath52 and @xmath53 where @xmath54 is the luminosity distance of an individual absorber from the qso and @xmath55 luminosity distance to the qso from the observer .
the luminosity distance between two objects at different redshifts can be calculated analytically for cosmological models in which @xmath16 .
we return to this point in section [ sec - omegal ] below .
if the proximity effect is indeed caused by enhanced ionization of the igm in the vicinity of qsos , one may expect to observe a larger deficit of lines relative to the ly-@xmath1 forest near high luminosity qsos than near low luminosity qsos . in figure [ fig - npred](a ) , we plot the fractional deficit of lines with respect to the number predicted by equ .
[ eq : dndzdw ] versus distance from the qso for this hst / fos sample combined with the high redshift objects observed with the multiple mirror telescope ( mmt ) presented in papers i and ii . we divide our qso sample into high and low luminosity objects at the median lyman limit luminosity of the combined mmt and hst / fos sample , log(l@xmath56 ) @xmath57 31 .
high luminosity objects show a marginally more pronounced proximity effect than low luminosity objects : 4.9@xmath50 for qsos with log(l@xmath56 ) @xmath58 31 versus 3.2@xmath50 for qsos with log(l@xmath56 ) @xmath59 31 . in panel
( b ) , we plot the line deficit within 2 h@xmath60 mpc as a function of log(l@xmath56 ) .
the lack of a significant difference in the line deficit between high and low luminosity qsos may indicate the presence of clustering , if absorption features cluster more strongly around more luminous qsos with deeper potential wells .
we will address the issue of clustering further below .
the bdo method of measuring the background can result in poor statistics at low redshift due to the low line density in the low redshift ly-@xmath1 forest .
we will quote results from this method , but we will generally the maximum likelihood method for measuring @xmath3 as presented by kf93 , which consists of constructing a likelihood function of the form @xmath61 \label{eq : maxlike},\ ] ] where @xmath62^{-(\beta-1 ) } , \label{eq : fnz}\ ] ] and the indicies @xmath63 and @xmath64 denote sample absorption lines and quasars , respectively . using the values of @xmath41 and @xmath43 from a separate maximum likelihood analysis on the ly-@xmath1 forest excluding regions of the spectra affected by the proximity effect ( dobrzycki et al .
2001 , paper iv ) , and a value of @xmath32 from studies with high resolution data , eg .
@xmath65 from hu et al .
( 1995 ) , the search for the best - fit value of @xmath3 consists of finding the value that maximizes this function , fixing the other parameters .
if the line density is low throughout a single ly-@xmath1 forest spectrum , it becomes difficult to distinguish any proximity effect , even in a large sample of spectra .
the absence of a proximity effect in this model formally translates into the limit @xmath66 because in this scenario , the qso has no additional effect on its surroundings and therefore generates no relative line underdensity .
the errors quoted in the values of log[@xmath3 ] are found from the fact that in solving for log[@xmath3 ] alone , the logarithm of the likelihood function , @xmath68 , is distributed as @xmath25 with one degree of freedom . in the case of an ill - defined solution
, the likelihood function is very broad and the formal error approaches infinity . if a proximity effect is weak but not absent in the data , a maximum likelihood solution is sometimes possible , but with no well - defined 1@xmath50 upper limit on log[@xmath3 ] . in other words ,
if an upper limit of infinity is quoted , the data can not rule out the nonexistence of a proximity effect to within 1@xmath50 confidence . using a constant equivalent width threshold results in the loss of a large amount of spectral information . in the case of a large equivalent width threshold , of course ,
many weak lines are discarded ; and in the case of a small threshold , regions of spectra where the signal - to - noise ratio ( s / n ) does not permit the detection of lines all the way down to the specified threshold are lost and only the highest s / n spectral regions are used .
the technique of measuring the statistics @xmath41 and @xmath42 has been expanded to allow for a threshold that varies with s / n across each qso spectrum ( bahcall et al .
1993,1996 , weymann et al .
1998 , scott et al .
we will use this variable threshold information to measure @xmath3 as well .
the results of this analysis are given in table [ table - jnu ] . before we begin the discussion of the results , some words about the normalization values listed in table [ table - jnu ] are in order . in the bdo method for measuring @xmath3 ,
lines are binned in @xmath45 and compared to the ionization model given by equ .
[ eq : dndx ] , for an assumed value of @xmath32 . in this case , the normalization listed in table [ table - jnu ] is the parameter in equ .
[ eq : dndzdw ] , found from the number of lines in the sample and the maximum likelihood value of @xmath41 : @xmath69 where @xmath70 is the total number of lines observed with rest equivalent width greater than @xmath71 , the limiting equivalent width of the line sample .
for the maximum likelihood solutions for @xmath3 , we convert line equivalent widths to column densities using the ly-@xmath1 curve of growth and an assumed value of @xmath72 , the characteristic doppler parameter of the lines .
as we will demonstrate , different values of @xmath32 and @xmath72 have only a small effect on the value of @xmath3 found .
the normalization is given by @xmath73 where @xmath74 is the limiting column density across each qso spectrum corresponding to a limiting equivalent width .
this quantity can be held constant , as in the bdo method , or it can be allowed to vary across each qso spectrum . in both of these formulations for the normalization ,
a proximity region around the qso is neglected and that proximity region is either defined by a velocity cut , eg .
@xmath75 - 3000 km s@xmath5 , or by a cut in @xmath45 , eg .
@xmath76 .
we also use the standard bdo method to find @xmath77= -22.04^{+0.43}_{-1.11}$ ] and @xmath78 for equivalent width thresholds of 0.32 and 0.24 respectively .
figures [ fig : chi2](a ) and ( d ) illustrate the @xmath25 of the binned data compared to the bdo ionization model as a function of assumed @xmath79 ) for these two thresholds . the bdo ionization model is expressed in terms of the number of lines per coevolving coordinate : @xmath80 where @xmath81 .
this @xmath25 curve is very broad , which is reflected in the large error bars and indicates the difficulty in isolating the optimal mean intensity of a weak background using this technique .
figures [ fig : dndx](a ) and ( d ) show the binned data and the ionization model for the values of @xmath79 ) listed above , those that give the lowest @xmath25 between the binned data and the model , ie . the minima of the curves in figures [ fig : chi2](a ) and ( d ) .
we executed the maximum likelihood search for @xmath3 , using two different fixed equivalent width thresholds , 0.24 and 0.32 as well as for the case of a variable threshold across all the spectra .
the uncertainty in @xmath41 does not translate directly into a large uncertainty in @xmath3 .
changing the value of @xmath41 alters the maximum likelihood normalization , @xmath44 , according to equ .
[ eq : mlnorm ] . from the sample of lines with rest equivalent widths greater than 0.32 we find @xmath77= -22.11^{+0.51}_{-0.40}$ ] for @xmath82 .
varying @xmath41 by @xmath83 gives @xmath77= -22.21 $ ] and @xmath84 with similar uncertainties .
the data used here are not of sufficient resolution to fit voigt profiles to the absorption features and derive hi column densities and dopper parameters .
we therefore choose to fix the values of @xmath32 and @xmath72 to those found from work on high resolution data , rather than allow them to freely vary in our analysis .
for the 0.32 fixed equivalent width threshold , we tested several pairs of values of @xmath85 where @xmath72 is in km s@xmath5 : ( 1.46,35 ) and ( 1.46,25 ) where the value of @xmath32 is taken from hu et al .
( 1995 ) ; as well as ( 1.45,25 ) and ( 1.70,30 ) found from low redshift ly-@xmath1 forest spectra taken with the goddard high resolution spectrograph ( ghrs ) on hst by penton et al .
( 2000a , b ) .
in addition , dav & tripp ( 2001 ) have found some evidence for @xmath32 increasing to 2.04 at @xmath86 from high resolution echelle data from the space telescope imaging spectrograph aboard the hst .
we test this value as well .
the likelihood functions for the maximum likelihood solutions listed in rows 2 - 6 , 8 - 12 , 14 , and 18 of table [ table - jnu ] are shown in figure [ fig : like1 ] .
the binned data and ionization models are plotted in figure [ fig : dndxl1 ] .
the values of @xmath3 derived for these various pairs of values of @xmath32 and @xmath72 are not significantly different from one another , though the results in table [ table - jnu ] indicate that varying @xmath32 has a larger impact on the inferred @xmath3 than does varying @xmath72 .
the solution for @xmath87 differs from the @xmath65 solution by @xmath88 . in the analysis that follows
, we adopt the values 1.46 and 35 km s@xmath5 .
the models of haardt & madau ( 1996 ) predict that the uv background arising from qsos drops by over an order of magnitude from @xmath89 to @xmath90 .
we therefore divide the sample into low and high redshift subsamples at @xmath91 and use both the bdo method and the maximum likelihood method for finding @xmath3 .
these results , also listed in table [ table - jnu ] , confirm some evolution in @xmath3 , though not at a high level of significance .
for the bdo solutions , we find log[@xmath3 ] at @xmath8 is equal to @xmath92 and log[@xmath3 ] at @xmath9 is equal to @xmath93
. the restrictive 1@xmath50 upper limit for log[@xmath3 ] at @xmath9 arises from the steeply rising @xmath25 as a function of log[@xmath3 ] shown in figure [ fig : chi2 ] .
this , in turn arises from the single line in the highest @xmath94 bin moving to the next bin for larger values of @xmath3 , resulting in a drastic change in the @xmath25 with respect to the photoionization model .
we do not consider this to be a reliable indicator of the uncertainty in @xmath3 at @xmath9 .
the maximum likelihood technique gives more robust estimates of the uncertainties . from this analysis ,
we find log[@xmath3 ] at @xmath8 is found to be @xmath95 , while at @xmath9 it is -21.98@xmath96 .
these results are shown in figures [ fig : lowzcomp](a ) and [ fig : allzcomp ] . including associated absorbers , damped ly-@xmath1 absorbers , or blazars in the proximity effect analysis appears to have little effect on the results .
one might expect associated absorbers to reduce the magnitude of the observed proximity effect and hence cause @xmath3 to be overestimated .
the value found including the 45 associated absorbers in our sample is indeed larger , log[@xmath3]@xmath97 , versus log[@xmath3]@xmath98 , but not significantly so .
likewise , if the intervening dust extinction in damped ly-@xmath1 absorbers is significant , including these objects in our analysis could cause us to overestimate the magnitude of the proximity effect and hence underestimate @xmath3 .
however , the inclusion of these 7 objects only negligibly reduces the value of @xmath3 derived .
qso variability on timescales less than @xmath99 years would be expected to smooth out the proximity effect distribution ( bdo ) .
however , the inclusion of 6 blazars in the sample , all at @xmath8 , resulted in no discernible change in @xmath3 . the sample used in the analysis of hi ionization rates discussed below includes all of these objects . for each solution
, we calculate the @xmath25 with respect to the ionization model expressed by equ .
[ eq : dndx ] , and the probability that the observed @xmath25 will exceed the value listed by chance for a correct model , q@xmath100 ( press et al .
we also execute a kolmogorov - smirnov ( ks ) test for each solution .
the ks test provides a measure of how well the assumed parent distribution of lines with respect to redshift , given by equ .
[ eq : fnz ] , reflects the true redshift distribution of lines ( cf .
murdoch et al .
1986 , press et al . 1992 ) .
the ks probability , q@xmath101 , indicates the probability that a value of the ks statistic larger than the one calculated could have occurred by chance if the assumed parent is correct . the ks probability associated with each solution for @xmath3
is listed in column 10 of table [ table - jnu ] .
we tested our maximum likelihood methods , including our treatment of the variable equivalent width thresholds by running our analysis on a simulated data set .
each of the 151 spectra in this simulated data set had a redshift equal to that of an object in our data set .
all objects including those showing associated absorption , damped ly-@xmath1 absorption , or blazar activity are included in this simulated set .
each spectrum is created using a monte carlo technique by which lines are placed in redshift and column density space according to equ .
[ eq : dndzdnh1 ] .
a background of known mean intensity modifies the column densities of the lines according to the bdo formulation given by equ .
[ eq : column ] . the same analysis done on the data , consisting of the line - finding algorithm and the maximum likelihood searches for @xmath41 and @xmath3 , is then used on the simulated spectra in order to recover the input @xmath79 ) .
three different values of log[@xmath3 ] are input , -21 , -22 , and -23 , and the results are listed in table [ table - sim ] . in order to understand the possible range of recovered log[@xmath3 ] , we repeated the input log[@xmath3]@xmath102 simulation in the constant threshold case nine additional times , resulting in @xmath103 ergs s@xmath5 @xmath6 hz@xmath5 sr@xmath5 .
in addition , since we observe the background to evolve with redshift from @xmath91 to @xmath90 , we implement a model in which @xmath3 varies as a power law in @xmath104 over the redshift range of the data .
this relationship is defined by the best fit to a power law variation of @xmath3 with redshift : @xmath77= 0.017\log(1+z)-21.87 $ ] .
we recover this using both the constant threshold and the variable threshold analyses , at all redshifts and at @xmath8 and @xmath9 separately .
the results of this exercise are shown in table [ table - sim ] and in figure [ fig : sims ] .
these simulation results indicate that both the constant and variable threshold analyses can overestimate the background by up to a factor of 3 - 5 , though the uncertainties for the variable threshold solutions are consistently lower , as a factor of @xmath105 more lines are used in these solutions .
we separated the first of the input log[@xmath3]@xmath102 simulated data samples into high and low redshift subsamples at @xmath91 , in order to determine if the change in @xmath3 as a function of redshift could be falsely introduced in a case there the input background is constant with redshift .
for both the constant and variable threshold treatments , this is not the case .
the value found for the low redshift subsample is actually larger than the value found for the high redshift subsample in both treatments . in the case of the varying input @xmath77
$ ] , the values recovered for the high redshift subsample and for the entire redshift range of the data are overestimates .
the slope of the linear relationship between @xmath77 $ ] and @xmath106 is quite small , 0.017 , resulting in a variable input @xmath77 $ ] that is actually nearly constant with redshift .
the solution for @xmath8 matches the input well for both the constant and variable threshold cases . at @xmath9 ,
the variable threshold solution overestimates the input by a larger factor , @xmath573 , or 1.6@xmath50 , than does the constant threshold solution , @xmath572 , or less than 1@xmath50 . in paper ii
, we argued that curve - of - growth effects are likely to come into play in the proximity effect analysis and to play a larger role for cases in which @xmath3 is large and the proximity effect signature is small .
here we find that the input @xmath3 is recovered most effectively by the constant and variable threshold cases for the largest input value of @xmath77 $ ] , @xmath107 .
however , nearly every case tested with these simulations results in a value of @xmath3 larger than the input value , especially when a variable equivalent width threshold is used .
the only case where the difference is significant is the input @xmath77=-23 $ ] , variable threshold case .
the recovered @xmath77 $ ] , -22.47 , is 4@xmath50 larger than the input .
we will return to the discussion of the variable threshold in section [ sec - varthr ] below . as described in paper ii , solving for the hi ionization rate , @xmath108 instead of @xmath3 avoids the assumption that the spectral indicies of the qsos and the background are identical .
we modified our maximum likelihood code to use the values of @xmath1 for each qso listed in table [ tab - flux ] to measure this quantity and the results are listed in table [ table - gamma ] . for objects with no available measured value of @xmath1 , we use @xmath109 , the extreme ultraviolet spectral index measured from a composite spectrum of 101 hst / fos qso spectra by zheng et al . (
the result for lines above a constant 0.32 rest equivalent width threshold is @xmath110 .
this result is not substantially changed if we instead use @xmath111 , the value found from a composite of 184 qso spectra from hst / fos , ghrs , and stis by telfer et al .
( 2001 ) , giving @xmath112 .
we also find little change in the result if we assume @xmath109 or @xmath111 for all qsos .
the variable threshold data result in a high hi ionization rate , and this is discussed further in the following section .
the constant threshold result is plotted in figure [ fig : gam ] .
evolution in the uv background is more apparent in the hi ionization rate than in the solutions for @xmath3 .
the result at @xmath9 is 6.5 times larger than that at @xmath8 .
the values of @xmath3 implied by these solutions for @xmath113 and a global source spectral index @xmath114 are also listed in table [ table - gamma ] .
we also parametrize the evolution of the hi ionization rate as a power law : @xmath115 and solve for the parameters @xmath116 and @xmath117 in both the constant and variable threshold cases .
the values we find are shown as the dashed line in figure [ fig : gam ] also listed in table [ table - gamma ] .
hm96 parametrize their models of the hi ionization rate with the function : @xmath118 we combine our data set with that of scott et al .
2000b to solve for the parameters @xmath119 , @xmath120 , @xmath121 , and @xmath122 .
we find @xmath123= ( @xmath124 , 0.35 , 2.07 , 1.77 ) for @xmath32=1.46 and @xmath123 = ( @xmath125 , 1.45 , 2.13 , 1.42 ) for @xmath32=1.7 , while the parameters found by hm96 for @xmath126 are ( @xmath12 , 0.43 , 2.30 , 1.95 ) .
these results are also represented by the solid curves in figure [ fig : gam ] , while the hm96 parametrization is shown by the dotted line for comparison .
the variable threshold analysis yielded some unexpected results .
as seen in the majority of the simulations , the values of @xmath79 ) found were consistently larger than the values found using a constant equivalent width threshold , indicating that the inclusion of weaker lines suppresses the proximity effect .
this is to be expected if clustering is occurring ( loeb & eisenstein 1995 ) , which in itself is to be expected to be more prominent at low redshift than at high redshift .
however , the suppression of the proximity effect by the inclusion of weak lines is somewhat counterintuitive from the perspective of the curve of growth .
most of the lines included in a constant threshold solution are on the flat part of the curve of growth .
therefore , though the ionizing influence of the quasar may be translated directly into a change in the hi column density , as predicted by the bdo photoionization model , this will not necessarily result in a corresponding change in the line equivalent width .
the solution for @xmath8 is nearly a factor of 3 larger than the the solution found in the case of a constant , 0.32 equivalent width threshold .
the solution for @xmath9 is a factor of @xmath127 larger than the constant threshold solution , with no well - defined 1@xmath50 upper limit due to the flattening of the likelihood function towards high @xmath79 ) this likelihood function for the total sample shows two peaks , the most prominent at log[@xmath79)]@xmath128 , the solution listed in table [ table - jnu ] , and a secondary peak at log[@xmath79 ) ] @xmath129 .
this behavior is also exhibited , even more dramatically , in the solutions for the hi ionization rate , as discussed above .
we conducted a jackknife resampling experiment ( babu & feigelson 1996 , efron 1982 ) to determine the source of these likelihood function peaks at large log(@xmath113 ) , or log[@xmath79 ) ] .
two objects , 0743 - 6719 ( @xmath130 ) and 0302 - 2223 ( @xmath131 ) , are found from jackknife experiments to produce all of this effect . in the jackknife experiment
, we perform the maximum likelihood calculation of @xmath79 ) n times , where n is the number of objects in the high redshift subsample . in each calculation , one object from the total sample is removed .
the results of this experiment are shown in the histogram in figure [ fig : jack ] .
the removal of 0743 - 6719 or 0302 - 2223 results in the two values of @xmath113 that are well - defined and that are in reasonable agreement with the value calculated at high redshift in the constant threshold case . removing only the one line from 0743 - 6719 nearest the ly-@xmath1 emission line with @xmath132 and observed equivalent width equal to 0.23 results in @xmath133 s@xmath5 .
this object was part of the hst key project sample ( jannuzi et al .
1998 ) and they cite no evidence of associated aborption in its spectrum . removing only the one line from 0302 - 2223 nearest the ly-@xmath1 emission line with @xmath134 and observed equivalent width equal to 0.27 results in @xmath135 s@xmath5 .
this object shows an absorption system at @xmath136 and is classified as an associated absorber .
no metal absorption is seen at @xmath134 , though this absorber is within 5000 km s@xmath5 of the qso , the canonical associated absorber region . removing both of these lines gives @xmath137 s@xmath5 . due to the small equivalent widths of both of these lines
they are not included in the constant threshold analysis , and the solutions for @xmath3 and @xmath113 for @xmath9 are well - defined .
it appears that this method has some trouble reliably recovering the background from a sample of absorption lines above an equivalent width threshold allowed to vary with s / n . as the method works well for the constant threshold case , we contend that the photoionization model , expressed in equ .
[ eq : dndx ] , used to create the likelihood function must not be an adequate model for the proximity effect when weak lines are included in the analysis .
liske & williger ( 2001 ) introduce a method for extracting @xmath3 from qso spectra based on flux statistics .
we shall return to this topic in future work .
as the results listed in table [ table - jnu ] indicate , the inclusion of the four blazars and one bl lac object , all at @xmath8 , in our sample does not change the result significantly . however , there is much observational evidence that radio loud and radio quiet quasars inhabit different environments , namely that radio loud quasars reside in rich clusters while radio quiet quasars exist in galaxy environments consistent with the field ( stockton 1982 , yee & green 1984 , 1987 , yee 1987 , yates , miller , & peacock 1989 , ellingson , yee , & green 1991 , yee & ellingson 1993 , wold et al .
2000 , smith , boyle , & maddox 2000 ) .
if there is a corresponding increase in the number of ly-@xmath1 absorption lines in the spectra of radio loud objects , this could cause the proximity effect to be suppressed , and the measured log[@xmath3 ] to be artificially large .
we have therefore divided our sample into radio loud and radio quiet subsamples using the ratio of radio to uv flux to characterize the radio loudness , @xmath138/log[s(1450 } \ ; \mbox{\aa})].\ ] ] the value of rl for each object in our sample is listed in table [ tab - flux ] .
a histogram of these values and the distribution of rl with @xmath139 for the sample objects are shown in figure [ fig : rl ] .
the division between radio loud and radio quiet was chosen to be rl=1.0 .
the resulting values of log[@xmath3 ] for these subsamples are listed in table [ table - jnu ] .
there is no significant trend for log[@xmath3 ] to appear larger for radio loud objects than for radio quiet objects .
we performed the maximum likelihood calculation for the case of a non - zero cosmological constant .
this means that the observer - qso and absorber - qso luminosity distances that appear in the relationship between @xmath140 and @xmath139 ( bdo ) must be calculated numerically from the expression : @xmath141 where @xmath142 ( peebles , 1993 ) as this integral can not be reduced to an analytical form for @xmath143 .
the calculations in the sections above assume ( @xmath144,@xmath28 ) = ( 1.0,0.0 ) . here
, we perform the maximum likelihood search for @xmath79 ) using ( @xmath144,@xmath28 ) = ( 0.3,0.7 ) . for a qso at @xmath145 with a lyman limit flux density of 0.1 @xmath146jy , an absorber at @xmath147 , and an assumed background of log[@xmath3]@xmath102 .
, this ( @xmath144,@xmath28 ) results in a value of @xmath140 that is @xmath148% smaller than that inferred in the @xmath16 case . unlike all the other solutions performed
, we ignore redshift path associated with metal lines and use all redshifts between @xmath149 and @xmath150 .
this does not change the results significantly , but cuts down the computation time substantially .
the results are listed in table [ table - jnu ] and are plotted in figure [ fig : lowzcomp ] . for comparison , we also give the solutions for @xmath3 found using the standard parameters , ( @xmath144,@xmath28 ) = ( 1.0,0.0 ) , with this redshift path neglected .
we find that ( @xmath144,@xmath28 ) = ( 0.3,0.7 ) , does not change the value of @xmath3 derived significantly from the value found using ( @xmath144,@xmath28 ) = ( 1.0,0.0 ) .
we performed a slightly modified re - analysis of the scott et al .
( 2000b ) sample of objects at @xmath17 and found little effect at high redshift as well .
the solution found for ( @xmath144,@xmath28 ) = ( 1.0,0.0 ) was log[@xmath3]@xmath151 , while for ( @xmath144,@xmath28 ) = ( 0.3,0.7 ) , we find log[@xmath3]@xmath152 for these data . in the case of a size distribution of ly-@xmath1 absorbers that is constant in redshift , the evolution of the number of ly-@xmath1 absorption lines per unit redshift is given by : @xmath154^{-0.5 } , \label{equ : noevol}\ ] ] ( sargent et al . 1980 ) where @xmath155 equals the absorber cross section times the absorber comoving number density times the hubble distance , @xmath156 .
a plot of @xmath153 versus @xmath139 for non - evolving ly-@xmath1 absorbers in ( @xmath157,@xmath28 ) = ( 1.0,0.0 ) and ( 0.3,0.7 ) cosmologies is shown in figure [ fig : noevol ] .
it is clear that non - evolving models are too shallow to fit points at @xmath20 , so the normalization is found from a fit to the fos data .
the fos data at @xmath158 are consistent with a non - evolving population for @xmath159 .
the data are less consistent with a non - evolving concordance model in which @xmath160 , though not significantly so .
the number density evolution of ly-@xmath1 absorbers over the redshift range @xmath161 can not be approximated with a single power law .
there is a significant break in the slope of the line number density with respect to redshift , near @xmath26 ( weymann et al .
1998 , paper iv ) though kim , cristiani , & dodorico ( 2001 ) argue that the break occurs at @xmath162 .
dav et al .
( 1999 ) show from hydrodynamical simulations of the low redshift ly-@xmath1 forest , that the evolution of the line density is sensitive mainly to the hi photoionization rate , but also to the evolution of structure ( cf .
their figure 7 ) .
the flattening of @xmath153 observed by weymann et al .
( 1998 ) is mostly attributed to a dramatic decline in @xmath163 with decreasing @xmath139 .
dav et al .
( 1999 ) derive an expression for the density of ly-@xmath1 forest lines per unit redshift as a function of the hi photoionization rate : @xmath164^{\beta-1 } h^{-1}(z ) , \label{equ : dave}\ ] ] where @xmath165 is the normalization at some fiducial redshift which we choose to be @xmath90 and @xmath166 can be expressed by equ .
[ equ : hmgam ] .
we fit the fos and mmt absorption line data , binned in @xmath153 as presented in paper iv and scott et al .
( 2000a , paper i ) , to this function in order to derive the parameters describing @xmath166 implied by the evolution in ly-@xmath1 forest line density .
we observe flattening of @xmath153 at @xmath158 , but not to the degree seen by weymann et al .
( 1998 ) in the key project data .
as described in paper iv , we find @xmath167 , for lines above a 0.24 threshold , while weymann et al .
( 1998 ) measure @xmath168 .
see paper iv for more discussion of the significance and underlying causes of this difference .
we find @xmath169 and @xmath170 for @xmath171 and lines with rest equivalent widths above 0.24 and 0.32 respectively .
these fits to equ .
[ equ : dave ] are shown in figure [ fig : dndz](a ) . in panel
( b ) , we plot @xmath166 , as expressed in equ . [
equ : hmgam ] , evaluated using the parameters found from the fit to equ .
[ equ : dave ] above .
the hm96 solution and the solution derived from the full fos and mmt data sets are represented by the thick and thin solid lines respectively .
the small values of @xmath121 derived from @xmath153 above translate into ionization rates that do not decrease dramatically with decreasing redshift and result from the less pronounced flattening of @xmath153 relative to the key project .
these fits are particularly insensitive to the normalization , @xmath119 , so the errors on this parameter are large .
these fits should therefore not be interpreted as measurements of @xmath166 as reliable as those found directly from the absorption line data .
but we find them instructive nonetheless .
the observed @xmath166 falls short of the ionization rate needed to fully account for the change in the ly-@xmath1 line density with redshift , indicating that if the value of @xmath41 at low redshift is indeed slightly larger than that found by the key project , @xmath153 may still be consistent with a non - evolving population of ly-@xmath1 absorbers in the sense noted above , but the formation of structure in the low redshift universe must play a significant role in determining the character of the ly-@xmath1 forest line density .
kf93 performed a similar measurement with a small subsample of this total sample- the hst quasar absorption line key project data of bahcall et al .
we compare our result to that from sample 2 of kf93 , which was constructed from the bahcall et al .
( 1993 ) data excluding one bal quasar and all heavy element absorption systems . the key project sample has since been supplemented ( bahcall et al .
1996 , jannuzi et al . 1998 ) and those data have been included when appropriate in the complete archival sample of fos spectra presented in paper iii .
the mean intensity kf93 derive from their sample 2 ( @xmath172 km s@xmath5 , @xmath32=1.48 , @xmath41=0.21 ) is @xmath173 ergs s@xmath5 @xmath6 hz@xmath5 sr@xmath5 .
this result is lower than ours for @xmath8 by a factor of @xmath174 , though the errors are large on both results are large enough that they are consistent .
we use 162 lines in our low redshift solution for @xmath3 , 65 more than kf93 .
several authors have examined the sharp cutoffs observed in the hi disks of galaxies in the context of using these signatures to infer the local ionizing background ( maloney 1993 , corbelli & salpeter 1993 , dove & shull 1994 ) .
the truncations are modeled as arising primarily from photoionization of the disk gas by the local extragalactic background radiation field .
using 21 cm observations ( corbelli , scheider , & salpeter 1989 , van gorkom 1993 ) to constrain these models , limits on the local ionizing background are placed at @xmath175 @xmath6 s@xmath5 , where @xmath176 and where @xmath177 for an isotropic radiation field .
additionally , narrow - band and fabry - perot observations of h@xmath1 emission from intergalactic clouds ( stocke et al .
1991 , bland - hawthorn et al .
1994 , vogel et al .
1995 , donahue , aldering , & stocke 1995 ) place limits of @xmath178 @xmath6 s@xmath5 , or @xmath179 ergs s@xmath5 @xmath6 hz@xmath5 sr@xmath5 for @xmath180 , while results from measurements of galactic high velocity clouds ( kutyrev & reynolds 1989 , songaila , bryant , & cowie 1989 , tufte , reynolds , & haffner 1998 ) imply @xmath181 @xmath6 s@xmath5 , though the ionization of high velocity clouds may be contaminated by a galactic stellar contribution .
tumlinson et al . ( 1999 ) have reanalyzed the 3c273/ngc3067 field using the h@xmath1 imaging data from stocke et al .
( 1991 ) as well as new ghrs spectra of 3c273 , in order to model the ionization balance in the absorbing gas in the halo of ngc3067 . from this analysis
, they derive the limits , @xmath182 @xmath6 s@xmath5 , or @xmath183 ergs s@xmath5 @xmath6 hz@xmath5 sr@xmath5 at @xmath184 .
weymann et al .
( 2001 ) have recently reported an upper limit of @xmath185 @xmath6 s@xmath5 , or @xmath186 ergs s@xmath5 @xmath6 hz@xmath5 sr@xmath5 from fabry - perot observations of the intergalactic hi cloud , 1225 + 01 , for a face - on disk geometry .
if an inclined disk geometry is assumed , this lower limit becomes @xmath187 ergs s@xmath5 @xmath6 hz@xmath5 sr@xmath5 .
these results are summarized in figure [ fig : allzcomp ] .
it is encouraging that the proximity effect value is consistent with the limits on the background set by these more direct estimates which are possible locally .
haardt & madau ( 1996 ) calculated the spectrum of the uv background as a function of frequency and redshift using a model based on the integrated emission from qsos alone .
the qso luminosity function is drawn from pei ( 1995 ) .
the opacity of the intergalactic medium is computed from the observed redshift and column density distributions of ly-@xmath1 absorbers given by equ .
[ eq : dndzdnh1 ] .
the effects of attenuation and reemission of radiation by hydrogen and helium in ly-@xmath1 absorbers are included in these models .
their result for @xmath126 and @xmath180 at @xmath90 is @xmath188 ergs s@xmath5 @xmath6 hz@xmath5 sr@xmath5 .
fardal et al . (
1998 ) compute opacity models for the intergalactic medium ( igm ) based on high resolution observations of the high redshift ly-@xmath1 forest from several authors .
shull et al .
( 1999 ) extend the models of fardal et al .
( 1998 ) to @xmath90 , treating opacity of low redshift ly-@xmath1 forest from observations made with hst / ghrs ( penton et al .
2000a , b ) and with hst / fos ( weymann et al .
1998 ) . like haardt & madau ( 1996 )
, they also incorporate the observed redshift distribution of lyman limit systems with log(n@xmath189 ) @xmath190 ( stengler - larrea et al . 1995 , storrie - lombardi et al .
their models also allow for a contribution from star formation in galaxies in addition to agn .
the qso luminosity function again is taken to follow the form given by pei ( 1995 ) with upper / lower cutoffs at 0.01/10 l@xmath191 .
qso uv spectral indicies are assumed to equal 0.86 , while the ionizing spectrum at @xmath192 has @xmath180 .
the contribution to the background from stars was normalized to the h@xmath1 luminosity function observed by gallego et al .
( 1995 ) and the escape fraction of photons of all energies from galaxies was taken to be @xmath193 . the full radiative transfer model described in fardal et al .
( 1998 ) was used to calculate the contribution to the mean intensity by agn , but not the contribution from stars , as they were assumed to contribute no flux above 4 ryd , the energies at which the effects of igm reprocessing become important .
these authors find @xmath194 ergs s@xmath5 @xmath6 hz@xmath5 sr@xmath5 at @xmath195 , with approximately equal contributions from agn and stars , a value somewhat lower than our result for @xmath8 , but which is allowed within the errors .
we estimate the contribution to the uv background from star - forming galaxies using the galaxy luminosity function of the canada - france redshift survey ( lilly et al .
1995 ) . at @xmath24
, we derive @xmath196 ergs s@xmath5 @xmath6 hz@xmath5 sr@xmath5 , assuming @xmath197 .
the hm96 models for the qso contribution give @xmath198 ergs s@xmath5 @xmath6 hz@xmath5 sr@xmath5 at @xmath24 . these estimates , and the range of measured @xmath3 in this paper
, @xmath199 ergs s@xmath5 @xmath6 hz@xmath5 sr@xmath5 imply an escape fraction of uv photons from galaxies between 4% and 70% .
the @xmath3 inferred from @xmath153 in section [ sec - dndz ] implies escape fractions well over 100% .
bianchi et al . ( 2001 ) make updated estimates of the mean intensity of the background with contributions from both qsos and star - forming galaxies .
their models incorporate various values of the escape fraction of lyman continuum photons from galaxies which are constant with redshift and wavelength .
our new results at @xmath158 are most consistent with their models of the qso contribution alone , though some contribution from galaxies , ie .
a small f@xmath200 , is allowed within the uncertainties . at @xmath201 ,
recent results from steidel , pettini , & adelberger ( 2001 ) on the lyman - continuum radiation from high redshift galaxies suggest that these sources become a more important component of the uv background at high redshift .
drawing on lessons learned from our work on high redshift objects in paper ii , we have made corrections for quasar systemic redshifts before performing the proximity effect analysis , as discussed in [ sec - zsys ] .
this correction , @xmath202 km s@xmath5 , was made to qso redshifts measured from ly-@xmath1 emission for objects for which no systemic redshift measurement was available .
for the low redshifts considered in this paper , redshifts measured from [ oiii ] , mgii , or balmer emission lines were deemed suitable as qso systemic redshift measurements .
we have removed known gravitational lenses from the sample . as discussed above ,
we perform the proximity effect analysis omitting and including spectra that show associated absorption and damped ly-@xmath1 absorption and determined that neither of these populations significantly biases our results .
because we are working with low redshift data where line densities are low , we expect that blending has not contributed as strong a systematic effect as in the high redshift sample of paper ii .
the curve - of - growth effects discussed in paper ii may still be present , since many lines in the sample have equivalent widths which place them on the flat part of the curve of growth . however , the effects of clustering may be even more important at low redshift than at high redshift .
loeb & eisenstein ( 1995 ) showed how the fact that quasars reside in the dark matter potentials of galaxies and small groups of galaxies can influence the proximity effect signature .
the peculiar velocities of matter clustered in these potentials can result in ly-@xmath1 absorption at redshifts greater than the quasar emission redshift .
we found that including associated absorbers in our sample did not significantly change our results .
recently , pascarelle et al . (
2001 ) report evidence for a lower incidence of ly-@xmath1 absorption lines arising in the gaseous halos of galaxies in the vicinities of qsos than in regions far from qsos .
they argue that galaxy - qso clustering may lead proximity effect measurements to overestimate @xmath3 at @xmath8 by a up to a factor of 20 .
while we agree that most systematic effects in this type of analysis , including clustering , will lead to overestimates of @xmath3 , the agreement between our results and the direct measurements discussed in section [ sec - direct ] give us confidence that our results are not biased by this large a factor .
the hydrodynamic simulations of the low redshift ly-@xmath1 forest of dav et al .
( 1999 ) indicate that , at low redshift , structures of the same column density correspond to larger overdensities and more advanced dynamical states than at high redshift . for a @xmath203 cosmology , an equivalent width limit of 0.32
corresponds to an overdensity of @xmath204 at @xmath22 , while at @xmath205 , this limit corresponds to @xmath206 .
this may have implications on the clustering of ly-@xmath1 absorption lines around qsos and hence on the values of @xmath3 derived from the proximity effect .
it is possible that we are seeing this clustering effect in the variable threshold solution at @xmath9 , in which the two highest @xmath45 lines in the sample are responsible for the inability to isolate a reasonable maximum likelihood @xmath3 .
we have analyzed a set of 151 qsos and 906 ly-@xmath1 absorption lines , the subset of the total data set presented in paper iii that is appropriate for the proximity effect .
the primary results of this paper are as follows : \(2 ) the value of @xmath3 is observed to increase with redshift over the redshift range of the sample data , @xmath7 . dividing the sample at @xmath207 , we find @xmath208 ergs s@xmath5 @xmath6 hz@xmath5 sr@xmath5 , at low redshift and @xmath209 ergs s@xmath5 @xmath6 hz@xmath5 sr@xmath5 at high redshift .
\(3 ) the inclusion of blazars at @xmath8 has no significant effect on the result .
there is no significant difference between the values of @xmath3 derived from radio loud ( rl @xmath58 1.0 ) and radio quiet ( rl @xmath59 1.0 ) objects , indicating that the observed richness of quasar environments does not distinctly bias the proximity effect analysis .
\(4 ) using information measured and gathered from the literature on each qso s uv spectral index and solving for the hi ionization rate , yields @xmath13 s@xmath5 for @xmath8 and @xmath14 s@xmath5 for and @xmath9 . solving directly for the parameters @xmath123 in the hm96 parametrization of @xmath163 using the hst / fos data presented by bechtold et al .
( 2001 ) combined with the high redshift , ground - based data presented by scott et al .
( 2000a , b ) results in @xmath123 = ( @xmath210 , 0.35 , 2.07 , 1.77 ) for @xmath65 and @xmath211 , 1.45 , 2.13 , 1.42 ) for @xmath212 for @xmath213 .
\(5 ) allowing for a varying equivalent width threshold across each qso spectrum results in consistently higher values of @xmath3 than are found from the constant threshold treatments . at @xmath9 ,
the variable threshold solution is not well - constrained .
jackknife experiments indicate that this is due the objects 0743 - 6719 and 0302 - 2223 , namely the highest @xmath45 absorption lines in each of their spectra .
\(7 ) the @xmath8 result is in agreement with the range of values of the mean intensity of the hydrogen - ionizing background allowed by a variety of local estimates , including h@xmath1 imaging and modeling of galaxy hi disk truncations . to within the uncertainty in the measurement
, this result agrees with the one previous proximity effect measurement of the low redshift uv background ( kf93 ) .
these results are consistent with calculated models based upon the integrated emission from qsos alone ( hm96 ) and with models which include both qsos and starburst galaxies ( shull et al .
the uncertainties do not make a distinction between these two models possible .
\(8 ) the results presented here tentatively confirm the igm evolution scenario provided by large scale hydrodynamic simulations ( dav et al .
this scenario , which is successful in describing many observed properties of the low redshift igm , is dependent upon an evolving @xmath3 which decreases from @xmath215 to @xmath216 .
however , the low redshift uv background required to match the observations of the evolution of the ly-@xmath1 forest line density is larger than found from the data , indicating that structure formation is playing a role in this evolution as well .
our results and the work of others are summarized in figure [ fig : allzcomp ] .
we find some evidence of evolution in @xmath3 , though it appears that even larger data sets , especially at @xmath8 and/or improved proximity effect ionization models will be required to improve the significance .
the authors thank the anonymous referee for a careful review of the paper and for helpful suggestions .
this research has made use of 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 .
this project was supported by stsci grant no.ar-05785.02-94a and stsci grant no .
go 066060195a .
js acknowledges support of the national science foundation graduate research fellowship and the zonta foundation amelia earhart fellowship .
js , jb , and mm received financial support from nsf grant ast-9617060b .
ad acknowledges support from nasa contract no.nas8-39073 ( cxc ) .
vpk acknowledges partial support from an award from the william f. lucas foundation and the san diego astronomers association .
aldcroft , t. l. , bechtold , j. , & elvis , m. 1994 , , 93 , 1 appenzeller , i. , krautter , j. , mandel , h. , bowyer , s. , dixon , w. v. , hurwitz , m. , barnstedt , j. , grewing , m. , kappelmann , n. , & krmer , g. 1998 , , 500 , l9 babu , g. j. & feigelson , e. d. 1996 , astrostatistics , ( london : chapman & hall ) bahcall , j. n. , jannuzi , b. t. , schneider , d. p. , hartig , g. f. , bohlin , r. , junkkarinen , v. 1991 , , 377 , l5 bahcall , j. n. , bergeron , j. , boksenberg , a. , hartig , g. f. , jannuzi , b. t. , kirhakos , s. , sargent , w. l. w. , savage , b. d. , schneider , d. p. , turnshek , d. a. , weymann , r. j. , & wolfe , a. m. 1993 , , 87 , 1 bahcall , j. n. , bergeron , j. , boksenberg , a. , hartig , g. f. , jannuzi , b. t. , kirhakos , s. , sargent , w. l. w. , savage , b. d. , schneider , d. p. , turnshek , d. a. , weymann , r. j. , & wolfe , a. m. 1996 , , 457 , 19 bajtlik , s. , duncan , r. c. , & ostriker , j. p. 1988
, , 327 , 570 ( bdo ) barthel , p. d. , tytler , d. , & thompson , b. 1990 , , 82 , 339 basu , d. 1994 , ap&ss , 222 , 91 bechtold , j. 1994 , , 91 , 1 ( b94 ) bechtold , j. , dobrzycki , a. , wilden , b. , morita , m. , scott , j. , dobrzycka , d. , & tran , k.- v. , aldcroft , t. l. 2001 , , in press ( paper iii , astro - ph/0111486 ) bi , h. & davidsen , a. 1997 , , 479 , 523 bianchi , s. , cristiani , s. , & kim , t. -s .
2001 , , 376 , 1 bland - hawthorn , j. , taylor , k. , veilleux , s. , & shopbell , p. l. 1994 , , 437 , l95 bolton , j. g. , peterson , b. a. , wills , b. j. , wills , d. 1976 , , 210 , l1 browne , i. w. a. , savage , a. , & bolton , j. g. 1975 , , 173 , 87p bryan , g. l. , machacek , m. , anninos , p. , & norman , m. l. 1999 , , 517 , 13 bunker , a. j. , marleau , f. r. , & graham , j. r. 1998 , , 116 , 2086 burbidge , e. m. & kinman , t. d. 1966 , , 145 , 654 burstein , d. & heiles , c. 1982 , , 87 , 1165 buson , l. m. & ulrich , m .- h .
1990 , a&a 240 , 247 carswell , r. f. , webb , j. k. , baldwin , j. a. , & atwood , b. 1987 , , 319 , 709 cen , r. , miralda - escud , j. , ostriker , j. p. , & rauch , m. 1994 , , 437 , l9 chen , h.- w. , lanzetta , k. m. , webb , j. k. , & barcons , x. 1998 , , 498 , 77 cheng , f. h. , you , j. h. , & yan , m. 1990 , , 358 , 18 cooke , a. j. , espey , b. , & carswell , r. f. 1997 , , 284 , 552 corbelli , e. , schneider , s. e , & salpeter , e. e. 1989 , , 97 , 390 corbelli , e. & salpeter , e. e. 1993 , 419 , 104 corbin , m. r. 1997 , , 113 , 245 corbin , m. r. & boroson , t. a. 1996 , , 107 , 69 cristiani , s. & koehler , b. 1987 , a&as , 68 , 339 cristiani , s. , dodorico , s. , fontana , a. , giallongo , e. , & savaglio , s. 1995 , , 273 , 1016 dav , r. , hernquist , l. , katz , n. , & weinberg , m. 1999 , , 511 , 521 dobrzycki , a. , bechtold , j. , scott , j. , & morita , m. 2001 , , in press ( paper iv , astro - ph/0111487 ) donahue , m. , aldering , g. , & stocke , j. t. 1995 , , 450 , l45 dove , j. b. & shull , j. m. 1994 , , 423 , 196 efron , b. 1982 , the jackknife , the bootstrap , and other resampling plans , ( philadelphia : society for industrial and applied mathematics ) ellingson , e. , yee , h. k. c. , & green , r. f. 1991 , , 371 , 49 fabricant , d. , cheimets , p. , caldwell , n. , & geary , j. 1998 , , 110 79 falomo , r. , pesce , j. e. , & treves , a. 1993 , , 411 , l63 fardal , m. , giroux , m. , & shull , j. m. 1998 , , 115 , 2206 fernndez - soto , a. , barcons , x. , carballo , r. , & webb , j. k. 1995 , , 277 , 235 gallego , j. m. , zamorano , j. , aragn - salamanca , a. , & rego , m. 1995 , , 455 , l1 gaskell , c. m. 1982 , , 263 , 79 giallongo , e. , cristiani , s. , fontana , a. , & trvese , d. 1993 , , 417 , 137 giallongo , e. , cristiani , s. , dodorico , s. , fontana , a. , & savaglio , s. 1996 , , 466 , 46 green , r. f. , schmidt , m. , & liebert , j. 1986 , , 61 , 305 green , p. j. 1996 , , 467 , 61 haardt , f. & madau , p. 1996
, , 461 , 20 ( hm96 ) hamann , f. , zuo , l. , & tytler , d. 1995 , , 444 , l69 hernquist , l. , katz , n. , weinberg , d. , miralda - escud , j. 1996 , , 457 , l51 hewitt , a. & burbidge , g. 1987 , , 63 , 1 hu , e. m. , kim , t. -s . , cowie , l. l. , & songaila , a. 1995 , , 110 , 2553 impey , c. d. , petry , c. e. , & flint , k. p. 1999 , , 524 , 5361 jannuzi , b. t. , bahcall , j. n. , bergeron , j. , boksenberg , a. , hartig , g. f. , kirhakos , s. , sargent , w. l. w. , savage , b. d. , schneider , d. p. , turnshek , d. a. , weymann , r. j. , & wolfe , a. m. 1998 , , 118 , 1 jauncey , d. l. , wright , a. e. , peterson , b. a. , & condon , j. j. 1978 , , 219 , l1 kim , t.- s. , hu , e. m. , cowie , l. l. , songaila , a. 1997 , , 114 , 1 kim , t.- s. , cristiani , s. , & dodorico , s. 2001 , , 373 , 757 kinman , t. d. & burbidge , e. m. 1967 , , 148 , l59 kinney , a. l. , bohlin , r. c. , blades , j. c. , & york , d. g. 1991 , , 75 , 645 knezek , p. m. & bregman , j. n. 1998 , , 115 , 1737 + kulkarni , v. p. & fall , s. m. 1993 , , 413 , l63 ( kf93 ) kutyrev , a. s. & reynolds , r. j. 1989 , , 344 , l9 lanzetta , k. m. , turnshek , d. a. , sandoval , j. 1993 , , 84 , 109 lanzetta , k. m. , bowen , d. v. , tytler , d. , & webb , j. k. 1995 , , 442 , 538 lanzetta , k. m. , webb , j. k. , & barcons , x. 1996 , , 456 , l17 laor , a. , bahcall , j. n. , jannuzi , b. t. , schneider , d. p. , green , r. f. , & hartig , g. f. 1994 , , 420 , 110 laor , a. , bahcall , j. n. , jannuzi , b. t. , schneider , d. p. , & green , r. f. 1995 , , 99 , 1 lawrence , c. r. , zucker , j. r. , readhead , a. c. s. , unwin , s. c. , pearson , t. j. , xu , w. 1996 , , 107 , 541 lilly , s. j. , tresse , l. , hammer , f. , crampton , d. , & le fvre , o. 1995 , , 455 , 108 liske , j. & williger , g. m. 2001 , , 328 , 653 lockman , f. j. & dickey , j. m. 1995 , adil fl 01l ( ncsa astronomy digital image library ) loeb , a. & eisenstein , d. j. 1995 , , 448 , 17l lu , l. , wolfe , a. m. , & turnshek , d. a. 1991 , , 367 , 19 lu , l. , sargent , w. l. w. , wonble , d. s. , & takada - hidai , m. 1996 , , 472 , 509 lynds , c. r. , hill , s. j. , heere , k. , & stockton , a. 1966 , , 144 , 1244 lynds , r. & wills , d. 1968 , , 153 , l23 lynds , c. r. 1971 , , 164 , l73 maloney , p. 1993
, , 414 , 41 m@xmath31dowell , j. c. , canizares , c. , elvis , m. , lawrence , a. , markoff , s. , mathur , s. , & wilkes , b. 1995 , , 450 , 585 m@xmath31intosh , d. h. , rix , h. -w . , rieke , m. , foltz . c. b. 1999 , , 517l , 73 miralda - escud , j. , cen , r. , ostriker , j. p. & rauch , m. 1996 , , 471 , 582 morris , s. l. & ward , m. j. 1988 , 230 , 639 morris , s. l. , weymann , r. j. , savage , b. d. , & gilliland , r. l. 1991 , , 377 , l21 murdoch , h. s. , hunstead , r. w. , pettini , m. , & blades , j. c. 1986 , , 309 , 19 netzer , h. , kazanas , d. , wills , b. j. , wills , d. , mingsheng , h. , brotherton , m. s. , baldwin , j. a. , ferland , g. j. , & browne , i. w. a. 1994 , , 430 , 191 nishihara , e. , yamashita , t. , yoshida , m. , watanbe , e. , okumura , s .- i . , mori , a. , & iye , m. 1997 , , 488 , l27 ortiz - gil , a. , lanzetta , k. m. , webb , j. k. , barcons , x. , & fernndez - soto , a. 1999 , , 523 , 720 osmer , p. s. , porter , a. c. , & green , r. f. 1994 , , 436 , 678 pascarelle , s. m. , lanzetta , k. m. , chen , h. -w . , & webb , j. k. 2001 , , 560 , 101 peebles , p. j. e. 1993 , principles of physical cosmology , ( princeton : princeton university press ) pei , y. 1995 , , 438 , 623 penton , s. v. , stocke , j. t. , & shull , j. m. 2000a , , 130 , 121 penton , s. v. , shull , j. m. , & stocke , j. t. 2000b , , 544 , 140 press , w. h. , teukolsky , s. a. , vetterling , w. t. , & flannery , b. p. 1992
, numerical recipes in fortran 77 , the art of scientific computing , second edition , cambridge : cambridge university press de robertis , m. 1985 , , 289 , 67 sanduleak , n. & pesch , p .
1984 , , 55 , 517 sargent , w. l. w. , young , p. , boksenberg , a. , & tytler , d. 1980 , , 42 , 41 savaglio , s. , cristiani , s. , dodorico , s. , fontana , a. , giallongo , e. , & molaro , p. 1997
, , 318 , 347 schmidt , m. 1977 , , 217 , 358 scott , j. , bechtold , j. , & dobrzycki , a. 2000a , , 130 , 37 ( paper i ) scott , j. , bechtold , j. , dobrzycki , a. , & kulkarni , v. 2000b , , 130 , 67 ( paper ii ) di serego - alighieri , s. , danziger , i. j. , morganti , r. , & tadhunter , c. n. 1994 , , 269 , 998 shull , j. m. , roberts , d. , giroux , m. l. , penton , s. v. , fardal , m. a. 1999 , , 118 , 1450 smith , h. e. , burbidge , e. m. , baldwin , j. a. , tohline , j. e. , wampler , e. j. , hazard , c. , & murdoch , h. s. 1977 , , 215 , 427 smith , r. j. , boyle , b. j. , & maddox , s. j. 2000 , , 313 , 252 songaila , a. , bryant , w. , & cowie , l. l. 1989 , , 345 , l71 stark , a. a. , gammie , c. f. , wilson , r. w. , bally , j. , linke , r. a. , heiles , c. , & hurwitz , m. 1992 , , 79 , 77 steidel , c. c. & sargent , w. l. w. 1991 , , 382 , 433 steidel , c. c. , pettini , m. , & adelberger , k. l. 2001 , , 546 , 665 stengler - larrea , e. et al .
1995 , , 444 , 64 stocke , j. t. , case , j. , donahue , m. , shull , j. m. , & snow , t. p. 1991
, , 374 , 72 stockton , a. 1982 , , 257 , 33 stockton , a. & mackenty , j. w. 1987 , , 316 , 584 storrie - lombardi , l. j. , m@xmath31mahon , r. g. , irwin , m. j. , & hazard , c. 1994 , , 427 , l13 telfer , r. c. , zheng , w. , kriss , g. a. , & davidsen , a. f. 2001 , , in press ( astro - ph/0109531 ) theuns , t. , leonard , a. , efstathiou , g. 1998 , , 297 , l49 tufte , s. l. , reynolds , r. j. , & haffner , l. m. 1998 , , 504 , 773 tumlinson , j. , giroux , m. l. , shull , j. m. , & stocke , j. t. 1999 , , 118 , 2148 tytler , d. , boksenberg , a. , sargent , w. l. w. , young , p. , & kunth , d. 1987 , , 64 , 667 tytler , d. & fan , x. -m .
1992 , , 79 , 1 ulrich , m.- h. 1976 , , 206 , 364 van gorkom , j. h. 1993 , in the environment and evolution of galaxies , ed .
j. m shull & h. a. thronson ( dordrecht : kluwer ) , 343 vogel , s. n. , weymann , r. , rauch , m. , & hamilton , t. 1995 , , 441 , 162 weymann , r. j. , carswell , r. f. , & smith , m. g. 1981 , , 19 , 41 weymann , r. j. , jannuzi , b. t. , lu , l. , bahcall , j. n. , bergeron , j. , boksenberg , a. , hartig , g. f. , kirhakos , s. , sargent , w. l. w. , savage , b. d. , schneider , d. p. , turnshek , d. a. , & wolfe , a. m. 1998 , , 506 , 1 weymann , r. j. , vogel , s. n. , veilleux , s. , & epps , h. 2001 , , 561 , 559 wilkes , b. j. 1986 , , 218 , 331 williger , g. m. , baldwin , j. a. , carswell , r. f. , cooke , a. j. , hazard , c. , irwin , m. j. , m@xmath31mahon , r. g. , & storrie - lombardi , l. j. 1994 , , 428 , 574 wills , d. & wills , b. j. 1976 , , 31 , 143 wold , m. , lacy , m. , lilje , p. b. , & serjeant , s. 2000 , , 316 , 267 yates , m. g. , miller , l. , & peacock , j. a. 1989 , , 240 , 129 yee , h. k. c. & ellingson , e. 1993 , , 411 , 43 yee , h. k. c. 1987 , , 94 , 1461 yee , h. k. c. & green r. f. 1987 , , 94 , 618 yee , h. k. c. & green r. f. 1984 , , 280 , 79 young , p. , sargent , w. l. w. , & boksenberg , a. 1982 , , 252 , 10 zhang , y. , anninos , p. , & norman , m. l. 1995 , , 453 , l57 zheng , w. & sulentic , j. w. 1990 , , 350 , 512 zheng , w. & malkan , m. a. 1993 , , 415 , 517 zheng , w. , kriss , g. a. , davidsen , a. f. , lee , g. , code , a. d. , bjorkman , k. s. , smith , p. s. , weistrop , d. , malkan , m. a. , baganoff , f. k. , & peterson , b. m. 1995 , , 444 , 632 zheng , w. , kriss , g. a. , telfer , r. c. , grimes , j. p. , & davidsen , a. f. 1997 , , 475 , 469 zotov , n. 1985 , , 295 , 94 llcccccccc 0003 + 1553 & opt.var .
& 0.4497 & 0.4502 & 0.4503 & & ( 1 ) & ( 2 ) & ( 3 ) & + 0003 + 1955 & opt.var .
& 0.0264 & 0.0264 & 0.0261 & & ( 1 ) & ( 1 ) & ( 4 ) & + 0007 + 1041 & opt.var .
& 0.0902 & 0.0890 & 0.089 & 0.0895 & ( 1 ) & ( 1 ) & ( 5 ) & ( 6 ) + 0015 + 1612 & rqq & 0.5492 & & & & ( 1 ) & & & + 0017 + 0209 & liner & 0.3994 & & & & ( 1 ) & & & + 0024 + 2225 & & 1.1081 & 1.1096 & & & ( 1 ) & ( 7 ) & & + 0026 + 1259 & sy1 & 0.1453 & 0.1463 & 0.1452 & 0.1458 & ( 1 ) & ( 1 ) & ( 5 ) & ( 6 ) +
0042 + 1010 & & 0.5854 & 0.583 & 0.586 & 0.584 & ( 1 ) & ( 8 ) & ( 8 ) & ( 8 ) + 0043 + 0354 & bal ? & 0.3803 & & & & ( 1 ) & & & + 0044 + 0303 & sy1?&0.6219 & 0.6222 & & & ( 1 ) & ( 2 ) & & + 0050 + 1225 & compact , sy1 & 0.0594 & & & & ( 1 ) & & & + 0100 + 0205 & opt.var . &
0.3937 & & 0.3936 & & ( 1 ) & & ( 3 ) & + 0102 - 2713 & & 0.7763 & & & & ( 1 ) & & & + 0107 - 1537 & & 0.8574 & & & & ( 1 ) & & & + 0112 - 0142 & & 1.3739 & 1.3727 & & & ( 1 ) & ( 1 ) & & + 0115 + 0242 & opt.var . &
0.6652 & 0.6700 & & & ( 1 ) & ( 9 ) & & + 0117 + 2118 & & 1.4925 & 1.499 & 1.504 & 1.499 & ( 1 ) & ( 10)&(11)&(11 ) + 0121 - 5903 & sy1 & 0.0461 & 0.0462 & 0.044 & & ( 1 ) & ( 1 ) & ( 5 ) & + 0122 - 0021 & opt.var.,lpq & 1.0710 & 1.0895 & & & ( 1 ) & ( 12 ) & & + 0137 + 0116 & opt.var . &
0.2622 & & 0.2631 & 0.2644 & ( 1 ) & & ( 1 ) & ( 1 ) + 0159 - 1147 & opt.var.,sy1 & 0.6683 & 0.6696 & & & ( 1 ) & ( 13 ) & & + 0214 + 1050 & opt.var . & 0.4068 & & 0.407 & & ( 1 ) & & ( 14 ) & + 0232 - 0415 & opt.var . &
1.4391 & 1.4434 & & & ( 1 ) & ( 1 ) & & + 0253 - 0138 & & 0.8756 & & & & ( 1 ) & & & + 0254 - 3327b & opt.var . & 1.916 & & & & ( 15 ) & & & + 0302 - 2223 & dlas & 1.4021 & & & & ( 1 ) & & & + 0333 + 3208 & opt.var.,lpq & 1.2642 & 1.264 & & & ( 1 ) & ( 7 ) & & + 0334 - 3617 & & 1.1085 & & & & ( 1 ) & & & + 0349 - 1438 & & 0.6155 & 0.615 & & 0.6206 & ( 1 ) & ( 16 ) & & ( 1 ) +
0355 - 4820 & & 1.0058 & 1.005 & & & ( 1 ) & ( 2 ) & & + 0403 - 1316 & opt.var.,hpq & 0.5705 & & 0.571&&(1 ) & & ( 14 ) & + 0405 - 1219 & opt.var.,hpq&0.5717 & 0.5730 & 0.573 & 0.5731 & ( 1 ) & ( 16)&(14)&(16 ) + 0414 - 0601 & opt.var . & 0.7739 & 0.773 & 0.774 & & ( 1 ) & ( 2 ) & ( 5 ) & + 0420 - 0127 & blazar , hpq & 0.9122 & 0.9162 & & & ( 1 ) & ( 13 ) & & + 0439 - 4319 & & 0.5932 & & & & ( 1 ) & & & + 0454 - 2203 & dlas , lpq & 0.5327 & 0.5350 & 0.534 & & ( 1 ) & ( 2 ) & ( 14 ) & + 0454 + 0356 & dlas & 1.3413 & 1.3490 & & & ( 1 ) & ( 10 ) & & + 0518 - 4549 & sy1 & 0.0355 & 0.0341 & & 0.0339 & ( 1 ) & ( 1 ) & & ( 17 ) + 0537 - 4406 & bl lac , hpq & 0.8976 & 0.8926 & & & ( 1)&(18 ) & & + 0624 + 6907 & & 0.3663 & 0.3687 & 0.3710 & 0.3698 & ( 1 ) & ( 1 ) & ( 1 ) & ( 1 ) + 0637 - 7513 & sy1 & 0.6522 & 0.6565 & & 0.6570 & ( 1 ) & ( 18 ) & & ( 18 ) + 0710 + 1151 & opt.var . &
0.7712 & & & & ( 1 ) & & & + 0742 + 3150 & sy1 & 0.4589 & 0.462 & 0.461 & 0.4620 & ( 1 ) & ( 19)&(14)&(10 ) + 0743 - 6719 & opt.var . &
1.5109 & 1.5089 & & 1.511 & ( 1 ) & ( 20 ) & & ( 21 ) + 0827 + 2421 & blazar , hpq & 0.9363 & 0.94 & & 0.942 & ( 1 ) & ( 7 ) & & ( 7 ) + 0844 + 3456 & sy1 & 0.0637 & 0.0646 & 0.064 & & ( 1 ) & ( 1 ) & ( 5 ) & + 0848 + 1623 & opt.var . & & 1.9220 & & & & ( 7 ) & & + 0850 + 4400 & & 0.5132 & 0.5142 & & 0.5150 & & ( 1 ) & & ( 1 ) + 0859 - 1403 & blazar & 1.3338 & 1.3381 & & 1.341&(1 ) & ( 13 ) & & ( 21 ) + 0903 + 1658 & opt.var . & 0.4108 & 0.4106 & 0.4114 & & ( 1 ) & ( 22)&(22 ) & + 0907 - 0920 & & 0.630 & & & & & & & + 0916 + 5118 & & 0.5520 & 0.5525 & & 0.5536 & & ( 1 ) & & ( 1 ) + 0923 + 3915 & opt.var.,sy1,lpq & 0.6986 & 0.6990 & & & ( 1 ) & ( 24 ) & & + 0935 + 4141 & & 1.937 & & & & & & & + 0945 + 4053 & lpq & 1.2479 & 1.2506 & & & ( 1 ) & ( 19 ) & & + 0947 + 3940 & sy1 & 0.2057 & & 0.2059 & & ( 1 ) & & ( 25 ) & + 0953 + 4129 & sy1 ? & 0.2331 & & 0.247 & 0.2326 & ( 1 ) & & ( 25)&(25 ) + 0954 + 5537 & blazar , hpq & 0.9005 & 0.9025 & & & ( 1 ) & ( 1 ) & & + 0955 + 3238 & opt.var.,sy1.8 & 0.5281 & & 0.531 & 0.5309 & & & ( 14)&(10 ) + 0958 + 5509 & & 1.7569 & 1.7582 & & & ( 10)&(7 ) & & + 0959 + 6827 & & 0.7663 & 0.7724 & & & ( 1 ) & ( 1 ) & & + 1001 + 0527 & & 0.1589 & 0.1605 & & 0.160 & ( 1 ) & ( 1 ) & & ( 25 ) + 1001 + 2239 & & 0.9766 & & & & ( 1 ) & & & + 1001 + 2910 & agn & 0.3285 & & & 0.3293 & ( 1 ) & & & ( 1 ) + 1007 + 4147 & & 0.6110 & 0.6125 & & & ( 1 ) & ( 13 ) & & + 1008 + 1319 & & 1.3012 & 1.2968 & & & ( 1 ) & ( 1 ) & & + 1010 + 3606 & sy1 & 0.0785 & & 0.079 & & ( 1 ) & & ( 5 ) & + 1026 - 004a & & 1.4349 & & & & ( 1 ) & & & + 1026 - 004b & & 1.5253 & & & & ( 1 ) & & & + 1038 + 0625 & opt.var.,lpq & 1.2667 & 1.272 & & & ( 1 ) & ( 7 ) & & + 1049 - 0035 & sy1 & 0.3580 & 0.360 & & 0.3605 & ( 1 ) & ( 5 ) & & ( 10 ) + 1055 + 2007 & opt.var . &
1.1136 & 1.1165 & & & ( 1 ) & ( 13 ) & & + 1100 + 7715 & opt.var.,agn & 0.3120 & & 0.324 & 0.339 & ( 1 ) & & ( 25)&(25 ) + 1104 + 1644 & opt.var.,sy1 & 0.6294 & & 0.630 & 0.6307 & ( 1 ) & & ( 5 ) & ( 6 ) + 1114 + 4429 & sy1 & 0.1448 & 0.1442 & 0.143 & & ( 1 ) & ( 1 ) & ( 25 ) & + 1115 + 4042 & sy1 & 0.1545 & 0.1552 & & 0.156 & ( 1 ) & ( 1 ) & & ( 25 ) + 1116 + 2135 & e2,sy1 ? & & & 0.1768 & 0.1756 & & & ( 25)&(25 ) + 1118 + 1252 & opt.var . &
0.6823 & & & & ( 1 ) & & & + 1127 - 1432 & blazar , lpq & 1.1824 & 1.2121 & & & ( 1 ) & ( 18 ) & & + 1130 + 1108 & & 0.5065 & & 0.5110 & 0.5104 & ( 1 ) & & ( 1 ) & ( 1 ) + 1136 - 1334 & sy1 & 0.5551 & 0.5571 & & 0.5604 & ( 1 ) & ( 18 ) & & ( 18 ) + 1137 + 6604 & opt.var.,lpq & 0.6449 & 0.6448 & 0.646 & & ( 1 ) & ( 13)&(5 ) & + 1138 + 0204 & & 0.3789 & & 0.3820 & 0.3831 & ( 1 ) & & ( 1 ) & ( 1 ) + 1148 + 5454 & opt.var . &
0.9688 & 0.9777 & & & ( 1 ) & ( 10 ) & & + 1150 + 4947 & opt.var . & 0.3334 & 0.333 & 0.333 & 0.333 & ( 1 ) & ( 26)&(26)&(26 ) + 1156 + 2123 & & 0.3464 & & 0.3475 & 0.3459 & ( 1 ) & & ( 1 ) & ( 1 ) + 1156 + 2931 & blazar , hpq & 0.7225 & 0.7281 & & & ( 1 ) & ( 1 ) & & + 1206 + 4557 & & 1.1596 & 1.164 & & & ( 1 ) & ( 7 ) & & + 1211 + 1419 & rqq , sy1 & 0.0802 & 0.0805 & 0.0807 & 0.0810 & ( 1 ) & ( 1 ) & ( 25)&(25 ) + 1214 + 1804 & & 0.3719 & & & 0.3726 & ( 1 ) & & & ( 1 ) + 1215 + 6423 & & 1.2981 & & & & ( 1 ) & & & + 1216 + 0655 & opt.var . &
0.3312 & 0.3302 & 0.334 & 0.3374 & ( 1 ) & ( 25)&(5 ) & ( 25 ) + 1219 + 0447 & agn & 0.0953 & 0.0931 & & & ( 1 ) & ( 1 ) & & + 1219 + 7535 & sb(r)ab pec , sy1 & 0.0701 & 0.0713 & 0.071&&(1 ) & ( 1 ) & ( 5 ) & + 1226 + 0219 & blazar , sy1,lpq & 0.156 & & 0.157&0.158 & ( 1 ) & & ( 27)&(27 ) + 1229 - 0207 & dlas , blazar , lpq & 1.0406 & 1.0439 & & & ( 1 ) & ( 13 ) & & + 1230 + 0947 & & 0.4176 & & 0.4162 & 0.4153 & ( 1 ) & & ( 1 ) & ( 1 ) + 1241 + 1737 & & 1.2807 & 1.282 & & & ( 1 ) & ( 7 ) & & + 1247 + 2647 & agn & 2.0394 & & & & ( 10 ) & & & + 1248 + 3032 & & 1.0607 & & & & ( 1 ) & & & + 1248 + 3142 & & & 1.029 & & & & ( 28 ) & & + 1248 + 4007 & & 1.0256 & 1.033 & & & ( 1 ) & ( 7 ) & & + 1249 + 2929 & & 0.8205 & & & & ( 1 ) & & & + 1250 + 3122 & & 0.7779 & & & & ( 1 ) & & & + 1252 + 1157 & opt.var . &
0.8701 & & & & ( 1 ) & & & + 1253 - 0531 & bl lac , hpq & 0.5367 & 0.5366 & 0.5356 & 0.536&(1 ) & ( 29)&(29)&(29 ) + 1257 + 3439 & opt.var . &
1.3760 & 1.376 & & & ( 1 ) & ( 7 ) & & + 1258 + 2835 & & 1.3611 & & & & ( 1 ) & & & + 1259 + 5918 & & 0.4679 & 0.4717 & & 0.4853 & ( 1 ) & ( 25 ) & & ( 25 ) + 1302 - 1017 & e4?,opt.var . & 0.2770 & 0.2867 & 0.278 & 0.2868 & ( 1 ) & ( 12)&(5 ) & ( 6 ) + 1305 + 0658 & & 0.6009 & 0.5999 & & & ( 1 ) & ( 1 ) & & + 1309 + 3531 & sab , sy1 & 0.1841 & & 0.184 & 0.183 & ( 1 ) & & ( 25)&(25 ) + 1317 + 2743 & & 1.0082 & 1.016 & & & ( 1 ) & ( 7 ) & & + 1317 + 5203 & blazar & 1.0550 & 1.0555 & & & ( 1)&(7 ) & & + 1318 + 2903 & opt.var . &
0.5469 & & & & ( 1 ) & & & + 1320 + 2925 & & 0.9601 & 0.972 & & & ( 1 ) & ( 7 ) & & + 1322 + 6557 & sy1 & 0.1676 & & & 0.1684 & ( 1 ) & & & ( 25 ) + 1323 + 6530 & & 1.6227 & 1.6233 & & & ( 1 ) & ( 30 ) & & + 1327 - 2040 & & 1.1682 & 1.170 & & & ( 1 ) & ( 18 ) & & + 1328 + 3045 & dlas & 0.8466 & 0.8508 & & & ( 1 ) & ( 13 ) & & + 1329 + 4117 & & 1.9351 & & & & ( 10 ) & & & + 1333 + 1740 & & 0.5464 & 0.5546 & & & ( 1 ) & ( 25 ) & & + 1351 + 3153 & & 1.3170 & 1.3382 & & & ( 1 ) & ( 31 ) & & + 1351 + 6400 & sy1 & 0.0886 & 0.0884 & 0.087 & 0.089 & ( 1 ) & ( 1 ) & ( 25)&(25 ) + 1352 + 0106 & & 1.1200 & & & & ( 1 ) & & & + 1352 + 1819 & sy1 & 0.1508 & 0.1514 & 0.1572 & 0.1538 & ( 1 ) & ( 1 ) & ( 25)&(25 ) + 1354 + 1933 & opt.var . &
0.7190 & 0.718 & 0.719 & & ( 1 ) & ( 7 ) & ( 5 ) & + 1356 + 5806 & & 1.3741 & 1.370&&&(1 ) & ( 7 ) & & + 1401 + 0952 & & 0.4363 & & & & ( 1 ) & & & + 1404 + 2238 & sy & 0.0966 & 0.0978 & & 0.098 & ( 1 ) & ( 1 ) & & ( 25 ) + 1407 + 2632 & & 0.95 & 0.946 & & 0.958 & ( 1 ) & ( 32 ) & & ( 32 ) + 1415 + 4509 & & 0.1145 & 0.1142 & 0.1143 & 0.1139 & ( 1 ) & ( 1 ) & ( 25)&(25 ) + 1416 + 0642 & & 1.4339 & & & 1.442 & ( 1 ) & & & ( 21 ) + 1424 - 1150 & & 0.8033 & 0.8037 & & & ( 1 ) & ( 18 ) & & + 1425 + 2645 & opt.var . &
0.3634 & & & 0.3644 & ( 1 ) & & & ( 10 ) + 1427 + 4800 & sy1 & 0.2215 & & 0.2203 & 0.2246 & ( 1 ) & & ( 25)&(25 ) + 1435 - 0134 & & 1.3099 & & & & ( 1 ) & & & + 1440 + 3539 & compact & 0.0764 & 0.0772 & 0.0777 & 0.0772 & ( 1 ) & ( 1 ) & ( 25)&(25 ) + 1444 + 4047 & e1 ? & 0.2659 & & 0.2672 & 0.267 & ( 1 ) & & ( 3 ) & ( 5 ) + 1512 + 3701 & sy1 ? & 0.3704 & 0.3734 & 0.371 & 0.3715 & ( 1 ) & ( 2 ) & ( 5 ) & ( 6 ) + 1517 + 2356 & & 1.9037 & & & & ( 10 ) & & & + 1517 + 2357 & & 1.834 & & & & & & & + 1521 + 1009 & & 1.3210 & 1.332 & & & ( 1 ) & ( 7 ) & & + 1538 + 4745 & & 0.7704 & 0.7711 & & & ( 1 ) & ( 7 ) & & + 1544 + 4855 & & 0.3985 & & & 0.4010 & ( 1 ) & & & ( 2 ) + 1555 + 3313 & & 0.9402 & 0.9427 & & & ( 1 ) & ( 31 ) & & + 1611 + 3420 & blazar , lpq & 1.3968 & 1.3997 & & & ( 1 ) & ( 33 ) & & + 1618 + 1743 & opt.var . &
0.5549 & 0.5560 & 0.555 & & ( 1 ) & ( 14)&(13 ) & + 1622 + 2352 & opt.var . & 0.9258 & 0.925 & & & ( 1 ) & ( 7 ) & & + 1626 + 5529 & sy1 & 0.1315 & 0.1325 & 0.132 & 0.133 & ( 1 ) & ( 1 ) & ( 25)&(25 ) + 1630 + 3744 & & 1.4712 & 1.478 & 1.474 & 1.478 & ( 1 ) & ( 10)&(11)&(27 ) + 1634 + 7037 & & 1.3338 & 1.338 & 1.336 & 1.342 & ( 1 ) & ( 10)&(11)&(27 ) + 1637 + 5726 & lpq & 0.7499 & 0.750&&0.751 & ( 1 ) & ( 7 ) & & ( 5 ) + 1641 + 3954 & opt.var.,hpq & 0.5946 & 0.5954 & 0.593&&(1 ) & ( 14)&(2 ) & + 1704 + 6048 & opt.var . & 0.3694 & 0.3704 & 0.372 & & ( 1 ) & ( 2 ) & ( 5 ) & + 1715 + 5331 & & 1.9371 & 1.932 & & & ( 10)&(7 ) & & + 1718 + 4807 & & 1.0809 & 1.0828 & & & ( 1 ) & ( 7 ) & & + 1803 + 7827 & bl lac & 0.6840 & & 0.6797 & & ( 1 ) & & & ( 23 ) + 1821 + 6419 & sy1 & 0.2957 & & 0.297 & & ( 1 ) & & ( 5 ) & + 1845 + 7943 & opt.var.,blrg,sy1 & 0.0567 & 0.0548 & & & ( 1 ) & ( 1 ) & & + 2112 + 0556 & & 0.4585 & & & 0.460 & ( 1 ) & & & ( 5 ) + 2128 - 1220 & opt.var.,lpq,sy1 & 0.4988 & 0.5000 & 0.499 & 0.5028 & ( 1 ) & ( 2 ) & ( 14)&(6 ) + 2135 - 1446 & e1,opt.var.,sy1 & 0.2016 & & 0.200 & 0.199 & ( 1 ) & & ( 14)&(34 ) + 2141 + 1730 & opt.var.,lpq,sy1 & 0.2124 & & 0.211 & & ( 1 ) & & ( 14 ) & + 2145 + 0643 & opt.var.,lpq & 0.9997 & 1.000 & & & ( 1 ) & ( 7 ) & & + 2155 - 3027 & opt.var.,bl lac & 0.116 & & & & & & & + 2201 + 3131 & lpq & 0.2953 & 0.2981 & 0.295 & 0.2979 & ( 1 ) & ( 16)&(5 ) & ( 16 ) + 2216 - 0350 & opt.var.,lpq & 0.8997 & 0.900 & & & ( 1)&(7 ) & & + 2223 - 0512 & opt.var.,hpq,bl lac & 1.4037 & & & & ( 1 ) & & & + 2230 + 1128 & blazar , hpq & 1.0367 & 1.0379 & & & ( 1 ) & ( 13 ) & & + 2243 - 1222 & opt.var.,hpq & 0.6257 & 0.6297 & & & ( 1 ) & ( 17 ) & & + 2251 + 1120 & opt.var . & & 0.322 & 0.326 & 0.3255 & & ( 34)&(5 ) & ( 10 ) + 2251 + 1552 & blazar , hpq & 0.8557 & & & & ( 1 ) & & & + 2251 - 1750 & opt.var.,sy1 & 0.0651 & 0.0637 & 0.064 & & ( 1 ) & ( 1 ) & ( 5 ) & + 2300 - 6823 & & 0.5149 & 0.511 & 0.516 & 0.512 & ( 1 ) & ( 35)&(35)&(35 ) + 2340 - 0339 & & 0.8948 & 0.893 & & & ( 1 ) & ( 7 ) & & + 2344 + 0914 & opt.var.,sy1 & 0.6710 & 0.6722 & 0.673 & 0.6731 & ( 1 ) & ( 16 ) & ( 5 ) & ( 16 ) +
2352 - 3414 & opt.var . & 0.7060 & 0.7063 & & & ( 1 ) & ( 2 ) & & + lcccc 0112 - 0142 & 18.0 & 1 & 13dec1996 & 1200 + 0137 + 0116 & 17.1 & 1 & 13dec1996 & 1200 + 0232 - 0415 & 16.4 & 1 & 13dec1996 & 1200 + 0349 - 1438 & 16.2 & 1 & 12dec1996 & 900 + 0414 - 0601 & 15.9 & 1 & 19dec1995 & 400 + 0454 - 2203 & 16.1 & 1 & 19dec1995 & 400 + 0624 + 6907 & 14.2 & 1 & 19dec1995 & 465 + 0827 + 2421 & 17.2 & 3 & 15feb1997 & 1200 + 0850 + 4400 & 16.4 & 1 & 19dec1995 & 300 + 0859 - 1403 & 16.6 & 2a & 12dec1996 & 3600 + 0916 + 5118 & 16.5 & 1 & 19dec1995 & 350 + 0923 + 3915 & 17.9 & 2b & 14jan1996 & 1800 + 0954 + 5537 & 17.7 & 2c & 20apr1996 & 3600 + 0959 + 6827 & 16.4 & 2b & 14jan1996 & 1800 + 1001 + 2910 & 15.5 & 2a & 12dec1996 & 3600 + 1008 + 1319 & 16.2 & 2a & 10dec1996 & 1800 + 1130 + 1108 & 16.9 & 2d & 14jan1996 & 3600 + 1138 + 0204 & 17.6 & 2e & 12dec1996 & 2400 + 1156 + 2123 & 17.5 & 2e & 12dec1996 & 1800 + 1156 + 2931 & 17.0 & 2a & 10dec1996 & 1800 + 1214 + 1804 & 17.5 & 2f & 21apr1996 & 1800 + 1230 + 0947 & 16.1 & 2f & 21apr1996 & 3600 + 1305 + 0658 & 17.0 & 2c & 20apr1996 & 3600 + lcccccrcccc 0003 + 1553 & 3.88 & 0.46 & [email protected] & [email protected] & 1.94 ( 1450 ) & 2.24 & ( 2)&(1b)&(1b)&(1b ) + 0003 + 1955 & 3.99 & 2.04 & [email protected] & [email protected] & 8.43 ( 1450)&-0.44 & ( 3)&(1a)&(1a)&(1a ) + 0007 + 1041 & 5.62 & & [email protected] & [email protected] & 1.47 ( 1450 ) & 0.00 & & ( 1a)&(1a)&(1a ) + 0015 + 1612 & 4.07 & & [email protected] & [email protected] & 0.11 ( 1450 ) & 0.00 & & ( 1b)&(1b)&(1b ) + 0017 + 0209 & 3.05 & & [email protected] & [email protected] & 0.31 ( 1450 ) & 0.00 & & ( 1b)&(1b)&(1b ) + 0024 + 2225&3.60 & & [email protected] & [email protected] & 0.79 ( 1450 ) & 2.40 & & ( 1c)&(1c)&(1c ) + 0026 + 1259 & 4.56 & & [email protected] & [email protected] & 2.22 ( 1450)&-0.04 & & ( 1b)&(1b)&(1b ) + 0042 + 1010 & 5.52 & & [email protected] & [email protected] & 0.09 ( 1450 ) & 2.99 & & ( 1c)&(1c)&(1c ) + 0043 + 0354&3.18 & & [email protected] & [email protected] & 0.97 ( 2093 ) & 0.00 & & ( 1c)&(1c)&(1c ) + 0044 + 0303 & 2.88 & 1.16 & [email protected] & [email protected] & 0.79 ( 1450 ) & 1.94 & ( 2)&(1c)&(1c)&(1c ) + 0050 + 1225&1.46 & & [email protected] & [email protected] & 2.56 ( 1450 ) & 0.06 & & ( 1a)&(1a)&(1a ) + 0100 + 0205 & 2.92 & & [email protected] & [email protected] & 0.45 ( 1450 ) & 0.00 & & ( 1b)&(1b)&(1b ) + 0102 - 2713 & 1.93 & & & 0.18 & 0.29 ( 1285 ) & 0.00 & & ( 4 ) & & ( 1b ) + 0107 - 1537 & 1.73 & & [email protected] & [email protected] & 0.16 ( 1450 ) & 0.00 & & ( 1c)&(1c)&(1c ) + 0112 - 0142&4.32 & & & 0.17 & 0.29 ( 1326 ) & 3.83 & & ( 4 ) & & ( 1c ) + 0115 + 0242&3.32 & & [email protected] & [email protected] & 0.08 ( 1450 ) & 4.08 & & ( 1c)&(1c)&(1c ) + 0117 + 2118&4.75 & 0.39 & [email protected] & [email protected] & 1.88 ( 1307 ) & 0.00 & ( 2)&(1c)&(1c)&(1c ) + 0121 - 5903 & 3.05 & & [email protected] & [email protected] & 2.91 ( 1450 ) & 0.00 & & ( 1a)&(1a)&(1a ) + 0122 - 0021 & 3.57 & & [email protected] & [email protected] & 0.86 ( 1450 ) & 3.13 & & ( 1c)&(1c)&(1c ) + 0137 + 0116 & 3.00 & & [email protected] & [email protected] & 0.07 ( 1450 ) & 3.97 & & ( 1b)&(1b)&(1b ) + 0159 - 1147 & 1.77 & & [email protected] & [email protected] & 1.33 ( 1450 ) & 3.01 & & ( 1c)&(1c)&(1c ) + 0214 + 1050 & 6.96 & & [email protected] & [email protected] & 1.22 ( 1450 ) & 2.57 & & ( 1b)&(1b)&(1b ) + 0232 - 0415 & 2.42 & 0.59 & & & & 2.73 & ( 2 ) & & & + 0253 - 0138&5.61 & & [email protected] & [email protected] & 0.78 ( 1450 ) & 0.00 & & ( 1c)&(1c)&(1c ) + 0254 - 3327b&2.32 & & & 0.28 & & 3.08 & & ( 4 ) & & + 0302 - 2223&1.87&0.31&[email protected] & [email protected] & 0.88 ( 1318 ) & 0.00 & ( 2)&(1c)&(1c)&(1c ) + 0333 + 3208 & 13.5 & & [email protected] & [email protected] & 0.81 ( 1450 ) & 3.38 & & ( 1c)&(1c)&(1c ) + 0334 - 3617&1.40 & & [email protected] & [email protected] & 0.14 ( 1450 ) & & & ( 1c)&(1c)&(1c ) + 0349 - 1438 & 3.83 & & [email protected] & [email protected] & 2.11 ( 1450 ) & 2.53 & & ( 1c)&(1c)&(1c ) + 0355 - 4820&1.16 & 0.39 & [email protected] & [email protected] & 0.70 ( 1450 ) & 2.91 & ( 5)&(1c)&(1c)&(1c ) + 0403 - 1316&3.65 & & [email protected] & [email protected] & 0.39 ( 1450 ) & 4.35 & & ( 1c)&(1c)&(1c ) + 0405 - 1219 & 3.74 & 2.05 & [email protected] & [email protected] & 4.14 ( 1450 ) & 2.68 & ( 2)&(1c)&(1c)&(1c ) + 0414 - 0601 & 5.14 & 0.34 & [email protected] & [email protected] & 0.70 ( 1450 ) & 2.66 & ( 2)&(1c)&(1c)&(1c ) + 0420 - 0127&7.10 & & [email protected] & [email protected] & 0.20 ( 1450 ) & 3.89 & & ( 1c)&(1c)&(1c ) + 0439 - 4319 & 2.30 & & [email protected] & [email protected] & 0.33 ( 1450 ) & 2.95 & & ( 1c)&(1c)&(1c ) + 0454 + 0356&7.39 & 0.38 & [email protected] & [email protected] & 1.26 ( 1336 ) & 2.50 & ( 2)&(1c)&(1c)&(1c ) + 0454 - 2203 & 2.99 & 0.38 & [email protected] & [email protected] & 1.28 ( 1450 ) & 2.77 & & ( 1b)&(1b)&(1b ) + 0518 - 4549 & 4.12 & & [email protected] & [email protected] & 0.13 ( 1450 ) & 5.06 & & ( 1a)&(1a)&(1a ) + 0537 - 4406&4.02 & 0.05 & [email protected] & [email protected] & 0.36 ( 1450 ) & 4.05 & ( 2)&(1c)&(1c)&(1c ) + 0624 + 6907 & 7.01 & & [email protected] & [email protected] & 5.26 ( 1450 ) & 0.00 & & ( 1b)&(1b)&(1b ) + 0637 - 7513 & 9.22 & 0.53 & [email protected] & [email protected] & 0.49 ( 1450 ) & 4.10 & ( 2)&(1c)&(1c)&(1c ) + 0710 + 1151&11.0 & & [email protected] & [email protected] & 1.22 ( 1450 ) & 4.12 & & ( 1c)&(1c)&(1c ) + 0742 + 3150&4.89 & 0.35 & [email protected] & [email protected] & 1.03 ( 1450 ) & 2.96 & ( 2)&(1b)&(1b)&(1b ) + 0743 - 6719 & 11.9 & 0.24 & & & & 3.46 & ( 2 ) & & & + 0827 + 2421&3.51 & & [email protected] & [email protected] & 0.59 ( 1450 ) & 3.17 & & ( 1c)&(1c)&(1c ) + 0844 + 3456&3.31 & & [email protected] & [email protected] & 4.94 ( 2495 ) & 0.00 & & ( 1c)&(1c)&(1c ) + 0848 + 1623&29.7 & & 0.46 & 0.15 & 0.19 ( 1450 ) & 0.00 & & ( 6 ) & & ( 11 ) + 0850 + 4400 & 2.53 & & [email protected] & [email protected] & 0.56 ( 1450 ) & 0.00 & & ( 1b)&(1b)&(1b ) + 0859 - 1403&5.71 & 0.60 & & & & 3.29 & ( 2 ) & & & + 0903 + 1658&3.61 & & [email protected] & [email protected] & 0.17 ( 1450 ) & 2.79 & & ( 1b)&(1b)&(1b ) + 0907 - 0920&4.57 & & [email protected] & [email protected]&0.11 ( 1822 ) & 0.00 & & ( 1c)&(1c)&(1c ) + 0916 + 5118 & 1.40 & & [email protected] & [email protected] & 0.82 ( 1450 ) & 0.00 & & ( 1c)&(1c)&(1c ) + 0923 + 3915&1.53 & & [email protected] & [email protected] & 0.77 ( 1450 ) & 4.83 & & ( 1c)&(1c)&(1c ) + 0935 + 4141&1.32 & & & 0.55 & & 0.00 & & ( 4 ) & & + 0945 + 4053 & 1.44 & & [email protected] & [email protected] & 0.15 ( 1450 ) & 4.07 & & ( 1c)&(1c)&(1c ) + 0947 + 3940 & 1.61 & & [email protected] & [email protected] & 1.25 ( 1450 ) & 0.00 & & ( 1b)&(1b)&(1b ) + 0953 + 4129 & 1.28 & & [email protected] & [email protected] & 1.58 ( 1450 ) & 0.10 & & ( 1b)&(1b)&(1b ) + 0954 + 5537&0.94 & & [email protected] & [email protected] & 0.18 ( 1450 ) & 3.51 & & ( 1c)&(1c)&(1c ) + 0955 + 3238&1.62 & 0.38 & [email protected] & [email protected] & 0.87 ( 1774 ) & 2.99 & ( 2)&(1c)&(1c)&(1c ) + 0958 + 5509&0.84 & 0.31 & & & & 0.00 & ( 2 ) & & & + 0959 + 6827 & 3.93 & & [email protected] & [email protected] & 1.10 ( 1720 ) & 1.99 & & ( 1c)&(1c)&(1c ) + 1001 + 0527&2.41 & & [email protected] & [email protected] & 0.55 ( 1450 ) & 0.26 & & ( 1b)&(1b)&(1b ) + 1001 + 2239 & 2.82 & & [email protected] & [email protected] & 0.12 ( 1450 ) & 3.17 & & ( 1c)&(1c)&(1c ) + 1001 + 2910 & 1.93 & & [email protected] & [email protected] & 1.88 ( 1450 ) & 0.00 & & ( 1b)&(1b)&(1b ) + 1007 + 4147 & 1.23 & 0.72 & [email protected] & [email protected] & 1.02 ( 1450 ) & 2.92 & ( 2)&(1c)&(1c)&(1c ) + 1008 + 1319&3.79 & 0.58 & & & & 0.00 & ( 2 ) & & & + 1010 + 3606 & 1.24 & & [email protected] & [email protected] & 1.00 ( 1450 ) & 0.00 & & ( 1a)&(1a)&(1a ) + 1026 - 004a & 4.85 & & & 0.11 & 0.19 ( 1328 ) & 0.00 & & ( 4 ) & & ( 1c ) + 1026 - 004b & 4.85 & & & 0.15 & 0.24 ( 1285 ) & 0.00 & & ( 4 ) & & ( 1c ) + 1038 + 0625&2.81 & & [email protected] & [email protected] & 1.00 ( 1361 ) & 3.09 & & ( 1c)&(1c)&(1c ) + 1049 - 0035&3.87 & 0.35 & [email protected] & [email protected] & 1.07 ( 1450 ) & 0.00 & ( 2)&(1b)&(1b)&(1b ) + 1055 + 2007 & 1.94 & & [email protected] & [email protected] & 0.34 ( 1450 ) & 3.64 & & ( 1c)&(1c)&(1c ) + 1100 + 7715&3.04 & & [email protected] & [email protected] & 1.33 ( 1450 ) & 2.76 & & ( 1b)&(1b)&(1b ) + 1104 + 1644 & 1.55 & & [email protected] & [email protected] & 1.22 ( 1450 ) & 2.66 & & ( 1c)&(1c)&(1c ) + 1114 + 4429&1.80 & & [email protected] & [email protected] & 0.35 ( 1450 ) & 0.00 & & ( 1b)&(1b)&(1b ) + 1115 + 4042&1.86 & & [email protected] & [email protected] & 1.35 ( 1450 ) & 0.00 & & ( 1b)&(1b)&(1b ) + 1116 + 2135 & 1.27 & & [email protected] & [email protected] & 2.87 ( 1450 ) & 0.01 & & ( 1b)&(1b)&(1b ) + 1118 + 1252&2.28 & & [email protected] & [email protected] & 0.14 ( 1450 ) & 2.75 & & ( 1c)&(1c)&(1c ) + 1127 - 1432&4.07 & & [email protected] & [email protected] & 0.49 ( 1450 ) & 4.78 & & ( 1c)&(1c)&(1c ) + 1130 + 1108&3.47 & & [email protected] & [email protected] & 0.62 ( 1450 ) & 0.00 & & ( 1b)&(1b)&(1b ) + 1136 - 1334 & 3.51 & 0.60 & [email protected] & [email protected] & 0.83 ( 1450 ) & 3.36 & ( 2)&(1b)&(1b)&(1b ) + 1137 + 6604&1.00 & 1.05 & [email protected] & [email protected] & 1.17 ( 1450 ) & 2.98 & ( 2)&(1c)&(1c)&(1c ) + 1138 + 0204&2.37 & & [email protected] & [email protected] & 0.35 ( 1450 ) & 0.00 & & ( 1b)&(1b)&(1b ) + 1148 + 5454 & 1.19 & 0.97 & [email protected] & [email protected] & 1.35 ( 1450)&-0.13 & ( 2)&(1c)&(1c)&(1c ) + 1150 + 4947 & 2.01 & & [email protected] & [email protected] & 0.26 ( 1450 ) & 3.44 & & ( 1b)&(1b)&(1b ) + 1156 + 2123 & 2.56 & & [email protected] & [email protected] & 0.49 ( 1450 ) & 2.23 & & ( 1b)&(1b)&(1b ) + 1156 + 2931 & 1.58 & 0.57 & [email protected] & [email protected] & 1.33 ( 1450 ) & 3.04 & ( 2)&(1c)&(1c)&(1c ) + 1206 + 4557 & 1.27 & 0.45 & [email protected] & [email protected] & 1.69 ( 1450 ) & 0.00 & ( 2)&(1c)&(1c)&(1c ) + 1211 + 1419 & 2.70 & & [email protected] & [email protected] & 2.37 ( 1450)&-0.37 & & ( 1a)&(1a)&(1a ) + 1214 + 1804&2.74 & & [email protected] & [email protected] & 0.52 ( 1450 ) & 0.00 & & ( 1b)&(1b)&(1b ) + 1215 + 6423&2.10 & & [email protected] & [email protected] & 0.18 ( 1340 ) & 3.18 & & ( 1c)&(1c)&(1c ) + 1216 + 0655 & 1.57 & & [email protected] & [email protected] & 1.44 ( 1450 ) & 0.44 & & ( 1b)&(1b)&(1b ) + 1216 + 503a&1.87 & & & 0.35 & 0.58 ( 1326 ) & 0.00 & & ( 4 ) & ( 1c)&(1c ) + 1219 + 0447&1.68 & & [email protected] & [email protected]&0.15 ( 2457 ) & 0.00 & & ( 1c)&(1c)&(1c ) + 1219 + 7535&3.13 & & [email protected] & [email protected] & 2.21 ( 1450 ) & 0.45 & & ( 1a)&(1a)&(1a ) + 1226 + 0219&1.81 & 7.40 & [email protected] & [email protected] & 26.9 ( 1330 ) & 4.26 & ( 7)&(1a)&(1a)&(1a ) + 1229 - 0207&2.34 & 0.23 & [email protected] & [email protected] & 0.57 ( 1450 ) & 3.25 & & ( 1c)&(1c)&(1c ) + 1230 + 0947&1.81 & & [email protected] & [email protected] & 0.96 ( 1450 ) & 0.00 & & ( 1b)&(1b)&(1b ) + 1241 + 1737 & 1.81 & 0.25 & & & & 2.16 & ( 2 ) & & & + 1247 + 2647&1.03 & 0.76 & & & & -0.07 & ( 2 ) & & & + 1248 + 3032 & 1.23 & & [email protected] & [email protected] & 0.09 ( 1450 ) & 3.19 & & ( 1c)&(1c)&(1c ) + 1248 + 3142&1.27 & & & 0.26 & & 0.00 & & ( 4 ) & & ( 8 ) + 1248 + 4007 & 1.44 & 0.57 & [email protected] & [email protected] & 0.65 ( 1450 ) & 0.00 & ( 2)&(1c)&(1c)&(1c ) + 1249 + 2929&1.14 & & & 0.22 & & 0.00 & & ( 4 ) & & ( 8 ) + 1250 + 3122 & 1.24 & & & 0.33 & 0.54 ( 1279 ) & 0.00 & & ( 4 ) & ( 1b)&(1b ) + 1252 + 1157 & 2.34 & & [email protected] & [email protected] & 0.54 ( 1450 ) & 3.12 & & ( 1c)&(1c)&(1c ) + 1253 - 0531&2.12 & 1.43 & [email protected] & [email protected] & 0.30 ( 1450 ) & 4.47 & ( 2)&(1c)&(1c)&(1c ) + 1257 + 3439&1.13 & & & 0.51 & 0.94 ( 1450 ) & 1.14 & & ( 4 ) & & ( 9 ) + 1258 + 2835&0.93 & & [email protected] & [email protected] & 0.34 ( 1331 ) & 0.00 & & ( 1c)&(1c)&(1c ) + 1259 + 5918 & 1.37 & 1.02 & [email protected] & [email protected] & 1.63 ( 1450 ) & 0.00 & ( 2)&(1b)&(1b)&(1b ) + 1302 - 1017 & 3.37 & 0.99 & [email protected] & [email protected] & 3.47 ( 1450 ) & 2.34 & ( 2)&(1b)&(1b)&(1b ) + 1305 + 0658 & 2.16 & & [email protected] & [email protected] & 0.23 ( 1450 ) & 3.13 & & ( 1c)&(1c)&(1c ) + 1309 + 3531&2.55 & & [email protected] & [email protected] & 1.12 ( 1450 ) & 1.58 & & ( 1b)&(1b)&(1b ) + 1317 + 2743 & 1.18 & 0.73 & [email protected] & [email protected] & 1.40 ( 1450 ) & 0.00 & ( 2)&(1c)&(1c)&(1c ) + 1317 + 5203&1.90 & & [email protected] & [email protected] & 0.66 ( 1450 ) & 2.70 & & ( 1c)&(1c)&(1c ) + 1318 + 2903 & 1.14 & 0.26 & [email protected] & [email protected] & 0.56 ( 1450 ) & 0.00 & ( 2)&(1b)&(1b)&(1b ) + 1320 + 2925 & 1.17 & & [email protected] & [email protected] & 0.36 ( 1450 ) & 0.00 & & ( 1c)&(1c)&(1c ) + 1322 + 6557 & 1.92 & & [email protected] & [email protected] & 1.01 ( 1450 ) & 0.00 & & ( 1b)&(1b)&(1b ) + 1323 + 6530&1.99 & & & 0.11 & & 3.02 & & ( 4 ) & & + 1327 - 2040&7.53 & 0.19 & [email protected] & [email protected] & 0.82 ( 1450 ) & 2.62 & ( 2)&(1c)&(1c)&(1c ) + 1328 + 3045&1.16 & & [email protected] & [email protected] & 0.24 ( 1450 ) & 4.49 & & ( 1c)&(1c)&(1c ) + 1329 + 4117&0.97 & 0.95 & & & & 0.00 & ( 2 ) & & & + 1333 + 1740 & 1.75 & 0.51 & [email protected] & [email protected] & 1.01 ( 1450 ) & 1.39 & ( 2)&(1b)&(1b)&(1b ) + 1351 + 3153&1.29 & & [email protected] & [email protected] & 0.11 ( 1319 ) & 2.88 & & ( 1c)&(1c)&(1c ) + 1351 + 6400&2.10 & & [email protected] & [email protected] & 4.36 ( 2531 ) & 1.10 & & ( 1c)&(1c)&(1c ) + 1352 + 0106 & 2.25 & 0.07 & [email protected] & [email protected] & 1.05 ( 1450 ) & 0.00 & ( 2)&(1c)&(1c)&(1c ) + 1352 + 1819 & 2.03 & & [email protected] & [email protected] & 0.71 ( 1450 ) & 0.00 & & ( 1b)&(1b)&(1b ) + 1354 + 1933 & 2.21 & 0.40 & [email protected] & [email protected] & 0.77 ( 1450 ) & 3.53 & ( 2)&(1c)&(1c)&(1c ) + 1356 + 5806&1.40 & & [email protected] & [email protected] & 0.59 ( 1344 ) & 2.34 & & ( 1c)&(1c)&(1c ) + 1401 + 0952&1.96 & & [email protected] & [email protected] & 0.31 ( 1450 ) & 0.72 & & ( 1b)&(1b)&(1b ) + 1404 + 2238&1.99 & & [email protected] & [email protected] & 0.86 ( 2413 ) & 0.29 & & ( 1c)&(1c)&(1c ) + 1407 + 2632 & 1.47 & 0.83 & [email protected] & [email protected] & 1.38 ( 1450 ) & 0.00 & ( 2)&(1c)&(1c)&(1c ) + 1415 + 4509 & 1.13 & & [email protected] & [email protected] & 1.32 ( 1790 ) & 0.00 & & ( 1b)&(1b)&(1b ) + 1416 + 0642&6.24 & & [email protected] & [email protected] & 0.40 ( 1308 ) & 3.67 & & ( 1c)&(1c)&(1c ) + 1424 - 1150 & 7.54 & & [email protected] & [email protected] & 0.83 ( 1450 ) & 2.59 & & ( 1c)&(1c)&(1c ) + 1425 + 2645&2.55 & 0.15 & [email protected] & [email protected] & 0.48 ( 1450 ) & 2.43 & ( 2)&(1b)&(1b)&(1b ) + 1427 + 4800 & 1.88 & & [email protected] & [email protected] & 0.86 ( 1450 ) & 0.00 & & ( 1b)&(1b)&(1b ) + 1435 - 0134 & 3.66 & 0.82 & & & & 0.00 & ( 5 ) & & & + 1440 + 3539&1.00 & & [email protected] & [email protected] & 4.96 ( 1857)&-0.58 & & ( 1b)&(1b)&(1b ) + 1444 + 4047 & 1.27 & 0.89 & [email protected] & [email protected] & 1.59 ( 1450 ) & 0.00 & ( 2)&(1b)&(1b)&(1b ) + 1512 + 3701 & 1.39 & 0.57 & [email protected] & [email protected] & 0.95 ( 1450 ) & 2.75 & ( 2)&(1b)&(1b)&(1b ) + 1517 + 2356&3.91 & & & 0.51 & & 0.00 & & ( 4 ) & & + 1517 + 2357&3.91 & & & 0.08 & & 0.00 & & ( 4 ) & & + 1521 + 1009 & 2.88 & 1.65 & & & & 0.00 & ( 2 ) & & & + 1538 + 4745&1.64 & 0.34 & [email protected] & [email protected] & 1.34 ( 1450 ) & 1.28 & ( 2)&(1c)&(1c)&(1c ) + 1544 + 4855 & 1.60 & 0.10 & [email protected] & [email protected] & 0.95 ( 1450 ) & 0.00 & ( 2)&(1b)&(1b)&(1b ) + 1555 + 3313&2.35 & & [email protected] & [email protected]&0.07 ( 1450 ) & 3.03 & & ( 1c)&(1c)&(1c ) + 1611 + 3420&1.65 & & & 0.18 & 0.30 ( 1322 ) & 4.88 & & ( 4 ) & ( 1c)&(1c ) + 1618 + 1743 & 4.14 & & [email protected] & [email protected] & 1.13 ( 1450 ) & 2.70 & & ( 1b)&(1b)&(1b ) + 1622 + 2352 & 4.46 & & [email protected] & [email protected] & 0.21 ( 1450 ) & 3.54 & & ( 1c)&(1c)&(1c ) + 1626 + 5529&1.83 & & [email protected] & [email protected] & 1.30 ( 1450 ) & 0.00 & & ( 1b)&(1b)&(1b ) + 1630 + 3744 & 1.07 & 0.84 & & & & 0.00 & ( 2 ) & & & + 1634 + 7037 & 4.55 & 1.96 & & & & 0.00 & ( 2 ) & & & + 1637 + 5726&1.90 & & [email protected] & [email protected] & 0.70 ( 1450 ) & 3.98 & & ( 1c)&(1c)&(1c ) + 1641 + 3954&1.02 & 0.61 & [email protected] & [email protected] & 0.67 ( 1450 ) & 3.92 & ( 2)&(1c)&(1c)&(1c ) + 1704 + 6048&2.32 & 0.90 & [email protected] & [email protected] & 1.68 ( 1450 ) & 2.86 & ( 2)&(1b)&(1b)&(1b ) + 1715 + 5331 & 2.71 & & 0.43 & 0.58 & 0.29 ( 1450 ) & 0.53 & & ( 10 ) & & ( 2 ) + 1718 + 4807&2.27 & & [email protected] & [email protected] & 4.09 ( 1450 ) & 1.52 & & ( 1c)&(1c)&(1c ) + 1803 + 7827&3.92 & & [email protected] & [email protected] & 1.16 ( 1450 ) & 3.35 & & ( 1c)&(1c)&(1c ) + 1821 + 6419&3.98 & 1.86 & [email protected] & [email protected] & 8.37 ( 2204 ) & 1.10 & ( 2)&(1c)&(1c)&(1c ) + 1845 + 7943&4.17 & & [email protected] & [email protected] & 0.58 ( 1450 ) & 3.88 & & ( 1a)&(1a)&(1a ) + 2112 + 0556 & 6.48 & 0.29 & [email protected] & [email protected] & 0.67 ( 1450 ) & 0.00 & ( 2)&(1b)&(1b)&(1b ) + 2128 - 1220 & 4.75 & 0.35 & [email protected] & [email protected] & 2.02 ( 1450 ) & 2.99 & ( 11)&(1b)&(1b)&(1b ) + 2135 - 1446&4.71 & & [email protected] & [email protected] & 0.88 ( 1450 ) & 3.17 & & ( 1b)&(1b)&(1b ) + 2141 + 1730&8.20 & & [email protected] & [email protected] & 1.43 ( 1450 ) & 2.84 & & ( 1b)&(1b)&(1b ) + 2145 + 0643 & 4.90 & & [email protected] & [email protected] & 1.14 ( 1450 ) & 3.58 & & ( 1c)&(1c)&(1c ) + 2201 + 3131&9.02 & 0.60 & [email protected] & [email protected] & 4.93 ( 1450 ) & 3.64 & ( 2)&(1b)&(1b)&(1b ) + 2216 - 0350&5.66 & 0.18 & [email protected] & [email protected] & 0.71 ( 1450 ) & 3.43 & ( 2)&(1c)&(1c)&(1c ) + 2223 - 0512&5.47 & 0.16 & & & & 4.35 & ( 2 ) & & & + 2230 + 1128&5.42 & & [email protected] & [email protected] & 0.64 ( 1450 ) & 4.39 & & ( 1c)&(1c)&(1c ) + 2243 - 1222 & 4.94 & & [email protected] & [email protected] & 1.25 ( 1450 ) & 3.32 & & ( 1c)&(1c)&(1c ) + 2251 + 1120&5.08 & & 1.2 & 1.46 & 0.49 ( 1450 ) & 3.06 & & ( 12 ) & & ( 2 ) +
2251 + 1552 & 6.38 & 0.09 & [email protected] & [email protected] & 1.15 ( 1450 ) & 3.94 & ( 2)&(1c)&(1c)&(1c ) + 2251 - 1750&2.77 & & [email protected] & [email protected] & 4.32 ( 2507 ) & 0.07 & & ( 1c)&(1c)&(1c ) + 2300 - 6823 & 3.69 & & [email protected] & [email protected] & 0.22 ( 1450 ) & 3.18 & & ( 1b)&(1b)&(1b ) + 2340 - 0339 & 3.61 & & [email protected] & [email protected] & 1.36 ( 1450 ) & 2.24 & & ( 1c)&(1c)&(1c ) + 2344 + 0914 & 5.76 & 0.34 & & 0.22 & 0.41 ( 1450 ) & 3.52 & ( 2)&(4 ) & ( 9 ) & ( 9 ) + 2352 - 3414 & 1.08 & & [email protected] & [email protected] & 0.77 ( 1450 ) & 2.70 & & ( 1c)&(1c)&(1c ) + llclcccccc 1 & 259 & 0.82 , 13.6 & 1.46 & & bdo & -22.04@xmath217 & 2.13 & 0.95 & 0.80 + 1 & 259 & 0.82 , 6.73 & 1.46 & 35 & ml & -22.11@xmath218 & 1.21 & 0.29 & 0.80 + 1 & 259 & 0.82 , 9.61 & 1.46 & 25 & ml & -22.12@xmath219 & 1.01 & 0.41 & 0.80 + 1 & 259 & 0.82 , 9.31 & 1.45 & 25 & ml & -22.13@xmath220 & 0.78 & 0.58 & 0.80 + 1 & 259 & 0.82 , 11.8 & 1.70 & 30 & ml & -21.74@xmath221 & 1.34 & 0.23 & 0.80 + 1 & 259 & 0.82 , 38.0 & 2.04 & 25 & ml & -21.47@xmath222 & 1.10 & 0.35 & 0.80 + 2 & 289 & 0.15 , 31.3 & 1.46 & & bdo & -22.06@xmath223 & 2.62 & 0.91 & 0.30 + 2 & 289 & 0.15 , 12.0 & 1.46 & 35 & ml & -22.03@xmath224 & 1.32 & 0.24 & 0.30 + 2 & 289 & 0.15 , 13.9 & 1.46 & 25 & ml & -22.04@xmath221 & 1.34 & 0.23 & 0.30 + 2 & 289 & 0.15 , 13.6 & 1.45 & 25 & ml & -22.06@xmath225 & 1.48 & 0.18 & 0.30 + 2 & 289 & 0.15 , 17.6 & 1.70 & 30 & ml & -21.69@xmath226 & 1.47 & 0.18 & 0.30 + 2 & 289 & 0.15 , 31.1 & 2.04 & 25 & ml & -21.42@xmath227 & 0.88 & 0.50 & 0.30 + 1a & 162 & 1.50 , 10.1 & 1.46 & & bdo & -22.87@xmath228 & 1.51 & 0.98 & 0.64 + 1a & 162 & 1.50 , 4.92 & 1.46 & 35 & ml & -22.18@xmath229 & 0.17 & 0.98 & 0.64 + 1a & 162 & 1.50 , 3.67 & 1.46 & 35 & ml & -21.72@xmath230 & 1.02 & 0.40 & 0.62 + 1a & 162 & 1.50 , 3.71 & 1.46 & 35 & ml & -21.88@xmath231 & 0.98 & 0.43 & 0.62 + 1b & 97 & -0.87 , 53.0 & 1.46 & & bdo & -22.02@xmath232 & 2.44 & 0.87 & 0.98 + 1b & 97 & -0.87 , 26.1 & 1.46 & 35 & ml & -21.98@xmath96 & 2.25 & 0.03 & 0.98 + 1b & 97 & -0.87 , 21.5 & 1.46 & 35 & ml & -21.76@xmath233 & 1.31 & 0.24 & 0.95 + 1b & 97 & -0.87 , 21.5 & 1.46 & 35 & ml & -21.95@xmath234 & 1.27 & 0.26 & 0.95 + 3 & 214 & 0.28 , 9.97 & 1.46 & 35 & ml & -21.57@xmath235 & 0.47 & 0.82 & 0.70 + 4 & 208 & 1.04 , 5.76 & 1.46 & 35 & ml & -22.15@xmath236 & 1.47 & 0.19 & 0.65 + 5 & 373 & 0.60 , 7.93 & 1.46 & 35 & ml & -21.74@xmath237 & 0.97 & 0.44 & 0.96 + 6 & 301 & 0.89 , 6.57 & 1.46 & 35 & ml & -22.17@xmath224 & 0.98 & 0.43 & 0.97 + 7 & 415 & 0.67 , 7.72 & 1.46 & 35 & ml & -21.82@xmath238 & 0.93 & 0.46 & 0.98 + 7a & 213 & 0.79 , 7.28 & 1.46 & 35 & ml & -22.22@xmath239 & 0.29 & 0.94 & 0.64 + 7b & 202 & 0.72 , 7.29 & 1.46 & 35 & ml & -21.60@xmath240 & 1.15 & 0.33 & 0.98 + 8 & 422 & 0.69 , 7.64 & 1.46 & 35 & ml & -21.85@xmath241 & 0.82 & 0.55 & 0.97 + 8a & 220 & 0.84 , 7.10 & 1.46 & 35 & ml & -22.23@xmath242 & 0.46 & 0.83 & 0.56 + 8b & 202 & 0.72 , 7.29 & 1.46 & 35 & ml & -21.60@xmath240 & 1.15 & 0.33 & 0.98 + 9 & 906 & 0.61 , 9.26 & 1.46 & 35 & ml & -21.21@xmath243 & 0.55 & 0.76 & 0.91 + 9a & 474 & 0.63 , 9.23 & 1.46 & 35 & ml & -21.79@xmath244 & 0.76 & 0.59 & 0.87 + 9b & 432 & 1.05 , 6.40 & 1.46 & 35 & ml & -20.82@xmath245 & 0.33 & 0.91 & 0.71 + cclccc -23.0 & all & 0.32 & 1.41,7.81 & -22.74@xmath246 & 5.13 + -23.0 & all & variable & 1.15,8.94 & -22.47@xmath247 & 5.73 + -22.0 & all & 0.32 & 1.17,8.27 & -21.32@xmath248 & 13.6 + -22.0 & @xmath249 & 0.32 & 0.95,8.74 & -20.81@xmath250 & 9.53 + -22.0 & @xmath251 & 0.32 & 1.79,5.21 & -21.64@xmath252 & 6.70 + -22.0 & all & variable & 1.48,6.71 & -21.63@xmath253 & 3.21 + -22.0 & @xmath249 & variable & 0.75,10.0 & -21.34@xmath254 & 11.3 + -22.0 & @xmath251 & variable & 1.52,6.09 & -21.63@xmath255 & 1.30 + -21.0 & all & 0.32 & 1.44,7.25 & -20.81@xmath256 & 1.56 + -21.0 & all & variable & 1.13,8.72 & -20.81@xmath257 & 0.73 + @xmath258 & all & 0.32 & 0.99,9.46 & -21.54@xmath259 & 3.35 + @xmath258 & @xmath249 & 0.32 & 0.51,11.1 & -21.80@xmath260 & 4.63 + @xmath258 & @xmath251 & 0.32 & 1.90,4.74 & -21.54@xmath261 & 1.55 + @xmath258 & all & variable & 1.38,7.32 & -21.56@xmath262 & 0.57 + @xmath258 & @xmath249 & variable & 0.84,10.1 & -21.83@xmath263 & 3.28 + @xmath258 & @xmath251 & variable & 2.48,2.67 & -21.37@xmath264 & 0.82 + llcccccc 1 & 0.69,7.65 & 1.46 & 35 & -12.17@xmath265 & 0.49 & 0.81 & -21.56 + 1a & 0.85,7.11 & 1.46 & 35 & -12.70@xmath239 & 0.38 & 0.88 & -22.09 + 1b & 0.72,7.29 & 1.46 & 35 & -11.88@xmath266 & 0.48 & 0.81 & -21.28 + 2 & 0.61,9.27 & 1.46 & 35 & -11.27@xmath267 & 0.78 & 0.58 & -20.67 + 2a & 0.63,9.24 & 1.46 & 35 & -12.23@xmath268 & 1.17 & 0.31 & -21.62 + 2b & 1.05,6.40 & 1.46 & 35 & -9.089@xmath269 & 1.17 & 0.31 & -18.48 + 1 & 0.69 , 7.21 & 1.46 & 35 & -12.67,1.73 & 1.01 & 0.40 & + 2 & 0.61 , 9.04 & 1.46 & 35 & -10.86,3.04 & 0.47 & 0.82 & + | in paper iii of our series @xmath0a uniform analysis of the ly-@xmath1 forest at @xmath2 " , we presented a set of 270 quasar spectra from the archives of the faint object spectrograph ( fos ) on the hubble space telescope ( hst ) .
a total of 151 of these spectra , yielding 906 lines , are suitable for using the proximity effect signature to measure @xmath3 , the mean intensity of the hydrogen - ionizing background radiation field , at low redshift . using a maximum likelihood technique and the best estimates possible for each qso s lyman limit flux and systemic redshift
, we find @xmath4 ergs s@xmath5 @xmath6 hz@xmath5 sr@xmath5 at @xmath7 .
this is in good agreement with the mean intensity expected from models of the background which incorporate only the known quasar population .
when the sample is divided into two subsamples , consisting of lines with @xmath8 and @xmath9 , the values of @xmath3 found are 6.5@xmath10 ergs s@xmath5 @xmath6 hz@xmath5 sr@xmath5 , and 1.0@xmath11 ergs s@xmath5 @xmath6 hz@xmath5 sr@xmath5 , respectively , indicating that the mean intensity of the background is evolving over the redshift range of this data set . relaxing the assumption that the spectral shapes of the sample spectra and the background are identical , the best fit hi photoionization rates are found to be @xmath12 s@xmath5 for all redshifts , and @xmath13 s@xmath5 and @xmath14 s@xmath5 for @xmath8 and @xmath9 , respectively . the inclusion of blazars , associated absorbers , or damped ly-@xmath1 absorbers , or the consideration of a @xmath15 cdm cosmology rather than one in which @xmath16 has no significant effect on the results .
the result obtained using radio loud objects is not significantly different from that found using radio quiet objects .
allowing for a variable equivalent width threshold gives a consistently larger value of @xmath3 than the constant threshold treatment , though this is found to be sensitive to the inclusion of a small number of weak lines near the qso emission redshifts .
this work confirms that the evolution of the number density of ly-@xmath1 lines is driven by a decrease in the ionizing background from @xmath17 to @xmath18 as well as by the formation of structure in the intergalactic medium . |
as of now , most of hadrons can be well described by the classical constituent quark model in the picture of @xmath3 for mesons and @xmath4 for baryons . however , according to the quantum chromodynamics , the exotic states ( such as the multiquark states , molecule states etc . )
are also allowed to exist in our universe .
therefore , searching and explaining these exotic states arouse great interest among researchers .
in the experiments , a series of charmoniumlike and bottomoniumlike states referred to @xmath5 have been observed belle03,lhc13,ba05,belle07,bes13,belle110,bes14,bes111,belle08,belleprd , belle100,lhc14,belle14,ab12,ia12,iab . especially , those charged @xmath6 states are even more exotic since they have a minimal quark content of @xmath7 ( @xmath8 ) or @xmath9 ( @xmath10 ) @xcite .
later , some neutral @xmath6 states ( including @xmath11 , @xmath12 and @xmath13 ) were reported by experiments @xcite , which provide important informations of confirming and understanding the exotic @xmath6 states . on theoretical aspects ,
these exotic states are interpreted as a hadronic molecule , a tetraquark , hadrocharmonium or just a cusp effect .
lx14,ae14,sl14,mn10,mn14,zsl05,zsl10,ln12,hh06,de06,nb06,qw13,eb13,dy13,cf14,fa14,iv12,db11,rdm07,lm14,zl14,wei11,wei12,wei15 , _ et al .
_ moreover , several hidden charm baryons composed by @xmath14 have been predicted and investigated @xcite .
these studies enriched the picture of exotic states .
recently , belle collaboration claimed that a new charged charmoniumlike @xmath15 was observed in the invariant mass spectrum of @xmath16 with a significance of 6.2@xmath17 @xcite .
its mass and width are @xmath18 mev/@xmath19 and @xmath20 mev @xcite , respectively .
meanwhile , the quantum number of @xmath0 was determined to be @xmath21 since other hypotheses with @xmath22 were excluded @xcite . in ref .
zl14 , the calculations show that @xmath2 is a strong candidate of the lowest axial - vector tetraquark state within the framework of the color - magnetic interaction . in refs .
@xcite , using the qcd sum rule approach , the relevant results also support the tetraquark interpretation of @xmath2 . besides
, the @xmath2 was described as a molecule - like state in @xcite .
the above informations indicate that the @xmath2 is an ideal candidate for investigating the nature of exotic charmoniumlike states . as of now , the charmoniumlike @xmath5 states
are only observed in four ways @xcite , i.e. , the @xmath23 annihilation ( @xmath24 or @xmath25 ) , @xmath26 fusion process ( @xmath27 ) , @xmath28 meson decay ( @xmath29 ) and hidden - charm dipion decays of higher charmonia or charmoniumlike states . therefore , searching for the charmoniumlike states through other production process is an important topic , which will be useful in confirming and understanding these exotic @xmath5 states .
for example , ke _ et al . _ suggested to search for the charged @xmath30 by the nucleon - antinucleon scattering @xcite , while the production of neutral @xmath31 and @xmath32 states in @xmath33 reaction were investigated in refs .
@xcite . moreover , in refs .
lxh08,he09,lin13,lin14 , the meson photoproduction process were proposed to be an effective way to search for the charmoniumlike states .
soon after , according to the theoretical predictions obtained in ref .
@xcite , an experiment of searching for the @xmath34 through @xmath35 was carried out by the compass collaboration @xcite .
unfortunately , no signal of exclusive photoproduction of the @xmath34 state and its decay into @xmath36 was found .
thus it is important to discuss whether there are other charmoniumlikes that have a discovery potential through @xmath37 channel .
besides , a more accuracy theoretical prediction is necessary . usually , for the meson photoproduction process , the mesonic reggeized treatment will play important role at high photo energies .
the exchange of dominant meson regge trajectories already used to successfully describe the meson photoproduction in refs .
since a high photon beam energy is required for the production of charmoniumlike states through meson photoproduction process , the reggeized treatment will be necessary to ensure the result accuracy . in this work , within the frame of an effective lagrangian approach and the regge trajectories model
, we systematically study the production of charged @xmath2 by meson photoproduction process in order to provide a reliable theoretical results and shed light on our understanding of the properties and production mechanism of charged @xmath38 state .
this paper is organized as follows .
after an introduction , we present the investigate method and formalism .
the numerical result and discussion are given in sec .
iii . in sec .
iv , we discuss the upper limit of decay width of @xmath39 . finally , this paper ends with a brief conclusion .
since the @xmath2 have a strong coupling with @xmath40 belle14,zl14,wei15 , the photoproduction process @xmath41 may be an ideal reaction channel of searching and studying production of the charged @xmath0 .
moreover , considering the signal of @xmath0 are mainly from the contributions of @xmath42 exchange , while the contributions from @xmath43 and @xmath44 exchange can be negligible process by analyzing the hera data .
besides , in refs .
@xcite , it is found that the contributions of @xmath43 and @xmath45 exchange in the @xmath46 reaction are very small .
thus , in the present work we only consider the contribution from the one pion exchange . here , the @xmath47 stand for the virtual photon . ] , the process as depicted in fig .
1 are regard as the source of signal of @xmath0 . besides , the reaction @xmath48 via pomeron exchange ( as shown in fig.2 )
are also calculated , which is considered to be the background for the @xmath0 photoproduction . to investigate @xmath15 production , an effective lagrangian approach and the regge trajectories model in terms of hadrons will be used in the follows .
reaction ( a ) and @xmath49 reaction ( b ) through @xmath50 exchange .
, title="fig : " ] reaction ( a ) and @xmath49 reaction ( b ) through @xmath50 exchange . , title="fig : " ] fig . 1
show the basic tree level feynman diagram for the production of @xmath15 in @xmath51 reaction via pion exchange . to gauge the contributions of these diagrams , we need to know the effective lagrangian densities for each interaction vertex . for the interaction vertex of @xmath52
, we use the effective pseudoscalar coupling @xcite,@xmath53where @xmath54 and @xmath42 stand for the fields of nucleon and pion meson , while @xmath55 is the pauli matrix .
the coupling constant of the @xmath52 interaction was given in many theoretical works , and we take @xmath56 @xcite . process through the pomeron exchange .
, title="fig : " ] process through the pomeron exchange .
, title="fig : " ] as mentioned above , the spin - parity of @xmath0 has been determined by belle collaboration to be @xmath57 @xcite .
thus the relevant effective lagrangian for the vertex and @xmath58 denote @xmath2 and @xmath59 , respectively .
] of @xmath60 read as @xcite,@xmath61where @xmath6 , and @xmath58 denote the fields of @xmath62 and @xmath63 meson , respectively . with the effective lagrangians above
, the coupling constant @xmath64 can be determined by the partial decay widths @xmath65 , @xmath66 , \end{aligned}\ ] ] with@xmath67where @xmath68 is the k@xmath69llen function with @xmath70 . as of now , no relevant experiment datas about @xmath71 can be available @xcite .
however , in ref .
@xcite , the authors obtained the partial decay width @xmath72 mev with the qcd sum rule approach , which allow us to estimate the lower ( upper ) limit of the decay width of @xmath73 , @xmath74 mev . with @xmath75 mev/@xmath76 and @xmath77 mev @xcite
, we get @xmath78 mev , which correspond to three typical partial decay width @xmath79 mev , respectively . for the interaction vertex of @xmath80 , we need to derive it by the vector meson dominance ( vmd ) mechanism @xcite . in the vmd mechanism for photoproduction
, a real photon can fluctuate into a virtual vector meson , which subsequently scatters off the target proton .
thus within the frame of vmd mechanism , we get the lagrangian of depicting the coupling of the intermediate vector meson @xmath63 with a photon as follows,@xmath81where @xmath82 and @xmath83 are the mass and the decay constant of @xmath63 meson , respectively . with the above equation
, one gets the expression for the @xmath84 decay,@xmath85where @xmath86 indicate the three - momentum of an electron in the rest frame of the @xmath63 meson , while @xmath87 is the electromagnetic fine structure constant .
thus , in the light of the partial decay width of @xmath84 @xcite@xmath88we get the constant @xmath89 . in fig .
2 , we present the feynman diagram for the @xmath90 process through pomeron exchange , which is considered as the main background contributions to @xmath41 process . to depict the pomeron exchange process , the relevant formulas which were used in refs .
lxh08,ad87,ma96 are adopted in this work .
the pomeron - nucleon coupling is described as follow,@xmath91where @xmath92 is the exchanged pomeron momentum squared .
@xmath93 gev@xmath94 stands for the coupling constant between a single pomeron and a light constituent quark . for the vertex of @xmath95 , with an on - shell approximation for keeping the gauge invariance , we have@xmath96with@xmath97where @xmath98 gev@xmath94 is the effective coupling constant between a pomeron and a charm quark within @xmath63 meson , while @xmath99 gev@xmath94 denotes a cutoff parameter in the form factor of pomeron .
after the above preparations , the invariant scattering amplitude @xmath101 for the @xmath102 reaction by exchanging a @xmath42 meson read as,@xmath103 \nonumber \\ & & \times \frac{1}{q_{\pi } ^{2}-m_{\pi } ^{2}}f_{\pi nn}(q_{\pi } ^{2})f_{z\psi \pi } ( q_{\pi } ^{2})\end{aligned}\]]where @xmath104 and @xmath105 are the form factors for the vertices of @xmath52 and @xmath60 , respectively .
we have the following definitions for both form factors,@xmath106and@xmath107where @xmath108 is the cutoff parameter for the @xmath52 vertex . in the next calculations , we take the typical value of @xmath109 gev as used in refs . @xcite
. process through @xmath50 meson or pionic regge trajectory exchange . here , the numerical results ( the blue solid line and the red dashed line ) correspond to the partial decay width @xmath110 mev , while the bands stand for the uncertainties with the variation of @xmath111 from 40.2 to 131.4 mev . ] process through @xmath50 meson or pionic regge trajectory exchange .
the notation of the lines and bands as in fig . 3.,title="fig : " ] process through @xmath50 meson or pionic regge trajectory exchange .
the notation of the lines and bands as in fig . 3.,title="fig : " ] as mentioned above , a higher photon beam energy is required for the production of charmoniumlike states through meson photoproduction process .
thus , to better describe the photoproduction of @xmath0 at high photon energies we introduce a pion reggeized treatment by replacing the feynman propagator @xmath112 with the regge propagator as follows @xcite,@xmath113}\frac{e^{-i\pi \alpha _ { \pi } ( t)}}{\sin [ \pi \alpha _ { \pi } ( t)]},\]]where @xmath114 is the slope of the trajectory and the scale factor @xmath115 is fixed at 1 gev@xmath94 , while @xmath116 and @xmath117 are the mandelstam variables .
in addition , the pionic regge trajectory @xmath118 read as gg11,he14,rj71@xmath119the unpolarized differential cross section for the @xmath0 photoproduction shown in fig .
1(a ) then reads@xmath120where @xmath121 and @xmath122 are the three - momentum of initial photon and final @xmath0 state , while @xmath123 denotes the angle of the outgoing @xmath0 state relative to the photon beam direction in the c.m
the total cross section can be easily obtained by integrating the above equation . in fig .
3 , the total cross section @xmath124 through @xmath42 meson or pionic regge trajectory exchange are presented with @xmath125 gev . since the total cross section is proportional to the partial decay width @xmath126
, thus we note that the cross section changes by a factor of 3 to 4 when the partial width @xmath126 is varied from @xmath127 to @xmath128 mev . besides
, it is found that the total cross section through the reggeized treatment is about five times smaller than that of result through a @xmath42 exchange , which indicate that the reggeized treatment can lead to a lower cross section of the @xmath0 photoproduction at high photon energies .
moreover , we note that the peak position of total cross section was moved to the higher energy point when the reggeized treatment is used in the calculations . fig .
4 are the differential cross section for the @xmath129 process by exchanging the @xmath42 meson or pionic regge trajectory at different energies , respectively . from fig .
4 one can see that , relative to the results related to the @xmath42 exchange , the differential cross section by exchanging the pionic regge trajectory are very sensitive to the @xmath123 angle and gives a considerable contribution at forward angles . with the feynman rules and above lagrangian densities
, we obtain the invariant scattering amplitude @xmath130 for the @xmath131 process through @xmath42 exchange ( as depicted in fig .
1(b ) ) as follows,@xmath132where @xmath133 are the propagators of the @xmath134 taking the breit - wigner form @xcite,@xmath135just as above practice , by replacing the feynman propagator @xmath136 with the regge propagator @xmath137 , we can get the scattering amplitude for the @xmath48 process through the pionic regge trajectory exchange . since the pomeron can mediate the long - range interaction between a confined quark and a nucleon , thus @xmath48 via the pomeron exchange ( as described in fig .
2 ) are the mainly background contribution to the @xmath51 reaction . the invariant scattering amplitudes @xmath138 and @xmath139 for figs .
2(a ) and 2(b ) can be written , respectively , as@xmath140@xmath141with @xmath142where @xmath143 is the pomeron trajectory . here ,
the concrete values @xmath144 and @xmath145 gev@xmath146 are adopted . considering the size of the hadrons , the monopole form factor for the off - shell intermediate nucleon is introduced as in the bonn potential model @xcite:@xmath147where @xmath148 and @xmath149 are the cut - off parameter and four - momentum of the intermediate nucleon , respectively . for the value of @xmath148 ,
we will discuss it in the next section .
it is worth mentioning that the form factor is phenomenological and has a great uncertainty .
thus the dipole form factor is deserved to be discussed and compared with the monopole form . combining the signal terms and background amplitudes , we get the total invariant amplitude@xmath150
thus the total cross section of the @xmath48 reaction could be obtained by integrating the invariant amplitudes in the three body phase space,@xmath151where the three - body phase space is defined as @xcite@xmath152
with the fowl code in the cern program library , the total cross section including both signal and background contributions can be calculated . in these calculations ,
the cutoff parameters @xmath148 related to the pomeron term is a free parameter .
thus we first need to give a constraint on the value of @xmath148 .
fig . 5 ( a ) and fig . 5 ( b )
present the variation of cross section from the background contributions for @xmath90 with monopole and dipole form factor , respectively .
it is obvious that the pomeron exchange contributions with dipole form factor are more sensitive to the values of the cutoff @xmath148 than that of monopole form factor .
thus monopole form factor is adopted in the following calculation . at present , no experiment data is available for the @xmath90 process .
however , we notice that the similar reaction @xmath153 and @xmath154 have been measured by some experiment @xcite , where the measured cross sections of these two process are about 1 nb and 10 nb , respectively . here
, we naively think that the cross section of @xmath155 may be equal to or less than that of @xmath156 , while it greater than that of @xmath154 .
thus we constrain the cutoff to be @xmath157 gev as used in ref .
@xcite , which will be used in our calculations . process with monopole form factor at the different values of the cutoff parameter @xmath148 .
( b ) is same as the ( a ) , but for the case of dipole form factor.,title="fig : " ] process with monopole form factor at the different values of the cutoff parameter @xmath148 .
( b ) is same as the ( a ) , but for the case of dipole form factor.,title="fig : " ] reaction . here ,
@xmath158 and @xmath159 denote the results via the pomeron and @xmath50 exchange , respectively , while @xmath160 is the total cross section of @xmath161 .
the variation of @xmath162 and @xmath160 to @xmath163 with several typical partial width values @xmath164 mev are also presented . ]
reaction . here ,
@xmath158 and @xmath165 denote the results via the pomeron exchange and pionic regge trajectory exchange , respectively , while @xmath160 is the total cross section of @xmath166 .
the variation of @xmath165 and @xmath160 to @xmath167 with several typical partial width values @xmath164 mev are also presented . ]
to better understand that the effects of reggeized treatment on the final results , we calculate the total cross section of the @xmath90 reaction without or with reggeized treatment as presented in fig .
6 and fig . 7 , respectively .
6 show the total cross sections for @xmath168 reaction including both @xmath42 exchange and pomeron exchange contributions by taking @xmath169 gev and @xmath157 gev .
we notice that the line shape of total cross section goes up very rapidly and has a peak around @xmath170 gev . in this energy region , the cross section of signal are larger than that of background when taking the partial width values @xmath171 or @xmath172 mev .
in contrast , fig . 7 present the total cross sections for @xmath131 reaction including both pionic regge trajectory exchange and pomeron exchange contributions by taking @xmath173 gev and @xmath157 gev . it is found that the total cross section shows a peak at center of mass energy @xmath174 gev , and the contributions from signal are driven down when using the reggeized treatment .
we note that the cross section of signal just a little bit higher than that of background at center of mass energy @xmath174 gev even a larger partial decay width value ( @xmath175 mev ) are adopted . to demonstrate the feasibility of searching for the charged charmoniumlike @xmath15 through the @xmath48 reaction , we further give the dalitz plot and invariant mass spectrum for the @xmath48 process .
8 present the dalitz plot and @xmath176 invariant mass spectrum for the @xmath48 process with the reggeized treatment at different center of mass energy , where the numerical results are obtained by taking the partial decay width @xmath177 mev . from dalitz
plot we notice that there exist a vertical band and a horizontal band , which are from the signal and background contributions , respectively .
moreover , one notice that the signal of @xmath0 with @xmath178 is more explicit than that with @xmath179 or 12 gev , which is consistent with the result in fig .
thus we can conclude that the @xmath178 gev is the best energy window for searching for the charged @xmath0 via the @xmath168 process . by analyzing the @xmath176 invariant mass spectrum in fig .
8 , one finds that the number of events of @xmath176 can reach up to 500@xmath1802 gev@xmath94 at @xmath178 gev when taking 50 million collisions of @xmath181 . moreover , we take the center of mass energy @xmath178 gev as one of the inputs to calculate the dalitz plot and @xmath176 invariant mass spectrum related to the smaller partial decay width , which are presented in fig .
9 . from dalitz plot in fig .
9 we notice that there exist an clear vertical band which related to the @xmath0 signal when taking partial decay width @xmath171 mev . since the signal and background contribution do not interfere with each other as shown in dalitz plot , the signal of @xmath0 can also be distinguished from the background .
thus we can expect about 375@xmath1802 gev@xmath94 events for the production of @xmath176 in 50 million collisions of @xmath181 at @xmath178 gev if taking @xmath171 mev , which is enough to meet the requirements of the experiment .
however , we also see that the signal of @xmath0 are submerged in the background and will be difficult to distinguish it from the background if taking @xmath182 mev . for comparison
, we calculate the dalitz plot and @xmath176 invariant mass spectrum for the @xmath48 process without the reggeized treatment at @xmath179 gev , as presented in fig . 10 .
one finds that a vertical band related to the signal of @xmath0 appears in dalitz plot even if the lowest partial decay width ( @xmath183 mev ) is adopted , which is obvious different from that with reggeized treatment .
the @xmath36 mass spectrum presented by the compass collaboration in @xcite , which studied exclusive photoproduction of a @xmath36 state at a nuclear target in the range from 7 gev to 19 gev in the centre - of - mass energy of the photon - nucleon system , does not exhibit any statistically significant structure at about @xmath185 gev .
nevertheless it can be used for estimation of an upper limit for the value @xmath186 .
a sum of two exponential functions for a continuum and a breit - wigner curve for a possible contribution of @xmath187 photoproduction was fitted to the mass spectrum in the range from 3.4 gev to 6.0 gev .
the mass @xmath188 mev and the width @xmath189 mev were used as the fixed parameters . doing this we ignore possible contribution of any other resonances like @xmath190 and their interference with @xmath38 . the @xmath36 mass spectrum with the fitting curve
is shown in fig .
the obtained from the fit possible number of @xmath38 events is @xmath191 .
it can be converted to the upper limit @xmath192 events corresponding to a confidence level of cl = 90% . according to the normalization used in compass this limit corresponds to the result @xmath193
this result can be used for estimation of an upper limit for the partial width @xmath194 of the decay @xmath195 based on the reggeized treatment .
the production cross section , averaged over the @xmath163-range covered by compass , is about @xmath196
so @xmath197assuming @xmath198 mev , we obtain an upper limit @xmath199 mev .
photoproduction of the @xmath0 state could also be tested using the data on the hermes experiment .
it covers the range @xmath200 gev@xmath201 gev @xcite where the difference between production cross sections calculated through pionic regge trajectory exchange and virtual pion exchange is maximal .
state obtained by compass @xcite .
the fitted function is shown as a red solid line .
dashed blue line corresponds to the continuum description . ]
in this work , we study the charged @xmath2 production in @xmath131 reaction with an effective lagrangian approach and the regge trajectories model . since the charmoniumlike @xmath38 was only observed in @xmath28 meson decay process , it is an interesting and important topic to study the charmoniumlike @xmath2 by different processes . through of analysis and comparison ,
our numerical results indicate : * the effect of introducing the reggeized treatment has been to significantly reduce the magnitude of cross section for the @xmath2 photoproduction .
the total cross section for the @xmath202 via pionic regge trajectory exchange is smaller than that of without reggeized treatment and the predictions in refs .
lxh08,he09,lin13,lin14 . *
we finds that the differential cross section for the @xmath203 by exchanging the pionic regge trajectory are very sensitive to the @xmath123 angle and gives a considerable contribution at forward angles , which can be checked by further experiment and may be an effective way to examine the validity of the reggeized treatment . *
the total cross section for the @xmath168 process with reggeized treatment is lower than that of without reggeized treatment .
the calculations indicate that the partial decay width @xmath204 is a key parameter in studying the production of @xmath2 via @xmath181 collision . adopting the partial decay width predicted in ref .
@xcite by assuming that the @xmath38 is a tetraquark state , we find that the signal of @xmath15 can also be distinguished from the background at @xmath178 gev if taking @xmath171 mev , but not for the case of taking @xmath205 mev . in ref .
@xcite , by assuming the @xmath2 as an axial - vector molecule - like state , the partial decay width @xmath206 mev was obtained with qcd sum rule . if the predicted @xmath207 mev in ref .
@xcite is reliable , then the signal of @xmath0 produced in @xmath181 collision will be difficult to be distinguished from background .
thus the experiment of the meson photoproduction of @xmath2 may provide a useful adjunctive information for the confirmation of the inner structure of @xmath2 . * the peak position of total cross section for the @xmath208 process was moved to the higher energy point when adding the reggeized treatment , which means that a higher beam energy is necessary for the meson photoproduction of @xmath38 .
the results show that @xmath209 gev is the best energy window for searching for the @xmath2 via @xmath181 collision .
all these calculations can be checked in the future experiment . * using data on exclusive photoproduction of a @xmath36 state from compass we estimated the upper limit for the value of @xmath210 production cross section multiplied by the branching ratio of the @xmath1 decay to be above @xmath211 pb ( cl=90 % ) . according to the reggeized treatment it corresponds to the upper limit of @xmath212 of about 37 mev , which is coincidence with the prediction of @xmath213 mev by assuming the @xmath2 as a molecule - like state in @xcite . since the reggeized treatment used in this work has been proven to be more precise than the general effective lagrangian approach in the pion and kaon photoproduction @xcite
, our theoretical results may provide a valuable information , both for searching for the @xmath2 via @xmath181 collision or explaining the lack of observation of @xmath2 in experiment .
therefore , the more experiment about the photoproduction of @xmath38 are suggested , which will be important to improve our knowledge of the nature of @xmath2 and the regge theory .
the authors would like to acknowledge the compass collaboration for allowing us to use the data of @xmath36 mass spectrum .
meanwhile , x. y. w. is grateful dr .
qing - yong lin for the valuable discussions and help .
this work is partly supported by the national basic research program ( 973 program grant no .
2014cb845406 ) , the national natural science foundation of china ( grant no .
11175220 ) and the one hundred person project of the chinese academy of science ( grant no .
y101020br0 ) . | in this work , inspired by the observation of charmoniumlike @xmath0 , we study the photoproduction of charged charmoniumlike @xmath0 with an effective lagrangian approach and the regge trajectories model .
the numerical results indicate that the reggeized treatment can lead to a lower total cross section of the @xmath0 photoproduction and the peak position of cross section was moved to the higher energy point when the reggeized treatment was added .
moreover , using the data from the compass experiment and presented theoretical predictions , an upper limit of the decay width of @xmath1 is estimated .
the relevant results not only shed light on the further experiment of searching for the charmoniumlike @xmath2 state via meson photoproduction , but also provide valuable informations for having a better comprehension of the nature of charmoniumlike @xmath2 state . |
precision frequency measurements @xcite of a stabilized laser source are of great importance in the field of metrology @xcite as well as atomic , molecular @xcite and optical physics @xcite . here
we show that weak values @xcite in an optical deflection measurement experiment @xcite can produce frequency shift resolutions down to 129 @xmath0 7 @xmath1 with only 2 mw of continuous wave optical power . by performing a weak measurement of the deflection of an infrared laser source that has passed through a weakly dispersive prism , we are able to measure a change in optical frequency comparable to precision fabry - perot interferometers @xcite .
this technique is relatively simple , requiring only a few common optical components and operating at atmospheric pressure .
additionally , we show that this technique has low noise over a large range of response frequencies , making it desirable for many applications such as doppler anemometry @xcite , tests of the isotropy of light propagation @xcite or laser locking without the use of high finesse fabry - perot interferometers @xcite or atomic lines .
first developed as a way to understand preselected and postselected quantum measurements and how they relate to time - reversal symmetry in quantum mechanics , the _ weak value _
@xmath2 of an operator @xmath3 was introduced in a seminal 1988 paper by aharonov , albert and vaidman ( aav ) @xcite .
the weak value is given by @xmath4 , where @xmath5 are the preselected and postselected states of the system , respectively .
this quantity , which is likened to the expectation value of @xmath3 , can have seemingly strange behavior , particularly in the limit where the preselected and postselected states are nearly orthogonal . while numerous experiments have validated the initial claims of the aav paper @xcite , there are new developments concerning the interpretation of preselected and postselected weak measurements @xcite .
weak values are a result of a so - called weak measurement , _
i.e. , _ a measurement which gains only partial information about the state of a system .
unlike von neumann measurements , a weak measurement disturbs the measured state of the system only minimally .
for this reason , weak measurements have been useful in reconsidering hardy s paradox @xcite as well as making meaningful , sequential measurements of noncommuting observables @xcite .
furthermore , due to the denominator of @xmath2 , there can be a large amplification of the weak value when the preselected and postselected states are nearly orthogonal ; as a result , there have been a number of experimental results published in the field of optical metrology @xcite .
there is also a vast array of results , both theoretical @xcite and experimental @xcite , which have gone a long way to further our understanding of the weak measurement process .
we describe here the frequency amplification experiment shown in fig . 1 by further developing the ideas of ref . .
although the actual experiment uses a classical beam , we choose to characterize the weak value effect one photon at a time ; this is valid , owing to the fact that we consider here a linear system with a coherent laser beam modeled as a linear superposition of fock states @xcite . in this experiment
, a single - mode gaussian beam of frequency @xmath6 and radius @xmath7 passes through an optical isolator , resulting in linearly polarized light .
we assume that the radius is large enough to ignore divergence due to propagation .
light then enters a sagnac interferometer containing a 50:50 beam splitter ( bs ) , a mirror and a prism .
the beam travels clockwise and counter - clockwise through the interferometer , denoted by the system states given by \{@xmath8 } ; we write the photon meter states in the position basis as \{@xmath9 } , where @xmath10 denotes the transverse , _ horizontal _ direction .
initially , the interferometer ( including the prism ) is aligned such that the split photon wave function spatially overlaps ( _ i.e. , _ the photons travel the same path whether by @xmath11 or @xmath12 ) .
after the interferometer is aligned , the photons traversing each path receive a small , constant momentum kick in the _ vertical _ direction ; this vertical kick is controlled by the interferometer mirror and results in a misalignment . due to its spatial asymmetry about the input bs
, this momentum kick creates an overall phase difference @xmath13 between the two paths . by adjusting the interferometer mirror
, we can control the amount of light that exits the interferometer into the dark port . while the amplified signal ultimately depends on the value of @xmath13 , and therefore on the magnitude of the misalignment , the _ signal to noise ratio _ ( snr ) is unaffected ( discussed below ) .
we then let @xmath14 represent the small momentum kick given by the prism to the beam ( after alignment ) in the horizontal @xmath10-direction .
the _ system _ and _ meter _ are entangled via an impulsive interaction hamiltonian @xcite ( resulting in a new state @xmath15 ) such that a measurement of the horizontal position of the photon after it exits the interferometer gives us some information about which path the photon took .
we consider a horizontal deflection that is significantly smaller than the spread of the wave packet we are trying to measure , _
i.e. , _ @xmath16 . in this approximation
, we find that the postselected state of the photons exiting the dark port is given by @xmath17 , \label{state}\ ] ] where the weak value , defined above , is given by @xmath18 for small @xmath13 .
there are two interesting features of eq .
( [ state ] ) .
first , the probability of detecting a photon has been reduced to @xmath19 , and yet the snr of an ensemble of measurements is nearly quantum limited @xcite despite not measuring the vast majority of the light .
second , the weak value ( which can be arbitrarily large in theory ) appears to amplify the momentum kick @xmath14 given by the prism ; the resulting average position is given by @xmath20 , where the angular brackets denote an expectation value .
we can compare this to the standard deflection caused by a prism measured at a distance @xmath21 which is given by @xmath22 , where @xmath23 is the wavenumber of the light . in order to predict the deflections @xmath24 or @xmath25
, we must know the form of @xmath14 . for a prism oriented such that it imparts the minimum deviation on a beam ,
the total angular deviation is given by @xmath26-\gamma$ ] , where @xmath27 is the index of refraction of the material and @xmath28 is the angle at the apex of the prism @xcite .
however , we are only interested in the small , frequency - dependent angular deflection @xmath29^{-2}-[n(\omega)]^2\}^{-1/2}$ ] , where @xmath30 ( @xmath31 ) is the index change in the prism ( angular deflection of the beam ) for a given frequency change of the laser . the small _ momentum kick _ is expressed as @xmath32
. we can then write the amplified deflection as @xmath33^{-2}-[n(\omega)]^2}}. \label{deflection}\ ] ] the frequency - dependent index @xmath27 of fused silica , which was used in this experiment , can be modeled using the sellmeier equation @xcite .
we can therefore calculate the expected @xmath25 using eq .
( [ deflection ] ) .
however , to compute the ultimate sensitivity of this weak value frequency measurement , we must include possible noise sources .
if we consider only shot - noise from the laser , the snr for small @xmath13 is approximated by @xmath34 as shown in ref . , where @xmath35 is the number of photons used in the interferometer .
note that @xmath35 is not the number of photons striking the detector , which is given by @xmath36 . by setting @xmath37 and using modest values for @xmath35 , @xmath7 and @xmath6
, we find that frequency sensitivities well below 1 khz are possible
. however , other sources of noise , such as detector dark current , radiation pressure and environmental perturbations will reduce the sensitivity of the device .
in our experimental setup ( shown in fig . 1 ) , we used a fiber - coupled 780 nm external cavity diode laser with a beam radius of @xmath38 m .
the frequency of the laser was modulated with a 10 hz sine wave using piezo controlled grating feedback .
the frequency control was calibrated using saturation absorption spectroscopy of the hyperfine excited states of the rubidium d2 line@xcite : @xmath39 @xmath40 transitions of rubidium 85 and the @xmath41 @xmath42 transitions of rubidium 87 .
linearly polarized light was divided before the interferometer using a 50:50 bs ( although an imbalanced ratio here would be ideal for practical applications ) .
the light in one port was measured with a photodiode and used to lock the power at 2 mw with an acousto - optic modulator before the fiber .
the interferometer was approximately @xmath43 cm in length ; the mirror used to adjust @xmath13 was approximately 6 cm from the input bs ( measured counter clockwise ) and the prism , made of fused silica , was approximately 5 cm from the input bs ( measured clockwise ) .
although the prism was not symmetrically placed in the interferometer as described in the theory above , the results are the same aside from a global offset in position which can be subtracted off during processing .
the interferometer was first aligned to minimize light in the dark port and then , using the aforementioned mirror , misaligned to allow a small percentage of the light ( @xmath442 - 5% ) into the dark port .
the position of this light was measured using a split detector ( new focus model 2921 ) .
the signal was passed through two 6 db / octave bandpass filters centered at 10 hz and amplified by a factor of about @xmath45 . for fig . 2 , we measured the peak of the deflection in each 100 ms cycle , repeated 25 times
; we computed the average and the standard deviation of this set as we varied the change in the optical frequency .
we find that the amplified deflection is a linear function of oscillating optical frequency given by about 720 @xmath0 11 pm / mhz .
compared to the unamplified deflection of about 9.1 pm / mhz given by the expression for @xmath24 , this gives an amplification factor of 79 @xmath0 1.2 and a computed @xmath46 of 1.3% ; this agrees with the measured @xmath46 of 2 - 5% if we include the extra light present in the signal due to phase front distortions from imperfect optics .
a characteristic noise scan was taken and plotted in fig . 3 with and without frequency modulation .
the signal was passed directly from the split detector into the oscilloscope before performing a fast fourier transform .
data was taken with and without a 7.4 mhz optical frequency modulation to show the noise floor over a large bandwidth .
the noise at higher frequencies was similarly flat .
second , to test the range over which this device could function , we optimized the interferometer at the low - frequency end of the laser s tuning range and obtained a snr of approximately 19 with the 7.4 mhz optical frequency modulation .
we then tuned to the high - frequency end of the laser s tuning range ( @xmath47 ghz ) , without adjusting or recalibrating the interferometer , and obtained a snr of 10 .
in fact , this range can be much larger so long as the weak value condition @xmath16 is satisfied ; for our beam radius and optical frequency , we could in principle measure over a range of 5 thz , or about 10 nm . for our experimental parameters
, we can measure below 1 mhz of frequency change with a snr around 1 , as shown in fig .
it should be noted that , although the time between measurements is a full 100 ms , our filtering limits the laser noise to time scales of about 30 ms . for analysis
, we take this as our integration time in estimating @xmath35 for each measurement .
the resulting sensitivity for our apparatus is 129 @xmath0 7 @xmath48 ; _
e.g. , _ if we had integrated for 1 s instead of 30 ms , this device could measure a 129 khz shift in frequency with a snr of 1 .
the error in frequency comes from the calibration described above . using eq .
( [ snr ] ) , we find that the ideal ultimate sensitivity is approximately 67 @xmath48 .
this implies that this apparatus , operating at atmospheric pressure with modest frequency filtering , is less than a factor of two away from the shot - noise limit in sensitivity .
this is no longer surprising since we now understand the fact that weak value experiments amplify the signal , but not the technical noise @xcite .
with only 2 mw of continuous wave input power , we have measured a frequency shift of 129 @xmath0 7 @xmath48 ; we have shown that the system is stable over our maximum tuning range of 140 ghz without recalibration and is nearly shot - noise - limited . with more optical power , longer integration and a more dispersive element such as a grating or a prism with @xmath49 , the sensitivity of this device can measure frequency shifts lower than 1 khz , although a higher sensitivity comes at the cost of maximum tuning range . compare this to commercially available fabry - perot interferometers , which report typical resolutions down to 5 mhz and free spectral ranges of only 1 - 5 ghz .
more sensitive fabry - perot interferometers exist , yet they require a host of custom equipment to reduce environment noise , including vacuum systems and vibration damping .
moreover , an important advantage of this technique is that a large percentage ( @xmath4490% ) of the light used in the interferometer can then be sent off to another experiment ( as indicated in fig . 1 ) , allowing for real - time frequency information during data collection .
while this device can not compare to the absolute frequency sensitivity of frequency combs @xcite , we believe that this method is a simple solution for high - resolution , relative frequency metrology and will serve as a valuable laser - locking tool .
this research was supported by us army research office grant no .
w911nf-09 - 0 - 01417 , darpa expansion grant no .
n00014 - 08 - 1 - 120 and the university of rochester .
31ifxundefined [ 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 10.1109/19.278596 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1038/416233a [ * * , ( ) ] link:\doibase 10.1103/physreva.78.012503 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.090401 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.60.1351 [ * * , ( ) ] link:\doibase 10.1103/physrevd.40.2112 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.104.240401 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.102.173601 [ * * , ( ) ] @noop * * , ( ) \doibase doi : 10.1016/s1386 - 1425(03)00222 - 1 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1038/nature09250 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.94.220405 [ * * , ( ) ] link:\doibase 10.1103/physreva.82.011802 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1038/463890a [ * * , ( ) ] link:\doibase 10.1103/physrevlett.68.2981 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.102.020404 [ * * , ( ) ] link:\doibase 10.1103/physreva.76.062105 [ * * , ( ) ] link:\doibase 10.1126/science.1152697 [ * * , ( ) ] link:\doibase 10.1103/physreva.80.041803 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physreva.81.033813 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.010405 [ * * , ( ) ] @noop _ _ , ed .
( , ) @noop * * , ( ) _ http://steck.us/alkalidata/ _ ( ) | we demonstrate an experiment which utilizes a sagnac interferometer to measure a change in optical frequency of 129 @xmath0 7 @xmath1 with only 2 mw of continuous wave , single mode input power .
we describe the measurement of a weak value and show how even higher frequency sensitivities may be obtained over a bandwidth of several nanometers .
this technique has many possible applications , such as precision relative frequency measurements and laser locking without the use of atomic lines . |
laser induced desorption of atoms and molecules from solid surfaces is a vital phenomenon to investigate fundamentals such as surface electronic structures and dynamics.@xcite besides , laser desorption is an essential technique for pulsed laser deposition used in thin film growth @xcite and mass spectrometry of protein employing matrix - assisted laser desorption ionization.@xcite when desorption flux is small , the velocity distribution of desorbed atoms is directly governed by the desorption mechanism .
when the desorption flux is large enough , on the other hand , the post - desorption collision between desorbed particles may become significant and modify the velocity distribution in the vicinity of the surface after the desorption .
manifestations of the collision effect in laser induced desorption have been reported both by experiments and simulations as the modifications of the angular and velocity distribution of desorbing atoms and molecules.@xcite cowin _
_ investigated the angular dependence of the translational temperature of d@xmath9 desorbed from tungsten surfaces under a pulsed laser irradiation.@xcite the translational temperature of d@xmath9 desorbed in the surface normal direction was higher than those in oblique directions .
they attributed the variation of the translational temperature to the collision effect .
used the direct monte carlo simulation of desorbing atoms to investigate the collision effect .
they showed that even in the sub - monolayer regime the collision noticeably modifies the final angular , velocity and rotational - energy distributions.@xcite however , there has not been any model that can quantitatively estimate the degree of modification by the post - desorption collision in laser desorption so far .
( surface normal ) direction at the surface and at the end of the knudsen layer .
@xmath10 , @xmath11 and @xmath6 denote the surface temperature , the translational temperature of the desorbed atoms at the end of the knudsen layer and the stream velocity , respectively.[kl ] ] knudsen layer formation theory has been developed in rarefied gas dynamics to model the steady flow of the strong evaporation from the surface.@xcite as shown in fig .
[ kl ] , the initial velocity distribution of thermally desorbed species at the surface is well described by a
half - range maxwell - boltzmann velocity distribution.@xcite in the knudsen layer theory , as a result of intensive post - desorption collisions , the half - range velocity distribution at the surface is thermally equilibrated to a full - range maxwell - boltzmann velocity distribution with a stream velocity at some distance from the surface .
this thermalization layer is defined as the knudsen layer .
the theory analytically predicts for monoatomic gas that the ratio of the translational temperature at the end of knudsen layer @xmath11 to the surface temperature @xmath10 and the mach number of the desorption flow at the end of the knudsen layer become 0.65 and 1.0 , respectively.@xcite kelly and dreyfus discussed that the knudsen layer formation theory may be applicable to the pulsed desorption flow with a large desorption flux .
however , it is not straightforward because the pulsed desorption involves complex time evolution of the density and velocity distributions of desorbed atoms.@xcite sibold and urbassek have shown by means of the monte carlo simulation of the boltzmann equation that the pulsed desorption flow at an intense flux is well characterized by the above values predicted by the knudsen layer formation theory .
although previous experimental studies recognized the collision effect in laser desorption,@xcite the formation of the knudsen layer in laser desorption has been discussed only by theory and simulations.@xcite so far , any experimental confirmation of knudsen layer formation in laser desorption has not been presented . for an experimental verification of the theory , a systematic observation of the translational temperature and stream velocity of the desorption flow as a function of the desorption flux
is strongly required . in the present paper , we investigated the laser induced thermal desorption ( litd ) of xe from an au(001 ) surface by means of the time - of - flight ( tof ) measurement as a function of the wide range of desorption flux by varying the initial xe coverage @xmath0 .
we found that at @xmath0 close to 0 ml , the tof was well analyzed by a maxwell - boltzmann velocity distribution .
hence , the desorption at @xmath0 close to 0 ml is rationalized by the thermal desorption followed by the collision - free flow . at @xmath0 close to monolayer
, we observed that the peak positions of the tof spectra shift towards smaller values . assuring
that the desorption is only thermally activated , we regard this modification of the tof as the manifestation of the collision effect . at larger @xmath0 ,
the peak positions of the tof spectra become constant .
we deduced the mach number @xmath12 of the desorption flow to be 0.96 at large @xmath0 under the assumption that @xmath13 .
the obtained value of @xmath12 and the saturating behavior of u at @xmath7 ml both well coincide with the previously reported values by the knudsen layer theory and simulations .
the facts suggest the formation of the knudsen layer in litd at large @xmath0 .
furthermore , we tentatively estimated the knudsen number of the initial desorption flow . ]
the au specimen and the physisorbed xe layers were prepared in the following manner . an au disk with 001 orientation was cut out from a single - crystal rod and was chemically and mechanically polished .
the au disk was put into an ultra high vacuum ( uhv ) chamber with a base pressure of @xmath14 pa .
the sample surface was cleaned by several cycles of ar@xmath15 sputtering and annealing at 700 k in the uhv by the electron bombardment .
the cleanliness of the specimen was confirmed by the observation of a clear reconstructed ( 5@xmath1620 ) pattern @xcite by low - energy electron diffraction and no contamination in the auger electron spectrum .
the sample temperature was monitored with a chromel - alumel type thermocouple directly spot - welded to the side of the au disk .
the disk was cooled down and kept at 23 k by a closed - cycle he compressor type refrigerator during all litd experiments .
the xe gas was dosed onto the au surface by backfilling the chamber .
the xe coverage @xmath0 was monitored by the litd yield .
the litd yield showed a linear dependence on the xe exposure in the present experimental condition .
the saturation coverage at the sample temperature of 80 k was defined to be 1 ml @xmath17 atoms / m@xmath2 as in ref . .
@xmath0 was varied by varying the dosage of xe gas onto the au surface .
we note that the desorption flux is defined as @xmath0 devided by the mean desorption period @xmath18 . on the basis of the result employing the litd simulation ,
@xmath18 was shown to be constant at about 4 ns at @xmath0 concerned in the present study.@xcite hence , the desorption flux can be varied by varying @xmath0 .
are 10 , 6.7 , 1.7 and 0.3 ml from top to bottom , respectively .
the dashed curves and the solid curves are the maxwell - boltzmann ( m - b ) velocity distribution and the shifted m - b velocity distribution fitted to the experimental results , respectively .
the vertical dashed line and the arrows indicate the peak position of each spectrum . [ tof ] ] the tof measurement of the litd of xe from an au surface was carried out in the following manner . the experimental setup is schematically shown in fig .
an arf excimer laser ( lambda physik ) was used as a pulsed laser source .
the wavelength was 193 nm and the time duration was 8 ns .
the incidence angle was 25@xmath19 from the surface normal direction .
the irradiated area on the sample disk was @xmath200.1 cm@xmath2 .
the laser pulse was guided on to the surface through an aperture and a quartz window .
the laser power absorbed by the sample was fixed at 80 mj / cm@xmath2 in all experiments taking the reflectivity into consideration . at this laser power absorbed by the sample ,
the litd of xe is the dominant desorption process as shown in ref . .
the reflected light escapes the uhv chamber through another quartz window . a quadrupole mass spectrometer ( qms : balzers qma125 )
was placed in the surface normal direction at a distance of 10 cm from the au specimen .
the qms was tuned for a high - sensitive detection of xe with a low mass selectivity .
the ion current of the qms was amplified with a fast current amplifier ( keithley 427 ) and sent to an oscilloscope ( tektronixs : tds620b ) .
the desorption of xe from the au(001 ) surface was induced with only one laser pulse in each litd experiment .
the respective tof was recorded using the oscilloscope with a single acquisition mode . of the desorbed xe atoms from au surfaces following the pulsed laser irradiations as a function of the xe coverage @xmath0 .
the data are obtained by analyzing the time - of - flight spectrum with the shifted maxwell - boltzmann velocity distribution .
the solid line is a guide for eyes . @xmath6
become constant at above 4 ml which is denoted by the arrow.[tud ] ]
we obtained a series of tof spectra of desorbed xe from au surfaces following pulsed laser irradiation as shown in fig .
the tof spectrum of xe from the au surface at @xmath3 ml is shown at the bottom of fig .
the vertical dashed line at 0.42 ms shows the peak position of the spectrum .
the spectrum was well analyzed with a maxwell - boltzmann velocity distribution in a flux from a thermal source described as , @xcite @xmath22 where @xmath23 , @xmath24 and @xmath25 are the mass of a xe atom , the boltzmann constant and the velocity of xe atoms , respectively , and @xmath26 and @xmath5 are fitting parameters . in the analysis , we convert eq .
( [ mb ] ) to a tof function for the density sensitive detector , which results in the form @xmath27 . by fitting eq .
( [ mb ] ) to the tof of xe at @xmath28 ml , we obtained @xmath5 of 255 k as illustrated by the dashed curve at the bottom of fig .
[ tof ] . at the top and middle of fig .
[ tof ] , the tof spectra of desorbed xe at @xmath29 to 1.7 ml are shown .
we note that the peak positions of those tof spectra , which are indicated by the arrows , are shifted towards smaller values with increasing @xmath0 compared with that of @xmath0 = 0.3 ml .
it was found that the peak positions of the tof spectra become constant at about 0.34 ms at @xmath30 ml .
we first discuss that the observed desorption proceeds only via thermal activation . in the previous study,@xcite we investigated the same system by varying the photon fluence at @xmath31 ml . at the small fluence region where the laser power is below 20 mj / cm@xmath2 , we observed a desorption of xe only at a photon energy of 6.4 ev and no desorption at 2.3 ev .
we assigned this photodesorption as a non - thermal desorption via a transient xe@xmath32 formation . at a large fluence region where laser power exceeds 60 mj / cm@xmath2 we observed , on the other hand , that thermal desorption is the dominant desorption process regardless of the photon energy .
these observations are confirmed by comparing the experimentally observed @xmath5 and xe desorption yield with the results of litd simulations .
therefore , we regard that the initial desorption is only thermally activated in the present experiment . in order to analyze the tof spectra at @xmath33 ml
, we introduce a shifted maxwell - boltzmann velocity distribution which is characterized by @xmath6 .
generally , the velocity distribution in a desorption flux shows an angular distribution and is expressed by the elliptical distribution which involves an angular dependent translational temperature as described in ref . .
in the present experiment , the detector was fixed in the surface normal direction .
therefore , the translational temperature only in this direction is detected in the present experiment . in this case , the shifted maxwell - boltzmann velocity distribution in a flux from a thermal source is expressed as,@xcite @xmath34 under the condition of @xmath35 , eq .
( [ smb ] ) is identical to eq .
( [ mb ] ) . in the analysis ,
( [ smb ] ) was also converted to a tof function for the density sensitive detector .
the tof spectra at @xmath4 ml were well analyzed using eq .
( [ smb ] ) . however , we found it difficult to obtain a unique set of @xmath5 and @xmath6 .
the spectra were well reproduced with an arbitrary value of @xmath5 from 165 to 310 k by adjusting @xmath6 from 125 to 0 m / s .
we note that , with a fixed value of @xmath5 between 165 and 310 k , the obtained value of @xmath6 monotonically increases with increasing @xmath0 from 0.3 to 4 ml and becomes constant at @xmath7 ml .
for example , fig .
[ tud ] shows u as a function of @xmath0 obtained by using eq .
( [ smb ] ) with @xmath5 fixed at 165 k , of which the fitting curves are shown as solid curves in fig .
[ tof ] . at @xmath36 ,
the observed feature of the tof can be rationalized with a model employing thermal desorption followed by collision - free flow . at the instance of thermal desorption ,
the desorbed xe gas is in thermal equilibrium with the surface at the temperature @xmath10 . here
, it is reasonable to consider the half - range maxwell - boltzmann velocity distribution for desorbed xe expressed as @xcite @xmath37 where @xmath26 and @xmath38 are a normalization factor and the velocity vector of xe atoms in the cartesian coordinate , respectively .
we note that in eq .
( [ mbv ] ) the velocity component in the @xmath39 direction has a distribution only at @xmath40,@xcite assuming the @xmath39 axis to be the surface normal direction .
given the desorption period is much shorter than the flight time and the irradiation diameter is much smaller than the flight distance , eq .
( [ mbv ] ) can be , as a velocity distribution in a flux from a thermal source , transformed to the same form as eq .
( [ mb ] ) .
thus , we notice that @xmath41 at @xmath36.@xcite for a qualitative analysis , we estimated the surface temperature during the present litd experiment using the first order desorption rate equation , assuming that the desorption activation energy is 240 mev@xcite and that the desorption period is 4 ns.@xcite the calculated results show that the required surface temperature is 260 k , which is in good agreement with the obtained value of @xmath42 k at @xmath28 ml . the agreement indicates that the model of the thermal desorption followed by the collision - free flow well describes the observed tof at @xmath36 , in good accordance with the results by cowin @xcite and wedler @xcite . at @xmath43 ml ,
we discuss that the desorption flow is described as an intense flow , where the post - desorption collision modifies the velocity distribution.@xcite several theoretical studies have speculated that the so - called knudsen layer is formed in the vicinity of the surface as a result of collisions.@xcite as schematically drawn in fig .
[ kl ] , in the knudsen layer model , the initial velocity distribution is in thermal equilibrium with the surface at @xmath10 except that there is no distribution at @xmath44 as in eq .
( [ mbv ] ) . as a result of a significant number of post - desorption collisions ,
the half - range velocity distribution at @xmath45 becomes thermally equilibrated at some distance from the surface to a full - range maxwell - boltzmann velocity distribution with a stream velocity @xmath6.@xcite the model also requires that , at the end of the knudsen layer , the temperature of the flow becomes identical ( @xmath11 ) in all degrees of freedom .
hence , at the end of the knudsen layer , the velocity distribution may be described as @xcite @xmath46 d\mathbf{v},\label{smbv}\ ] ] where @xmath47 , @xmath48 and @xmath11 are the velocity component in the cartesian coordinate , stream velocity and the temperature at the end of the knudsen layer , respectively . under the condition that the flight distance is sufficiently longer than the thickness of the knudsen layer besides the conditions described above , eq .
( [ smbv ] ) is converted , as a velocity distribution from a thermal source , to the identical form to eq .
( [ smb ] ) .
thus , we see that @xmath49 and @xmath50 at @xmath43 ml .
equation ( [ smb ] ) was shown to well describe the experimental results of the tofs at @xmath43 ml .
this fact , along with the convergence feature of @xmath6 , indicates that the experimental result at @xmath43 ml may be rationalized with the knudsen layer formation model .
we further examine the results at @xmath43 ml by quantitatively comparing the obtained results of @xmath48 with previous theoretical reports on knudsen layer formation in the steady strong evaporation @xcite and the pulsed desorption @xcite from plane surfaces .
we compare @xmath48 in terms of the mach number @xmath12 at the end of the knudsen layer .
@xmath12 is defined as @xmath51 where @xmath52 and @xmath53 are the local velocity of sound and the heat capacity ratio of the gas , respectively , the latter of which is 5/3 for monoatomic gas as in the present case .
.[tablem]mach number @xmath12 obtained in the present study and previously reported values by simulation and theory on the knudsen layer . [ cols="<,^,^,^ " , ]
lastly , we note the observed feature at @xmath54 ml , where the knudsen layer is not formed .
thus , the velocity distributions are not necessarily well fitted with eq .
( [ smb ] ) . nevertheless , they were very well analyzed with eq .
( [ smb ] ) in all @xmath0 at @xmath54 ml .
the result is in good agreement with the simulated results by sibold and urbassek,@xcite although any analytical formulation for the flow at @xmath54 ml has not been presented so far .
hence , further theoretical and experimental works are required for elucidating the mechanism for the modification of velocity distribution at @xmath54 ml due to the post - desorption collision as pointed out in ref . .
in conclusion , we observed the tofs of xe from a au(001 ) surface by litd at a wide range of @xmath0 . at @xmath55
ml , the tof is well analyzed by the maxwell - boltzmann velocity distribution at 255 k indicating thermal desorption followed by the collision - free flow . at @xmath33
ml , the tof is well analyzed by the shifted maxwell - boltzmann velocity distribution , which is regarded as the manifestations of the post - desorption collision effects . at @xmath43
ml , the value of @xmath6 under the assumption that @xmath13 became constant at around 125 m / s , which corresponds to @xmath56 , being in good agreement with the prediction by the knudsen layer formation theory . in the litd experiment , we found that the collision effect already appears at kn @xmath57 and that the knudsen layer is formed at kn @xmath8 , which corresponds to a mean collision number of greater than 2.6 per atom for general systems .
this work was supported by grant - in aid for scientific research ( a ) of japan society for the promotion of science ( jsps ) .
a. i. acknowledges support from a research assistant of global coe program `` the physical sciences frontier '' , mext , japan and tepco memorial foundation .
40ifxundefined [ 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 _
( , , ) .
@noop * * , ( ) .
@noop * * , ( ) .
@noop * * ( ) . @noop * * , ( ) .
@noop * * , ( ) .
@noop * * , ( ) .
* * , ( ) .
@noop * * , ( ) .
@noop * * , ( ) .
@noop * * , ( ) . in @noop
_ , ( , , ) p. . in @noop
_ _ , ( , , ) p. .
@noop * * , ( ) .
@noop * * , ( ) .
@noop * * , ( ) .
@noop * * , ( ) .
@noop * * , ( ) .
@noop * * , ( ) . @noop * * , ( ) .
@noop * * , ( ) .
@noop * * , ( ) .
@noop * * , ( ) .
@noop * * , ( ) . | laser induced thermal desorption of xe atoms into vacuum from a metal surface following the nano - second pulsed laser heating was investigated by the time - of - flight ( tof ) measurement .
the desorption flow was studied at a wide range of desorption flux by varying the initially prepared xe coverage @xmath0 ( 1 ml @xmath1 atoms / m@xmath2 ) . at @xmath3 ml , the tof of xe was well represented by a maxwell - boltzmann velocity distribution , which is in good agreement with thermal desorption followed by collision - free flow . at @xmath4 ml , the peak positions of the tof spectra were shifted towards the smaller values and became constant at large @xmath0 , which were well fitted with a shifted maxwell - boltzmann velocity distribution with a temperature @xmath5 and a stream velocity @xmath6 . with @xmath5 fixed at 165 k , @xmath6 was found to increase from 80 to 125 m/ s with increasing @xmath0 from 1.2 to 4 ml . at @xmath7
ml , the value of @xmath6 become constant at 125 m/ s. the converging feature of @xmath6 was found to be consistent with analytical predictions and simulated results based on the knudsen layer formation theory .
we found that the knudsen layer formation in laser desorption is completed at knudsen number kn @xmath8 . |
let @xmath0 be an integer @xmath1-matrix such that @xmath2 is in the row span of @xmath0 .
this defines a _ projective toric variety _
@xmath3 in @xmath4 .
its _ dual variety _
@xmath5 is the closure in the projective space dual to @xmath4 of the set of hyperplanes that are tangent to @xmath3 at a regular point . the toric variety @xmath3 is called _ non - defective _ if its dual variety @xmath5 has codimension one . in this case , the _
@xmath0-discriminant _ is the irreducible homogeneous polynomial @xmath6 which defines the hypersurface @xmath5 .
alternatively , the dual variety @xmath5 can be thought of as the set of singular hypersurfaces in @xmath7 with newton polytope prescribed by the matrix @xmath0 .
the study of these objects is an active area of research in computational algebraic geometry , with the fundamental reference being the monograph by gelfand , kapranov and zelevinsky @xcite .
our main object of interest in this paper is the _ tropical @xmath0-discriminant _ @xmath8 .
this is the polyhedral fan in @xmath9 which is obtained by tropicalizing @xmath5 .
while it is generally difficult to compute the dual variety @xmath5 from @xmath0 , we show that its tropicalization @xmath8 can be computed much more easily . in theorem
[ thm : tropical ] , we derive an explicit description of the tropical @xmath0-discriminant @xmath8 , and in theorem [ thm : initial ] we present an explicit combinatorial formula for the extreme monomials of @xmath6 . without loss of generality , we assume that the columns of the matrix @xmath0 span the integer lattice @xmath10 , and that the point configuration given by the columns of @xmath0 is not a pyramid .
these hypotheses ensure that the toric variety @xmath3 has dimension @xmath11 and that the dual variety @xmath5 is not contained in any coordinate hyperplane .
a key player in this paper is the tropicalization of the kernel of @xmath0 .
as shown in @xcite and @xcite , this tropical linear space is subdivided both by the _ bergman fan _ of the matroid dual to @xmath0 , i.e. , the _ co - bergman fan _
@xmath12 of @xmath0 , and by the _ nested set fans _ of its lattice of flats @xmath13 . in other words , let @xmath14 denote the geometric lattice whose elements are the sets of zero - entries of the vectors in @xmath15 , ordered by inclusion .
we write @xmath16 for the set of proper maximal chains in @xmath14 and represent these chains as @xmath17-element subsets @xmath18 of @xmath19 .
we obtain the following descriptions of the tropicalization of the kernel of @xmath0 : @xmath20 the union on the right hand side in fact is the finest in the hierarchy of unimodular simplicial fan structures provided by the _ nested set fans _ @xcite .
the tropical linear space ( [ eq : cobergmanfan ] ) is a subset of @xmath9 .
we obtain the tropical @xmath0-discriminant by adding this tropical linear space to the ( classical ) row space of the @xmath1-matrix @xmath0 : [ thm : tropical ] for any @xmath21-matrix @xmath0 as above , the tropical @xmath0-discriminant @xmath8 equals the minkowski sum of the co - bergman fan @xmath22 and the row space of @xmath0 .
theorem [ thm : tropical ] is the tropical analogue of kapranov s _ horn uniformization _ @xcite . by definition ,
the tropical discriminant @xmath8 inherits the structure of a fan from the grbner fan of the ideal of @xmath5 and , in the non - defective case , also from the secondary fan of @xmath0 @xcite . in general ,
neither of these two fan structures refines the other , as we shall see in examples [ ex : k_4 ] and [ ex : bjsst ] .
the tropicalization of a variety retains a lot of information about the geometry of the original variety @xcite . in theorem
[ thm : initial ] below , our tropical approach leads to a formula for the extreme monomials of the @xmath0-discriminant @xmath6 , and , a fortiori , for the degree of the dual variety @xmath5 .
an alternating product formula for the extreme monomials of @xmath6 was given in @xcite under the restrictive assumption that @xmath3 is smooth .
our formula ( [ eq : degree1 ] ) is positive , it is valid for any toric variety @xmath3 regardless of smoothness , and its proof is self - contained .
[ thm : initial ] if @xmath3 is non - defective and @xmath23 a generic vector in @xmath9 then the exponent of @xmath24 in the initial monomial @xmath25 of the @xmath0-discriminant @xmath6 equals @xmath26 where @xmath27 is the subset of @xmath16 consisting of all chains such that the row space of the matrix @xmath0 has non - empty intersection with the cone @xmath28 . here , the @xmath0-discriminant @xmath6 is written as a homogeneous polynomial in the variables @xmath29 , and @xmath30 is the @xmath23-lowest monomial @xmath31 which appears in the expansion of @xmath6 in characteristic zero .
theorem [ thm : initial ] generalizes to the defective case , when we take @xmath6 as the _ chow form _ of the dual variety @xmath5 .
this is stated in theorem [ thm : initial2 ] .
aiming for maximal efficiency in evaluating ( [ eq : degree1 ] ) with a computer , we can replace @xmath32 with the corresponding maximal nested sets of the geometric lattice @xmath33 , or with the corresponding maximal cones in the bergman fan @xmath12 .
our maple implementation of the formula ( [ eq : degree1 ] ) is discussed in section [ sec : subdivisions ] .
this paper is organized as follows . in section [ sec : tropical ] , we review the construction of the tropicalization @xmath34 of a projective variety @xmath35 , and we show how the algebraic cycle underlying any initial monomial ideal of @xmath35 can be read off from @xmath34 . in section [ sec : monomials ] , we discuss general varieties which are parametrized by a linear map followed by a monomial map . theorem [ thm : tuv ] gives a combinatorial description of the tropicalization of the image of such a map .
the dual variety @xmath5 of any toric variety @xmath3 admits such a parametrization .
this is derived in section [ sec : horn ] , and it is used to prove theorem [ thm : tropical ] and theorem [ thm : initial ] in the general form of theorem [ thm : initial2 ] .
we also obtain in corollary [ cor : dimension ] a characterization of the dimension of @xmath5 . in section [
sec : subdivisions ] we discuss computational issues , and we examine the connection between the tropical discriminant of @xmath0 and regular polyhedral subdivisions .
in particular , we consider the problem of characterizing @xmath36-equivalence of regular triangulations in combinatorial terms . finally , section [ sec : resultants ] is devoted to the case when @xmath0 is an _ essential cayley configuration_. the corresponding dual varieties @xmath5 are resultant varieties , and we compute their degrees and initial cycles in terms of mixed subdivisions .
* acknowledgement * : we thank the forschungsinstitut fr mathematik at eth zrich for hosting alicia dickenstein and bernd sturmfels in the summer of 2005 .
we are grateful to jenia tevelev and josephine yu for comments on this paper .
_ tropical algebraic geometry _ is algebraic geometry over the tropical semi - ring @xmath37 with arithmetic operations @xmath38 and @xmath39 .
it transfers the objects of classical algebraic geometry into the combinatorial context of polyhedral geometry .
fundamental references include @xcite .
tropicalization is an operation that turns complex projective varieties into polyhedral fans .
let @xmath40 be an irreducible projective variety of dimension @xmath41 and @xmath42 $ ] its homogeneous prime ideal . for @xmath43 ,
we denote by @xmath44 the initial ideal generated by all initial forms @xmath45 , for @xmath46 , where @xmath45 is the subsum of all terms @xmath47 in @xmath48 which have _ lowest _ weight @xmath49 .
the _ tropicalization _
@xmath34 of @xmath35 is the set @xmath50 let @xmath51 be the field of puiseux series , i.e. , series with complex coefficients , real exponents and well ordered supports .
the elements of @xmath52 are also known as _
transfinite puiseux series _ , and they form an algebraically closed field of characteristic zero . the _ order _ of a non - zero element @xmath53 in @xmath54 is the smallest real number @xmath55 such that @xmath56 appears with non - zero coefficient in @xmath53 . for a vector @xmath57 in @xmath58
we write @xmath59 .
the points in the tropicalization @xmath34 are precisely the orders of @xmath60-valued points on the variety @xmath35 ( see @xcite , ( * ? ? ?
* theorem 2.1.2 ) , ( * ? ? ? * theorem 2.1 ) ) .
the set @xmath34 carries the structure of a polyhedral fan .
namely , it is a subfan of the grbner fan of @xmath61 ; see @xcite . by a result of bieri and groves @xcite ,
the fan @xmath34 is pure of dimension @xmath62 . in @xcite
it was shown that @xmath34 is connected in codimension one , and a practical algorithm was given for computing @xmath34 from polynomial generators of @xmath61 .
we will view the tropicalization @xmath34 of a projective variety as an @xmath63-dimensional fan in _ tropical projective space _
@xmath64 , which is an @xmath65-dimensional real affine space .
every maximal cone @xmath66 of the fan @xmath34 comes naturally with an _ intrinsic multiplicity _
@xmath67 , which is a positive integer .
the integer @xmath67 is computed as the sum of the multiplicities of all monomial - free minimal associated primes of the initial ideal @xmath68 in @xmath69 $ ] , where @xmath23 is in the relative interior of the cone @xmath66 .
[ speyerdiss ] a geometric description of the intrinsic multiplicity @xmath67 arises from the beautiful interplay of degenerations and compactifications discovered by tevelev @xcite and studied by speyer ( * ? ? ?
* chapter 2 ) and hacking ( unpublished ) .
let @xmath70 denote the toric variety associated with the fan @xmath34 .
consider the intersection @xmath71 of @xmath35 with the dense torus @xmath72 in @xmath4 , and let @xmath73 be the closure of @xmath71 in @xmath70 . by
* , and 2.7 ) , the variety @xmath73 is complete and the multiplication map @xmath74 is faithfully flat . if follows that the intersection of @xmath73 with a codimension @xmath75 orbit has codimension @xmath75 in @xmath73 . in particular
, the orbit @xmath76 corresponding to a maximal cone @xmath66 of @xmath34 intersects @xmath73 in a zero - dimensional scheme @xmath77 .
the intrinsic multiplicity @xmath67 of the maximal cone @xmath66 in the tropical variety @xmath34 is the length of @xmath77 .
we list three fundamental examples which will be important for our work . * ( 1 ) * let @xmath35 be a * hypersurface * in @xmath4 defined by an irreducible polynomial @xmath48 in @xmath69 $ ] .
then @xmath34 is the union of all codimension one cones in the normal fan of the newton polytope of @xmath48 .
the intrinsic multiplicity @xmath67 of each such cone @xmath66 is the lattice length of the corresponding edge of the newton polytope of @xmath48 . *
( 2 ) * let @xmath78 be the * toric variety * defined by an integer @xmath1-matrix @xmath0 as above .
its tropicalization @xmath79 is the linear space spanned by the rows of @xmath0 . *
( 3 ) * let @xmath35 be a * linear subspace * in @xmath80 or in @xmath4 .
the tropicalization @xmath34 is the bergman fan of the matroid associated with @xmath35 ; see @xcite and ( [ tlsrep1 ] ) below . in the last two families of examples ,
all the intrinsic multiplicities @xmath67 equal @xmath81 .
the tropicalization @xmath34 can be used to compute numerical invariants of @xmath35 .
first , the dimension of @xmath34 coincides with the dimension of @xmath35 . in theorem
[ thm : mult ] below , we express the multiplicities of the minimal primes in the initial monomial ideals of @xmath61 in terms of @xmath34 .
equivalently , we compute the algebraic cycle of any initial monomial ideal @xmath68 .
this formula tells us the degree of the variety @xmath35 , namely , the degree is the sum of the multiplicities of the minimal primes of @xmath82 .
let @xmath83 denote the codimension of the irreducible projective variety @xmath35 in @xmath4 .
assume that @xmath35 is not contained in a coordinate hyperplane , and let @xmath61 be its homogeneous prime ideal in @xmath69 $ ] .
if @xmath23 is a generic vector in @xmath9 , the initial ideal @xmath68 is a monomial ideal of codimension @xmath84 .
every minimal prime over @xmath68 is generated by a subset of @xmath84 of the variables .
we write @xmath85 for the monomial prime ideal indexed by the subset @xmath86 .
assume that the cone @xmath87 meets the tropicalization @xmath34 .
we may suppose that the generic weight vector @xmath88 satisfies that the image of @xmath23 in @xmath89 does not lie in @xmath34 and that the intersection of the cone @xmath87 with @xmath34 is finite and contained in the union of the relative interiors of its maximal cones .
let @xmath66 be a maximal cone of the tropical variety and @xmath90 where @xmath91 and @xmath92 are the corresponding linear spaces , which are defined over @xmath93 .
we associate with @xmath94 the _ lattice multiplicity _ of the intersection of @xmath95 and @xmath96 , which is defined as the absolute value of the determinant of any @xmath97-matrix whose columns consist of a @xmath98-basis of @xmath99 and a @xmath98-basis of @xmath100 .
here is the main result of this section .
[ thm : mult ] for @xmath88 a generic weight vector , a prime ideal @xmath101 is associated to the initial monomial ideal @xmath68 if and only if the cone @xmath87 meets the tropicalization @xmath34 .
the number of intersections , each counted with its associated lattice multiplicity times the intrinsic multiplicity , is the multiplicity of the monomial ideal @xmath68 along the prime @xmath101 .
we work over the puiseux series field @xmath102 , we write @xmath103 for the @xmath104-dimensional projective space over the field @xmath52 , and we consider @xmath35 as a subvariety of @xmath103 .
the translated variety @xmath105 is defined by the prime ideal @xmath106.\ ] ] clearly , a point of the form @xmath107 lies in @xmath34 if and only if there is a point in @xmath108 with order @xmath109 .
let @xmath95 be a general linear subspace of dimension @xmath84 in @xmath110 which is defined over @xmath111 .
the intersection @xmath112 is a finite set of reduced points in @xmath103 .
the number of these points is the degree @xmath113 of @xmath35 .
there is a flat family over @xmath52 with generic fiber @xmath108 and special fiber the scheme @xmath114 . since @xmath95 is generic , the intersection of @xmath95 with this family is still flat
then , all points in the intersection @xmath108 with @xmath95 are liftings of points in @xmath115 ( and all points can be lifted ) .
moreover , the multiplicity of the @xmath101-primary component of the initial ideal @xmath68 equals the number of points in @xmath116 of the form @xmath117 where @xmath118 for all @xmath75 , @xmath119 for @xmath120 and @xmath121 for @xmath122 , i.e. , the number of points with values in the intersection of @xmath34 with the cone @xmath123 . by our genericity assumption ,
each such intersection point @xmath124 lies on some maximal cone @xmath66 of @xmath34 and it is counted with its multiplicity , which is the product of the intrinsic multiplicity @xmath67 times the lattice multiplicity of the transversal intersection of rational linear spaces in ( [ eq : ll ] ) .
this product can be understood by means of the flat family discussed in remark [ speyerdiss ] .
namely , it follows from the @xmath72-invariance of the multiplication map @xmath125 that the scheme - theoretic fiber of @xmath125 over any point of @xmath76 is isomorphic to @xmath126 , where @xmath127 is the stabilizer of a point in @xmath76 .
since @xmath70 is normal , @xmath127 is a torus @xmath128 .
by @xcite , @xmath129 is the intersection of @xmath72 with the flat degeneration of @xmath35 in @xmath4 given by the one parameter subgroup of @xmath72 specified by the rational vector @xmath23 .
our construction above amounts to computing the intersection of @xmath130 with the corresponding degeneration of @xmath95 .
the lattice index is obtained from the factor @xmath128 , and @xmath67 is obtained from the factor @xmath77 .
their product is the desired intersection number . the algebraic cycle of the variety @xmath35
is represented by its _ chow form _ @xmath131 , which is a polynomial in the bracket variables @xmath132 = [ \gamma_1 \cdots \gamma_c]$ ] ; see ( * ? ? ?
theorem [ thm : mult ] implies that the @xmath23-leading term of the chow form @xmath131 equals @xmath133^{u_\gamma } $ ] , where @xmath134 is the ( correctly counted ) number of points in @xmath135 .
we discuss this statement for the three families of examples considered earlier . *
( 1 ) * if @xmath35 is a hypersurface then @xmath136 and the bracket variable @xmath132 $ ] is simply the ordinary variable @xmath24 for @xmath137 .
the @xmath23-leading monomial of the defining irreducible polynomial @xmath138 equals @xmath139 where @xmath140 is the number of times the ray @xmath141 intersects the tropical hypersurface @xmath34 , counted with multiplicity . *
( 2 ) * if @xmath142 is a toric variety then @xmath143 and theorem [ thm : mult ] implies the familiar result ( * ? ? ?
8.3.3 ) that the initial cycles of @xmath3 are the regular triangulations of @xmath0 .
indeed , @xmath144 intersects the row space of @xmath0 if and only if the @xmath145-simplex @xmath146 appears in the regular triangulation @xmath147 of @xmath0 induced by @xmath23 . for
a precise definition of @xmath147 see section [ sec : subdivisions ] .
the intersection multiplicity is the lattice volume of @xmath148 . hence @xmath149^{{\rm
vol}(\bar \gamma ) } $ ] . * ( 3 ) * if @xmath35 is a linear space then its ideal @xmath61 is generated by @xmath84 linearly independent linear forms in @xmath69 $ ] .
the tropical variety @xmath34 is the bergman fan , to be discussed in section [ sec : monomials ] , and the grbner fan of @xmath61 is the normal fan of the associated matroid polytope @xcite . for fixed generic @xmath23
, there is a unique @xmath84-element subset @xmath150 of @xmath151 $ ] such that @xmath152 intersects @xmath34 .
the intersection multiplicity is one , and the corresponding initial ideal equals @xmath153 .
in this section we examine a class of rational varieties @xmath35 whose tropicalization @xmath34 can be computed combinatorially , without knowing the ideal @xmath61 . we consider a rational map @xmath154 that factors as a linear map @xmath155 followed by a laurent monomial map @xmath156 .
the linear map is specified by a complex @xmath157-matrix @xmath158 , and the laurent monomial map is specified by an integer @xmath159-matrix @xmath160 .
thus the @xmath161-th coordinate of the rational map @xmath162 equals the following monomial in linear forms : @xmath163 let @xmath164 denote the zariski closure of the image of @xmath48 . observe that if all row sums of @xmath165 are equal then @xmath48 induces a rational map @xmath166 , and the closure of its image is an irreducible projective variety , which we also denote by @xmath164 .
our goal is to compute the tropicalization @xmath167 of the variety @xmath164 in terms of the matrices @xmath168 and @xmath165 . in particular
, we will avoid any reference to the equations of @xmath164 .
the general framework of this section will be crucial for our proofs of the results on @xmath0-discriminants and their tropicalization stated in the introduction .
we list a number of special cases of varieties which have the form @xmath164 . *
( 1 ) * if @xmath169 , and @xmath170 then @xmath48 is the linear map @xmath171 , and @xmath172 is the image of @xmath168 .
we denote this * linear subspace * of @xmath173 by @xmath174 .
its tropicalization @xmath175 is the bergman fan of the matrix @xmath168 , to be discussed in detail below . *
( 2 ) * if @xmath176 and @xmath177 then @xmath48 is the monomial map specified by the matrix @xmath165 .
hence @xmath164 coincides with the * toric variety * @xmath178 which is associated with the transpose @xmath179 of the matrix @xmath165 .
its tropicalization is the column space of @xmath165 . *
( 3 ) * let @xmath180 , suppose the rows of @xmath168 are linearly independent , and suppose the matrix @xmath165 has constant row sum
. then @xmath181 is a * rational curve*. every rational projective curve arises from this construction , since every binary form is a monomial in linear forms .
our next theorem implies that @xmath167 consists of the rays in @xmath182 spanned by the rows of @xmath165 .
theorem [ thm : tuv ] can also be derived from ( * ? ? ?
* proposition 3.1 ) , but we prefer to give a proof that is entirely self - contained .
[ thm : tuv ] the tropical variety @xmath167 equals the image of the bergman fan @xmath175 under the linear map @xmath183 specified by the matrix @xmath165 . in what follows we consider all algebraic varieties over the field @xmath184 and we use the characterization of the tropical variety @xmath167 in terms of @xmath60-valued points of the ideal of @xmath164 . extending scalars , we consider the map @xmath185 .
let @xmath186 be any point in the image of that map .
for @xmath187 we set @xmath188 and @xmath189 .
the vector @xmath190 lies in the linear space @xmath174 , and hence the vector @xmath191 lies in @xmath175 .
the order of the vector @xmath192 is the vector @xmath193 , since the order of its @xmath161th coordinate @xmath194 equals @xmath195 .
hence @xmath196 lies in @xmath197 , the image of the bergman fan @xmath175 under the linear map @xmath165 .
the image of @xmath48 is zariski dense in @xmath164 , i.e. , there exists a proper subvariety @xmath35 of @xmath164 such that @xmath198 contains the image of @xmath48 . by the bieri - groves theorem
@xcite , @xmath34 is a polyhedral fan of lower dimension inside the pure - dimensional fan @xmath167 . from this
it follows that @xmath167 is the closure of @xmath199 in @xmath200 . in the previous paragraph we showed that @xmath199 is a subset of @xmath197 .
since the latter is closed , we conclude that @xmath201 .
it remains to show the converse inclusion @xmath202 .
take any point @xmath203 , and choose @xmath204 such that @xmath205 .
there exists a @xmath60-valued point @xmath206 in the linear space @xmath174 such that @xmath207 .
then the point @xmath208 lies in @xmath209 , and its order clearly equals @xmath210 . hence @xmath211 as desired .
[ intrmultyuv ] the intrinsic multiplicity @xmath67 of any maximal cone @xmath66 in a sufficiently fine fan structure on the tropical variety @xmath167 is a lattice index which can be read off from the matrix @xmath165 .
namely , suppose @xmath66 lies the image of a maximal cone @xmath212 of the bergman fan @xmath175 .
then @xmath67 is the index of the subgroup @xmath213 of @xmath214 .
this follows from remark [ speyerdiss ] using standard arguments of toric geometry .
theorem [ thm : tuv ] and remark [ intrmultyuv ] furnish a combinatorial construction for the tropicalization of any variety which is parameterized by monomials in linear forms .
using the results of section [ sec : tropical ] , theorem [ thm : tuv ] can now be applied to compute geometric invariants of such a variety , for instance , its dimension , its degree and its initial cycles . to make this computation effective , we need an explicit description of the bergman fan @xmath175 . luckily ,
the relevant combinatorics is well understood , thanks to @xcite , and in the remainder of this section we summarize what is known .
let @xmath215 denote the _ matroid _ associated with the rows of the @xmath216-matrix @xmath168 .
thus @xmath215 is a matroid of rank at most @xmath217 on the ground set @xmath218 = \{1,2,\ldots , r\}$ ] .
bases _ of @xmath215 are the maximal subsets of @xmath218 $ ] which index linearly independent rows of @xmath168 .
fix a vector @xmath219 .
then the _ @xmath23-weight _ of a basis @xmath220 of @xmath215 is @xmath221 . consider the set of all bases of @xmath215 that have maximal @xmath23-weight .
this collection is the set of bases of a new matroid which we denote by @xmath222 .
note that each @xmath222 has the same rank and the same ground set as @xmath223 .
an element @xmath224 of @xmath218 $ ] is a _ loop _ of @xmath222 if it does not lie in any basis of @xmath215 of maximal @xmath23-weight .
we can now describe @xmath175 in terms of the matroid @xmath215 : @xmath225 this representation endows our tropical linear space with the structure of a polyhedral fan .
namely , if @xmath226 , then the set of all @xmath227 such that @xmath228 is a relatively open convex polyhedral cone in @xmath229 .
the collection of these cones is denoted @xmath230 and is called the _ bergman fan _ of the matroid @xmath215 . depending on the context
, we may also write @xmath231 and call it the bergman fan of the matrix @xmath168 . we now recall the connection between bergman fans and _ nested set complexes_. the latter encode the structure of wonderful compactifications of hyperplane arrangement complements in the work of de concini and procesi @xcite , and they were later studied from a combinatorial point of view in @xcite .
a subset @xmath232 $ ] is a _
flat _ of our matroid @xmath215 if there exists a vector @xmath233 such that @xmath234 : u_i \,{=}\ , 0 \}$ ] .
the set of all flats , ordered by inclusion , is the _ geometric lattice _
a flat @xmath236 in @xmath237 is called _ irreducible _ if the lower interval @xmath238 does not decompose as a direct product of posets .
denote by @xmath239 the set of irreducible elements in @xmath240 . in other contexts ,
the irreducible elements of a lattice of flats of a matroid were named _
connected elements _ or _ dense edges_. the matroid @xmath215 is _ connected _ if the top rank flat @xmath241 $ ] is irreducible , and we assume that this is the case . otherwise , we artificially add @xmath242 to @xmath239 .
we call a subset @xmath243 _ nested _ if for any set of pairwise incomparable elements @xmath244 in @xmath245 , with @xmath246 , the join @xmath247 is not contained in @xmath239 .
the nested subsets in @xmath239 form a simplicial complex , the _ nested set complex _ @xmath248 .
see ( * ? ? ?
2.3 ) for further information .
feichtner and yuzvinsky ( * ? ? ?
* eqn ( 13 ) ) introduced the following natural geometric realization of the nested set complex @xmath248 .
namely , the collection of cones @xmath249 forms a unimodular fan whose face poset is the face poset of the nested set complex of @xmath240 . here
@xmath250 denotes the incidence vector of a flat @xmath251 .
we consider this fan in the tropical projective space @xmath89 , and we also denote it by @xmath248 . it was shown in ( * ? ? ?
* thm 4.1 ) that the _ nested set fan _
@xmath252 is a simplicial subdivision of the bergman fan @xmath253 , and hence of our tropical linear space @xmath175 .
the bergman fan need not be simplicial , so the nested set fan can be finer than the bergman fan .
however , in many important cases the two fans coincide @xcite .
what we defined above is the coarsest in a hierarchy of nested set complexes associated with the geometric lattice @xmath254 .
namely , for certain choices of subsets @xmath255 in @xmath240 , the same construction gives a nested set complex @xmath256 which is also realized as a unimodular simplicial fan .
such @xmath255 are called _ building sets _ ; there is one nested set fan for each building set @xmath255 in @xmath257 , see ( * ? ? ?
2.2 , 2.3 ) .
if two building sets are contained in another , @xmath258 , then @xmath259 is obtained from @xmath260 by a sequence of stellar subdivisions ( * ? ? ?
* thm 4.2 ) .
the smallest building set is @xmath261 , the case discussed above , and the largest building set is the set of all flats , @xmath262 . in the letter case , the corresponding nested set complex
@xmath263 is the _ order complex _ or _ flag complex _
@xmath264 of the lattice , i.e. , the simplicial complex on @xmath265 whose simplices are the totally ordered subsets .
again , there is a realization of @xmath264 as a unimodular fan with rays generated by the incidence vectors of flats in @xmath240 .
we call this fan the _ flag fan _ of @xmath240 and denote it by @xmath264 for ease of notation . summarizing the situation , we have the following sequence of subdivisions each of which can be used to compute the tropicalization of a linear space .
[ threesubdivisions ] given a matrix @xmath168 and the matroid @xmath215 of rows in @xmath168 , the tropical linear space @xmath175 has three natural fan structures : the bergman fan @xmath253 is refined by the nested set fan @xmath266 , which is refined by the flag fan @xmath267 .
we present an example which illustrates the concepts developed in this section .
[ ex : uv ] let @xmath268 , @xmath269 , @xmath270 , and consider the map @xmath271 from @xmath272 to @xmath272 whose coordinates are the following monomials in linear forms :
@xmath273 so @xmath48 is as in ( [ eq : linmonmap ] ) with @xmath274 the projectivization of the variety @xmath275 is a surface in @xmath276 , since @xmath168 has rank @xmath277 .
we construct the irreducible homogeneous polynomial @xmath278 which defines this surface .
the matroid of @xmath168 is the graphic matroid of @xmath279 with one edge removed . from (
* example 3.4 ) we know that the bergman complex @xmath231 is the complete bipartite graph @xmath280 , embedded as a @xmath281-dimensional fan in @xmath282 . the tropical surface @xmath167 is the image of @xmath231 under @xmath165 .
this image is a @xmath281-dimensional fan in @xmath283 .
it has seven rays : six of them are images of the rays of @xmath231 , the last one is the intersection of the images of two @xmath281-dimensional faces that occurs due to the non - planarity of @xmath280 .
hence the newton polytope of the polynomial @xmath284 is @xmath277-dimensional with @xmath285 vertices , @xmath286 edges , and @xmath287 facets ; see figure [ fig : np ] .
the six extreme monomials of @xmath284 can be computed ( even by hand ) using theorem [ thm : mult ] , namely , by intersecting the rays @xmath288 with @xmath167 in @xmath289 .
this computation reveals in particular that the degree of the polynomial @xmath284 is @xmath290 .
using linear algebra , it is now easy to determine all @xmath291 monomials in the expansion of @xmath284 .
in this section we return to the setting of the introduction , and we prove theorem [ thm : tropical ] and theorem [ thm : initial ] . recall that @xmath0 is an integer @xmath292-matrix such that @xmath293 is in the row span of @xmath0 , i.e. , the column vectors @xmath294 lie in an affine hyperplane in @xmath295 .
we also assumed that the vectors @xmath296 span @xmath10 .
we identify the matrix @xmath0 with the point configuration @xmath297 .
the convex hull of the configuration @xmath0 is a @xmath298-dimensional polytope with @xmath299 vertices .
the projective toric variety @xmath3 is defined as the closure of the image of the monomial map @xmath300 @xmath301 .
equivalently , @xmath3 is the set of all points @xmath302 such that @xmath303 for all @xmath304 with @xmath305 .
let @xmath306 denote the projective space dual to @xmath4 .
the point @xmath307 in @xmath308 corresponds to the hyperplane @xmath309 .
the _ dual variety _
@xmath5 is defined as the closure in @xmath310 of the set of points @xmath311 such that the hyperplane @xmath312 intersects the toric variety @xmath3 at a regular point @xmath313 and contains the tangent space @xmath314 of @xmath3 at @xmath313 .
kapranov @xcite showed that reduced discriminantal varieties are parametrized by monomials in linear forms .
this parametrization , called the _ horn uniformization _ , will allow us to determine the tropical discriminant @xmath315 via the results of section [ sec : monomials ] .
we denote by @xmath316 the projectivization of the kernel of the linear map given by @xmath0 , an @xmath317-dimensional projective subspace of @xmath4 , and we denote by @xmath318 the dense torus of @xmath3 .
the following result is a variant of ( * ? ? ?
* theorem 2.1 ) ; see also @xcite .
[ prop : hdparametrization ] the dual variety @xmath5 of the toric variety @xmath3 is the closure of the image of the map @xmath319 which is given by @xmath320 consider the unit point @xmath321 on the toric variety @xmath3 .
the hyperplane @xmath312 contains both the point @xmath322 and the tangent space @xmath323 at this point if and only if @xmath324 lies in the kernel of @xmath0 .
this follows by evaluating the derivative of the parametrization @xmath325 of @xmath3 at @xmath326 .
if @xmath327 is any point in the dense torus of @xmath3 , then the tangent space at that point is gotten by translating the tangent space at @xmath328 as follows : @xmath329 the hyperplane @xmath312 contains @xmath313 if and only if @xmath330 contains @xmath328 , and @xmath312 contains @xmath314 if and only if @xmath331 contains @xmath323 .
these two conditions hold , for some @xmath332 in the dense torus of @xmath3 , if and only if @xmath333 .
proposition [ prop : hdparametrization ] shows that the dual variety @xmath5 of the toric variety @xmath3 is parametrized by monomials in linear forms . in the notation of section [ sec : monomials ] we set @xmath334 , @xmath335 , @xmath336 , and the two matrices are @xmath337 where @xmath338 is an @xmath339-matrix whose columns span the kernel of @xmath0 over the integers .
thus the rows of @xmath338 are _ gale dual _ to the configuration @xmath0 .
[ lem : connecttosec3 ] the variety @xmath340 defined by ( [ chooseuandv ] ) as in section [ sec : monomials ] is equal to the cone over the dual variety @xmath341 of the toric variety @xmath342 .
let @xmath343 be the rational map defined by ( [ chooseuandv ] ) as in section [ sec : monomials ] , and set @xmath344 and @xmath345 .
then ( [ eq : linmonmap ] ) equals @xmath346 which equals the @xmath161-th coordinate of @xmath347 if we write @xmath348 as the image of @xmath338 .
we are now ready to prove theorem [ thm : tropical ] .
_ proof of theorem [ thm : tropical ] .
_ let us first note that the _ co - bergman fan _ @xmath22 of the rank @xmath113 configuration given by the columns of @xmath0 equals the bergman fan of the rank @xmath349 configuration given by the rows of @xmath338 .
thus @xmath22 , as it appears in ( [ eq : cobergmanfan ] ) is the bergman fan of the matroid dual to the matroid given by the columns of @xmath0 .
the support of the co - bergman fan @xmath22 is the tropicalization of the linear space @xmath350 .
now , if @xmath168 is taken as in ( [ chooseuandv ] ) then we have the following decomposition in @xmath351 : @xmath352 the image of this fan under the linear map @xmath353 is the ( minkowski ) sum of @xmath22 and the image of @xmath354 .
of course , the latter is the row space of @xmath0 .
hence our assertion follows from lemma [ lem : connecttosec3 ] and theorem [ thm : tuv ] .
@xmath355 similarly , the tropicalization of the reduced version of the dual variety can be derived from theorem [ thm : tuv ] .
let @xmath338 be again an @xmath356-matrix whose columns span the kernel of @xmath0 over the integers .
the _ reduced dual variety _
@xmath357 is the closure of the image of the rational morphism @xmath358 , whose @xmath161-th coordinate equals @xmath359 [ crl : reduceddualvariety ] given a gale dual @xmath338 of @xmath0 , the tropicalization of the reduced dual variety @xmath360 is the image of the co - bergman fan @xmath22 under the linear map @xmath361 .
we often do not distinguish between the reduced and unreduced version of dual varieties and their tropicalizations , and denote both by @xmath5 and @xmath8 , respectively .
we illustrate theorem [ thm : tropical ] for the case when @xmath362 is the veronese surface , regarded as the projectivization of the variety of all symmetric @xmath363 matrices of rank @xmath364 .
[ ex : symm3x3 ] we take @xmath365 , and we fix the matrix @xmath366 note that the more usual matrix @xmath367 defines the same veronese embedding of @xmath368 into @xmath369 , but the columns of @xmath370 do not span @xmath371 and the monomial parametrization @xmath372 is two - to - one .
points @xmath373 in @xmath369 are identified with symmetric @xmath363-matrices @xmath374 a point @xmath109 is in @xmath316 if and only if the corresponding matrix @xmath168 has zero row and column sums .
if this holds , and @xmath375 is any point in @xmath376 , then the symmetric @xmath363-matrix @xmath236 corresponding to @xmath377 is singular because it satisfies @xmath378 hence @xmath347 parametrizes rationally the hypersurface of singular symmetric matrices @xmath236 , and the @xmath0-discriminant equals the classical discriminant @xmath379 . the tropicalization of @xmath5 is obtained as follows .
we choose a gale dual @xmath338 of @xmath0 , @xmath380 note that the matroid given by the columns of @xmath0 is self - dual .
the bergman fan @xmath381 is a @xmath281-dimensional fan in @xmath382 , or , equivalently , a graph on the @xmath383-sphere .
we depict this graph in figure [ fig_b3x3 ] on the left .
it has nine vertices , corresponding to the six singletons @xmath384 and the three circuits @xmath385 , @xmath386 , @xmath387 .
its image under @xmath361 is the tropical discriminant @xmath8 , a @xmath281-dimensional fan in @xmath277-dimensional real space or , equivalently , a graph on the @xmath281-sphere .
we depict this graph in figure [ fig_b3x3 ] on the right .
the rays @xmath385 , @xmath386 , and @xmath387 of @xmath388 map into the relative interiors of the @xmath281-dimensional cones @xmath389 , @xmath390 , and @xmath391 of @xmath8 , respectively .
more precisely , the ray @xmath385 of @xmath388 is mapped to the negative of the image of the ray @xmath81 , which in turn is the dividing ray in the @xmath281-dimensional cone spanned by the images of the rays @xmath281 and @xmath383 .
the rays @xmath386 and @xmath387 are mapped similarly . since @xmath0 is non - defective , the tropical discriminant @xmath8 is the union of codimension @xmath81 cones in the normal fan of the newton polytope of @xmath6 . as @xmath8 is the @xmath81-skeleton of a triangular prism ,
we conclude that the newton polytope of @xmath6 is a bipyramid .
its five vertices correspond to the five terms in the determinant of @xmath236 .
returning to the general case , we note that the dimension of the image of @xmath347 is at most @xmath392 , so the dual variety @xmath5 is a proper subvariety of @xmath4 .
if the dimension of @xmath5 is less than @xmath393 , that is , if @xmath5 is not a hypersurface , we say that the toric variety @xmath3 and its point configuration @xmath0 , are _
defective_. in the non - defective case , there is a unique ( up to sign ) irreducible polynomial @xmath394 with integer coefficients which vanishes on @xmath5 .
the polynomial @xmath394 is the _
@xmath0-discriminant _ as defined in ( * ? ?
what follows , the dual variety @xmath5 itself will be referred to as the @xmath0-discriminant , even if @xmath0 is defective . by the bieri - groves theorem @xcite ,
the dimension of the @xmath0-discriminant @xmath5 coincides with the dimension of the tropical @xmath0-discriminant @xmath8 .
theorem [ thm : tropical ] furnishes a purely combinatorial formula for that dimension .
[ cor : dimension ] the dimension of the @xmath0-discriminant @xmath5 in @xmath4 is one less than the largest rank of any matrix @xmath395 where @xmath396 runs over @xmath397 .
here @xmath16 is the subset of @xmath19 defined in the introduction . that definition is now best understood using the matroid - theoretic concepts which we reviewed in the second half of section [ sec : monomials ] , where we take @xmath398 to be the @xmath399-matrix @xmath338 as in ( [ chooseuandv ] ) and , hence , @xmath215 to be the rank @xmath400 matroid associated with the rows of @xmath338 .
in fact , @xmath215 is the matroid dual to the matroid given by the columns of the @xmath401-matrix @xmath0 , and @xmath13 coincides with the lattice of flats @xmath402 .
the set @xmath16 corresponds to the facets of the flag complex @xmath267 . in light of theorem
[ threesubdivisions ] , one could reformulate corollary [ cor : dimension ] with @xmath66 ranging over the facets of the nested set complex @xmath403 or the bergman fan @xmath404 . we are now prepared to state and prove the general version of theorem [ thm : initial ] .
let @xmath405 denote the set of all proper chains of length @xmath406 in @xmath407 , where we identify flats of the matroid in @xmath14 with their incidence vectors , and hence the chains in @xmath405 with @xmath408-element subsets of @xmath19 .
equivalently , @xmath405 is the set of @xmath408-element subsets of the elements of @xmath409 .
we write @xmath410 for the initial ideal , with respect to some @xmath88 , of the homogeneous prime ideal @xmath411 of the @xmath0-discriminant @xmath5 .
[ thm : initial2 ] suppose that the @xmath0-discriminant @xmath5 has codimension @xmath84 and let @xmath412
. if @xmath23 is a generic vector in @xmath9 then the multiplicity of the initial monomial ideal @xmath413 along the prime @xmath414 equals @xmath415 where @xmath416 is the subset of @xmath405 consisting of all chains @xmath66 such that @xmath417 theorem [ thm : initial ] is the special case of theorem [ thm : initial2 ] when @xmath0 is non - defective , i.e. , @xmath418 and @xmath419 is the principal ideal generated by @xmath420 . in that case , the initial monomial ideal @xmath421 is generated by the initial monomial @xmath25 . _
proof of theorem [ thm : initial2 ] .
_ according to theorem [ thm : mult ] , the prime @xmath101 is associated to @xmath422 if and only if the polyhedral cone @xmath423 meets the tropicalization @xmath8 , which was described in theorem [ thm : tropical ] as @xmath424 . the collection of cones @xmath425 for @xmath426 forms a unimodular triangulation of the co - bergman fan @xmath427 .
this was proved by ardila and klivans @xcite , and we discussed it in theorem [ threesubdivisions ] , calling @xmath428 the flag fan @xmath267 of the matroid given by the rows of a gale dual @xmath338 of @xmath0 .
therefore , ( [ theyintersect ] ) characterizes when @xmath429 meets @xmath430 for some @xmath431 .
the multiplicity of this intersection is precisely the stated @xmath432-determinant .
this can be derived from remarks [ speyerdiss ] and [ intrmultyuv ] .
our degree formula for the @xmath0-discriminant can now be rephrased in the following manner which is more conceptual and geometric .
[ cor : mult ] a monomial prime @xmath101 is associated to @xmath433 if and only if the cone @xmath87 meets the fan @xmath424 .
the number of intersections , counted with multiplicity , is the multiplicity of @xmath421 along @xmath101 .
we start this section with a brief discussion of computational issues .
the formula for the extremal terms of @xmath6 in theorem [ thm : initial ] gives rise to a practical method for computing the newton polytope of the @xmath0-discriminant @xmath6 .
indeed , the co - bergman fan of the matrix @xmath0 can be computed efficiently by gluing local bergman fans , as explained in ( * ? ? ?
* algorithm 5.5 ) .
see examples 5.7 , 5.8 and 5.9 in @xcite for some non - trivial computations
. extending the software used for those computations , we wrote a maple program for evaluating the formula ( [ eq : degree1 ] ) in theorem [ thm : initial ] .
the input for our program consists of three positive integers @xmath434 , and a @xmath1-matrix @xmath0 which is assumed to be non - defective .
the output is a list of initial monomials @xmath25 of the @xmath0-discriminant @xmath6 , for @xmath435 randomly chosen vectors @xmath23 in @xmath436 .
our maple implementation is available upon request from any of the authors
. note that in case @xmath0 is non - defective , it is possible to recover @xmath437 up to constant from its newton polytope @xmath438 by solving a linear system of equations .
namely , consider a generic polynomial @xmath439 with exponents in the lattice points of @xmath438 .
imposing the condition that @xmath439 vanishes on the image of the rational parametrization of @xmath6 given in proposition [ prop : hdparametrization ] and lemma [ lem : connecttosec3 ] translates into a system of linear equations in the coefficients of @xmath439 whose solution space is one - dimensional .
[ mixeddiscriminant ] let @xmath440 and @xmath441 , and consider the matrix @xmath0 below .
the corresponding @xmath0-discriminant is the _
mixed discriminant _ of the two bivariate polynomials @xmath442 whose exponent vectors are read respectively from the first four and last four columns of the last two rows of the matrix .
the non - vanishing of @xmath6 expresses the condition that the intersection @xmath443 is transversal .
what follows is the output of our maple program on this input . on a fast workstation ,
our code takes about half a second to compute the co - bergman fan .
afterwards it takes about one second per initial monomial .
so the total running time for this matrix is about @xmath435 seconds , where @xmath435 is the number of iterations specified by the user : .... [ 1 1 1 1 0 0 0 0 ] [ ] [ 0 0 0 0 1 1 1 1 ] a : = [ ] [ 2 3 5 7 11 13 17 19 ] [ ] [ 19 17 13 11 7 5 3 2 ] computing the co - bergman fan of a .... done .
time elapsed = 0.340 the number of maximal cones in the co - bergman fan is 57 48 of these cones map to codimension one in the tropical discriminant .
what follows are 3 pairs of weight vectors and initial monomials : 28 35 35 28 [ 446 , 773 , 680 , 37 , 925 , 963 , 765 , 380 ] , x1 x4 x5 x8 34 29 2 39 22 [ 439 , 464 , 454 , 360 , 303 , 279 , 591 , 583 ] , x1 x4 x5 x6 x8 2 39 22 22 39 2 [ 801 , 447 , 685 , 447 , 765 , 775 , 358 , 498 ] , x1 x2 x4 x5 x7 x8 .... from this output we see that this @xmath0-discriminant is a polynomial of degree @xmath444 . we discuss the geometric meaning of this example in section [ sec : resultants ] .
gelfand , kapranov and zelevinsky @xcite established the relationship between the @xmath0-discriminant and the _ secondary fan _ of a point configuration .
the secondary fan parametrizes the regular polyhedral subdivisions of @xmath0 .
it is shown in @xcite that , in the non - defective case , the a - discriminant divides the principal @xmath0-determinant .
hence , the newton polytope of @xmath6 is a minkowski summand of the secondary polytope of @xmath0 , which in turn implies that the secondary fan @xmath445 is a refinement of the normal fan of the newton polytope of @xmath6 .
the tropical descriminant @xmath8 being the codimension 1-skeleton of the normal fan of the newton polytope of @xmath6 , we obtain that @xmath8 is a subfan of the secondary fan @xmath445 . for a given @xmath446 , the corresponding regular subdivision @xmath147 , i.e.
, the support cell in @xmath445 , is obtained as follows : lift the point configuration @xmath296 from @xmath295 into @xmath447 by extending with the coordinates of @xmath23 to @xmath448 in @xmath447 .
the cells of @xmath147 are the subsets of @xmath0 corresponding to the lower facets of the convex hull of the lifted point configuration .
in fact , we conjecture that membership in @xmath8 depends only on the regular subdivision specified by the vector @xmath23 , even in the defective case : [ conj : secfan ] for any point configuration @xmath0 , the tropical discriminant @xmath8 is a union of cones in the secondary fan @xmath445 .
another notion that arose in the work of gelfand , kapranov and zelevinsky is the notion of _ @xmath36-equivalence _ of regular triangulations of a non - defective point configuration : let @xmath147 and @xmath449 be two regular triangulations which are neighbors in the secondary fan @xmath445 .
this means that their cones in @xmath445 share a common face of codimension one .
we call @xmath147 and @xmath449 _ @xmath36-equivalent _ if they specify the same leading monomial of the @xmath0-discriminant , i.e. , @xmath450 .
the @xmath36-equivalence classes of regular triangulations of a point configuration @xmath0 define a partition of the set of maximal cones in the secondary fan @xmath445 .
the @xmath36-equivalence classes are in bijection with the connected components of the complement @xmath451 of the tropical discriminant .
we return to example [ ex : symm3x3 ] to illustrate the relation of the tropical discriminant and the secondary fan of a point configuration as well as the notion of @xmath36-equivalence .
[ ex : symm3x3rev ] let @xmath0 be the @xmath452-matrix in example [ ex : symm3x3 ] whose toric variety is the veronese surface in @xmath369 and whose tropical discriminant @xmath8 was depicted in figure [ fig_b3x3 ] .
we now identify @xmath8 as a subfan of the secondary fan @xmath445 .
both of these two - dimensional fans are drawn as planar graphs in figure [ fig_secfan3x3 ] .
the graph @xmath445 is dual to the three - dimensional associahedron , and its @xmath453 regions are labeled with the @xmath453 regular triangulations of the configuration @xmath0 .
the tropical discriminant @xmath8 is the subgraph which is indicated by solid lines .
edges of @xmath445 that do not belong to @xmath8 are dashed .
the @xmath453 regular triangulations of @xmath0 occur in five @xmath36-equivalence classes corresponding to the open cells in the complement of @xmath8 .
the magnification on the left in figure [ fig_secfan3x3 ] shows a portion of the ( dual ) secondary polytope of @xmath0 with corresponding polyhedral subdivisions of the 6-point configuration indicated next to the faces .
we present two more examples which illustrate the various natural fan structures on the tropical discriminant . in the non - defective case
, the fan structure inherited from the secondary fan refines the fan structure dual to the newton polytope of @xmath6 .
the following example shows that a proper subdivision can occur .
[ ex : k_4 ] let @xmath454 , and consider the non - defective configuration @xmath455 here , @xmath456 . modulo the row space of @xmath0 , the codimension 1-skeleton of the secondary fan is two - dimensional .
it is the the cone over a planar graph with eight vertices and @xmath457 edges .
the tropical discriminant
@xmath8 corresponds to the induced subgraph on six of the vertices , namely , the images of the six dimensional weights @xmath458 and @xmath459 . here , _ the secondary fan strictly refines the grbner fan on @xmath8_. the latter is combinatorially a complete graph @xmath279 with vertices @xmath460 and @xmath461 , while the former has the edges @xmath462 and @xmath463 subdivided by the vertices @xmath464 and @xmath465 , respectively . for defective configurations ,
the grbner fan structure can properly refine the secondary fan structure on @xmath8 , as the following example shows .
[ ex : bjsst ] we take @xmath466 , and consider the defective configuration @xmath467 the tropical discriminant @xmath8 is a @xmath287-dimensional fan in @xmath468 , regarded as a @xmath285-dimensional polyhedral complex .
combinatorially , it is an immersion of the complete tripartite hypergraph @xmath469 .
_ the grbner fan subdivision _ has @xmath470 maximal cones and it _ strictly refines the secondary fan subdivision _ which has only @xmath471 cones .
indeed , the vector @xmath472 lies in the relative interior of a maximal cone of the secondary fan subdivision which breaks into three maximal cones in the grbner fan subdivision .
this example was verified by applying the software gfan @xcite to the equations defining @xmath5 .
in particular , we found that @xmath473 is the codimension two primary ideal generated by the determinant of the @xmath363-matrix @xmath474 plus the square of the ideal of @xmath475-minors of the @xmath476-matrix @xmath477 .
there is yet another way to interpret the @xmath0-discriminant and its tropicalization . for @xmath478 under the conditions imposed above
, the @xmath0-discriminant is the closure of the set of points @xmath479 $ ] in @xmath4 such that the hypersurface in @xmath7 defined by the laurent polynomial @xmath480 has a singular point in @xmath7 . in the tropical world , for @xmath481 , the regular polyhedral subdivision @xmath147 of @xmath0 is geometrically dual to the tropical hypersurface defined by the tropical polynomial @xmath482 .
hence , we may consider such tropical hypersurface to be _ singular _ whenever the vector @xmath23 lies in the tropical discriminant @xmath8 . in this sense ,
our object of study in this article is the _ space of singular tropical hypersurfaces _ and theorem [ thm : initial ] gives a refined formula for the degree of that space .
the tropical discriminant @xmath8 is the polyhedral space which parametrizes all singular tropical hypersurfaces with fixed monomial support @xmath0 .
this relates our results to the celebrated work of mikhalkin @xcite on gromov - witten invariants . for @xmath483 ,
the @xmath0-discriminant @xmath5 parametrizes singular curves on the toric surface @xmath3 .
our formula for the degree of @xmath5 is consistent with the lattice paths count in @xcite for the number of nodal curves on @xmath3 of genus @xmath484 through @xmath485 fixed points , where @xmath439 is the number of interior points in the lattice polytope @xmath486
. it would be interesting to explore possible applications of our combinatorial approach to gromov - witten theory .
the work of gathmann and markwig @xcite offers an algebraic setting for such a study ( see also the recent preprint @xcite , which addressed this question after the first version of our paper ) .
one of the main applications of @xmath0-discriminants is the study of resultants in _
elimination theory_. the configurations @xmath0 which arise in elimination theory have a special combinatorial structure arising from the _ cayley trick_. see @xcite for a geometric introduction . based on the results of the earlier sections
, we here study the combinatorics and geometry of _ tropical resultants _ , and we generalize the positive degree formula for resultants in @xcite to resultant varieties of arbitrary codimension .
let @xmath487 be finite subsets of @xmath488 .
their _ cayley configuration _ is defined as @xmath489 where @xmath490 is the standard basis of @xmath491 . to be consistent with our notation in sections 15
, we can regard @xmath0 as a @xmath1-matrix with @xmath492 and @xmath493 . as in @xcite and @xcite ,
the cayley configuration @xmath0 represents the following system of @xmath217 laurent polynomial equations in @xmath62 unknowns : @xmath494 here @xmath495 are coordinates on @xmath496 and we use multi - index notation @xmath497 .
our earlier examples include the following cayley configurations : * in example [ mixeddiscriminant ] we have @xmath498 , and the system ( [ cayleysystem ] ) takes the form @xmath499 * in example [ ex : k_4 ] we have @xmath500 , @xmath501 , @xmath502 and @xmath503 .
the @xmath0-discriminant is the sylvester resultant @xmath504 * in example [ ex : bjsst ] we have @xmath505 , @xmath506 and @xmath507 , and the system ( [ cayleysystem ] ) consists of three quadratic equations in one unknown @xmath53 : @xmath508 the variety @xmath5 of all solvable systems of this form has codimension two .
returning to the general case , we say that the cayley configuration @xmath0 is _ essential _ if the minkowski sum @xmath509 has affine dimension at least @xmath510 for every subset @xmath511 of @xmath512 with @xmath513 .
the _ resultant variety _ of the cayley configuration @xmath0 is the zariski closure in @xmath4 of the set of all points @xmath514 whose corresponding system ( [ cayleysystem ] ) has a solution @xmath53 in @xmath496 .
the following result is a generalization of proposition 1.7 in ( * ? ?
* a ) and of proposition 5.1 in @xcite .
[ prop : cayley ] the resultant variety of any cayley configuration @xmath0 contains the @xmath0-discriminant @xmath5 . if @xmath515 and the cayley configuration @xmath0 is essential then the resultant variety and the @xmath0-discriminant coincide .
consider the hypersurface in @xmath516 defined by the equation @xmath517 if @xmath518 is a singular point on this hypersurface then @xmath519 is a solution to ( [ cayleysystem ] ) .
this proves the inclusion . if @xmath0 is essential then a linear algebra argument as in (
* a ) shows that every solution @xmath53 of ( [ cayleysystem ] ) arises in this way .
the hypothesis that @xmath0 be essential is necessary for the equality of the resultant variety and the @xmath0-discriminant , even when @xmath520 , the situation of classical elimination theory .
the following simple example illustrates the general behavior .
[ ex : nonessential ] let @xmath521 , @xmath522 , and @xmath523 @xmath524 .
the cayley matrix @xmath0 represents a toric @xmath383-fold @xmath3 in @xmath525 .
it is not essential since @xmath526 is one - dimensional .
the system ( [ cayleysystem ] ) equals @xmath527 the resultant variety has codimension one , with equation @xmath528 , but the @xmath0-discriminant @xmath5 has codimension three .
in fact , we have @xmath529 in this case .
let @xmath531 denote the incidence variety consisting of all pairs @xmath532 in @xmath533 such that ( [ cayleysystem ] ) holds .
let @xmath534 be the projection to the first factor . by proposition [ prop : cayley ] ,
the resultant variety @xmath5 coincides with the closure of @xmath535 .
looking at the second projection @xmath536 , which is surjective and whose fibers are linear spaces of dimension @xmath537 , we deduce that @xmath531 is irreducible and has dimension @xmath538 .
then , @xmath539 .
given a generic choice of coefficients @xmath24 for the first @xmath62 polynomials , it follows from the essential hypothesis and bernstein s theorem , that the first @xmath62 equations in ( [ cayleysystem ] ) have a common solution @xmath519 .
we can freely choose all but one of the coefficients of the last @xmath84 polynomials so that @xmath53 solves ( [ cayleysystem ] ) .
this implies that @xmath540 , and the lemma follows .
corollary [ cor : dimension ] asserts that there exists a chain @xmath541 of @xmath542-vectors representing the supports of vectors in the kernel of @xmath0 such that the rank of the matrix @xmath543 is precisely @xmath544 .
we present an explicit way of choosing such a chain . by performing row operations
, we can assume that each set @xmath545 contains the origin .
set @xmath546 .
let @xmath547 for @xmath548 such that @xmath549 are linearly independent .
such elements exist because the family of supports is essential .
now , for any other element @xmath550 in @xmath551 , we can find an element @xmath552 in @xmath553 with support corresponding to the origin in each @xmath545 for @xmath161 from @xmath81 to @xmath62 , union the variables corresponding to @xmath549 and @xmath554 , plus the origin of @xmath555 in case @xmath556 .
choose any such @xmath557 and let @xmath558 be the support of @xmath552 ; it will have @xmath559 non - zero coordinates .
add a new point @xmath560 in @xmath338 .
we can assume that the support of @xmath561 equals the union of their supports .
let @xmath562 be the associated support vector .
we continue in this manner , adding a new point in @xmath338 at a time , and considering a new element in the chain of support vectors , but avoiding to pick all of @xmath563 and all of each of @xmath564 .
this produces precisely @xmath565 vectors @xmath566 in @xmath16 .
it is straightforward to check that the rank of the submatrix of @xmath567 given by the first @xmath217 and the last @xmath568 columns has maximal rank @xmath569 .
note that this is just a @xmath542 matrix . adding the last @xmath62 columns of @xmath354 containing the information about the specific supports @xmath570 will increase the rank by @xmath62 , as a consequence of the fact that the family is essential .
therefore , the rank of the matrix @xmath543 is precisely @xmath571 .
we identify @xmath572 with the disjoint union of the sets @xmath573 .
thus a generic vector @xmath88 assigns a height to each point in any of the @xmath545 , and it defines a _ tight coherent mixed subdivision _
@xmath574 of the minkowski sum @xmath575 ( cf .
@xcite ) . when @xmath576 , the initial form with respect to @xmath23 of the mixed resultant is described in ( * ? ? ?
* theorem 2.1 ) in terms of the sum of volumes of suitable _ mixed cells _ of the tight coherent mixed decomposition ( tcmd ) induced by @xmath23 .
we next generalize this result to resultant varieties of arbitrary codimension @xmath84 . for a classical study of resultant ideals of dense homogeneous polynomials
we refer to @xcite .
[ thm : sec6 ] a codimension @xmath84 monomial prime @xmath577 is a minimal prime of the monomial ideal @xmath413 only if @xmath578 consists of one point each from @xmath84 of the @xmath555 . the multiplicity of @xmath101 is the total volume of all mixed cells in the tight coherent mixed subdivision @xmath579 which use the points of @xmath578 in their decomposition .
the resultant variety @xmath5 is irreducible ( by proposition [ prop : cayley ] ) , and it has codimension @xmath84 ( by lemma [ rescodimc ] ) .
this implies ( by ( * ? ? ?
* theorem 1 ) ) that every minimal prime of the initial monomial ideal @xmath413 has codimension @xmath84 .
let @xmath580 be such a minimal prime .
after relabeling we may assume that each @xmath581 in @xmath101 is a coefficient of one of the last @xmath84 laurent polynomials in ( [ cayleysystem ] ) .
the proof for the case @xmath576 is given in @xcite , and the proof for @xmath582 uses the same general technique .
we write @xmath583 for the @xmath161-th laurent polynomial in ( [ cayleysystem ] ) , but with @xmath24 replaced by @xmath584 .
let @xmath585 be the algebraic closure of the field of rational functions over @xmath586 in the coefficients of the first @xmath62 laurent polynomials @xmath587 , let @xmath588 denote the vector all coefficients of the last @xmath84 laurent polynomials @xmath589 , and consider the polynomial ring @xmath590 $ ] .
let @xmath591 denote the mixed volume of the newton polytopes of the polynomials @xmath587 .
then , by bernstein s theorem , the system @xmath592 has @xmath591 distinct roots @xmath593 in @xmath594 . for any @xmath595 ,
the ideal @xmath596 is generated by linear forms in @xmath590 $ ] .
the intersection of these ideals , @xmath597 , is an ideal of codimension @xmath84 and degree @xmath591 in @xmath590 $ ] .
geometrically , we obtain @xmath511 by embedding the prime ideal of @xmath5 into @xmath590 $ ] and then replacing @xmath24 by @xmath584 .
this is the higher codimension version of the _ product formula for resultants _ ( * ? ? ?
( 14 ) ) .
the ideal @xmath511 represents a flat family , and its special fiber @xmath598 at @xmath599 coincides with the special fiber of the image of @xmath421 in @xmath590 $ ] .
in particular , @xmath101 is an associated prime of @xmath598 , and it contains one of the ideals @xmath600 . since the generators of @xmath601 are @xmath84 linear forms in disjoint groups of unknowns @xmath602 , we see that @xmath101 contains one unknown from each group .
this proves the first statement in theorem [ thm : sec6 ] .
after relabeling we may assume that @xmath603 is a coefficient of @xmath604 for @xmath605 .
each root @xmath606 corresponds to a mixed cell @xmath607 in the tcmd of the small minkowski sum @xmath608 defined by the restriction of @xmath23 . by the genericity of @xmath23 , the mixed cell @xmath607 corresponds to a unique cell @xmath609 in the tcmd @xmath574 of the big minkowski sum @xmath610 , and every mixed cell of @xmath574
arises in this manner .
the reasoning above implies that the mixed cell @xmath609 uses the points of @xmath578 in its decomposition if and only if @xmath611 in @xmath590 $ ] .
this completes the proof . the first assertion in theorem [ thm : sec6 ]
can also be derived more easily , namely , from the fact that for any @xmath612-element subset @xmath511 of @xmath613 , the mixed resultant of the configurations @xmath614 vanishes on @xmath5 and only contains unknowns @xmath615 with @xmath616 . however , for the multiplicity count in the second assertion we need the `` product formula '' developed above .
theorem [ thm : sec6 ] has the following corollary .
let @xmath505 , @xmath506 and @xmath507 as in example [ ex : bjsst ] , and choose @xmath620 which represents the _ reverse lexicographic term order_. then @xmath621 this ideal has seven associated primes , of which three are minimal : @xmath622 , @xmath623 , and @xmath624 .
they correspond to the three mixed cells @xmath625 , @xmath626 and @xmath627 of the tcmd @xmath574 of @xmath628 .
each mixed cell has volume two , which implies that the degree of @xmath5 is @xmath629 .
let @xmath630 and @xmath631 and take the @xmath545 to be the four subtriangles of the square with vertices @xmath632 , @xmath542 , @xmath633 and @xmath634 . here
( [ cayleysystem ] ) is a system of four equations in two unknowns @xmath635 and @xmath636 which can be written in matrix form as @xmath637 the resultant variety @xmath638 has codimension @xmath281 and degree @xmath639 .
this is the sum over the mixed areas of the @xmath640 pairs of triangles .
the mixed area is @xmath281 for each such pair . using computer algebra , we find that the prime ideal of the resultant variety @xmath5 is generated by the @xmath641 determinant in ( [ eq:4by4 ] ) together with ten additional polynomials of degree six in the @xmath24 .
theorem [ thm : sec6 ] gives a combinatorial recipe for constructing all the initial monomial ideals of this prime ideal .
we conclude with a brief discussion of the _ tropical resultant _ @xmath8 .
the results in section [ sec : subdivisions ] characterize this polyhedral fan in terms of regular subdivisions of @xmath0 , and we will now rephrase this characterization in terms of coherent mixed subdivisions of @xmath642 .
theorem 5.1 in @xcite implies that every regular subdivision @xmath147 of @xmath0 corresponds uniquely to a coherent mixed subdivision ( cmd ) , which we denote by @xmath574 .
note that @xmath147 is a polyhedral cell complex of dimension @xmath643 while @xmath574 has only dimension @xmath62 , and @xmath147 is a triangulation if and only if @xmath574 is a tcmd .
every cell @xmath644 of a cmd @xmath574 decomposes uniquely as a minkowski sum @xmath645 , where @xmath646 for all @xmath161 .
we write @xmath647 for the corresponding cell of @xmath574 .
we say that the cell @xmath644 is _ fully mixed _ if each @xmath648 has affine dimension at least one .
using the techniques in the proof of theorem [ thm : sec6 ] , we can derive the following : specializing to the classical case @xmath576 , when @xmath5 is the hypersurface defined by the mixed resultant , we can now recover the combinatorial results in @xcite . in particular , the positive formula for the extreme monomials of the sparse resultant given in ( * ? ? ?
* theorem 2.1 ) can now be recovered as a special case of theorem [ thm : initial ] .
thus , one way to look at sections 15 in the present paper is that these extend all results in @xcite from essential cayley configurations with @xmath520 to arbitrary configurations @xmath0 .
a perspective on how this relates to the approach of @xcite is given by the points ( a),(b),(c),(d ) found at the end of the introduction in @xcite .
j. richter - gebert , b. sturmfels , t. theobald : _ first steps in tropical geometry _ ; in : `` idempotent mathematics and mathematical physics , '' proceedings vienna 2003 , ( g.l .
litvinov and v.p .
maslov , eds . ) , american math .
society , contemporary mathematics 377 ( 2005 ) , pp . | tropical geometry is used to develop a new approach to the theory of discriminants and resultants in the sense of gelfand , kapranov and zelevinsky .
the tropical @xmath0-discriminant is the tropicalization of the dual variety of the projective toric variety given by an integer matrix @xmath0 .
this tropical algebraic variety is shown to coincide with the minkowski sum of the row space of @xmath0 and the tropicalization of the kernel of @xmath0 .
this leads to an explicit positive formula for all the extreme monomials of any @xmath0-discriminant . |
recently discovered iron - based superconductors@xcite induce great interest in scientific community because of rather high @xmath8 , proximity to the spin - density wave state and multiband nature of electronic transport.@xcite@xmath10@xcite however , these materials also encourage potential technical applications due to high upper critical fields @xmath11 and critical current densities @xmath12.@xcite@xmath10yang h in the family of iron - based superconductors , fech ( ch = s , se , and te , fech-11 type ) materials have the simplest crystal structure , nearly isotropic high @xmath13 and rather high @xmath12,@xcite@xmath14@xcite but their relatively low @xmath8 impedes prospects for applications .
superconducting @xmath8 was raised up to about 32 k in a@xmath1fe@xmath2se@xmath3 ( a = k , rb , cs , and tl , fech-122 type ) iron selenide superconductors with rather high @xmath13 ( @xmath4 56 t for @xmath15 at 1.6 k).@xcite@xmath14@xcite preliminary results indicate that the @xmath12 of k@xmath1fe@xmath2se@xmath3 is much lower when compared to iron arsenides or binary fech-11 type iron selenides.karpinski@xmath14@xcite@xmath14@xcite@xmath14@xcite@xmath10@xcite post annealing and quenching treatment can induce metallic and superconducting state in as - grown and insulating k@xmath1fe@xmath2se@xmath3 crystals,@xcite yet current carrying characteristics of such materials are not known . in this work , we report on the significant enhancement of critical current density in k@xmath1fe@xmath2se@xmath3 single crystals obtained via post - annealing and quenching process .
we also give detailed insight into the vortex pinning mechanism .
main pinning sources are the 3d normal cores whereas dominant vortex interaction with pinning centers is via spatial variations in @xmath8 .
details of crystal growth and structure characterization were reported elsewhere.@xcite the as - grown crystals were sealed into pyrex tube under vacuum ( @xmath4 10@xmath16 pa ) .
the samples were annealed at 400 @xmath17c for 1h and quenched in the air as reported previously.@xcite .
crystals were claved and cut into rectangular bars .
magnetization measurements were performed in a quantum design magnetic property measurement system ( mpms - xl5 ) .
fe@xmath2se@xmath3 crystals taken in @xmath18 = 0.1 ( ac ) and 1 mt ( dc ) field , respectively .
( c ) magnetization hysteresis loops of as - grown and quenched samples at 1.8 k for @xmath15 .
( d ) superconducting critical current densities @xmath19 of as - grown and quenched samples . ]
calculated volume fractions from ac susceptibility at 1.8 k are rather similar , 75% for as grown and 88% for quenched crystal .
however , the quenched crystal shows a very steep transition at 31 k and saturates at about 10 k whereas for as - grown sample the diamagnetic signal increases gradually with slightly lower @xmath8 ( fig .
the single sharp peak of 4@xmath20 in quenched crystals ( fig .
1(a , b ) ) indicates more homogeneous superconducting state . the calculated volume fraction from dc susceptibility ( fig .
1(b ) ) significantly increased after quenching , consistent with previous results.@xcite hence , post - annealing and quenching process singnificantly advances superconducting volume fraction in quenched k@xmath1fe@xmath2se@xmath3 .
the small volume fraction estimated from the fc curve suggests possible strong magnetic flux pinning effects .
magnetic hyshteresis loops ( mhl ) of quenched sample are much bigger and more symmetric ( fig .
the pinning force is enhanced significantly and the bulk pinning is dominant when compared to the as - grown sample .
the mhl of as - grown crystal is small and asymmetric , suggesting that the surface barrier may be important.@xcite@xmath14@xcite moreover , there is no fishtail effect up to 5 t which has been observed in s - doped k@xmath1fe@xmath2se@xmath21s@xmath1 single crystal with s = 0.99 at low field and in feas-122 single crystals at high field.@xcite@xmath14@xcite@xmath14@xcite@xmath10@xcite the in - plane critical current density @xmath22 for a rectangularly - shaped crystal with dimension @xmath23 when @xmath15 is@xcite@xmath14@xcite @xmath24 where a and b ( @xmath25 ) are the in - plane sample size in cm , @xmath26 is the difference between the magnetization values for increasing and decreasing field at a particular applied field value ( measured in emu/@xmath27 ) , and @xmath22 is the critical current density in a/@xmath7 . as shown in fig .
1(d ) , the calculated @xmath28 for quenched sample from fig .
1(c ) is enhanced about 50 times when compared to as - grown sample .
this can not be simply ascribed to the improvement of the superconducting volume fraction , because the volume fraction of quenched crystal is only about 4 times larger than the volume fraction of the as - grown crystal .
critical current values in quenched crystal are higher than that in k@xmath1fe@xmath2se@xmath3 crystals grown using the one - step technique and are the highest known @xmath0 among fech-122 type materials.@xcite the quenched sample also exhibits better performance at high field .
the @xmath0 for quenched sample is still larger than 10@xmath6 a/@xmath7 at 4.8 t whereas for as - grown sample , it has decreased about one order of magnitude .
the @xmath29 is also larger than for k@xmath1fe@xmath2se@xmath30s@xmath31 with @xmath32 = 0.99.@xcite fe@xmath2se@xmath3 crystal for @xmath15 .
( b ) magnetic field and temperature dependencies of superconducting critical current densities @xmath19 for quenched k@xmath1fe@xmath2se@xmath3 crystal determined from mhls . ] the temperature dependent symmetric curves for all mhls imply that the bulk pinning dominates in the crystal at all temperatures .
the hysteresis area decreases with the temperature suggesting gradual decrease of @xmath0 as the temperature is increased ( fig .
the current carrying performance of quenched crystals is superior at all temperatures and fields when compared to crystals prepared using the one - step technique.@xcite ( @xmath33 ) at various temperatures .
solid line is the fitting curve using @xmath34 .
inset shows @xmath35 as a function of @xmath36 .
the fitting result using @xmath37 is shown as solid lines .
( b ) reduced temperature dependence of @xmath36(@xmath38 ) with the solid line standing for the fitting result obtained by using the ( 1-@xmath38)@xmath39 law .
( c ) reduced temperature dependence of the @xmath12(@xmath38 ) at zero field .
the dotted , dashed and solid lines show the @xmath40 , @xmath41 and the fitting result using @xmath42 , respectively ( see text ) .
the measured and estimated @xmath36 are shown as closed and open circles in inset of ( a ) and ( b ) . ] in order to explain the mechanism of flux pinning in quenched sample , we studied the temperature and field dependencies of the vortex pinning force @xmath43 .
based on the dew - huges model,@xcite if there is a dominant pinning mechanism then the normalized vortex pinning forces @xmath44 from different measurement temperatures should overlap and a scaling law of the form @xmath45 will be observed .
here @xmath33 is the reduced field @xmath46 and @xmath35 corresponds to the maximum pinning force .
the irreversibility field @xmath47 is the magnetic field where @xmath48 extrapolates to zero .
the indices @xmath49 and @xmath50 provide the information about the pinning mechanism . as shown in fig .
3(a ) , the normalized curves of @xmath51 for @xmath52 @xmath53 22 k present a temperature independent scaling law . using the scaling function @xmath54 , we estimate @xmath49 = 0.86(1 ) and @xmath50 = 1.83(2 ) , respectively . the value of @xmath55 ( @xmath56 ) @xmath57 0.32 is consistent with the peak positions ( @xmath58 0.33 ) of the experimental curves at different temperatures .
those values are close to expected values for core normal point - like pinning ( @xmath49 = 1 , @xmath50 = 2 , and @xmath55 = 0.33).@xcite moreover , for @xmath52 @xmath59 20 k , the @xmath60 can be estimated by @xmath35 location at @xmath61 = 0.33 .
partial @xmath51 curves measured between 10 and 20 k also exhibit the same scaling law , suggesting that core normal point - like pinning mechanism is dominant above 10 k. these point - like pinning center could come from the random distribution of fe vacancies after quenching , similar to feas-122 type materials.@xcite@xmath14@xcite@xmath14@xcite on the other hand , the @xmath35 obeys the @xmath62 scaling with @xmath63 1.67(1 ) ( inset of fig .
3(a ) ) , close to the theoretical value ( @xmath64 = 2 ) for the core normal point - like pinning.@xcite moreover , as shown in fig .
3(b ) , the temperature dependence of @xmath47 can be fitted by using @xmath65 where @xmath66 and we obtained @xmath67 = 1.21(1 ) , close to the characteristic value of 3d giant flux creep ( @xmath67 = 1.5).@xcite similar index has been observed in overdoped ba(fe@xmath68co@xmath1)@xmath3as@xmath3.@xcite given the presence of 3d core pinning in quenched k@xmath1fe@xmath2se@xmath3 single crystals , it is important to distinguish between the case of @xmath69 and @xmath70 pinnings . for type - ii superconductors ,
vortices interact with pinning centers either via the spatial variations in the @xmath71 ( `` @xmath9 pinning '' ) or by scattering of charge carriers with reduced mean free path @xmath72 near defects ( `` @xmath70 pinning'').@xcite these two pinning types have different temperature dependence and therefore result in different relationship between @xmath73 and @xmath66 in the single vortex - pinning regime ( low - field and zero - field regions ) . for @xmath69 pinning , @xmath74 while for @xmath70 pinning , @xmath75.@xcite as shown in fig .
3(c ) , the @xmath76 is between the two curves corresponding to @xmath9 and @xmath70 pinnings , respectively , but much closer and similar in shape to the @xmath9-pinning curve . using @xmath77 ,
the experimental data can be fitted very well with @xmath78 = 0.74(2 ) , suggesting that both @xmath9 and @xmath70 pinnings play roles in the quenched k@xmath1fe@xmath2se@xmath3 single crystals , but the former mechanism is dominant .
it also implies that the main pinning centers lead to the distribution of @xmath8 in their vicinity or even might be non - superconducting like y@xmath3o@xmath79 and y - cu - o precipitates in yba@xmath3cu@xmath79o@xmath80 thin films.@xcite even though the @xmath0 of quenched k@xmath1fe@xmath2se@xmath3 single crystals is still one or two order(s ) smaller than that of other iron pnictide superconductors,@xcite@xmath14@xcite@xmath10prozorov post - annealing and quenching technique is an effective way to increase the @xmath0 of k@xmath1fe@xmath2se@xmath3 .
in summary , we report giant increase in the @xmath0 of k@xmath1fe@xmath2se@xmath3 single crystals by post - annealing and quenching technique .
we demonstrate that quenched k@xmath1fe@xmath2se@xmath3 crystals carry the highest observed @xmath0 among fech-122 type materials and exhibit good performance at high field .
detailed analysis of vortex pinning mechanism points out to the presence of a 3d point - like normal core pinning in quenched samples .
moreover , the analysis of temperature dependence of @xmath0 at zero field indicates that the @xmath9 pinning is dominant at measured temperature range .
work at brookhaven is supported by the u.s .
doe under contract no .
de - ac02 - 98ch10886 and in part by the center for emergent superconductivity , an energy frontier research center funded by the u.s .
doe , office for basic energy science .
j. karpinski , n. d. zhigadlo , s. katrych , z. bukowski , p. moll , s. weyeneth , h. keller , r. puzniak , m. tortello , d. daghero , r. gonnelli , i. maggio - aprile , y. fasano , .
fischer , k. rogacki , and b. batlogg , physica c * 469 * , 370 ( 2009 ) .
a. yamamoto , j. jaroszynski , c. tarantini , l. balicas , j. jiang , a. gurevich , d. c. larbalestier , r. jin , a. s. sefat , m. a. mcguire , b. c. sales , d. k. christen , and d. mandrus , appl .
. lett . * 94 * , 062511 ( 2008 ) . | the critical current density @xmath0 of k@xmath1fe@xmath2se@xmath3 single crystals can be enhanced by more than one order of magnitude , up to @xmath4 2.1@xmath510@xmath6 a/@xmath7 by post annealing and quenching technique .
a scaling analysis reveals the universal behavior of the normalized pinning force as a function of the reduced - field for all temperatures , indicating the presence of a single vortex pinning mechanism .
the main pinning sources are three dimensional ( 3d ) point - like normal cores .
the dominant vortex interaction with pinning centers is via spatial variations in critical temperature @xmath8 ( `` @xmath9 pinning '' ) . |
the anti - correlation between the equivalent widths of broad emission lines and agn luminosity ( the `` baldwin effect '' ; hereinafter be ) was first discovered by baldwin ( 1977 ) for the civ @xmath01549 broad emission line in high redshift agns .
it was initially hoped to be able to use the effect to calibrate the agn luminosity to be able to use agns as cosmological standard candles ( baldwin et al .
1978 ) , but the large dispersion of this relationship rendered this impossible ( baldwin et al .
1989 ; zamorani et al . 1992 ) .
the be is now well established for nearly all broad emission lines and the slope of the be steepens with increasing ionization potential ( zheng & malkan 1993 ; dietrich et al .
several mechanisms have been proposed to explain this effect ( see shields 2007 for a review ) . among these
perhaps the most widely accepted one is that the ionizing continuum softens with increasing luminosity so there are relatively fewer ionizing photons for broad emission line formation in high - luminosity agns .
this model can reproduce the ionization energy
be slope relationship fairly well ( korista et al . 1998 ) and the assumption has observational support ( binette et al .
1989 ; zheng & malkan 1993 ; wang & lu 1998 ; korista et al .
1998 ) . some theoretical models ( netzer 1985 , 1987 ; netzer , laor , & gondhalekar 1992 ;
wandel 1999a , b ) could produce a softer ionizing spectrum in high - luminosity sources , but the standard thin disc model ( shakura & sunyaev 1973 ) they adopt suffers from many problems ( antonucci 2002 ; gaskell & klimek 2003 ; gaskell 2008 ; lawrence 2012 ; antonucci 2012 ) . in this sense , the explanation of broad line be is still elusive and controversial .
an outlier of the ionization energy
be slope relationship is nv @xmath01240 , which has an ionization energy of 97.7ev but shows no be at all ( dietrich et al .
it has been proposed that this can be explained by a dependence of metallicity on agn luminosity , which in turn is a combination of eddington ratio ( @xmath2 ) and black hole mass ( korista et al .
1998 ; hamann & ferland 1993 , 1999 ; dietrich et al . 1999 ; dietrich & wilhelm - erkens 2000 ) . in local galaxies ,
gas metallicity correlates well with the mass of galaxies ( tremonti et al .
2004 ) , and more massive galaxies have larger black hole masses ( @xmath3 ) according to the @xmath3-@xmath4 relationship . for type1 agn population , the brighter agns would have higher @xmath3on average ( kollmeier et al .
2006 ; steinhardt & elvis 2010a , b ; lusso et al .
2012 ) and thus higher metallicities .
recent studies of the be for civ , mgii and feii , however , challenge this picture by showing that the be might instead be driven by the ew s correlation with eddington ratio ( baskin & laor 2004 ; bachev et al . 2004 ; warner , hamann & dietrich 2004 ; zhou et al .
2006 ; dong et al . 2009a , b ) or with @xmath3(netzer , laor & gondhalekar 1992 ; wandel , peterson & malkan 1999 ; shields 2007 ; kovacevic et al .
2011 ) .
the be of narrow emission lines is much less well studied , and some results are still controversial .
steiner ( 1981 ) discovered a strong be for [ oiii ] @xmath05007 ( especially in agns with strong optical feii ) and wills et al .
( 1993 ) found the narrow lines in high - luminosity agns are very weak .
meanwhile , croom et al . (
2002 ) found a significant be for [ nev ] @xmath03425 and [ oii ] @xmath03727 but obtained a null result for [ neiii ] @xmath03870 and [ oiii ] @xmath05007 using the 2df sample .
subsequent work shows , however , that [ oiii ] @xmath05007 does show a be ( dietrich et al .
2002 , netzer et al .
hnig et al .
( 2008 ) and keremedjiev et al .
( 2009 ) , used spitzer data to find that the bes of different mid - ir narrow emission lines have a nearly constant slope .
several mechanisms have been proposed to explain the nlr be .
nlrs follow a size luminosity relationship ( schmitt et al . 2003b ; bennert et al .
2002 ; greene et al .
the size of the nlr may grow beyond the size of the host galaxy in high - luminosity agns , thus turning from ionization - bounded to matter - bounded , and so producing the be ( croom et al .
2002 ) . as the ew is proportional to the covering factor ( cf ) and
it is found that the cf contributes to much of the variance of the ew ( baskin & laor 2005 ) , a luminosity- dependent cf is also a possible cause of the be ( shields et al .
1995 ; stern & laor 2012b ; stern & laor 2012c ) . it has recently been proposed that the ew of narrow emission lines is dependent on the inclination to the accretion disc ( risaliti et al .
2011 ) , at least in the highest ew sources . in principle
, this is a potential cause of the be too . to make progress in understanding the nlr
be , we need to measure narrow emission lines of a wide range of ionization potentials and critical densities , and determine the nature of their bes more accurately . in this paper
, we use a well - defined sample drawn from sdss dr4 and employ the technique of composite spectra to investigate the be for prominent narrow emission lines in the optical band and to study the origin of the nlr be .
we use a cosmology with @xmath5 = 70 kms@xmath6mpc@xmath6 , @xmath7 = 0.3 , and @xmath8 = 0.7 throughout this paper .
we need accurate emission line and continuum measurements to ensure reliable determinations of equivalent widths and continuum luminosities . from the spectral data set of the sloan digital sky survey
fourth data release ( adelman - mccarthy et al . 2006 ) , we have selected 4178 seyfert 1 galaxies and quasars ( i.e. , type 1 agns ) as described in dong et al .
we apply a redshift cutoff of 0.8 so that the redshift of the spectrum can be accurately determined using [ oiii ] @xmath05007 . to ensure high - quality spectra we require a median signal - to - noise ratio ( s / n ) of @xmath9 per pixel in the optical . to minimize the host - galaxy contamination ( see the appendix of dong et al .
2011 ) , we restrict the weak stellar absorption features , such that the rest - frame ews of cak ( 3934 ) , cah + h@xmath10 ( 3970 ) , and h@xmath11 ( 4102 ) absorption features are undetected at @xmath12 significance .
this criterion ensures that the host contamination is less than 10% around 4200(dong et al .
one may note that dong et al .
( 2011 ) do not consider very young stellar population ( i.e. , emission line galaxies ) .
but previous analysis have shown that in massive galaxies , the optical continuum is not dominated by very young stellar population although the uv continuum may be ( schawinski et al .
2007 ) . after removing duplications and sources with too many bad pixels in the h@xmath13 + [ oiii ] region
, we obtain 4178 type 1 agns . including radio - loud agns
may influence the measurement of ew of emission thus produce false effect for two reasons : firstly , the jet may interact with the ism to enhance the narrow line emission ( labiano et al .
2007 ) and secondly , it might also enhance the continuum through beaming if the jet points close to our line of sight . by matching with the first catalog ( becker et al . 1995 ) using the method of lu et al .
( 2010 ) , we reject 499 radio - loud agns so that our final sample consists of 3677 sources .
we give a brief description of our spectrum fitting process here ; the details can be found in dong et al .
( 2011 ) . to model the spectrum
, we used a code based on the mpfit package ( markwardt 2009 ) and fit the agn featureless continuum , the feii multiplets , and other emission lines simultaneously .
the agn continuum is represented locally by a power - law , for the region of 42005600 and for the h@xmath1 region ( if present ) .
the feii template by vron - cetty et al .
( 2004 ) that we use , is constructed using the identification and measurement of feii lines in i zw 1 .
it has two separate sets of templates in _ analytical _ forms , one for the broad - line system and the other for the narrow - line system . within each system ,
the relative velocity shifts and relative strength are assumed to be the same as those in i zw 1 .
broad feii lines share the same profile as broad h@xmath13 , while each narrow feii lines is modeled with a gaussian . during the fitting ,
the normalization and redshift of each system are taken as free parameters .
the broad balmer lines are fitted with as many gaussians as is statistically justified .
all narrow emission lines , except for the [ oiii]@xmath144959 , 5007 doublet lines , are fitted with a single gaussian .
each line of the [ oiii ] doublet is modeled with two gaussians , one accounting for the line core and the other for a possible blue wing , as seen in many objects .
since the sources in our sample do not suffer from significant host galaxy contamination , we do not apply starlight corrections to individual spectra . for each source
, we use @xmath15 and fwhm of broad h@xmath13 line to obtain the @xmath3 using the virial mass estimates by the dibai method ( dibai 1977 ) using the formalism of wang et al .
the typical statistical scatter about the @xmath3 obtained by reverberation - mapping is about 0.4 dex , and it may also be subject to more systematic errors . ( krolik 2001 ; collin et al .
2006 ; shen et al .
2008 ; fine et al . 2008 ; marconi et al . 2008
; denney et al . 2009 ; rafiee & hall 2011a ; steinhardt 2011 ) .
the error of @xmath2 is of similar magnitude as that of @xmath3 . a convenient way to explore the correlations between ews and other parameters like
luminosity is to make composite spectra for different parameter bins .
we normalize individual spectrum to the mean flux around 4200and then construct the geometric composite as vanden berk et al .
( 2001 ) . to get an accurate continuum measurement
, we need to subtract the broad emission lines , especially , h@xmath1 and h@xmath13 and feiiemission from the spectrum . for the broad component of h@xmath13 and feii we subtract them from the original spectrum ; but for the broad component of h@xmath1 , we leave it un - subtracted when making composite spectra . this is because the broad h@xmath1 is highly blended with [ nii ] @xmath06583 and [ nii ] @xmath06548 , so deblending in individual spectrum is not reliable while deconvolving the blends in the high sn composite spectrum is easier . the fitting algorithm used to model
the composite spectrum is the same as described in section 2.2 .
more specifically , we fit the blends using one gaussian for the narrow component of h@xmath1 , [ nii ] @xmath06583 and [ nii ] @xmath06548 , two gaussians for [ sii ] @xmath166717 , 6731 and three for h@xmath17 .
the broad - line - subtracted spectra are shown in fig . 1 .
the embedded panels show the fitting result for the h@xmath17 + [ nii ] blends .
first , we want to explore if the be exists in the prominent narrow lines [ nev ] @xmath03425 , [ oii ] @xmath03727 , [ neiii ] @xmath03870 , h@xmath18 , [ oiii ] @xmath05007 , [ oi]@xmath06300 , h@xmath19 , [ nii ] @xmath06583and [ sii ] @xmath166717 , 6731 . measuring narrow lines in individual spectrum
are subject to the s@xmath20n limit and the ew-@xmath21 relation of all our lines shows a large dispersion ( typical 0.2 dex , this can be seen clearly in the middle panel of fig . 4 ) , so we turn to composite spectra for reliable measurements .
we divide our sample into intervals of 0.3 dex in @xmath22 luminosity starting at log @xmath21 = 43.2 to 45.9 @xmath23 . and make a composite spectrum in each luminosity bin as described in section 2.3 . from fig.1
we can see two clear and remarkable results : * with increasing luminosity , the narrow lines vanish .
* the slope of the observed continuum become bluer with increasing luminosity . to see the dependencies of the ews on the luminosity more clearly , in fig . 2 we plot the log ew that derived from composite spectra for each luminosity bin for each of the lines listed above against log @xmath21 in order to get more qualitative results .
for each line , we show a weighted linear regression of the logarithm of the ew on the logarithm of the luminosity and we list all the be slopes of the narrow emission lines together with their ionization energies and critical densities in table.1 . when fitting the relationship , we add an error to each data point that would help to reduce the normalized @xmath24 to @xmath251 .
these added errors are the same for individual narrow line , and they account for the potential other systematic errors .
the be slopes of the high - ionization lines [ neiii ] @xmath03870 and [ nev ] @xmath03425 are @xmath26 , @xmath27 respectively which are consistent with the results of keremedjiev et al .
( 2009 ) who obtained @xmath28 , and @xmath29 respectively for the neiii @xmath0 15.56@xmath30 m , and nev @xmath0 14.32@xmath30 m lines .
our [ oiii ] @xmath05007 be slope of @xmath31 is similar to the values of @xmath32 found by kovacevic et al .
( 2011 ) and steeper than the @xmath33 found by dietrich et al .
the difference might be due to different sample we employ .
the error of slopes we give here include only the statistic error of the fitting , but do not include the measuring error and intrinsic dispersion of the ew , which may reach 0.2 dex typically .
in addition to finding the be in high - ionization lines , we also find the be in the low - ionization lines of [ oi]@xmath06300 [ sii ] @xmath166717 , 6731 and [ oii ] @xmath03727 .
the narrow recombination line : h@xmath18 also shows a be with a similar slope as for the forbidden lines . for comparison , the _ broad _
h@xmath13 line shows an inverse be with a slope of 0.16 for our sample .
this is consistent with greene & ho ( 2005 ) and croom et al .
( 2002 ) who reported @xmath34 and @xmath35 respectively .
we can see in fig . 2 that the ew of the narrow component of h@xmath1 increases with luminosity too .
these indicate that the behaviors of broad lines and narrow lines are different .
the correlation coefficients between ew([oiii ] ) and @xmath15 , z , @xmath3 , @xmath2 are -0.12 , -0.05 , -0.01 , -0.21 , respectively .
these are similar to the results in zhang et al .
( 2011 ) . for our flux - limited sample ,
the redshift , luminosity , black hole mass , and eddington ratio are correlated with each other .
the correlation coefficients between @xmath15 and @xmath36 , @xmath3 , @xmath2 are 0.81 , 0.67 , 0.15 respectively . after controlling for @xmath36 or @xmath2 in a partial correlation analysis , ew([oiii ] )
correlates with @xmath15 with @xmath37=-0.14 and -0.16 .
this indicates that the be of narrow lines is not a secondary effect of correlations between the ew and redshift or @xmath2but an independent phenomenon .
we note that as the correlation strength is not strong , one possibility is that there are other factors that regulate the ew as discussed in zhang et al .
( 2011 ) . also , measurement error in ew([oiii ] ) may act to smear the correlation .
the measurement error could be introduced by limited s@xmath20n and fitting process . in spite of these uncertainties ,
the be of narrow lines does exist and may shed light on physical process in agns as explored in detail below .
a dependence of the be slope on ionization energy is found for broad emission lines in agns ( see introduction ) and this is the most compelling evidence for a luminosity - dependent ionizing spectrum .
the nlr lies far from the nucleus , has a complex geometry and may contain dust ( netzer & laor 1993 ; tomono et al .
2001 ; radomski et al .
2003 ; schweitzer et al .
this makes the response of the narrow - line flux to changes in the sed complicated .
we plot the slope of be against the ionization energies of the narrow lines in the left panel of fig . 3 .
we also plot the keremedjiev et al .
( 2009 ) data for [ siv ] 10.51@xmath30m:@xmath38 , [ neii ] 12.81@xmath30m:@xmath39 , [ neiii ] 15.56@xmath30m:@xmath28 and [ nev ] 14.32@xmath30m:@xmath29 in purple crosses and the broad - line be from dietrich et al .
( 2002 ) in blue rectangles for comparison .
we can see that the narrow - line be slopes do not correlate with the ionization - energy ( @xmath40 meaning the two - sided probability that a correlation is not present is 78% . ) but cluster around -0.2 with a dispersion of @xmath41 .
[ nii ] @xmath06583 , h@xmath19 and [ oii ] @xmath03727 have slopes of @xmath42 , @xmath43 , and @xmath44 so they deviate from -0.2 significantly .
possible reasons for this are discussed in section 4.1 .
it is well known that the nlr is stratified that the high - ionization lines rise from the inner part of the nlr while low - ionization lines rise further away ( see , for example , veilleux et al .
1991 ; robinson et al .
1994 ; bennert et al .
2006a , b ; kraemer et al . 2009 ) .
the central electron temperature , density , and ionization parameter are , in general , higher in seyfert 1s than in seyfert 2s ( gaskell 1984 ; schmitt 1998 ; bennert et al .
the lines of lower critical density are not as strong near the nuclei as the higher critical density lines .
a possible explanation of this is that the change in critical density marks a transition between difference types of clouds . to see
if this is a factor in the nlr be we therefore plot the ews against critical density in the right panel of fig .
2 using the same symbols of panel ( a ) . we again fail to find any correlation in this plot ( @xmath45 using a spearman rank correlation analysis ) .
using our sample of 3677 radio - quiet agns from sdss dr4 that have little contamination by host galaxy , we find that , in contrast with the be of broad emission lines , the narrow lines show a nearly constant be slope irrespective of ionization potential or critical density .
the [ oii ] @xmath03727 , [ nii ] @xmath06583 and h@xmath19 lines show significantly different be slopes from other lines .
although the slope of different lines cluster around -0.2 , we do find some lines that show large deviations .
firstly , [ nii ] @xmath06583 show a much flatter slope ( @xmath42 ) than other lines ( @xmath46 5% in two - sided ks test , meaning if the two samples are drawn from the same distribution , we expect to see a difference as large or larger than we see here only 5 in 100 times . ) . this flattening was also found for nv @xmath01240 , which has an ionization potential of 77.7 ev but no be .
one plausible explanation of a weaker or absent be for the nitrogen lines is that the increase of metallicity in brighter agn will enhance nitrogen abundance relative to carbon , oxygen etc .
, because nitrogen is a secondary element whose abundance scale as @xmath47 .
this trend could compensate for the decrease of ew in higher - luminosity sources thus making the be slope flatter .
this explanation is supported by a number of recent studies .
the metallicities of the blr ( hamann & ferland 1993 ; nagao et al . 2006a ; juarez et al .
2009 ) and nlr ( e.g. , nagao et al . 2006b
; matsuoka et al . 2009 ) are both found to correlate with the luminosity of the agn .
so in principal , the @xmath48l relationship could produce the flattening of the be of [ nii ] @xmath06583 we observe here .
however , this explanation has a major problem because niv ] and niii ] , which would be expected to deviate from the ionization energy
slope relationship for the same reason , lie on it ( dietrich et al .
2002 ) . despite this serious problem to be resolved , occam s razor suggests that an enhanced abundance in high - luminosity agns is the simplest and most plausible explanation of the flatter be slope of [ nii ] @xmath06583 .
the second significant deviation from a be slope of -0.2 is the steep slopes of h@xmath19 and [ oii ] @xmath03727 . both of these are star formation ( sf ) indicators in star - forming galaxies ( kenicutt et al .
1998 , ho 2005 ) and their ew may reach several hundred in starburst galaxies .
so even though the agns we selected show no absorption feature in the continuum , the emission lines are still possible subject to sf contamination because of their large ew . in low - luminosity sources
whose continuum is low , the emission line from sf may contribute significantly to the total line flux , and thus enhance the ew . in type-2 agns , the sf contribution of h@xmath19 is estimated to be more than 60% ( brinchmann et al .
the h@xmath1 emitting region of type-1 agns , however , may have a higher fraction of the emission originating with the agn ( zhang et al .
2008 ) but still be heavily influenced by sf .
furthermore , a variety of studies have established a correlation between the strength of agn activity and star formation in the local universe ( e.g. , rowan - robinson 1995 ; croom et al . 2002 ; netzer et al .
2007 , 2009 ; shao et al . 2010 ) .
the steepest correlation is @xmath49 ( netzer 2009 ) meaning @xmath50 , a be slope steeper than -0.2 .
so it makes sense that the sf in host galaxies would steepen the be slopes of h@xmath19 and [ oii ] @xmath03727 .
however , the [ oii ] @xmath03727 and h@xmath19 bes can not be attributed entirely to sf agn relationship . for a plausible range of ionization parameters , densities , and ionizing spectra ,
the intensity of [ oii ] @xmath03727 is proportional to that of [ oiii ] @xmath05007 ( 10% - 30% , ferland & osterbrock 1986 ; ho et al .
1993a , 1993b ) . since [ oiii ]
@xmath05007shows a significant be , [ oii ] @xmath03727 is unlikely to show a radically different trend .
so it is safe to conclude that the [ oii ] @xmath03727 and h@xmath19 be could be partly ( but not totally ) produced by sf contamination . a third effect we can see in fig .
1 is the optical continuum getting bluer towards higher luminosity .
this would leverage the continuum and lower the ew if the line flux remains unchanged .
this could arise if the nlr sees a filtered spectral - energy distribution ( sed ) ( kraemer et al .
1998 ; groves et al .
2004a , b ) so that a change in the ionizing spectrum of an agn would not change the sed the nlr sees much .
the [ nev ] @xmath03425 , [ oii ] @xmath03727 and [ neiii ] @xmath03870 lines are most likely to be influenced by a continuum shape difference effect because of their shorter wavelength .
a possible explanation of the bluer color of the uv - optical sed is the anisotropy of the continuum emission . because the accretion disc is optically thick , we will see a dimmer continuum when viewing it edge - on .
( this is the combined result of the @xmath51 projection effect and the disk equivalent of `` limb darkening '' . )
we could also preferentially be seeing the inner , high - temperature part of the disc when viewing face - on .
these effects will combine to give a higher observed luminosity with a face - on viewing angle .
the reddening of spectrum depends on viewing angle too ( keel 1980 ; de zotti & gaskell 1985 ; zhang et al . 2008 ) .
gaskell et al .
( 2004 ) used radio orientations to get agn reddening curves and found the continuum shape is profoundly affected by reddening for all but the bluest agns . because of these effects sources with small inclination would both show large luminosity and have low reddening .
an alternative explanation of continuum shape difference is host galaxy contamination .
shen et al . (
2011 ) made composite spectra of different @xmath15 luminosity bins and found that the uv part of all the composite spectra are similar while the optical part flatten with decreasing luminosity .
they interpreted this trend as due to host galaxy contamination in the optical region of the spectrum in low - luminosity agns .
similar argument is given by stern & laor et al .
( 2012a ) .
even though we have rejected objects with significant stellar light contributions , weak absorption lines can be spotted on the final composite spectrum of the lowest luminosity bin because of its extremely high signal to noise ratio .
our data can not distinguish different mechanisms that give rise to the continuum shape difference , and this is beyond the scope of this paper .
after correcting the deviations listed above , our conclusion that the slopes of be for different narrow lines are nearly constant is further strengthened .
it has been argued on both observational and theoretical grounds that there is a softening of the ionizing continuum with increasing luminosity ( see , for example , binette et al . 1989 ; netzer , laor , & gondhalekar 1992 ; zheng & malkan 1993 ; wang & lu 1998 ; korista et al .
if this is indeed the case , an important prediction of this is the ionization potential be slope relationship ( korista et al .
it is successful in explaining the broad line be , but a major failing is that it fails to explain the constant be slope of narrow lines .
it is already known that the nlrs of different agns are similar in the sense that the line ratios show less than 0.5 dex difference from object to object .
( koski 1978 ; veilleux & osterbrock 1987 ; veilleux 1991a , 1991b , 1991c ; veron - cetty & veron 2000 ; dopita et al . 2002 ; gorjian et al .
kraemer et al . ( 2000 ) , dopita et al . (
2002 ) and groves et al .
( 2004a , b ) proposed a nlr model where dust regulates the incident ionizing spectrum so as to keep the ionization parameter in the nlr constant . in this model ,
variation of the sed is filtered by dust in the nlr and thus the continuum shape seen by the nlr is dominated by the effect of the dust rather than by intrinsic changes in the sed .
this leads to more stable conditions in the nlr .
this could be the cause of the lack of a dependence of the slope of the be on the ionization energy .
a luminosity - dependent covering factor is a natural explanation of both a blr and nlr be .
however , to explain the ionization dependence , the variation in covering factor must have a different dependence for lines of different ionizations .
this would not be a surprise for the blr since there is strong radial ionization stratification and the highest ionization lines are an order of magnitude closer to the black hole than the lowest ionization lines ( see gaskell 2009 for a review ) .
shields , ferland , & peterson ( 1995 ) suggested that the broad - line be could be caused by a luminosity - dependent covering factor for clouds that are optically - thin to photons with wavelength less than 912 . the optically - thin clouds which have small column densities , preferentially emit high - ionization lines .
however , comparison of line profiles shows that the blr be is due to changes in the _ low velocity _ blr gas ( francis et al .
1992 ) rather than changes in the very broad component .
furthermore , snedden & gaskell ( 2007 ) argued that optically - thin gas does _ not _ make a substantial contribution to the blr .
nevertheless , given the strong radial ionization stratification of the blr , and the decreasing covering factor with ionization ( e.g. , francis et al . 1992 ) , an additional luminosity - dependence of the covering factor could explain the be and why the high - ionization lines should show a steeper slope than low - ionization lines . while this could be consistent with the blr , it is not obvious how such an explanation can be reconciled with the lack of an ionization - dependence we find here for the be .
the nlr size is correlated with the luminosity of an agn as @xmath52 ( bennert et al .
2002 ; schmitt et al . 2003b ; bennert et al .
2006a , b ; greene et al . 2011 ) .
so in high - luminosity agns , the nlr may possibly turn from ionization - bounded to matter - bounded , and the luminosity of narrow lines would cease increasing with agn luminosity .
this model could in principle explain part of the be of nlr ( croom et al .
2002 ) but it would be too rash a conclusion that it is the origin of the be because the bulk of the nlr emission originates from within the central few tens or hundreds of pc ( schmitt et al .
2003a ) so is unlikely to exceed the scale of the galactic bulge .
besides , the nlr has no natural size because it has no definite edge , so some cut in surface brightness or line - ratio is needed to define the size ( schmitt et al .
2003b ; bennert et al .
2006a , b ; greene et al . 2011 ) .
this makes the interpretation of size complicated .
netzer et al .
( 2004 ) argue that that @xmath52 is theoretically sound yet this relationship must break down for @xmath53 exceeding a few kpc .
they found that high-@xmath36 agns have nlr sizes no larger than 10kpc .
it is also well known that the the nlr is stratified so that the high - ionization lines rise from the inner part of nlr while low - ionization lines from further away ( veilleux et al .
1995 ; bennert et al .
2006a , b ) .
so the `` disappearing nlr effect '' , if exists , could preferentially influence low - ionization and low - critical density lines .
this again stands against the constant slope we find .
because different lines all seem to share a similar be slope , it is natural to think that it might be the continuum , rather than the emission lines , which is the cause .
jiang et al . ( 2006 ) proposed that the continuum variation can produce a weak be ( @xmath54 ) if the light - crossing time for the region emitting narrow @xmath55 exceeds the variability timescale for the x - ray continuum and the amplitude of variability anti - correlates with the luminosity .
shu et al . ( 2012 ) found a strong anti - correlation between the ew of the narrow @xmath55 line and @xmath56 ( @xmath57 ( where @xmath58 means time - averaged values ) consistent with the x - ray be expected in an individual agn if the narrow - line flux remains constant while the continuum varies . for a nlr
whose light - crossing time is about @xmath59 yr , the narrow - line luminosity can safely be assumed to be constant .
the amplitude of variability of the ionizing continuum can be obtained from the monitoring of agns after correction for the constant host galaxy light contribution .
as is well known , the amplitude of variability of an agn increases with time .
this can be most readily seen from the structure functions of agns ( the variance as a function of time interval between observations ) .
we are interested in variability on the longest timescales ( timescales similar to the nlr light - crossing time ) .
we do not , of course , have monitoring on such long timescales , but we can estimate the amplitude from structure functions .
cid fernandez et al .
( 2000 ) have presented structure functions for bright agns .
these are all consistent with the amplitude of variability rising as the time interval increases to a characteristic time of a year or a few years and then remaining constant .
collier & peterson ( 2001 ) got a similar result for lower - luminosity agns ( but with a shorter characteristic time ) and also find that the forms of the uv structure functions are similar to the optical ones .
there are indications that the structure function increases gradually on much longer timescales , but data on this part of the structure function are limited . observed variability thus gives a lower limit to variability on the light - crossing timescale of the nlr . in order to minimize the host galaxy light contribution one needs to go to as short a wavelength as possible
this means using @xmath60-band observations or space ultraviolet observations .
for example , for ngc 4151 merkulova ( 2006 ) found an _ observed _ peak - to - peak @xmath60-band amplitude of a factor of 7.5 .
uv variability can be a lot larger .
for example , the peak - to - peak amplitude at 1300 for fairall 9 is a factor of 25 ( koratkar & gaskell 1989 ) .
inspection of the results of long - term uv monitoring of a number of agns with the _ iue _ satellite ( koratkar & gaskell 1989 , koratkar & gaskell 1991a , b , clavel et al .
1991 , and obrien et al .
1998 ) give a median uv peak - to - peak variability of a factor of 7 , but the upper quartile is a factor of 13 . after correction for host - galaxy light , the amplitudes of optical variability are similarly large .
for example , the ngc 4151 photometry of lyuty & doroshenko ( 1999 ) gives a peak - to - peak amplitude of a factor of about 25 to 30 .
continuum anisotropy ( wang & lu 1999 ) can produce an effect similar to that caused by the actual variability of the continuum .
it was recently found by risaliti et al .
( 2011 ) that the optically - thick disc emission responsible for the continuum and isotropic [ oiii]emission will produce the ew([oiii ] ) distribution very well .
variability and continuum anisotropy are , in practice , indistinguishable , so we can consider them together .
we made monte - carlo simulations to explore whether continuum variability can explain the be slope . in a flux - limited survey
, the luminosity will show an approximately log - normal distribution .
we generated an artificial sample of sources with a similar distribution of @xmath15 as the observed agn sample ( 0.35 dex here ) .
we set the ew of the artificial - sources to be the mean ew of the whole sample , and we add a gaussian of 0.22 dex to the ew to account for the intrinsic dispersion .
uv and optical agn variability is approximately log - normal
i.e. , it looks normal when plotted in magnitudes ( see gaskell 2004 for a discussion of log - normal variation of agns ) .
the continuum variation was simulated by adding a gaussian with @xmath61 ( 0.18 dex ) to the continuum while the emission line flux was kept unchanged .
this is an over - simplification , but it can help us to gain some insight into the effect that continuum variability brings .
we generated 4000 sources in each round and measured the slope of be using the 4000 sources .
an example of the simulation is shown in fig .
4 . with a peak - to - peak continuum variation of a factor of three
, the simulation can produce a be slope of @xmath62 while a factor of six variation produces a be slope of @xmath63 .
the simulated distribution of ew-@xmath15 can be seen to be very similar to what is observed .
obviously , this model will produce a similar slope for every narrow line if the continuum change with similar amplitude in the wavelength range we concern .
it should be noted that due to variation in continuum slope , not all lines show the same be slope , but depends on wavelength .
we assume a constant narrow line flux during the continuum variability .
this is an approximation for the continuum variations on time scales of much shorter than the light travel time over the nlr because at such short time scale , the nlr has little response to the continuum variations . for variations on
longer time scales , one needs to properly convolve the continuum variations with the response of emission lines ( transfer function ) . by considering the latter response
, the variations of equivalent width will be somewhat smaller , but this is equivalent to requiring a somewhat larger continuum variability amplitude .
conversely , since continuum variability or the equivalent of continuum variability will inevitably produce a be , the be slope we observe could give an upper - limit on the few @xmath64yrs to @xmath59yrs variability of agn optical continuum .
if the variability amplitudes exceed the factor of three a steeper be will emerge .
this might seem at first sight to be at variance with the amplitudes of uv variability and structure functions observed , but it must be remembered that the equivalent width is the ratio of the line flux to the _ observed _ optical continuum and the observed continuum has a substantial starlight contamination .
an example of how the apparent continuum variability is substantially less than the real variability is shown in fig . 4 of gaskell et al ( 2008 ) .
host galaxy light limits the apparent peak - to - peak continuum variability to about a factor of three ( i.e. , a rms variability of a few tenths of a magnitude ) .
this is just what is observed for the pg agns studied by cid fernandez et al .
since the structure functions are relatively flat after a year or so ( cid fernandez et al . 2000 ) , the variability effect on the be should be apparent in a few years .
an obvious test of this is to reobserve agns after a few years . in summary
, the combination of a 50% variation of @xmath15and a log - normal distribution of luminosity will naturally produce a -0.20 slope of be for every narrow line , as we observe .
the model we employ is obviously an oversimplification . to make a more realistic model , we need a dedicated treatment of sample selection effects and to make a more realistic assumption of the variability including amplitude , variation form , and their dependence on wavelength etc , about the light curve . despite these simplifications , our results do show that continuum variation will inevitably produce a similar be for each narrow line .
meanwhile , while the model predicts the same be slope for each narrow line , the differences in slope are still significant due to the small error bars .
there must therefore be some additional factor at work .
it has been shown in fig . 3 that the differences do not correlate with the ionization potential or critical density , so there must be other factors taking effect . a deeper exploration of other factors like ionization slope
, nlr geometry and a more realistic model are needed to make progress , but these are beyond the scope of this paper .
we have constructed a sample of 3677 @xmath65 radio - quiet agns from sloan digital survey data release 4 that span three orders of magnitude in luminosity to explore the relationship between equivalent width of narrow lines and @xmath15 .
we have computed composite spectra for each @xmath15 bin to enhance the s / n ratio .
h@xmath17 and h@xmath66 as well as feiiwere subtracted to get an accurate measurement of narrow emission lines and continuum .
we find that most narrow lines show a similar be slope of about -0.2 while the large deviation of [ nii ] @xmath06583 , [ oii ] @xmath03727 , h@xmath19 and [ nev ] @xmath03425 might be explained by a metallicity effect , sf contamination , or the uv - optical continuum difference .
the slope does not show any correlation with ionization energy and critical density .
we propose that the combination of a 50% variation of the continuum near 5100 and a log - normal distribution of observed luminosity distribution will naturally produce a -0.2 slope of be for every narrow line .
we thank our referee : robert antonucci for inspiring comments that help to improve the paper significantly .
and we thank brent groves for helpful discussion on nlr model .
this work is supported by chinese nsf grants nsf-10703006 , nsf-10973013 , nsf-11073019 , and nsf-11233002 , the gemini - conicyt fund of chile through project n32070017 and fondecyt of chile through project n 1120957 .
funding for the sloan digital sky survey ( sdss ) has been provided by the alfred p. sloan foundation , the participating institutions , the national aeronautics and space administration , the national science foundation , the u.s .
department of energy , the japanese monbukagakusho , and the max planck society .
the sdss is managed by the astrophysical research consortium ( arc ) for the participating institutions .
the sdss web site is http://www.sdss.org/.
dietrich , m. , appenzeller , i. , wagner , s. j. , et al . 1999 , , 352 , l1 dietrich , m. , & wilhelm - erkens , u. 2000 , , 354 , 17 dietrich , m. , hamann , f. , shields , j. c. , constantin , a. , vestergaard , m. , chaffee , f. , foltz , c. b. , & junkkarinen , v. t. 2002 , apj , 581 , 912 dong , x .- b . , wang , t
g . , wang , j .-
, et al .
2009 , , 703 , l1 dong , x. , wang , j. , wang , t. , et al .
2009 , the starburst - agn connection , 408 , 83 dong , x .- b . ,
wang , j .-
, ho , l. c. , et al .
2011 , , 736 , 86 kraemer , s. b. , ruiz , j. r. , & crenshaw , d. m. 1998 , , 508 , 232 kraemer , s. b. , crenshaw , d. m. , hutchings , j. b. , et al .
2000 , , 531 , 278 kraemer , s. b. , trippe , m. l. , crenshaw , d. m. , et al .
2009 , , 698 , 106 netzer , h. 1985 , , 216 , 63 netzer , h. 1987 , , 225 , 55 netzer , h. , laor , a. , & gondhalekar , p. m. 1992 , , 254 , 15 netzer , h. , & laor , a. 1993 , , 404 , l51 netzer , h. , shemmer , o. , maiolino , r. , oliva , e. , croom , s. , corbett , e. , & di fabrizio , l. 2004 , apj , 614 , 558 netzer , h. , lutz , d. , schweitzer , m. , et al .
2007 , , 666 , 806 netzer , h. 2009 , , 399 , 1907 | the anti - correlations between the equivalent widths of emission lines and the continuum luminosity in agns , known as the baldwin effect are well established for broad lines , but are less well studied for narrow lines . in this paper
we explore the baldwin effect of narrow emission lines over a wide range of ionization levels and critical densities using a large sample of broad - line , radio - quiet agns taken from sloan digital sky survey ( sdss ) data release 4 .
these type1 agns span three orders of magnitude in continuum luminosity .
we show that most narrow lines show a similar baldwin effect slope of about -0.2 while the significant deviations of the slopes for [ nii ] @xmath06583 , [ oii ] @xmath03727 , [ nev ] @xmath03425 , and the narrow component of h@xmath1 can be explained by the influence of metallicity , star - formation contamination and possibly by difference in the shape of the uv - optical continuum .
the slopes do not show any correlation with either the ionization potential or the critical density .
we show that a combination of 50% variations in continuum near 5100 and a log - normal distribution of observed luminosity can naturally reproduce a constant baldwin effect slope of -0.2 for all narrow lines .
the variations of the continuum could be due to variability , intrinsic anisotropic emission , or an inclination effect . |
massive star formation must eventually differ substantially from low mass star formation because the outcomes are so different .
massive stars form in groups inside dense clusters of lower mass stars and whereas in low mass star formation , more isolated single stars or binary stars are formed .
both modes of star formation are initiated in cold molecular clouds so a key question is at what point of evolution do the two types of star formation diverge . for several decades ,
high quality data have been available for nearby low mass star formation .
spatially resolved molecular line observations in many species and transitions can be readily obtained . by combining this data with radiative transfer codes ,
it has proved possible to extract quantitative dynamical information from these high quality line profiles .
in contrast , massive star formation is intrinsically rarer and of shorter duration than low mass star formation which makes observational targets statistically much more distant .
however , comparable quality data to that of low mass star formation is now becoming available .
major observational spectral line surveys in the millimetre regime are now being carried out of massive star forming regions across the galaxy , for example in nh@xmath0 ( hops , * ? ? ?
* ) , co @xcite , cs @xcite , and multiple lines malt90 @xcite .
g333 is one of the most massive gmcs in the 4th quadrant of the galaxy and has been characterised by @xcite as being of radius 67 pc , distance 3.6 kpc and mass @xmath1 . in a companion paper to this work , lo et al 2015 present data from an observational survey of g333 on the nanten and mopra telescopes .
this work is part of a series of observational studies of this region @xcite .
the data revealed that three bright molecular line sources exhibit signature signs of infall .
these are designated g333.60.2 , g333.10.4 and g332.80.5 . within a mopra beam size of the outflow sources
are radio - detected hii regions , 1.2-mm dust emission clumps @xcite , and masers and _ iras _ sources , features which are all consistent with the presence of high mass star formation .
the mass and temperature of each source were confined with a detailed analysis of the observed luminosities . in particular , the masses were constrained by using the ( 10 ) line flux as the co(10 ) lines were found to be optically thick . therein results from a first attempt at radiative transfer modelling of the sources
were presented . here
we expand on the findings from the observations and earlier modelling by presenting radiative transfer modelling of the three sources in agreement with the observational results to quantify limits for the turbulent velocity , infall velocities , and chemical abundances .
as the three high mass star forming regions ( hmsfr ) are located within a massive giant molecular cloud ( gmc ) complex , the analysis was hampered by the multiplicity of sources present and the confusion of their association with features both spatially and kinematically .
the outflows are thus less well - defined than in low mass star forming regions ( lmsfr ) and the clean , detailed analysis that is possible there is not possible here . in particular , g333.60.2 presents a complex field of features in co(10 ) and presents very strong absorption in ( 10 ) and so also presents a particular challenge for modelling . despite these complications ,
many physical parameters can be constrained from the observations .
these data and their analysis are presented in full in @xcite . here
we reproduce the tables summarizing the major results in table [ tab : derivedproperties ] .
we will constrain ourselves to models that are in agreement with the core and outflow properties determined by @xcite .
the observational evidence presented by @xcite for outflows from the three sources is robust .
outflows are ultimately powered by infall but the dynamical signature of infall is much harder to detect for several reasons .
firstly , the gas undergoing the most rapid infall is located close to the central protostellar source - disk system which is typically unresolved by single dish observations .
gas in the outer regions of the cloud may have small infall motions but these can easily be masked by effects such as turbulence or bulk oscillations of the cloud .
secondly , infall takes place over a spherical volume in contrast to the highly directional outflows so that gas moving in a range of directions is typically present within a beam .
thirdly , effects such as rotation @xcite , turbulence and bulk oscillations @xcite can complicate the identification of an infall signature . to demonstrate that infall is also present therefore requires careful consideration of the radiative transfer effects taking place in potential sources . [ cols="<,^,^,^,^,^,^ " , ]
data from three massive star - forming cores in the g333 giant molecular cloud , obtained with mopra was used to investigate the dynamics and chemistry of these sources .
self - consistent 3-d molecular line radiative transfer modelling of the observed lines with the mollie code has confirmed the presence of infall in the sources .
the modelled infall velocities are 1.5 to 3.5 kms@xmath2 with infall rate of order @xmath3 m@xmath4yr@xmath2 .
derived time - scales suggest the three cores may be undergoing accelerated collapse as shown in @xcite , possibly due to external pressure from dynamical activities among high - mass star forming regions .
there is evidence of significant molecular depletion due to freeze - out , with the remaining gas phase molecules being subject to strong fractionation effects in species .
the turbulent and infall velocities associated with these hmsfrs are an order of magnitude larger than those expected in the low - mass case .
the highly supersonic velocity structure in turbulence and infall will very naturally lead to localised shocks that could supply the @xmath5c@xmath6 ions needed to fractionate co and hco@xmath6 .
the many similarities of these massive cores with their low mass counterparts is supportive of view of massive star formation being a scaled - up version of low mass star formation .
the principal and crucial difference is that the internal velocity fields are highly supersonic in the massive star formation case , and this gives a natural way to fragment the cloud into many individual star forming units , as required to form the dense clusters of low mass stars seen associated with young massive stars .
nl s postdoctoral fellowship is supported by a conicyt / fondecyt postdoctorado , under project no . 3130540 .
nl acknowledges partial support from the alma - conicyt fund for the development of chilean astronomy project 31090013 , center of excellence in astrophysics and associated technologies ( pfb 06 ) and centro de astrofsica fondap15010003 .
mpr acknowledges support from a science foundation ireland research frontiers grant .
lb acknowledges support from conicyt project pfb 06 .
the mopra telescope and atca are part of the australia telescope and are funded by the commonwealth of australia for operation as national facility managed by csiro .
the unsw - mops digital filter bank used for the observations with the mopra telescope was provided with support from the university of new south wales , monash university , university of sydney and australian research council .
this work has made use of the simbad database , operated at cds , strasbourg , france and also nasa s astrophysics data system .
we have also made use of the nist recommended rest frequencies for observed interstellar molecular microwave transitions , by frank j. lovas .
we thank the referee for the constructive comments on improving this work .
m. g. , braiding c. , glueck c. , goldsmith p. , hawkes j. , hollenbach d. j. , kulesa c. , martin c. l. , pineda j. l. , rowell g. , simon r. , stark a. a. , stutzki j. , tothill n. j. h. , urquhart j. s. , walker c. , walsh a. j. , wolfire m. , 2013 , pasa , 30 , 44 p. a. , cunningham m. r. , bains i. , muller e. , wong t. , burton m. g. , 2007 , in elmegreen b. g. , palous j. , eds , iau symposium vol . 237 of iau symposium , turbulence in the g333 molecular cloud .
pp 429429 c. h. , walsh a. j. , lowe v. , voronkov m. a. , ellingsen s. p. , breen s. l. , purcell c. r. , barnes p. j. , burton m. g. , cunningham m. r. , hill t. , jackson j. m. , longmore s. n. , peretto n. , urquhart j. s. , 2015 , mnras , 448 , 2344 | three bright molecular line sources in g333 have recently been shown to exhibit signatures of infall .
we describe a molecular line radiative transfer modelling process which is required to extract the infall signature from mopra and nanten2 data .
the observed line profiles differ greatly between individual sources but are reproduced well by variations upon a common unified model where the outflow viewing angle is the most significant difference between the sources .
the models and data together suggest that the observed properties of the high - mass star - forming regions such as infall , turbulence , and mass are consistent with scaled - up versions of the low - mass case with turbulent velocities that are supersonic and an order of magnitude larger than those found in low - mass star - forming regions . using detailed radiative transfer modeling ,
we show that the g333 cores are essentially undergoing a scaled - up version of low mass star formation .
this is an extension of earlier work in that the degree of infall and the chemical abundances are constrained by the rt modeling in a way that is not practical with a standard analysis of observational data .
we also find high velocity infall and high infall mass rates , possibly suggesting accelerated collapse due to external pressure
. molecular depletion due to freeze - out onto dust grains in central regions of the cores is suggested by low molecular abundances of several species .
strong evidence for a local enhancement of -bearing species towards the outflow cloud cores is discussed , consistent with the presence of shocks caused by the supersonic motions within them .
[ firstpage ] stars : formation ism : clouds ism : molecules ism : structure radio lines : ism . |
there are many motivations to introduce an extended higgs sector .
examples are , including obsolete ones , fine - tuning in the weak - scale higgs mass @xcite suggesting supersymmetry @xcite , @xmath0-violation @xcite , grand unification @xcite , extended gauge symmetry @xcite , strong @xmath0 problem @xcite , string theory @xcite , vacuum stability @xcite , light fermion mass / weak - scale hierarchy @xcite , fermion flavour structure @xcite , neutrino mass @xcite , dark matter ( see e.g. @xcite ) , cosmic rays @xcite , baryogenesis @xcite , inflation @xcite , novel collider phenomenology @xcite , as well as anomalies seen in precision ( see e.g. @xcite ) and accelerator ( see e.g. @xcite ) experiments . in this article , we entertain the possibility that an intriguing outcome from the lep experiment is in fact pointing to an extended higgs sector , which may be directly inspected at the large hadron collider ( lhc ) . specifically , we take as our framework the two higgs doublet model ( 2hdm ) of type - i with @xmath7 in the vicinity of the @xmath2-boson mass .
the two main points of this paper are the following .
( 1 ) the charged higgs produced in association with the lightest neutral higgs can be discovered as an excess in the @xmath8 process at the lhc .
moreover , this could happen even at the early stage with @xmath9 in a substantial portion of the parameter space .
( 2 ) with the lightest neutral higgs mass assumed to be @xmath10 , the charged higgs one - loop effect can enhance @xmath11 enough to produce the excess observed at the recent lhc analyses @xcite .
this enhancement is particularly welcome since the @xmath12 production at hadron colliders is generically suppressed due to higgs mixing .
this light charged higgs scenario was previously advocated in @xcite , whose contents we recap here .
the lep elecroweak working group has performed an analysis of @xmath2 decay branching ratios . in their report
, @xmath13 appears to be higher than @xmath14 and @xmath15 , to the level of 2.8 standard deviations @xcite .
should this difference be real , it would violate the charged current lepton universality , that is predicted by the gauge invariance of the standard model ( sm ) and has been confirmed in many indirect tests . as a resolution of this puzzle , one of the authors pointed out that a significant portion of the apparent lepton non - universality can be attributed to production of light charged higgs pairs which predominantly decay to the @xmath16 final states @xcite .
one of the most natural ways to realise a @xmath1 that has mass @xmath17 and is thus light enough to be on - shell produced at lep , is to employ the type - i two higgs doublet model ( 2hdmi ) . in this model ,
all the @xmath1fermion couplings are suppressed like @xmath18 , the ratio of the two vacuum expectation values ( vevs ) . with the choice of a high enough @xmath5 , this enables @xmath1 to escape from the severe flavour changing neutral current ( fcnc ) constraints as well as any process that probes the @xmath1fermion interaction such as @xmath19 , @xmath20 , and @xmath21 . presuming the above lep anomaly to be real
, the following points make this scenario more appealing .
first , invoking a light charge higgs ( discussed also in @xcite ) is the only proposed solution that does not spoil the other precision lepton universality tests using lepton or meson decays @xcite .
second , the requirement of a light charged higgs plus @xmath22 singles out type - i out of the four types of 2hdms @xcite in which tree - level fcnc is forbidden by a @xmath23 symmetry @xcite .
note that these assumptions already fix many aspects of the model that otherwise offer more than one option , i.e.the yukawa coupling structure , the charged higgs mass , and the viable range of @xmath5 .
these combine to predict a distinct set of collider physics signatures . in this paper
, we shall consider @xmath0-violation in the higgs sector , which can in general occur in a class of 2hdms . in this way
, we can explore higgs mixing patterns which received relatively less attention .
a practical advantage is that one can analyse both the scalar and the pseudo - scalar higgses in a unified manner : switching between the two sectors reduces to tuning the higgs mixing angles , which determine the composition of each neutral mass eigenstate . within the context of the minimal supersymmetric standard model ( mssm ) , there are earlier studies of @xmath0-violating higgs phenomenology . in particular , the general @xmath0-violating potential and mass spectrum
can be found in the paper of a. pilaftsis and c. wagner @xcite , where the importance of the scalar pseudo - scalar transitions in higgs phenomenology was emphasised . in the same article ,
the couplings of the charged higgs boson to @xmath0-violating higgs bosons were also discussed .
this article is organised as follows .
we begin by giving an overview of the model in the next section .
section [ sec : strategy ] is a summary of our search strategy , which leads to the main phenomenological results presented in section [ sec : res ] .
section [ sec : h1 ] is devoted to the diphoton signal of the lightest neutral higgs .
we deliver our conclusion in section [ sec : conclusions ] .
in addition , we present : the ghost lagrangian for our setup of 2hdm in appendix [ appe : b ] , the relation between our parametrisation and perturbativity in appendix [ appe : perturbativity ] , and differential cross - sections of our main process in appendix [ appe : diff ] .
let us give a brief description of the type - i 2hdm that we have chosen .
the type of a 2hdm refers to one of the multiple ways to organise the yukawa couplings .
there are four types in which quarks or leptons of each charge couple to only one higgs doublet @xcite , so that neutral higgses are naturally prevented from mediating fcnc @xcite . to this end
, we impose the softly broken @xmath23 symmetry , under which the two higgs doublets @xmath24 transform as @xmath25 whereas the fermion fields remain unchanged .
this allows only the following yukawa couplings , @xmath26 between @xmath27 and fermions , so all the fermions acquire mass from the vev of @xmath27 @xcite .
the scalar potential including the soft @xmath23-breaking terms reads @xmath28 \\ &
-\frac{1}{2}\left\{m_{11}^2(\phi_1^\dagger\phi_1 ) \!+\!\left[m_{12}^2 ( \phi_1^\dagger\phi_2)\!+\!{\rm h.c.}\right ] \!+\!m_{22}^2(\phi_2^\dagger\phi_2)\right\}. \end{split}\ ] ] in unitary gauge , the higgs doublet fields are expanded around the minimum like @xmath29 ,
\quad \phi_2 = \left [ \begin{array}{c } -i { \cos\!\beta}h^+ \\ \frac{1}{\sqrt{2 } } { ( v_2 + \eta_2 + i { \cos\!\beta}\,\eta_3 ) } \end{array } \right ] .\ ] ] one can take a basis of @xmath24 such that @xmath30 and @xmath31 are both real .
their quadrature sum is subject to the constraint , @xmath32 , where @xmath33 is the sm higgs vev .
their ratio is parametrised by @xmath34 in terms of the pseudo - scalar and charged higgs mixing angle @xmath35 .
we introduce @xmath0-violation in the higgs sector by assuming @xmath36 and @xmath37 , which enables the neutral scalar and pseudo - scalar components to mix together . the real symmetric mass - squared matrix @xmath38 with respect to the basis @xmath39 ,
is diagonalised by the mass eigenstates @xmath40 like @xmath41 using a @xmath42 orthogonal matrix @xmath43 .
it is parametrised in terms of the mixing angles @xmath44 in the form , @xmath45 where @xmath46 and @xmath47 .
the neutral higgs mass eigenvalues are assumed to be in the order , @xmath48 this determines the following physical domain of the mixing angles @xcite : @xmath49 the two @xmath0-violating mass matrix elements are related by @xcite @xmath50 which translates into the constraint , @xmath51 we shall remain in the parameter volume that is compatible with the mass ordering ( [ eq : mass ordering ] ) in conjunction with ( [ eq : m3 ] ) , in order to avoid double counting of physically identical parameter sets .
two more dimensionful quantities can be derived from the scalar potential : the charged higgs mass , @xmath52 and the auxiliary parameter , @xmath53 that sets the mass scale of charged / neutral higgs particles except @xmath12 .
we have the eight physical quantities determined from the higgs potential , @xmath54 in this work , we opt to invert this dependency and express the parameters appearing in ( [ eq : pot ] ) in terms of those in ( [ eq : input ] ) @xcite .
the mass parameters are then given by @xmath55 with @xmath56 and @xmath57 fixed by ( [ eq : mu ] ) .
the quartic couplings are [ eq : lambda ] @xmath58 , \end{split } \\
\begin{split } \lambda_2&=\frac{1}{s_\beta^2v^2 } [ s_1 ^ 2c_2 ^ 2m_1
^ 2 + ( c_1c_3-s_1s_2s_3)^2m_2 ^ 2 \\ & + ( c_1s_3+s_1s_2c_3)^2m_3 ^ 2-c_\beta^2\mu^2 ] , \end{split } \\ \begin{split } \lambda_3&=\frac{1}{c_\beta s_\beta v^2 } \{c_1s_1[c_2 ^ 2m_1 ^ 2+(s_2 ^ 2s_3 ^ 2-c_3 ^ 2)m_2 ^ 2 \\ & + ( s_2 ^ 2c_3 ^ 2-s_3 ^ 2)m_3
^ 2 ] + s_2c_3s_3(c_1 ^ 2-s_1 ^ 2)(m_3 ^ 2-m_2 ^ 2)\ } \\ & + \frac{1}{v^2}(2m_{h^\pm}^2-\mu^2 ) , \end{split } \\
\lambda_4&=\frac{1}{v^2 } [ s_2 ^ 2m_1 ^ 2+c_2 ^ 2s_3 ^ 2m_2 ^ 2+c_2 ^ 2c_3 ^
2m_3 ^ 2 + \mu^2 - 2m_{h^\pm}^2 ] , \\ { \mathrm{re}}\lambda_5&=\frac{1}{v^2 } ( -s_2 ^ 2m_1 ^ 2-c_2 ^ 2s_3 ^ 2m_2 ^ 2-c_2 ^ 2c_3 ^ 2m_3 ^ 2+\mu^2 ) , \\
\begin{split } \label{eq : lambda5i } { \mathrm{im}}\lambda_5&=\frac{-1}{c_\beta s_\beta v^2 } \{c_\beta[c_1c_2s_2m_1 ^ 2-c_2s_3(c_1s_2s_3+s_1c_3)m_2 ^ 2 \\ & + c_2c_3(s_1s_3-c_1s_2c_3)m_3 ^ 2 ] + s_\beta[s_1c_2s_2m_1 ^ 2 \\ & + c_2s_3(c_1c_3\!-\!s_1s_2s_3)m_2 ^ 2 \!-\!c_2c_3(c_1s_3\!+\!s_1s_2c_3)m_3 ^ 2]\ } , \end{split}\end{aligned}\ ] ] where @xmath59 , @xmath60 . in this approach ,
the solutions for @xmath61 may turn out to be non - perturbatively large , unless one takes care to set the input parameters in ( [ eq : input ] ) to sensible values . to avoid entering a nonsensical regime
, we shall check that the sizes of the higgs quartic couplings meet the conditions , @xmath62 ) , within the parameter space explored in this work .
the interactions of higgs bosons with other particles are described by the lagrangian , @xmath63 with @xmath64 from ( [ eq : yukawas ] ) .
.feynman rules .
the momentum directions are indicated in figure [ fig : hhc via w ] .
the chirality projection matrices are defined by @xmath65 . [ cols="^,^ " , ] one expands the covariant derivatives , replaces @xmath24 by ( [ eq : higgs expansion ] ) , and diagonalises the fermion and higgs mass terms , to arrive at the feynman rules in table [ tab : feynman rules ] .
they will be used to calculate and interpret the results that we present in the following sections .
a remark is in order regarding the parametrisation in the @xmath0-conserving limit .
one can enforce @xmath0-conservation simply by setting @xmath66 .
this eliminates mixing between the @xmath0-odd and the @xmath0-even degrees - of - freedom , as can be seen in ( [ eq : rotation matrix ] ) .
the mixing between @xmath67 and @xmath68 with the angle @xmath69 , results in the lighter and the heavier @xmath0-even neutral eigenstates , @xmath70 and @xmath71 .
the @xmath0-odd neutral higgs @xmath72 is identified with @xmath73 . in this case
, the relation ( [ eq : rel m13-m23 ] ) becomes trivial and therefore it no longer constrains @xmath74 to be determined as a function of the other parameters . as a result , we are left with two less mixing angles and one more free mass parameter in the @xmath0-conserving higgs sector than in the @xmath0-violating case .
conventionally , these seven free parameters are chosen to be : @xmath75 @xcite .
having established the framework , one should fix the range of parameters to study . for future references ,
we first collect the values which were chosen for the reasons explained below : @xmath76 this set shall be used by default throughout the analysis unless specified otherwise , in combination with one of the two higgs mixing configurations , @xmath77 the charged higgs mass is more or less predetermined by the lep data .
the lower bound is provided by the direct search , and the upper limit comes from the requirement that the @xmath1 pair production should restore lepton universality in leptonic @xmath2 decays .
we take @xmath7 which reduces the apparent discrepancy between @xmath78 and the average of @xmath79 and @xmath80 by 1 @xmath81 @xcite .
the latest update from the opal collaboration reports the direct search limit , @xmath82 , for @xmath83 at 95% confidence level ( cl ) @xcite . in view of the recent evidences from atlas @xcite and cms @xcite
, we set the lightest neutral higgs mass to @xmath3 . the heavier higgs mass parameter is taken to be fairly higher than @xmath84 .
non - degenerate mass spectrum is needed for @xmath0-violating higgs mixing .
the last dimensionful parameter @xmath85 is set to a representative value of the weak scale order .
the chosen @xmath86 is a mass at which the sm higgs production is strongly constrained at the lhc . not far from there
lies also @xmath87 . in this range ,
the rates of the @xmath88 and @xmath89 channels are limited to be roughly below half their sm predictions @xcite . in our model ,
the upper bounds must be reinterpreted as arising from processes mediated by @xmath90 and @xmath91 whose interactions are different from those of the sm higgs . for this , we require that @xmath92 a potentially important constraint is provided by the @xmath93 parameter @xcite . with the higgs mass spectrum assumed in this work , one can check that the one - loop correction from the higgs bosons to @xmath93 is small enough to pass the electroweak precision tests @xcite . for safety of a light charged higgs , it is crucial to choose a high enough @xmath5 .
the strongest bound is given by @xmath94 , which requires @xmath95 @xcite .
this range of @xmath5 also suppresses @xmath96 below 4% .
this maximal top - quark branching fraction coincides roughly with the numerical values of the 95% cl limits from the light charged higgs searches at the lhc @xcite .
however , these lhc constraints on @xmath97 are based on the assumption that @xmath98 , and therefore become weaker in our setup where @xmath99 . in principle , one could use any @xmath5 larger than the fcnc bound , as long as it lets the charged higgs decay inside the detector .
in practice , it is bounded from above by perturbativity through ( [ eq : lambda ] ) . obviously , this upper limit is a function of the other parameters . in appendix [ appe :
perturbativity ] , it is shown that @xmath85 can always be adjusted so that perturbativity holds .
one should not regard this as a fine - tuning but rather an artifact of the employed parametrisation .
if one started by setting the parameters appearing in ( [ eq : pot ] ) and then derived the quantities in ( [ eq : input ] ) , then one could trivially satisfy perturbativity by keeping @xmath61 inside the range ( [ eq : perturbativity ] ) . even in this reverse approach
, one could naturally obtain a high @xmath5 , as one can see from the example of inert doublet model where an exact @xmath23 symmetry leads to infinite @xmath5 @xcite . among the three remaining higgs mixing angles , we scan @xmath100 and @xmath101 which determine the composition of the lightest neutral higgs .
in ( [ eq : rotation matrix ] ) , one can notice that @xmath102 is the fraction of the @xmath0-odd component in @xmath12 .
we fix @xmath103 to @xmath104 to have an intermediate level of @xmath0-violation . for illustrative purposes ,
we select two benchmark points , labelled @xmath105 and @xmath106 in ( [ eq : benchmark points ] ) .
the fraction of @xmath72 in @xmath12 is chosen to be small at @xmath105 and large at @xmath106 .
given the model and the input parameters , we consider which charged higgs production mechanism is most useful .
we first calculate the cross - sections of the following standard channels : @xmath107 the last two entries are meant to imply their conjugate processes as well .
the final state in ( [ eq : tbhc ] ) can be made through either top pair production or gluon - gluon fusion @xcite .
+ in figures [ fig : sp ] , we plot the cross - sections against @xmath5 for the two benchmark points .
one observes the general tendency that each production rate is depleted for high @xmath5 . as one can expect from table [ tab : feynman rules ] , processes ( [ eq : thc ] ) and ( [ eq : tbhc ] ) die out as @xmath5 grows , while the others approach their individual asymptotic cross - sections even though many of them are small .
( 180,120)(0,-60 ) ( 25,35)(50,0 ) ( 25,-35)(50,0 ) ( 50,0)(100,0)34.5 ( 140,40)(100,0)4 ( 100,0)(140,-40)4 ( 75,8)[b]@xmath2 ( 127,12)[l]@xmath108 ( 127,-12)[l]@xmath109 ( 148,40)[l]@xmath110 ( 148,-40)[l]@xmath111 ( 17,35)[r]@xmath112 ( 17,-35)[r]@xmath113 an outstanding exception is the production associated with a neutral higgs ( [ eq : hhc ] ) , which is not suppressed even for high @xmath5 . the diagrams for this process are shown in figure [ fig : hhc via w ] , which scale according to the @xmath114@xmath1@xmath2 vertices in table [ tab : feynman rules ] . in the @xmath12@xmath1@xmath2 interaction ,
the pseudo - scalar coupling is proportional to @xmath102 and independent of @xmath35 , and the scalar coupling is proportional to @xmath115 in the limit of @xmath116 . combining these two behaviours , one can expect that the @xmath117 production becomes efficient unless @xmath118 is large and @xmath119 is small .
the results plotted in figures [ fig : sp ] do not depend on @xmath85 .
therefore , they remain valid even if one adjusts @xmath85 for the perturbativity of @xmath61 in the high @xmath5 regime .
this means that the production mechanism ( [ eq : hhc ] ) is of the most general interest among the displayed channels , and we shall focus on it in what follows . note that the same type of diagram was previously considered in the contexts of the mssm @xcite and a fermiophobic higgs scenario @xcite .
the next step should be to select the decay products of @xmath1 and @xmath12 .
the two dominant branching ratios of @xmath111 are @xmath120 as long as @xmath121 , beyond which the charged higgs decays mediated by a virtual top becomes non - negligible @xcite .
these branching fractions are essentially independent of any of the input parameters .
this is a notable feature of the 2hdmi , stemming from the universal scaling behaviour of the @xmath1fermion couplings : they are all proportional to @xmath122 , as shown in table [ tab : feynman rules ]
. as we are mainly interested in a charged higgs that is light enough to have been produced at lep energies , we can regard ( [ eq : brs of h+ ] ) as good approximations and simply choose the decay product with the highest rate , @xmath16 .
+ as for @xmath12 , we plot its branching fractions in the vicinity of @xmath123 , in figures [ fig : brh1 ] .
one finds that the @xmath124 mode is dominant at both benchmark points .
this tends to be the case unless both @xmath125 and @xmath102 are vanishingly small .
at the point where @xmath126 , the tree - level @xmath12fermion coupling vanishes , causing its fermiophobia .
considering these dominant decay modes of @xmath1 and @xmath12 , we are going to concentrate on the final state , @xmath4 for numerical simulation .
one can also notice that the branching fractions of the @xmath6 and @xmath127 modes can differ significantly from their individual sm values .
albeit not directly related to the charged higgs search , these changes greatly affect the phenomenology of @xmath12 .
we shall come back to this issue in section [ sec : h1 ] .
the main process of our concern is @xmath128 , followed by the decays , @xmath129 and @xmath130 at the lhc . among these three subprocesses ,
the charged higgs production and decay revealed their outstanding stability under variation of @xmath5 over a wide range , in the preceding section .
we shall envisage two experimental conditions at the lhc , which we call the `` early '' and the `` full '' searches .
the `` early search '' denotes the centre - of - mass energy of the current run , @xmath131 , and the integrated luminosity of @xmath132 .
we assume that the `` full search '' will deliver @xmath133 at @xmath134 .
the kinematics of the @xmath4 state can be fully reconstructed with high efficiency as it has only one neutrino .
we assume a @xmath135-reconstruction efficiency @xmath136 and a @xmath137-tagging efficiency @xmath138 , plus the following set of selection cuts : @xmath139 which we choose considering the operation of atlas . assuming these cuts , we calculated the sm background .
the cross - section is @xmath140 at @xmath131 and @xmath141 at @xmath134 . along with the associated production of @xmath142
, we also study a similar process with @xmath1 replaced by @xmath2 , i.e.@xmath143 .
as we shall see , the parameter space covered by this higgs - strahlung is complementary to the region accessible to the @xmath142 state .
we select the final state , @xmath144 with @xmath145 , and assume a @xmath2-reconstruction efficiency @xmath136 .
the cuts in ( [ eq : cuts ] ) are used again , and the same cut on @xmath135 is applied on the transverse momenta of the light leptons , @xmath146 . using these parameters , we obtain the higgs - less sm background cross - section to be @xmath147 at @xmath131 and @xmath148 at @xmath134 .
we used lanhep @xcite to implement the model , and calchep @xcite plus the cteq6 m parton distribution functions @xcite , to perform the phenomenological analysis .
the passarino - veltman functions appearing in the effective vertices of @xmath114@xmath149@xmath149 and @xmath114@xmath150@xmath150 are calculated by calling the looptools package @xcite .
we first check the rates of our main processes with @xmath151 . in this @xmath0-conserving case
, one has the lighter and the heavier @xmath0-even higgses , @xmath152 and @xmath153 , and the @xmath0-odd higgs , @xmath154 .
the production cross - sections of @xmath70 , @xmath71 , and @xmath72 , each in association with a charged higgs , are shown in figure [ fig : cpc xs vs angles ] as functions of @xmath155 .
for each type of neutral higgs , there are three curves that respectively represent @xmath156 , @xmath111 , and @xmath1 productions in the order of increasing thickness .
the higher @xmath111 rate relative to @xmath156 reflects the fact that the lhc is a @xmath157 collider , not @xmath158 .
the shape of each curve is essentially determined by the dependence of the corresponding @xmath2@xmath1@xmath114 coupling on @xmath100 and @xmath35 , presented in table [ tab : feynman rules ] .
these patterns will be useful for understanding the simulation results to follow . from the plot
, one can expect @xmath159 events that occur through a charged - neutral higgs associated production with the `` full luminosity '' at @xmath160 , unless the mixing angle is such that the given channel is highly suppressed .
our analysis is of relevance to a light charged higgs , even if one is not interested in a scenario where charged higgs bosons have been on - shell produced at lep energies . in this regard ,
it is legitimate to consider changes that would be caused by varying @xmath7 .
obviously , the above production rates will be reduced as @xmath7 grows , as is shown in figure [ fig : cpc xs vs mhc ] .
we take the range of @xmath7 up to @xmath161 , beyond which the changed higgs decay pattern undergoes a qualitative change @xcite , thereby invalidating our assumptions given in ( [ eq : brs of h+ ] ) . for @xmath162 ,
each cross - section decreases roughly down to half the value in panel [ fig : cpc xs vs angles ] .
these curves illustrate the phase - space suppression that takes place for higher @xmath7 .
we emphasise that this reduction of statistics is the only essential information that one needs to understand how the significances would change in the following quantitative analysis with @xmath163 , if one took a different @xmath7 . for the simulation of charged higgs search , we turn on @xmath0-violation allowing for mixing among all the three neutral higgses , @xmath70 , @xmath71 , and @xmath72 . in figures [ fig : significance cpv ]
, the contours of the significance @xmath164 are shown for each process , which have been evaluated by means of the method described in @xcite .
there are two complications in comparison to the @xmath0-conserving limit .
one is that the additional mixing between the @xmath0-odd and the @xmath0-even components affects the significance of each process .
the other is that @xmath165 are no longer independent of one another but constrained by ( [ eq : m3 ] ) . in combination with the latter relation ,
the mass ordering ( [ eq : mass ordering ] ) excludes the upper - right and the lower - left rectangular regions from the parameter domain .
another theoretical constraint that rules out the upper - left and the lower - right triangular regions , is perturbativity ( [ eq : perturbativity ] ) .
we have also checked numerically that imposing perturbativity guarantees the tree - level unitarity in higgs - higgs scattering in the parameter space under consideration .
in addition to the above theoretical requirements , we impose the experimental constraints in ( [ eq : h2 exclusion ] ) in order to respect the current higgs production limits from the lhc .
this rules out the dark - shadowed strip on the plane .
the bounds on @xmath90 and @xmath91 apply as long as their masses are in the range between 160 and 600 gev .
all these conditions being fulfilled , the question to ask is how much portion of the remaining unshaded region the lhc can cover . as in the preceding plot , the @xmath2@xmath1@xmath12 coupling largely determines the shape of the @xmath142 contours . as long as one moves along the horizontal axis , @xmath12 is purely @xmath0-even , and therefore its interactions not involving other neutral higgses reduce to those of @xmath70 in the @xmath0-conserving higgs sector .
and @xmath91 do not reduce to those of @xmath71 and @xmath72 due to non - vanishing @xmath103 . ] for instance , the @xmath2@xmath1@xmath12 coupling vanishes at the point @xmath166 , as is the case in figure [ fig : cpc xs vs angles ] .
this explains why the significance decreases as one approaches this point .
it should be meaningful to interpret the interesting behaviour of @xmath164 around the point @xmath167 , although it belongs to the @xmath168 exclusion strip .
for instance , there might be a way to circumvent the constraint ( [ eq : h2 exclusion ] ) by taking a different set of parameters . in the neighbourhood of this point ,
one finds a sudden drop of significance even though the @xmath2@xmath1@xmath12 coupling stays nearly maximal .
this is because the @xmath12@xmath137@xmath169 coupling is suppressed there as can be seen in table [ tab : feynman rules ] . at the origin , @xmath12 becomes fermiophobic , and its dominant decay modes are @xmath170 and @xmath171 .
the fermiophobia of @xmath12 generically gives rise to an enhanced @xmath172 without the charged higgs loop contribution . including the @xmath1-loop , the diphoton branching fraction can range from zero up to @xmath173 .
[ we give a related discussion in section [ sec : h1 ] .
] therefore , the possibility of observing the @xmath142 state in the central low - significance hole via an alternative decay channel of @xmath12 such as the @xmath6 mode , is highly dependent on the other model parameters .
as one departs away from the horizontal axis , @xmath174 , @xmath12 acquires more of the @xmath0-odd component .
this reinforces the production rate of @xmath142 and thereby the significance , since the @xmath2@xmath1@xmath72 coupling is not suppressed for any @xmath100 as can be seen in figure [ fig : cpc xs vs angles ] .
overall , figures [ fig : significance cpv ] demonstrate that there is an ample portion of the parameter space in which the lhc has a good chance to discover the @xmath142 production .
obviously , the chance is substantially better in the `` full luminosity '' case , shown in figure [ overlap_14tev ] . with the exception of the narrow corners around both ends of the horizontal axis and the hole encircling the fermiophobic point , the discovery of @xmath142 through the @xmath4 final state should be possible at any point in the unshaded physically sensible domain defined by ( [ eq : mass ordering ] ) , ( [ eq : perturbativity ] ) , and ( [ eq : h2 exclusion ] ) . in the same figures
is presented the @xmath12-strahlung process , which could also be detected depending on the mixing angles and the lhc operation condition .
one can understand the shape of the @xmath175 contours by looking at the @xmath12@xmath176@xmath177 coupling in table [ tab : feynman rules ] . in figure [ overlap_07tev ] for `` early searches ''
, one finds no common portion of the parameter space with @xmath178 for both processes .
this setup predicts that it is not possible to discover both channels during the first data taking period in the @xmath179 final states .
should this occur , this scenario will have to be modified to survive . in figure [ overlap_14tev ] , one finds a much broader region in which @xmath12-strahlung could be discovered .
part of this region is shadowed in light grey where the `` full lhc '' could observe both this process and @xmath142 production .
this section is devoted to phenomenology of the lightest neutral higgs .
in particular , we show that the charged higgs loop can exert a marked influence on the @xmath181 decay mode , which is of the utmost importance in the neutral higgs search at the lhc .
we shall consider the following standard higgs production mechanisms : @xmath182 in figures [ fig : np ] , we plot the cross - sections of these channels at @xmath131 against @xmath5 for each of the two benchmark points .
we have explicitly verified that all these production mechanisms are mostly unaffected in the range of @xmath84 between @xmath183 and @xmath184 and in the neighbourhood of the chosen @xmath7 .
all the production rates are lower than those in the sm .
this is because of the suppression of the couplings in table [ tab : feynman rules ] . combining production and decay
, we obtain an intriguing result : @xmath12 may have escaped from the past and present searches due to the suppression of both its production and the branching ratios to massive vector boson pairs .
as this scenario generically predicts lower rates into fermion pairs , it could be disfavoured by the recent hint in the @xmath124 mode from the tevatron @xcite if it persists .
obviously , one can ease the reduction of @xmath12 production by taking the higgs mixing pattern to the sm - like limit , @xmath185 . in this case , the charged - neutral higgs associated production will be suppressed as shown in figures [ fig : significance cpv ] . recall that we are assuming a hierarchical neutral higgs mass spectrum and that all the experimental evidences from the lhc and the tevatron are caused by the lightest of the three neutral higgs particles , which is also supposed to be produced in association with @xmath1 .
relaxing one or more of these hypotheses may lead to a different consequence , although we do not explore this possibility in this paper . on the other hand ,
branching fractions of the two loop - induced decay modes , @xmath186 and @xmath187 , have possibilities of being enhanced as well as diminished , depending on the input parameters .
if they are increased , the suppressed higgs production can be alleviated .
the decay into the digluon final state arises from the fermion loop , dominated by the top - quark contribution .
although the same type of loop is present also in the sm , its value can be different due to the variation of the @xmath12@xmath188@xmath189 coupling which is determined by the higgs mixing angles as in table [ tab : feynman rules ] .
in particular , an interesting role is played by the @xmath0-odd component @xmath72 in @xmath12 that could enhance the digluon branching fraction .
its effects are twofold : ( a ) @xmath72 does not couple to a vector boson pair unlike the @xmath0-even higgs @xmath70 ;
( b ) the digluon to difermion partial width ratio of @xmath72 is higher than the corresponding quantity of @xmath70 , i.e.@xmath190 when @xmath191 .
this effect stems from the different lorentz structures of the scalar and the pseudo - scalar couplings with fermions .
the potential enhancement of the @xmath127 decay mode might be of interest at a future lepton collider experiment .
the decay amplitude of @xmath186 has two more components , the @xmath2- and the @xmath1-loops , in addition to the fermion loop . in the sm ,
the dominant contribution comes from the @xmath2-loop . like the fermion loop
, the @xmath2-loop can vary as a function of @xmath192 .
one might attempt to increase the diphoton branching fraction by invoking the same mechanism as with the digluon mode .
as @xmath12 becomes pseudo - scalar - like , the fermion loop alone does receive the relative enhancement that is given by ( [ eq : agg / hgg ] ) with @xmath127 replaced by @xmath6 and each numerator including only the fermion loop .
however , this kills the @xmath2-loop at the same time , which is the dominant contribution in the sm .
therefore , a high @xmath0-odd fraction in @xmath12 is not really helpful in increasing @xmath193 .
nevertheless , it is known that the charged higgs loop can greatly affect the @xmath6 mode when @xmath7 is low @xcite .
in particular , this mechanism can lead to a high enough diphoton rate to be consistent with the excess surrounding the higgs mass of @xmath3 recently observed by atlas @xcite and cms @xcite .
the @xmath12@xmath111@xmath156 coupling depends on more parameters than the other loops , in particular on @xmath85 , with which one can play in order to engineer the amplitude .
the maximal permitted size of the @xmath1-contribution is mostly controlled by perturbativity and unitarity .
in passing , we remark that change of @xmath85 affects also @xmath194 and therefore the region excluded by ( [ eq : h2 exclusion ] ) , as we will see in the following plots .
+ to further study this potential variation , we plot the branching fraction of @xmath186 in figures [ h1-aa ] and [ h1-aa mu=180 ] .
indeed , it can deviate substantially from its sm value depending on the mixing angles .
one finds a valley between the two curves labelled `` a '' on each of the two plots , along which the branching ratio is highly suppressed . across this valley
, the @xmath186 amplitude flips its sign .
as the plots show , the contours are deformed if @xmath85 is changed , since they depend on the @xmath1-loop .
one can also notice that part of the `` d '' contour of a high branching fraction lies on the verge of the perturbativity border , demonstrating that a sizeable @xmath1-loop contribution involves large quartic higgs couplings .
in conjunction with gluon - gluon fusion , this brings us to the cross - section of @xmath195 in figures [ gg - aa ] and [ gg - aa mu=180 ] . with our crude leading - order approximation
, it is estimated to be about @xmath196 in the sm . in the entire region of panel [ gg - aa ] ,
the cross - section is smaller than its sm value due to the suppressed @xmath12 production .
nevertheless , one still has the option to change @xmath85 that affects the charged higgs loop .
for instance , the cross - section at @xmath197 is approximately @xmath198 for @xmath199 . at this point ,
the @xmath2- and the @xmath1-loops interfere destructively .
taking @xmath200 instead , one can make the interference constructive by flipping the sign of the @xmath1-contribution .
this pushes the cross - section up above the sm prediction , as displayed in figure [ gg - aa mu=180 ] . in this plot
, the rate of @xmath195 can be comparable to its sm value even when the mixing angles are fairly removed from the sm - like point , @xmath185 . in combination with figures [ fig : significance cpv ]
, this exhibits the exciting possibility that the lhc can discover a light charged higgs in the foreseeable future while the recent evidences for higgs decays to @xmath6 are partly due to the charged higgs loop effect .
as one can expect , the charged higgs loop effect weakens for high @xmath7 .
for instance , @xmath193 for @xmath162 is roughly half its value for @xmath163 , when one keeps @xmath197 and @xmath200 .
we demonstrated the superb potential of the lhc to discover a light charged higgs close to the @xmath2-boson mass in the @xmath0-violating 2hdmi . for this purpose
, we exploited the property of the @xmath201 process that its cross - section is stable against variation of @xmath5 , which can be anywhere above the lower bound placed by the most stringent fcnc constraints .
we set the higgs masses to their favoured values , @xmath202 and @xmath203 , and found that the lhc can spot an excess in the @xmath4 events , except in small corners and holes of the parameter space .
we also examined the @xmath12-strahlung channel from @xmath2 , which offers access to the complementary parameter volume to that of @xmath204 . with regard to @xmath0-violation
, we observed interesting consequences of the scalar pseudo - scalar mixing .
in particular , @xmath205 can be considerably affected by the composition of @xmath12 .
this result might be relevant to higgs scrutiny at prospective linear collider experiments .
motivated by the recent evidences for higgs production at the lhc , we studied @xmath206 . due to the higgs mixing matrix elements appearing at interaction vertices , the production of @xmath12
is generically suppressed in comparison to the sm .
nonetheless , the @xmath1-loop amplitude can make a constructive interference with the standard @xmath2-loop to raise @xmath207 . depending on the parameter choice
, this effect can be significant enough to increase the diphoton production rate from gluon - gluon fusion up to or even above the sm prediction .
alternatively , one can opt to enhance the @xmath12 production itself by letting the higgs mixing pattern approach the sm - like limit . in this case , the other processes with @xmath12 going into the fermion and the massive vector boson pairs are also restored , the price to pay being suppression of the @xmath204 rate . related to this last point , it may be beneficial to study further possibilities of the light charged higgs scenario in which one makes assumptions different from those underlying the present article .
gmp would like to thank per osland and alexander pukhov for helpful discussions .
jp thanks eung jin chun for valuable comments and kazuki sakurai for the help with access to the literature .
we acknowledge financial support from german research foundation dfg through grant no .
sto876/21 and bmbf .
assuming the t hooft - feynman gauge , we choose a basis for the @xmath208 by expanding the higgs doublets as @xmath209 where @xmath210(@xmath211 ) is the goldstone boson of @xmath212(@xmath213 ) .
we remark that the @xmath73 pseudo - scalar field is orthogonal to the neutral goldstone boson @xmath211 . as for the goldstone boson mass spectrum ,
it is possible to find a convenient way to write the mass matrix .
having each @xmath214 the same group representation of the sm higgs , following the notation of @xcite , in the gauge - goldstone matrix follows from the four gauge bosons @xmath215 , @xmath213 , and the three goldstone bosons @xmath216 . ]
basis we find the following representation of the co - variant derivative : @xmath217 while the vector boson ( and ghost ) mass matrix is @xmath218 , the goldstones mass matrix is : @xmath219 therefore we get @xmath220 the mass matrix in equation ( [ gb_mass_matrix ] ) shows that the goldstones have a mass that is equivalent to the sm - ones , as expected since the gauge sector has not been extended .
as we have already intimated , the ghost mass matrix and interactions are defined by means of the same matrix @xmath221 via @xmath222 notice that the @xmath223 and the @xmath224 of equation ( [ gb_mass_matrix ] ) have different numbers of zero - eigenvalues , but their non - zero eigenvalues are in a one - to - one correspondence ; furthermore , the eigenvalues of the gauge - fixing mass matrix are the same of the gauge boson mass matrix , as expected . then , the ghost lagrangian is defined as @xmath225 c^b\ , \nonumber \\ & - & \bar{c}^a \left [ \mathcal{d}_2^a \cdot \left ( \mathcal{d}_2^b + \mathcal{s}_2^b \right)^t \right ] c^b\ , , \end{aligned}\ ] ] where the matrices @xmath226 represent the link between the fluctuations ( goldstones ) of the higgses around their vevs and the gauge bosons ; a convenient combination of scalar fields .
consistently , the sum of the two contributions disarms these connections : ghost fields do not link to charged higgs . ]
way to write this matrices is @xmath227 @xmath228 where @xmath229 finally , the ghost fields ( @xmath230 ) read as @xmath231 where @xmath232
in order to properly realise the high @xmath233 scenario , we must carefully consider the perturbativity of the couplings .
firstly , from equation ( [ eq : m3 ] ) we see that the @xmath234 approximation leads to the following definition of the heaviest neutral scalar mass : @xmath235 from equations ( [ eq : lambda ] ) we can conclude that all the couplings are finite and perturbatively small except for @xmath236 and @xmath237 .
is apparently divergent , by plugging the definition of @xmath238 in ( [ eq : lambda5i ] ) it is easy to see that the result is always finite .
if @xmath30 exactly vanishes , one can rotate away the phase of @xmath239 to render @xmath0-conservation manifest @xcite .
] in fact , for @xmath240 or equivalently @xmath241 and @xmath242 , the two potentially diverging couplings become : @xmath243 , \\
\lambda_3&\simeq\frac{m_1 ^ 2-m_2 ^ 2}{v^2\cos{\beta}}\ , \frac{r_{22 } r_{12}}{r_{33}}.\end{aligned}\ ] ] hence , no matter how large @xmath5 is , we can always keep @xmath236 and @xmath237 perturbatively small by choosing @xmath244 respectively .
this shows that there is no strict upper limit on @xmath5 from perturbativity .
however , we do not make use of this mechanism for drawing our conclusions in the main part of the present paper .
with respect to @xmath245 of the @xmath137-quark and the @xmath135-lepton at the indicated centre - of - mass energies , for the benchmark point @xmath105 . each curve is the sum of @xmath246 and that for the conjugate final state .
the shadowed range ( red online ) shows the selection cuts set for the signal - to - background analysis.,scaledwidth=48.0% ] in figure [ fig : spectra ] , we show the @xmath245 spectra for the @xmath137-quark and the @xmath135-lepton , extracted from the process @xmath247 at the lhc .
they were evaluated at the benchmark point @xmath105 . in the @xmath248 distributions
, one finds that the selection cuts in ( [ eq : cuts ] ) act on regions with suppressed differential cross - sections .
this means that we retain the major portion of events at our disposal even after applying the cuts . even in the @xmath249 distributions ,
the excluded area below each curve is relatively small compared to the total cross - section .
s. weinberg , _ gauge hierarchies _ , _ phys .
_ * b82 * ( 1979 ) 387 ; m. veltman , _ the infrared ultraviolet connection _ , _ acta phys .
_ * b12 * ( 1981 ) 437 ; c. h. llewellyn smith and g. g. ross , _ the real gauge hierarchy problem _ , _ phys .
_ * b105 * ( 1981 ) 38 .
see e.g. j. l. rosner , _ an @xmath251 interpretation of an @xmath252 event _ , _ phys .
_ * d55 * ( 1997 ) 3143 [ http://arxiv.org/abs/hep-ph/9607467[hep-ph/9607467 ] ] ; s. king , s. moretti , and r. nevzorov , _ exceptional supersymmetric standard model _ , _ phys .
_ * b634 * ( 2006 ) 278 , [ http://arxiv.org/abs/hep-ph/0511256[hep-ph/0511256 ] ] .
see e.g.l .
basso , s. moretti , and g. m. pruna , _ theoretical constraints on the couplings of non - exotic minimal @xmath253 bosons _ , _ jhep _ * 1108 * ( 2011 ) 122 , [ http://arxiv.org/abs/1106.4762[arxiv:1106.4762 ] ] .
see e.g.m .
gonderinger , y. li , h. patel , and m. j. ramsey - musolf , _ vacuum stability , perturbativity , and scalar singlet dark matter _ , _ jhep _ * 01 * ( 2010 ) 053 , [ http://arxiv.org/abs/0910.3167[arxiv:0910.3167 ] ] .
see e.g.e .
ma , _ naturally small seesaw neutrino mass with no new physics beyond the tev scale _ , _ phys .
_ * 86 * ( 2001 ) 25022504 , [ http://arxiv.org/abs/hep-ph/0011121[hep-ph/0011121 ] ] ; s. m. davidson and h. e. logan , _ dirac neutrinos from a second higgs doublet _ , _ phys .
_ * d80 * ( 2009 ) 095008 , [ http://arxiv.org/abs/0906.3335[arxiv:0906.3335 ] ] .
r. barbieri , l. j. hall , and v. s. rychkov , _ improved naturalness with a heavy higgs : an alternative road to lhc physics _ , _ phys .
_ * d74 * ( 2006 ) 015007 , [ http://arxiv.org/abs/hep-ph/0603188[hep-ph/0603188 ] ] ; m. gustafsson , e. lundstrom , l. bergstrom , and j. edsjo , _ significant gamma lines from inert higgs dark matter _ , _ phys .
lett . _ * 99 * ( 2007 ) 041301 , [ http://arxiv.org/abs/astro-ph/0703512[astro-ph/0703512 ] ] .
see e.g.j .- o .
gong , h. m. lee , and s. k. kang , _ inflation and dark matter in two higgs doublet models _ , http://arxiv.org/abs/1202.0288[arxiv:1202.0288 ] .
see e.g.r .
schabinger and j. d. wells , _ a minimal spontaneously broken hidden sector and its impact on higgs boson physics at the large hadron collider _ , _ phys .
_ * d72 * ( 2005 ) 093007 , [ http://arxiv.org/abs/hep-ph/0509209[hep-ph/0509209 ] ] .
cheung , c .- h .
chou , and o. c. kong , _ muon anomalous magnetic moment , two higgs doublet model , and supersymmetry _ , _ phys . rev .
_ * d64 * ( 2001 ) 111301 , [ http://arxiv.org/abs/hep-ph/0103183[hep-ph/0103183 ] ] .
a. j. buras , m. v. carlucci , s. gori , and g. isidori , _ higgs - mediated fcncs : natural flavour conservation vs. minimal flavour violation _ , _ jhep _ * 1010 * ( 2010 ) 009 , [ http://arxiv.org/abs/1005.5310[arxiv:1005.5310 ] ] . p. ko , y. omura , and c. yu , _
chiral @xmath254 flavor models and flavored higgs doublets : the top fb asymmetry and the @xmath255 _ , _ jhep _ * 1201 * ( 2012 ) 147 , [ http://arxiv.org/abs/1108.4005[arxiv:1108.4005 ] ] . , g. aad _
_ , _ search for the standard model higgs boson in the diphoton decay channel with @xmath256 fb@xmath257 of @xmath157 collisions at @xmath258 tev with atlas _ , _ phys .
_ * 108 * ( 2012 ) 111803 , [ http://arxiv.org/abs/1202.1414[arxiv:1202.1414 ] ] .
, g. aad _
_ , _ combined search for the standard model higgs boson using up to 4.9 fb@xmath257 of @xmath157 collision data at @xmath259 tev with the atlas detector at the lhc _ , _ phys .
_ * b710 * ( 2012 ) 49 , [ http://arxiv.org/abs/1202.1408[arxiv:1202.1408 ] ] .
r. dermisek , _ light charged higgs and lepton universality in @xmath2 boson decays _ , http://arxiv.org/abs/0807.2135[arxiv:0807.2135 ] ; r. dermisek , _ light charged and @xmath0-odd higgses in mssm - like models _ , _ aip conf .
_ * 1078 * ( 2009 ) 226 , [ http://arxiv.org/abs/0809.3545[arxiv:0809.3545 ] ] ; k. j. bae , r. dermisek , d. kim , h. d. kim , and j .- h .
light higgs scenario in bmssm and lep precision data _ , http://arxiv.org/abs/1001.0623[arxiv:1001.0623 ] .
a. filipuzzi , j. portoles , and m. gonzalez - alonso , _ @xmath260 flavor symmetry and lepton universality violation in @xmath261 _ , http://arxiv.org/abs/1203.2092[arxiv:1203.2092 ] .
see e.g.v .
d. barger , j. hewett , and r. phillips , _ new constraints on the charged higgs sector in two higgs doublet models _ , _ phys .
_ * d41 * ( 1990 ) 3421 .
a. pilaftsis and c. e. m. wagner , _ higgs bosons in the minimal supersymmetric standard model with explicit @xmath0 violation _ , _ nucl .
_ * b553 * ( 1999 ) 3 , [ http://arxiv.org/abs/hep-ph/9902371[hep-ph/9902371 ] ] .
s. l. glashow and s. weinberg , _ natural conservation laws for neutral currents _
, _ phys .
_ * d15 * ( 1977 ) 1958 .
a. w. el kaffas , w. khater , o. m. ogreid , and p. osland , _ consistency of the two higgs doublet model and @xmath0 violation in top production at the lhc _ , _ nucl .
_ * b775 * ( 2007 ) 45 , [ http://arxiv.org/abs/hep-ph/0605142[hep-ph/0605142 ] ] .
see e.g.s .
kanemura , y. okada , e. senaha , and c .- p .
yuan , _ higgs coupling constants as a probe of new physics _ , _ phys .
_ * d70 * ( 2004 ) 115002 , [ http://arxiv.org/abs/hep-ph/0408364[hep-ph/0408364 ] ] .
atlas collaboration , _ an update to the combined search for the standard model higgs boson with the atlas detector at the lhc using up to 4.9 fb@xmath257 of @xmath157 collision data at @xmath259 tev _ , http://cdsweb.cern.ch/record/1430033[atlas-conf-2012-019 ] ; cms collaboration , m. pieri _ et al .
_ , _ searches for the standard model higgs boson at cms _ , http://arxiv.org/abs/1205.2907[arxiv:1205.2907 ] .
h. e. haber and h. e. logan , _ radiative corrections to the @xmath265 vertex and constraints on extended higgs sectors _ , _ phys .
rev . _ * d62 * , 015011 ( 2000 ) [ http://arxiv.org/abs/hep-ph/9909335[hep-ph/9909335 ] ] .
d. toussaint , _ renormalization effects from superheavy higgs particles _ , _ phys .
_ * d18 * , 1626 ( 1978 ) ; s. bertolini , _ quantum effects in a two higgs doublet model of the electroweak interactions _ , _ nucl
. phys . _ * b272 * , 77 ( 1986 ) ; w. hollik , _ nonstandard higgs bosons in @xmath266 radiative corrections _ , _ z. phys . _ * c32 * , 291 ( 1986 ) ; _ radiative corrections with two higgs doublets at lep / slc and hera _ , _ z. phys . _ * c37 * , 569 ( 1988 ) .
atlas collaboration , g. aad _
_ , _ search for charged higgs bosons decaying via @xmath267 in top quark pair events using @xmath157 collision data at @xmath259 tev with the atlas detector _ , http://arxiv.org/abs/1204.2760[arxiv:1204.2760 ] ; cms collaboration , s. chatrchyan _ et al .
_ , _ search for a light charged higgs boson in top quark decays in @xmath157 collisions at @xmath259 tev _ , http://arxiv.org/abs/1205.5736[arxiv:1205.5736 ] .
j. l. diaz - cruz and o. a. sampayo , _ contribution of gluon fusion to the production of charged higgs at hadron colliders _ , _ phys .
_ * d50 * ( 1994 ) 6820 .
a. djouadi , w. kilian , m. muhlleitner , and p. zerwas , _ production of neutral higgs boson pairs at lhc _ , _ eur .
j. _ * c10 * ( 1999 ) 45 , [ http://arxiv.org/abs/hep-ph/9904287[hep-ph/9904287 ] ] .
a. g. akeroyd and m. a. diaz , _ searching for a light fermiophobic higgs boson at the tevatron _ , _ phys .
_ * d67 * ( 2003 ) 095007 , [ http://arxiv.org/abs/hep-ph/0301203[hep-ph/0301203 ] ] ; a. akeroyd , m. a. diaz , and f. j. pacheco , _
double fermiophobic higgs boson production at the cern lhc and lc _ , _ phys .
_ * d70 * ( 2004 ) 075002 , [ http://arxiv.org/abs/hep-ph/0312231[hep-ph/0312231 ] ] .
i. f. ginzburg , _ light charged higgs at lhc _ , http://arxiv.org/abs/1205.5890[arxiv:1205.5890 ] .
a. v. semenov , _ lanhep : a package for automatic generation of feynman rules in gauge models _ , http://arxiv.org/abs/hep-ph/9608488[hep-ph/9608488 ] .
a. pukhov , _ calchep 2.3 : mssm , structure functions , event generation , batchs , and generation of matrix elements for other packages _ , http://arxiv.org/abs/hep-ph/0412191[hep-ph/0412191 ] ; http://theory.sinp.msu.ru/~pukhov/calchep.html[http://theory.sinp.msu.ru/ pukhov / calchep.html ] .
p. m. nadolsky , h .-
lai , q .- h .
cao , j. huston , j. pumplin , d. stump , w .- k .
tung and c .- p .
yuan , _ implications of cteq global analysis for collider observables _ , _ phys .
_ * d78 * ( 2008 ) 013004 [ http://arxiv.org/abs/0802.0007[arxiv:0802.0007 ] ] .
t. hahn and m. perez - victoria , _ automatized one loop calculations in four - dimensions and @xmath268-dimensions _ , _ comput .
_ * 118 * ( 1999 ) 153 , [ http://arxiv.org/abs/hep-ph/9807565[hep-ph/9807565 ] ] .
s. i. bityukov and n. v. krasnikov , _ on the observability of a signal above background _ , _ nucl .
instrum .
_ * a452 * ( 2000 ) 518 ; s. bityukov , s. erofeeva , n. krasnikov , and a. nikitenko , _ program for evaluation of significance , confidence intervals and limits by direct calculation of probabilities _ , in _ statistical problems in particle physics , astrophysics and cosmology : proceedings of phystat05 , oxford , uk , 1215 sep 2005 _
( l. lyons and m. k. nel , eds . ) , pp .
106107 , imperial college press , 2005 .
, _ combined cdf and d0 search for standard model higgs boson production with up to 10.0 fb@xmath257 of data _ , http://arxiv.org/abs/1203.3774[arxiv:1203.3774 ] .
n. bernal , d. lopez - val , and j. sola , _ single higgs - boson production through @xmath6 scattering within the general 2hdm _ , _ phys .
_ * b677 * ( 2009 ) 39 , [ http://arxiv.org/abs/0903.4978[arxiv:0903.4978 ] ] ; d. lopez - val and j. sola , _ single higgs - boson production at a photon - photon collider : general 2hdm versus mssm _ , _ phys .
_ * b702 * ( 2011 ) 246 , [ http://arxiv.org/abs/1106.3226[arxiv:1106.3226 ] ] .
a. arhrib , r. benbrik , and n. gaur , _
@xmath270 in inert higgs doublet model _ , http://arxiv.org/abs/1201.2644[arxiv:1201.2644 ] .
m. e. peskin and d. v. schroeder , _ an introduction to quantum field theory _ , addison - wesley , 1995 . | the large hadron collider is shown to have great scope for a light charged higgs discovery , in the context of the @xmath0-violating type - i two higgs doublet model .
this scenario with similar masses of @xmath1 and @xmath2 was suggested by the puzzling departure from charged current lepton universality found in the lep data . with the lightest neutral higgs mass set to @xmath3
, the charged - neutral higgs associated production mechanism can cause a significant excess in the @xmath4 events over a vast range of @xmath5 as long as the higgs mixing pattern avoids a few limiting cases .
thanks to the low @xmath1 mass , the charged higgs loop can play a striking role in neutral higgs decays into @xmath6 , thereby compensating for a suppressed gluon - gluon fusion rate .
the effect of scalar pseudo - scalar mixing on loop - induced higgs signals is also discussed . |
i would like to acknowledge y. jia and c.d .
l for fruitful discussions and kind encouragement , h.y .
cheng , c.k .
chua and x.h .
liu for useful discussions on the harmonic oscillator wave functions , y.m .
wang for reading the manuscript and sharing the results of @xmath82 in light - cone sum rules , and z.t .
wei for the collaboration in the light - front quark model and the comment on the sensitivity to light - front wave functions .
s. godfrey and j. l. rosner , phys .
d * 64 * , 074011 ( 2001 ) [ erratum - ibid .
d * 65 * , 039901 ( 2002 ) ] ; and many references therein ; for another new insight into hindered magnetic dipole transition , see : y. jia , j. xu and j. zhang , arxiv:0901.4021 [ hep - ph ] .
w. jaus , phys .
d * 41 * , 3394 ( 1990 ) ; phys .
d * 44 * , 2851 ( 1991 ) ; h. y. cheng , c. y. cheung and c. w. hwang , phys . rev .
d * 55 * , 1559 ( 1997 ) ; h. m. choi , c. r. ji and l. s. kisslinger , phys . rev .
d * 65 * , 074032 ( 2002 ) . h. y. cheng and c. k. chua , phys .
d * 69 * , 094007 ( 2004 ) [ arxiv : hep - ph/0401141 ] . c. d. lu , w. wang and z. t. wei , phys .
d * 76 * , 014013 ( 2007 ) [ arxiv : hep - ph/0701265 ] ; w. wang , y. l. shen and c. d. lu , eur .
j. c * 51 * , 841 ( 2007 ) [ arxiv:0704.2493 [ hep - ph ] ] ; w. wang and y. l. shen , phys .
d * 78 * , 054002 ( 2008 ) ; w. wang , y. l. shen and c. d. lu , phys .
d * 79 * , 054012 ( 2009 ) [ arxiv:0811.3748 [ hep - ph ] ] ; x. x. wang , w. wang and c. d. lu , phys .
d * 79 * , 114018 ( 2009 ) [ arxiv:0901.1934 [ hep - ph ] ] ; c. h. chen , y. l. shen and w. wang , arxiv:0911.2875 [ hep - ph ] .
h. w. ke , x. q. li and z. t. wei , phys . rev .
d * 80 * , 074030 ( 2009 ) [ arxiv:0907.5465 [ hep - ph ] ] .
h. w. ke , x. q. li and z. t. wei , arxiv:0912.4094 [ hep - ph ] .
h. y. cheng and c. k. chua , arxiv:0909.4627 [ hep - ph ] . | hindered magnetic dipole transitions @xmath0 are studied in the covariant light - front approach . compared with the allowed magnetic dipole transitions , we find that results for hindered magnetic dipole transitions are sensitive to heavy quark mass and shape parameters of the light - front wave functions .
it is possible to tune the parameters so that the predictions of branching fractions of @xmath1 are consistent with the recent experimental data , but the relevant decay constant of @xmath2 is much smaller than that of @xmath3 .
we also generalize the investigation to the charmonium sector and find the the same conclusion .
since the photon energy in allowed magnetic dipole ( m1 ) transitions is limited , hindered m1 transitions with changes of principal quantum numbers are believed to offer better chances to discover the pseudoscalar quarkoniums such as @xmath2 .
several years ago , the cleo collaboration set the upper bounds for this kind of decays @xcite @xmath4 in 2008 the babar collaboration observed a peak in the photon energy spectrum at @xmath5 mev in radiative @xmath6 decay @xcite .
this is viewed as the first observation of the @xmath2 meson which is the lowest bottomonium ground state .
the branching ratio of this radiative decay is @xmath7 the measured mass @xmath8 gives the mass split @xmath9 subsequently the @xmath2 meson has also been observed in the radiative @xmath10 decay @xcite @xmath11 the ratio of the two branching fractions is measured as @xmath12 the recently updated results by the cleo collaboration @xcite are consistent with measurements by the babar collaboration taking into account the uncertainties @xmath13 decay widths of m1 transitions can be expressed in a well - known formula @xmath14 @xmath15 is the fine - structure constant , @xmath16 denotes the charge in unit of @xmath17 of the transition quark , and @xmath18 denotes the quark mass .
@xmath19 is the energy of the photon in the vector meson rest frame . for allowed transitions @xmath20 ,
the emitted photon is typically soft .
one can expand the spherical bessel function @xmath21 by @xmath19 .
radiative corrections and nonrelativistic corrections can be systematically studied in the effective field theory @xcite .
in particular the predicted branching ratio of @xmath22 is well consistent with the experimental data .
the emitted photon in hindered m1 transitions is rather energetic and the convergence of the expansion in @xmath19 becomes poor . on the other aspect ,
the decay amplitude is zero at the leading power in @xmath19 since the different wave functions are orthogonal .
higher power terms contribute and they are sensitive to the treatments of the nonperturbative dynamics in transition form factors . for more than twenty years , hindered m1 transitions have received extensive interests but theoretical predictions vary over several orders @xcite .
the recent experimental data could offer an opportunity to investigate the dynamics in the quarkonia transition and in particular might be helpful to constrain the form of the wave functions .
in this work , we use the covariant light - front quark model ( lfqm ) to investigate the m1 transitions and examine whether the commonly - used light - front wave functions ( lfwf ) can simultaneously explain the experimental data for both allowed and hindered m1 transition .
the light - front qcd may be the only potential candidate to reconcile the low energy quark model and the high energy parton model .
the lfqm @xcite can give a full treatment on spins of hadrons using the so - called melosh transformation .
physical quantities are represented as the overlap of lfwf .
these wave functions are expressed in terms of the internal variables of the quark and gluon degrees and thus are manifestly lorentz invariant . to preserve the covariance and remove the dependence of physical quantities on the direction of the light - front , jaus proposed the covariant lfqm in which the zero - mode contributions are systematically included @xcite .
the application of the covariant lfqm to the s - wave and p - wave decay constants and form factors is very successful @xcite . under this framework ,
the transition @xmath23 has already been studied in ref .
@xcite ( see also ref .
@xcite ) . in the following
we employ the light - front decomposition of the momentum @xmath24 , where @xmath25 .
these momenta can be expressed in terms of the internal variables @xmath26 as : @xmath27 with @xmath28 . with these internal variables
, one can define some useful quantities @xmath29 here @xmath30 can be interpreted as the energy of the quark / antiquark and @xmath31 can be viewed as kinetic invariant mass of the meson system .
the transition amplitude of @xmath32 is usually parametrized as @xmath33 where the photon is firstly taken as off - shell @xmath34 and the convention @xmath35 is adopted .
feynman diagrams for this process are given in fig .
[ fig : feyn ] , in which the photon is emitted from the quark shown in ( a ) or from the antiquark shown in ( b ) . due to the charge conjugation invariance
, these two diagrams provide identical contributions .
it is straightforward to evaluate these two diagrams : @xmath36 where @xmath37 , @xmath38 and @xmath39,\\ s_{\mu\nu}^b&=&{\rm tr}\left[\left(\gamma_\nu-\frac{p_{1\nu}-p_{2\nu}}{w_v}\right)(-p\!\!\!\slash_2+m_2 ) \gamma_\mu(-p\!\!\!\slash'_2+m_2 ) \gamma_5(p\!\!\!\slash_1+m_1)\right].\end{aligned}\ ] ] functions @xmath40 , depending on the four momentum of the internal quarks , are the wave functions for vector and pseudoscalar mesons . in the absence of singularity in the @xmath40 , integrating over the minus component of the internal momentum will pick up the pole in the propagators .
then the functions @xmath41 are replaced by @xmath42 which only depends on the plus component and the transverse momentum and @xmath43 is replaced by @xmath44 @xmath45 where @xmath46 denotes the momentum distribution inside the meson . in order to preserve the covariance ,
it is suggested that the zero mode contributions should be added in the covariant lfqm .
the inclusion of the zero - mode contribution corresponds to a proper way to integrate out the minus component and after the integration expressions for the form factor @xmath47 are given as follows @xcite @xmath48 and the decay width of @xmath49 is then evaluated as @xmath50 in the @xmath51 limit , the heavy quark inside the quarkonium moves nonrelativistically .
integrating out the hard off - shell degrees of freedom , one reaches an effective field theory known as nonrelativistic qcd @xcite to deal with the low energy dynamics . in a heavy quarkonium ,
the square of the transverse momentum @xmath52 is of the order @xmath53 and the momentum fraction of the heavy quark @xmath54 .
the kinetic invariant mass @xmath55 is roughly @xmath56 . at the leading power in @xmath57 , the form factor @xmath58 in eq .
( [ eq : formfactor - lfqm ] ) is reduced to @xmath59 this formula , with the lfwf discussed in the following , could formally reproduce the leading power behavior shown in eq .
( [ eq : m1-space ] ) in the momentum space . .
] for hindered m1 transition , the transition amplitude is from the higher power corrections and one needs a specific approach to calculate them .
as we can see from the above discussion , the form factors in the covariant lfqm are expressed as the overlap of the lfwf .
it contains the nonperturbaitve coupling of the quark pair with a meson . except for some limited cases ,
these quantities can not be derived from the first principle .
practically one often adopts a phenomenological form .
the commonly - used wave function for the @xmath60-wave mesons with lowest principle number is the gaussian - type @xmath61 in the case of the wave functions of the @xmath62 mesons , we adopt the solutions of the harmonic - oscillator problem @xmath63 the fourier transformations , with the jacobi determinant @xmath64 , are derived as @xmath65 this kind of lfwf for the @xmath66 mesons has been used in ref .
@xcite while its light - cone form is studied in ref .
. it would be meaningful to use the allowed and hindered m1 transitions to constrain the forms of the lfwf and in particular to test the harmonic - oscillator wave functions given above , as most of the previous studies in the covariant lfqm mainly focus on mesons without any radial excitation .
.input parameters in the light - front wave functions ( in units of gev ) .
[ cols="^,^,^,^,^",options="header " , ] as another check , we will extend the above discussion into the charmonium sector : @xmath67 .
the relevant experimental data is given as @xcite @xmath68 again with proper inputs the results in table [ tab : results - etac ] could be consistent with these data , but the relevant decay constant of @xmath69 is much smaller that of its vector partner @xmath70 .
our result for @xmath71 @xmath72 is approaching the experimental upper bound .
the first uncertainties are from @xmath73gev and the second uncertainties are from @xmath74 gev . in ref .
@xcite , the inputs @xmath75 gev , @xmath76 gev and @xmath77 gev are adopted and prediction of @xmath78 is consistent with the data reported by the cleo collaboration @xcite .
if we adopt these parameters , the branching ratio of @xmath79 is larger than the experimental data by two orders @xmath80 inspired by these inconsistencies , one may get the same conclusion as in the bottomonium sector .
in conclusion , we studied the hindered m1 transitions @xmath79 and the @xmath81 in the covariant light - front quark model . compared with the allowed m1 transitions , theoretical results for hindered channels
are found to strongly depend on quark masses and the shape of the light - front wave functions .
because of this sensitivity , these transitions offer good laboratories to study the dynamics between the transitions and constrain the forms of the light - front wave functions .
using the harmonic - oscillator wave functions and the experimental data , our results show that the decay constants of the pseudoscalar mesons are required to be much smaller than the decay constants of their vector partners . with the precise decay constants measured in various other channels in the future
, different forms of the wave functions could be directly tested in the channels studied here .
a more practical way in the future is to incorporate more radiative transitions between heavy quarkonium states as in ref .
@xcite .
_ note added _ : during the long - time revision of this work , a similar paper @xcite appears .
one of the important differences is : we have shown that hindered m1 transitions are sensitive to the shape of the wave functions as they should . |
electrons are described by the dirac equation where the u(1 ) maxwell electromagnetic field ( emf ) @xmath0 is coupled to the charge current @xmath1 as described by the lagrangian ( in the natural unit where @xmath2 ; @xmath3 ) @xcite @xmath4 \psi.\ ] ] where @xmath5 is the covariant derivative , @xmath6 is the electron mass .
note that the spin is encoded by 4 component nature of the spinors @xmath7 and @xmath8 and the @xmath9 gamma matrices @xmath10 , but the charge and charge current alone determine the electromagnetic properties of the electrons , which are given by @xmath11 in condensed matter physics , on the other hand , the low energy phenomena compared with the mass gap @xmath12 are considered , and only the positive energy states described by the two - component spinor are relevant .
then , the relativistic spin - orbit interaction originates when the negative energy states ( positron stats ) are projected out to derive the effective hamiltonian or lagrangian .
the projection to a subspace of the hilbert space leads to the nontrivial geometrical structure which is often described by the gauge theory .
this is also the case for the dirac equation , and the resultant gauge field is su(2 ) non - abelian gauge field corresponding to the zeeman effect ( time - component ) and the spin - orbit interaction ( spatial components ) as described below .
the effective lagrangian for the positive energy states can be derived by the expansion with respect to @xmath13 @xcite @xmath14 \psi , \label{eq : lag}\ ] ] where @xmath7 is now the two - component spinor and @xmath15 , and @xmath16 @xmath17 are the gauge covariant derivatives with @xmath18 being the quantity proportional to the bohr magneton @xcite .
@xmath0 is the maxwell emf , and the su(2 ) gauge potential are defined as @xmath19 and @xmath20 represent the pauli matrices .
the su(2 ) gauge field is coupled to the 4-component spin current @xmath21.\end{aligned}\ ] ] namely , the zeeman coupling and the spin - orbit interaction can be regarded as the gauge coupling between the 4-spin current and the su(2 ) gauge potential .
( the spin current is the tensor quantity with one suffix for the direction of the spin polarization while the other for the direction of the flow . )
note that the system has no su(2 ) gauge symmetry since the `` vector potential '' @xmath22 is given by the physical field strength @xmath23 and @xmath24 , i.e. , the relation @xmath25 automatically holds .
this fact is connected to the absence of the conservation law for the spin density and spin current density in the presence of the relativistic spin - orbit interaction . in the spherically symmetric systems , the total angular momentum , i.e.
, the sum of the orbital and spin angular momenta , is conserved , but the rotational symmetry is usually broken by the periodic or disorder potential @xmath26 in condensed matter systems .
therefore , it is usually assumed that the conservation law of spin is lost by the spin - orbit interaction .
however , it is noted that the spin and spin current densities are `` covariantly '' conserved as described by the `` continuity equation '' @xcite @xmath27 replacing the usual derivative @xmath28 by the covariant derivative @xmath29 .
this suggest that the conservation law holds in the co - moving frame , but the crucial issue is how to translate this law to the laboratory frame , which is the issue addressed in this paper . note again that the su(2 ) gauge symmetry is absent in the present problem , and hence the lagrangian like @xmath30 , which usually leads to the generalized maxwell equation and also to the conservation law of 4 spin current including both the matter field and gauge field @xcite , is missing . instead
, we will regard @xmath22 as the frozen background gauge field , and focus on the quantum dynamics of noninteracting electrons only . in this paper , we derive the hidden conservation law by defining the `` twisted '' spin and spin current densities which satisfy the continuity equation with the usual derivative @xmath28 .
the description is intended to be pedagogical and self - contained . for this purpose ,
the theoretical techniques developed in high energy physics is useful .
the essential idea is to take into account the effect of the background gauge field in terms of the noncommutative geometry generalizing the concept of `` product '' .
this is achieved by extending the usual lie algebra to hopf algebra .
usually , a conservation law is derived from symmetry of an action , i.e. , noether s theorem .
the symmetry in the noncommutative geometry is called as a `` twisted '' symmetry , and this symmetry and the corresponding generalized noether s theorem have been studied in the high energy physics .
seiberg and witten proposed that an equivalence of a certain string theory and a certain field theory in noncommutative geometry @xcite .
since then , the noncommutative geometry have been attracted many researchers . on the other hand
, it is known that the poincare symmetry is broken in a field theory on a noncommutative geometry .
it is a serious problem because the energy and momentum can not be defined .
m. chaichian , _ et al_. proposed the twisted symmetry in the minkowski spacetime , and alleged that the twisted poincare symmetry is substituted for the poincare symmetry @xcite .
moreover , g. amelino - camelia , _ et al_. discussed noether s theorem in the noncommutative geometry @xcite . as we will discuss in detail later , a certain type of a noncommutative geometry space is equal to a spin - orbit coupled system
. therefore , a global su(2 ) gauge symmetry in the noncommutative geometry space gives a noether current corresponding to the `` twisted '' spin and spin current in the spin - orbit coupling system .
this enables us to derive the generalized noether s theorem for the twisted spin and spin current densities .
now some remarks about the application is in order .
spintronics is an emerging field of electronics where the role of charge and charge current are replaced by the spin and spin current aiming at the low energy cost functions @xcite .
the relativistic spin - orbit interaction plays the key role there since it enables the manipulation of spins by the electric field .
however , this very spin - orbit interaction introduces the spin relaxation which destroys the spin information in sharp contrast to the case of charge where the information is protected by the conservation law .
therefore , it has been believed that the spintronics is possible in a short time - scale or the small size devices .
the discovery of the conservation law of twisted spin and spin current densities means that the quantum information of spin is preserved by this hidden conservation law , and could be recovered .
actually , it has been recently predicted that the adiabatic change in the spin - orbit interaction leads to the recovery of the spin moment called spin - orbit echo @xcite .
therefore , the conservation law of the twisted spin and spin current densities is directly related to the applications in spintronics .
the plan of this paper follows ( see fig . [
fig : flow ] ) . in section
[ sec : noether ] , we review the conventional noether s theorem , and describe briefly its generalization to motivate the use of hopf algebra and deformation quantization . in section [ sec : hopf ] , the hopf algebra is introduced , and section [ sec : star ] gives the explanation of the deformation quantization with the star product . the gauge interaction is compactly taken into account in the definition of the star product .
these two sections are sort of short review for the self - containedness and do not contain any original results except the derivation of the star product with gauge interaction .
section [ section : ts ] is the main body of this paper . by combining the hopf algebra and the deformation quantization , we present the derivation of the conserved twisted spin and spin current densities .
section [ sec : conc ] is a brief summary of the paper and contains the possible new directions for future studies .
the readers familiar with the noncommutative geometry and deformation quantization can skip sections [ sec : hopf ] , [ sec : star ] , and directly go to section [ section : ts ] .
in this section , we discuss noether s theorem @xcite , and its generalization as a motivation to introduce the hopf algebra and deformation quantization . in section [ subsec :
con ] , we will recall noether theorem , and rewrite it using the so - called `` coproduct '' , which is an element of the hopf algebra . in section [ subsec : general ] , we will sketch a derivation of generalized noether theorem .
we start with the action @xmath31 given by @xmath32 where @xmath33 represents a range of the spacetime coordinate @xmath34 ( @xmath35 ) with a dimension dim , i.e. , @xmath36 is the dimension of the space , @xmath37 describes a lagrangian density , and @xmath38 @xmath39 represents light speed .
we introduce a field @xmath40 with internal degree of freedom @xmath41 , and infinitesimal transformations : @xmath42 where we characterize the transformations by the subscript ; specifically , @xmath43 represents a general infinitesimal transformation .
hereafter , we will employ einstein summation convention , i.e. , @xmath44 with vectors @xmath45 and @xmath46 @xmath47 , and the minkowski metric : @xmath48 .
we define the variation operator of the action as follow : @xmath49 where we characterize this variation by @xmath43 , because this variation is derived from the infinitesimal transformations eqs .
( [ eq : tfx ] ) and ( [ eq : tff ] ) . since the integration variable @xmath50 can be replaced by @xmath34 , eq .
( [ def_action.eq ] ) is @xmath51\nonumber\\ & = & \int d^{\rm dim}xh_{\delta\omega}(x){\cal l}^\prime(x ) + \int d^{\rm dim}xh_\omega(x)\left[{\cal l}^\prime(x ) -{\cal l}(x)\right],\end{aligned}\ ] ] where @xmath52 and @xmath53 .
therefore , we obtain the following equation through partial integration : @xmath54 + { \cal o}((\delta_\zeta x)^2 ) , \label{eq : variationaction}\end{aligned}\ ] ] where we have introduced the so - called lie derivative : @xmath55 and we replaced @xmath56 by @xmath57 due to @xmath58 . hereafter , we assume that the action is invariant under the infinitesimal transformations eqs .
( [ eq : tfx ] ) and ( [ eq : tff ] ) . in the case where the lagrangian density is a function of @xmath40 and @xmath59 , i.e. , @xmath60 $ ] ,
the lie derivative of the lagrangian is given by @xmath61- { \cal l}\left[\phi_r(x),\partial_\mu\phi_r(x)\right]\nonumber\\ & = & \left\ { \frac{\partial { \cal l}}{\partial\phi_r}\delta^{\rm l}_\zeta\phi_r+ \frac{\partial { \cal l}}{\partial\left(\partial_\mu\phi_r\right ) } \partial_\mu\delta^{\rm l}_\zeta\phi_r\right\}\nonumber\\ & = & \left(\frac{\partial{\cal l}}{\partial\phi_r } -\partial_\mu\frac{\partial { \cal l}}{\partial(\partial_\mu\phi_r)}\right ) \delta^{\rm
l}_\zeta\phi_r + \partial_\mu \left(\frac{\partial{\cal l}}{\partial(\partial_\mu\phi_r ) } \delta^{\rm l}_\zeta\phi_r\right ) , \label{lie.eq}\end{aligned}\ ] ] and the variation of the action is calculated by @xmath62 if we require that @xmath63 and @xmath64 vanish on the surface @xmath65 , we obtain the euler - lagrange equation . on the other hand ,
if we require that fields @xmath40 satisfy the euler - lagrange equation , we obtain continuity equation @xmath66 with a noether current @xmath67 hereafter let us discuss an infinitesimal global u(1)@xmath68su(2 ) gauge transformation and infinitesimal translation and rotation transformations , which are denoted by @xmath69 in this paper .
variations in terms of @xmath69 are defined by @xmath70 with an infinitesimal parameter @xmath71 , and symmetry generators @xmath72 and @xmath73 . 1 . for the global u(1)@xmath68su(2 ) gauge transformation , @xmath74 , @xmath75 , and @xmath76 ( @xmath77 ; @xmath78 ) , where @xmath79 , and @xmath80 with the planck constant @xmath81 and pauli matrices @xmath82 .
2 . for the translation , @xmath83 , @xmath84 and @xmath85 with an infinitesimal parameter @xmath86 and the momentum operator @xmath87 ( @xmath88 ; @xmath78 ) .
3 . for the rotation , @xmath89 , @xmath90 , and @xmath91 , which corresponds to the angular momentum tensor ( @xmath88 ; @xmath78 ) . for these transformations , equation
@xmath92 is satisfied .
this can be seen explicitly as follows .
the variations of space coordinates of the global u(1)@xmath68su(2 ) and the translation transformations are given by @xmath93 or @xmath94 , respectively , and thus @xmath92 is trivial .
the variation of the rotation transformation is given by @xmath95 , therefore @xmath96 .
we consider a variation of the lagrangian density ; @xmath97 note that eq .
( [ eq : vl ] ) is correct for any infinitesimal transformation .
here we consider the global u(1)@xmath68su(2 ) gauge transformation and/or the translation and rotation transformations @xmath98 . because @xmath92 , we obtain the following equation : @xmath99 from eqs.([eq : variationaction ] ) and ( [ eq : variationlagrangian ] ) , one can see @xmath100 where @xmath101 denotes that the type of the variation in eq .
( [ eq : variationaction ] ) is restricted to the global u(1)@xmath68su(2 ) or poincare transformations .
( for simplicity we omitted the subscript @xmath33 in the integral ) .
finally , for @xmath102 , the variation of the action is equal to the variation of lagrangian .
this fact will be used later in section [ section : ts ] where the variation of the lagrangian density instead of the action will be considered .
now , we would like to introduce a hopf algebra for the purpose of generalizing noether s theorem @xcite . at first , we rewrite noether s theorem in section [ sec : noether ] by using the hopf algebra , and next , we introduce a twisted symmetry @xcite . for simplicity ,
we only consider the global u(1)@xmath68su(2 ) gauge symmetry and the poincare symmetry .
we assume that the lagrangian density is written as @xmath103 with a field @xmath104 , a hermitian conjugate @xmath105 , and an single - particle lagrangian density operator @xmath106 , which is a 2@xmath682 matrix ; the overline represents the complex conjugate .
the action can be rewritten as @xmath107 where `` @xmath108 '' represents the trace in the spin space , @xmath109 , @xmath110 , and @xmath111 represents the convolution integral : @xmath112 with smooth two - variable functions @xmath113 and @xmath114 .
the variation operator @xmath98 of the action can be also rewritten as @xmath115\nonumber\\ & = & { \rm tr}\int d^{\rm dim}x_1d^{\rm dim}x_2\left [ \tilde{{\cal l}}(x_2,x_1 ) i\vartheta\xi g(x_1,x_2)-i\vartheta\xi \tilde{{\cal l}}(x_2,x_1)g(x_1,x_2)\right]\nonumber\\ \label{eq : variationxi}\end{aligned}\ ] ] with @xmath116 ; in addition , we assumed that the single - particle lagrangian density operator is invariant under the infinitesimal transformation @xmath98 . here
, we introduce grassmann numbers @xmath117 and @xmath118 ; an integral is defined by @xmath119 .
the variation of the right - hand side of eq.([eq : variationxi ] ) can be rewritten as follow : @xmath120\nonumber\\ & = & -{\rm i}{\rm tr}\int d\theta_1d\theta_2d^{\rm dim}x_1d^{\rm dim}x_2 \mu\circ(\mu\otimes\mbox{id})\nonumber\\ & & \circ \left[\left ( \theta_1\tilde{\cal l}(x_2,x_1)\otimes\vartheta\theta_2\xi + \vartheta\theta_2\xi\otimes\theta_1\tilde{\cal l}(x_2,x_1 ) \right)\otimes g(x_1,x_2)\right]\nonumber\\ & = & -{\rm i}{\rm tr}\int d\theta_1d\theta_2d^{\rm dim}x_1d^{\rm dim}x_2 \mu\circ(\mu\otimes { \rm id})\left[\triangle(\vartheta\theta_2\xi)\circ \left(\theta_1\tilde{\cal l}(x_2,x_1)\otimes g(x_1,x_2)\right)\right]\nonumber\\ & \equiv&{\rm \hat{t}r}\left[\triangle(\vartheta\theta_2\xi)\circ ( ( \theta_1\tilde{\cal l})\otimes g)\right],\label{eq : variableofactionbycoproduct}\end{aligned}\ ] ] where @xmath121 and @xmath122 represent a tensor product and a product of operators , respectively .
the operator @xmath123 denotes the transformation of the tensor product to the usual product @xmath124 , and @xmath125 represents a coproduct : @xmath126 where @xmath43 and @xmath127 represent a certain operator and the identity map , respectively .
these operators constitutes the hopf algebra as will be explained in the next section .
moreover , we have defined @xmath128 .
we emphasize here that the variation is written by the coproduct @xmath125 , which is important to formulate the generalized noether theorem in the presence of the gauge potential .
the coproduct determines an operation rule of a variation operator ; for example , the coproduct ( [ eq : coproductz ] ) represents the leibniz rule .
a twisted symmetry transformation is given by deformation of the coproduct .
we now sketch the concept of the twisted symmetry in deformation quantization @xcite .
first , we assume that the variation of action @xmath129 is zero , i.e. , @xmath43 represents the symmetry transformation of the system corresponding to the action @xmath130 .
next , we consider the action @xmath131 with external gauge fields @xmath132 .
usually , external gauge fields breaks symmetries of @xmath130 , i.e. , @xmath133 .
here we introduce a map : @xmath134 , which will be defined in section [ sec : twistedelement ] .
the basic idea is to generalize the `` product '' taking into account the gauge interaction . using this map ,
the variation is rewritten as @xmath135 . on the other hand ,
when the twisted symmetry @xmath136 can be defined , we obtain the following equation : @xmath137 namely , @xmath138 corresponds to a symmetry with external gauge fields . in the expression for the variation of action in terms of the hopf algebra eq.([eq : variableofactionbycoproduct ] ) , we can replace @xmath139 by @xmath140 corresponding to the change from @xmath141 to @xmath138 as shown in section [ section : ts ] .
this is achieved by using the hopf algebra and the deformation quantization , which will be explained in sections [ sec : hopf ] and [ sec : star ] , respectively .
therefore , we can generalize the noether s theorem and derive the conservation law even in the presence of the gauge field @xmath132 .
here we introduce a hopf algebra . first , we rewrite the algebra using tensor and linear maps . secondly
, a coalgebra is defined using diagrams corresponding to the algebra .
finally , we define a dual - algebra and hopf algebra .
we define the algebra as a @xmath142-vector space @xmath143 having product @xmath123 and unit @xmath144 . here
, @xmath142 represents a field such as the complex number or real number . in this paper
, we consider @xmath143 as the space of functions or operators .
a space of linear maps from a vector space @xmath145 to a vector space @xmath146 is written as @xmath147 .
a product @xmath123 is a bilinear map : @xmath148 , i.e. , @xmath149 and a unit is a linear map : @xmath150 , i.e. , @xmath151 with @xmath152 and @xmath153 .
here @xmath123 and @xmath144 satisfies @xmath154 @xmath155 @xmath156 with @xmath157 and @xmath158 .
the product @xmath123 has the association property , which is written as @xmath159 . because the left - hand side and the right - hand side of the previous equation give the following equations : @xmath160 and @xmath161 for all @xmath162 , then @xmath163 is equal to the association property @xmath164 .
this property is illustrated as the following diagram : @xmath165 here @xmath166 denotes that this graph is the commutative diagram .
the unit @xmath144 satisfies the following equation : @xmath167 .
since the left - hand side and the right - hand side of the previous equation give the following equations @xmath168 and @xmath169 for all @xmath170 and @xmath153 , and @xmath171 , then the unit can be written as @xmath172 .
note that @xmath173 , where @xmath174 represents the equivalence relation , i.e. , @xmath175 denotes that @xmath176 and @xmath177 are identified .
this property is illustrated as : @xmath178 algebra is defined as a set @xmath179 .
a coalgebra is defined by reversing the direction of the arrows in the diagrams corresponding to the algebra .
thus , we will define a coproduct @xmath180 and counit @xmath181 with a @xmath142-vector space @xmath143 .
a coproduct is a bilinear map from @xmath143 to @xmath182 : @xmath183 and satisfies co - association property : @xmath184 namely , @xmath185 ( compare the diagram corresponding to the association property and that corresponding to the co - association property ) . a counit @xmath186 is a linear map from @xmath143 to field @xmath142 : @xmath187 and satisfies the following diagram : @xmath188 namely , @xmath189 where @xmath173 .
since @xmath125 and @xmath186 are linear maps , @xmath125 and @xmath186 satisfy @xmath190 @xmath191 with @xmath192 and @xmath153 .
note that @xmath173 and @xmath193 .
a coalgebra is defined as a set @xmath194 .
for example , in the vector space @xmath195 , we define a coproduct @xmath196 and @xmath197 , and a counit @xmath198 and @xmath199 .
the set @xmath200 is coalgebra , because this set satisfies the equations : @xmath201 and @xmath202 .
because the coproduct and counit are linear map , we only check the above equations with respect to @xmath203 and @xmath204 . for @xmath203 , @xmath205 and @xmath206 therefore , @xmath207 .
moreover , @xmath208 and @xmath209 therefore @xmath210 . for @xmath211 , @xmath212 and @xmath213 therefore , @xmath214 .
finally , @xmath215 and @xmath216 therefore , @xmath217 .
namely , the set @xmath200 is the coalgebra .
note that @xmath218 corresponds to the product with a constant : @xmath219 , where we have used the coproduct @xmath197 at the final equal sign .
here @xmath220 are smooth functions , @xmath221 is included in the function space , and @xmath222 .
@xmath223 represents the leibniz rule : @xmath224 , where we have used the coproduct @xmath225 at the last equal sign .
@xmath226 and @xmath227 represent the filtering action to a constant function : @xmath228 and @xmath229 , respectively .
a dual - algebra is the set of an algebra and a coalgebra , i.e. , the set of @xmath230 . on a dual - algebra
, we define a @xmath231-product as @xmath232 with @xmath233 .
we define an antipode @xmath234 which satisfies the following equation : @xmath235 where @xmath236 corresponds to the identity mapping , i.e. , @xmath237 is an inverse of unit . for example , @xmath238 in the set @xmath239 is defined as @xmath240 and @xmath241 . for @xmath203 , @xmath242 and @xmath243 therefore
, we obtain @xmath244 . for @xmath204 , @xmath245 and @xmath246 therefore , we obtain @xmath247 .
namely , @xmath239 is the hopf algebra . a dual - algebra with an antipode @xmath237 , i.e. , @xmath248 ,
is called a hopf algebra . by using the approach similar to a coproduct and counit
, we can define a codifferential operator @xmath249 from a diagram of the differential @xmath250 .
the differential @xmath204 is the linear map : @xmath251 and satisfies leibniz rule @xmath252 which is illustrated as @xmath253 a codifferential operator @xmath254 is a linear map ; @xmath255 , and satisfies the following diagrams : @xmath256 namely , a codifferential operator @xmath254 satisfies @xmath257 . in section [ sub : l_infty ] , the codifferential operator will be introduced .
in this section , we explain the deformation quantization using the noncommutative product encoding the commutation relationships . at first , in section [ sec : wr ] , we introduce the so - called wigner representation and wigner space , and show that a product in the wigner space is noncommutative
. this product is called moyal product and it guarantees the commutation relationship of the coordinate and canonical momentum .
next , we add spin functions and background gauge fields to the wigner space , and rewrite the coordinates of wigner space as a set of spacetime coordinates @xmath258 , mechanical momenta @xmath259 , and spins @xmath260 ( @xmath259 includes the background gauge fields ) . to generalize the moyal product for the deformed wigner space , which is a set of function defined on @xmath261 , we explain the general constructing method of the noncommutative product in section [ sec : sp ] ; the noncommutative product is the generalized moyal product , which is called `` star product '' . this constructing method is given as a map from a poisson bracket in the wigner space to the noncommutative product ( see section [ sec : sp ] ) , and we see the condition of this deformation quantization map in section [ sec : sp ] .
this map is described by the path integral of a two - dimensional field theory , which is called the topological string theory . in section [ sec : topologicalstringtheory ] , we explain this topological string theory , and in section [ sec : edqtst ] , we discuss the perturbative treatment of this theory . in section [ sec : diagramrules ] , we summarize the diagram technique
. finally , in section [ sec : twistedelement ] , we construct the star product in @xmath261 space .
we note that the star product guarantees the background gauge structure .
we start with the introduction of the wigner representation . from equation ( [ eq : action ] ) , a natural product is the convolution integral : @xmath262 where @xmath263 with a two spacetime arguments function space @xmath264 .
here we introduce the center of mass coordinate @xmath258 and the relative coordinate @xmath265 as follows : @xmath266 moreover we employ the following fourier transformation : @xmath267 now we define the wigner space : @xmath268~|~f\in{\cal g}\}$ ] @xcite . in this space ,
the convolution is transformed to the so - called moyal product @xcite : @xmath269 because @xmath270 & = & \int\frac{d^{\rm dim}p}{(2\pi\hbar)^{\rm dim } } e^{ip_\nu\xi^\nu/\hbar } \left\{f(x , p ) e^{\frac{{\rm i}\hbar}{2}\left({\overleftarrow \partial}_{x^\nu}{\overrightarrow \partial}_{p_\nu}-{\overleftarrow \partial}_{p_\nu}{\overrightarrow \partial}_{x^\nu}\right)}g(x , p)\right\}\nonumber\\ & = & \int\frac{d^{\rm dim}p}{(2\pi\hbar)^{\rm dim}}d^{\rm dim}\xi_1d^{\rm dim}\xi_2e^{ip_\nu\xi^\nu/\hbar } \big\{e^{-ip_\nu\xi_1/\hbar}f(x+\xi_1/2,x-\xi_1/2)\nonumber\\ & & \times e^{\frac{{\rm i}\hbar}{2}\left({\overleftarrow \partial}_{x^\nu}{\overrightarrow \partial}_{p_\nu}-{\overleftarrow \partial}_{p_\nu}{\overrightarrow \partial}_{x^\nu}\right)}e^{-ip_\nu\xi^\nu_2/\hbar } g(x+\xi_2/2,x-\xi_2/2)\big\}\nonumber\\ & = & \int\frac{d^{\rm dim}p}{(2\pi\hbar)^{\rm dim}}d^{\rm dim}\xi_1d^{\rm dim}\xi_2 e^{ip_\nu\left(\xi^\nu-\xi^\nu_1-\xi^\nu_2\right)/\hbar}\nonumber\\ & & \times f(x+\xi_1/2,x-\xi_1/2)e^{\frac{1}{2 } \left({\overleftarrow\partial}_{x^\nu}\xi^\nu_2- \xi^\nu_1{\overrightarrow\partial}_{x^\nu}\right ) } g(x+\xi_2/2,x-\xi_2/2)\nonumber\\ & = & \int d^{\rm dim}\xi_1d^{\rm dim}\xi_2\delta(\xi-\xi_1-\xi_2)\nonumber\\ & & \times f\left(x+\frac{\xi_1+\xi_2}{2 } , x-\frac{\xi_1-\xi_2}{2}\right ) g\left(x-\frac{\xi_1-\xi_2}{2 } , x-\frac{\xi_1+\xi_2}{2}\right)\nonumber\\ & = & \int d^{\rm dim}x_+d^{\rm dim}x_-\delta(\xi - x_+ ) f\left(x+\frac{x_+}{2},x_-\right ) g\left(x_-,x-\frac{x_+}{2}\right)\nonumber\\ & = & \int d^{\rm dim}x_-f(x+\xi/2,x_- ) g(x_-,x-\xi/2)\nonumber\\ & = & f*_cg\end{aligned}\ ] ] with @xmath271 and @xmath272 . in the wigner space , the position operator @xmath273 and the momentum operator @xmath87 becomes @xmath274 and @xmath275 because @xmath276= \int d^{\rm dim}\xi e^{-\frac{{\rm i}}{\hbar}p_\nu\xi^\nu } \left(x^\mu+\xi^\mu/2\right)g(x+\xi/2,x-\xi/2 ) = \left(x^\mu+\frac{{\rm i}\hbar}{2}\partial_{p_\mu}\right ) g(x , p ) = x^\mu\star_{\cal m}g(x , p),\ ] ] @xmath277= \int d^{\rm dim}\xi e^{-\frac{{\rm i}}{\hbar}p_\nu\xi^\nu } g(x+\xi/2,x-\xi/2)\left(x^\mu-\xi^\mu/2\right ) = g(x , p)\star_{\cal m}x^\mu,\ ] ] @xmath278= \int d^{\rm dim}\xi e^{-\frac{{\rm i}}{\hbar}p_\nu\xi^\nu } \frac{\hbar}{{\rm i}}\partial_{x_1^\mu}g(x_1,x_2 ) = p_\mu\star g(x , p),\end{aligned}\ ] ] and @xmath279= \int d^{\rm dim}\xi e^{-\frac{i}{\hbar}p_\nu\xi^\nu } \frac{\hbar}{{\rm i}}\partial_{x_2^\mu}g(x_1,x_2 ) = g(x , p)\star p_\mu.\end{aligned}\ ] ] the commutation relationship of operators is @xmath280_{\star_{\cal m}}:=x^\mu\star_{\cal m}p_\nu -p_\nu\star_{\cal m}x^\mu= { \rm i}\hbar\delta^\mu_\nu$ ] , which corresponds to the canonical commutation relationship of operators : @xmath281={\rm i}\hbar\delta^\mu_\nu$ ] .
to add the spin arguments in @xmath282 , we will employ the following bilinear map : @xmath283 with @xmath284 and @xmath285 .
note that the spin operator @xmath286 is characterized by the commutation relation @xmath287={\rm i}\epsilon^{abc}\hat{s}^c$ ] @xmath288 with the levi - civita tensor @xmath289 , and the star product ( [ eq : m+s ] ) reproduces the relation , i.e. , the operator @xmath290 satisfies @xmath291_\star={\rm i}\epsilon^{abc}s^c$ ] . to obtain the map @xmath134 , we introduce the variables transformation @xmath292 where @xmath293 with @xmath294 , the electric charge @xmath295 , a u(1 ) gauge field @xmath0 , and a su(2 ) gauge field @xmath22 .
their fields are treated as real numbers , and the integral over @xmath296 can be replaced by an integral over @xmath297 .
this transformation induces the following transformations of differential operators : @xmath298 @xmath299 where @xmath300 .
we expand @xmath301 in terms of @xmath302 as @xmath303 we define the bilinear map @xmath304 corresponding to the commutation relation in terms of the phase space @xmath305 , and expand it in terms of @xmath302 as @xmath306 from eqs .
( [ eq : transition1 ] ) and ( [ eq : transition2 ] ) , @xmath307 is given as follows : @xmath308 with @xmath309 .
note that @xmath310 is the poisson bracket .
a constitution method of higher order terms @xmath311 with @xmath312 is called a deformation quantization , which is given by kontsevich @xcite , as will be described in the next subsection . in this subsection , we explain the kontsevich s deformation quantization method @xcite .
we define a star product as @xmath313 with @xmath314 @xcite . here
@xmath315 is called the two - cochain ( @xmath316 represents the function space ) .
we require that the star product satisfies the association property @xmath317 , which limits forms of @xmath318 and @xmath319 .
we note that the association property is necessary for the existence of the inverse with respect to the star product .
for example , the inverse of the lagrangian is a green function , which always exists as @xmath320 with a wave function @xmath7 .
now , we define a @xmath259-cochain space @xmath321 with @xmath322 , where @xmath316 represents a function space such as the wigner space @xmath282 ; we define a multi - vector space @xmath323 , where @xmath324 represents a manifold such as a classical phase space ( dimension @xmath325 ) , @xmath326 denotes a tangent vector bundle with a tangent vector space @xmath327 at @xmath328 ( @xmath34 is a coordinate at @xmath259 ; @xmath329 represents a certain coefficient ) , @xmath330 denotes a @xmath142-th completely antisymmetric tensor product , ( for example , @xmath331 ) , and @xmath332 represents the section ; for example , @xmath333 is defined as a set of tangent vector at each position @xmath328 .
the poisson bracket @xmath334 is element of @xmath335 , where @xmath336 is called the poisson structure @xmath337 .
the deformation quantization is the constitution method of higher order cochains @xmath338 from the poisson bracket @xmath339 . in other words ,
the deformation quantization is the following map @xmath340 : @xmath341 where @xmath342 satisfies the jacobi identity and @xmath343 satisfies the association property , as shown in fig .
[ fig : dq](a ) . in the following sections
, we will generalize the two - cochain @xmath344 and the second order differential operator @xmath345 to the so - called @xmath346 algebra ( the definition is given in section [ sub : l_infty ] ) . in the section [ sec :
ce ] , we will introduce the two - cochain @xmath344 and second order differential operator @xmath345 , and the @xmath259-cochain @xmath347 and @xmath142-th order differential operator @xmath348 .
we will show that these operators satisfy certain conditions , and @xmath347 and @xmath348 are embedded in a differential graded lie algebra ( d.g.l.a ) ( the definition is shown in section [ sec : ce ] ) .
moreover , in section [ sub : l_infty ] , the d.g.l.a will be embedded in the @xmath346 algebra ( see fig . [
fig : dq](b ) ) . in the @xmath346 algebra ,
the jacobi identity and the association property are compiled in the following equation @xmath349 where @xmath350 , and @xmath254 is called the codifferential operator , which will be introduced in section [ sub : l_infty ] .
namely , in the @xmath346 algebra , the deformation quantization is a map from @xmath339 to @xmath351 holding the solution of eq .
( [ eq : qe ] ) ( figure [ fig : dq](c ) ) .
such a map is uniquely determined in the @xmath346 algebra . in this paper
, we will identify the tensor product @xmath121 with the direct product @xmath68 , i.e. , @xmath352 : @xmath353 with @xmath354 and @xmath355 ( @xmath175 denotes that @xmath176 and @xmath177 are identified ; @xmath356 represents the ordered pair , i.e. , it is a set of @xmath113 and @xmath114 , and @xmath357 ) . with the jacobi identity to @xmath344 with the association property .
( b ) : enlargement of algebras .
the two vector space @xmath345 and two cochain space @xmath344 generalize to multi - vector space @xmath358 and cochain space @xmath359 , respectively .
these spaces are compiled in the d.g.l.a ; finally , @xmath346 algebra is introduced by using the d.g.l.a .
( c ) : the deformation quantization is redefined as the map on the @xmath346 algebra.,width=453 ] from eq .
( [ eq : dq0 ] ) , the association property is given by the following equation : @xmath360 with @xmath361 .
( the symbol `` @xmath362 '' represents the usual commutative product , and @xmath363 . ) because @xmath364 is the poisson bracket , which is bi - linear differential operator , we define @xmath365 as a differential operator on a manifold @xmath324 ; moreover , we also assume that @xmath259-cochains are differential operators and products of functions . here ,
@xmath366 and @xmath367 represent a space of smooth functions on a manifold @xmath324 and a space of multilinear differential maps from @xmath368 to @xmath366 , respectively .
degree of @xmath369 is defined by @xmath370 now , we introduce a coboundary operator @xmath371 @xcite ; @xmath372 with @xmath369 ; note that @xmath373 , and thus , @xmath374 is the boundary operator .
the gerstenhaber bracket is defined as @xmath375_c:{\cal c}^k({\cal a};{\cal a})\otimes { \cal c}^{k^\prime } ( { \cal a};{\cal a})\rightarrow { \cal c}^{k+k^\prime-1}({\cal a};{\cal a})$ ] @xcite : @xmath376_c ( \underbrace{f_0,f_1,\cdots , f_{k+k^\prime-2 } } _ { k+k^\prime-1 } ) \nonumber\\ & & : = \sum^{k-1}_{r=0}(-1)^{r(k^\prime-1 ) } \beta^k(\underbrace{f_0,\cdots , f_{r-1 } , \beta^{k^\prime}(f_r,\cdots , f_{r+k^\prime-1}),f_{r+k^\prime},\cdots , f_{k+k^\prime-2}}_{k})\nonumber\\ & & -\sum_{r=0}^{k^\prime-1}(-1)^{(k-1)(r+k^\prime-1 ) } \beta^{k^\prime}(\underbrace{f_0,\cdots , f_{r-1 } , \beta^k(f_r,\cdots , f_{r+k-1}),f_{r+k},\cdots , f_{k+k^\prime-2}}_{k^\prime}),\nonumber\\\end{aligned}\ ] ] where @xmath377 and @xmath378 .
note that @xmath373 , and thus , @xmath374 is the boundary operator . by using the coboundary operator and the gerstenhaber bracket , eq .
( [ eq : ap ] ) is rewritten as @xmath379_c\label{eq : ce}\ ] ] with @xmath380 ( @xmath381 ) .
for example , eq . ( [ eq : ap ] ) for @xmath382 is given as : @xmath383 where we have used @xmath384 and @xmath385 .
the coboundary operator for @xmath386 is given by : @xmath387 moreover , the gerstenhaber bracket in terms of @xmath388 is given by @xmath389_c(f , g , h)= \beta_i(\beta_j(f , g),h)-\beta_i(f,\beta_j(g , h))+ \beta_j(\beta_i(f , g),h)-\beta_j(f,\beta_i(g , h)).\label{eq : gb}\ ] ] using the above eqs .
( [ eq : ce0]-[eq : gb ] ) , we can check the equivalence between eq .
( [ eq : ap ] ) and eq .
( [ eq : ce ] ) . equation ( [ eq : ce ] ) is called the cohomology equation , and the star product is constructed by using solutions of the cohomology equation . if we add eq .
( [ eq : ce ] ) with respect to @xmath390 , we obtain the following equation : @xmath391_c=0\label{eq : mcc}\ ] ] with @xmath392 ; @xmath393 , @xmath394 . here , we identify the vector fields @xmath395 with anti - commuting numbers @xmath396 ( @xmath397 ) , @xmath398 ; thus the poisson bracket @xmath399 is rewritten by @xmath400 .
now , we define the batalin - vilkovisky ( bv ) bracket : @xmath401_{\rm bv}:=-\sum_{i=1}^d\left ( \frac{\alpha_1\overleftarrow{\partial}}{\partial x^i } \frac{\overrightarrow{\partial}\alpha_2}{\partial \tilde{\eta}_i}- \frac{\alpha_1\overleftarrow{\partial}}{\partial \tilde{\eta}_i } \frac{\overrightarrow{\partial}\alpha_2}{\partial x^i } \right)\ ] ] with @xmath402 . by using the bv bracket ,
the jacobi identity is rewritten as @xmath403_{\rm bv}=0,\label{eq : bve}\ ] ] with @xmath402 ; for @xmath404 , @xmath405 and @xmath406 . by using the bv bracket , the jacobi identity is rewritten as @xmath403_{\rm bv}=0,\label{eq : bve}\ ] ] where @xmath407 , i.e. , @xmath408 .
now , we generalize the differential @xmath409 and bv - bracket @xmath375_{\rm bv}$ ] for @xmath410 and @xmath411 as follows ( @xmath412 ) : @xmath413_{\rm bv}:{\cal t}^k\bigotimes { \cal t}^{k^\prime}\rightarrow { \cal t}^{k+k^\prime-1}\nonumber\\ & & [ \alpha^k,\alpha^{k^\prime}]_{\rm bv}:=- \sum_{i=1}^d\left(\frac{\alpha^k\overleftarrow{\partial}}{\partial\tilde{\eta_i } } \frac{\overrightarrow{\partial}\alpha^{k^\prime}}{\partial x^i}- \frac{\alpha^k\overleftarrow{\partial}}{\partial x^i } \frac{\overrightarrow{\partial}\alpha^{k^\prime } } { \partial \tilde{\eta}_i}\right)\end{aligned}\ ] ] with @xmath414 , and @xmath415 ; degree of @xmath416 is defined by @xmath417 the cochain algebra is defined by the set of the differential operator @xmath374 , the gerstenhaber bracket @xmath375_c$ ] and @xmath418 , i.e. , @xmath419_c,{\cal c})$ ] ; in addition , the multi - vector algebra is defined by the set of the differential operator @xmath420 , bv bracket @xmath375_{\rm bv}$ ] and @xmath421 , i.e. , @xmath422_{\rm bv},{\cal t})$ ] .
the cochain algebra and the multi - vector algebra satisfy the following common relations : @xmath423 @xmath424= [ \partial\gamma_1,\gamma_2]+ ( -1)^{{\rm deg}(\gamma_1 ) } [ \gamma_1,\partial\gamma_2],\label{eq : dgla2}\ ] ] @xmath425= -(-1)^{{\rm deg}(\gamma_1){\rm deg}(\gamma_2 ) } [ \gamma_2,\gamma_1]\label{eq : dgla3}\ ] ] @xmath426 + ( -1)^{{\rm deg}(\gamma_3)({\rm deg}(\gamma_1)+ { \rm deg}(\gamma_2))}[\gamma_3,[\gamma_1,\gamma_2]]+ ( -1)^{{\rm deg}(\gamma_1)({\rm deg}(\gamma_2)+ { \rm deg}(\gamma_3))}[\gamma_2,[\gamma_3,\gamma_1]]=0\label{eq : dgla4}\ ] ] with @xmath427 , @xmath428 , and @xmath375\equiv[~,~]_{(c\ { \rm or}\ bv)}$ ] .
therefore , the two algebra can be compiled in the so - called the differential graded lie algebra ( d.g.l.a ) @xmath429,{\cal g})$ ] , where @xmath430 is a graded @xmath142-vector space with @xmath431 has a degree @xmath432 ( @xmath433 ; @xmath434 is the set of integers ) , and d.g.l.a . has the linear operator @xmath204 and the bi - linear operator @xmath375 $ ] : @xmath435:~{\cal g}^k\bigotimes{\cal g}^l \rightarrow{\cal g}^m,~ x_k\in{\cal g}^k,~x_l\in{\cal g}^l,~x_m\in{\cal g}^m , \nonumber\\ & & \ \ \ \ \ \ { \rm deg}([x_k , x_l])={\rm deg}(x_k)+{\rm deg}(x_l)={\rm deg}(x_m),\end{aligned}\ ] ] where @xmath204 and @xmath375 $ ] satisfy eqs .
( [ eq : dgla1 ] ) , ( [ eq : dgla2 ] ) , ( [ eq : dgla3 ] ) and ( [ eq : dgla4 ] ) . in d.g.l.a . ,
( [ eq : mcc ] ) and ( [ eq : bve ] ) are compiled in the so - called maurer - cartan equation @xcite : @xmath436=0\label{eq : mc}\ ] ] with @xmath437 .
therefore , the deformation quantization @xmath340 is a map : @xmath438=0,~i=1,2.\end{aligned}\ ] ] namely , the deformation quantization is a map holding a solution of the maurer - cartan equation ( [ eq : mc ] ) . in the section [ sub : l_infty ]
, we will introduce a @xmath346 algebra , and will redefine the deformation quantization ; in the @xmath346 algebra , the maurer - cartan equation ( [ eq : mc ] ) is rewritten as @xmath439 ( @xmath254 and @xmath440 will be defined in [ sub : l_infty ] ) .
now we define a commutative graded coalgebra @xmath441 .
first , we define a set @xmath442 , where @xmath443 with a graded @xmath142-vector space @xmath444 ( @xmath445 ) , @xmath125 and @xmath254 represent the coproduct and codifferential operator , respectively ; moreover , @xmath446 denotes cocommutation ( definition is given later ) .
the coproduct , cocommutation and codifferential operator satisfy the following equations : @xmath447 with @xmath448 , where @xmath449 and @xmath450 .
@xmath254 represents a codifferential operator adding one degree : @xmath451 with @xmath452 for @xmath453 ( the explicit form of @xmath254 is given later ; @xmath454 ) . by using @xmath446
, we define the commutative graded coalgebra @xmath441 from @xmath442 ; the identify relation @xmath174 is defined as @xmath455 , i.e. , @xmath456 and @xmath457 are identified .
now , we define the commutative graded tensor algebra : @xmath458|x\in{\cal v}\},\ ] ] where @xmath459=\left\{y~|~y\in{\cal v},~x\sim y\right\}$ ] , and @xmath460 with @xmath461 ; a product in @xmath441 is defined by @xmath462 $ ] .
namely , in @xmath441 , @xmath463 with @xmath464 .
( let us recall that the derivation of the exterior algebra from the tensor space ; @xmath465 and @xmath441 correspond to the tensor space and the exterior algebra , respectively . ) moreover , in the case that @xmath466 , the commutative graded coalgebra @xmath441 is called the @xmath346 algebra . for the @xmath346 algebra ,
the coproduct and codifferential operator are uniquely determined by using multilinear operators : @xmath467 as follows : @xmath468 where @xmath469 represents a sign with a replacement @xmath470 . from the condition @xmath466
, we can identify @xmath471 with @xmath472)$ ] in d.g.l.a .
if we put @xmath473 , @xmath474 for @xmath475 is equal to the maurer - cartan equation eq .
( [ eq : mc ] ) in d.g.l.a @xcite , where @xmath476 with @xmath477 .
therefore , the deformation quantization is a map : @xmath478 with @xmath479 to constitute such a map @xmath340 , we introduce the @xmath480 map @xmath481 , which is defined as the following map holding degrees of coalgebra : @xmath482 moreover , the @xmath346 map satisfies the following equations : @xmath483 @xmath484 a form of such a map is limited as @xcite : @xmath485 where @xmath486 is a map from @xmath441 to @xmath487 holding degrees ; @xmath488 here we define @xmath489 , which satisfies @xmath490 .
the map @xmath481 holds solutions of maurer - cartan equations @xmath491 and @xmath492 ; from @xmath493 and the definition of the @xmath346 map : @xmath494 , we obtain the following equation : @xmath495 which means that the @xmath346 map transfers a solution of the maurer - cartan equation from another solution .
now , we return to the deformation quantization .
the multi - vector space @xmath358 , is embedded in @xmath496 ; @xmath497 , where @xmath498 for @xmath499 , @xmath500 , @xmath501_{\rm bv}$ ] , and @xmath502 for @xmath503 ; @xmath446 replaces the wedge product `` @xmath504 '' with the product `` @xmath362 '' . for the cochain space @xmath359 , it is also embedded in @xmath496 ; @xmath505 , where @xmath506 for @xmath507 , @xmath508 , @xmath509_c$ ] , and @xmath510 for @xmath503 .
the star product is given by @xmath511 , which is identified as the map @xmath512 with @xmath513 .
here we summarize the main results of the succeeding sections without explaining their derivations .
the map @xmath514 is given by a path integral of a topological field theory having super fields : @xmath515 ; and scalar fields : @xmath516 , @xmath517 , and @xmath518 ; and one - form fields : @xmath519 , @xmath520 , @xmath521 , and @xmath522 ; and grassmann fields @xmath523 ; on a disk @xmath524 @xcite .
these fields are defined in section [ sec : topologicalstringtheory ] . using these fields , the map @xmath525 is given as follows : @xmath526 for any function @xmath527 , which depend on @xmath34
; in this paper , @xmath34 represents the coordinate in the classical phase space .
here @xmath528 , and @xmath6 is defined by @xmath529 , which is common and independent of @xmath530 ( @xmath531 ) . the operator @xmath532 is defined as @xmath533 \delta_x(\psi(\infty))\\ & \equiv & \int_{1=t_1>t_2>\cdots > t_m=0 }
f_1(\psi(t_1))\prod_{k=2}^{m-1 } \partial_{i_k}\left [ f(\psi(t_{k}))a^{+i_k}(t_k ) \right ] f_m(\psi(0))\delta_x(\psi(\infty))\label{eq : patho}\end{aligned}\ ] ] for @xmath6 , @xmath534 , and @xmath535 , where @xmath536\ ] ] with a hodge operator @xmath537 , ( @xmath538 ) ; we will introduce the explicit definition in section [ sec : perturbationtst ] .
moreover , for @xmath539 ( @xmath540 is an integer number ; @xmath541 ) , @xmath542 where the subscript @xmath543 means that the fields ( @xmath544 ) go to ( @xmath545 ) .
these results lead to the diagram technique in section [ sec : diagramrules ] and the explicit expression of the star product in section [ sec : twistedelement ] . in this section ,
we expound the fields : @xmath132 , @xmath7 , @xmath39 , @xmath546 , @xmath547 , @xmath548 , @xmath186 , @xmath549 , and @xmath258 .
the simplest topological string theory is defined the following action : @xmath550 with local coordinates @xmath551 on a disk @xmath552 ( we consider that the disk is the upper - half plain in the complex one , i.e. , @xmath553 ) , where @xmath554 and @xmath555 are u(1 ) gauge fields and scalar fields , respectively ; @xmath556 is a gauge strength ( @xmath557 and @xmath558 ) .
the other fields @xmath39 , @xmath546 , @xmath547 , @xmath548 , @xmath186 , and @xmath559 are introduced in section [ sec : gf ] ; we discuss the gauge fixing method using the so - called bv - brst formalism @xcite ( where the bv refers to batalin and vilkovisky ; best refers to becchi , rouet , stora and tyutin ) . in section
[ sec : cgc ] , we discuss the gauge invariance of the path integral , and introduce the sd operator . in section [ sec : pathintegraltst ] , we see that correspondence of the deformation quantization and topological string theory . here , we quantize the action ( [ eq : ts ] ) using the path integral . roughly speaking ,
the path integral is the gauss integral around a solution of an equation of motion . in many cases , a general action
@xmath237 has no inverse .
therefore , we will add some extra fields , and obtain the action @xmath560 having inverse , which is called as the quantized action .
now , we discuss a general field theory .
we assume that a general action @xmath237 is a function of fields @xmath561 , i.e. , @xmath562 $ ] ; each field @xmath561 is labeled by a certain integer number , which is called as a ghost number @xmath563 ( it is defined below ) .
@xmath564 denotes that the fields fixed on the solution of the classical kinetic equation : @xmath565 , and the subscript of the fields represents a number of fields .
because the gauss integral is an inverse of a hessian , a rank of the hessian should be equal to the number of the fields . here
a hessian is defined by : @xmath566:= \frac{\overrightarrow{\delta}}{\delta\phi^i}s \frac{\overleftarrow{\delta}}{\delta\phi^j},\ ] ] where @xmath567 , and @xmath568 ; for a boson @xmath561 , @xmath569 is a ghost number @xmath563 ; for a fermion @xmath561 , @xmath569 is @xmath570 .
we define the rank of the hessian @xmath571 and the number of the fields @xmath561 by @xmath572 and @xmath573 , respectively . generally speaking , @xmath574 , because an action has some symmetries @xmath575 with nontrivial symmetry generators @xmath576 , where is satisfies the following equation : @xmath577 with @xmath578 .
the nontrivial symmetry generator decrease the rank of hessian from the number of fields . to define the path integral , we should add @xmath579 virtual fields @xcite .
the additional fields are called as ghost fields @xmath580 and antifields @xmath581 ( @xmath582 ) , and these fields are labeled by ghost numbers . for @xmath580 , the ghost number is defined by @xmath583 .
the fields and ghost fields have antifields @xmath584 .
the antifields corresponding to @xmath585 and @xmath580 are described as @xmath586 and @xmath587 , respectively .
the ghost number of @xmath588 is defined by @xmath589 .
statistics of the anti - fields is opposite of fields , i.e. , if the fields are fermions(bosons ) , the anti - fields are bosons(fermions ) .
( here we only consider the so - called irreducible theory . for a general theory ,
see references @xcite . ) using these fields , we will transform the action @xmath590\mapsto
s_{\rm gh}[\psi]$ ] , where @xmath591 with @xmath582 and @xmath592 ( @xmath434 represents the set of integers ) , @xmath593 are fields , @xmath580 represents ghost fields , and @xmath588 denotes anti - fields of the fields @xmath594 .
hereafter we write a function space created by the fields and anti - fields as @xmath595 .
it is known that @xmath560 is given by @xmath596 note that the anti - fields will be fixed , and @xmath597 ( see section [ sec : pathintegraltst ] ) . for the topological string theory , the fields @xmath598 are u(1 ) gauge fields @xmath599 and scalar fields @xmath600 with @xmath558 and @xmath557 ; namely , @xmath601 . since @xmath602 and @xmath603 , the fields number @xmath604 is @xmath605 .
the action ( [ eq : ts ] ) has the u(1 ) gauge invariance : @xmath606 with @xmath607 represents a scalar function ( @xmath558 ) .
therefore , the topological string theory has @xmath608 linear - independent nontrivial symmetry generators . here
we replace the scalar fields @xmath609 with ghost fields @xmath610 ( brst transformation ) .
moreover , we add antifields @xmath611 ; since the gauge transformation does not connect to @xmath7 and the other fields , we does not add @xmath612 ( the space of fields and ghost fields has @xmath608 symmetry generators , and the space of anti - fields and the anti - ghosts also have @xmath608 symmetry generators corresponding to u(1 ) gauge symmetry ; see figure [ fig : hessian ] ) : @xmath613 in this case , @xmath614 ; on the other hand , the action is a function of @xmath605 fields @xmath615 , @xmath616 ghost fields @xmath610 and @xmath608 anti - fields @xmath611 .
therefore , a rank of the hessian corresponding to @xmath617 is calculated by @xmath618 : = \frac{\delta}{\delta\psi^\alpha(\sigma_2)}s \frac{\delta}{\delta\psi^\beta(\sigma_1)}$ ] the first column and first raw represent the right - hand side and the left - hand side of variation functions , respectively .
@xmath619 represents a non - trivia noether current ( @xmath620 ) .
the hessian is block diagonal matrix ; the ranks of the upper left and the lower right parts are @xmath621 .
therefore , the total rank of the hessian is @xmath622 ( @xmath530 is fixed ) .
, width=453 ] since @xmath623 ( antifields will be fixed ) , the field number of the path - integral of @xmath617 is equal to the rank of the hessian of @xmath617 ; hence , the path - integral of the action @xmath617 become well - defined . finally , the gauge invariance action is written by @xmath624 with @xmath625 and @xmath626 , where we define @xmath627 using a hodge operator @xmath628 : @xmath629 ( the definition of the hodge operator is depend on the geometry of the disk @xmath552 ; we will introduce the explicit definition in section [ sec : perturbationtst ] ) , which is also called as the anti - field . in this section
, we will add interaction terms : @xmath630 , where @xmath114 represents an expansion parameter , and we will see that @xmath560 is uniquely fixed except a certain two form @xmath342 by a gauge invariance condition .
note that @xmath342 satisfies the jacobi s identity .
therefore , we can identify @xmath342 with the poisson bracket .
first we discuss the gauge invariant condition .
if we identify the fields and anti - fields with coordinates @xmath18 and canonical momentum @xmath259 , i.e. , @xmath631 , and we also identify the action @xmath237 and the hamiltonian h : @xmath632 . in the analytical mechanics , @xmath633 represents a transform along the surface @xmath634 , i.e. , @xmath635 holds the hamiltonian .
similarly , we can define a gauge transformation , which holds the action @xmath237 , using the poisson bracket in the two - dimension field theory .
it is known as the batalin - vilkovisky ( bv ) bracket @xcite ; the definitions of the bracket are @xmath636 with @xmath637 .
the bv bracket has the ghost number @xmath221 , then a bv - brst operator @xmath638 adds one ghost number .
the bv bracket satisfies the following equations : @xmath639 with @xmath640 . using the bv - brst operator ,
the gauge invariance of action @xmath237 is written as @xmath641 , i.e. , @xmath642 which is called the classical master equation .
we use this equation and eqs .
( [ eq : bv1 ] ) and ( [ eq : bv2 ] ) ; we obtain @xmath643 , which corresponds to the condition of the brst operator : @xmath644 ( @xmath645 is the brst operator ) .
therefore , the dv - brst operator is the generalized brst one .
next , we discuss generalization of the topological field theory .
let us write a generalized action @xmath560 as @xmath646 where @xmath114 is an expansion parameter .
using gauge invariance condition ( [ eq : gaugeinvariance ] ) , @xmath647 ( @xmath648 ) is given by a solution of the following equation : @xmath649 the general solution is given by @xcite @xmath650,\\ ( s_{n>1})_{\rm gh}&=&0\end{aligned}\ ] ] with @xmath651 , @xmath652 and @xmath653 ( @xmath654 ) , where @xmath655 is a function of @xmath7 , and satisfies the following equation : @xmath656 here , if we identify @xmath600 with @xmath657 , this equation is the jacobi identity of poisson bracket .
therefore , we can identify the poisson bracket with the topological string theory .
now we discuss the path integral of the topological string theory @xmath658 with @xmath659 , and an observable quantity operator @xmath660 .
note that this path - integral does not include integrals in terms of the anti - fields .
therefore , we must fix the anti - fields ; then , we consider that the anti - field @xmath661 is a function of the field @xmath662 , i.e. , @xmath663 and @xmath664 namely , the path integral is defined by @xmath665 a choice of @xmath666 is corresponding to the gauge fixing in the gauge theory .
the path integral must be independent to the gauge choice ( gauge invariance ) . to obtain a gauge invariant condition
, we take the variation of the path integral in terms of anti - fields , and obtain the following gauge invariant condition @xcite : @xmath667 where we have introduced the schwinger - dyson ( sd ) operator : @xmath668 where @xmath669 is defined as follows : if @xmath594 represents a boson , @xmath670 ; if @xmath594 represents a fermion , @xmath671 .
equation ( [ eq : qme ] ) is called the quantum master equation .
it is known that the following two conditions are equivalence : @xmath672 where @xmath673 is called the gauge - fixing fermion ( an example will be shown later ) . to perform the path integral , we generalize the classical action @xmath560 to a quantum action @xmath674 .
the correction terms @xmath675 ( @xmath676 ) are calculated from the master equation : @xmath677 or @xmath678 in the case where @xmath679 , we can put @xmath680 .
fortunately , the topological string theory satisfies @xmath679 .
therefore , we do not have to be concerned about the quantum correction of the action .
finally , we consider the gauge fixing . here
we employ the lorentz gauge : @xmath681 and we add the integral of the lorentz gauge to @xmath560 . however , the path integral should hold gauge invariance , i.e. , the path integral should be independent of gauge fixing term . then
, the gauge fixing can be written gauge - fixed fermion : @xmath682 where we introduced @xmath616 fields @xmath683 ( @xmath684 ) , and anti - fields @xmath685 are given by @xmath686 now , we employ the lagrange multiplier method , and introduce @xmath616 scalar fields @xmath687 .
the gauge - fixed action is written by @xmath688 the other anti - fields are also fixed by this gauge - fixing fermion : @xmath689 gauge fixed action @xmath690 is written by @xmath691.\end{aligned}\ ] ] here we perform the following variable transformations : @xmath692 where @xmath693 ; @xmath694 .
for any scalar field @xmath695 ( @xmath696 ) , @xmath697 is called as the super field , where @xmath698 and @xmath699 represent a one - form field and a two - form field , respectively . by using the super fields , the gauge fixed action @xmath690 can be rewritten as @xmath700,\ ] ] where @xmath701 .
this is the final result in this section .
hereafter , we write @xmath702 $ ] and @xmath703 .
we return to the discussion about the deformation quantization . here
we see that the equivalence of the deformation quantization and the topological string theory , and introduce the perturbation theory of the topological string theory , which is equal to kontsevich s deformation quantization @xcite . here
we summarize correspondence between path integral with @xmath346 map .
first we note that the map : @xmath704 is isomorphic , because @xmath705 .
sd operator satisfies the conditions of codifferential operator @xmath254 in @xmath480 algebra , where the vector space and the degree of the space correspond to @xmath706 and the ghost number , respectively .
the path integral @xmath707 gives the deformation quantization @xmath512 .
the master equation @xmath708 with @xmath709 is corresponding to the @xmath346 map s condition @xmath710 for @xmath711 with a positive integer @xmath6 , @xmath712 is given by @xmath713 where @xmath714 is the expansion of @xmath715 , and is defined as @xmath716 and @xmath660 is chosen to satisfy @xmath717 where @xmath718 .
we put @xmath660 as follow : @xmath719_{(m-2)}\delta_x(x(\infty)),\\ & \equiv & \int_{1=t_1>t_2>\cdots > t_m=0 } f_1(\psi(1))\prod_{k=2}^{m-1 } \partial_{i_k}\left [ f(\psi(t_{k}))a^{+i_k}(t_k ) \right ] f_m(\psi(0))\delta_x(\psi(\infty))\end{aligned}\ ] ] where the subscript @xmath720 denotes that @xmath721 forms are picked up from the products of super fields , and @xmath722 represents the surface of the disk @xmath552 , i.e. , @xmath723 is the parameter specifying the position on the boundary @xmath724 ( @xmath725 . to be exact ,
the action and fields include gauge fixing terms , ghost fields and anti - field .
finally , the deformation quantization is given as follow : @xmath726 now we see that the perturbation theory of the topological string theory . first , we write the action as @xmath727 .
the first term is defined as @xmath728 , \label{eq : s0}\ ] ] where @xmath729 , and we have expanded @xmath600 around @xmath657 .
the path integral of an observable quantity @xmath730 is given by @xmath731 where @xmath732 .
this expansion corresponds to the summation of all diagrams by the contractions of all pairs in terms of fields and ghost fields . from equation ( [ eq : s0 ] ) ,
propagators are inverses of @xmath733 here we assume that the disk is the upper complex plane : @xmath734 with @xmath735 , and the boundary is @xmath736 .
( @xmath737 represents the real number space , and @xmath738 denotes a complex number . ) the hodge operator @xmath739 is defined by @xmath740 where @xmath741 represents the complex conjugate of @xmath738 .
moreover , @xmath742 where @xmath743 with the complex number plane @xmath744 , and @xmath745 .
now , we calculate green functions of @xmath746 and @xmath747 , because the green functions are inverses of these operators : @xmath748 where @xmath749 .
the solution depends on the boundary condition . in the case
that @xmath738 and @xmath750 satisfy the neumann boundary condition , a solution is a function of @xmath751 on the other hand , @xmath738 and @xmath750 satisfy the dirichlet boundary condition , a solution is a function of @xmath752 the neumann boundary condition is @xmath753 , and the dirichlet boundary condition is @xmath754 .
the propagators are given by @xmath755 and so on . from these propagators
, we can obtain diagram rules corresponding to the deformation quantization . in section [ sec : diagramrules ]
, we will introduce exact diagram rules . to obtain the star product , we choice @xmath756 from the perturbation theory of the topological string theory ,
we can obtain the following diagram rules of the star product , which is first given by kontsevich @xcite : @xmath757 where @xmath758 , @xmath759 and @xmath760 are defined as follows : @xmath761 is a set of the graphs @xmath758 which have @xmath762 vertices and @xmath763 edges .
vertices are labeled by symbols `` @xmath221 '' , `` @xmath764 '' , @xmath765 , `` @xmath766 '' , `` @xmath767 '' , and `` @xmath768 '' .
edges are labeled by symbol @xmath769 , where @xmath770 , @xmath771 , and @xmath772 .
@xmath769 represents the edge which starts at `` @xmath142 '' and ends at `` @xmath773 '' .
there are two edges starting from each vertex with @xmath770 ; @xmath767 and @xmath768 are the exception , i.e. , they act only as the end points of the edges .
hereafter , @xmath774 and @xmath775 represent the set of the vertices and the edges , respectively .
@xmath759 is the operator defined by : @xmath776\times \nonumber\\ & & \left[\left ( \prod _ { e\in e_{\gamma } , e=(*,l)}\partial _ { i(e ) } \right)f\right]\times \left[\left ( \prod _ { e\in e_{\gamma } , e=(*,r)}\partial _ { i(e ) } \right)g\right],\end{aligned}\ ] ] where , @xmath31 is a map from the list of edges @xmath777 to integer numbers @xmath778 . here
@xmath779 ; @xmath325 represents a dimension of the manifold @xmath780 .
@xmath759 corresponds to the graph @xmath758 in the following way : the vertices `` @xmath221 '' , `` @xmath764 '' , @xmath765 , `` @xmath766 '' , correspond to the poisson structure @xmath781 . @xmath768 and @xmath767 correspond to the functions @xmath113 and @xmath114 , respectively .
the edge @xmath782 represents the differential operator @xmath783 acting on the vertex @xmath773 . the simplest diagram for @xmath784 is shown in fig .
@xmath785 , which corresponds to the poisson bracket : @xmath786 .
the higher order terms are the generalizations of this poisson bracket .
figure @xmath787 shows a graph @xmath788 with @xmath789 corresponding to the list of edges @xmath790 corresponding to poisson bracket .
( b ) : a graph @xmath791 correspond to the list of edges : @xmath792 .
, width=226 ] in addition , the operator @xmath793 is given by @xmath794 we put the coordinates for the vertices in the upper - half complex plane @xmath795 @xmath796 represents the complex plain ; @xmath797 denotes the imaginary part of @xmath798 .
therefore , @xmath768 and @xmath767 are put at @xmath799 and @xmath221 , respectively .
we associate a weight @xmath760 with each graph @xmath800 as @xmath801 where @xmath802 is defined by @xmath803 @xmath259 and @xmath18 are the coordinates of the vertexes `` @xmath142 '' and `` @xmath773 '' , respectively .
@xmath804 represents the complex conjugate of @xmath805 .
@xmath806 denotes the space of configurations of @xmath766 numbered pair - wise distinct points on @xmath807 : @xmath808 here we assume that @xmath807 has the metric : @xmath809 with @xmath810 ; @xmath811 is the angle which is defined by @xmath812 and @xmath813 , i.e. , @xmath814 with the metric @xmath815 . for example , @xmath816 corresponding to fig .
@xmath785 is calculated as : @xmath817 where we have included the factor `` @xmath764 '' arising from the interchange between two edges in @xmath332 .
@xmath760 corresponding to the fig .
@xmath787 is @xmath818 where @xmath819 and @xmath820 are the coordinates of vertexes `` @xmath221 '' and `` @xmath764 '' , respectively . here
, we have used the following facts : @xmath821 generally speaking , the integrals are entangled for @xmath822 graphs , and the weight of these are not so easy to evaluate as eq .
@xmath823 .
note that the above diagram rules also define the twisted element as the following relation : @xmath824 . from eq .
( [ eq : su(2)poisson ] ) , the poisson structure corresponding to our model is @xmath825 where the symbols @xmath530 and @xmath826 represent indexes of the phase space @xmath827 .
we separate the poisson structure as follows : @xmath828 here , for @xmath829 , @xmath284 ( @xmath830 ) , where @xmath831 are functions @xmath258 and @xmath259 .
because @xmath832 is constant and @xmath833 and @xmath834 are functions of @xmath835 and @xmath836 , and any function @xmath113 is written as @xmath837 ( @xmath838 only depends on @xmath258 and @xmath259 ) , then we obtain additional diagram rules : two edges starting from @xmath839 connect with both vertices `` @xmath767 '' and `` @xmath768 '' .
at least one edge from vertices @xmath832 or @xmath834 connect with vertices `` @xmath767 '' or `` @xmath768 '' .
a number of the edges entering @xmath833 is one or zero .
we also separate the graph @xmath758 into @xmath840 , @xmath841 and @xmath842 . here
, we define the numbers of vertices @xmath832 , @xmath833 , and @xmath834 as @xmath843 , @xmath844 , and @xmath845 , respectively .
@xmath842 is the graph consisted by vertices corresponding to @xmath839 , and `` @xmath767 '' and `` @xmath768 '' , and edges starting from these vertices .
we consider @xmath842 as a cluster , and define @xmath841 as the graph consisted by the vertices corresponding to @xmath833 , which acts on the cluster corresponding to @xmath846 .
@xmath847 is the rest of the graph @xmath758 without @xmath848 and @xmath842 . here
, we label vertexes @xmath846 , @xmath848 and @xmath849 by `` @xmath850 '' , `` @xmath851 '' and `` @xmath852 '' , respectively . the edge starting from `` @xmath142 '' and ending to `` @xmath853 '' represents @xmath854 .
next , we calculate weight @xmath855 and the operator @xmath856 corresponding to @xmath857 , and later those for @xmath858 . 0.5 cm _ separation of graph @xmath758 _
we now sketch the proof of @xmath859 , where @xmath860 for @xmath861 , and @xmath862 is given by eq .
( [ w1 ] ) .
and @xmath839 , respectively ; the dotted arrow , waved arrow , and real arrow represent @xmath863 , @xmath864 , and @xmath865 , respectively .
, width=113 ] from the additional rule a1 , each operator corresponding to vertexes @xmath834 and edges @xmath866 acts on @xmath113 and @xmath114 independently .
thus @xmath867 .
secondly we consider the graph which consists of four vertexes corresponding to @xmath868 , @xmath839 , and `` @xmath767''and `` @xmath768 '' as shown in fig .
[ graph0.eps ] .
we also assume that one edge of the vertex corresponding to @xmath868 connects with a vertex corresponding to @xmath839 . in this case , from additional diagram rule a3 , another edge of the vertex has to connect with `` @xmath767 '' or `` @xmath768 '' .
since we can exchange the role `` @xmath768 '' and `` @xmath767 '' by the variable transformation @xmath869 , ( @xmath810 ) , we assume that one edge of the vertex corresponding to @xmath868 connect with `` @xmath767 '' .
the weight @xmath760 in this case is given by eq .
( [ w2 ] ) , i.e. , the integrals for the weight is given by replacing coordinate of the vertex corresponding to @xmath839 with coordinate of `` @xmath768 '' in @xmath807 .
this result can be expanded to every graph though a graph includes the vertices @xmath832 .
for example , we illustrate the calculation of a six vertices graph , which only includes @xmath832 and @xmath834 , in fig .
[ gd2 ] . at first
we make the cluster having only vertices @xmath834 , @xmath113 and @xmath114 ( fig .
[ gd2](b ) ) , which is corresponding to the following operator : @xmath870 .
the edges from the vertices act on the cluster independently ( fig .
[ gd2](c ) ) ; we obtain the following operator : @xmath871 .
the position of each vertex corresponding to @xmath868 and @xmath839 can be move independently in integrals , and the entangled integral does not appear .
therefore the weight @xmath872 of a graph @xmath873 only depends on the number of vertexes corresponding to @xmath868 and @xmath839 , and @xmath874 holds generally . from additional rule a2 , we can similarly discuss about a graph @xmath847 , and obtain @xmath875 .
finally , we can count the combination of @xmath843 , @xmath844 and @xmath845 , and it is given by @xmath876 .
therefore we obtain the eq : @xmath859 . , where the dotted arrow and real arrow represent the derivative with respect to @xmath259 and @xmath258 , respectively , and the white circle and the white triangle represent @xmath877 and @xmath839 , respectively .
we rewrite the graph @xmath878 as the graph @xmath879 which is given by the cluster represented by the big circle and the operators into it , where the big circle represents the graph @xmath880 .
, width=453 ] the summation of each graph is easy , and we can derive the star product : @xmath881 , where twisted element @xmath304 is written as follow : @xmath882 because the action @xmath131 including a global u(1)@xmath68su(2 ) gauge field , the action @xmath131 is written as @xmath883 , thus , the map @xmath884 is given by @xmath885 , i.e. , @xmath886 the inverse map is given by the replacement of i by @xmath887 in the map ( [ eq : f_02a ] ) .
in this section , we will derive the twisted spin density , which corresponds to the spin density in commutative spacetime without the background su(2 ) gauge field .
first , we will derive a general form of the twisted spin current in section [ subsec : dtsws ] , which is written by using the twisted variation operator .
this operator is constituted of the coproduct and twisted element ; the coproduct reflects the action rule of the global su(2 ) gauge symmetry generator , and the twisted element represents gauge structure of the background gauge fields . in section [ sub : rdm ] , at first , we will calculate the twisted spin density of the so - called rashba - dresselhaus model in the wigner representation using the general form of the twisted spin density , and next , we will find the twisted spin operator in real spacetime using correspondence between operators in commutative spacetime and noncommutative phasespace . the lagrangian density in the wigner space is given by @xmath888 where @xmath6 is the electric mass , @xmath889 for any functions @xmath113 and @xmath114 .
the variation corresponding to the infinitesimal global su(2 ) gauge transformation is defined as @xmath890 where @xmath548 represents an infinitesimal parameter .
therefore , the variation of the lagrangian density @xmath891 is given by @xmath892 here we introduce the grassmann numbers @xmath893 ( @xmath894 ) , the product @xmath123 , and the coproduct @xmath895 , where the coproduct satisfies @xmath896 for any functions @xmath113 and operator @xmath186 .
the equation ( [ eq : dl ] ) can be rewritten as @xmath897 where we used @xmath898 with the kronecker delta @xmath899 here we introduce the following symbols : @xmath900 where the coproduct in the differential operator space is defined as @xmath901 here vectors @xmath902 corresponding to following operators : @xmath903 and @xmath904 , or @xmath903 and @xmath905 ( @xmath445 ) , and @xmath906 are bases of a vector space @xmath907 ( @xmath142 represents a scalar ) .
the coproduct @xmath125 in the vector space @xmath908 satisfies the coassociation law : @xmath909 because @xmath910 and @xmath911 it represents the leibniz rule with respect to the differential operator @xmath28 .
for instance ,
@xmath912 ; it corresponds to the following calculation : @xmath913 the variation ( [ eq : dl2 ] ) is rewritten as @xmath914 if we replace @xmath915 with @xmath916 in equation ( [ eq : dl3 ] ) , the integrals in terms of @xmath34 and @xmath259 of the right - hand side of eq .
( [ eq : dl3 ] ) become zero because @xmath917 does not include the su(2 ) field , which breaks the global su(2 ) gauge symmetry , in the case that the parameter @xmath548 is constant .
therefore , for the action @xmath918 , @xmath919 is the infinitesimal su(2 ) gauge transformation with background su(2 ) gauge fields .
because @xmath920 , we can write @xmath921 in the case that the infinitesimal parameter depends on the spacetime coordinate , this equation can be written as @xmath922 therefore , we obtain the twisted noether current @xmath923 in particular , the twisted spin @xmath924 is conserved quantity . here , we assumed that the su(2 ) gauge is static one .
however , we do not use this condition in the derivation of the twisted noether charge and current density .
then , we can derive the virtual twisted spin with a time - dependent su(2 ) gauge : @xmath925 . in this case
, we only use the time - dependent su(2 ) gauge field strength @xmath926 , which has non - zero space - time components @xmath927 ( @xmath928 ) . here
, we discuss the adiabaticity of the twisted spin . in the case that su(2 ) gauge fields have time dependence , the twisted spin is not conserved .
now , we assume that @xmath929 ( @xmath830 ) with constant fields @xmath930 ; @xmath931 is an adequate slowly function dependent on time .
because @xmath925 includes @xmath932^{-1 } \end{array } \right)$ ] with @xmath933 , the difference between @xmath925 and @xmath934 comes from only that between inverse of field strength : @xmath935^{-1 } -\lambda^{-2}[c_i , c_j]^{-1}\sim ( \dot{\lambda}/\lambda)^2[c_i , c_j]^{-1}$ ]
. therefore we obtain @xmath936 this means that @xmath934 is the adiabatic invariance .
namely , for the infinitely slow change in @xmath931 during the time period @xmath937 , @xmath934 remains constant while @xmath938 is finite .
this fact is essential for the spin - orbit echo proposed in @xcite .
here we apply the formalism developed so far to an explicit model , i.e. , the so - called rashba - dresselhaus model given by @xmath939 with a potential @xmath940 , where @xmath342 and @xmath343 are the rashba and dresselhaus parameters , respectively . completing square in terms of @xmath941
, we obtain @xmath942 , @xmath943 , @xmath944 , @xmath945 , @xmath946 , and @xmath947 , where @xmath948 .
to calculate the twisted symmetry generator @xmath949 , we first consider the @xmath950 .
@xmath951 is given by @xmath952 and @xmath953 is given by @xmath954 similarly , @xmath955 is given by @xmath956 we note that the operators @xmath957 have each inverse operator , which are denoted by @xmath958 , respectively .
here , the overline @xmath959 represents the complex conjugate .
because the twisted variation is @xmath960 , the infinitesimal parameter @xmath961 becomes an operator @xmath962 .
it is calculated by using the operator formula @xmath963\cdots]]\ ] ] for any operators @xmath964 and @xmath965 . in the calculation
, one will use the following formula in midstream : @xmath966 and so on , for any operator @xmath967 . from these results of the calculations ,
we obtain the twisted spin as follow @xmath968 where @xmath969 is defined as @xmath970\circ\left [ e^{-\frac{{\rm i}}{2}\big ( { \partial_x}\otimes\partial_y+ { \partial_y}\otimes\partial_x\big ) \circ\big(\frac{1}{2m(\alpha^2-\beta^2 ) } \otimes \sigma^z\big ) } \right.\nonumber\\ & & \left .
+ e^{-\frac{{\rm i}}{2}\big ( { \partial_x}\otimes\partial_y+ { \partial_y}\otimes\partial_x\big ) \circ\big ( \sigma^z\otimes\frac{1}{2m(\alpha^2-\beta^2 ) } \big ) } \right ] + e^{\frac{{\rm i}}{2}(\alpha+\beta)\sigma_- \big(\partial_{p_x}+\partial_{p_y}\big ) } \otimes \sin^2\left ( \frac{{\rm i}\hbar \big(\partial_x-\partial_y\big)}{8\sqrt{2}m(\alpha+\beta ) } \right),\end{aligned}\ ] ] where @xmath971 .
finally , we will rewrite the twisted spin as an operator form in commutative spacetime . roughly speaking , the operator in the commutative spacetime . and
the one in the noncommutative wigner space have the following relations ( the left - hand side represents operators on the wigner space ; the right - hand side represents operators on the commutative spacetime ) : @xmath972 because @xmath280_\star:=x^\mu\star p_\nu - p_\nu\star x^\mu= { \rm i}\hbar\delta_\nu^\mu$ ] is equal to the commutation of the operator form : @xmath281={\rm i}\hbar\delta_\nu^\mu$ ] .
the equivalence of @xmath973 and the twisted spin in the operator form on the commutative spacetime @xmath974 can be confirmed using the wigner transformation in terms of @xmath974 .
the operator form of the twisted spin is given by @xmath975 where @xmath976\end{aligned}\ ] ] with @xmath977 .
this operator in eq .
( [ eq : tso ] ) , when integrated over @xmath978 , is the conserved quantity for any potential configuration @xmath940 as long as @xmath342 , @xmath343 are static and the electron - electron interaction is neglected .
in this paper , we have derived the conservation of the twisted spin and spin current densities .
also the adiabatic invariant nature of the total twisted spin integrated over the space is shown . here
we remark about the limit of validity of this conservation law .
first , we neglected the dynamics of the electromagnetic field @xmath0 which leads to the electron - electron interaction .
this leads to the inelastic electron scattering , which is not included in the present analysis , and most likely gives rise to the spin relaxation .
this inelastic scattering causes the energy relaxation and hence the memory of the spin will be totally lost after the inelastic lifetime .
this situation is analogous to the two relaxation times @xmath979 and @xmath980 in spin echo in nmr and esr .
namely , the phase relaxation time @xmath980 is usually much shorter than the energy relaxation time @xmath979 , and the spin echo is possible for @xmath981 .
similar story applies to spin - orbit echo @xcite where the recovery of the spin moment is possible only within the inelastic lifetime of the spins .
however , the generalization of the present study to the dynamical @xmath0 is a difficult but important issue left for future investigations .
also the effect of the higher order terms in @xmath13 in the derivation of the effective lagrangian from dirac theory requires to be scrutinized .
another direction is to explore the twisted conserved quantities in the non - equilibrium states . under the static electric field ,
the system is usually in the current flowing steady state .
usually this situation is described by the linear response theory , but the far from equilibrium states can in principle be described by the non - commutative geometry @xcite .
the nonperturbative effects in this non - equilibrium states are the challenge for theories , and deserve the further investigations .
the author thanks y.s .
zhang , k. richter , v. krueckl , j. nitta , and s. onoda for useful discussions .
this work was supported by priority area grants , grant - in - aids under the grant number 21244054 , strategic international cooperative program ( joint research type ) from japan science and technology agency , and by funding program for world - leading innovative r and d on science and technology ( first program ) .
m. chaichian , p. kulish , k. nishijima and a. tureanu . on a lorentz - invariant interpretation of noncommutative space - time and its implications on noncommutative qft .
physics letters b , * 604*(1 - 2 ) , 98102 ( 2004 ) .
g. amelino - camelia , g. gubitosi , a. marcian , p. martinetti , f. mercati , d. pranzetti and r. a. tacchi .
first results of the noether theorem + for hopf - algebra spacetime symmetries .
progress of theoretical physics supplement , * 171 * , 6578 ( 2007 ) .
d. happel .
hochschild cohomology of finite dimensional algebras . in m .-
malliavin , editor , _
sminaire dalgbre paul dubreil et marie - paul malliavin _ ,
1404 of _ lecture notes in mathematics _ , pages 108126 .
springer berlin / heidelberg .
isbn 978 - 3 - 540 - 51812 - 9 ( 1989 ) . | the quantization scheme by noncommutative geometry developed in string theory is applied to establish the conservation law of twisted spin and spin current densities in the spin - orbit coupled systems . starting from the pedagogical introduction to hopf algebra and deformation quantization ,
the detailed derivation of the conservation law is given . |
the cabibbo - kobayashi - maskawa ( ckm ) matrix element @xmath11 describes the rich phenomena of flavor - changing transitions between the two heavy quarks @xmath12 and @xmath13 .
its precise extraction has become a very important issue in heavy flavor physics .
generally @xmath0 is extracted by studying either exclusive or inclusive semileptonic @xmath1 decays . since long distance contributions are involved in these decays , theoretical estimates of the nonperturbative parameters are of crucial importance for a precise extraction of @xmath0 .
heavy quark symmetry ( hqs ) @xcite and effective field theories of heavy quarks play a special role in such estimates . in this short note
, we are going to provide a more precise extraction for @xmath0 from both exclusive and inclusive semileptonic @xmath1 decays within the framework of hqeft of qcd .
this effective field theory can directly be derived from qcd@xcite by integrating out the small components but with carefully treating the quark and antiquark components .
recently , it has been shown that the hqeft can actually be regarded as a large component qcd@xcite .
of particular , this effective field theory has been put forward and successfully applied to various hadron processes @xcite .
the resulting effective lagrangian and the heavy quark expansion ( hqe ) in the hqeft of qcd appear to be different from the usual heavy quark effective theory ( hqet ) @xcite in which the quark and antiquark components were dealt with separately , i.e. , there is no quark - antiquark coupling terms considered in hqet . the application of the usual hqet in extracting @xmath0 has been discussed by several groups and summarized in ref.@xcite .
we shall further emphasize in this note the relevant new features of hqeft with respect to the usual hqet .
our paper is organized as follows : @xmath0 extraction from exclusive and inclusive b decays will be discussed in section [ exclusive ] and section [ inclusive ] , respectively .
the previously obtained values for @xmath0 will be updated by using the most recent experimental measurements . and
a brief conclusion will be presented in section [ conclusion ] .
the @xmath14 differential decay rates are @xmath15\vert v_{cb } \vert^2 { \cal f}^2(\omega ) , \\ \label{widthb2d } \frac{d\gamma(b\rightarrow dl\nu)}{d\omega}&=&\frac{g^2_f}{48\pi^3}(m_b+m_{d})^2 m^3_{d}(\omega^2 - 1)^{3/2 } \vert v_{cb } \vert^2 { \cal g}^2(\omega ) \end{aligned}\ ] ] with @xmath16 = \eta_{v } ( 1+\delta ) , \end{aligned}\ ] ] where the qcd radiative corrections to two loops give the short distance coefficients @xmath17 and @xmath18 @xcite .
@xmath19 is the product of the four - velocities of the @xmath1 and @xmath20 mesons , @xmath21 .
the weak transition form factors @xmath22 , @xmath23 and @xmath24 can be expanded in powers of @xmath25 and represented by the heavy quark spin - flavor independent wave functions .
then based on eqs.([widthb2ds ] ) and ( [ widthb2d ] ) , @xmath0 can be precisely extracted as long as the form factors are reliably evaluated in some theoretical framework .
the hqe of the transition form factors are studied in detail up to order @xmath2 in hqeft framework in ref.@xcite .
when the contributions of operators containing two gluon field strength tensors are omitted , at the zero recoil point @xmath26 the relevant form factors can be written as @xmath27 ^ 2 -\frac{1}{8m^2_b\bar{\lambda}^2 } ( f_{1}+3f_{2 } \nonumber\\ & & -2\bar{\lambda}\varrho_{1}-6\bar{\lambda}\varrho_{2 } ) -\frac{1}{8m_c^2\bar{\lambda}^2}(f_{1}-f_{2}-2\bar{\lambda}\varrho_{1 } + 2\bar{\lambda}\varrho_{2 } ) \nonumber\\ & & + \frac{1}{4m_b m_c \bar{\lambda}^2}(f_{1}+f_{2}-2\bar{\lambda}\varrho_{1 } -2\bar{\lambda}\varrho_{2 } ) , \\
\label{formfactorhz } h_{+}&= & 1+\frac{1}{8\bar{\lambda}^2}(\frac{1}{m_b}-\frac{1}{m_c})^2 [ ( \kappa_1 + 3\kappa_2)^2-(f_1 + 3f_2 ) + 2\bar{\lambda}(\varrho_1 + 3\varrho_2 ) ] , \\
\label{formfactorhf } h_{-}&= & 0 , \end{aligned}\ ] ] where the parameters in the rhs . of eqs.([formfactorha1])-([formfactorhf
] ) are defined in hqeft @xcite . for simplicity ,
when the variable @xmath19 is not written explicitly , we refer to the zero recoil values of relevant functions , i.e. , @xmath28 , @xmath29 , etc .
the binding energy of a heavy meson @xmath30 is defined as @xmath31 it is seen from eqs.([formfactorha1 ] ) and ( [ formfactorhz ] ) that automatically both the form factors @xmath32 and @xmath33 in hqeft of qcd do not receive @xmath25 order correction at the zero recoil point . another point to be emphasized is eq.([formfactorhf ] ) .
the vanished value of @xmath34 in hqeft arises from the fact of partial cancellation between the @xmath25 correction in the current expansion and the @xmath25 correction coming from the insertion of the effective lagrangian into the transition matrix elements @xcite .
such a cancellation is not observed in hqet because in the latter framework the quark - antiquark couplings are not taken into account explicitly .
though the so - called luke s theorem in hqet protects the weak transition matrix elements from @xmath25 order correction at zero recoil point , it does not protect @xmath34 from such correction @xcite . as a result , in hqet framework
, only the @xmath35 decay rate at zero recoil is strictly protected against @xmath25 order correction while @xmath36 decay rate is not .
this is the main reason besides the experimental considerations for the conclusion that the @xmath36 decay is not as favorable as @xmath37 decay for @xmath0 extraction in hqet . in hqeft ,
eq.([formfactorhf ] ) advocates reliable extraction of @xmath0 from both @xmath38 and @xmath36 decays , because both rates of these decays do not receive @xmath25 order correction , as can be seen from eqs.([widthb2ds])-([formfactorhf ] ) .
however , we note that the current world averages for @xmath39 and @xmath40 are @xcite @xmath41 so the current experimental data for @xmath40 receive larger errors than those for @xmath39 , which may lead to large experimental uncertainty for the result of @xmath0 extracted from @xmath36 . now
the value of @xmath0 depends on our estimates of @xmath42 and @xmath43 .
first of all , it has been shown @xcite that some of the nonperturbative parameters appearing in the rhs .
of eqs.([formfactorha1 ] ) -([formfactorhf ] ) can be related to the heavy meson masses .
explicitly , we have @xmath44 thus some parameters can be determined from the heavy meson masses .
a detailed study has been presented in ref.@xcite . here
we would not repeat the analysis .
it is easily read from ref.@xcite that by using @xmath45 one obtains the following values for @xmath46 , @xmath47 and @xmath48 , @xmath49 : @xmath50 which agree well with the sum rule results : @xmath51 , @xmath52 and @xmath53 @xcite . then up to order @xmath2 only two parameters @xmath54 and @xmath55 remain unknown . to have an estimation for them , we recall the definition of the nonperturbative parameters .
they are defined by the matrix elements as follows @xcite : @xmath56 \nonumber\\ & & + \frac{1}{\bar{\lambda}}tr[\kappa_{\alpha\beta}(v , v^{\prime } ) \bar{\cal m}^{\prime}\gamma p_{+}\frac{i}{2}\sigma^{\alpha\beta}{\cal m } ] , \nonumber\\ < m^{\prime}_{v^{\prime}}\vert { \bar{q}^{(+)}_{v^{\prime } } } \gamma & & \frac{1}{iv\cdot d } ( i{d\hspace{-0.25cm}\slash_{\bot}})(-iv\cdot d ) ( i{d\hspace{-0.25cm}\slash_{\bot } } ) { q^{(+)}_v}\vert m_v > = -\varrho_1(\omega)\frac{1}{\bar{\lambda}}tr[\bar{\cal m}^{\prime}\gamma { \cal m } ] \nonumber\\ & & + \frac{1}{\bar{\lambda}}tr[\varrho_{\alpha\beta}(v , v^{\prime})\bar{\cal m}^{\prime}\gamma p_{+ } \frac{i}{2}\sigma^{\alpha\beta}{\cal m } ] , \nonumber \\
< m^{\prime}_{v^{\prime}}\vert { \bar{q}^{(+)}_{v^{\prime } } } \gamma & & \frac{1}{iv\cdot d } ( i{d\hspace{-0.25cm}\slash_{\bot}})^2 \frac{1}{iv\cdot d } ( i{d\hspace{-0.25cm}\slash_{\bot}})^2 { q^{(+)}_v}\vert m_v > = -\chi_1(\omega)\frac{1}{\bar{\lambda}^2 } tr[\bar{\cal m}^{\prime}\gamma { \cal m } ] \nonumber\\ & & + \frac{1}{\bar{\lambda}^2 } tr[\chi_{\alpha\beta}(v , v^{\prime})\bar{\cal m}^{\prime}\gamma p_{+ } \frac{i}{2}\sigma^{\alpha\beta}{\cal m } ] + \cdots , \end{aligned}\ ] ] where the ellipsis in the last equation represents the contributions of operators containing two gluon field strength tensors .
the lorentz tensor @xmath57 is decomposed into scalar factors as @xmath58 , and @xmath59 , @xmath60 are decomposed similarly .
the functions @xmath48 and @xmath49 are defined as @xmath61 as will be further emphasized in the next section , the operator @xmath62 in the matrix elements was considered to give contribution of order the binding energy @xcite , i.e. , @xmath63 , which has also been confirmed in the simple case by sum rule calculation @xcite . now
if taking this simple replacement , we get from eqs.([wavefunctiondef ] ) @xmath64 note that our estimates in eqs.([estfrommass ] ) and ( [ rho1estimate ] ) are consistent with the relation among them in eq.([definef1 ] ) . in eq.([estfrommass ] ) @xmath49 almost equals zero .
of course one reason for this may be the possible partial cancellation between @xmath65 and @xmath66 in eq.([definef2 ] ) . but more importantly
, we notice that @xmath65 , @xmath55 ( and therefore @xmath49 ) are parameters characterizing the chromomagnetic type operators , and the contributions of these chromomagnetic operators are generally believed to be small .
if using the same method of estimation as that for @xmath54 and @xmath67 , one gets @xmath68 it can be seen from the above equations that the two terms in eq.([definef2 ] ) do have partial cancellation .
however , eqs.([rho2estimate ] ) , ( [ chi2estimate ] ) and ( [ definef2 ] ) can not hold simultaneously .
this indicates that eqs.([rho2estimate ] ) and ( [ chi2estimate ] ) are only rough estimates for the chromomagnetic type parameters since these parameters are very small in magnitude .
nevertheless , as @xmath55 must be small and can not influence the final result for @xmath0 as significantly as @xmath54 does , here we would first take eq.([rho2estimate ] ) in extracting @xmath0 but leave more rigid determination of @xmath55 for future work .
fig.1 presents @xmath69 as a function of @xmath70 at the fixed quark mass difference @xmath71gev .
the curves are relatively flat in the region around @xmath72gev , which indicates a reliable extraction of @xmath69 around this point .
this value of @xmath70 is in agreement with that in eq.([estfrommass ] ) .
varying @xmath73 and @xmath74 in the ranges @xmath75 and @xmath76 , and taking @xmath77 and @xmath78 , we get @xmath79 from eqs.([exp1 ] ) and ( [ deltas1 ] ) , we then obtain @xmath80 for @xmath36 decay , analogously , we obtain @xmath81 and @xmath82 comparing eqs.([vcbbds ] ) and ( [ vcbbd ] ) , we see that the results of @xmath0 extracted from @xmath14 decays have similar theoretical errors . but
large difference exists between the central values of our present results , eqs.([vcbbds ] ) and ( [ vcbbd ] ) .
however , due to the large experimental error for @xmath36 channel , the @xmath0 values extracted above are compatible with each other . a better determination of @xmath0 from @xmath36 decay
is expected when a more precise result for the quantity @xmath40 is obtained .
note that in ref.@xcite @xmath54 and @xmath55 are not estimated as above .
there @xmath83 is assumed so that the central values for @xmath0 extracted from @xmath84 decays can be close to each other .
indeed , @xmath83 together with the experimental averages ( [ exp1 ] ) , ( [ exp2 ] ) lead to @xmath85 from @xmath38 decay and @xmath86 from @xmath36 decay .
inclusive semileptonic @xmath1 decays is the other alternative to determine @xmath0 . in the usual hqet
the light quark in a hadron is generally treated as a spectator , which does not affect the heavy hadron properties to a large extent .
however , this treatment may be one possible reason for the failure of hqet in some applications .
for example , the world average value for bottom hadron lifetime ratio @xmath87 can not be explained well in the usual framework of hqet @xcite .
the dynamics of inclusive @xmath1 decays is also analyzed in detail in refs.@xcite within the hqeft framework . instead of simply applying the equation of motion for infinitely heavy free quark , @xmath88
, we treat the heavy quark in a hadron as a dressed particle , which means that the residual momentum @xmath89 of the heavy quark within a hadron is considered to comprise contributions from the light degrees of freedom .
this simple picture is adopted to conveniently take into account the effects of light degrees of freedom and the binding effects of heavy and light components of the hadron but not deal with the complex dynamics of hadronization directly .
as the light degrees of freedom within the heavy hadron is relativistic , one has @xmath90 explicitly , the momentum of a heavy hadron @xmath91 is represented as @xmath92 with @xmath93 being the momentum depending on the heavy flavor and suppressed by the inverse of the heavy quark mass .
this directly leads to the relation between @xmath94 and the binding energy , @xmath95 or @xmath96 to acquire a good convergence of hqe , we perform the expansion in terms of @xmath97 ( or say , equivalently , in terms of @xmath98 .
then the @xmath99 decay rate is found to be @xcite @xmath100 where @xmath101 , @xmath102 , @xmath103 and @xmath104 are functions of the mass square ratios @xmath105 , @xmath106 and @xmath107 .
here the calculation is performed up to nonperturbative order @xmath108 and perturbative order @xmath109 .
the function @xmath110 characterizes qcd radiative corrections .
its two - loop results were obtained in refs.@xcite . @xmath111 and @xmath112
can be determined from the meson masses via eq.([massb ] ) .
@xmath47 is often extracted from the known @xmath113 mass splitting @xmath114 which is consistent with eq.([estfrommass ] ) and the sum rule result @xcite .
there are several points to be mentioned for eq.([incwidth ] ) .
firstly , in deriving eq.([incwidth ] ) the effects of light degrees of freedom are explicitly accounted for in the picture of a dressed heavy quark in a hadron .
secondly , it is seen that the next leading order contributions vanish in our hqe in terms of the inverse dressed heavy quark mass , @xmath115 .
furthermore , our hqe in terms of @xmath116 ( or @xmath115 ) has a good convergence .
it is found that the @xmath117 order contributions induce only @xmath118 corrections to the total width @xmath119 .
therefore we conclude that the higher order nonperturbative corrections can be safely neglected .
finally , now one needs only to treat the dressed quark mass @xmath101 instead of considering the uncertainties arising from the two quantities @xmath73 and @xmath120 separately .
note that these are the features of hqe in hqeft .
they can not be observed in the hqe in the usual hqet , where one assumes @xmath121 to be zero or of higher order of @xmath122 . in the latter case
the next leading order corrections can be absent only when the hqe is performed in terms of @xmath122 , and the heavy quark mass @xmath73 and the binding energy @xmath120 have to be treated separately .
one result of this is that in hqet the theoretical prediction of the total decay width strongly depends on the value of bottom quark mass @xmath73 and may have larger uncertainties than in hqeft . using eq.([incwidth ] )
, @xmath0 can be extracted from experimental data for inclusive decay rates .
fig.3 shows the obtained values of @xmath123 as a function of the energy scale @xmath124 and the parameters @xmath125 , @xmath46 .
it is seen that the extracted value of @xmath0 depends on the energy scale @xmath124 weakly .
so the main uncertainties come from @xmath125 and @xmath46 .
the curves in fig.3(b ) and fig.3(c ) have minimal value points , around which the values of @xmath126 are favorable because they are less sensitive to the variation of @xmath125 and @xmath46 . varying @xmath125 and @xmath46 in the regions @xmath127 , @xmath128 , these favorable values of @xmath126 change in the range : @xmath129 where the central value is obtained at @xmath130 and @xmath131 .
then the @xmath132 lifetime @xmath133 and the most recent cleo data for @xmath134 branching ratio @xmath135 @xcite yield latexmath:[\[\begin{aligned } @xmath0 extracted using different values of branching ratios is presented in fig.4 . as mentioned above
, the values at @xmath137gev should be favorable because in this region the curves become less sensitive to the mass @xmath125 .
interestingly , as shown in fig.3(c ) , when choosing @xmath138gev we find that the curve of @xmath126(or the resultant @xmath0 ) as a function of @xmath139 reaches its minimal ( or maximal ) point at @xmath140 .
this value for @xmath46 is again in good agreement with that we used in section [ exclusive ] and with that obtained from sum rule calculation @xcite .
note that the mass @xmath125 discussed here arises from the charm quark propagator in the hqe of the matrix elements
. it should be the pole mass of charm quark .
thus it is not surprising that the value of @xmath125 obtained here is larger than the value for constituent mass quoted in section [ exclusive ] .
we have presented a more precise extraction of @xmath0 by studying the exclusive and inclusive semileptonic @xmath1 decays up to the order of @xmath2 in hqeft of qcd .
the nonperturbative parameters in hqeft up to the same order are estimated consistently from various considerations .
it has been shown that in hqeft of qcd @xmath0 can be reliably extracted from exclusive decays @xmath14 with similar theoretical uncertainties , because neither of their differential decay rates receives @xmath25 order corrections at zero recoil point . in studying inclusive
@xmath1 decays , we treat the heavy quark in a hadron as a dressed particle whose residual momentum comprises some effects from the light degrees of freedom .
this enables us to consider the effects of light components in the hadron but still has a simple physical picture in application . in this framework ,
the @xmath141 total decay rate is protected from @xmath115 order correction , and the @xmath117 order correction is very small .
furthermore , @xmath73 and @xmath120 only appear in the form of dressed quark mass : @xmath101 , which must also reduce the uncertainties in our calculations as @xmath142 $ ] . using the most recent experimental data for the b - meson decay rates , we have arrived at the determination for @xmath0 with @xmath143 where the result extracted from @xmath144 decay receives a larger experimental uncertainty than that from @xmath145 decay but a similar theoretical uncertainty as the latter
. the result obtained from @xmath37 decay agrees quite well with that from inclusive @xmath146 decay .
e. v. shuryak , phys . lett .
* b 93 * ( 1980 ) 134 ; nucl . phys . * b 198 * ( 1982 ) 83 .
s. nussinov , w. wetzel , phys . rev . *
d 36 * ( 1987 ) 130 .
n. isgur , m. wise , phys .
* b 232 * ( 1989 ) 113 ; * b 237 * ( 1990 ) 527 ; * b 206 * ( 1988 ) 681 .
y. l. wu , mod .
. lett . * a 8 * ( 1993 ) 819 .
y. l. wu , y. a. yan , m. zhong , y. b. zuo , w. y. wang , mod .
. lett . * a 18 * ( 2003 ) 1303 .
w. y. wang , y. l. wu , y. a. yan , int .
phys . * a 15 * ( 2000 ) 1817 .
y. a. yan , y. l. wu , w. y. wang , int .
phys . * a 15 * ( 2000 ) 2735 .
w. y. wang , y. l. wu , int . j. mod .
phys . * a 16 * ( 2001 ) 377 .
y. l. wu , y. a. yan , int . j. mod
. phys . * a 16 * ( 2001 ) 285 . w. y. wang , y. l. wu , phys
b 515 * ( 2001 ) 57 . w. y. wang , y. l. wu , phys . lett . *
b 519 * ( 2001 ) 219 . w. y. wang , y. l. wu , m. zhong , phys . rev . *
d 67 * ( 2003 ) 014024 .
m. zhong , y. l. wu , w. y. wang , int
phys . * a 18 * ( 2003 ) 1959 .
f. berto , m. a. sanchis - lozano , nuov .
cim . * a 112 * ( 1999 ) 1181 .
f. berto , m.a .
sanchis - lozano , nucl .
* 93 * ( 2001 ) 212 .
m. a. sanchis - lozano , int .
j. mod . phys . * a 16 * ( 2001 ) 4189 .
h. georgi , phys . lett .
* b 240 * ( 1990 ) 447 .
for review see e.g. , m. neubert , phys .
* 245 * ( 1994 ) 259 .
k. hagiwara et al .
, phys . rev .
* d 66 * ( 2002 ) 010001 - 701 ; and references therein .
a. czarnecki , phys .
* 76 * ( 1996 ) 4124 .
a. f. falk , m. neubert , phys .
* d 47 * ( 1993 ) 2965 .
m. neubert , c. sachrajda , nucl . phys .
* b 483 * ( 1997 ) 339 ; m. neubert , hep - ph/9707217 ; and references therein . m. s. baek , j. lee , c. liu , h. s. song , phys . rev . * d 57 * ( 1998 ) 4091 .
m. luke , m. j. savage , m. b. wise , phys .
* b 345 * ( 1995 ) 301 .
m. lu , m. luke , m. j. savage , b. h. smith , phys . rev . *
d 55 * ( 1997 ) 2827 .
b. c. barish et al .
( cleo ) , phys .
* 76 * ( 1996 ) 1570 . | the more precise extraction for the ckm matrix element @xmath0 in the heavy quark effective field theory ( hqeft ) of qcd is studied from both exclusive and inclusive semileptonic @xmath1 decays .
the values of relevant nonperturbative parameters up to order @xmath2 are estimated consistently in hqeft of qcd .
using the most recent experimental data for @xmath1 decay rates , @xmath3 is updated to be @xmath4 from @xmath5 decay and @xmath6 from @xmath7 decay as well as @xmath8 from inclusive @xmath9 decay . #
1@xmath10#1 : 12.15.hh , 12.39.hg , 13.20.he @xmath0 , heavy quark effective field theory |
in physical theories on a fixed background spacetime , as in newtonian theory or special relativity , it is not difficult to formulate asymptotic fall - off conditions on fields of spatially bounded systems .
for example , the gravitational potential due to a newtonian star is usually required to decay to zero at infinity of euclidean space , with the decay rate being compatible with laplace s equation . in general relativity
no a priori given background space exists .
the metric itself is both a dynamical field and a quantity which determines distances .
one expects that in a suitable coordinate system far away from a system of bodies the metric should have a form @xmath0 small quantities , where @xmath1 is minkowski metric .
what , however , does it mean `` far away '' , what is `` infinity '' ?
can one formulate suitable boundary conditions in a coordinate - free manner ?
what is the decay of a _ radiative _ gravitational field ?
after several important contributions to the gravitational radiation theory in the late 1950 s and early 1960 s by pirani , bondi , robinson , trautman and others , a landmark paper by bondi et al @xcite appeared in which radiative properties of isolated ( spatially bounded ) axisymmetric systems were studied along outgoing null hypersurfaces @xmath2 , with @xmath3 representing a retarded time function .
an ansatz was made that the metric along @xmath2 can be expanded in inverse powers of @xmath4 , @xmath5 where @xmath4 denotes a suitable parameter along null generators ( parametrized by coordinates @xmath6 ) on the hypersurfaces @xmath2 . under the assumption ( [ rse1 ] )
einstein s vacuum equations were shown to determine uniquely formal power series solution of the form ( [ rse1 ] ) , provided that a free `` news function '' @xmath7 is specified .
the news function contains all information about radiation at infinity @xmath8 .
it enters the fundamental `` bondi mass - loss formula '' for the total mass @xmath9 of an isolated system at retarded time @xmath3 .
field equations imply that @xmath9 is a monotonically decreasing function of @xmath3 if @xmath10 a natural interpretation is that gravitational waves carry away positive energy from the system and thus decrease its mass . in the work of bondi
et al @xcite as well as in the important generalizations by sachs , newman and penrose , the decay of radiative fields was studied in preferred coordinate systems . in 1963
penrose @xcite formulated a beautiful _ geometrical _ framework for description of the `` radiation zone '' in general relativity in terms of conformal infinity .
penrose s definition of asymptotically flat radiative spacetimes avoids such problems as `` distances '' or `` suitable coordinates '' , and incorporates a clear definition of what is infinity .
it is inspired by the work on radiation theory mentioned above , and by the properties of conformal infinity in minkowski spacetime .
in contrast to an euclidean space , in minkowski spacetime one can go to infinity in various directions : moving along timelike geodesics we come to the future ( or past ) timelike infinity @xmath11 ( or @xmath12 ) ; along null geodesics ( cf .
( [ rse1 ] ) ) we reach the future ( past ) null infinity @xmath13 ; spacelike geodesics lead to spatial infinity @xmath14 .
minkowski spacetime can be compactified and mapped into a finite region by an appropriate conformal transformation .
thus one obtains the well - known _ penrose diagram _ in which the three types of infinities are mapped into the boundaries of the compactified spacetime - see fig . 1 .
it is generally accepted that penrose s definition forms the only rigorous , geometrical basis for the discussion of gravitational radiation from isolated systems .
it enables us to use techniques of local geometry `` at infinity '' and to define covariantly such fundamental quantities as the total ( bondi ) mass of an isolated system .
despite its rigour and elegance , penrose s _ definition _ might turn to be of a limited importance if no interesting radiative spacetimes _ exist _ which satisfy the definition . in section 5
we shall describe special exact radiative spacetimes which represent `` uniformly accelerated '' sources in general relativity and admit @xmath15 as required ; however , at least four points on @xmath15 - those in which worldlines of the sources start and end - are singular .
there are no other _ explicit _ exact radiative solutions describing finite sources available at present and this situation will probably not change soon .
nevertheless , thanks to the work of friedrich @xcite and christodoulou and klainerman @xcite we know that globally non - singular ( including _ all _ @xmath15 ) asymptotically flat exact solutions of einstein s equations really exist .
the key idea of friedrich is in realizing that penrose s treatment of infinity not only permits to use the methods of local differential geometry at @xmath15 but also to analyze global existence problems of solutions of einstein s equations in physical spacetime @xmath16 by solving initial value problems in unphysical spacetime @xmath17 . by using this approach
friedrich established that formal bondi - type expansions ( [ rse1 ] ) converge locally at @xmath18 .
he succeeded to show that one can formulate the `` hyperboloidal initial value problem '' for einstein s vacuum equations in which initial data are given on a hyperboloidal spacelike hypersurface @xmath19 which intersects @xmath15 ( see fig .
it can then be proven that hyperboloidal initial data , which are sufficiently close to minkowskian hyperboloidal data ( i.e. to the metric induced on the hypersurface @xmath19 in minkowski spacetime by the standard minkowski metric ) evolve to a vacuum spacetime which is smooth on @xmath20 and @xmath21 as required by penrose s definition . despite the deep and complicated work by friedrich ,
one still did not have what one really would wish .
let us first note that initial data _
`` sufficiently close '' _ to minkowskian data in friedrich s result we mentioned above , do not mean any approximation - `` smallness '' of data is understood in a functional sense in an appropriate sobolev space .
one can not hope to prove global existence of smooth general solutions developed from general , arbitrarily _
strong _ data .
vacuum data representing strong gravitational waves could , due to nonlinearities , lead to creation of black holes with singularities inside .
nevertheless , ultimately one would like to have also results on the evolution of strong data .
more importantly , however , we would like to have initial data given on a _ standard _ spacelike cauchy hypersurface which does not intersect @xmath15 but `` ends '' at spatial infinity ( cf .
hypersurface @xmath22 in fig .
1 ) , rather than data given on a hyperboloidal initial hypersurface .
it could well happen that a spacetime evolved from data on @xmath19 is smooth `` above '' @xmath19 ( in the shaded region in fig .
1 ) , however , it does not satisfy penrose s requirements of asymptotic flatness `` bellow '' @xmath19 .
a remarkable progress in proving rigorously the existence of general , asymptotically flat radiative spacetimes was achieved by christodoulou and klainerman .
their treatise @xcite contains the first really global general existence statement for full , nonlinear einstein s vacuum equations with vanishing cosmological constant : any smooth asymptotically flat initial data set ( determined by the first and the second fundamental form on a cauchy hypersurface ) which is `` near flat ( minkowski ) data '' leads to a unique , smooth and geodesic complete development solution of einstein s vacuum equations .
this solution is `` globally asymptotically flat '' in the sense that the curvature tensor decays to zero at infinity in all directions .
the christodoulou - klainerman theorem involves a `` global smallness assumption '' which requires appropriate ( integral ) norms of curvature tensor of initial data integrated over the initial cauchy hypersurface to be small .
the theorem demonstrates the existence of singularity - free , asymptotically flat radiative vacuum spacetimes . in hydrodynamics , even for arbitrary small initial data ( decaying at infinity ) , an analogous theorem is not true since shocks arise . in general relativity
the effect of nonlinear terms , which could have led to formation of singularities , is excluded owing , apparently , to the covariance and algebraic properties of the einstein vacuum equations .
that singularities could have had well developed is evident : the collision of arbitrarily weak , vacuum plane gravitational waves leads ( due to nonlinearities ) to the formation of a singularity .
this , of course , does not contradict the christodoulou - klainerman result because initial data for plane waves are not asymptotically flat .
the work of friedrich , christodoulou , klainerman and others demonstrates rigorously that the general picture of null infinity is compatible with the vacuum einstein field equations
. however , important open questions remain . in classical papers by bondi et al @xcite , newman and penrose and others , the decay of the curvature ( characterized by the weyl tensor ) along outgoing null geodesics at infinity exhibits `` the peeling - off '' property : the fall - off of various components of the weyl tensor is related to their petrov algebraic type . to be more specific , certain complex linear combinations of the weyl tensor in the orthonormal frame , @xmath23 , behave as @xmath24 as @xmath25 .
( in particular , @xmath26 has the same algebraic structure as the weyl tensor of a plane wave radiative field of a bounded system resembles asymptotically that of a plane wave . )
this decay of the curvature can be shown to follow from a sufficient differentiability ( smoothness ) of the conformally rescaled ( unphysical ) metric @xmath27
. a sufficient smoothness of null infinity is thus commonly assumed .
the results of christodoulou and klainerman , however , show a weaker peeling .
they were only able to prove that the asymptotically flat vacuum initial data lead to @xmath28 ( _ not _
@xmath29 ) at null infinity .
an increasing evidence against the proposal of a smooth @xmath15 has led chruciel et al @xcite to introduce the concept of a _ polyhomogeneous _ @xmath15 .
the metric is called polyhomogeneous if at large _
r _ it admits an expansion in terms of @xmath30 rather than @xmath31 ( as it has been assumed in the works of bondi and others - cf . eq .
( [ rse1 ] ) ) .
the hypothesis of polyhomogeneity of @xmath15 has been shown to be formally consistent with einstein s vacuum equations . under appropriate assumptions on the asymptotic form of the polyhomogeneous metric
, one can demonstrate that the bondi mass - loss law can be formulated , and the peeling - off property of the curvature holds , with the first two terms identical to the standard peeling , the third term being @xmath32 .
at null infinity the conformally rescaled ( unphysical ) metric is not smooth .
although a substantial progress in understanding the existence of radiative solutions of vacuum field equations and the asymptotic structure of corresponding radiative spacetimes has been achieved , we have seen that open problems remain .
curiously enough , in the case of vacuum einstein s equations with a _ non - vanishing cosmological constant _ a more complete picture is known for some time already . by using his regular conformal field equations ,
friedrich demonstrated @xcite that initial data sufficiently close to de - sitter data develop into solutions of einstein s equations with a positive cosmological constant , which are asymptotically simple ( with a smooth conformal infinity ) , as required in the original framework of penrose .
later friedrich @xcite also discussed the existence of asymptotically simple solutions to the einstein vacuum equations with a negative cosmological constant . in his more recent investigations @xcite ,
friedrich constructed the new - finite but `` wider '' than the point @xmath33 - representation of spacelike infinity .
this construction enables one to make much deeper analysis of the initial data in the region where null infinity touches spacelike infinity .
good chances now exist to obtain clear criteria determining which data lead to the smooth and which just to the polyhomogeneous null infinity .
the ultimate goal of rigorous work on the existence and asymptotics of solutions of the einstein equations is _ physics _ :
one hopes to be able to consider astrophysical sources , to relate their behaviour to the characteristics of the far fields .
one would like to have under control various ( both analytical and numerical ) approximation procedures .
a still more ambitious program is to consider strong initial data so as to be able to analyze such issues as cosmic censorship . in the following
we shall briefly discuss three classes of _ explicit _ radiative solutions of einstein s equations : cylindrical waves , robinson - trautman and type n spacetimes , and the boost - rotation symmetric spacetimes .
the latter represent the only known examples describing moving , radiating objects ; except for points of null infinity where particles start and end , the null infinity is smooth .
we here closely follow our recent , more detailed reviews @xcite in which also other classes of radiative spacetimes are analyzed , such as plane waves and gravitational waves representing inhomogeneous cosmological models .
despite the fact that cylindrically symmetric waves can not describe exactly the radiation from bounded sources , they even recently played an important role in clarifying a number of complicated issues , such as testing the quasilocal mass - energy , testing codes in numerical relativity , investigation of the cosmic censorship , and quantum gravity . in work with ashtekar and schmidt @xcite , we considered gravitational waves with a space - translation killing field ( `` generalized einstein - rosen waves '' ) . in
( 2 + 1)-dimensional framework the einstein - rosen subclass forms a simple instructive example of explicitly given spacetimes which admit a smooth global null ( and timelike ) infinity even for strong initial data .
4-dimensional vacuum gravity which admits a spacelike hypersurface killing vector @xmath34 is equivalent to 3-dimensional gravity coupled to a scalar field . in 3 dimensions , there is no gravitational radiation .
hence , the local degrees of freedom are all contained in the scalar field .
one therefore expects that cauchy data for the scalar field will suffice to determine the solution . for data which fall off appropriately
, we thus expect the 3-dimensional lorentzian geometry to be asymptotically flat in the sense of penrose , i.e. that there should exist a 2-dimensional boundary representing null infinity . in general cases ,
this is analyzed in @xcite . restricting ourselves to the einstein - rosen waves by assuming that there is a further spacelike , hypersurface orthogonal killing vector @xmath35 which commutes with @xmath36
, we find the 3-metric given by @xmath37 the field equations become @xmath38 thus , we can first solve the axisymmetric wave equation for @xmath39 on minkowski space and then solve for @xmath40 the only unknown metric coefficient by quadratures . by analyzing the asymptotic behavior of the solutions we can conclude that _ cylindrical waves in ( 2 + 1)-dimensions give an explicit model of the bondi - penrose radiation theory which admits smooth null and timelike infinity for arbitrarily strong initial data_. there is no other such model available .
the general results on the existence of @xmath15 in 4 dimensions assume weak data .
these spacetimes have attracted increased attention in the last decade most notably in the work by chruciel , and chruciel and singleton @xcite . in these studies
the robinson - trautman spacetimes have been shown to exist globally for all positive `` times '' , and to converge asymptotically to a schwarzschild metric .
interestingly , the extension of these spacetimes across the `` schwarzschild - like '' event horizon can only be made with a finite degree of smoothness .
these studies are based on the derivation and analysis of an asymptotic expansion describing the long - time behaviour of the solutions of the nonlinear parabolic robinson - trautman equation . in our work
@xcite we studied robinson - trautman spacetimes with a positive cosmological constant @xmath41 .
the results proving the global existence and convergence of the solutions of the robinson - trautman equation can be taken over from the previous studies since @xmath41 does not explicitly enter this equation .
we have shown that , starting with arbitrary , smooth initial data at @xmath42 , these cosmological robinson - trautman solutions converge exponentially fast to a schwarzschild - de sitter solution at large retarded times ( @xmath43 ) .
the interior of a schwarzschild - de sitter black hole can be joined to an `` external '' cosmological robinson - trautman spacetime across the horizon @xmath44 with a higher degree of smoothness than in the corresponding case with @xmath45 . in particular , in the extreme case with @xmath46 , in which the black hole and cosmological horizons coincide , the robinson - trautman spacetimes can be extended smoothly through @xmath44 to the extreme schwarzschild - de sitter spacetime with the same values of @xmath41 and @xmath47 .
however , such an extension is not analytic ( and not unique ) .
we have also demonstrated that the cosmological robinson - trautman solutions represent explicit models exhibiting the cosmic no - hair conjecture .
as far as we are aware , these models represent the only exact analytic demonstration of the cosmic no - hair conjecture under the presence of gravitational waves .
they also appear to be the only exact examples of black hole formation in nonspherical spacetimes which are not asymptotically flat . since diverging , non - twisting robinson - trautman spacetimes of type n have singularities , there has been hope that if one admits a nonvanishing twist a more realistic radiative spacetime may exist .
stephani @xcite , however , indicated , by constructing a general solution of the linearized equations , that singularities at infinity probably exist .
more recently , finley et al @xcite found an approximative twisting type n solution up to the third order of iteration on the basis of which they suggested that it seems that the twisting , type n fields can describe a radiation field outside bounded sources .
however , employing the newman - penrose formalism and maple we succeeded in discovering a nonvanishing quartic invariant in the 2nd derivatives of the riemann tensor @xcite , which shows that solutions of both stephani and finley et al contain singularities at large @xmath4 .
mac alevey @xcite argued that an approximate solution at any finite order can be calculated without occurrence of singularities .
it is very likely , however , that a corresponding exact solution must contain singularities since mason @xcite proved that the only vacuum algebraically special spacetime that is asymptotically simple is the minkowski space . even if a radiative solution with complete smooth null infinity may be out of reach , it is of interest to construct radiative solutions which at least admit a global null infinity in the sense that its smooth cross sections exist although this null infinity is not necessarily complete . the only explicit examples of such solutions are spacetimes with boost - rotation symmetry .
i reviewed these spacetimes representing `` uniformly accelerated objects '' ( see fig .
2 ) in various places ( see e.g. @xcite and references therein ) ; here i shall just mention some new results . the penrose diagram of these spacetimes is schematically illustrated in fig .
the unique role of the boost - rotation symmetric spacetimes is exhibited by a theorem @xcite which roughly states that in axially symmetric , locally asymptotically flat electrovacuum spacetimes ( in the sense that a null infinity satisfying penrose s requirements exists , but it need not necessarily exist globally ) , the only additional symmetry that does not exclude radiation is the _ boost _ symmetry .
where all funtions describing metric and electromagnetic field tensor @xmath49 are independent of @xmath50 . assuming asymptotic expansions of these functions at large @xmath4 with @xmath3 , @xmath51 , @xmath50 fixed to guarantee asymptotic flatness , and using the outgoing radiation condition and the field equations , one finds these expansions to have specific forms .
for example , @xmath52 @xmath53 where the ` coefficients ' @xmath54 , @xmath55 , ... are functions of @xmath3 and @xmath51 .
the expansions are needed to further orders
see @xcite for their forms .
let us only recall that the decrease of the bondi mass , @xmath56 is given by @xmath57 where @xmath58 , @xmath59 , @xmath60 , @xmath61 are gravitational and electromagnetic news functions .
now one writes down the killing equations and solves them asymptotically in @xmath62 .
one arrives at the following theorem @xcite : suppose that an axially symmetric electrovacuum spacetime admits a `` piece '' of @xmath63 in the sense that the bondi - sachs coordinates can be introduced in which the metric takes the form ( [ ds ] ) , with the asymptotic form of the metric and electromagnetic field given by ( [ jednarovnice ] ) .
if this spacetime admits an additional killing vector forming with the axial killing vector a two - dimensional lie algebra then the additional killing vector has asymptotically the form @xmath64\ , \label{bbotr}\ ] ] where @xmath65 is a constant .
for @xmath66 it generates asymptotically translations ( function @xmath67 has then a specific form ) . for @xmath68
it is the boost killing field .
the case of translations is analyzed in detail in @xcite .
theorem 1 , precisely formulated and proved there , states that if asymptotically translational killing vector is spacelike , then null infinity is singular at some @xmath69 ; if it is null , null infinity is singular at @xmath70 or @xmath71 .
the first case corresponds to cylindrical waves , the second case to a plane wave propagating along the symmetry axis .
we refer to @xcite for the case when there is also a cosmic string present along the symmetry axis .
the case of timelike killing vector is described by theorem 2 ( proved also in @xcite ) : `` if an axisymmetric electrovacuum spacetime with a non - vanishing bondi mass admits an asymptotically translational killing vector and a complete cross section of @xmath63 , then the translational killing vector is timelike and spacetime is thus stationary . ''
the case of the boost killing vector @xmath72 is thoroughly analyzed in @xcite .
the general functional forms of the news functions ( both gravitational and electromagnetic ) , and of the mass aspect and total bondi mass of boost - rotation symmetric spacetimes are there given .
recently these results were obtained @xcite by using the newman - penrose formalism and under more general assumptions ( for example , @xmath15 could in principle be polyhomogeneous ) .
the general structure of the boost - rotation symmetric spacetimes with hypersurface orthogonal killing vector was analyzed in detail in @xcite .
their radiative properties , including explicit construction of radiation patterns and of bondi mass for the specific boost - rotation symmetric solutions were investigated in several works we refer to the reviews @xcite and @xcite for details .
there also the role of the boost - rotation symmetric spacetimes in such diverse fields like numerical relativity and quantum production of black - hole pairs is noticed and references are given . here
i would like to mention yet a recent progress in understanding specific boost - rotation symmetric spacetimes with killing vectors which are _ not _ hypersurface orthogonal .
this is the _ spinning _
@xmath73-metric ( see e.g. @xcite ) .
it was discovered by plebaski and demaiski as a generalization of the standard @xmath73-metric which is known to represent uniformly accelerated non - rotating black holes . in @xcite we first transformed the metric into weyl coordinates , and then found a transformation which brings it into the canonical form of the radiative spacetimes with the boost - rotation symmetry : where functions @xmath75 and @xmath76 are given in terms of @xmath77 in a somewhat complicated but explicit manner .
this metric can represent two uniformly accelerated , spinning black holes , either connected by a conical singularity , or with conical singularities extending from each of them to infinity .
the behaviour of curvature invariants clearly indicates the presence of a non - vanishing radiation field ( see figure 5 in @xcite ) .
the spinning @xmath73-metric is the only explicitly known example with two killing vectors which are not hypersurface orthogonal , in which one can give arbitrarily strong initial data on hyperboloid above the roof " ( @xmath78 ) which evolve into the radiative spacetime with smooth @xmath63 .
friedrich , h. ( 1998 ) _
einstein s equation and geometric asymptotics _ , in _ gravitation and relativity : at the turn of the millenium _ ( proceedings of the gr-15 conference ) , eds .
n. dadhich and j. narlikar , inter - university centre for astronomy and astrophysics press , pune chruciel , p. , maccallum , m. a. h. and singleton , p. b. ( 1995 ) _ gravitational waves in general relativity xiv .
bondi expansions and the ` polyhomogeneity ' of @xmath79 _ , phil . trans .
a350 * , 113 bik , j. ( 2000 ) _ selected solutions of einstein s field equations : their role in general relativity and astrophysics _ , in _ einstein s field equations and their physical meaning _ , ed .
b. g. schmidt , springer verlag chruciel , p. t. ( 1992 ) _ on the global structure of robinson - trautman space - times _ , proc .
a * 436 * , 299 ; chruciel , p. t. , singleton , d. b. ( 1992 ) _ non - smoothness of event horizons of robinson - trautman black holes _ ,
phys . * 147 * , 137 , and references therein bik , j. ( 1997 ) _ radiative spacetimes : exact approaches _ , in _ relativistic gravitation and gravitational radiation _ ( proceedings of the les houches school of physics ) , eds .
j .- a . marck and j .- p .
lasota , cambridge university press , cambridge | the question of existence of general , asymptotically flat radiative spacetimes and examples of explicit classes of radiative solutions of einstein s field equations are discussed in the light of some new developments .
the examples are cylindrical waves , robinson - trautman and type n spacetimes and especially boost - rotation symmetric spacetimes representing uniformly accelerated particles or black holes . |
neutrinos of different flavor suffer different refraction in matter @xcite .
the energy shift between @xmath3 and @xmath0 or @xmath1 is @xmath5 with @xmath6 the fermi constant , @xmath7 the baryon density , and @xmath8 the electron fraction .
@xmath9 is caused by the charged - current @xmath3-electron interaction that is absent for @xmath0 and @xmath1 . for a matter density @xmath10
g @xmath11 we have @xmath12 ev , yet this small energy shift is large enough to be of almost universal importance for neutrino oscillation physics . in normal matter , @xmath13 and @xmath14 leptons appear only as virtual states in radiative corrections to neutral - current @xmath0 and @xmath1 scattering , causing a shift @xmath15 between @xmath0 and @xmath1 .
it has the same effect on neutrino dispersion as real @xmath14 leptons with an abundance @xcite @xmath16 = 2.7\times10^{-5}\ , , \label{eq : ytau}\ ] ] where @xmath17 was assumed . for the neutron abundance we have used @xmath18 ,
but it provides only a 2.5% correction so that its exact value is irrelevant .
a large nonstandard contribution to @xmath19 can arise from radiative corrections in supersymmetric models @xcite , but we will here focus on the standard - model effect alone .
this `` mu tau matter effect '' modifies oscillations if @xmath20 . for propagation through the earth and for @xmath21@xmath22 ,
this occurs for neutrino energies @xmath23 tev .
the oscillation length then far exceeds @xmath24 so that @xmath25 is irrelevant for the high - energy neutrinos that are searched for by neutrino telescopes . alternatively ,
tau matter effect can be important at the large densities encountered by neutrinos streaming off a supernova ( sn ) core @xcite . for @xmath26
the condition @xmath27 implies @xmath28 .
numerical sn density profiles @xcite reveal that this occurs far beyond the shock - wave radius during the accretion phase , but retracts close to the neutrino sphere after the explosion has begun . to illustrate this point we show in fig .
[ fig : profiles ] the same matter density profiles as in ref .
@xcite at 1 ms post bounce ( red line ) and at 1 s post bounce ( blue line ) . as a green horizontal band
we indicate the condition @xmath29 for a typical range of sn neutrino energies , whereas the yellow and light - blue bands indicate the densities corresponding to the h - resonance ( driven by @xmath30 ) and the l - resonance ( driven by @xmath31 ) .
the @xmath0 , @xmath1 , @xmath32 and @xmath33 fluxes from a sn are virtually identical , leaving the @xmath34-resonance moot , whereas the h- and l - resonances cause well - understood consequences that are completely described by the energy - dependent swapping probabilities for @xmath3 and @xmath4 with some combination @xmath35 of the @xmath13 and @xmath14 flavor @xcite .
therefore , the traditional view has been that genuine three - flavor effects play no role for sn neutrino oscillations unless mu and tau neutrinos are produced with different fluxes @xcite . in a recent series of papers @xcite it was recognized , however , that the traditional picture was not complete : neutrino - neutrino interactions cause large collective flavor transformations in the sn region out to a few 100 km ( gray shaded region in fig .
[ fig : profiles ] ) . with the exception of refs .
@xcite , only two - flavor conversions driven by @xmath36 and the small @xmath37 have thus far been studied .
we here extend our previous numerical solutions @xcite to the case of three neutrino flavors .
our main results can be summarized as follows : ( i ) a two - flavor treatment indeed captures the full effect if one ignores @xmath25 and if the ordinary msw resonances occur outside of the collective neutrino region .
( ii ) including @xmath25 strongly modifies the @xmath3 or @xmath4 survival probabilities , influencing the neutrino signal from the next galactic sn .
( iii ) the effect depends sensitively on a possible deviation from maximal @xmath38 .
the purpose of our paper is to provide a first illustration of these findings that no doubt need to be refined in future .
our work is organized as follows . in sec .
[ sec : eoms ] we present the equations of motion which are solved for the three - flavor neutrino fluxes in a simplified scenario for the sn environment . in sec .
[ sec : smallmutau ] we consider the limit of a vanishing @xmath34 matter effect , while our results when it is significant are described in sec .
[ sec : largemutau ] .
we conclude in sec .
[ sec : conclusion ] . at 1
ms and 1 s post bounce of the numerical sn models described in ref .
@xcite ( solid lines ) .
the dashed lines represent the simplified matter profile of eq .
( [ eq : profile ] ) for @xmath39 km@xmath40 and @xmath41 km@xmath40 , used in our numerical calculations in figure [ fig : evolution ] . as horizontal bands
we indicate the conditions @xmath27 , @xmath42 , and @xmath43 for a typical range of sn neutrino energies .
the gray shaded range of radii corresponds to the region of collective neutrino transformations . within the radius @xmath44 the collective oscillations
are of the synchronized type.[fig : profiles ] ]
mixed neutrinos are described by matrices of density @xmath45 and @xmath46 for each ( anti)neutrino mode .
the diagonal entries are the usual occupation numbers whereas the off - diagonal terms encode phase information .
the equations of motion ( eoms ) are @xmath47\ , , \label{eq : eoms}\ ] ] where the hamiltonian is @xcite @xmath48 @xmath49 being the velocity .
the matrix of vacuum oscillation frequencies is @xmath50 in the mass basis .
the matter effect is represented , in the weak interaction basis , by @xmath51 .
for antineutrinos the only difference is @xmath52 . in spherical symmetry
the eoms can be expressed as a closed set of differential equations along the radial direction @xcite .
we solve them numerically as previously described @xcite , now using @xmath53 matrices instead of polarization vectors .
the factor @xmath54 in the hamiltonian implies `` multi - angle effects '' for neutrinos moving on different trajectories @xcite .
however , for realistic sn conditions the modifications are small , allowing for a single - angle approximation .
we implement this approximation by launching all neutrinos with @xmath55 relative to the radial direction @xcite . as a further simplification we use a monochromatic spectrum ( @xmath26 ) , ignoring the `` spectral splits '' caused by collective oscillation effects @xcite .
oscillation effects require flavor - dependent flux differences .
one expects @xmath56 .
the equal parts of the fluxes drop out of the eoms , so as initial condition we use @xmath57 and @xmath58 with @xmath59 . for the neutrino parameters we use @xmath60 , @xmath61 , @xmath62 , @xmath63 , and a vanishing dirac phase @xmath64 ,
all consistent with measurements @xcite .
we consider the entire allowed range @xmath65 because our results depend sensitively on @xmath38 .
we use a fixed matter profile of the form @xmath66 , implying a radial variation of the weak potential of @xmath67 where @xmath68 km is our nominal neutrino - sphere radius and @xmath69 . in fig .
[ fig : profiles ] we show this profile ( dashed lines ) for two different values of @xmath70 km@xmath40 and @xmath41 km@xmath40 . for the former case ,
the h - resonance is at @xmath71 km , the l - resonance at @xmath72 km , and the @xmath34-resonance at @xmath73 km .
for the latter they are at @xmath74 km , @xmath75 km , and @xmath76 km . as the @xmath34 resonance ,
although this would be correct only for a small vacuum mixing angle in the 23-subsystem . ]
the strength of the neutrino - neutrino interaction can be parametrized by @xmath77 where the fluxes are taken at the neutrino - sphere radius @xmath78 . as in our previous work
@xcite we shall assume @xmath79 km@xmath40 .
in the single - angle approximation where all neutrinos are launched with @xmath55 relative to the radial direction @xcite , the radial dependence of the neutrino - neutrino interaction strength can be explicitly written as @xmath80 while the @xmath81 scaling of @xmath82 for @xmath83 is generic , the overall strength @xmath84 depends on the neutrino fluxes and on their angular divergence , i.e. , on the true radius of the neutrino sphere .
our @xmath68 km is not meant to represent the physical neutrino sphere , it is only a nominal radius where we fix the inner boundary condition for our calculation .
the collective neutrino oscillations are of the synchronized type within the `` synchronization radius . '' for our chosen @xmath84 and for the assumed excess @xmath3 flux of 25% we find @xmath85 km as indicated in fig . [
fig : profiles ] .
collective flavor transformations occur at @xmath86 .
therefore , the @xmath34 matter effect can be important only if it is sufficiently large for @xmath86 .
figure [ fig : profiles ] illustrates that the region where the @xmath34-resonance takes place depends on the time after bounce . for realistic values of the matter density profile and neutrino - neutrino interaction ,
one expects @xmath87 to lie far beyond the collective region at early times .
this can be inferred from the relative position of @xmath44 and the intersection of the 1 ms profile and the green band . at later times though the proto neutron star contracts and @xmath87 moves to smaller radii .
eventually @xmath87 becomes smaller than @xmath44 , at which point @xmath25 becomes irrelevant . in order to mimic these different situations we will us a simple power - law matter profile of the form in eq .
( [ eq : profile ] ) . in other words
, we will use a mu - tau matter potential of the form @xmath88 with a fixed @xmath89 given by eq .
( [ eq : ytau ] ) and a variable coefficient @xmath90
. therefore early and late times can be reproduced by considering large and small values of @xmath90 , respectively , as can be seen in fig .
[ fig : profiles ] . in other words
, we will always assume that the ordinary msw resonances are far outside of the collective neutrino region , whereas the @xmath34 resonance can lie at smaller ( vanishing @xmath34 matter effect ) or larger ( large @xmath34 matter effect ) radii than @xmath44 .
as a first case we consider the traditional assumption of a vanishing @xmath34 matter effect , which we account for using a value of @xmath91 .
we assume an inverted @xmath36 and use a non - maximal value @xmath92 .
our numerical calculations for this case are shown in the top row of fig .
[ fig : evolution ] .
the first two panels correspond to the radial evolution of the fluxes of the weak interaction eigenstates of neutrinos and antineutrinos , respectively , whereas in the last two panels we show the evolution of the propagation eigenstates .
these are the eigenstates of @xmath93 , i.e. , of that part of the hamiltonian eq .
( [ eq : hamiltonian ] ) that does not include the neutrino - neutrino interactions . in the collective neutrino region , we observe the usual pair conversion of the @xmath3 and @xmath4 fluxes into the @xmath13 and @xmath14 flavors .
had we chosen a maximal @xmath94 mixing angle , the appearance curves for these flavors would be identical . for larger distances
the evolution consists of ordinary msw transformations that are best pictured in the basis of instantaneous propagation eigenstates in matter ( last two panels ) . beyond the collective transformation region ,
all neutrinos and antineutrinos stay fixed in their propagation eigenstates . in the weak - interaction basis , on the other hand , this implies fast oscillations because we have a fixed energy , preventing kinematical decoherence between different energy modes . in the panels for neutrino and antineutrino interaction states , for radii beyond the dense - neutrino region we show as thick lines the average evolution as well as the envelopes of the fast - oscillating flavor fluxes .
another way of describing this evolution is by the level crossing schemes of fig .
[ fig : crossing ] .
the upper panel represents the case of vanishing @xmath25 , corresponding to fig .
5d of ref .
the central panel represents the case with large @xmath25 and a 23-mixing angle in the first octant and is similar to fig . 2 of ref .
@xcite . in such plots
one shows the neutrino energy levels as a function of the matter density .
the continuation of this diagram to negative densities gives us the energy levels of antineutrinos : the neutrino energy at a negative density really means the antineutrino energy at the corresponding positive density . for vanishing density ( vacuum ) , we have the three vacuum mass eigenstates that are identical for neutrinos and antineutrinos .
the upper ( blue ) line corresponds to propagation eigenstate 2 , the middle ( green ) line to 1 , and the bottom ( red ) line to 3 , a scheme representing the inverted hierarchy case .
these lines represent the propagation eigenstates that are adiabatically connected for different densities . while in vacuum
the propagation eigenstates coincide with the mass eigenstates , at large densities they correspond to weak interaction eigenstates .
for vanishing @xmath25 and at the low energies relevant to our problem , the @xmath13 and @xmath14 flavor are not distinguishable so that any convenient linear combination can be chosen as interaction eigenstates .
it is convenient to introduce the states @xmath95 and @xmath96 that correspond to a vanishing 23-mixing angle , i.e. , they diagonalize the 23-subsystem . if the small 13-mixing angle were to vanish , the 3-mass eigenstate would coincide with @xmath96 . in the upper panel of fig .
[ fig : crossing ] and using the @xmath97 basis , the 2-state connects adiabatically to @xmath3 and @xmath98 , whereas the 3-state connects adiabatically to @xmath4 and @xmath96 .
( upper panel ) and a large @xmath25 with 23-mixing in the first octant ( central panel ) or the second octant ( lower panel ) .
the arrows indicate the transitions caused by collective flavor transformations .
[ fig : crossing],title="fig : " ] ( upper panel ) and a large @xmath25 with 23-mixing in the first octant ( central panel ) or the second octant ( lower panel ) .
the arrows indicate the transitions caused by collective flavor transformations .
[ fig : crossing],title="fig : " ] ( upper panel ) and a large @xmath25 with 23-mixing in the first octant ( central panel ) or the second octant ( lower panel ) .
the arrows indicate the transitions caused by collective flavor transformations . [
fig : crossing],title="fig : " ] at the neutrino sphere , the fluxes are prepared in @xmath3 and @xmath4 eigenstates , which in the case of inverted mass hierarchy coincide with the propagation ( or matter ) eigenstates @xmath99 and @xmath100 , respectively . in the absence of neutrino - neutrino interactions , since the l - resonance is always adiabatic , the @xmath3 s leave the star as @xmath101 . in the case of
@xmath4 the evolution depends on @xmath102 @xcite . for values larger than @xmath103
they propagate also adiabatically ( msw transformation ) and escape as @xmath104 , whereas for values smaller than @xmath105 the transition at the h - resonance is strongly non - adiabatic : there is a jump of matter eigenstates from @xmath100 to @xmath106 and the @xmath4 s leave the star as @xmath107 . as a consequence ,
the survival probability is @xmath108 and @xmath109 or @xmath110 for large and small @xmath37 , respectively . in the presence of neutrino - neutrino interactions ,
important collective effects take place in the inner sn layers , where the neutrino density is high .
we observe in the first two panels of fig .
[ fig : evolution ] that collective pair transformations convert the @xmath3 and @xmath4 fluxes to @xmath96 and @xmath111 as indicated by the arrows in the upper panel of fig .
[ fig : crossing ] .
the consequences for the subsequent evolution are dramatic . in the case of @xmath3
a fraction equal to @xmath112 stays in @xmath113 and evolves as in the absence of neutrino - neutrino interactions , while the rest of @xmath3 are transformed to @xmath114 . as a consequence
, the final @xmath3 flux , normalized to the initial @xmath4 one , is expected to be approximately @xmath115 , see thick line in the upper left panel in fig .
[ fig : evolution ] . in the case of antineutrinos
the effect of the collective pair conversion is to interchange the eigenstates in which @xmath4 and @xmath111 arrive at the h - resonance .
now @xmath4 enters the resonance as @xmath116 instead of @xmath117 .
therefore , for @xmath118 the resonance is adiabatic and the @xmath4 s leave the star as @xmath107 , leading to a final normalized flux of approximately @xmath119 , see the thick line in the second panel in fig .
[ fig : evolution ] . instead , if @xmath120 again there is a jump of matter eigenstates from @xmath106 to @xmath100 at the h - resonance .
in this case @xmath4 leaves the star as @xmath104 , leading to a normalized @xmath4 flux equal to @xmath102 .
the impact of collective effects is easier to understand if we follow the previous literature @xcite and observe that , in a two - flavor system , the impact of ordinary matter can be transformed away by going into a rotating reference frame for the polarization vectors .
collective conversions proceed in the same way as they would in vacuum , except that the effective mixing angle is reduced . therefore ,
assuming an inverted hierarchy ( ih ) for the atmospheric mass splitting and a normal hierarchy ( nh ) for the solar splitting , we should consider the level scheme as in the upper left panel of fig .
[ fig : levels ] .
the mass eigenstates now approximately coincide with the interaction eigenstates because the 23-mixing angle was removed by going to the primed states , and the mixing angles involving @xmath3 are effectively made small by the presence of matter .
of course , this level scheme does not adiabatically connect to the true vacuum situation .
the initial state consists of @xmath3 and @xmath4 and thus essentially of @xmath121 and @xmath107 .
collective conversions driven by @xmath36 then transform @xmath122 pairs to @xmath123 pairs in the familiar two - flavor way . if both hierarchies are normal , we begin in the lowest - lying state and nothing happens . ] in the hypothetical case where both hierarchies are inverted ( upper right panel in fig .
[ fig : levels ] ) , we begin in the highest state and @xmath36 drives us directly to the lowest state . finally , if the atmospheric hierarchy is normal and the solar one is inverted ( lower right panel in fig .
[ fig : levels ] ) , collective transformations driven by @xmath124 take us to the lowest state .
we have numerically solved the evolution of the three - flavor system with a realistic sn matter profile and found that the results confirm this simple picture . in a two - flavor treatment ,
the much smaller @xmath124 leads to collective transformations at a much larger radius than @xmath36 . in a three - flavor treatment
, @xmath36 therefore acts first and takes us directly to the lowest - lying state if the atmospheric hierarchy is inverted
. otherwise only the hypothetical case of the lower - right panel in fig .
[ fig : levels ] is an example where @xmath125 plays any role .
we have numerically verified that normal @xmath30 combined with inverted @xmath31 is the only case where @xmath31 drives collective transformations . since @xmath31 is measured to be normal , the previous two - flavor treatments based on @xmath126 and @xmath37 fortuitously capture the full effect .
we conclude that in the limit of a vanishing @xmath34 effect the collective flavor transformations and the subsequent msw evolution factorize and that the collective effects are correctly treated in a two - flavor picture . of course , this situation may change if the matter profile is so shallow that the ordinary msw effects occur in the same region as the collective phenomena @xcite .
next we calculate the flavor evolution for the same model , now including a significant @xmath25 , i.e. we assume a large @xmath90 . in this case
the flavor content of the neutrino and antineutrino fluxes emerging from the sn surface depend on the strength of @xmath25 as well as the choice of @xmath38 , as can be seen in the corresponding panels of fig .
[ fig : evolution ] .
this dependence is best illustrated with the help of the contour plot fig .
[ fig : contours ] where we show the @xmath3 and @xmath4 fluxes emerging from the sn , averaged over fast vacuum oscillations . and
@xmath90 for the @xmath3 ( top ) and @xmath4 ( bottom ) fluxes emerging from the sn surface for both normal ( right ) and inverted ( left ) mass hierarchy .
all fluxes are normalized to the initial @xmath127 flux .
we show values averaged over fast vacuum oscillations .
[ fig : contours ] ] if @xmath25 is so large that the mu - tau effect is strong in the region of collective neutrino oscillations , there are two stable limiting cases , depending on the 23 mixing angle . if the mixing angle is sufficiently non - maximal and in the first octant , the collective oscillations transform the initially prepared @xmath3 and @xmath4 fluxes to the propagation eigenstates as indicated by the arrows in the middle panel of fig .
[ fig : crossing ] , i.e. , we observe pair transformations to @xmath128 .
this behavior is understood if we assume that in the @xmath34 system we can once more go to a rotating frame and now simply imagine that the 23 mixing angle is effectively small by the impact of the @xmath34 matter effect . in this case @xmath129 . since collective quasi - vacuum oscillations
take us to the lowest - lying state , the @xmath130 state in the inverted hierarchy , we are effectively taken to @xmath128 pairs . instead , if the 23 mixing angle is in the second octant , @xmath0 and @xmath1 switch roles , explaining that now @xmath131 and @xmath132 .
these are only heuristic explanations .
we expect that they can be made precise in a true analytic three - flavor treatment of collective neutrino oscillations along the lines of ref .
@xcite . for intermediate values of @xmath25 and for 23
mixing angles near maximal , the final fluxes depend sensitively on parameters . for intermediate values of @xmath25 , there are also nontrivial effects for the normal hierarchy .
the collective effects do not place the ensemble into propagation eigenstates , preventing a simple interpretation . the sensitive dependence for intermediate @xmath25
is also illustrated in fig .
[ fig : cuts ] where we show the emerging average @xmath3 and @xmath4 fluxes as functions of @xmath90 for two values of @xmath38 , one in the first and the other in the second octant . in fig .
[ fig : cuts2 ] we show the same @xmath3 and @xmath4 fluxes as functions of @xmath133 for @xmath134 km@xmath40 .
one can notice how the fall of @xmath135 is not exactly centered at @xmath136 but slightly shifted to smaller values .
this is due to second - order corrections to the @xmath34 resonance condition .
( top ) and @xmath4 ( bottom ) , normalized to the initial @xmath127 flux , emerging from the sn as a function of @xmath90 for a 23 mixing angle in the first ( red line ) or second ( blue line ) octant .
these curves represent cuts through the inverted hierarchy contour plots of fig .
[ fig : contours ] at the indicated values of @xmath133.[fig : cuts],title="fig : " ] ( top ) and @xmath4 ( bottom ) , normalized to the initial @xmath127 flux , emerging from the sn as a function of @xmath90 for a 23 mixing angle in the first ( red line ) or second ( blue line ) octant .
these curves represent cuts through the inverted hierarchy contour plots of fig .
[ fig : contours ] at the indicated values of @xmath133.[fig : cuts],title="fig : " ] ( top ) and @xmath4 ( bottom ) , normalized to the initial @xmath127 flux , emerging from the sn as a function of @xmath133 for @xmath134 km@xmath40 .
these curves represent cuts through the inverted hierarchy contour plots of fig .
[ fig : contours ] at the indicated value of @xmath90.[fig : cuts2],title="fig : " ] ( top ) and @xmath4 ( bottom ) , normalized to the initial @xmath127 flux , emerging from the sn as a function of @xmath133 for @xmath134 km@xmath40 .
these curves represent cuts through the inverted hierarchy contour plots of fig .
[ fig : contours ] at the indicated value of @xmath90.[fig : cuts2],title="fig : " ] this dependence on the @xmath38 octant leads to a clear imprint on the final survival probability .
let us first consider the first octant . in the case of @xmath3
a fraction equal to @xmath137 stays in @xmath113 .
however the presence of the @xmath34-resonance in the neutrino channel makes the rest of the @xmath3 to be transformed to @xmath138 .
their subsequent evolution would depend on the adiabaticity of the @xmath34-resonance , but it has been shown to be always adiabatic @xcite . as a consequence , the final @xmath3 flux is expected to be approximately @xmath139 , see thick line in the left panel of the second row in fig .
[ fig : evolution ] . in the case of antineutrinos
the situation is completely analogous to the case of vanishing @xmath25 so that @xmath140 or @xmath102 , depending on the value of @xmath37 .
if @xmath38 belongs to the second octant , then the @xmath34-resonance lies in the antineutrino channel .
the crucial point is that now all @xmath4 are transformed to @xmath141 before reaching the @xmath34-resonance , see the lower panel in fig .
[ fig : crossing ] . taking into account that @xmath142 does not encounter the h - resonance , the survival probability will be always @xmath143 , independently of the value of @xmath37 . on the other hand neutrinos
do not feel the @xmath34-resonance and therefore their propagation is the same as in the vanishing @xmath25 case .
we present in table [ tab : barnuep ] a summary of the cases discussed so far .
one can see the importance of the presence of collective neutrino effects , as well as the dependence on the strength of the mu - tau matter effect .
.summary of the approximate values of the @xmath4 survival probability for an inverted hierarchy , including or not collective effects . here
a small ( large ) mixing angle @xmath37 stands for @xmath144 ( @xmath118 ) , while a small ( large ) @xmath145 represents @xmath87 being smaller ( larger ) than @xmath44 .
[ tab : barnuep ] [ cols="^,^,^,^,^,^ " , ] another interesting feature concerns the position of @xmath44 in the presence of a large @xmath34 matter effect . as can be seen comparing the first two rows of fig .
[ fig : evolution ] , the radius where collective neutrino transformations begin is slightly larger ( @xmath146 km ) when we include a significant @xmath25 .
we have checked that , while for a small @xmath25 the position of @xmath44 is independent of @xmath38 , for a large @xmath25 the onset of bipolar transformations is delayed for nonzero values of @xmath38 .
this effect is largest for maximal mixing ( @xmath147 ) and symmetric relative to @xmath147 .
a full understanding of this variation presumably requires an analytic three - flavor treatment in the spirit of ref .
at the relatively low energies relevant for sn neutrinos , charged mu and tau leptons can not be produced so that mu- and tau - flavored neutrinos are not distinguishable in the sn or in detectors .
( in the inner core of a sn the temperatures may be high enough to produce a significant thermal muon density , but this would not affect the emission from the neutrino sphere . ) the impact of the small second - order difference between the @xmath0 and @xmath1 refractive index does not produce observable effects as long as one only considers the traditional msw flavor conversion @xcite .
the picture changes if one includes the unavoidable effect of collective neutrino transformations in the region above the neutrino sphere . if the matter density is large enough that @xmath145 is comparable to or larger than @xmath148 , the survival probability of @xmath3 and @xmath4 can be completely modified and depends sensitively on the mixing angle @xmath38 . in future one
should also include non - monochromatic energy spectra , leading to spectral split phenomena that could be more complicated than the previously studied two - flavor cases .
one should also explore the impact of realistic angular distributions and of a non - zero dirac phase in the neutrino mixing matrix .
lower - mass progenitors may collapse with a o - ne - mg core and , on the computer , explode easily because there is very little mass in the envelope @xcite . even at core
bounce and immediately afterward , the density profile is so shallow that the ordinary h- and l - resonances may occur within the collective neutrino region @xcite . in this case
the effects discussed here are irrelevant because the mu - tau matter effect is negligible .
probably our effects are also negligible during the cooling phase of an iron - core sn
. however , flavor oscillation effects are probably largest during the accretion phase of an iron - core sn where the flavor dependence of the spectra and fluxes is more pronounced than during the cooling phase @xcite . when it is important , the mu - tau matter effect adds one more layer of complication to the already vexed problem of collective sn neutrino oscillations .
it was previously recognized that `` ordinary '' collective oscillations are almost completely insensitive to the smallness of @xmath37 as long as it is not exactly zero .
here we have found the opposite for the large mixing angle @xmath38 that is often assumed to be maximal .
even small deviations from maximal 23-mixing can imprint themselves in the collective oscillation effect .
both results are counter - intuitive and opposite to ordinary flavor oscillations .
we thank e. akhmedov for an illuminating correspondence , a. mirizzi for comments on the manuscript , and b. dasgupta and a. dighe for helpful comments and for making their manuscript @xcite available before completion .
this work was partly supported by the deutsche forschungsgemeinschaft ( grant tr-27 `` neutrinos and beyond '' ) , by the cluster of excellence `` origin and structure of the universe '' ( garching and munich ) , by the european union ( contracts no .
rii3-ct-2004 - 506222 and mrtn - ct-2004 - 503369 ) , and by the spanish grants fpa2005 - 01269 ( mec ) and acomp07 - 270 ( generalitat valenciana ) .
ae was supported by an fpu grant from the spanish government .
sp and rt were supported by mec contracts ( _ ramn y cajal _ and _ juan de la cierva _ , respectively ) .
l. wolfenstein , phys .
d * 17 * , 2369 ( 1978 ) .
botella , c.s .
lim and w.j .
marciano , phys .
d * 35 * , 896 ( 1987 ) .
e. roulet , phys .
b * 356 * , 264 ( 1995 ) [ arxiv : hep - ph/9506221 ] .
akhmedov , c. lunardini and a.yu .
smirnov , nucl .
b * 643 * , 339 ( 2002 ) [ arxiv : hep - ph/0204091 ] .
a. arcones , h.t .
janka and l. scheck , astron .
astrophys . * 467 * , 1227 ( 2007 ) [ arxiv : astro - ph/0612582 ] .
dighe and a.yu .
smirnov , phys .
d * 62 * , 033007 ( 2000 ) [ arxiv : hep - ph/9907423 ] .
s. pastor and g.g .
raffelt , phys .
lett . *
89 * , 191101 ( 2002 ) [ arxiv : astro - ph/0207281 ] .
sawyer , arxiv : hep - ph/0408265 .
sawyer , phys .
d * 72 * , 045003 ( 2005 ) [ arxiv : hep - ph/0503013 ] . h. duan , g.m .
fuller and y.z .
qian , phys .
d * 74 * , 123004 ( 2006 ) [ arxiv : astro - ph/0511275 ]
. h. duan , g.m .
fuller , j. carlson and y.z .
qian , phys .
d * 74 * , 105014 ( 2006 ) [ arxiv : astro - ph/0606616 ] .
s. hannestad , g.g .
raffelt , g. sigl and y.y.y .
wong , phys .
d * 74 * , 105010 ( 2006 ) [ arxiv : astro - ph/0608695 ] .
raffelt and g. sigl , phys .
d * 75 * , 083002 ( 2007 ) [ arxiv : hep - ph/0701182 ] .
h. duan , g.m .
fuller , j. carlson and y.z .
qian , phys .
d * 75 * , 125005 ( 2007 ) [ arxiv : astro - ph/0703776 ] .
raffelt and a.yu .
smirnov , phys .
d * 76 * , 081301 ( 2007 ) [ arxiv:0705.1830 ] .
raffelt and a.yu .
smirnov , phys .
d * 76 * , 125008 ( 2007 ) [ arxiv:0709.4641 ] .
a. esteban - pretel , s. pastor , r. toms , g.g .
raffelt and g. sigl , phys .
d * 76 * , 125018 ( 2007 ) [ arxiv:0706.2498 ] .
h. duan , g.m .
fuller and y.z .
qian , phys .
d * 76 * , 085013 ( 2007 ) [ arxiv:0706.4293 ] .
h. duan , g.m .
fuller , j. carlson and y.z .
qian , phys .
lett . * 99 * , 241802 ( 2007 ) [ arxiv:0707.0290 ] .
fogli , e. lisi , a. marrone and a. mirizzi , j. cosmol .
astropart .
* 12 * , 010 ( 2007 ) [ arxiv:0707.1998 ] .
h. duan , g.m .
fuller , j. carlson and y.z .
qian , phys . rev .
lett . * 100 * , 021101 ( 2008 ) [ arxiv:0710.1271 ] . b. dasgupta and a. dighe , arxiv:0712.3798 [ hep - ph ] .
g. sigl and g.g .
raffelt , nucl .
b * 406 * , 423 ( 1993 ) .
m. maltoni , t. schwetz , m.a .
trtola and j.w.f .
valle , new j. phys .
* 6 * , 122 ( 2004 ) [ arxiv : hep - ph/0405172v6 ] .
fogli , e. lisi , a. marrone and a. palazzo , prog .
phys . * 57 * , 742 ( 2006 ) [ arxiv : hep - ph/0506083 ] .
gonzlez - garca and m. maltoni , physics reports , in press ( 2008 ) [ arxiv:0704.1800 ] .
kitaura , h .-
janka and w. hillebrandt , astron .
astrophys . * 450 * , 345 ( 2006 ) [ arxiv : astro - ph/0512065 ] .
keil , g.g .
raffelt and h.t .
janka , astrophys . j. * 590 * , 971 ( 2003 ) [ arxiv : astro - ph/0208035 ] . | we study three - flavor collective neutrino transformations in the dense - neutrino region above the neutrino sphere of a supernova core .
we find that two - flavor conversions driven by the atmospheric mass difference and the 13-mixing angle capture the full effect if one neglects the second - order difference between the @xmath0 and @xmath1 refractive index .
including this `` mu tau matter term '' provides a resonance at a density of @xmath2 that typically causes significant modifications of the overall @xmath3 and @xmath4 survival probabilities .
this effect is surprisingly sensitive to deviations from maximal 23-mixing , being different for each octant . |
suppose given a based map @xmath9 of simply connected cw complexes .
denote by @xmath10 the path component of the space of ( unbased ) maps @xmath11 consisting of those maps that are homotopic to @xmath4 .
then evaluation at the basepoint of @xmath5 gives a based map @xmath12 .
we refer to this map as the _ evaluation map_. we define the _ @xmath3th evaluation subgroup of @xmath4 _ to be the subgroup @xmath13 of @xmath2 .
the famous gottlieb groups @xmath14 occur as the special case in which @xmath15 and @xmath16 @xcite .
the gottlieb groups of a space have been much studied by homotopy theorists ( see @xcite for a survey of results and references ) .
while many general results are known , explicit computation of @xmath17 appears difficult and is limited to a small number of sporadic examples .
one reason that accounts in part for this difficulty is the fact that a map of spaces @xmath18 does not necessarily induce a corresponding homomorphism of gottlieb groups , since in general @xmath19 . in particular , attempts to study @xmath17 via a cell decomposition of @xmath5 are frustrated , since it is not clear what effect a cell attachment may have on the gottlieb groups .
one tool for studying @xmath17 that attempts to circumvent this problem is the so - called _ @xmath8-sequence of a map _ , introduced by woo and lee in @xcite .
the @xmath8-sequence of a map @xmath9 is a sequence @xmath20 of groups and homomorphisms that derives from the long exact homotopy sequence of the map @xmath4 . in this sequence
, @xmath21 is the gottlieb group of @xmath5 , @xmath22 is the evaluation subgroup of @xmath4 introduced above , and the third term @xmath23 is a suitably defined relative " term .
the @xmath8-sequence arises as follows .
one has the following commutative diagram of spaces : @xmath24^{f _ * } \ar[d]_{\omega } & \map(x , y ; f)\ar[d]^{\omega}\\ x \ar[r]_{f } & y}\ ] ] now pass to the corresponding induced homomorphisms of homotopy groups . in a standard way , the induced homomorphisms of homotopy groups @xmath25 and
@xmath26 can be fitted into the long exact homotopy sequences of the maps @xmath27 and @xmath4 respectively .
then the evaluation maps induce maps of each term in the long exact sequences , resulting in a homotopy ladder .
the @xmath8-sequence of the map @xmath4 is then the image of the top long exact homotopy sequence in that of the bottom .
a portion of the @xmath8-sequence is shown here : @xmath28 \ar@{->>}[d]^{\omega_{\ # } } & \pi_{n}(\map(x , x;1 ) ) \ar@{->>}[d]^{\omega_{\ # } } \ar[r]^{(f_*)_{\ # } } & \pi_{n}(\map(x , y;f ) ) \ar@{->>}[d]^{\omega_{\ # } } \\
g^{rel}_{n+1}(y , x;f ) \ar@{^{(}->}[d ] \ar[r ] & g_{n}(x ) \ar@{^{(}->}[d ] \ar[r]^{f_{\ # } } & g_{n}(y , x;f ) \ar@{^{(}->}[d ] \\
\pi_{n+1}(f ) \ar[r ] & \pi_{n}(x ) \ar[r]^{f_{\ # } } & \pi_{n}(y ) } \ ] ] the homomorphisms in the @xmath8-sequence are restrictions of those in the long exact homotopy sequence of the map @xmath4 . therefore , the @xmath8-sequence forms a chain complex ( consecutive compositions are trivial ) .
some general conditions are known under which the @xmath8-sequence is exact ( e.g. @xcite ) , but in general it is not exact ( e.g. @xcite ) . in this paper
we bring the techniques of rational homotopy theory to bear on problems and questions concerning evaluation subgroups ( of a map ) in general , and the @xmath8-sequence in particular .
our main goal is to expand the range of application of these techniques in this area . to this end
, we are primarily concerned with establishing a suitable framework for considering such questions . at the same time
, we obtain a number of results of interest in their own right .
the paper is organized as follows : our main results are established in sections @xmath29 and @xmath30 .
the basic result in section [ sec : deriv spaces ] is theorem [ thm : equivalence of squares ] , in which we identify the map induced on rational homotopy groups by the evaluation map @xmath31 .
we describe this induced homomorphism as the map induced on homology by a map of complexes of derivations of the sullivan minimal models of @xmath5 and @xmath32 .
theorem [ thm : equivalence of squares ] has a number of immediate corollaries .
for instance , we obtain a characterization of the rationalized evaluation subgroups of a map ( corollary [ cor : rational eval subgp of f ] ) that extends a well known characterization , due to flix and halperin , of the ( rationalized ) gottlieb groups of a space in terms of derivations of its minimal model . in section [ sec : long exact sequences ] , we extend and amplify the basic result of theorem [ thm : equivalence of squares ] .
we show that several related long exact sequences of rational homotopy groups are naturally expressed as long exact sequences in homology of derivation complexes of sullivan minimal models .
in particular , we obtain a description , within the framework of derivation spaces , of the @xmath8-sequence of a map after rationalization .
the results of this section are of interest in their own right and in several cases ( e.g. theorem [ thm : long exact seq of eval fib ] ) they are independent of the @xmath8-sequence . in section [ sec : examples ] , we use the framework established in sections @xmath29 and @xmath30 to study questions concerning the rationalized @xmath8-sequence . just as in the integral setting , the rationalized @xmath8-sequence is not exact in general .
example [ ex : non - exact g - seq ] gives a simple example for which the rationalized @xmath8-sequence fails to be exact at each of the three types of term that occur .
since non - exactness rationally implies non - exactness integrally , this example provides a new , complete example of the failure of exactness of the @xmath8-sequence .
this example also shows that the framework established in sections @xmath29 and @xmath30 provides an effective setting in which to carry out explicit computations . by way of contrast , in theorem [ thm : exact g - seq ] we give one set of conditions under which the rationalized @xmath8-sequence is exact . in theorem
[ thm : h^omega=0 ] we show that under certain circumstances , the rationalized @xmath8-sequence may be exact at all occurrences of one type of term , while failing to be exact at the other types of term .
this same result establishes a relationship , albeit under rather restrictive circumstances , between the ( vanishing of a certain type of the ) @xmath33-homology of a map @xmath34 and the ( vanishing of ) negative - degree derivations of the rational cohomology of @xmath5 .
the last development that we present in section [ sec : examples ] is a connection between the @xmath8-sequence of certain fibre inclusion maps and a well - known conjecture in rational homotopy theory .
let @xmath35 be a fibration with base an odd - dimensional sphere
. for certain types of fibre space @xmath5 , a conjecture of halperin asserts that the fibration should be rationally tncz ( see conjecture [ conj : halperin ] below ) . in theorem
[ thm : halperin g - trivial ] , we show that this is the case exactly when the rationalized @xmath8-sequence of the fibre inclusion @xmath36 reduces to a certain short exact sequence . in this way
, we obtain an equivalent phrasing of the conjecture of halperin , in terms of the ideas studied in this paper . finally ,
in a technical appendix , we give full details for several results from rational homotopy theory used in the proof of theorem [ thm : equivalence of squares ] .
we next discuss existing results on rational homotopy , function spaces , and the gottlieb groups of a space that relate to this paper .
our aim in this discussion is to indicate the basic results that exist in the area . at the same time
, we identify how these results relate to our work , and our points of departure from them . from results in rational homotopy theory
, we can identify three
tributary streams " that flow into this work precisely at theorem [ thm : equivalence of squares]our first main result at varying levels of generality .
there is also a fourth tributary flowing into our work , from outside rational homotopy theory , that includes the work on the @xmath8-sequence . the first tributary of antecedent results in rational homotopy theory concerns the rationalized gottlieb groups of a space .
in @xcite , flix and halperin gave a characterization of the rationalized gottlieb groups of a space , in terms of derivations of the sullivan minimal model . at its most specialized level , theorem [ thm : equivalence of squares ] retrieves this characterization ( corollary [ cor : rational gottlieb group ] ) and extends it to a similar characterization of the rationalized evaluation subgroups of a general map ( corollary [ cor : rational eval subgp of f ] ) .
flix and halperin went on to prove a remarkable result concerning the rationalized gottlieb groups of a finite complex ( * ? ? ?
* th.iii ) .
their result significantly extends results of gottlieb from @xcite , and relates the rationalized gottlieb groups with the rational lusternik - schnirelmann category .
unfortunately , no analogous result seems forthcoming for the rationalized evaluation subgroups of a map .
nonetheless , our characterization of the rationalized evaluation subgroups of a map is as effective for concrete computations as is the earlier characterization of the rationalized gottlieb groups .
the second tributary concerns the rational homotopy type of @xmath37the classifying space for fibrations with fibre @xmath5 .
although we are not concerned with this classifying space as such , the connection arises because we have an isomorphism of homotopy groups @xmath38 , and also the gottlieb groups of @xmath5 are obtained as the image in homotopy groups of the connecting homomorphism of the corresponding classifying fibration ( * ? ? ?
* th.2.6 ) . at a higher level of generality than the rationalized gottlieb groups ,
theorem [ thm : equivalence of squares ] includes an identification of the map induced on rational homotopy groups by the evaluation map @xmath39 , and in particular it identifies the rational homotopy groups of @xmath40 .
whilst the identification of these groups we give is familiar in rational homotopy theory , our proof for this obtained by restricting that of theorem [ thm : equivalence of squares]is the first direct and detailed one that has actually appeared in the literature .
we support this assertion as follows : in ( * ? ? ?
* sec.11 ) , sullivan sketched ( with no proof ) a model for the rational homotopy type of @xmath37 , from which the description of the rational homotopy groups of @xmath40 contained in theorem [ thm : equivalence of squares ] follows .
a justification of sullivan s model for @xmath37 may be gleaned by collating results from a number of sources spread through the literature ( e.g. @xcite ) . but to date , no direct proof of sullivan s model has been given .
even amongst those articles that focus specifically on the rational homotopy
_ groups _ , either of @xmath37 or of @xmath41and that therefore avoid the technical problems of dealing with the rational homotopy _
type_we still do not find complete details .
meier ( * ? ? ? *
( 1.4 ) , ( 2.6 ) ) outlines the basic idea , but is actually focussed on a special kind of situation in which the minimal model can be replaced by its cohomology .
grivel @xcite focusses on the same special case as meier , and quotes sullivan s model directly .
flix and thomas ( * ? ? ?
* sec.2.3 ) give exactly the description of the rational homotopy groups of @xmath40 contained in theorem [ thm : equivalence of squares ] , but no details of the proof are given .
as well as including a direct and detailed proof for the case of @xmath40 , theorem [ thm : equivalence of squares ] extends this identification of the rational homotopy groups to the general case of @xmath10 .
it thus provides a natural framework for the study of rational homotopy groups of function spaces . finally , the third tributary from rational homotopy theory consists of a model , due to sullivan and haefliger , for the rational homotopy type of @xmath10 . at its full level of generality ,
theorem [ thm : equivalence of squares ] identifies the map induced on rational homotopy groups by @xmath42 , a general evaluation map .
the precursor to the identification we give , and to our general line of proof , is the approach of thom in @xcite , although obviously there is no reference to minimal models in his work .
pursuing the river analogy a little further , we may think of thom s result , if not as the source , then at least as somewhere in the headwaters .
we emphasize the connection between theorem [ thm : equivalence of squares ] and thom s approach by retrieving a basic result of his in corollary [ cor : thom ] .
coming further downstream , sullivan also described in ( * ? ? ?
* sec.11 ) a model for the space of sections of a fibration homotopic to a given section . by specializing to the trivial fibration @xmath43 ,
this yields a model for the function space @xmath10and more generally a model for the rational homotopy _ type _ of the general evaluation map @xmath33 .
a detailed proof for sullivan s model in this case was given by haefliger @xcite .
now this model should in principle determine the rational homotopy groups of the function space ( see @xcite and @xcite , where it is used quite effectively ) . however , the model in question is a ( non - minimal ) dg algebra model .
therefore , the homomorphism induced by @xmath33 on rational homotopy groups which is exactly the information we require to proceed with our development is available only indirectly , at best . by focussing on the rational homotopy _
groups_as opposed to the rational homotopy type , we have arrived in theorem [ thm : equivalence of squares ] at an entirely new characterization of the map induced on rational homotopy groups by @xmath31 .
furthermore , we have been able to give a direct proof that avoids many of the technical complexities of haefliger s work and is completely independent of it .
the remaining tributary flowing into our work comes from outside rational homotopy theory , and concerns classical results on the gottlieb groups of a space , and more recent results on evaluation subgroups of a map and the @xmath8-sequence .
we have already mentioned some of the results in this area . in his original work on evaluation subgroups
, gottlieb observed that a map of spaces does not necessarily induce a map of gottlieb groups , and gave conditions under which it does ( * ? ? ?
* sec.1 ) .
gottlieb briefly mentions the evaluation subgroups of a map in @xcite , but did not study them as such .
a number of basic properties of the evaluation subgroup of a map are established in @xcite . as we mentioned earlier ,
woo and lee introduced the @xmath8-sequence of a map in @xcite .
results on the @xmath8-sequence basically fall into one of three areas : conditions under which the @xmath8-sequence is exact ( e.g. ( * ? ? ?
* th.12 ) and @xcite ) , examples of non - exactness ( e.g. @xcite and @xcite ) , and extensions and generalizations of evaluation subgroups and the @xmath8-sequence ( e.g. @xcite ) . these results give the stepping - off point for our work in sections 3 and 4 .
using our description of the rationalized @xmath8-sequence ( theorem [ thm : g - seq of f ] ) , we extend the known exactness results to several new cases .
we also give new instances of non - exactness .
more significantly , by focussing on the rational setting , our methods make the production of such examples straightforward . on the other hand , our results and examples are not restricted to amplifying previous results in this area . in corollary
[ cor : g - sequence halperin ] , we suggest a different kind of result that relates properties of the @xmath8-sequence to the triviality of a fibration . furthermore , some results in section 3including theorem [ thm : long exact seq of f _ * ] and theorem [ thm : long exact seq of eval fib]are of interest independently of any relation to the @xmath8-sequence .
we finish this introduction by setting some notation and terminology . throughout this paper , @xmath5 and @xmath32 will denote simply connected cw complexes of finite type . by _ vector space _ we mean a rational graded vector space . by _ algebra _ , we mean the kind of commutative graded algebras over the rationals that arise in rational homotopy .
that is , they are non - negatively graded , connected ( @xmath44 ) and usually simply connected ( @xmath45 ) , with cohomology of finite type . for a vector space @xmath46
, we denote the free commutative graded algebra generated by @xmath46 by @xmath47 .
we use the acronym dg to denote differential graded : thus , dg vector space , dg algebra , and so - forth . for a dg algebra ,
the differential is of degree @xmath48 . in other situations , however , particularly when we consider the complex of derivations of a dg algebra , the differential is of degree @xmath49 .
we will generally refer to a dg vector space whose differential is of degree @xmath49 as a chain complex .
if @xmath50 is a map , either topological or algebraic , then @xmath51 denotes pre - composition by @xmath4 and @xmath27 denotes post - composition by @xmath4 . in any setting in which it is appropriate , we use @xmath52 to denote the map induced on homology ( or cohomology ) by @xmath4 , and @xmath26 to denote the map induced on homotopy groups by the map of spaces @xmath4 .
a map of dg algebras is called a _ quasi - isomorphism _ if it induces an isomorphism on cohomology .
we use @xmath33 in a generic way to denote an evaluation map , and we denote the identity map of a topological space or the identity homomorphism of an algebra by @xmath53 .
we denote the rationalization of a space @xmath5 by @xmath54 and of a map @xmath4 by @xmath55 ( cf .
@xcite ) .
we assume that the reader is familiar with the basics of rational homotopy .
our general reference for this material is @xcite .
we recall , in particular , that a space @xmath5 has a _ minimal model _
@xmath56 , which is a certain type of dg algebra .
namely , @xmath56 is a free algebra @xmath47 with a decomposable differential , that is , @xmath57 .
furthermore , a map of spaces @xmath58 induces a map of minimal models @xmath59 .
we refer to this induced map as the _ sullivan minimal model _ of the map @xmath4 .
it is a complete rational homotopy invariant for a map , and in principle all rational homotopy theoretic information about @xmath4 can be retrieved from it .
passing to cohomology , for example , gives @xmath60 , which corresponds to the homomorphism of rational cohomology algebras induced by @xmath4 .
the results of this paper illustrate how deeper information about a space or map may be retrieved from the minimal model by making correspondingly more sophisticated constructions with the model .
our purpose in this and the next section is to give a unified description in rational homotopy theory of all the terms involved in the definition of the @xmath8-sequence .
informally stated , we show that the homology theory of derivation complexes of sullivan minimal models provides an algebraic model for the rational homotopy theory of function spaces at the level of homotopy groups .
we focus on the following commutative square that appears in the homotopy ladder from which the @xmath8-sequence arises : @xmath61^{(f_*)_\ # } \ar[d]_{\omega_\ # } & \pi_n\big(\map(x , y ; f)\big)\ar[d]^{\omega_\#}\\ \pi_n(x ) \ar[r]_{f_\ # } & \pi_n(y)}\ ] ] it turns out that identifying the rationalization of this commutative square is sufficient to arrive not only at a characterization of the rationalized evaluation subgroups of @xmath4 , but also at a description of the rationalization of the @xmath8-sequence .
furthermore , our identification of the rationalization of this square allows us to conclude several subsidiary results of interest .
we say two maps of vector spaces @xmath62 and @xmath63 are _ equivalent _ if there exist isomorphisms @xmath64 and @xmath65 which make the diagram @xmath66^{f } \ar[d]_{\alpha}^{\cong } & v \ar[d]_{\beta}^{\cong}\\ u ' \ar[r]_{g } & v'}\ ] ] commutative .
this notion of equivalence for vector space maps extends in the obvious way to sequences of vector space maps , commutative squares of vector space maps , and any other diagram of vector space maps .
suppose a dg algebra @xmath67 is isomorphic to @xmath68 in degree zero , that is , @xmath69 .
then the map @xmath70 that sends all elements of positive degree to zero , and is the identity in degree zero , is an augmentation
. this will be the situation in all cases of interest to us here , and thus we refer to @xmath71 as _ the _ augmentation . here , as in the sequel , we regard @xmath68 as the trivial dg algebra concentrated in degree zero and with trivial differential .
thus @xmath72 is a dg algebra map . given dg algebras @xmath73 and @xmath74 and a ( fixed ) dg algebra map @xmath75 ,
@xmath76-derivation _ of degree @xmath3 to be a linear map @xmath77 that _ reduces degree by @xmath3 _ and satisfies the derivation law @xmath78
. we will only consider derivations of positive degree , that is , those that reduce degree by some positive integer . when @xmath79 we require additionally that @xmath80 .
let @xmath81 denote the vector space of @xmath76-derivations of degree @xmath3 , for @xmath82 .
finally , define a linear map @xmath83 by @xmath84 .
a standard check now shows that @xmath85 and thus @xmath86 is a chain complex . in order to cut down on cumbersome notation
, we will usually suppress the differential from our notation , and write @xmath87 for the homology in degree @xmath3 of the chain complex @xmath86 .
that is , @xmath88 denotes the homology represented by @xmath89-cycles of @xmath90 that reduce degree by @xmath3 .
a special case of the preceding that is of interest to us is the one in which @xmath91 and @xmath92 . in this case , the chain complex of derivations @xmath93 is just the usual complex of derivations on the dg algebra @xmath94 .
note once again that we restrict the derivations in degree @xmath53 to the cycles and that the complex is zero in non - positive degrees .
pre - composition with the dg algebra map @xmath95 thus gives a map of chain complexes @xmath96 .
furthermore , post - composition by the augmentation @xmath97 induces dg vector space maps @xmath98 and @xmath99 .
all of the above can be applied to a map of minimal models .
suppose @xmath34 is a map of spaces , and @xmath100 is the corresponding sullivan minimal model of the map @xmath4 .
then we have a commutative square of chain complexes @xmath101^{(\mathcal{m}_{f})^ * } \ar[d]_{\varepsilon _ * } & \der_*(\mathcal{m}_{y } , \mathcal{m}_{x } ; \mathcal{m}_{f } ) \ar[d]^{\varepsilon_*}\\ \der_*(\mathcal{m}_{x } , \q ; \varepsilon ) \ar[r]^{\widehat{(\mathcal{m}_{f})^ * } } & \der_*(\mathcal{m}_{y } , \q ; \varepsilon)}\ ] ] in this square , @xmath102 is the augmentation .
both horizontal maps are obtained by pre - composing with the same map @xmath103 , but in different contexts .
since we will need to distinguish between these two maps notationally in the sequel , we have used an extra decoration on the bottom one .
the main result of this section is the following : [ thm : equivalence of squares ] let @xmath5 and @xmath32 be simply connected cw complexes of finite type , with @xmath5 finite . for @xmath104 ,
the commutative square obtained by rationalizing ( [ eq : homotopy square ] ) is equivalent to the square obtained by passing to homology in degree @xmath3 from ( [ eq : derivation square ] ) .
that is , the commutative squares @xmath105^{(f_*)_\#\otimes\q } \ar[d]_{\omega_\#\otimes\q } & & \pi_n\big(\map(x , y ; f)\big)\otimes\q\ar[d]^{\omega_\#\otimes\q}\\ \pi_n(x)\otimes\q \ar[rr]_{f_\#\otimes\q } & & \pi_n(y)\otimes\q}\ ] ] and @xmath106 ^ -{h((\mathcal{m}_{f})^ * ) } \ar[d]_-{h(\varepsilon _ * ) } & & h_{n}\big(\der(\mathcal{m}_{y } , \mathcal{m}_{x } ; \mathcal{m}_f)\big)\ar[d]^-{h(\varepsilon _ * ) } \\ h_{n}\big(\der(\mathcal{m}_{x } , \q ; \varepsilon)\big)\ar[rr]^-{h(\widehat{(\mathcal{m}_{f})^ * } ) } & & h_{n}\big(\der(\mathcal{m}_{y } , \q ; \varepsilon)\big)}\ ] ] are equivalent for each @xmath107 .
we prove this result below .
first , we comment on some ingredients of the statement and proof , and give some immediate consequences .
[ rem : indecomposables ] in rational homotopy theory there is a standard way to identify the rational homotopy groups of a space @xmath5 , and more generally the homomorphism of rational homotopy groups induced by a map of spaces .
namely , the rational homotopy groups are identified with the dual of the vector space of indecomposables of the minimal model , thus @xmath108 . for maps ,
@xmath109 is identified with the dual of the map of vector spaces of indecomposables @xmath110 induced by the sullivan minimal model @xmath111 of a map @xmath34 ( see ( * ? ? ? * sec.15(d ) ) for details ) . from the bottom maps in the two squares of theorem
[ thm : equivalence of squares ] , we obtain a superficially different description of @xmath109 .
but it is easy to see that this agrees with the standard one : note that the derivation law implies @xmath112 , while the minimality of @xmath113 implies that @xmath114 in the chain complex @xmath115 .
thus we have @xmath116 . the agreement between our description of @xmath109 and
the standard one is obvious from this isomorphism .
theorem [ thm : equivalence of squares ] also contains and depends upon basic results concerning the rationalization of function space components and evaluation subgroups due to several authors .
we consider this material here . when @xmath5 is a finite complex , the function space @xmath117 has the homotopy type of a cw complex by the result of milnor @xcite .
in fact , by hilton - mislin - roitberg ( * ? ? ? * ch .
2.5 ) the components @xmath118 are nilpotent complexes . moreover , given a rationalization @xmath119 of @xmath32 the induced map @xmath120 is a rationalization of @xmath121 ( * ? ? ?
ii , th . 3.11 ) . by (
* th.2.3 ) , rationalization in the initial variable @xmath122 induces a weak equivalence @xmath123 .
these results , together with the naturality of the various maps involved , imply the following result .
[ thm : rationalization ] let @xmath34 be a map between simply connected complexes of finite type with @xmath5 finite .
let @xmath124 denote the rationalization of @xmath4 .
the commutative squares @xmath105^{(f_*)_\#\otimes\q } \ar[d]_{\omega_\#\otimes\q } & & \pi_n\big(\map(x , y ; f)\big)\otimes\q\ar[d]^{\omega_\#\otimes\q}\\ \pi_n(x)\otimes\q \ar[rr]_{f_\#\otimes\q } & & \pi_n(y)\otimes\q}\ ] ] and @xmath125^{((f_\q)_*)_\ # } \ar[d]_{(\omega)_\ # } & & \pi_n\big(\map(x_\q , y_\q ; f_\q)\big)\ar[d]^{(\omega)_\#}\\ \pi_n(x_\q ) \ar[rr]_{(f_\q)_\ # } & & \pi_n(y_\q)}\ ] ] are equivalent for each @xmath126 the result @xmath127 for @xmath5 a simply connected finite complex due to lang @xcite is an easy consequence of theorem [ thm : rationalization ] , as is its generalization @xmath128 ( c.f .
@xcite ) for @xmath129 simply connected complexes of finite type with @xmath5 finite .
the corresponding localization result for the relative gottlieb group can be deduced from the preceding discussion , as well .
[ thm : relative gottlieb rationalization ] let @xmath34 be a map between simply connected complexes of finite type with @xmath5 finite .
then @xmath130 for @xmath131 consider , as in the introduction , the long exact homotopy sequence of the induced map @xmath132 and its relative homotopy group @xmath133 for @xmath131 we also have @xmath134 and the relative group @xmath135 the results cited above and the @xmath136-lemma , imply the maps induced on the various function spaces involved by the rationalizations @xmath137 and @xmath138 induce a rationalization homomorphism @xmath139 the maps @xmath140 and @xmath141 also induce a rationalization @xmath142 by naturality , we obtain a commutative diagram @xmath143^{r } \ar[d]^{\omega_\sharp } & \pi_n((f_\q ) _ * ) \ar[d]^{\omega_\sharp } \\
\pi_n(f ) \ar[r]^{r } & \pi_n(f_\q ) . }
\ ] ] thus @xmath144 combining the preceding , we obtain the following : [ cor : rationalization of the g - sequence ] let @xmath34 be a map between simply connected complexes of finite type with @xmath5 finite .
then the rationalization of the @xmath8-sequence of @xmath4 @xmath145 is equivalent to the @xmath8-sequence of the rationalization of @xmath4 @xmath146 in the next section , we identify the rational @xmath8-sequence in the context of rational homotopy theory .
meanwhile , from theorem [ thm : equivalence of squares ] we immediately retrieve minimal model descriptions of the rationalized gottlieb groups of a space , and of the rationalized evaluation subgroups of a map .
the first of these is well - known : [ cor : rational gottlieb group ] let @xmath5 be a simply connected finite complex .
the rationalized @xmath3th gottlieb group @xmath147 is isomorphic to the image of the induced homomorphism @xmath148 for @xmath126 this is easily translated into the standard minimal model description of the rationalized gottlieb groups given by flix and halperin ( see ( * ? ? ?
* sec.29(c ) ) ) .
specifically , they describe a gottlieb element of the minimal model @xmath149 as a linear map @xmath150 that extends to a derivation of @xmath113 satisfying @xmath151 .
such a derivation @xmath152 is a cycle in @xmath153 , and the class that it represents has non - zero image under @xmath154 precisely when the original linear map @xmath150 is non - zero . on recalling that @xmath155
, we see the two descriptions agree .
[ cor : rational eval subgp of f ] let @xmath34 be a map between simply connected complexes of finite type with @xmath5 finite .
the rationalized @xmath3th evaluation subgroup @xmath156 of the map @xmath4 is isomorphic to the image of the induced homomorphism @xmath157 for @xmath126 this identification of the rationalized evaluation subgroups of the map @xmath4 can be conveniently phrased in a way comparable to the flix - halperin description of the rationalized gottlieb groups : an evaluation subgroup element of the minimal model @xmath100 , with @xmath158 , is a linear map @xmath159 that extends to an @xmath103-derivation @xmath160 satisfying @xmath161 .
whenever such a @xmath162 is non - zero , it is a non - zero element in the image of @xmath154 . in view of the preceding remarks ,
we introduce the following vocabulary and notation .
[ def : eval subgp of dg map ] suppose @xmath95 is a map of dg algebras .
we define the _ evaluation subgroup of @xmath76 _ as the image of the map @xmath163 we denote it by @xmath164 . in the special case in which @xmath165 and @xmath92 , we refer to the _ gottlieb group of @xmath94 _ , and use the notation @xmath166 . from the previous discussion , we see that @xmath167 and @xmath168 .
we will define vector space isomorphisms @xmath169 , @xmath170 , @xmath171 , and @xmath172 to give the following equivalence of commutative squares : @xmath173 ^ -{h((\mathcal{m}_{f})^ * ) } \ar'[d]^-{h(\varepsilon_*)}[dd ] & & h_{n}\big(\der(\mathcal{m}_{y},\mathcal{m}_{x};\mathcal{m}_{f})\big ) \ar[dd]^-{h(\varepsilon _ * ) } \\
\pi_n\big(\map(x , x ; 1)\big)\otimes\q \ar[ru]^{\phi}_{\cong}\ar[rr]_(0.5){(f_*)_\#\otimes1 } \ar[dd]_{\omega_\ # } & & \pi_n\big(\map(x , y ; f)\big)\otimes\q \ar[ru]^{\phi_f}_{\cong } \ar[dd]^(0.3){\omega_\#}\\ & h_{n}\big(\der(\mathcal{m}_{x } , \q;\varepsilon)\big ) \ar'[r]_(0.9){h(\widehat{(\mathcal{m}_{f})^*})}[rr ] & & h_{n}\big(\der(\mathcal{m}_{y } , \q;\varepsilon)\big ) \\
\pi_n(x)\otimes\q \ar[ru]^{\psi_x}_{\cong } \ar[rr]_{f_\#\otimes1 } & & \pi_n(y)\otimes\q\ar[ru]^{\psi_y}_{\cong } } \ ] ] we obtain @xmath174 as the rationalization of a natural homomorphism @xmath175 which we now define .
a representative of a homotopy class @xmath176 determines , via the exponential correspondence , a map @xmath177 that satisfies @xmath178 , where @xmath179 is the inclusion .
the map @xmath180 is often called an _ affiliated map _ for ( a representative of ) @xmath64 . passing to minimal models , we obtain a map @xmath181 with @xmath182 ( equals , not just up to dg homotopy see proposition [ prop : restricted dg homotopy ] ) .
now @xmath183 is a _ formal _ space , which means there is a quasi - isomorphism of dg algebras @xmath184 . in turn
, this gives a quasi - isomorphism @xmath185 .
recall that @xmath47 denotes the free algebra generated by @xmath46 .
that is , polynomial on generators of even degree , and exterior on generators of odd degree .
write @xmath186 as @xmath187 if @xmath3 is even , or as @xmath188 if @xmath3 is odd . given @xmath189
we may write @xmath190 thus defining a linear map @xmath191 that reduces degree by @xmath3 . a standard check using
the fact that @xmath192 is a dg algebra map shows that @xmath193 is an @xmath194-derivation that is a @xmath195-cycle . set @xmath196
\in h_{n}\big(\der(\mathcal{m}_{y } , \mathcal{m}_{x } ; \mathcal{m}_f)\big)$ ] . to show @xmath197 is well - defined ,
suppose that @xmath198 are homotopic representatives of @xmath64 with affiliated maps @xmath199 respectively .
then the homotopy @xmath200 from @xmath201 to @xmath202 gives a homotopy @xmath203 , from @xmath180 to @xmath8 , by setting @xmath204 .
further , since @xmath205 is a _ based homotopy _ , the homotopy @xmath206 satisfies @xmath207 , where @xmath208 denotes the inclusion @xmath209 and @xmath210 denotes the homotopy @xmath211 that is stationary at @xmath4 . a basic result of rational homotopy theory
says that homotopic maps have dg homotopic sullivan minimal models ( see ( * ? ? ?
* ch.12 ) .
so the homotopy @xmath206 gives a dg homotopy @xmath212 between minimal models for @xmath180 and @xmath8 .
translating the restriction on @xmath206 into minimal model terms allows us to assume that @xmath213 is such that @xmath214 for an element @xmath215 .
this translation is intuitively plausible , but its justification requires some technical details , which we provide in proposition [ prop : restricted dg homotopy ] . since the dg homotopy @xmath213 is from @xmath216 to @xmath217 , then at @xmath218 we have @xmath219 , and from @xmath220 , we have @xmath221 . to establish well - definedness , we must show these differ by a boundary in @xmath222 . to this end , use ( [ eq : dg homotopy k ] ) to write separate expressions for @xmath223 and @xmath224 . since a dg homotopy is an algebra map ,
these expressions agree . by equating them and collecting like terms we obtain equations @xmath225 for each @xmath226 . by substituting @xmath227 for @xmath215
, we obtain derivations @xmath228 .
on the other hand , use ( [ eq : dg homotopy k ] ) to write separate expressions for @xmath229 and @xmath230 , with the latter obtained by applying @xmath231 to both sides of ( [ eq : dg homotopy k ] ) . since a dg homotopy respects differentials
, these expressions agree . by equating them and collecting like terms we obtain equations
@xmath232 for each @xmath226 . with the previous substitution , this gives @xmath233 , that is , @xmath234 for each @xmath226 .
it follows that the difference of derivations @xmath235 is a @xmath195-boundary in @xmath236 .
hence @xmath237 is well - defined .
it is not difficult to show that @xmath237 is a homomorphism .
but since the proof requires some technical notions from rational homotopy theory , we postpone it to the appendix ( proposition [ prop : phi homomorphism ] ) .
as we stated before , the map of vector spaces @xmath170 is now obtained as the rationalization of the ( group ) homomorphism @xmath237 .
the map @xmath169 is defined in the same way , specializing to the case in which @xmath6 and @xmath7 .
the maps @xmath171 and @xmath172 are the standard minimal model identification of the rational homotopy groups of a space , as discussed in remark [ rem : indecomposables ] .
next , we show @xmath174 is surjective .
denote by @xmath238_{f_{\q}}$ ] the subset of the set of homotopy classes of maps @xmath239 consisting of classes represented by a map that restricts to @xmath240 on @xmath241 . by theorem [ thm : rationalization ] , we identify @xmath242 with @xmath243 , and hence with @xmath238_{f_{\q}}$ ] .
now suppose given @xmath244
\in h_{n}\big(\der(\mathcal{m}_{y},\mathcal{m}_{x } ; \mathcal{m}_f)\big)$ ] .
use @xmath152 to define @xmath245 for @xmath215 .
since @xmath152 is an @xmath194-derivation that is a cycle , this defines a dg algebra map @xmath246 .
now lift @xmath76 through the surjective quasi - isomorphism @xmath247 as in ( * ? ? ?
* lem.12.4 ) , to obtain a map @xmath248 that satisfies @xmath249 . by the standard correspondence between maps of minimal models and maps of rational spaces ,
this gives a map @xmath250 that satisfies @xmath251 . using , for example , ( * ? ? ?
* th.9.7 ) , we can adjust @xmath180 into a homotopic map @xmath252 that satisfies @xmath253 , so that @xmath254 represents a class of @xmath238_{f_{\q}}$ ] .
as described at the start of this paragraph , @xmath254 corresponds to a homotopy class @xmath255 .
evidently , we have @xmath256 .
finally , we show @xmath174 is injective . since @xmath174 is a vector space homomorphism
, it is sufficient to show that @xmath257 is zero whenever @xmath258 . using the identification of the previous paragraph ,
let @xmath259 be an affiliated map for @xmath64 .
suppose that @xmath260 for @xmath261 .
using @xmath262 , define a map @xmath263 by @xmath264 a routine check verifies that @xmath265 is a dg algebra map .
furthermore , it is a dg homotopy from @xmath266 to the map @xmath267 given by @xmath268 .
now this latter map is a sullivan model of an affiliated map for @xmath269 .
therefore , the dg homotopy translates into a homotopy between affiliated maps @xmath270 for @xmath64 and @xmath271 .
it follows that @xmath272 , and thus @xmath174 is injective .
we observe that , strictly speaking , we have not justified that the homotopy between the maps @xmath239 is relative to @xmath241 , which corresponds to the homotopy between the maps @xmath273 being based .
however , a based map from a sphere is null - homotopic if and only if it is based null - homotopic ( cf .
@xcite ) .
commutativity of the cube diagram follows from the naturality of the homomorphism @xmath274 . by naturality ,
we mean the following : suppose given maps of spaces @xmath275 and @xmath276 .
the we have induced maps of function spaces @xmath277 and @xmath278 . for either case
we obtain a commutative square involving @xmath279 .
namely , we have @xmath280 in the first case , and @xmath281 in the second case , as is easily checked . since @xmath174 is obtained from @xmath274 by localization , the isomorphism @xmath174 has the same naturality property .
this is sufficient to conclude that the top , bottom , left , and right faces of the cube commute . for the evaluation map @xmath282
can be identified with @xmath283 , where @xmath284 and @xmath285 denote basepoints , and @xmath286 inclusion of the basepoint . likewise for the evaluation map @xmath287 , and then @xmath34 can be identified with @xmath288 ( cf . also remark [ rem : indecomposables ] ) .
finally , the front and rear faces commute because the squares ( [ eq : homotopy square ] ) and ( [ eq : derivation square ] ) are commutative .
we finish this section with two more immediate consequences of theorem [ thm : equivalence of squares ] .
the first retrieves a basic result of thom , in the rational homotopy setting .
( ( * ? ? ?
* th.2))[cor : thom ] let @xmath289 be an eilenberg - mac lane space , for @xmath46 a finite dimensional ( ungraded ) rational vector space .
if @xmath5 is a finite cw complex , and @xmath290 is any map , then @xmath291 .
the minimal model for @xmath32 is @xmath292 with zero differential , where @xmath293 denotes the dual vector space of @xmath46 .
it follows easily that @xmath294 .
notice that this result with the remarks on rationalization preceding theorem [ thm : rationalization]easily extends to yield the rational homotopy type of @xmath121 , in case @xmath32 is a rational @xmath206-space .
that is , a space whose rationalization is an @xmath206-space . for if @xmath32 is a rational @xmath206-space , then so too is @xmath295 , which is homotopy equivalent via translation by @xmath4to @xmath121 .
now a rational @xmath206-space is determined up to rational homotopy type by its rational homotopy groups .
furthermore , a rational @xmath206-space is a product of rational eilenberg - mac lane spaces .
hence @xmath121 has the rational homotopy type of a product of spaces @xmath296 , for @xmath297 a rational eilenberg - mac lane space as in the corollary . we give a further consequence of theorem [ thm : equivalence of squares ]
concerning the rational homotopy groups of certain function spaces .
define an _
@xmath298-space _ to be a finite simply connected complex with finite dimensional rational homotopy ( a _ rationally elliptic space _ ) such that @xmath299 .
this type of space features in the following well - known conjecture of halperin ( cf .
@xcite ) : [ conj : halperin ] suppose @xmath5 is an @xmath300-space .
then any fibration @xmath301 of simply connected spaces is tncz , that is , the fibre inclusion @xmath302 induces a surjection on rational cohomology .
the rational homotopy of @xmath303 for @xmath5 an @xmath298-space is directly related to conjecture [ conj : halperin ] ( see @xcite for details ) .
the following result extends ( * ? ? ?
* prop.2.6 ) ( also compare ( * ? ? ? * cor.4.6 ) and ( * ? ? ? * th.b ) ) : [ cor : grivel ] let @xmath0 be a map between @xmath298-spaces .
then for @xmath304 , @xmath305 the argument given by grivel for the case in which @xmath6 and @xmath7can be used word for word to show @xmath306 the result now follows from theorem [ thm : equivalence of squares ] .
in this section , we identify the rationalized long exact homotopy sequences of the maps @xmath34 and @xmath307 , and hence the rationalized @xmath8-sequence of @xmath4 .
our identifications flow from the following observation : suppose given vector space homomorphisms @xmath308 for each @xmath3 .
then , up to equivalence , _ there is a unique way to fit these into a three - term long exact sequence_. since we rely on this observation for our basic results , we make a formal statement of the fact .
[ lem : unique les ] suppose given equivalences of vector space homomorphisms @xmath309^{\phi_n } \ar[d]^{\alpha_n}_{\cong } & b_n \ar[d]^{\beta_n}_{\cong}\\ a'_n \ar[r]_{\phi'_n } & b'_n}\ ] ] for each @xmath3 .
then for any two long exact sequences of vector spaces containing the @xmath310 and the @xmath311 thus @xmath312 & c_{n+1}\ar@{.>}[d]^{\gamma_{n+1}}_{\cong } \ar[r]^{\partial_{n+1 } } & a_n \ar[r]^{\phi_n}\ar[d]^{\alpha_n}_{\cong } & b_n \ar[r]^{p_n}\ar[d]^{\beta_n}_{\cong } & c_n \ar[r]\ar@{.>}[d]^{\gamma_n}_{\cong } & \cdots\\ \cdots \ar[r ] & c'_{n+1 } \ar[r]^{\partial'_{n+1 } } & a'_n \ar[r]_{\phi'_n } & b'_n \ar[r]^{p'_n } & c'_n \ar[r ] & \cdots,}\ ] ] there exist isomorphisms @xmath313 that make the long exact sequences equivalent .
we define a map @xmath314 as follows : decompose @xmath315 and @xmath316 as @xmath317 and @xmath318 , where @xmath319 and @xmath320 are complements . for @xmath321 , define @xmath322 . for @xmath323 , set @xmath324 , where @xmath325 is such that @xmath326 .
it is straightforward to check that @xmath314 is a well - defined isomorphism .
indeed , it restricts to give isomorphisms @xmath327 and @xmath328 .
the required commutativity properties follow immediately .
there is nothing remarkable in this observation .
it is important , nonetheless , since it allows us to choose descriptions of the long exact homotopy sequences that we need in whatever way is most convenient for our purposes .
we begin with the long exact homotopy sequence of the map @xmath329 @xmath330 . from theorem [ thm : equivalence of squares ]
, we see that the map this induces on rational homotopy groups can be identified with @xmath331 .
since this map is a homomorphism induced on homology by a map of chain complexes , there is a standard way to fit it into a long exact sequence , which we now describe . given a map @xmath95 of chain complexes ,
define a _ relative _ chain complex @xmath332 as follows : @xmath333 , with differential @xmath334 of degree @xmath49 given by @xmath335 .
now define chain maps @xmath336 and @xmath337 by @xmath338 and @xmath339 . on passing to homology
, we obtain a long exact sequence of the following form : @xmath312 & h_{n+1}(\rel(\phi ) ) \ar[r]^-{h(p ) } & h_{n}(a ) \ar[r]^-{h(\phi ) } & h_{n}(b ) \ar[r]^-{h(j ) } & h_{n}(\rel(\phi ) ) \ar[r ] & \cdots,}\ ] ] in which @xmath340 denotes the @xmath3th homology of the chain complex @xmath332 .
we refer to this long exact sequence as the _ long exact homology sequence of the map @xmath76_. furthermore , this construction is natural . for suppose given a commutative square of dg vector spaces @xmath341^{\phi}\ar[d]_{\alpha } & b \ar[d]_{\beta}\\ a ' \ar[r]^{\phi ' } & b'.}\ ] ] then the obvious map @xmath342 is a chain map that satisfies @xmath343 and @xmath344 .
thus we obtain a homology ladder @xmath312 & h_{n+1}(\rel(\phi ) ) \ar[r]^-{h(p ) } \ar[d]^-{h(\alpha , \beta ) } & h_{n}(a ) \ar[r]^-{h(\phi)}\ar[d]^-{h(\alpha ) } & h_{n}(b ) \ar[r]^-{h(j)}\ar[d]^-{h(\beta ) } & h_{n}(\rel(\phi ) ) \ar[r]\ar[d]^-{h(\alpha , \beta ) } & \cdots\\ \cdots \ar[r ] & h_{n+1}(\rel(\phi ' ) ) \ar[r]^-{h(p ' ) } & h_{n}(a ' ) \ar[r]^-{h(\phi ' ) } & h_{n}(b ' ) \ar[r]^-{h(j ' ) } & h_{n}(\rel(\phi ' ) ) \ar[r ] & \cdots}\ ] ] in particular , we can apply this construction to the map of chain complexes @xmath345 induced by the minimal model @xmath346 of the map @xmath58 .
[ thm : long exact seq of f _ * ] the long exact sequence induced by @xmath347 on rational homotopy groups is equivalent to the long exact homology sequence of the map ( [ eq : mf^ * 1 ] ) .
specifically , this is a long exact sequence @xmath348 ^ -{h(j ) } & h_{n+1}(\rel((\mathcal{m}_{f})^ * ) ) \ar ` d[l ] ` [ llld]_(0.7){h(p ) } [ llld ] \\
h_{n}\big(\der(\mathcal{m}_{x } , \mathcal{m}_{x } ; 1)\big ) \ar[rr]^-{h((\mathcal{m}_{f})^ * ) } & & h_{n}\big(\der(\mathcal{m}_{y } , \mathcal{m}_{x } ; \mathcal{m}_{f})\big ) \ar[r]^-{h(j ) } & \cdots \\ & & \cdots \ar[r]^-{h(j ) } & h_{3}(\rel((\mathcal{m}_{f})^ * ) ) \ar ` d[l ] ` [ llld]_(0.7){h(p ) } [ llld ] \\
h_{2}\big(\der(\mathcal{m}_{x } , \mathcal{m}_{x } ; 1)\big ) \ar[rr]^-{h((\mathcal{m}_{f})^ * ) } & & h_{2}\big(\der(\mathcal{m}_{y } , \mathcal{m}_{x } ; \mathcal{m}_{f})\big ) } \ ] ] in which @xmath349 is the relative chain complex of the map ( [ eq : mf^ * 1 ] ) , as described above .
by theorem [ thm : rationalization ] we may assume @xmath34 is map of rational spaces .
theorem [ thm : equivalence of squares ] gives equivalences of vector space maps @xmath350_{\phi_{1}}^{\cong } \ar[rr]^-{(f_*)_\ # } & & \pi_{n}\big(\map(x , y;f)\big ) \ar[d]^{\phi_{f}}_{\cong}\\ h_{n}\big(\der(\mathcal{m}_{x } , \mathcal{m}_{x } ; 1)\big ) \ar[rr]_-{h\big ( ( \mathcal{m}_f)^{*}\big ) } & & h_{n}\big(\der(\mathcal{m}_{y } , \mathcal{m}_{x } ; \mathcal{m}_f)\big),}\ ] ] for each @xmath104 . the top horizontal maps are contained in the long exact sequence induced by @xmath351 on rational homotopy groups .
the bottom horizontal maps are contained in the long exact homology sequence of the map ( [ eq : mf^ * 1 ] ) . from lemma [ lem : unique les ] , these sequences are equivalent . when we refer to the long exact homotopy sequence of a map , we mean this in the sense of ( * ?
* chaps.3,4 ) : recall that given a map @xmath34 , this sequence is as follows : @xmath352 if @xmath4 is the inclusion of a subspace , then the groups @xmath353 are just the usual homotopy groups of a pair .
generally , @xmath353 is defined as homotopy classes of pairs @xmath354 such that the diagram @xmath355 \ar[r]^-{g_{1 } } & x \ar[d]^{f } \\
cs^{n-1 } \ar[r]^-{g_{2 } } & y}\ ] ] commutes . since @xmath356 is not necessarily abelian and we are interested in rationalizing this sequence , we stop at @xmath357 . on the other hand
, one can convert @xmath4 into a fibration and use the corresponding long exact sequence in homotopy .
either approach suits our purposes and indeed the same sequence of homotopy groups results from either . from the above , we see that if @xmath358 denotes the homotopy fibre of the map @xmath359 , then for @xmath104 we have @xmath360 where @xmath349 is the relative chain complex of the map ( [ eq : mf^ * 1 ] ) .
the preceding result specializes to give a description of the long exact sequence induced by a general map on rational homotopy groups .
the minimal model @xmath361 of the map @xmath34 also induces a map of chain complexes @xmath362 [ thm : long exact seq of f ] the long exact sequence induced by @xmath34 on rational homotopy groups is equivalent to the long exact homology sequence of the map ( [ eq : mf^ * 2 ] ) .
specifically , this is a long exact sequence @xmath363 ^ -{h(j ) } & h_{n+1}(\rel((\mathcal{m}_{f})^ * ) ) \ar ` d[l ] ` [ llld]_(0.7){h(p ) } [ llld ] \\
h_{n}\big(\der(\mathcal{m}_{x } , \q;\varepsilon)\big ) \ar[rr]^-{h((\mathcal{m}_{f})^ * ) } & & h_{n}\big(\der(\mathcal{m}_{y } , \q;\varepsilon)\big ) \ar[r]^-{h(j ) } & \cdots \\ & & \cdots \ar[r]^-{h(j ) } & h_{3}(\rel((\mathcal{m}_{f})^ * ) ) \ar ` d[l ] ` [ llld]_(0.7){h(p ) } [ llld ] \\
h_{2}\big(\der(\mathcal{m}_{x } , \q;\varepsilon)\big ) \ar[rr]^-{h((\mathcal{m}_{f})^ * ) } & & h_{2}\big(\der(\mathcal{m}_{y } , \q;\varepsilon)\big ) } \ ] ] in which @xmath349 is the relative chain complex of the map ( [ eq : mf^ * 2 ] ) . argue exactly as in the proof of theorem [ thm : long exact seq of f _ * ] , making use of theorem [ thm : equivalence of squares ] and lemma [ lem : unique les ] .
there is already a standard way to describe the long exact sequence induced by a map on rational homotopy groups , using minimal models .
this uses the notion of a so - called k - s model of the map @xmath111 ( * ? ? ?
* sec.15(d ) ) .
the description we give above , however , is better suited to our purposes . note that if @xmath180 denotes the homotopy fibre of the map @xmath34 , then for @xmath104 we have @xmath364 where @xmath349 is the relative chain complex of the map ( [ eq : mf^ * 2 ] ) .
it is perhaps interesting to compare the description given in theorem [ thm : long exact seq of f ] to the standard description of the long exact sequence in rational homotopy groups of a fibration .
we now turn our attention to the @xmath8-sequence , and identify it within our current framework .
suppose given a dg algebra map @xmath95 .
starting from this map , we can construct the following commutative square of dg vector spaces : @xmath365 ^ -{\phi^ * } \ar[d]_-{\varepsilon _ * } & \der_*(a , b ; \phi)\ar[d]^-{\varepsilon _ * } \\
\der_*(b , \q ; \varepsilon)\ar[r]^-{\widehat{\phi^ * } } & \der_*(a , \q ; \varepsilon).}\ ] ] in this diagram , @xmath72 denotes the augmentation of either @xmath366 or @xmath94 , and we have used a decoration to distinguish the lower horizontal map from the upper . on passing to homology and using the naturality of the relative chain complex construction , we obtain the following homology ladder : @xmath312 ^ -{h(j ) } & h_{n+1}(\rel(\phi^ * ) ) \ar[r]^-{h(p ) } \ar[d]^-{h(\varepsilon _ * , \varepsilon _ * ) } & h_{n}\big(\der(b , b ; 1)\big ) \ar[r]^-{h(\phi^*)}\ar[d]^-{h(\varepsilon _ * ) } & h_{n}\big(\der(a , b ; \phi)\big ) \ar[d]^-{h(\varepsilon _ * ) } \
\cdots\\ \cdots \ar[r]^-{h(j ) } & h_{n+1}(\rel(\widehat{\phi^ * } ) ) \ar[r]^-{h(\widehat{p } ) } & h_{n}\big(\der(b , \q ; \varepsilon)\big ) \ar[r]^-{h(\widehat{\phi^ * } ) } & h_{n}\big(\der(a , \q ; \varepsilon)\big)\ \cdots}\ ] ] for @xmath104 .
we supplement definition [ def : eval subgp of dg map ] with the following vocabulary .
[ def : rel eval subgp of dg map ] suppose @xmath95 is a map of dg algebras .
for @xmath367 we define the _ @xmath3th relative evaluation subgroup of @xmath76 _ as the image of the map @xmath368 we denote it by @xmath369 . then the image of the upper long exact sequence in the lower , of the ladder above , gives a ( not necessarily exact ) sequence @xmath370 ^ -{h(\widehat{j } ) } & g^{rel}_{n+1}(a , b ; \phi ) \ar ` d[l ] ` [ lld]_(0.7){h(\widehat{p } ) } [ lld ] \\
g_n(b ) \ar[r]^-{h(\widehat{\phi^ * } ) } & g_n(a , b ; \phi ) \ar[r]^-{h(\widehat{j } ) } & g^{rel}_{n}(a , b ; \phi ) \ar`d[l ] ` [ lld]_(0.7){h(\widehat{p } ) } [ lld ] \\ \cdots & \cdots \ar[r]^-{h(\widehat{j } ) } & g^{rel}_{3}(a , b ; \phi ) \ar ` d[l ] ` [ lld]_(0.7){h(\widehat{p } ) } [ lld ] \\ g_{2}(b ) \ar[r]^-{h(\widehat{\phi^ * } ) } & g_{2}(a , b ; \phi)\big ) } \ ] ] we refer to this sequence as _ the @xmath8-sequence of the map _ @xmath95 .
all of the above can be applied to the minimal model @xmath100 of the map @xmath34 . by doing so , and then collecting together previous results , we obtain the following result . [ thm : g - seq of f ] the rationalization of the @xmath8-sequence of the map @xmath58 , as far as the term @xmath371
, is equivalent to the @xmath8-sequence of the corresponding map of sullivan models .
starting from the cube displayed in the proof of theorem [ thm : equivalence of squares ] , we extend each of the four left - to - right maps into their respective long exact sequences .
this is then completed into an equivalence of ladders , by defining isomorphisms @xmath314 and @xmath372 to give a commutative square @xmath373 ^ -{\gamma_n } \ar[d]_-{\omega_\ # } & h_{n}\big(\rel(\phi^*)\big)\ar[d]^-{h(\varepsilon _ * , \varepsilon _ * ) } \\ \pi_n(f)\ar[r]^-{\widehat{\gamma_n } } & h_{n}\big(\rel(\widehat{\phi^*})\big)}\ ] ] for each @xmath374 .
these isomorphisms are defined as in lemma [ lem : unique les ] , using the top and bottom faces of the cube .
the one commutativity relation that needs checking , namely that of the displayed square , follows easily from the commutativity of the adjacent and parallel squares , together with the way in which the @xmath314 and @xmath372 are defined .
the result now follows , since whenever one has such an equivalence of ladders , the equivalence restricts to give an equivalence of the corresponding sequences of images .
in particular , we obtain the companion result to corollary [ cor : rational gottlieb group ] and corollary [ cor : rational eval subgp of f ] . [ cor : relative eval subgp of f ] let @xmath34 be a map between simply connected complexes of finite type with @xmath5 finite .
the rationalized @xmath3th relative evaluation subgroup @xmath375 of the map @xmath4 is isomorphic to the image of the induced homomorphism @xmath376 for @xmath131 we comment on the low - end terms in the @xmath8-sequence . in theorem [ thm : long exact seq of f ] and theorem [ thm : long exact seq of f _ * ] we terminate our long exact sequences at the terms corresponding to @xmath377 and @xmath357 respectively
this is because we need a simply connected hypothesis to ensure our combination of rationalization and minimal model techniques remains valid .
as a result , our algebraic description of the rationalized @xmath8-sequence terminates at the term corresponding to @xmath378 .
now in theorem [ thm : long exact seq of f ] , we require @xmath5 to be simply connected and finite . as is well - known , this implies @xmath379 for each @xmath208 .
therefore , under our hypotheses , the rationalized @xmath8-sequence of a map @xmath58 should be considered as @xmath136-term ( not necessarily exact ) sequences @xmath380 & g_{2n}(y , x;f)\otimes\q \ar[r ] & g_{2n}^{rel}(y , x;f)\otimes\q
\ar ` d[l ] ` [ lld ] [ lld ] \\
g_{2n-1}(x)\otimes\q \ar[r ] & g_{2n-1}(y , x;f)\otimes\q \ar[r ] & g_{2n-1}^{rel}(y , x;f)\otimes\q \ar[r ] & 0 } \ ] ] for @xmath104 .
our algebraic description of the rationalized @xmath8-sequence given by theorem [ thm : g - seq of f ] includes all these @xmath136-term sequences .
the sporadic " low - end term @xmath381 is best computed by using its characterization given in corollary [ cor : rational eval subgp of f ] . before turning to some applications of our algebraic description of the rationalized @xmath8-sequence , we give one more description of a long exact homotopy sequence . whilst not strictly necessary for our purposes , it is nonetheless interesting .
we will use @xmath382 to denote the _ based _ mapping space component .
then we have the fibration sequence @xmath383 & \map(x , y ; f ) \ar[r]^-{\omega } & y.}\ ] ] we will describe the long exact sequence on rational homotopy groups induced by this fibration .
recall that we have the augmentation @xmath384 for a dg algebra @xmath366 .
let @xmath385 denote the _ augmentation ideal _ , that is , the kernel of @xmath72 . given a dg algebra map @xmath75 , let @xmath386 be the dg algebra map which agrees with @xmath76 in positive degrees and vanishes in degree zero . a dg algebra map @xmath75 together with the short exact augmentation sequence @xmath380 & \widetilde{b } \ar[r]^{i } & b \ar[r]^{\varepsilon } & \q \ar[r ] & 0}\ ] ] of dg algebras gives rise to the short exact sequence of dg vector spaces @xmath380 & \der_{*}(a , \widetilde{b } ; \widetilde{\phi } ) \ar[r]^{i _ * } & \der_{*}(a , b ; \phi ) \ar[r]^{\varepsilon _ * } & \der_{*}(a , \q ; \varepsilon ) \ar[r ] & 0.}\ ] ] this in turn gives a long exact sequence on homology , in the usual way , of the form @xmath387 ^ -{\delta _ * } & h_{n}\big(\der(a , \widetilde{b};\widetilde{\phi})\big ) \ar[r]^-{h(i _ * ) } & h_{n}\big(\der(a , b ; \phi)\big ) \ar[r]^-{h(\varepsilon _ * ) } & h_{n}\big(\der(a , \q ; \varepsilon)\big)\cdots } \ ] ] for @xmath104 .
call this sequence the _ long exact derivation homology sequence _ of the dg algebra map @xmath388 .
[ thm : long exact seq of eval fib ] the long exact sequence induced by the evaluation fibration ( [ eq : eval fibn ] ) on rational homotopy groups is equivalent to the long exact derivation homology sequence ( [ eq : augment1 ] ) of the map @xmath389 .
specifically , this is a long exact sequence @xmath390 ^ -{h(\varepsilon _ * ) } & h_{n+1}\big(\der(\mathcal{m}_{y } , \q ; \varepsilon)\big ) \ar ` d[l ] ` [ lld]_(0.7){\delta _ * } [ lld ] \\
h_{n}\big(\der(\mathcal{m}_{y } , \widetilde{\mathcal{m}_{x}};\widetilde{\mathcal{m}_{f}})\big ) \ar[r]^-{h(i _ * ) } & h_{n}\big(\der(\mathcal{m}_{y } , \mathcal{m}_{x } ; \mathcal{m}_{f})\big ) \ar[r]^-{h(\varepsilon _ * ) } & \cdots \\ & \cdots \ar[r]^-{h(\varepsilon _ * ) } & h_{3}\big(\der(\mathcal{m}_{y } , \q ; \varepsilon)\big ) \ar ` d[l ] ` [ lld]_(0.7){\delta _ * } [ lld ] \\
h_{2}\big(\der(\mathcal{m}_{y } , \widetilde{\mathcal{m}_{x}};\widetilde{\mathcal{m}_{f}})\big ) \ar[r]^-{h(i _ * ) } & h_{2}\big(\der(\mathcal{m}_{y } , \mathcal{m}_{x } ; \mathcal{m}_{f})\big ) \ar[r]^-{h(\varepsilon _ * ) } & h_{2}\big(\der(\mathcal{m}_{y } , \q ; \varepsilon)\big).}\ ] ] let @xmath5 and @xmath32 be simply connected spaces with @xmath5 finite
. then the rational homotopy groups of the based function space @xmath391 are given by @xmath392 for @xmath104 .
we illustrate the effectiveness of the framework established in the previous two sections with examples .
first , we give a composite example that includes specific computation of many of the ingredients of the above .
our example is one in which the @xmath8-sequence of a map fails to be exact ( after rationalization ) at each of the three types of term that occur .
we begin with some notational conventions .
suppose that @xmath393 and @xmath394 are minimal algebras , with @xmath395 and @xmath396 for suitable graded vector spaces @xmath397 and @xmath46 .
let @xmath75 be a fixed dg algebra map .
since any linear map @xmath398 extends in a unique way to a @xmath76-derivation , we can view the space @xmath399 of negative degree linear maps as a subspace of @xmath400 .
( of course , in degree @xmath53 we must restrict to cycles ! ) with this view @xmath401 as graded spaces although the differential depends on the derivation structure .
this point of view , whereby a derivation is specified on generators and then extended to the whole algebra , is one that we will invariably adopt in any practical calculation .
now suppose given a basis @xmath402 for @xmath397 and an element @xmath403 with @xmath404 we will write @xmath405 for the @xmath76-derivation carrying @xmath406 to @xmath407 and vanishing on the other @xmath408 . thus any derivation can be expressed as a sum @xmath409 . when @xmath410 we write @xmath411 rather than @xmath412 for the derivation dual to @xmath408 .
[ ex : non - exact g - seq ] let @xmath413 be the map with coordinate functions @xmath414 obtained by pinching out the bottom cell and @xmath415 the inclusion .
we will use our framework from above to compute various terms from the long exact sequences corresponding to theorem [ thm : long exact seq of f _ * ] and theorem [ thm : long exact seq of f ] . denote @xmath416 by @xmath5 and @xmath417 by @xmath32 , thus @xmath34 .
our computation will show , using theorem [ thm : g - seq of f ] , that the @xmath8-sequence of @xmath4 is non - exact at the terms @xmath418 , @xmath419 , and @xmath420 .
first , @xmath421 , with differential given on generators by @xmath422 , and @xmath423 , and @xmath424 with differential @xmath425 , @xmath426 , @xmath427 , and @xmath428 . in both models , subscripts denote degrees .
then the sullivan model of @xmath4 , which we denote by @xmath429 , is given on generators by @xmath430 , @xmath431 , @xmath432 , and @xmath433 .
for degree reasons , @xmath434 unless @xmath435 or @xmath436 .
furthermore , @xmath437 is spanned by the derivations @xmath438 , @xmath439 , @xmath440 , and @xmath441 of degree @xmath30 , @xmath442 , @xmath443 , and @xmath436 respectively .
an easy computation reveals that @xmath444 , but that @xmath445 and @xmath441 are both ( non - exact ) cocycles .
it follows from theorem [ thm : equivalence of squares ] that @xmath446 further , it is evident that @xmath447 with the non - zero cohomology in degrees @xmath442 and @xmath436 generated by cocycles @xmath439 and @xmath441 , respectively .
given these generators , we see that @xmath448 that is , the homomorphism @xmath449 induced by the evaluation map on rational homotopy groups , is an isomorphism in degree @xmath436 and is zero in all other degrees .
it follows from theorem [ thm : equivalence of squares]see corollary [ cor : rational gottlieb group]that @xmath450 other than in degree @xmath436 , where we have @xmath451 . up to this point , our observations are both well - known , and
also easily obtained by a number of standard methods .
we now show that the rationalized @xmath8-sequence is non - exact at the @xmath452 term . recall that this term of the @xmath8-sequence , together with its adjacent terms , is obtained from the diagram @xmath453 ^ -{p } \ar[d]^-{(\varepsilon _ * , \varepsilon _ * ) } & \der_{11}(\mathcal{m}_x , \mathcal{m}_x ; 1 ) \ar[r]^-{\phi^*}\ar[d]^-{\varepsilon _ * } & \der_{11}(\mathcal{m}_y , \mathcal{m}_x ; \phi ) \ar[d]^-{\varepsilon_*}\\ \rel_{12}(\widehat{\phi^ * } ) \ar[r]^-{\widehat{p } } & \der_{11}(\mathcal{m}_x , \q ; \varepsilon ) \ar[r]^-{\widehat{\phi^ * } } & \der_{11}(\mathcal{m}_y , \q ; \varepsilon ) , } \ ] ] by passing to homology and then considering the image of the top sequence in the bottom . a brute force calculation will display the result , but we opt to argue at a more general level so as to indicate some reason for non - exactness .
it is evident that @xmath454 ) = 0 \in h_{11}\big(\der(\mathcal{m}_y , \q ; \varepsilon)\big)$]indeed , this latter term is zero , since it is isomorphic to @xmath455 . the key point for non - exactness here , however , is that in the top sequence we have @xmath456 ) \not= 0 \in h_{11}\big(\der(\mathcal{m}_y , \mathcal{m}_x ; \phi)\big)$ ] .
in fact , a straightforward check shows that @xmath457 . since @xmath458 , there are no non - zero boundaries in degree @xmath436 and hence @xmath456 ) \not= 0 $ ] .
consequently , @xmath459 $ ] can not be in the image of @xmath460 in the top sequence .
therefore , since @xmath154 is an isomorphism in degree @xmath436 , @xmath461 ) = [ x_{11}^*]$ ] can not be in the image of @xmath462 .
it follows from these facts that @xmath459 \in g_{11}(\mathcal{m}_{x})$ ] is a non - zero element in the kernel of @xmath463 and yet is not in the image of @xmath464 .
next consider the term @xmath465 : before passing to homology , the relevant diagram is the following : @xmath466 ^ -{\phi^*}\ar[d]^-{\varepsilon _ * } & \der_{4}(\mathcal{m}_y , \mathcal{m}_x ; \phi ) \ar[d]^-{\varepsilon _ * } \ar[r]^-{j } & \rel_{4}(\phi^ * ) \ar[d]^-{(\varepsilon _ * , \varepsilon _ * ) } \\
\der_{4}(\mathcal{m}_x , \q ; \varepsilon ) \ar[r]^-{\widehat{\phi^ * } } & \der_{4}(\mathcal{m}_y , \q ; \varepsilon ) \ar[r]^-{\widehat{j } } & \rel_{4}(\widehat{\phi^ * } ) , } \ ] ] the derivation @xmath467 is a cocycle , as is easily checked . under @xmath468 , we have @xmath469 ) = [ y_{4}^ * ] \not=0 $ ] .
since @xmath470 = h(\widehat{\phi^*})([x_{4}^*])$ ] , it follows that @xmath471 ) = 0 $ ] . as we noted above , however , @xmath472 .
therefore , @xmath470
\in g_{4}(\mathcal{m}_{y } , \mathcal{m}_{x } ; \phi)$ ] is a non - zero element in the kernel of @xmath473 that is not in the image of @xmath474 .
finally , consider the term @xmath475 : here , we begin with the following diagram : @xmath476 ^ -{\varepsilon _ * } \ar[r]^-{j } & \rel_{8}(\phi^ * ) \ar[d]^-{(\varepsilon _ * , \varepsilon _ * ) } \ar[r]^-{p } & \der_{7}(\mathcal{m}_x , \mathcal{m}_x ; 1 ) \ar[d]^-{\varepsilon _ * } \\
\der_{8}(\mathcal{m}_y , \q ; \varepsilon ) \ar[r]^-{\widehat{j } } & \rel_{8}(\widehat{\phi^ * } ) \ar[r]^-{\widehat{p } } & \der_{7}(\mathcal{m}_x , \q ; \varepsilon ) } \ ] ] we find that @xmath477 is a cocycle that has non - zero image in @xmath478 .
furthermore , it is evident that @xmath479 ) = 0$]indeed , @xmath480 . to see that @xmath481 $ ] is not in the image of @xmath482
, we will compute @xmath483 to be zero .
a general derivation @xmath484 can be written as @xmath485 for coefficients @xmath486 . to find the cocycles of this form
, we first observe that @xmath487 must vanish on the generators @xmath488 and @xmath489 , for degree reasons . on @xmath490 , we compute as follows : @xmath491 thus , if @xmath492 is a cocycle , then we must have @xmath493 . a similar computation shows
that @xmath494 , and thus that there are no non - zero cocycles in @xmath495 . in summary , we have computed that @xmath496 this last part of our computation is easily confirmed using the result of corollary [ cor : grivel ] .
it follows , of course , that @xmath497 and , in particular , that @xmath481 $ ] is not in the image of @xmath498 .
the first example of a non - exact @xmath8-sequence , given in @xcite , was in dimension one .
a higher dimensional example was produced later in @xcite . with the approach illustrated in the above example
, it is straightforward to produce higher dimensional examples of non - exact rationalized @xmath8-sequences .
observe that non - exact ordinary @xmath8-sequences are produced as a result , since if a sequence of abelian groups is not exact after tensoring with @xmath68 then it was not exact to begin with .
thus , non - exactness rationally implies non - exactness integrally . although the @xmath8-sequence in general is non - exact , there are certain situations in which it is perfectly well behaved , at least after rationalization .
we now mention one such situation .
we say that a space @xmath5 is an _
@xmath499-space _ if its rational cohomology algebra is a free graded algebra ( exterior algebra on the odd - degree generators tensored with a polynomial algebra on the even - degree generators ) .
equivalently , we could require that @xmath5 be an @xmath206-space after rationalization , whence such a space is also referred to as a _
rational @xmath206-space_. recall also the definition of an @xmath298-space from before corollary [ cor : grivel ] .
[ thm : exact g - seq ] let @xmath9 be any map from an @xmath298-space @xmath5 to an @xmath499-space @xmath32 that induces the zero homomorphism on rational homotopy groups .
then the rationalized @xmath8-sequence splits into short exact sequences @xmath380 & g_{n+1}(y , x;f)\otimes\q \ar[r]^-{j } & g^{rel}_{n+1}(y , x;f)\otimes\q \ar[r]^-{p } & g_n(x)\otimes\q \ar[r ] & 0}\ ] ] for each @xmath104 .
our assumption that @xmath500 means that the long exact sequence induced by @xmath4 on rational homotopy groups splits .
furthermore , since @xmath32 is an @xmath501-space , we have @xmath502 for each @xmath3 .
it follows that @xmath503 for each @xmath3 . from these general considerations ,
we have short exact sequences @xmath504 & g_{n+1}(y , x ; f ) \otimes\q \ar[r]^-{j } & \pi_{n+1}(f)\otimes\q \ar[r]^-{p } & \pi_n(x)\otimes\q \ar[r ] & 0}\ ] ] for @xmath104 . to sharpen this to the statement of the theorem , we work within our minimal model framework .
we use some results of halperin @xcite , on the rational homotopy of an @xmath298-space .
these state that @xmath5 has minimal model @xmath505 of the form @xmath506 where @xmath507 is even , @xmath508 is odd , @xmath509 and @xmath510 .
furthermore , the cohomology is evenly graded , and any cocycle in @xmath511 , the ideal of @xmath113 generated by the @xmath512 , is exact .
it is well - known that the minimal model of an @xmath501-space takes the form @xmath513 , with trivial differential .
the map @xmath514 has sullivan minimal model @xmath515 that is determined by the @xmath516 .
since each @xmath517 is a cocycle , it follows that each @xmath518 is a cocycle . from the results of halperin mentioned above , we can write @xmath519 , for suitable @xmath520 and @xmath521 .
the assumption that @xmath4 induces zero on rational homotopy groups translates into the further restriction that each @xmath522 is decomposable . the short exact sequences ( [ eq : short exact general ] ) , translated into our derivation setting , correspond to short exact sequences @xmath523 ^ -{h(\widehat{j } ) } & h_{n+1}\big(\rel(\widehat{\phi^*})\big ) \ar[r]^-{h(\widehat{p } ) } &
h_n\big(\der(\mathcal{m}_{x } , \q ; \varepsilon)\big ) \to 0}\ ] ] we first prove that each map @xmath524 restricts to give a surjection @xmath525 the gottlieb elements of @xmath113 are precisely the @xmath526 , dual to the odd - degree generators of @xmath113 . this can be seen from the description of the gottlieb elements given in corollary [ cor : rational gottlieb group ] , together with the facts about the minimal model of @xmath5 recalled earlier .
now write @xmath519 as above and , for each @xmath527 , define a derivation @xmath528 then @xmath529 , since the differential in @xmath530 is trivial . on the other hand
, we have @xmath531 with the last step following because @xmath532 is a @xmath89-cycle . consequently
, @xmath533 is a @xmath534-cycle . since @xmath535
, it follows that @xmath524 does restrict to the desired surjection .
the map @xmath536 is injective on rational homotopy groups , as we have already observed , and therefore restricts to an injection in the rationalized @xmath8-sequence .
so it only remains to show exactness at the @xmath537 terms .
suppose that @xmath538 ) \in \ker\big(h(\widehat{p } ) \colon g^{rel}_{n+1}(\mathcal{m}_{y } , \mathcal{m}_{x } ; \phi ) \to g_{n}(\mathcal{m}_{x})\big),\ ] ] for some cocycle @xmath539 .
the fact that @xmath540 ) = 0 $ ] implies that @xmath541 .
thus @xmath542 .
now define a derivation @xmath543 by setting @xmath544 on generators of @xmath530 of degree @xmath545 and @xmath546 on all other generators of @xmath530 .
it is easily seen that @xmath547 is a cycle .
indeed , @xmath548 on all generators of @xmath530 , since @xmath547 has non - zero image only in degree zero , and @xmath549 since @xmath550 .
thus @xmath551 ) \in g_{n+1}(\mathcal{m}_{y } , \mathcal{m}_{x } ; \phi)$ ] satisfies @xmath552 ) = [ ( 0 , \varepsilon \circ \overline{\psi } ) ] = [ ( 0 , \varepsilon \circ \psi ) ] = h(\varepsilon_*,\varepsilon_*)([(\theta , \psi)]).\ ] ] that is , @xmath553 and the rationalized @xmath8-sequence is exact at each @xmath537 term .
various conditions are known , under which the @xmath8-sequence of a map @xmath34 is exact .
for instance , it is exact when @xmath4 is null - homotopic @xcite , and when @xmath4 is a homotopy monomorphism @xcite .
the hypotheses of theorem [ thm : exact g - seq ] are well - suited for rational homotopy theory .
both types of space are well - known , and it is easy to give examples to which the theorem applies .
in fact , for fixed @xmath5 and @xmath32 , the maps to which it applies are classified up to rational homotopy by the decomposable rational cohomology of @xmath5 that occurs in those ( even ) degrees in which @xmath32 has a generator of rational cohomology .
we emphasize that the @xmath501-space @xmath32 must be allowed to have polynomial generators in rational cohomology , and hence be infinite - dimensional , otherwise the theorem reduces to the case in which the map @xmath4 is rationally null - homotopic .
furthermore , the hypothesis that @xmath4 be zero on rational homotopy groups is necessary .
for example , the map @xmath554 , given by inclusion of the bottom cell , has a non - exact rationalized @xmath8-sequence , as is easily confirmed by computations similar to those of example [ ex : non - exact g - seq ] since the @xmath8-sequence of a map @xmath9 is a boundary sequence , but not usually an exact sequence , it is natural to consider its homology .
this gives the so - called @xmath33-homology of @xmath4 @xcite .
in general , one obtains an @xmath33-homology group at each of the three types of term . in the following ,
we restrict our attention to the @xmath33-homology that occurs at the gottlieb group term @xmath14 , denoted @xmath555 in @xcite .
thus we consider the sub - quotients of the gottlieb groups @xmath14 defined by @xmath556 when @xmath32 is an @xmath499-space , the rational @xmath33-homology of @xmath557 is related to the the negative derivations on the rational cohomology of @xmath5 that are induced by derivations on the minimal model . to be precise ,
define a linear map of degree zero @xmath558 by the rule @xmath559)([\chi ] ) = [ \theta(\chi)]$ ] , for @xmath152 a cycle in @xmath560 and @xmath561 a cocycle in @xmath113 .
it is straightforward to check that @xmath562 is well - defined .
* prop.1.6 ) .
in fact , @xmath562 is a morphism of graded lie algebras . ) in the next result , and the example that follows it , we illustrate that the rationalized @xmath8-sequence may be exact at all occurrences of one type of term , while failing to be exact at the other types of term . in other words ,
the rational @xmath33-homology of a map may be zero at one type of term , yet non - zero at the other types of term .
since @xmath32 is an @xmath499-space , its minimal model @xmath565 has trivial differential .
let @xmath566 denote the minimal model of @xmath4 .
for a derivation @xmath567 , we have @xmath568 . using this observation
, we obtain a map of chain complexes @xmath569 defined by @xmath570 $ ] .
using the preceding observation , together with the free - ness of @xmath571 , it is straightforward to check that @xmath572 induces an isomorphism on passing to homology ( note that the right - hand term has trivial differential , and so is its own homology ) .
furthermore , the following diagram commutes : @xmath573 ^ -{h(\phi^ * ) } \ar[d]_{\varphi_{x } } & & h_{*}\big(\der(h^*(y;\q ) , \mathcal{m}_{x};\phi)\big ) \ar[d]^{\mu}_{\cong } \\
\der_{*}(h^*(x;\q ) , h^*(x;\q);1 ) \ar[rr]_{\big(h(\phi)\big)^ * } & & \der_{*}\big(h^*(y;\q ) , h^*(x;\q);h(\phi)\big ) } \ ] ] therefore , the assumption that @xmath574 implies that the top map @xmath575 in the above diagram is zero . a straightforward diagram chase using the homology ladder that defines the rationalized @xmath8-sequence now gives the result .
following theorem [ thm : exact g - seq ] we remarked that the cellular inclusion @xmath576 does not have an exact rationalized @xmath8-sequence .
however , it does satisfy the hypotheses of theorem [ thm : h^omega=0 ] , since here @xmath577 has the property that all derivations of the cohomology algebra are trivial .
the hypothesis on @xmath5 in theorem [ thm : h^omega=0 ] , that @xmath578 , deserves some comment .
first , we observe that the nature of the hypothesis distinguishes structure at the minimal model level from structure at the cohomology level .
this is a distinction that is made in rational homotopy for a wide variety of structures .
next , we observe that this condition is satisfied for many , if not all , @xmath300-spaces @xmath5 . indeed , conjecture [ conj : halperin]the long - standing conjecture of halperin concerning @xmath300-spaces is equivalent to the assertion that all negative - degree derivations on the cohomology algebra of an @xmath300-space are trivial ( see @xcite for details ) .
whenever this conjecture is true and it has been verified in many cases obviously we have @xmath578 .
therefore , theorem [ thm : h^omega=0 ] can be compared with theorem [ thm : exact g - seq ] , as a result with weaker hypotheses , and correspondingly weaker conclusion .
finally , we note that the map @xmath562 makes an appearance in a completely different context , in the work of belegradek and kapovitch @xcite .
our last set of results relate directly to conjecture [ conj : halperin ] .
first , we observe that for an inclusion of a summand of a product , the @xmath8-sequence behaves in a particularly nice way . since it is no harder to do so , we state and prove this result in the integral setting .
[ prop : split short g ] suppose that @xmath579 is the inclusion into the first summand .
then the @xmath8-sequence of @xmath580 is exact , and furthermore reduces to split short exact sequences @xmath380 & g_n(x ) \ar[r]^-{(i_1)_\ # } & g_n(x \times
b , x ; i_2 ) \ar[r]_-{(p_2)_\ # } & \pi_n(b ) \ar[r ] \ar@/_1pc/[l]_-{(i_2)_\ # } & 0 , } \ ] ] where @xmath581 is projection onto the second summand and the splitting is induced by inclusion into the second summand @xmath582 .
this follows from results in @xcite ( see also @xcite ) , but we give a brief argument here .
first , let @xmath583 be any fibre sequence .
hilton s excision homomorphism for relative homotopy groups gives an isomorphism @xmath584 ( * ? ? ?
* chap.3 ) .
thus we may view @xmath585 as a subgroup of @xmath586 and the @xmath8-sequence of the fibre inclusion @xmath527 as a subsequence the long exact homotopy sequence of the fibration .
now apply this remark to the trivial fibration @xmath587 .
the inclusion @xmath582 induces a splitting of the long exact homotopy sequence of this trivial fibration in the usual way .
the result now follows from the observation that @xmath588 , as is easily established from the definitions .
of course , this result and its proof can be rationalized , and it is in the rational setting that we will use it .
conjecture [ conj : halperin ] concerns fibrations @xmath301 with fibre an @xmath300-space and arbitrary base .
however , it is well - known how to reduce the conjecture to consideration of such fibrations with base an odd - dimensional sphere @xcite .
furthermore , in @xcite it is pointed out that , for such fibrations with base an odd - dimensional sphere , halperin s conjecture actually asserts that the fibration should be trivial . from these remarks , we see that a necessary condition for conjecture [ conj : halperin ] to be true is that any fibration @xmath589 with @xmath5 an @xmath300-space must have a fibre inclusion whose @xmath8-sequence reduces to the split short exact sequences corresponding via proposition [ prop : split short g ] to the inclusion @xmath590 .
perhaps surprisingly , the converse is true . 1 .
the fibration is rationally tncz , that is , @xmath592 is surjective .
2 . the rationalized @xmath8-sequence of the fibre inclusion reduces to split short exact sequences @xmath380 & g_n(x ) \otimes \q \ar[r]^-{(j)_\ # } & g_n(e , x ; j ) \otimes \q \ar[r]_-{(p)_\ # } & \pi_n(s^{2r+1 } ) \otimes \q \ar[r ] \ar@/_1pc/[l]_-{(i_2)_\ # } & 0}\ ] ] for each @xmath126 suppose the fibration @xmath35 has minimal model @xmath594 ^ -{i } & \lambda(u)\otimes\lambda v , d \ar[r]^-{\pi } & ( \lambda v , d ) , } \ ] ] where @xmath208 denotes the inclusion @xmath595 and @xmath596 is the projection . the hypothesis that @xmath597 is onto included in ( 2 ) , when translated into our derivation setting , gives the existence of a @xmath596-derivation @xmath598 that is a cocycle , and that satisfies @xmath599 . using this @xmath600 ,
define a linear map @xmath601 by setting @xmath602 for a typical element @xmath603 .
we will check that @xmath169 is actually a dg algebra isomorphism @xmath604 .
first , @xmath169 is an algebra map .
this follows from the fact that @xmath600 is a derivation .
suppose given two elements @xmath605 .
then we have @xmath606 on the other hand , we have @xmath607 these agree , since @xmath608 .
next , we check that @xmath169 commutes with differentials , that is , that @xmath609 .
this will follow from the fact that @xmath600 is a cocycle .
first observe that @xmath610 .
it is thus sufficient to check that @xmath611 for a typical element @xmath612 . to this end , write @xmath613 .
now we calculate @xmath614 and @xmath615 these agree since @xmath600 is a cocycle , whence @xmath616 .
it is evident that @xmath169 is an isomorphism , since we have @xmath617 and @xmath618 .
therefore , @xmath619 is a dg isomorphism and the fibration is rationally trivial .
[ cor : g - sequence halperin ] let @xmath5 be an @xmath298-space
. then @xmath5 satisfies conjecture [ conj : halperin ] if and only if the @xmath8-sequence of the fibre inclusion in every fibration of the form @xmath620 decomposes into split short exact sequences as in ( 2 ) of theorem [ thm : halperin g - trivial ] . in @xcite
, meier showed that halperin s conjecture for @xmath5 is equivalent to the collapsing of the rational serre spectral sequence for all fibrations with fibre @xmath5 and base an odd sphere . by (
* th.2.3 ) this latter condition is equivalent to the rational homotopy triviality of the fibration .
the result now follows from theorem [ thm : halperin g - trivial ] .
it is possible to develop the ideas leading to theorem [ thm : halperin g - trivial ] substantially beyond the application given above .
namely , one can use the splitting of the @xmath8-sequence of the fibre inclusion as a _ measure of how close the fibration is to being trivial_. now in the above results , we see that for fibrations of the form @xmath620 , with @xmath5 an @xmath300-space , this notion coincides with the notion of the fibration being tncz . furthermore ,
in this closely restricted setting , both coincide with the fibration actually being trivial . in general , however , this @xmath8-sequence point of view gives a new way of measuring how close a fibration is to being trivial .
we intend to develop these ideas in a subsequent paper . in this appendix , we give careful proofs of the results from dg algebra homotopy theory that are used to establish theorem [ thm : equivalence of squares ] .
since it is a technical appendix , we rely on a greater degree of familiarity with techniques from rational homotopy theory .
we use the notion of pullback in the dg algebra setting . by this , we mean the following .
suppose given dg algebra maps @xmath621 and @xmath276 .
then we form the _
( dg algebra ) pullback _ ( or _ fibre product _ , as it is called in @xcite ) as @xmath622 .
here @xmath623 denotes the direct sum of dg algebras . together with the projections , the pullback forms the following ( strictly ) commutative square of dg algebra maps : @xmath624 ^ -{p_1}\ar[d]_{p_2 } & a \ar[d]^{f}\\ b\ar[r]_{g } & c}\ ] ] this square possesses the usual universal property of pullbacks .
namely , suppose given dg algebra maps @xmath625 and @xmath626 that satisfy @xmath627
. then there exists a dg algebra map @xmath628 , which is the unique dg algebra map for which @xmath629 and @xmath630 . we emphasize that throughout this appendix we distinguish carefully between diagrams that are strictly commutative and ones that are commutative only up to dg homotopy .
indeed , it is precisely this distinction that calls for the proofs of this appendix .
we also use the so - called _ surjective trick _ , described in ( * ? ? ? * sec.12(b ) ) . given a dg algebra map @xmath633 , this manoeuvre results in a diagram @xmath634_{\eta } \ar@<0.75ex>[r]^-{\lambda } & b\otimes e(a ) \ar@<0.75ex>[l]^-{1\cdot\varepsilon } \ar@{>>}[ld]^{\gamma}\\ a}\ ] ] in which @xmath492 is a surjection , and both @xmath635 and @xmath636 are quasi - isomorphisms .
some parts of the diagram commute , thus @xmath637 and @xmath638 .
other compositions result in commutativity only up to dg homotopy . recall that the notion of dg homotopy is only defined for dg algebra maps from a minimal model . given any map @xmath639 from a minimal model into @xmath640
, we have @xmath641 , where @xmath642 denotes dg homotopy of maps from a minimal model .
in particular , we thus have @xmath643 .
now suppose given a map @xmath34 .
we choose and fix a minimal model @xmath644 for @xmath4 as follows ( ( * ? ? ?
* sec.12(c ) ) ) : let @xmath645 denote the map induced by @xmath4 on polynomial differential forms .
let @xmath646 @xmath647 and @xmath648 denote minimal models for @xmath5 and @xmath32 .
as in ( * ? ? ?
* sec.12(b ) ) , we convert @xmath649 into a surjection @xmath650 and lift @xmath651 through the surjective quasi - isomorphism @xmath652 , using ( * ? ? ?
* lem.12.4 ) , to obtain @xmath653 .
now set @xmath654 .
all this is summarized in the following diagram . @xmath655 ^ -{\gamma_x}_(0.6){\simeq } \ar@<1ex>[dl]^(0.6){\beta}\ar@<1ex>[dl]^(0.4){\simeq } \\
\mathcal{m}_y \ar@{.>}[rr]_{\mathcal{m}_f } \ar@/^1pc/[rrru]^-{\phi_{f}}\ar[d]_{\eta_y}^{\simeq } & & \mathcal{m}_x \ar[d]_{\eta_x}^{\simeq } \ar[ur]^-{\alpha}\ar[ur]^(0.3){\simeq}\\
a^*(y ) \ar[rr]_{a^*(f ) } & & a^*(x)}\ ] ] in this and subsequent diagrams , we indicate that a map is a quasi - isomorphism with the symbol @xmath656 . by construction , we have @xmath657 , @xmath658 , and @xmath659 .
remaining parts of the diagram only commute up to dg homotopy , however , thus we have @xmath660 , @xmath661 , and so - on .
now let @xmath662 be a representative of a homotopy class in @xmath663 .
let @xmath664 be an affiliated map for @xmath64 , that is , @xmath665 . since @xmath64 is a based map , we have @xmath666 , where @xmath667 denotes ( based ) inclusion into the second summand @xmath668 . in the following result
, we justify that the sullivan minimal model of any affiliated map , and a dg homotopy between two such , have the restricted form that we require of them for the definition and well defined - ness of @xmath170 . [ prop : restricted dg homotopy ] suppose given maps @xmath669 and @xmath670 a homotopy from @xmath180 to @xmath8 that is stationary on @xmath671 .
suppose that @xmath672 for @xmath18 and let @xmath644 be a fixed choice of minimal model for @xmath4 .
there is a dg homotopy @xmath673 from @xmath216 to @xmath217 , minimal models for @xmath180 and @xmath8 , of the form @xmath674 in particular , any map @xmath675 that satisfies @xmath676 has a minimal model @xmath677 of the form @xmath678 construct the following pullback : @xmath679 ^ -{p_1}\ar[d]_{p_2 } & \mathcal{m}_x \otimes \lambda(t , dt ) \otimes e(a^*(x\times i ) ) \ar@{>>}[d]^{\gamma}_{\simeq}\\ a^*(s^n\times x\times i)\ar@{>>}[r]_-{a^*(i ) } & a^*(x\times i)}\ ] ] here , @xmath680 denotes the inclusion @xmath681 and @xmath492 denotes the surjective quasi - isomorphism obtained by converting the quasi - isomorphism @xmath682 to a surjection . since @xmath208 is an inclusion , the induced map @xmath683 is a surjection . from lemma [ lem : basic pullback ] , we have that @xmath632 is a quasi - isomorphism .
now consider the following commutative diagram : @xmath684|(0.68)\hole^(0.3){\phi } \ar@/^1pc/@{>>}[rdd]^(0.7){\;\varepsilon\cdot\big(1\otimes1\otimes e(a^*(i))\big ) } \ar@/_2pc/@{>>}[ddd]|(0.46)\hole^(0.3){\gamma ' } \\ & & \\
\mathcal{m}_y \ar@{.>}[r]|(0.74)\hole_{\psi } \ar@{.>}[ruu]^{\phi_h } \ar@/_/[dr]_(0.4){a^*(h)\circ\eta_y } \ar@/^2pc/[rr]^(0.55){\big(j\otimes e(a^*(i))\big)\circ\phi_f } & p \ar[r]^-{p_1}\ar[d]^{p_2 } & \mathcal{m}_x \otimes \lambda(t , dt ) \otimes e(a^*(x\times i ) ) \ar@{>>}[d]^{\gamma}\\ & a^*(s^n\times x\times i)\ar[r]_-{a^*(i ) } & a^*(x\times i),}\ ] ] here , @xmath685 is the surjective quasi - isomorphism obtained by converting the quasi - isomorphism @xmath686 to a surjection , @xmath653 denotes the lift used above to obtain our fixed choice of minimal model for @xmath4 , and @xmath687 denotes the map @xmath688 . from the pullback
, we obtain the maps @xmath76 and @xmath600 indicated .
since both @xmath632 and @xmath685 are quasi - isomorphisms , it follows that @xmath76 is a quasi - isomorphism .
we claim that @xmath76 is also surjective . for suppose @xmath689 , so that @xmath690 .
since @xmath691 is surjective , we can pick @xmath692 with @xmath693 , and @xmath694 .
thus @xmath695 .
so now pick @xmath696 with @xmath697recall that @xmath698 is freely generated by the vector space @xmath699 and its suspension .
then @xmath700 ; so @xmath76 is indeed surjective .
now lift @xmath600 through the surjective quasi - isomorphism @xmath76 , to obtain a map @xmath701 with @xmath702 . as usual for the minimal model of a map , we now set @xmath703 .
we now check that this dg homotopy starts and ends at minimal models for @xmath180 and @xmath8 respectively .
let @xmath704 denote inclusion at @xmath705 , and @xmath706 the augmentation given by @xmath707 and @xmath708 .
we must check that @xmath709 is a minimal model for @xmath180 . for this , consider the following diagram : @xmath710|(0.67)\hole ^(0.5){\gamma ' } \ar@<1ex>[dd]^{\beta ' } \ar[rd]^(0.7){\big((1\otimes1)\cdot\varepsilon_0\big)\otimes e(a^*(j_0))\big ) } \\ & & \mathcal{m}_{s^n}\otimes\mathcal{m}_x\otimes e\big(a^*(s^n\times x)\big ) \ar@{>>}@/^3pc/[dd]^-{\gamma '' } \ar@<1ex>[d]^(0.6){\beta '' } \\
\mathcal{m}_y \ar[r]_-{\mathcal{h } } \ar@/^1pc/[ruu]^-{\phi_{h}}\ar[d]_{\eta_y } & \mathcal{m}_{s^n}\otimes\mathcal{m}_x\otimes \lambda(t , dt ) \ar[d]_{\eta ' } \ar[uu]^-{\alpha'}\ar[r]_{(1\otimes1)\cdot\varepsilon_0 } & \mathcal{m}_{s^n}\otimes\mathcal{m}_x \ar[u]^-{\alpha '' } \ar[d]_{\eta''}\\ a^*(y ) \ar[r]_-{a^*(h ) } & a^*(s^n\times x\times i)\ar[r]_{a^*(j_0 ) } & a^*(s^n\times x)}\ ] ] we have @xmath711 it follows that @xmath712 is a minimal model for @xmath180 .
a similar argument shows that @xmath713 is a minimal model for @xmath8 , where @xmath714 is the augmentation given by @xmath715 and @xmath716 . finally , we show that @xmath717 , as defined in the proof of theorem [ thm : equivalence of squares ] , is a homomorphism .
the key here is to translate addition in @xmath718 into addition of homotopy classes of affiliated maps .
denote by @xmath719_f$ ] the set of homotopy classes of affiliated maps @xmath180 that restrict to @xmath720 .
now suppose @xmath721 correspond to @xmath722_f$ ] .
then the sum @xmath723 corresponds to @xmath724_f$ ] , described as follows .
first consider the commutative diagram @xmath725 ^ -{i_2 } \ar[dl]^-{i_2 } \ar'[d]^-{f}[dd ] & & s^n\times x \ar[dl]^-{j_2 } \ar[dd]^-{g } \\ s^n\times x \ar[rr]_(0.4){j_1 } \ar[dd]_{f } & & p \ar[dd]^(0.3){(f\mid g)}\\ & y \ar[dl]_{1_y } \ar'[r]^-{1_y}[rr]\ar[dl]^{1_y } & & y \ar[dl]^{1_y } \\ y \ar[rr]_{1_y } & & y } \ ] ] in which the top and bottom squares are pushouts , and the front vertical map is induced as a map of pushouts from the back three .
the pushout @xmath407 can readily be identified with @xmath726 .
let @xmath727 denote the usual pinching comultiplication .
then we obtain the composition @xmath728 , which gives @xmath729 .
first consider the map @xmath734 .
as discussed in ( * ? ? ?
* sec.13(b ) ) , a dg algebra model for the pushout @xmath735 can be obtained as a pullback of minimal models .
specifically , by forming the pullback of minimal models @xmath736 ^ -{p_1}\ar[d]_{p_2 } & h^*(s^n;\q)\otimes\mathcal{m}_x \ar@{>>}[d]^{\varepsilon\cdot1}\\ h^*(s^n;\q)\otimes\mathcal{m}_x\ar@{>>}[r]_-{\varepsilon\cdot1 } & \mathcal{m}_x,}\ ] ] we obtain a dg algebra @xmath737 quasi - isomorphic to the minimal model of @xmath407 .
further , the same discussion can readily be extended to show that in the above situation , a minimal model for the induced map @xmath738 can be obtained using the map induced on the pullback of minimal models .
now this induced map is the map @xmath739 given by @xmath740 . under the identification of the pullback @xmath741 with @xmath742 , the composition @xmath743 evidently has minimal model of the described form .
p. hilton , g. mislin , and j. roitberg , _ localization of nilpotent groups and spaces _ , north - holland publishing co. , amsterdam , 1975 , north - holland mathematics studies , no .
15 , notas de matemtica , no . 55 .
j. oprea , _
gottlieb groups , group actions , fixed points and rational homotopy _ , lecture notes series , vol . 29 , seoul national university research institute of mathematics global analysis research center , seoul , 1995 . | let @xmath0 be a based map of simply connected spaces .
the corresponding evaluation map @xmath1 induces a homomorphism of homotopy groups whose image in @xmath2 is called the @xmath3th evaluation subgroup of @xmath4 .
the @xmath3th gottlieb group of @xmath5 occurs as the special case in which @xmath6 and @xmath7 .
we identify the homomorphism induced on rational homotopy groups by this evaluation map , in terms of a map of complexes of derivations constructed using sullivan minimal models .
our identification allows for the characterization of the rationalization of the @xmath3th evaluation subgroup of @xmath4 .
it also allows for the identification of several long exact sequences of rational homotopy groups , including the long exact sequence induced on rational homotopy groups by the evaluation fibration . as a consequence ,
we obtain an identification of the rationalization of the so - called @xmath8-sequence of the map @xmath4 .
this is a sequence in general not exact of groups and homomorphisms that includes the gottlieb groups of @xmath5 and the evaluation subgroups of @xmath4 .
we use these results to study the @xmath8-sequence in the context of rational homotopy theory .
we give new examples of non - exact @xmath8-sequences and uncover a relationship between the homology of the rational @xmath8-sequence and negative derivations of rational cohomology .
we also relate the splitting of the rational @xmath8-sequence of a fibre inclusion to a well - known conjecture in rational homotopy theory . |
the theoretical description of multiphase flows is essentially based on analyzing the reponse of a cloud of dispersed particles of different size ranges in a fluid .
these particles constitute dynamic phases and hence a multiphase flow .
a widely used multiphase system is a bubble column which is a reactor where a discontinuous gas phase in the form of bubbles , moves relative to a continuous phase .
bubble columns have a wide range of applications in chemical industries , biotechnology or in nuclear reactors @xcite , @xcite , @xcite , @xcite , nucreac .
the transient behavior is important at the start - up of these systems and its analysis is essential in order to characterize the dynamic performance of the columns . among the phenomena that occur in these systems
void wave propagation mechanisms are of great importance since many transient and steady states are controlled by the propagation of these waves and , in this sense among others , the dynamic characterization of multiphase flows is essential for the prevention of instabilities .
the ( in)stability of bubbly flows which are characterized by almost uniformly sized bubbles , is usually described in terms of the propagation properties of void fraction and pressure disturbances caused by natural or imposed fluctuations of the rate of air supply @xcite , @xcite , @xcite .
bubble size , rise velocity , size distribution and liquid and bubble velocity profile have a direct bearing on the performance of bubble columns . however , most of the time the dispersion devices deliver a dispersed phase with a given size distribution .
the importance of the size distributions is only scarcely evaluated , most of the time by direct empirical trials and its influence on the global behavior has still to be studied . actually , to our knowledge and from the theoretical point of view
, it has not been yet well established whether the stability of the motion of a swarm of bubbles is different for monodispersed or polydispersed bubble flows .
the main objective of this work is to investigate the effects that size polydispersity might produce on the stability of a bubble column .
we shall introduce the effect of polidispersity through the drag force in the hydrodynamic equations , using a method based on statistical concepts and on a point - force approximation @xcite . as we shall see below
, the corrections on the drag force factor , @xmath1 , due to polydispersity depend only on the first three moments of a given particle size distribution and they also have an effect on several properties of kinematic waves . in particular , we found that size polydispersity enhances the stability of void waves by a factor which varies between @xmath2 as a function of frequency and for a particular type of bubble column . in this way our model predicts effects that might be verified experimentally but this
, however , remain to be assessed . to this
end the paper is organized as follows . in section 2
we briefly review a hydrodynamical model for bubbly fluids introduced by biesheuvel and gorissen @xcite .
next , in section 3 we consider a dispersion of spherical air bubbles of different radi in water and we calculate the effect of size polidispersity on the gain factor , mean bubble velocity , kinematic wave velocities as a function of void fraction , for different wave frequencies .
in this section we summarize the main ideas and steps behind the hydrodynamical model for bubbly fluids introduced in ref .
the equations of motion for a swarm of bubbles in a bubble column have been derived in the literature by using standard methods of kinetic theory to average over an ensamble or realizations of the flow @xcite , @xcite . in ref .
@xcite a dispersion of equally sized air bubbles in a water column where the bubbles are small enough to remain spherical through the whole system , is considered .
they assumed that the air can be taken as an incompressible fluid where no mass transfer is allowed between the bubbles and the water , which is assumed to be an incompressible newtonian liquid .
the conservation equation for the mean number density of the gas bubbles , @xmath3 , and the conservation equation for the mean bubble momentum , @xmath4 ( kelvin impulse ) , were obtained for this system @xcite , @xmath5 @xmath6 + \mathbf{\nabla } _
{ \mathbf{x}}\cdot \left [ n\left ( \frac{4}{3}\pi a^{3}\rho _ { g}\mathbf{v+i}_{l}\right ) \right ] \nonumber \\ & & -\mathbf{\nabla } _ { \mathbf{x}}\cdot \left ( \mathbb{t}_{g}+\mathbb{t}% _ { l}\right ) \nonumber \\ & = & n\mathbf{f}_{d}+n\frac{4}{3}\pi a^{3}(\rho _ { l}-\rho _ { g})\mathbf{g . }
\label{2}\end{aligned}\]]@xmath7 is the fluid impulse , @xmath8 and @xmath9 are the fluid stresses ; @xmath10 is the drag force exerted by the fluid on the bubble and @xmath11 stands for the gravity field ; @xmath12 @xmath13 denote , respectively , the mass densities of water and air .
@xmath14 stands for the liquid s viscosity . in order to describe the flow parameters of the bubble swarm , eqs .
( [ 1 ] ) and ( [ 2 ] ) should be expressed in terms of the volume fraction of bubbles ( or void fraction ) @xmath15 and their velocity field @xmath16 .
following ref .
@xcite we assume that the uniform flow of bubbles is along the axial direction of the column with a mean axial rise velocity @xmath17 , therefore , @xmath18 .
the effect of hydrodynamic interactions between the bubbles on the mean frictional force may be represented by introducing a function @xmath19 into @xmath17 in the form @xmath20 , @xcite .
the magnitude of the terminal velocity , @xmath21 , of a single bubble of radius @xmath22 in a stagnant liquid is given by @xcite @xmath23 , where @xmath24 @xmath25 is the drag force factor and experiments suggest that wallis , @xmath26
. the mean fluid impulse is modelled by @xmath27where @xmath28 takes into account the effect of the hydrodynamic interactions . according to ref .
@xcite an expression for @xmath29 that renders reliable results up to large values of @xmath30 is @xmath31 . since in a nonuniform bubbly flow the stress
@xmath32 @xmath33 play the role of an effective pressure , they also assume that the _ kinetic _ contribution , @xmath34 , is proportional to the effective density of the bubbles , @xmath35 , and to the mean square of their velocity fluctuations @xmath36 , @xcite . here
@xmath37 stands for the limit of closest packaging of a set of spheres and is close to the value 0.62 .
thus , @xmath38 . furthermore
, if the non - uniformity is the main cause of an additional transfer of bubble momentum and fluid impulse associated with stress , biesheuvel and gorissen @xcite postulate that such a contribution to the stress should be given by the force @xmath39 .
therefore , taking into account both contributions to the stress , @xmath40 , where @xmath41 is the one dimensional nonuniform flow velocity and @xmath42 is an effective viscosity .
on the other hand , the mean frictional force is enhanced by an effective diffusive flux of bubbles due to their fluctuating motion .
this effect is similar to an steady drag force acting upon each one of the bubbles and proportional to the mean number density gradient .
therefore , using ( [ 2 ] ) this force is represented by @xmath43 $ ] .
substitution of the above expressions into eqs .
( [ 1 ] ) and ( [ 2 ] ) leads to the following closed set of one - dimensional equations of motion for the bubbly flow in a zero volume flux reference frame , @xmath44 @xmath45 + \frac{% \partial } { \partial z}\left [ \rho _ { ef}(\varepsilon ) v^{2}\right ] -\frac{% \partial } { \partial z}\mathbb{t } \nonumber \\ & = & -c_{d}\varepsilon f_{0}\left ( v+\frac{\mu _ { e}(\varepsilon ) } { \varepsilon \rho _ { ef}}\frac{\partial \varepsilon } { \partial z}\right ) -\varepsilon \left ( \rho _ { g}-\rho _ { l}\right ) g. \label{15}\end{aligned}\ ] ] these equations may be rewritten in a laboratory reference frame by considering the mean axial velocity of the dispersion , @xmath46 , defined by @xmath47 . here
@xmath48 and @xmath49 are the mean bubble and fluid axial velocity in the laboratory reference frame .
note that due to the incompressibility of both , liquid and gas , @xmath46 is only a function of time .
therefore @xmath50 and a galileo transformation of eqs .
( [ 14 ] ) and ( [ 15 ] ) gives @xmath51 @xmath52 \nonumber \\ & & + \frac{\partial } { \partial z}\left [ \varepsilon \left ( \rho _ { g}u_{g}+% \frac{1}{2}\rho _ { l}m_{0}\left ( u_{g}-u\right ) \right ) u_{g}\right ] \nonumber \\ & & -\frac{\partial } { \partial z}\left ( -p_{e}+\mu _ { e}\frac{\partial u_{g}}{% \partial z}\right ) -\varepsilon \rho
_ { g}\frac{\partial u}{\partial t } \nonumber \\ & = & -c_{d}\varepsilon f_{0}\left ( \left ( u_{g}-u\right ) + \frac{\mu _ { e}(\varepsilon ) } { \varepsilon \rho _ { ef}}\frac{\partial \varepsilon } { % \partial z}\right ) -\varepsilon \left ( \rho _ { g}-\rho _ { l}\right ) g , \label{18}\end{aligned}\]]together with the incompressibility condition @xmath53 consider a quiscent equilibrium state of the dispersion described by @xmath54 .
the deviations from this state will be denoted by @xmath55 and @xmath56 .
linearization of eqs .
( [ 17 ] ) - ( [ 19 ] ) around the reference state yields the wave - hierarchy equation@xmath57
\nonumber \\ & & = -\left [ \left ( \frac{\partial } { \partial t}+c_{0}\frac{\partial } { % \partial z}\right ) \delta \varepsilon -\nu _ { \varepsilon } \frac{\partial ^{2}\delta \varepsilon } { \partial z^{2}}\right ] \label{20}\end{aligned}\]]with lower and higher - order wave velocities given by @xmath58 and @xmath59 ^{1/2}. \label{22}\]]here @xmath60 and @xmath61 $ ] .
the primes ( @xmath62 ) denote derivatives with respect to @xmath15 and evaluated at the unperturbed state @xmath54 . for relatively low radial frequencies the wave propagation
is described by a linearized burgers / korteweg - de vries equation@xmath63 \frac{\partial ^{2}\delta \varepsilon } { \partial z^{2}}+\tau _ { \varepsilon } \nu _ { \varepsilon } ( u_{g_{0}}-c_{0})\frac{\partial ^{3}\delta \varepsilon } { % \partial z^{3 } } , \label{25}\]]with a solution @xmath64 where @xmath65 is the frequency of the void wave and @xmath66 -% \frac{\nu _ { \varepsilon } \omega ^{2}\left ( u_{g}-c_{0}\right ) } { c_{0}^{3}}% \left [ \tau _ { \varepsilon } ( c^{+}-c_{0})(c_{0}-c^{-})+\delta _ { \varepsilon } % \right ] . \label{27}\]]in terms of these quantities the so called gain factor , @xmath67 $ ] , where @xmath68 denotes the real part and @xmath69 the distance between two impedance probes in the experiments to measure @xmath70 @xcite .
the method developed by tam @xcite uses the concept of randomness of the bubble cloud and derives equations describing the average properties of the fluid motion .
these averages are taken over a statistical ensemble of particle configurations . a slow viscous flow past a large collection of spheres of a given size distribution is considered to derive a particle drag formula free from empirical assumptions .
the result essentially replaces the disturbance produced by a sphere in low reynolds number flow , by that of a point force located at the centre of the sphere .
the correction drag force factor is given by @xmath71 c_{d } , \label{28}\]]where @xmath72 ^{1/2}}{(1 - 3c)}. \label{29}\ ] ] @xmath73 are the moments of the size distribution @xmath74 and @xmath75 .
since the terminal velocity of a bubble depends on @xmath76 , it is reasonable to assume that in the polydispersed case @xmath77 should be replaced by @xmath78 .
substitution of this asumption into eqs .
( [ 17 ] ) - ( [ 19 ] ) , carrying out the linearization procedure described in the last section and using the explicit expressions of @xmath79 , one can show that these quantities scale as @xmath80 . if these polydispersed quantities are substituted into eq .
( [ 27 ] ) , one obtains an expression for the polydispersed gain factor @xmath81 $ ] .
to compare the monodisperesed and polydispersed results on the gain factor , mean bubble velocity , kinematic wave velocities as a function of void fraction , we used the the following parameter values for an air - water bubble column , @xmath82 @xmath83 , @xmath84 @xmath85 , @xmath86 @xmath87 @xmath88 @xmath89 , @xmath90 @xmath89 , @xmath91 @xmath87 @xmath92 , @xmath93 . in fig .
1 we plot @xmath70 and @xmath94 vs. @xmath15 for different frequencies @xmath95 and for a log - normal distribution @xmath74 with average and dispersion @xmath96 , @xmath97 , respectively . ,
@xmath94 vs. @xmath15 for waves with frequencies of @xmath98 , @xmath99 and @xmath100 .
the liquid is stagnant and the parameter values are those given in section 4 . ]
note that for values @xmath101 @xmath102 the attenuation rate drops significantly for for instance , the per cent difference defined by @xmath103 @xmath104 . for the range both cases .
for instance , for a frequency of @xmath105 this difference ranges from @xmath106 per cent , whereas for a frequency of @xmath100 it varies in the interval @xmath107 per cent .
this means that stability is larger in about 23 per cent for the latter case , a change that is significant in bubble reactors @xcite .
the quantities @xmath108 and @xmath109 are plotted as functions of @xmath15 in fig .
the curve for @xmath110 is always between that for @xmath111 and @xmath112 . according to the whitham stability criterion @xcite , when @xmath113 @xmath114 the uniform flow is unstable .
this occurs for both distributions , however , for the monodispersed case it occurs for @xmath115 , whereas for the polydispersed case the system is stable up to a larger value of the void fraction , e.g. @xmath116 . and @xmath117 as functions of the void fraction @xmath15 for the same parameter values as in fig . 1 . ]
summarizing , in this work we have analyzed the effects of size polydispersity in several features of the void fraction waves and their stability properties .
we found that the presence of a size distribution reinforces the stability of the waves , as shown in figs . 1 and 2 .
furthermore , the per cent difference may be quantified by estimating @xmath103 @xmath118 , @xmath119 amounts to a maximum percentual difference of @xmath120 .
it is convenient to emphasize once again , that the hydrodynamic model used in this work @xcite is idealized in many aspects .
for instance , compressibility and hydrodynamic interactions between bubbles and with the boundaries , have not been taken into account .
however , given the complexity of these effects and of the system itself , the simple dimensional model proposed by biesheuvel and gorissen seems to be a good first step in modeling the complex behavior of a bubble column .
it also ilustrates how some of the methodology and concepts of kinetic theory and statistical mechanics may be used to deal with complex phenomena in engineering systems .
we should also mention that in this work we have assumed an initial polydisperse size distribution and the coalescence of bubbles has not been considered @xcite .
however , this remains to be assessed .
some other important effects remain to be considered as well , like the bubble - bubble interaction mechanisms.nevertheless the approach followed here by including the influence of the distribution through the drag effects , considering a mean field approach , is an attempt to set a first framework for the bubble size distribution incorporation to further studies .
our results reinforce this point of view in the sense that a description of a bubble column based on the concept of randomness of a bubble cloud and average properties of the fluid motion , may be a useful approach that has not been exploited in engineering systems .
y. mercadier , contribution ltude des propagations de perturbations de taux de vide dans les coulements disphasiques eau - air bulles .
thsis , universit scientifique et medicale et institut national polytechnique des grenoble , france , 1981 | the occurrence of swarms of small bubbles in a variety of industrial systems enhances their performance .
however , the effects that size polydispersity may produce on the stability of kinematic waves , the gain factor , mean bubble velocity , kinematic and dynamic wave velocities is , to our knowledge , not yet well established .
we found that size polydispersity enhances the stability of a bubble column by a factor of about @xmath0 as a function of frequency and for a particular type of bubble column . in this way our modelpredicts effects that might be verified experimentally but this , however , remain to be assessed .
our results reinforce the point of view advocated in this work in the sense that a description of a bubble column based on the concept of randomness of a bubble cloud and average properties of the fluid motion , may be a useful approach that has not been exploited in engineering systems . |
the current sn classification scheme is based on both the properties of the spectrum and of the light curve .
the basic properties of stripped envelope sne ( ia , b , c ) light curves ( lcs ) are well reproduced by considering the diffusion of the radioactive decay energy of @xmath0ni and @xmath0co into homologously expanding sn ejecta ( arnett 1979 , 1980 , 1982 , 1996 - hereafter a79 , a80 , a82 , a96 ) .
core - collapse sne are divided into the type ib / c and type ii subclasses .
sn ib / ic are considered explosions of compact progenitors that have lost their outer hydrogen ( for sn ib ) and helium ( for sn ic ) envelopes .
a79 addressed what we would now call sn ib ; a80 and a82 pertained more directly to sn ia , but the principles are the same . for type ii sn ,
the following subtypes have been proposed : type iip , type iib , type iil and type iin .
sn iip explosions are the most common , and they are believed to be the result of the death of a massive red supergiant ( rsg ) progenitor star .
the lcs of sn iip are characterized by a long plateau that is indicative of the energy liberated by a recombination front in the extended hydrogen envelope of the progenitor star .
the late time decline rate of many sn iip lcs is consistent with that of the radiactive decay of @xmath0co .
arnett & fu ( 1989 ) presented a semi - analytical model that incorporates the effects of radiative diffusion with energy deposition from the radioactive decay of @xmath0co ( also pulsar and fallback energy ) plus h recombination in order to fit the observed lc of sn 1987a .
this model can produce a variety of sn iip - like lcs ( and probably sn iib , although this has not been explored in depth ) depending on the choice of h mass and the composition of the outer shell of the progenitor rsg star .
more accurate radiation hydrodynamics simulations of sn iip have been done yielding simultaneous lcs and spectra ( see examples in falk & arnett 1977 ; falk 1978 ; klein & chevalier 1978 ; kasen & woosley 2009 ; dessart , livne & waldman 2010 ; bersten et al . 2011 ; dessart & hiller 2011 ) as well as models of sn 1987a that have reproduced the observables in great detail ( blinnikov et al .
2000 ; dessart & hiller 2010 ) .
one - dimensional numerical lc models based on radiation hydrodynamics calculations are available for sn iil ( swartz et al . 1991 ; blinnikov & bartunov 1993 ) , but a simple semi - analytical approach , similar to that presented by a82 , a96 for the case of sn ia has not been provided . the simulations by swartz et al .
( 1991 ) were able to reproduce some of the characteristics of sn iil lcs under the assumption of o - ne - mg ( oxygen - neon - magnesium ) core - collapse , but the fits were not good for all phases of the observed lcs .
the lcs of sn iin present a diversity when it comes to shape , peak luminosity , duration and decline rate .
some of the recently discovered super - luminous supernovae ( slsne ) are classified as sn iin events , mainly due to the spectroscopic signatures of ejecta - csm interaction ( schlegel 1996 ) .
although there are numerical simulations of the interaction of sn ejecta with the csm and model lcs under optically - thin conditions ( falk & arnett 1977 ; van marle et al .
2010 ) the only attempt made so far to incorporate the effects of radiative diffusion of shock - deposited energy in an optically - thick csm envelope was done in the case of red supergiant ( rsg ) progenitors surrounded by a csm with a limited set of density profiles ( @xmath1 ) ( moriya et al .
2011 ) . the work done by moriya et al .
( 2011 ) using the 1-d radiation hydrodynamics code stella ( blinnikov & bartunov 1993 ; blinnikov et al . 1998 ; blinnikov et al .
2006 ) reproduced the lc of ultraviolet - rich sn iip sn 2009kf reasonably well and provided a tool to study the lcs of low - luminosity sn iin powered by the interaction of the sn ejecta with a cs wind . for higher luminosity events in the regime of slsne ,
moriya et al . (
2011 ) find that extraordinary mass - loss rates are required that are inconsistent with an rsg progenitor .
this and other factors motivate us to update the analytical models of a80 , 82 and determine approximate lcs for slsne powered by optically - thick sn ejecta - csm interaction for a variety of csm and progenitor characteristics .
this is important in order to produce a physically acceptable model for a slsn iin lc for which there is a smooth rise to maximum light ; a simplified model of instantaneous shock heating yields a lc that increases steeply to maximum in a very short time - scale . the large parameter space that is connected with sn iin ( properties of the sn progenitor , the sn ejecta and of the csm envelope )
provide a natural explanation for the observed diversity of sn iin lcs .
another motivation for developing an efficient approximate analytical model is to perform a qualitative study of this large parameter space . given that many of the recently discovered slsne have been classified as sn iil or sn iin events , a unified " model is sought in order to explain their properties ; at least the properties of their csm environments and ideally of their progenitors as well .
one of the interesting aspects of the slsn events that seem to be surrounded by dense , optically - thick shells is that the nature of the underlying supernova is hidden and thus remains obscure .
there seems to be too much mass to be associated with any underlying white dwarf thermonuclear explosion , but that leaves many unanswered questions .
what are the progenitor mass and radius , the ejecta composition and energy ? what is the explosion mechanism , core - collapse or something else ?
the collective enigmatic nature of these super - luminous events means that we have to return to rudimentary studies to explore the parameter space that may be appropriate .
for this reason we have generalized the basic light curve models of a80 , 82 using the first law of thermodynamics and the diffusion approximation .
where arnett first proposed his models in the context of radioactive decay , we have adapted his technique for an arbitrary prescription of the power input .
this allows us to explore radioactive decay , shell - shocks , magnetars and , within some limitations , combinations of these power inputs .
these basic models allow us to address questions such as : what is needed to make the light curve rise have a certain shape , and , independently , what shapes the decline ; what is needed to generate a nearly symmetric light curve , or a very asymmetric one ; what is the effect of the initial radius of the supernova progenitor ; what are the constraints on the density profiles in the supernova ejecta and the circumstellar matter ?
the most general question we propose to explore is whether or not the heterogeneity of the current sample of slsn requires very different progenitors and physics , or if there is some common theme expressed in different ways .
the goal of this study is not to compete with more realistic radiation hydrodynamics simulations but rather complement them by providing approximate solutions that will help us understand the importance of the parameters involved in the case of slsne .
benchmarking against more accurate , numerical results is discussed in this work as a means to illustrate the uncertainties and the limitations of our model .
we organize the paper as follows : in 2 we present the general physical assumptions of the basic lc model and we provide solutions in the cases of homologously expanding material and fixed sn photosphere . in
3 we present a variety of physically - motivated luminosity inputs and we develop model lcs for shock energy deposition resulting from sn ejecta - csm interaction incorporating the effects of radiative diffusion .
we also develop a hybrid lc model with ejecta - csm interaction plus @xmath0ni and @xmath0co radioactive decay . in
4 we provide a characteristic fit of the hybrid ejecta - csm interaction radioactive decay input to the lc of the slsn 2006gy and discuss the implications of this model for the nature of the event . applications to other slsne will be presented in a subsequent paper .
finally , in 5 we summarize our conclusions .
details of the derivations of the analytical models are given in appendices a and b.
following the prescriptions of a80 , 82 for type i sn lcs , we present a solution for a general heating input .
the initial assumptions for our model remain the same as in a80 , 82 , those being : 1 ) homologous expansion of the ejecta 2 ) centrally located power input source 3 ) radiation pressure dominant . under these assumptions
we consider the first law of thermodynamics : @xmath2 where @xmath3 is the specific internal energy , @xmath4 is the pressure , @xmath5 is the specific volume where @xmath6 is density , @xmath7 is the specific input energy generation rate , @xmath8 the output radiated luminosity and @xmath9 is the mass coordinate of the fluid element . in general , the temperature profile of the diffusion mass , @xmath10 , @xmath7 and @xmath8 are functions of position , @xmath11 , and time , @xmath12 , where @xmath13 is the dimensionless position coordinate for a fluid element relative to a fiducial radius taken to be the radius of the photosphere . for homologous expansion @xmath14 where @xmath15 is
the initial radius of the ejecta at the moment of shock breakout and @xmath16 is the characteristic expansion velocity of the ejecta .
the velocity @xmath16 is not necessarily the photospheric expansion velocity , @xmath17 , as measured from sn spectra , but we can use @xmath18 as an approximation in some cases . for the output luminosity we use the radiation diffusion approximation @xmath19 where @xmath20 is the mean free path with @xmath21 being the mean opacity that we take to be a constant , @xmath6 the density , and @xmath22 the speed of light . in the following analysis
we alter the two first criteria of a80 , 82 by considering @xmath23 0 ( fixed photospheric radius instead of homologous expansion ) and input sources that terminate due to their movement through the diffusion mass such as a forward shock that breaks out from an optically - thick csm envelope .
the assumption of constant opacity is a weakness of our analytical approach , but it was necessary in order to obtain a separable pde . in more realistic situations that are accounted for in radiation hydrodynamics models opacity is depth and time dependent .
we note that while moriya et al .
( 2011 ) use their radiation hydrodynamics code to compute the ionization state and opacities of the underlying explosion , they adopt a thomson electron scattering opacity for fully ionized solar metallicity material ( @xmath24 @xmath25 g@xmath26 ) within the ionization front of the csm .
we will use the same value for the optical opacity throughout this work .
this general lc model has been considered and expanded by blinnikov & popov ( 1993 ) for the case of a piecewise constant opacity in power - law density distributions for the sn ejecta .
as pointed out in their paper , the correct approach to the subject belongs to the class of moving boundary problems and not the usual eigenvalue formulation that has been considered so far for the analytical models .
a different approach has also been considered by popov ( 1995 ) where he adopts a mixed boundary condition for radiation ( not the radiative zero solution that is considered in the a80 , a82 , a96 models ) and obtains solutions for general heating terms in the form of green s functions but specifically focusing attention on radioactive heating .
as shown in appendix a , the general full solution for the output luminosity from the photosphere of the sn ejecta can be written as : @xmath27 } \int_0^t e^{[\frac{t'^{2}}{t_{d}^{2}}+\frac{2r_{0}t'}{vt_{d}^{2 } } ] } f(t ' ) [ \frac{r_{0}}{vt_{d}}+\frac{t'}{t_{d}}]dt ' + \frac{e_{th,0}}{t_{0 } } e^{-(t^{2}/t_{d}^{2}+2r_{0}t / vt_{d}^{2})},\ ] ] where we have introduced the effective light curve time scale , @xmath28 , and where @xmath29 is the diffusion time scale , @xmath30 is the initial central specific volume of the ejecta , @xmath31 is a constant arising from the separation of variables of the lc pde , @xmath32 is the expansion time - scale , @xmath33 is the initial luminosity input ( see appendix a ) , @xmath34 the total sn explosion energy and @xmath35 is a constant that accounts for the density profile of the diffusion mass .
a80 , 82 adopt @xmath36 13.8 as a good approximation for a variety of diffusion mass density profiles . the term @xmath37 ( see equation a9 in appendix a ) can be written as @xmath38 or @xmath39 ( as we have done in equation 3 ) as convenient .
for @xmath40 the additive term @xmath41 that appears in equation 3 vanishes and the solution for small initial radius is recovered , as presented in a80 , 82 .
arnett & fu ( 1989 ) and a96 also implicitly present this general solution and apply it to the case of the lc of sn 1987a for several power inputs ( pulsar heating , swept - up luminosity and fallback ) . the second term in equation 3
is an initial value term related to the internal energy , @xmath34 that the sn possesses at a given time . in equation 3 this term
is governed by the energy the ejecta have at time @xmath42 0 that they gained from the sn blast wave that subsequently diffuses from the optically - thick expanding envelope . in a80 , this term is used to explain sn iip lcs , for which the csm shock input energy from radioactive @xmath0ni and @xmath0co decay is small compared to the total energy of the ejecta at early times . due to the fast exponential decay
this fireball " term may be ignored for our purposes since it affects the lc only at very early times .
we note that this term does play a role in models for which the input ceases and diffusion controls the luminous output . in this context
, the coefficient of the exponential term becomes @xmath33 , the initial value of the luminosity at the time the power input ceases , after which the luminosity decays according to the exponential diffusion term ( 3.1 , 3.2.1 ) . in
3 we will explore several physical luminosity inputs to obtain the final generalized model lcs .
next , rather than homologous expansion , we assume @xmath43 , @xmath44 ( and thus @xmath45 ) for a fixed photosphere radius .
the motivation for this is the fact that around massive sn progenitors there can be an optically - thick csm shell , so that when the sn ejecta collide with that surrounding medium , a forward shock is formed that propagates into the nearly stationary csm and a reverse shock propagates into the ejecta , both depositing kinetic energy and heating the interacting media .
these shocks provide a natural source for the output luminosity of the event ( chevalier 1982 ; chevalier & fransson 1994 ; chugai & danziger 1994 ) .
although the diffusion problem that we solved above involves an homologously expanding sn photosphere , we assume the same principles hold for a stationary photosphere and diffusion mass .
this assumption modifies the final pde that we solve ( see appendix a ) and leads to the following general solution : @xmath46 where @xmath34 is , again , the initial internal energy from the sn blast wave that affects the lc at early times and that we ignore in our analysis . for small initial radius ( @xmath40 )
that term also goes to zero .
having the general solutions for the output sn lcs provided by equations 3 , and 4 for homologously expanding and fixed photosphere , respectively , we now consider several physical heat inputs as the source functions " for the integrals of these expressions in order to obtain the final output lc models . from the form of these general solutions it follows
that , in order to obtain a physical lc that rises smoothly to maximum and subsequently declines , the luminosity deposition function must be a smooth continuous function that does not monotonically increase .
for a constant or monotonically rising power input , the output luminosity will always increase with time with a monotonically increasing slope until the heat input vanishes at some point . for a monotonically declining input , the luminosity will increase to a maximum value and then decline , dominated by diffusion and cooling . as shown in 3.3 , plausible input sources can rise and then decline . at very late times
, the output luminosity will be the same as the input luminosity as the diffusion time becomes short compared to the total elapsed time .
this property is used as a diagnostic for radioactive decay of @xmath0ni and @xmath0co for some sn lcs where late - time photometric observations indicate a decline rate consistent with that of @xmath0co ( colgate & mckee 1969 ; colgate et al .
the same is expected to be the case for other continuous power inputs .
we may consider two general categories of power inputs : centrally located ones and moving ones .
a well - known input is the radioactive decay of @xmath0ni and @xmath0co that is believed to power the sn ia , sn ib / c and some sn ii lcs ( a80 , a82 ) , particularly the maximum of sn 1987a and sn iib .
although there can be some outward mixing , the newly - formed nickel is often taken to be confined near the center of the sn ejecta , around the core of the progenitor star .
this assumption agrees well with some simulations of sn ia , sn ib / c and sn ii events ( nomoto et al .
another centrally located input that has been considered for sne is radiation from a magnetic dipole associated with a pulsar ( ostriker & gunn 1971 ) as applied in the case of sn 1987 by arnett & fu ( 1989 ) or a magnetar as proposed recently as the power source for the slsne 2007bi and 2008es ( kasen & bildsten 2010 ; woosley 2010 ; see also maeda et al .
2007 ) .
moving inputs include the energy released in sn ejecta due to the recombination of h and he and shock heating .
arnett & fu ( 1989 ) presented a semi - analytical model that takes the effects of h recombination into account in order to reproduce the lc of sn 1987a .
the recombination front recedes into the expanding sn ejecta .
the effect of that is the creation of a plateau phase in the sn lc that is more pronounced for large initial radius and large mass of the h envelope . the analytical model presented in arnett & fu ( 1989 ) can therefore be used to reproduce some sn iip lc characteristics .
analytical models for sn iip lcs have also been provided by popov ( 1993 ) and kasen & woosley ( 2009 ) .
shock heating can be another important moving power source for some sn lcs .
sn iin show evidence for csm interaction in their spectra with prominent h and sometimes he emission features indicating the presence of a shock running into the csm and depositing kinetic energy ( chevalier 1982 ; chevalier & fransson 1994 ) .
a significant fraction of the recently discovered slsne are sn iin and show signs of csm interaction , so an analytical model that describes the output light curve that results from this power input might be helpful in understanding the basic properties of the csm involved in the process .
a peculiarity of this input is that its dynamics depend on the physical characteristics of the interacting media ( the sn ejecta and the csm shell ) .
in addition , it is not a continuous input ; once the reverse shock sweeps up all the available ejecta mass or the forward shock sweeps up all the available csm mass there is no further heating and all that is left is just the diffusion of the previously shock - deposited energy .
incorporating forward shock heating into an optically - thick csm envelope requires a different treatment in order to account for the movement of the input source , and the resulting pde is not formally separable . for the scope of this work
, we will assume that reverse and forward shock heating are both centrally located , but that they terminate ; forward shock heating terminates when the shock breaks out of the csm and reverse shock heating when the available sn ejecta mass has been swept - up .
with these assumptions , the pde becomes separable ( appendix b ) .
the assumption of centrally located power source for the case of the forward and the reverse shocks , although convenient , is not generally true and thus increases the uncertainties and limitations of this approximate model .
for this reason in 3.3.4 we compare our results with numerical results presented by chugai et al .
( 2004 ) , woosley , blinnikov & heger ( 2007 ) and moriya et al .
( 2011 ) for the same initial conditions and discuss the differences .
these approximate solutions give useful guidance , as we will see , but a proper solution calls for numerical radiation hydrodynamics .
a combination of power sources , centrally located and moving / terminated , is probably involved for most core - collapse sn .
we consider such a case by developing a hybrid radioactive decay and csm interaction diffusion model in 3.3.3 .
we start by considering the simple initial value problem where the initial luminosity @xmath33 is introduced at @xmath47 at the sn photosphere .
this problem was also considered by a80 and deals with the resulting sn lc if the energy input were just the initial sn shock energy for which @xmath48 in equation 3 .
this input produces a light curve with instantaneous rise to maximum light at @xmath49 followed by simple diffusion .
a96 also presented solutions of this expanding fireball " problem that has the following form in the homologous expansion case : @xmath50}.\ ] ] in the fixed photosphere case , this reduces to : @xmath51 these results are the same in form as the second terms of equations 3 and 4 for homologous expansion and fixed photospheric radius , respectively .
pure diffusion on the decline has the same form for any luminosity input that vanishes after some time interval
. similar diffusive decay from an initial value of luminosity at a given time pertains to models in which the power input is truncated ( 3.2.1 , 3.3.2 ) .
this simple solution has been considered by some authors as an interpretation for the decline of some slsne ( for an example , see smith & mccray 2007 for the lc of sn 2006gy ) .
model lcs powered by initial value luminosity input are shown in figure 1 for the cases of homologously expanding matter and stationary photospheres .
this model does not provide a natural explanation for a lc with a smooth extended rise .
centrally located power inputs such as the radioactive decay of @xmath0ni and @xmath0co as well as magnetar spin - down have been considered in previous work . in the following subsections
we summarize those past results in order to use them for comparison with the models that we will introduce . motivated by the fact that an instantaneous shock luminosity input does not provide an output sn lc with a smooth rise to maximum light , we consider a top - hat " model in which a constant input is provided for a finite time and then shuts off . to be specific , we consider the input to be a shock with constant luminosity , @xmath52 lasting for time , @xmath53 , and thus producing total energy @xmath54 .
therefore , the input shock luminosity that is provided to the expanding diffusion mass is @xmath55 for @xmath56 and @xmath57 otherwise .
this is the input luminosity function that we insert into equations 3 and 4 to calculate the output lc in the cases of homologously expanding matter and stationary matter with a photosphere of fixed radius , respectively .
we also assume that the shock luminosity deposition takes place deep within the diffusion mass , therefore we neglect any corrections attributable to the movement of the shock towards the photosphere .
direct integration yields @xmath58 , & $ t < t_{sh}$ , \cr \frac{e_{sh}}{t_{sh } } e^{-(t^{2}/2t_{d}^2 + 2r_{0}t / v t_{d}^{2 } ) } [ e^{(t_{sh}^{2}/2t_{d}^2+r_{0}t_{sh}/v t_{d}^{2})}-1 ] , & $ t >
t_{sh}$,}\ ] ] in the case of homologously expanding matter and @xmath59 , & $ t < t_{sh}$ , \cr \frac{e_{sh}}{t_{sh } } e^{-t / t_{0 } } [ e^{t_{sh}/t_{0}}-1 ] , & $ t > t_{sh}$,}\ ] ] in the case of fixed photospheric radius .
examples of model lcs for those two cases within the context of this type of luminosity input are shown in figure 2 . in this type of model the energy that powers the sn light curve is produced by the diffusion of shock - generated energy through an optically - thick csm shell of large initial radius .
smith & mccray ( 2007 ) adopted a l@xmath60 rise for their model of sn 2006gy based on the early portion of the diffusion models of a82 .
they did not self - consistently consider the input necessary to drive such a rise .
this constant power model is related to the top hat " magnetar - input model that is considered in kasen & bildsten ( 2010 ) with the exception that we solve for the general case of large initial radius and the diffusion time is defined somewhat differently ( see appendix a ) .
we note that in the case where @xmath15 is small , the kasen & bildsten result for small initial radius is recovered . while not especially realistic , this top - hat " model captures the essence of the shell - shock model on both the rise and decline .
diffusion of the radioactive decay energy of newly synthesized @xmath0ni and @xmath0co has long been considered the source for powering the lcs of sn ia , sn ib / c and some sn ii .
a79 , a80 , a82 were the first to analytically solve this problem and later works by valenti et al .
( 2008 ) and chatzopoulos , wheeler & vinko ( 2009 ) provide the generalized mathematical expressions for the output lc including the contribution due to nickel decay and positron heating ( assuming the same deposition function ) as well as the effects of gamma - ray leakage .
the output luminosity in this case is found to be : @xmath61}[(\epsilon_{ni}-\epsilon_{co } ) \int_0^t [ \frac{r_{0}}{vt_{d}}+\frac{t'}{t_{d}}]e^{[\frac{t'^{2}}{t_{d}^{2}}+\frac{2r_{0}t'}{vt_{d}^{2}}]}e^{-t'/t_{ni}}dt ' \nonumber \\ + \epsilon_{co } \int_0^t [ \frac{r_{0}}{vt_{d}}+\frac{t'}{t_{d}}]e^{[\frac{t'^{2}}{t_{d}^{2}}+\frac{2r_{0}t'}{vt_{d}^{2 } } ] } e^{-t'/t_{co}}dt'](1-e^{-at^{-2}}),\end{aligned}\ ] ] here @xmath15 is the initial radius of the progenitor , @xmath62 is the initial nickel mass , @xmath63 8.8 days , @xmath64 111.3 days , @xmath65 erg @xmath66 @xmath67 and @xmath68 erg @xmath66 @xmath67 are the energy generation rates due to ni and co decays respectively ( valenti et al . 2008 ) .
the factor @xmath69 accounts for the gamma - ray leakage , where large @xmath70 means that practically all gamma rays and positrons are trapped .
the gamma - ray optical depth of the ejecta is taken to be @xmath71 , where @xmath72 is the gamma - ray opacity of the sn ejecta .
within this model , for which the time - scales of the power inputs are known , the following expression for the mass of the sn ejecta can be obtained : @xmath73 where @xmath74 is the mass of the sn ejecta , and @xmath35 is an integration constant equal to about 13.8 .
this equation for the diffusion mass ( the mass of the sn ejecta in this case ) holds for both homologously expanding matter and fixed photosphere , but only under the assumption of a centrally located power input .
if the power input source moves within the diffusion mass , @xmath75 is itself time - dependent and the diffusion mass also changes with time , depending on the position of the source relative to the that of the photosphere .
although this model can provide good formal fits to observed lcs of some slsne , it can not be adopted as a general explanation due to the fact that in many cases the derived @xmath62 is greater than @xmath74 for reasonable choices of @xmath21 .
the usual assumption of a radioactive decay diffusion model is homologous expansion of the sn ejecta .
we alter this criterion and consider also the solution for a lc powered by the radioactive decays of @xmath0ni and @xmath0co , but for the case of fixed photospheric radius
. this may be the case for some sn iin progenitors that are surrounded by a dense extended optically - thick csm envelope .
in such case , the photosphere of the relative diffusion mass is coincident with the photosphere of the optically - thick csm which , in principle , does not expand homologously or moves with a velocity small compared to the velocity of the sn ejecta .
although those types of events are primarily powered by shock deposited energy there may still be some contribution from radioactive decay . in this case , we make use of equation 4 for a radioactive @xmath0ni and @xmath0co decay input and we arrive at the following solution : @xmath76(1-e^{-at^{-2}}).\end{aligned}\ ] ] recently , kasen & bildsten ( 2010 ) and woosley ( 2010 ) considered models of slsn lcs powered by the spin - down of a young magnetar ( see also ostriker & gunn 1971 ; arnett & fu 1979 ; maeda et al .
2007 ) . in such a model ,
the energy input by the magnetar is given by the dipole spin - down formula : @xmath77 where @xmath78 is the initial magnetar rotational energy , @xmath79 is the characteristic time scale for spin - down that depends on the strength of the magnetic field and @xmath80 2 for a magnetic dipole . for a fiducial moment of inertia ,
the initial period of the magnetar in units of 10 ms is given by @xmath81erg / s /@xmath82 .
the magnetic field of the magnetar can be estimated from @xmath83 and @xmath79 as @xmath84 , where @xmath85 is the magnetic field in units of @xmath86 g and @xmath87 is the characteristic time scale for spin - down in units of years . incorporating the magnetar spin - down deposition function presented here into equation 3 and including the effects of large initial radius , we arrive at the following solution : @xmath88 } \int_{0}^{x } e^{[z^{2}+\frac{r_{0}z}{vt_{d } } ] } \left[\frac{r_{0}}{vt_{d}}+z\right ] \frac{1}{(1+yz)^{2}}dz,\ ] ] where @xmath89 and @xmath90 with @xmath91 again being an effective " diffusion time . as was the case for the radioactive decay diffusion model , the mass of the sn ejecta for this model
is also given by equation 10 .
figure 3 shows an example of a radioactive decay diffusion model lc ( solid black curve ) compared with a magnetar spin - down model lc ( solid red curve ) where the parameters have been chosen so that the output lcs have approximately the same peak luminosity .
the parameters that were used for the models presented in figure 3 are @xmath92 10 @xmath93 , @xmath94 @xmath95 erg , @xmath79 = 30 days and for the homologously expanding case for @xmath96 40 days , @xmath97 cm and @xmath98 10,000 km s@xmath26 .
whereas the decay times of @xmath0ni and @xmath0co are known experimentally , the magnetar spin - down model contains two adjustable time - scales , @xmath79 and @xmath91 , instead of just @xmath91 as is the case for the radioactive decay diffusion model .
the magnetar model therefore provides more freedom for fitting to observed sn lcs .
sn iin and the recently discovered slsne show a variety of lc characteristics in terms of maximum luminosity , duration , shape and decline rate .
motivated by that , we attempt to use the ejecta - csm interaction scenario as introduced by chevalier ( 1982 ) and chevalier & fransson ( 1994 ) coupled with diffusion as treated by a80 , 82 in order to obtain model lcs for this type of events .
we adopt the scenario introduced by chevalier & fransson ( 1994 ) in which the progenitor star is embedded in a csm shell described by a power - law density profile @xmath99 , where @xmath100 is the density of the csm medium , @xmath101 is a scaling constant and @xmath102 the power - law exponent for the csm density profile . in general ,
@xmath103 , where @xmath104 is the density of the csm shell at @xmath105 .
we use as a fiducial value for @xmath106 where @xmath107 the radius of the progenitor star .
thus we set the density scale of the csm , @xmath104 , immediately outside the stellar envelope . for @xmath108 ,
a steady - wind csm model is recovered , where @xmath109 with @xmath110 the pre - sn wind mass - loss rate and @xmath111 the pre - sn wind velocity .
deviations from @xmath112 2 may indicate a different pre - sn mass - loss history resulting in csm clumps or shells .
values of @xmath102 close to zero may indicate the presence of csm bubbles " or shells formed by strong stellar winds as proposed for the csm environments around wolf - rayet stars ( chevalier & liang 1989 ; dwarkadas 2011 ) .
the density profile of the sn ejecta is taken to have double power - law profile where the outer ejecta have a power - law density profile , @xmath113 , where @xmath114 is a scaling parameter for the ejecta density profile , @xmath115^{(n-3)/2}/[(3-\delta)(n-3)m_{ej}]^{(n-5)/2}$ ] , @xmath116 is the power - law exponent of the outer component , and @xmath117 is the slope of the inner density profile of the ejecta ( values of @xmath118 0 , 2 are typical ) , @xmath119 is the total sn energy and @xmath74 is the total sn ejecta mass .
this assumption for the evolution and the density profile of the sn ejecta is supported by realistic numerical calculations .
the parameter @xmath116 varies depending on the nature of the progenitor and the presence or absence of a convection zone in the outer parts of the star .
a fiducial value for @xmath116 is 11.7 corresponding to the case for red supergiant progenitors ( matzner & mckee 1999 ) whereas lower values of @xmath116 correspond to the envelopes of more compact progenitors .
the interaction between those two media , the sn ejecta and the csm , each with power - law density profiles results in a forward / circumstellar and a reverse / ejecta shock , the dynamics of which are described by self - similar solutions presented in chevalier & fransson ( 1994 ) .
we use those similarity solutions to derive the following expression for the luminosity that is produced from this process ( neglecting the second , initial value , fireball " term in equation 3 ; see appendix b for the full derivation ) : @xmath120 where @xmath121 , @xmath122 and @xmath70 are constants that depend on the values of @xmath116 and @xmath102 and , for a variety of values , are given in table 1 of chevalier ( 1982 ) , @xmath123 , @xmath124 denote the heaviside step function that controls the termination of the forward and reverse shock respectively ( @xmath125 and @xmath126 are the termination time scales for the two shocks ) and @xmath127 is the initial time of the csm interaction that sets the initial value for the luminosity produced by shocks where @xmath128^{\frac{1}{2}}/x_{0}$ ] is the characteristic velocity of the sn ejecta , where @xmath129 is the dimensionless radius of the break in the sn ejecta density profile from the inner flat component ( described by @xmath117 ) to the outer , steeper component ( described by @xmath116 ) which is at radius @xmath130 .
the first and the second terms in equation 14 refer to the forward and reverse shock luminosity input , respectively .
the forward shock termination time - scale , @xmath125 , is given by the following expression , assuming that the input from the forward shock terminates when all the available csm has been swept up : @xmath131^{(s-3)/(n - s)}}{4 \pi \beta_{f}^{3-s}}|^{\frac{n - s}{(n-3)(3-s ) } } m_{csm}^{\frac{n - s}{(n-3)(3-s)}},\ ] ] where @xmath132 is the total mass of the csm .
once the forward shock breaks out from the optically - thick part of the csm envelope , the luminosity deposition from it is primarily in the uv / x - ray region of the spectrum , not the optical . the time @xmath125 can represent the breakout time if @xmath132 is taken to be the mass of the optically - thick csm , rather than the total mass .
we adopt this assumption below . assuming that the reverse shock terminates when all the ejecta are swept - up , the reverse shock termination time - scale , @xmath126 , is given by the following expression : @xmath133^{\frac{n - s}{s-3}}.\ ] ] where @xmath74 is the total ejecta mass . to verify the analytical result of equation 14 , we compare it to the lc that is calculated by a two - dimensional simulation of the collision of sn ejecta with a wind csm component ( @xmath108 ) as presented by van marle et al .
we adopt the same parameters for the csm and the sn ejecta components as in their model o01 ( @xmath134 11.7 , @xmath112 2 , @xmath135 200 km s@xmath26 and @xmath136 @xmath137 @xmath93 yr@xmath26 ) and present the results in figure 4 .
the agreement between our analytical result ( red solid curve ) and their simulation ( filled circles ) is remarkable given that they have used a different technique to calculate the radiated luminosity in their simulation . the spectrum produced by optically - thin ejecta - csm interaction , is expected to be a hard spectrum populated with emission lines produced by forward and reverse shocks ( chevalier & fransson 1994 ; nymark , chandra & fransson 2009 ) .
spectra with these characteristics have not been seen in slsne so far , so it is natural to assume that the effects of radiative diffusion of the shock - generated luminosity are important in explaining both the currently observed optical light curves and optical spectra of these events .
for this reason we develop an analytical model that couples the effects of a csm shock heat input as derived in the previous section , with the prescription of radiative diffusion as treated by a80 , 82 .
we will consider the case of a stationary csm photosphere for this problem and assume that the photosophere is somewhere in the csm envelope where it will not , in principle , be rapidly expanding . in these models
we ignore backwarming " of the photosphere by any shock that has propagated beyond the csm photosphere .
we include the effects of terminated shock luminosity input .
we start by assuming that the luminosity input from the forward and the reverse shock happens deep within the fixed photosphere of the csm and therefore we treat it as centrally located .
this assumption coupled with the nature of the csm we explore means that the typical shock crossing time scale , @xmath138 , is larger than the effective radiation diffusion time scale @xmath91 .
this statement is equivalent to the condition that the csm optical depth is smaller than the characteristic optical thickness of a radiation - dominated and radiation - mediated shock , @xmath139 , and thus in the regime considered by nakar & sari ( 2010 ) who calculate shock breakout lcs for a variety of sn progenitors ( see also ensman & burrows 1992 ) .
chevalier & irwin ( 2011 ) investigated situations in the opposite regime , @xmath140 , for a steady state wind ( @xmath112 2 ) , considering the effect of the propagation of a radiation - dominated forward shock into the mass - loss region for the cases where the characteristic radiation breakout radius ( @xmath141 , as defined in their work ) is either smaller or larger than the wind termination radius ( @xmath142 ) . in chevalier & irwin ( 2011 ) , diffusion occurs when the forward shock reaches @xmath141 and @xmath143 .
subsequently , the diffusion would proceed more rapidly than the shock .
this is the regime we consider .
we note that when the condition @xmath144 is satisfied , the shock is no longer radiation mediated .
ofek et al .
( 2010 ) used a related model to explain the lc of the iin sn ptf 09uj .
balberg & loeb ( 2011 ) presented a similar model , but for a less dense wind in the context of observational signatures of the uv / x - ray shock breakout lc .
both the forward and the reverse shocks move through the diffusion mass affecting the radiation diffusion time and changing the form of the output lc . accounting for the movement of the sources
makes the resulting pde unseparable and the problem hard to solve analytically , and we neglect this aspect in the current models .
we do , however , account for the shock propagation by terminating the shock luminosity input after the specific time scales that are given by equations 15 and 16 .
we assume that the forward shock input within the optically thick part of the csm terminates at a time given by equation 15 for @xmath145 , where @xmath146 is the mass of the optically - thick part of the csm .
this shock termination time is close to the time of forward shock breakout , @xmath147 , which formally occurs when the shock reaches optical depth @xmath148 .
the mass of the optically - thick csm is given by : @xmath149 and we use optical depth , @xmath150 to determine the radius of the photosphere , @xmath151 , where @xmath21 is the optical opacity and @xmath152 is the total radius of the csm determined by the following formula : @xmath153 where @xmath132 is the total mass of the csm envelope that we use as the basic parameter .
solving equation 19 with respect to @xmath152 , equation 18 for @xmath151 and setting the results back into equation 17 we obtain the final result for @xmath146 as an expression that depends on the basic model parameters @xmath132 , @xmath101 , @xmath102 , @xmath107 and @xmath21 . once the forward shock breaks out
, the bolometric luminosity will still invoke the full forward / reverse shock input given by equation 14 , but the optical luminosity will decline since the bulk of the energy will be emitted in the uv / x - ray region of the spectrum .
a portion of the optically - thin forward shock contribution is expected to be in the optical due to re - radiation and electron scattering ( chevalier & fransson 1994 ) , but we neglect this effect . with this model
we can also estimate the time and intensity of the rise of uv / x - ray flux due to forward shock breakout .
now we can implement the sn ejecta - csm interaction luminosity input from the forward and the reverse shock given by equation 14 into the first term of equation 4 that accounts for the diffusion through an optically - thick csm with a fixed photosphere and obtain the following final expression for the output model lc : @xmath154dt',\end{aligned}\ ] ] where , in the case of an optically - thick csm , @xmath155 .
after the times of the termination of the shock power inputs , the luminosity decays according to equation 6 where the initial value of the luminosity is now that for the corresponding terms in equation 20 at @xmath147 and @xmath126 , respectively . the analytic model described by equation 20 for the output lcs for sne powered by sn ejecta - csm interaction depends on the properties of the progenitor star ( @xmath119 , @xmath74 , @xmath107 , @xmath117 and @xmath116 ) and the properties of the csm ( @xmath102 , @xmath156 , @xmath21 and @xmath132 ) .
the fact that there are many unknown parameters involved in this problem follows from the complex nature of csm interaction due to the large variety of possible csm environments ( winds , shells of any density scale ) and the large variety of possible sn progenitors ( red or blue supergiant stars ) .
this natural variety can explain certain differences in the observed optical lc shape , duration and luminosity of sn iin . in the same context ,
the faster post - maximum decline of a sn iil lc can be the result of particular conditions in this large parameter space involved in sn ejecta - csm interaction , as we discuss next .
here we consider a hybrid model in which the luminosity input is provided by both sn ejecta - csm interaction and by the radioactive decays of @xmath0ni and @xmath0co . to find the combined luminosity input we add the radioactive decay luminosity input to the forward and
reverse shock luminosity input given by equation 14 .
figure 5 shows an example of this hybrid luminosity input for collision of sn ejecta ( @xmath157 erg , @xmath158 20 @xmath93 , @xmath159 cm , @xmath134 12 ) with an optically - thick massive csm shell ( @xmath112 0 , @xmath160 1 @xmath93 ) in the case where 1 @xmath93 of @xmath0ni is formed in the explosion , which might be the case for a hypernova or a pair - instability sn ( pisn ) .
the early behavior of the input is controlled by the relative contribution between the shock inputs and the radioactive decay , and the very late - time decay rate is equal to the @xmath0co radioactive decay rate . in the cases where
csm interaction is dominant , the early decline of the input simply scales as a power law function of time with the power being a function of the values of @xmath116 and @xmath102 as given by equation 14 . to obtain a model bolometric lc for this hybrid input that can be used for fits to observed sn lcs
we assume that the radioactive decay deposition takes place within the whole diffusion mass ( @xmath161 ) .
we assume the csm interaction luminosity input takes place just within @xmath146 , since in this context we assume the shocks to be frozen " at the interaction region .
we take the interaction region to be the interface between the edge of the progenitor star and the csm envelope and deep within the csm photosphere .
therefore , the final hybrid csm interaction plus radioactive decay lc model has the following form : @xmath162dt'+\frac{1}{t_{0 } ' } e^{-\frac{t}{t_{0 } ' } } \int_0^t e^{\frac{t'}{t_{0 } ' } } m_{ni } \nonumber \\ \times [ ( \epsilon_{ni}-\epsilon_{co})e^{-t'/t_{ni } } + \epsilon_{co}e^{-t'/t_{co}}]dt',\end{aligned}\ ] ] where @xmath75 and @xmath163 correspond to the diffusion time - scales through @xmath132 and @xmath161 respectively .
once again , the forward and reverse shock terms are replaced by pure diffusion decay terms analogous to equation 6 when shock input terminates .
figures 6 and 7 show some model sn lcs that result from sn ejecta - csm interaction with a shell ( @xmath112 0 ) and a steady - state wind ( @xmath112 2 ) , respectively .
examples are plotted for a variety of input luminosity combinations so that the effect on the final output lc can be illustrated .
the parameters used for the models plotted are listed in the captions of the figures .
some of the peaks of the resulting model sn lcs are not smooth due to the fact that we have used a simplistic heaviside function prescription for the termination of the forward and reverse shock luminosity input . in reality
this termination process will be a smoother function of time , thus making the peak of those lcs smoother , too .
note in figures 6 and 7 that when the shock inputs rise monotonically with time ( shell model ; figure 6 ) the lc on the rise has a monotonically increasing slope , but when the shock inputs decline monotonically with time ( wind model ; figure 7 ) , the lc on the rise has a monotonically decreasing slope , reminiscent of most observed sne .
the shape of the rising lc is thus a potentially strong constraint on the models .
given the large number of parameters involved in the hybrid model described by equation 21 , a careful survey of the parameter space is needed in order to investigate issues of parameter degeneracy and correlation .
we will present such an analysis in a follow - up paper that will illustrate fits of our model to observed slsn lcs . from a qualitative perspective , model lcs presented here will generally be brighter and briefer for higher values of @xmath119 , @xmath107 and @xmath104 and lower values of @xmath132 .
luminosities characteristic of slsne are more effectively produced considering interactions with dense shells ( @xmath112 0 ) rather than with steady - state winds ( @xmath112 2 ) .
models with constant density shells tend to be more sensitive to @xmath107 than those with winds because of the effect of the forward and reverse shocks .
interactions with steady - state winds are capable of producing lcs reminiscent of some normal luminosity sne iin , as we will show in the follow - up paper .
we now use the hybrid model presented in equation 21 to investigate the variety of lc shapes that we can obtain from that model and discuss implications for the sn ii classification scheme . as can be seen in figure 8
, a hybrid csm interaction plus radioactive decay input in which csm interaction dominates can produce a faster post - maximum decline rate that is consistent with the observed rapid post - maximum decline rate ( linear " in the logarithmic scale ) of sn iil .
this is due to the termination of the dominant luminosity input ( at @xmath164 in the case illustrated , for which the reverse shock luminosity dominates ) that will lead to a fast post - maximum decline dominated by cooling and diffusion of previously deposited shock luminosity .
furthermore , the differences seen between sn iil and sn iin in their spectra , namely the absence of p cygni absorption components in the balmer lines in the latter , maybe due to differences in the optical depth of the csm involved : a higher optical depth would obscure the effects of the underlying sn photosphere expansion thus leading to significant csm interaction and the presence of narrow balmer emission lines without a detectable broad absorption component , whereas a lower optical depth in the csm allows for the underlying sn expansion to be seen and the effects of the csm interaction are mainly imprinted in the lc of the event .
therefore , the sn iil class may represent a subset attributable to a particular restricted range in the large number of parameters involved in the more generic sn ejecta - csm interaction scenario and thus sn iil may be a subclass of sn iin .
the generic model presented here might also be used to explain the existence of approximately symmetric sn lcs like the ones observed for the transients scp06f6 ( barbary et al . 2009 ) , ptf09cwl , ptf09cnd and ptf09atu ( quimby et al . 2011 ) .
an example of such a symmetric " lc around the peak is also seen in figure 8 when csm interaction and radioactive decay are comparable .
this topic will be discussed in more detail in the follow - up paper .
this hybrid model also allows us to predict when the forward shock will break out of the csm photosphere based on the fit to the optical light curve .
the shock breakout will be followed by an x - ray / uv burst of radiation .
the model may also be used to estimate the intensity of that radiation and the subsequent uv / x - ray lc . in figure 9
we illustrate this by plotting the optical and uv / x - ray output lcs for a choice of parameters given in the captions for the case of sn ejecta interaction with a constant density shell ( @xmath112 0 ; left panel ) and a wind ( @xmath112 2 ; right panel ) . the various approximations associated with our analytic hybrid sn lc model and its limitations do not allow us to use it to address more specific emission characteristics like radio emission , thermalization of the radiation , the ionization state , the change in the opacity of the gas , and the evolution of the photosphere properties .
those aspects can be addressed only via numerical radiation hydrodynamics simulations and are beyond the scope of this project .
we can , however , address the limitations and uncertainties of our model bolometric sn lcs by benchmarking our results against existing numerical calculations for the same initial conditions .
to do so , we have used the results from three sets of numerical simulations : the woosley , blinnikov & heger ( 2007 ) ( hereafter wbh07 ) simulation of the lcs produced by a pulsational pair - instability supernova ( ppisn ) , the moriya et al .
( 2011 ) simulations of the interaction between sn ejecta from rsg progenitors with optically - thick csm winds and the chugai et al .
( 2004 ) simulation of the collision of sn ejecta with a 0.4 @xmath93 csm with density profile @xmath165 in order to reproduce the lc of sn 1994w .
we note that in our treatment we neglect the initial offset that the lcs will have in the time axis due to the fact that the first light will emerge after a diffusion time - scale .
this offset is shown in the simulations of wbh07 and moriya et al .
( 2011 ) and we can also explicitly calculate it in each case .
wbh07 considered the evolution of a 110 @xmath93 star that ends its life with an oxygen core of @xmath166 50 @xmath93 that is in the domain for which the core is unstable due to electron - positron pair production .
the collapse and subsequent pulse is not strong enough to disrupt the whole star , but only to eject 24.5 @xmath93 of its outer parts .
the remainder of the star ( with radius @xmath167 cm ) contracts , encountering pair - instability again after 6.8 years that leads to a second , more energetic pulse ( @xmath168 erg ) that ejects 5.1 @xmath93 .
the ejecta from this second pulse then collides with the slower moving ejecta from the first pulse producing a luminous sn - like output lc that was calculated using the code stella ( blinnikov & bartunov 1993 ; blinnikov et al .
1998 ; blinnikov et al .
2006 ) and presented in supplementary figure 6 of wbh07 .
the authors consider this ppisn phenomenon as an explanation for the slsn 2006gy .
to reproduce the wbh07 result , we use our hybrid model , presented by equation 21 above , for the same parameters ( @xmath169 @xmath168 erg , @xmath170 @xmath86 cm , @xmath158 5.1 @xmath93 , @xmath160 24.5 @xmath93 , @xmath92 0 @xmath93 ) .
the parameter @xmath104 is not explicitly given in wbh07 however we can estimate it , assuming steady - state mass loss . in this case @xmath171 and @xmath172 ( 24.5 @xmath173/(6.8 years ) , so for @xmath174 1,000 @xmath175 @xmath66 @xmath176 @xmath177 g @xmath178 .
this value is close to the dense shell values seen in supplementary figures 10 and 11 of wbh07 .
the density profiles of the sn ejecta ( second pulse ) and the csm ( ejecta from first pulse ) are also uncertain and not necessarily described by power laws in the context of a ppisn . for the purposes of our comparison , we will assume that the sn ejecta from the second pulse has a sn - like power law density profile with slope @xmath134 12 and for the csm we will consider two cases : one for a constant density shell ( @xmath112 0 ) and one for a steady state wind ( @xmath112 2 ) .
we also assume in our model that the photosphere is within a stationary csm while in the model of wbh07 the first shell which comprises the csm is expanding homologously at a speed of 100 - 1,000 @xmath175 @xmath66 .
the results of our comparison are shown in figure 10 .
the solid black and red curves correspond to the bolometric lcs for the original ejecta velocity and a doubled ejecta velocity , respectively , and the dotted black and red curves correspond to the ubvri lcs as calculated by wbh07 .
the dashed curves show our analytic lcs for @xmath112 0 ( green curve ) , for @xmath112 2 ( blue curve ) and for @xmath112 2 but for the values of @xmath74 and @xmath132 adjusted in order to provide a better fit to the results from the simulations ( yellow curve ) . as can be seen , given the uncertainties of the density profiles and the simplifying assumptions of our model ( centrally located power input and a fixed photosphere in the csm ) the two models agree in terms of rise time to maximum light and peak luminosity .
the post - maximum decline rates of our models are similar to those found by wbh07 , but of higher luminosity .
a somewhat better agreement is found if we decrease @xmath74 and @xmath132 by about 50 % , giving some indication of the uncertainty in our ability to estimate parameters .
bolometric and color lcs of sne produced by the interaction of sn ejecta ( from a rsg progenitor ) with a csm that is a steady state wind ( @xmath112 2 ) or with a @xmath112 1.5 density profile slope were presented by moriya et al . ( 2011 ) as results of 1-d numerical radiation hydrodynamics calculations .
we compare the output lcs from our analytic hybrid model to those presented by moriya et al .
( 2011 ) for four of their models : s13hw2r20m2e3 ( @xmath169 @xmath179 erg , @xmath170 @xmath180 cm , @xmath158 8 @xmath93 , @xmath160 0.65 @xmath93 , @xmath181 @xmath182 g @xmath178 ) , s15hw2r10m3e3 ( @xmath169 @xmath179 erg , @xmath170 @xmath180 cm , @xmath158 10 @xmath93 , @xmath160 0.031 @xmath93 , @xmath181 @xmath183 g @xmath178 ) , s15hw2r20m2e1 ( @xmath169 @xmath184 erg , @xmath170 @xmath180 cm , @xmath158 10 @xmath93 , @xmath160 0.65 @xmath93 , @xmath181 @xmath182 g @xmath178 ) and s15hw2r20m3e3 ( @xmath169 @xmath179 erg , @xmath170 @xmath180 cm , @xmath158 10 @xmath93 , @xmath160 0.065 @xmath93 , @xmath181 @xmath183 g @xmath178 ) .
the details for the parameters used for these models are given in table 2 of moriya et al .
figure 11 shows the result of this comparison .
the solid curves give the model lcs from the moriya et al .
( 2011 ) simulations and the dashed curves our analytical model lcs . our models again produce the rise time to maximum light and peak luminosity as well as the width of the main diffusion curve .
our post - maximum decline rates are again consistently similar to or slower than the ones found by the simulations .
chugai et al . ( 2004 ) sucessfully reproduced the plateau - like lc of the sn iin sn 1994w with radiation hydrodynamics simulations that involved contribution from both the radioactive decay of @xmath0ni and interaction with a dense csm .
more specifically they computed three models , of which their sn94w58 model provided the best fit to the data .
this model considered the explosion of a progenitor with a very extended radius ( @xmath170 @xmath185 cm ; their wind termination radius is several times larger ) with energy @xmath169
@xmath186 erg and @xmath158 7 @xmath93 that produced 0.015 @xmath93 of radioactive @xmath0ni within a 0.4 @xmath93 csm with density profile slope @xmath112 1 . given those parameters
the scale density at the base of the csm is @xmath181 @xmath187 g @xmath178 . using those parameters
, we plot our analytical lc model ( dotted curve ) and compare it to the chugai et al .
( 2004 ) numerical v - band ( solid green curve ) and u - band ( solid blue curve ) lcs in figure 12 .
we again see that the overall agreement between the models is very good .
our analytical model reproduces the plateau of sn 1994w in terms of duration and luminosity .
we also observe that the analytical lc model is closer to the u - band lc numerical model ; it reproduces the two luminosity breaks of the lc as a result of forward ( early break ) and reverse ( later break ) shock input termination .
the comparison between our analytic model lcs and radiation hydro lcs for the same conditions presented here illustrates the uncertainties involved .
the fact that we consider the forward and reverse shock power input to be centrally located and the photosphere to be fixed in the csm in order to obtain a separable pde leads to an overestimate of the diffusion time scale in our models because the relative shock and photosphere dynamics in reality will lead to an ever decreasing optical depth between the forward shock and the photosphere in the csm .
this will have the effect of producing a lc with a slower post - maximum decline rate in our models , which is what we see in figures 10 and 11 .
consequently , the overestimate of the diffusion time scales means we must employ smaller diffusion masses ( @xmath132 and @xmath74 ) to better fit a given lc . for a given lc , we will tend to underestimate the masses involved .
this is illustrated by our adjusted @xmath112 2 model to better reproduce the result of wbh07 .
the optical depth and the diffusion characteristics will also be affected by proper treatment of the acceleration of the csm by the precursor wave , the ionization state of the csm , and the position of the ionization front taking into account the decline in optical opacity beyond it , effects considered by moriya et al .
( 2011 ) . our comparison with the models of moriya et al .
( 2011 ) are a particularly stringent test , since the csm in their models was of rather low mass and our models are designed to represent the case of especially massive csm shells .
the breaks in luminosity seen in the two classes of models are of somewhat different , but closely , related nature . in our case , the first break in luminosity
occurs when the forward shock breaks out of the optically - thick csm or when the reverse shock has swept up all the available mass of sn ejecta . for moriya
et al . ( 2011 ) , these breaks are the result of the photosphere receding within the interaction region ; the late time radiation is due to left - over " thermal emission within the ejecta . in our models ,
the shock moves out to the photosphere ; in the models of moriya et al .
( 2011 ) the photosphere moves inward to meet the forward shock , thanks in large part to the acceleration of the csm by the radiation flow .
for substantially more massive csm the radiative acceleration will be less , and we would expect our models and those of moriya et al . to converge to greater similarity . as noted above
, we get quite good agreement with the results of chugai et al .
( 2004 ) for models with a more massive csm .
although the physics involved is accounted for in a more accurate way in the numerical simulations , our simple analytic results reproduce the basic lc features .
we understand the sign of the effect of neglecting the decrease in optical depth in front of the forward shock .
our models may thus be used as a first step in fitting observed sn lcs to get a basic understanding of the parameters involved before proceeding to more expensive numerical simulations .
slsn 2006gy stirred a lot of discussion among the sn community .
it was discovered by the rotse - iiib telescope of the texas supernova search project ( smith et al .
2007 ) , which obtained unfiltered photometry over the course of @xmath166 200d providing us with a well - constrained lc and explosion date .
we allow the explosion date to vary in a limited range in our fitting process that will be discussed below .
the very late - time decline rate of the lc of slsn 2006gy is reported to be consistent with the decline of @xmath0co ( smith et al .
a rich database of optical spectra were obtained for slsn 2006gy ( smith et al .
2007 , 2008 , 2010 ) that provides an extensive record of its spectral evolution .
slsn 2006gy showed strong balmer emission features with their narrow components associated with p cygni absorption indicative of photospheric expansion .
the h@xmath31 line profile evolved throughout the course of the lc of slsn 2006gy showing an evolution that is marked by three phases described in smith et al .
the full width at half - maximum ( fwhm ) of h@xmath31 around maximum light reveals characteristic velocities of @xmath166 4,000 km s@xmath26 .
we note , however , that this velocity information may not directly correspond to the bulk kinematic motion of the slsn ejecta .
smith et al .
( 2010 ) discuss the possibilities of line broadening due to electron scattering that yields information about the scattering optical depth and thermal motion of the electrons , but not about a true physical bulk expansion velocity associated with slsn 2006gy .
we will nevertheless use the value of @xmath166 4,000 km s@xmath26 as a fiducial velocity for our model fitting purposes .
spectra of the host galaxy of slsn 2006gy determined the redshift of the slsn to be @xmath188 0.074 , which means that the absolute visual peak magnitude of the event reached @xmath166 -22 m , making slsn 2006gy one of the brightest explosions ever discovered .
based on the information given by the optical lc and the spectra , slsn 2006gy was classified as a sn iin event ( smith et al .
2007 ) .
various models have been discussed to explain the nature of slsn 2006gy .
smith & mccray ( 2007 ) considered a shell - shock diffusion model for which the luminosity output around maximum light is reproduced by shock heating due to the interaction between the sn ejecta and a dense csm shell .
their simple model provided a decent fit for the decline of the lc , but failed to properly account for the rise of the lc since they considered an instantaneous shock input at maximum light .
they also considered a contribution from the radioactive decay of @xmath0co to account for the lc at very late times and estimated a @xmath0ni mass of about @xmath166 8 @xmath93 .
a pair - instability sn ( pisn ) scenario was discussed for slsn 2006gy in smith et al .
this model has difficulties in accounting for the large discrepancy between the total ejected and nickel mass and it turns out that it is not a good fit to the data , as we will show below .
other models considered interaction with an extensive csm envelope ( smith et al . 2007 ) or
cs clouds ( agnoletto et al . 2009 ) .
a popular picture today is that slsn 2006gy resulted from the explosion of a massive lbv - type star within a dense csm envelope , reminiscent to that of @xmath189 car ( smith et al .
another model for slsn 2006gy was considered by wbh07 in which the lc of the sn is due to interaction between shells ejected as a result of the ppisn process ( 3.3.4 ) .
we also note that another , exotic model , a quark - nova explosion has been considered for slsn 2006gy and other slsne ( ouyed et al .
2010 ) .
given that so far we are lacking a self - consistent lc model for slsn 2006gy that reproduces the whole lc and accounts for its spectral characteristics , we consider the rotse lc of slsn 2006gy converted to a pseudo - bolometric lc to test the hybrid model discussed in this paper .
we assume the bolometric correction bc=0 due to the fact that smith et al .
( 2007 ) do not provide such an estimate since there are not adequate simultaneous multi - band photometric observations throughout the course of the lc of the event .
therefore we accept the lack of a true observed bolometric lc of slsn 2006gy as one more uncertainty in our fit .
we also adopt @xmath190 0.72 mag yielding r - band extinction @xmath191 1.68 mag ( smith et al .
we fit equation 21 to the pseudo - bolometric lc of slsn 2006gy and present our result in figure 13 .
ideally the fit would be done by a chi - square minimization technique ( chatzopoulos , wheeler & vinko 2009 ) , but here we have chosen parameters by hand to illustrate the capacity of the model to represent the observations .
the model illustrated was obtained for the following choice of parameters : @xmath118 0 , @xmath134 12 , @xmath112 0 , @xmath192 erg , @xmath158 40 @xmath93 , @xmath160 5 @xmath93 ( @xmath193 4.9 @xmath93 ) , @xmath170 @xmath194 cm , @xmath181 @xmath195 g @xmath178 , @xmath92 2 @xmath93 . for these parameters we estimate the radius of the photosphere to be @xmath196 @xmath197 cm and the optical depth of the csm @xmath198 120 , which is consistent with the fact that the photosphere is within the optically thick shell .
as argued in the previous section where we presented a comparison with the radiation hydrodynamics results of woosley , blinnikov & heger ( 2007 ) for the lc of slsn 2006gy , our masses may be underestimated by @xmath166 50 % .
given the maximum luminosity of slsn 2006gy ( @xmath199 @xmath200 erg s@xmath26 ) and the radius of the photosphere we derived , we estimate a black - body temperature of @xmath201 13,000k .
this value for @xmath202 is consistent with the value @xmath203 11,000 - 12,000k that smith et al .
( 2007 ) derived from low - resolution spectra taken close to maximum light .
we note , however , that black - body emission generally provides a poor fit to the seds of sne iin due to flux dilution and line blanketing and to the fact that other emission mechanisms are dominant ( see for example dessart et al .
2009 on sn 1994w , miller et al .
2010 on sn 2008iy ; chatzopoulos et al .
2011 on sn 2008am ) .
the fit to the lc of slsn 2006gy given in figure 13 was based on a constant density shell . to illustrate the utility of our models to limit parameters
, we note that it was diffcult to provide an equally satisfactory fit based on a finite mass , steady - state " wind , with @xmath112 2 .
the csm luminosity input is generally too small to account for the observed luminosities even for large values of the csm scale density , @xmath104 .
extraordinary values for @xmath104 ( 10@xmath204 , 10@xmath205 , 10@xmath206 g @xmath178 ) imply a wind that starts at the stellar photosphere , but with a higher density than the outer envelope of the star .
this condition is unphysical and implies extraordinary mass - loss rates ( @xmath207100 solar masses per year ) .
even with extraordinary values for @xmath104 there are fitting issues because the input for the forward and reverse shocks in the @xmath112 2 case declines rapidly monotonically , thus forcing the rise to maximum to be very rapid .
figures 7 and 9 ( second panel ) illustrate @xmath112 2 cases for which the rise time is less than 15 - 20 days . for @xmath112 2 ,
it is very difficult to produce a longer rise time without , for instance , having a wind mass considerably in excess of the ejecta mass and an overall poor fit in detail .
wind models that roughly give the correct maximum luminosity also give a continuously - declining csm shock input that produces a very slow decline of the output luminosity until shock termination occurs that does not agree with the observations .
we thus find that while the models do have a large number of parameters , the observations considerably constrain the model parameters in practice .
we also plot in figure 13 ( dashed curve ) the expected uv / x - ray lc due to the forward shock propagation in the optically - thin part of the csm based on the parameters determined by the fit to the optical lc .
we see that the forward - shock luminosity , and associated uv / x - ray burst , is terminated at about 54 days after explosion in the rest - frame
. this may be consistent with the low x - ray flux detected by _
chandra _ at about 54 rest - frame days after explosion as reported by smith et al .
once the forward shock terminates upon reaching the outer edge of the shell , the uv / x - ray luminosity falls dramatically .
this result also implies that the uv / x - ray burst prior to termination of the forward shock would have been bright enough to be detectable .
we conclude that interaction of typical sn ejecta with a massive optically - thick csm shell plus a contribution from 2 @xmath93 of radioactive ni seems to reproduce the observed lc of slsn 2006gy . in this context ,
the progenitor of slsn 2006gy was probably a massive star that underwent significant and episodic mass loss .
the massive csm shell could have been the result of either lbv - type ejection , or a bubble due to interaction of stellar winds of previous epochs or the result of a ppisn .
we also note that our hybrid csm interaction plus radioactive decay model fit for slsn 2006gy is similar to the idea presented by agnoletto et al .
( 2009 ) in the sense that they also used arguments for combined csm interaction ( with cs clumps in their decription ) and radioactive decay of less nickel than previously implied ( 3 @xmath93 ) .
we derived semi - analytical models for the lcs of sn ii by incorporating the effects of diffusion as treated by a80 , 82 with a variety of inputs that are considered candidate sn lc powering mechanisms ( radioactive decays of @xmath0ni and @xmath0co , magnetar spin - down and ejecta - csm interaction ) .
we found solutions for the cases of a fixed ( in radius ) csm photosphere and homologously expanding matter , including the effects of terminated power input in the case of cs shock heating .
one implication from our results is that the principal factor that gives rise to the observed diversity in the shape , duration and luminosity of sn iin is the presence or absence of a csm enviroment and the physical properties that are associated with it ; mainly its optical depth , characteristic density and density profile .
sn iip , sn ib and sn ic are all the result of core - collapse explosions with the difference that in sn ib and sn ic events the h and he envelope of the progenitor star is lost and in sn iip the large outer h envelope is retained .
the csm enviroment around these events is probably of very low optical depth , so that the effects of ejecta - csm interaction are negligible and the only mechanism that gives rise to the optical luminosity output are shock heating of the ejecta ( for sn iip ) plus an internal source ( radioactive decays or magnetar spin - down ) combined with the effects of h and he recombination . for sn iip events where
the h envelope is retained , the effects of h recombination are more pronounced , while for sn ib and sn ic events the lcs lack an extended plateau phase due to the absence of the outer envelope .
some sn ib / c events develop sn iin characteristics in their nebular spectra at late times , indicating that the ejecta have finally reached the previously expelled h or he envelope and started to interact with it ( pastorello et al .
2008 ) .
on the other hand , in sn iil and sn iin events , the csm envelopes have higher optical depth so that the effects of ejecta - csm interaction dominate the output luminosity . in the case of sn
iil the optical depth is probably moderate since p cygni features are seen and the lc decline is faster , while for sn iin a high optical depth is implied that keeps the expansion of the sn ejecta obscured and produces a lc of longer duration .
the pronounced effects of the optically - thick csm environment on the lc of luminous sn iin manifest themselves by showing a variety in the lc shapes and durations of these events .
this leads us to argue that the sn iil classification could be a byproduct of the diversity of csm properties and not a separate class of sn explosions .
the faster decline and apparently symmetric shape of the lcs of some sn iil and slsne ( scp06f6 ; barbary et al .
2009 , ptf09cwl , ptf09cnd and ptf09atu ; quimby et al .
2011 ) may be well reproduced by models of shock heating , where the forward and reverse shocks terminate their contribution once they have swept up most of the available sn ejecta and csm mass .
if that is true , then strong pre - sn mass loss might constitute a significant source of diversity among sne iin , producing slsne at the high mass end , and putting such interacting sne into a different parameter space compared to sne that do not strongly interact .
this point of view was advocated by blinnikov & bartunov ( 1993 ) where they presented radiation - hydrodynamics models of sn iil in order to fit the lc of sn 1979c .
we applied our hybrid csm interaction plus radioactive decay model in the case of the lc of slsn 2006gy .
we found that the lc of this extraordinary event could be reasonably well reproduced by a model where interaction of @xmath166 40 @xmath93 of sn ejecta with a circumstellar shell of @xmath166 5 @xmath93 plus radioactive decay of @xmath166 2 @xmath93 of @xmath0ni provide the luminosity input .
although we derive a smaller @xmath0ni mass than previous authors , it remains the case that slsn 2006gy was a brilliant explosion and that the progenitor star must have undergone episodic mass - loss , a picture consistent with an lbv - type star ( smith et al .
we note that alternative scenarios for the nature of the massive csm shell around slsn 2006gy , such as a wind blown - bubble ( dwarkadas 2011 ) or a ppisn ( woosley , heger & blinnikov ; 2007 ) can not be excluded .
the models we present here are only approximate and have a large number of parameters , but they also provide the means to efficiently explore a large range of parameter space . in this way , parameters can be chosen to guide more elaborate and expensive numerical radiation - hydrodynamics calculations . in a subsequent paper
we will present fits to the lcs of many of the recently discovered slsne and other peculiar transient events .
the ultimate goal is to deduce the physical properties of the progenitors and the csm environments of a variety of sne using their observed optical lcs and spectra in order to constrain the mass - loss history and pre - sn evolution of massive stars .
we would like to thank roger chevalier , vikram dwarkadas , lars bildsten , rob robinson , dave arnett , sean couch , and the anonymous referee for useful discussions and comments .
we would especially like to thank sergei blinnikov and takashi moriya for providing their results from radiation hydrodynamics simulations of sn ejecta csm interaction that we have used for benchmarking against our analytical models .
this research is supported in part by nsf grants ast-0707669 and ast-1109801 and by the texas advanced research program grant astro - arp-0094 .
e. chatzopoulos would like to thank the propondis foundation of piraeus , greece for its support of his studies .
j. vinko received support from hungarian otka grant k76816 .
agnoletto , i. , et al .
2009 , , 691 , 1348 arnett , w. d. 1979 , , 230 , l37 arnett , w. d. 1980 , , 237 , 541 arnett , w. d. 1982 , , 253 , 785 arnett , d. 1996 , supernovae and nucleosynthesis , ed .
spergel , princeton series in astrophysics , princeton , nj : princeton university press arnett , w. d. , & fu , a. 1989 , , 340 , 396 balberg , s. , & loeb , a. 2011 , , 414 , 1715 barbary , k. , et al .
2009 , , 690 , 1358 bersten , m. c. , benvenuto , o. , & hamuy , m. 2011 , , 729 , 61 blinnikov , s. i. , & bartunov , o. s. 1993 , , 273 , 106 blinnikov , s. i. , & popov , d. v. 1993 , , 274 , 775 blinnikov , s. , lundqvist , p. , bartunov , o. , nomoto , k. , & iwamoto , k. 2000 , , 532 , 1132 chatzopoulos , e. , wheeler , j. c. , & vinko , j. 2009 , , 704 , 1251 chatzopoulos , e. , wheeler , j. c. , vinko , j. , et al .
2011 , , 729 , 143 chevalier , r. a. 1982 , , 258 , 790 chevalier , r. a. , & fransson , c. 1994 , , 420 , 268 chevalier , r. a. , & liang , e. p. 1989
, , 344 , 332 chevalier , r. a. , & fransson , c. 2001 , arxiv : astro - ph/0110060 chevalier , r. a. , & irwin , c. m. 2011 , , 729 , l6 chugai , n. n. , & danziger , i. j. 1994 , , 268 , 173 chugai , n. n. , blinnikov , s. i. , cumming , r. j. , et al . 2004 , , 352 , 1213 colgate , s. a. , & mckee , c. 1969 , , 157 , 623 colgate , s. a. , petschek , a. g. , & kriese , j. t. 1980 , , 237 , l81 dessart , l. , hillier , d. j. , gezari , s. , basa , s. , & matheson , t. 2009 , , 394 , 21 dessart , l. , & hillier , d. j. 2010 , , 405 , 2141 dessart , l. , livne , e. , & waldman , r. 2010 , , 408 , 827 dessart , l. , & hillier , d. j. 2011 , , 410 , 1739 dwarkadas , v. v. 2011 , , 23 ensman , l. , & burrows , a. 1992 , , 393 , 742 falk , s. w. , & arnett , w. d. 1977 , , 33 , 515 falk , s. w. 1978 , , 225 , l133 kasen , d. , & woosley , s. e. 2009 , , 703 , 2205 kasen , d. , & bildsten , l. 2010 , , 717 , 245 klein , r. i. , & chevalier , r. a. 1978 , , 223 , l109 maeda , k. , tanaka , m. , nomoto , k. , et al .
2007 , , 666 , 1069 matzner , c. d. , & mckee , c. f. 1999 , , 510 , 379 miller , a. a. , silverman , j. m. , butler , n. r. , et al .
2010 , , 404 , 305 moriya , t. , tominaga , n. , blinnikov , s. i. , baklanov , p. v. , & sorokina , e. i. 2011 , , 415 , 199 nakar , e. , & sari , r. 2010 , , 725 , 904 nomoto , k. , thielemann , f .- k .
, & yokoi , k. 1984 , , 286 , 644 nymark , t. k. , chandra , p. , & fransson , c. 2009 , , 494 , 179 ofek , e. o. , rabinak , i. , neill , j. d. , et al .
2010 , , 724 , 1396 ostriker , j. p. , & gunn , j. e. 1971 , , 164 , l95 ouyed , r. , kostka , m. , koning , n. , leahy , d. , & steffen , w. 2010 , arxiv:1010.5530 pastorello , a. , et al .
2008 , , 389 , 131 popov , d. v. 1993 , , 414 , 712 popov , d. v. 1995 , astronomy letters , 21 , 610 quimby , r. m. , et al .
2011 , , 474 , 487 schlegel , e. m. 1996 , , 111 , 1660 smith , n. 2006 , , 644 , 1151 smith , n. , et al .
2007 , , 666 , 1116 smith , n. , & mccray , r. 2007 , , 671 , l17 smith , n. , chornock , r. , li , w. , ganeshalingam , m. , silverman , j. m. , foley , r. j. , filippenko , a. v. , & barth , a. j. 2008 , , 686 , 467 smith , n. , chornock , r. , silverman , j. m. , filippenko , a. v. , & foley , r. j. 2010 , , 709 , 856 swartz , d. a. , wheeler , j. c. , & harkness , r. p. 1991
, , 374 , 266 utrobin , v. p. 1996
, , 306 , 219 valenti , s. , et al.2008 , , 383 , 1485 van marle , a. j. , smith , n. , owocki , s. p. , & van veelen , b. 2010 , , 407 , 2305 woosley , s. e. , blinnikov , s. , & heger , a. 2007 , , 450 , 390 woosley , s. e. 2010 , , 719 , l204
a80 , a96 presented a solution for light curves powered by radioactive decay of @xmath0ni based on the energy equation : @xmath208 where @xmath209 is the specific internal energy , @xmath210 is the pressure , @xmath5 is the specific volume , @xmath211 is the specific energy generation rate corresponding to the radioactive decay of @xmath0ni and the luminosity is given by the diffusion approximation : @xmath19 where @xmath21 is an appropriate opacity .
a80 , 82 solved this set of equations by assuming 1 ) homologous expansion , 2 ) a centrally - located power source , and 3 ) that radiation pressure was dominant .
a80 , 82 did not consider a general power input source as we need to consider here .
kasen & bildsten ( 2010 ) presented a related solution for the light curve based on an integrated version of the energy equation and a general power input , of which a magnetar was a specific example .
kasen & bildsten did not show explicitly how this global , integrated solution is related to the solution of a80 , 82 that was expressed in terms of local , specific quantities . here
we present the most general solution in the formulation of a80 , 82 , show how that solution is generalized to any power input , and how that solution relates to the global , integrated solution .
following a80 , 82 we adopt the same prescription for separation of variables , @xmath212^{4},\ ] ] where @xmath11 is the dimensionless radial variable , @xmath213 is the initial central temperature , @xmath16 is the scaling velocity of the homologous expansion @xmath214 , and @xmath15 is the initial radius . adopting @xmath215 constant
, we can then write , following the analysis of a80 , 82 , the luminosity at the surface , @xmath216 1 , as a function of time as : @xmath217 } \int_0^t e^{[\frac{t'^{2}}{t_{d}^{2}}+\frac{2r_{0}t'}{vt_{d}^{2 } } ] } f(t ' ) [ \frac{r(t')}{t_{0}r_{0}}]dt ' + \frac{e_{th,0 } i_{m}}{m \epsilon_{0 } b i_{th } } e^{-(t^{2}/t_{d}^{2}+2r_{0}t / vt_{d}^{2})},\ ] ] where @xmath218 is the ejecta mass , @xmath219 is the amplitude of the specific power input in erg @xmath67 @xmath66 , @xmath220 is the time - dependence of the power input , @xmath221 is the initial total thermal energy , @xmath91 is the lc time - scale which is the geometric mean of the diffusion time - scale @xmath222 and the expansion time - scale @xmath223 , therefore @xmath224 .
the factor @xmath37 can be written as @xmath38 or @xmath225 , as convenient to evaluate the limits @xmath226 0 and @xmath23 0 .
the density is assumed to scale as @xmath227^{3}$ ] , where @xmath228 is the central density at time 0 and the function @xmath31 appearing in the definition of @xmath75 , @xmath229,\ ] ] is a constant by separation of variables .
the parameter @xmath230 is assumed to be constant where @xmath231 represents the radial distribution of the power input , @xmath232 , and @xmath233 .
a80 , 82 show that @xmath234 where @xmath235 for a variety of density distributions and that @xmath236 , the initial amount of nickel injected in the radioactive decay model .
with these definitions and relations we can write for a general power input @xmath237 and , expressed in terms of luminosity @xmath238 where @xmath239 or , formally , @xmath240 } \int_0^t e^{[\frac{t'^{2}}{t_{d}^{2}}+\frac{2r_{0}t'}{vt_{d}^{2 } } ] } f(t ' ) [ \frac{r(t')}{t_{0}r_{0}}]dt ' + \frac{e_{th,0}}{t_{0 } } e^{-(t^{2}/t_{d}^{2}+2r_{0}t / vt_{d}^{2})}.\ ] ] this is the form of the output luminosity we present in equation 3 of the main text . returning to equation a1
, we can integrate the energy equation over mass to find @xmath241^{4 } \dot{\phi } i_{th } = l_{inp } - l\ ] ] where @xmath242 .
kasen & bildsten ( 2010 ) began with an approximate integrated form of the energy equation to write @xmath243 + \frac{1}{3 } a t^{4 } 4 \pi r^{2 } \frac{dr(t)}{dt } = l_{inp}-\frac{4 \pi r(t)^{2 } c}{3 \kappa \rho}\frac{\partial a t^{4}}{\partial r},\ ] ] where the second term on the right hand side is @xmath244 . using the same separation of variables for @xmath10 as in equation a3
, this equation can be written as @xmath241^{4 } \dot{\phi } \frac{\psi(x)}{3 } = l_{inp } - l.\ ] ] this equation agrees with equation a10 only if @xmath245 , which is only true if @xmath246 constant .
this means that the temperature as a function of radius has to be constant .
the solution of kasen & bildsten is thus not strictly speaking self - consistent , since a temperature gradient is required and employed to estimate the luminosity .
their solution is nevertheless a useful approach , given the uncertainties involved in modeling light curves driven by uncertain phenomena .
to derive an expression for the luminosities of the forward and reverse shocks we assume that all of their kinetic energy converts efficiently to radiation ( radiative shock approximation ) : @xmath247 where @xmath248 is the swept - up mass behind the shock and @xmath249 the shock velocity . in order to estimate a final expression for the luminosity presented in equation b1 , we need to know the dynamics of the forward and reverse shock ( radius , velocity and acceleration ) as a function of time .
chevalier ( 1982 ) considered the interaction between two media with power - law density profiles : the sn ejecta density profile @xmath113 , where @xmath114 is a scaling parameter for the ejecta density profile , @xmath250^{(n-3)/2}/[(3-\delta)(n-3)m_{ej}]^{(n-5)/2}$ ] , @xmath116 is the power - law exponent and @xmath117 is the slope of the inner density profile of the ejecta ( values of @xmath118 0 , 2 are typical ) and the csm density profile @xmath99 , where @xmath100 is the density of the csm medium , @xmath101 is a scaling constant and @xmath102 the power - law exponent . in general @xmath103 where @xmath104 is the density of the csm shell at @xmath105 .
we use as fiducial value @xmath106 , where @xmath107 the radius of the progenitor star .
thus we set the density scale of the csm , @xmath104 , immediately outside the stellar envelope .
using momentum conservation , chevalier ( 1982 ) and chevalier & fransson ( 2001 ) found the following self - similar solutions for the radii of the forward and the reverse shocks , respectively , as a function of time : @xmath251^{\frac{1}{n - s}}t^{\frac{n-3}{n - s}},\ ] ] and @xmath252^{\frac{1}{n - s}}t^{\frac{n-3}{n - s}},\ ] ] where @xmath121 , @xmath122 and @xmath70 are constants that depend on the values of @xmath116 and @xmath102 and , for a variety of values , are given in table 1 of chevalier ( 1982 ) .
more specifically the parameters @xmath121 and @xmath122 refer to the ratio of the shock radius to the radius of the contact discontinuity that forms as a result of sn ejecta - csm interaction ( @xmath253 and @xmath254 for the forward and the reverse shock , respectively , where @xmath255 is the radius of the contact discontinuity given by equation 3 of chevalier ( 1982 ) and @xmath256 and @xmath257 are the radii of the forward and the reverse shocks respectively ) . using @xmath258 , @xmath259 yields the velocity and the acceleration of the shock as a function of time .
given the shock dynamics derived above , we can calculate the swept - up mass behind the forward shock assuming that @xmath107 is smaller than @xmath260 for @xmath261 : @xmath262^{\frac{3-s}{n - s } } t^{\frac{(n-3)(3-s)}{n - s}},\ ] ] and the swept - up mass behind the reverse shock : @xmath263^{\frac{1}{n - s}}t^{\frac{n-3}{n - s}}}{v_{sn } t})^{3-n},\ ] ] where @xmath264 is the characteristic sn expansion velocity and @xmath265 is the radius of the sn photosphere which is assumed to expand homologously .
we have again assumed that @xmath107 is small .
substituting equations b2 through b5 and their derivatives in equation b1 yields the final results for the luminosity input from the forward and the reverse shocks ( @xmath266 and @xmath267 respectively ) : @xmath268 and @xmath269 where @xmath123 and @xmath124 denote the heaviside step function that controls the termination of the forward and reverse shock , respectively , and @xmath127 is the initial time of the csm interaction which sets the initial value for the luminosity produced .
the velocity , @xmath264 , in equation b5 is given by @xmath270^{\frac{1}{2}}}{x_{0}},\ ] ] where @xmath129 is the dimensionless radius of the break in the sn ejecta density profile from the inner flat component ( controlled by @xmath117 ) to the outer , steeper component ( controlled by @xmath116 ) which is at radius @xmath130 .
@xmath267 in equation b7 is independent of @xmath264 and hence of @xmath271 .
summing equations b6 and b7 , the total luminosity input from self - similar sn ejecta - csm interaction is @xmath272 the forward shock termination time - scale , @xmath125 , is given by the following expression , assuming that the input from the forward shock terminates when all the available csm has been swept up : @xmath131^{(s-3)/(n - s)}}{4 \pi \beta_{f}^{3-s}}|^{\frac{n - s}{(n-3)(3-s ) } } m_{csm}^{\frac{n - s}{(n-3)(3-s)}},\ ] ] where @xmath132 the total mass of the csm .
following the same assumption , the reverse shock termination time - scale @xmath126 is given by the following expression : @xmath133^{\frac{n - s}{s-3}}.\ ] ] after termination , the luminosity of each component decays in a manner analogous to equation 6 in the main text such that @xmath273 and @xmath274 | we present generalized supernova ( sn ) light curve ( lc ) models for a variety of power inputs including the previously proposed ideas of radioactive decay of @xmath0ni and @xmath0co and magnetar spin - down .
we extend those solutions to include finite progenitor radius and stationary photospheres as might be the case for sn that are powered by interaction of the ejecta with circumstellar matter ( csm ) .
we provide an expression for the power input that is produced by self - similar forward and reverse shocks that efficiently convert their kinetic energy into radiation .
we find that this ejecta - csm interaction luminosity that we derive is in agreement with results from multi - dimensional radiation hydrodynamics simulations in the case of an optically - thin csm .
we develop a semi - analytical model for the case of an optically - thick csm by invoking an approximation for the effects of radiative diffusion similar to that adopted by arnett ( 1982 ) for sn ii and compare this model to the results of numerical radiation hydrodynamics models .
this model can give complex light curves , but for monotonically declining shock input , the lcs have a smooth rise , peak and decline . in the context of this model , we provide predictions of the shock breakout of the forward shock from the optically - thick part of the csm envelope .
we also introduce a hybrid lc model that incorporates ejecta - csm interaction plus @xmath0ni and @xmath0co radioactive decay input .
we fit this hybrid model to the lc of the super - luminous supernova ( slsn ) 2006gy .
we find that shock heating produced by ejecta - csm interaction plus some contribution from radioactive decay provides a better fit to the lc of this event than previously presented models .
we also address the relation between sn iil and sn iin with ejecta - csm interaction models . the faster decline of sn iil can be reproduced by the diffusion of previously deposited shock power if the shock power input to the diffusive component vanishes when the reverse shock sweeps up the whole ejecta and/or the forward shock propagates through the optically - thick cs matter . a csm interaction with forward and reverse shock power input can produce the lcs of sn iin in terms of duration , shape and decline rate , depending on the properties of the csm envelope and the progenitor star .
this model can also produce lcs that are symmetric in shape around peak luminosity , which is the case for the observed lcs of some recently discovered peculiar transient events .
we conclude that the observed lc variety of sn iin and of some slsne is likely to be a byproduct of the large range of conditions relevant to significant ejecta - csm interaction as a power source .
0.57 in |
a complete understanding of transport behavior in many - body systems is one of the utmost challenges in fundamental studies of non - equilibrium statistical mechanics . in the classical domain
, it is widely believed that chaotic systems should show diffusive ( normal ) transport , whereas integrability should be associated with ballistic ( abnormal ) transport @xcite , although normal transport has also been verified for a non - chaotic system @xcite . in the quantum domain , the conditions that determine specific transport behaviors are still under debate , but here also , the main conjecture is of the correspondence integrable - ballistic and chaotic - diffusive @xcite .
several theoretical approaches have been undertaken to address this issue , including : attempts to derive the fourier law from a microscopic foundation by applying the hilbert space average method @xcite and by numerically studying transport of heat in finite chaotic and non - chaotic systems coupled to heat reservoirs @xcite ; analysis of the transport behavior of local magnetization in isolated finite systems at zero temperature @xcite ; new advances toward the problem of quantum thermalization @xcite ; comparisons of the results for conductivity in integrable and chaotic systems at finite temperature and in the thermodynamic limit @xcite , a subject of intense discussion here being the possibility of ballistic transport in non - integrable quasi - one - dimensional systems @xcite ; numerical studies of spin diffusion at long times at infinity temperature @xcite ; as well as studies of transport properties near the metal - insulator transition @xcite .
investigations of transport behavior in the particular case of quasi - one - dimensional spin-1/2 systems have been highly motivated by experiments in low - dimensional magnetic compounds , such as copper oxide ( cuprate ) systems , where ballistic behavior has been observed for heat conduction @xcite , and also for magnetization , as revealed by nuclear magnetic resonance ( nmr ) experiments @xcite .
these compounds are described by models of interacting spins-1/2 arranged in structures such as chains , two - leg ladders , and square lattices @xcite .
clean spin-1/2 heisenberg chains with only nearest - neighbor interactions are integrable models solved with the bethe ansatz method @xcite , whereas two - dimensional lattices are chaotic @xcite and two - leg spin ladder systems become chaotic when the interchain and intrachain interactions are of the same order @xcite . in ref .
@xcite , an isolated isotropic finite spin-1/2 heisenberg chain at zero temperature and with only nearest - neighbor interactions was considered in the analysis of transport of local magnetization in both cases : when the system was clean and therefore integrable , and when random on - site disorder led to the onset of quantum chaos .
free boundary conditions were assumed .
a bouncing behavior of the local magnetization was observed in the integrable regime and interpreted as a hint of ballistic transport , whereas for the chaotic system the local magnetization showed an exponential relaxation to equilibrium , which was considered a reflection of diffusive transport .
the first part of this paper is also dedicated to the investigation of transport of local magnetization in spin-1/2 heisenberg systems , but different ways to induce chaos are dealt with . the goal is to resolve whether non - integrability may have a unique correspondence with the exponential decay verified in @xcite .
the integrability - breaking terms considered are : on - site disorder described either by ( i ) randomly distributed zeeman energies @xcite or by ( ii ) a few defects placed on specific positions of the chain @xcite ; and couplings to more surrounding spins such as ( iii ) next - nearest neighbor interactions and ( iv ) interchain interactions , as typical of two - leg spin ladder systems .
an exponential decay is observed only for ( i ) and ( iv ) . for the finite systems studied ,
few defects lead to a behavior that sometimes resembles localization , the exponential behavior appearing only under very special conditions , while next - nearest neighbor interactions generate oscillations of the local magnetization that are similar to those seen for the integrable system , although faster .
these observations support the conjecture that chaos might not be a sufficient condition for normal transport .
the second part of this work focuses on the analysis of methods of coherent quantum control , the so - called dynamical decoupling ( dd ) schemes , as potential tools to manipulate transport behavior .
dd schemes consist of sequences of external control operations that average out unwanted contributions to the system hamiltonian .
these methods have long been applied in nmr spectroscopy @xcite , where the goal is to modify the nuclear spin hamiltonian to eliminate or scale selected internal interactions .
more recently , dd has addressed the removal of interactions between the target system and its environment @xcite and has been put in a general control - theoretic framework @xcite .
it has also been considered in studies of transport of information @xcite . here
, we introduce dd sequences that suppress the effects of terms leading to quantum chaos , the purpose being the approach to the ballistic transport behavior verified for the integrable system .
the paper is organized as follows .
ii explains how the identification of the chaotic regime is performed and describes the models to be considered .
iii presents the results for the transport of local magnetization for systems in different regimes .
iv introduces dd sequences to control transport behavior and shows , as an illustration , results for the sequence that cancels the effects of on - site disorder .
concluding remarks are given in sec .
for quantum systems , chaos may be identified by analyzing the distribution of spacings @xmath0 between neighboring energy levels @xcite .
quantum levels of integrable systems tend to cluster and are not prohibited from crossing , the typical distribution is poissonian : @xmath1 in contrast , non - integrable systems show levels that are correlated and crossings are strongly resisted , the level statistics is given by the wigner - dyson distribution .
the exact form of the distribution depends on the symmetry properties of the hamiltonian . in the case of systems with time reversal invariance
it is given by : @xmath2 to analyze the transition from integrability to chaos , the quantity @xmath3 , defined as @xmath4ds}{\int_0^{s_0}[p _ { p}(s)-p_{wd}(s)]ds } \label{eta},\ ] ] was introduced in @xcite , where @xmath5 is the first intersection point of @xmath6 and @xmath7 . for an integrable system @xmath8 , while for a chaotic system @xmath9 .
the critical value below which the system is considered chaotic is chosen to be @xmath10 @xcite .
we study homogeneous and isotropic spin-1/2 heisenberg chains with open boundary conditions , as described by the hamiltonian : @xmath11 above , @xmath12 is set equal to 1 ; @xmath13 is the spin operator at site @xmath14 , @xmath15 being the pauli operators ; and @xmath16 corresponds to the number of sites .
the parameter @xmath17 is the zeeman splitting of spin @xmath14 as determined by a static magnetic field in the @xmath18 direction .
the system is clean whenever all sites have the same energy splitting @xmath19 , and it is disordered when defects characterized by different energy splittings @xmath20 are present . @xmath21 and @xmath22 are the interaction strengths of nearest - neighbor ( nn ) and next - nearest - neighbor ( nnn ) couplings , respectively , and are assumed to be constant .
all calculations in this work are performed in the basis consisting of eigenstates of the total spin operator in the @xmath18 direction , @xmath23 . in this basis , the nn and nnn ising interactions , @xmath24 and @xmath25 ,
contribute to the diagonal elements of the hamiltonian , while the @xmath26-terms , @xmath27 and @xmath28 , constitute the off - diagonal elements .
the role of the @xmath26-terms is to transfer excitations through the system by exchanging the position of nearest and next - nearest neighboring spins pointing in opposite directions .
also considered here are two - leg spin ladder systems corresponding to two coupled spin chains as described by @xmath29 + h_{1,2 } \nonumber \\ & = & \sum_{m=1}^2 \left[\sum_{n=1}^{l/2 } \omega_{n , m } s_{n , m}^z + \sum_{n=1}^{l/2 - 1 } j \vec{s}_{n , m}.\vec{s}_{n+1,m } \right ] \nonumber \\ & + & \sum_{n=1}^{l/2 } j_{\bot } \vec{s}_{n,1}.\vec{s}_{n,2 } \ : , \label{ladder}\end{aligned}\ ] ] where @xmath17 and @xmath21 are as before , @xmath30 determines the chain in which the site is positioned , and @xmath31 characterizes the strength of the interchain interaction . in order to derive meaningful level spacing distributions , before diagonalizing the hamiltonian and unfolding the spectrum @xcite , all trivial symmetries of the system need to be identified .
the analysis of symmetries is necessary , because a poisson distribution may appear whenever wigner - dyson distributions from different symmetry sectors are mixed , which may lead to erroneously interpret the system as integrable . in both models considered here ,
@xmath32 is conserved , therefore , instead of diagonalizing matrices of dimension @xmath33 , we study the largest subspace . for @xmath16
even it corresponds to the sector with @xmath34 and dimension @xmath35 ^ 2 $ ] . depending on the parameter values , hamiltonians ( [ ham ] ) and ( [ ladder ] )
may also exhibit the following symmetries @xcite : invariance under lattice reflection , which leads to parity conservation ; and conservation of total spin @xmath36 , that is @xmath37=0 $ ] ( @xmath38 symmetry ) .
notice also that the heisenberg model with a magnetic field does not commute with the conventional time - reversal operator , however the distribution associated with its chaotic regime is still given by eq .
( [ wd ] ) , as discussed in @xcite .
in studies of transport properties , the most popularly used method is the green - kubo formula @xcite .
however , the application of this formula for the treatment of heat transport has been criticized and the use of the hilbert space average method to demonstrate the emergence of heat diffusion from microscopic models has been suggested as an alternative @xcite . this approach has been extended to the analysis of transport of magnetization in ref .
@xcite . here , as in @xcite , we study the transport of local magnetization as defined by @xmath39 where @xmath40 is the state of the system at instant @xmath41 written in the basis of @xmath32 .
the initial states considered come from the sector @xmath42 . for system ( [ ham ] ) , @xmath43 has all spins pointing up placed in the first half of the chain , while the remaining down - spins appear in the other half .
for the two - leg system , the initial state has all up - spins in one chain and all down - spins in the other .
we assume @xmath44 , which leads to @xmath45 in both cases .
_ integrable system . _
the clean system with only nearest - neighbor interaction , as described by @xmath46 ( [ ham ] ) with @xmath47 and @xmath48 , corresponds to an integrable model solved with the bethe ansatz method @xcite .
the dynamics for the local magnetization is shown on the left panel at the top of fig .
[ fig1:random ] .
the bouncing behavior suggests ballistic transport @xcite . in what follows ,
we compare this result with the time evolution of local magnetization for chaotic regimes induced by different ways . in the chain with only nearest - neighbor interactions , as given by @xmath46 ( [ ham ] ) with @xmath48 , chaos may be induced if one or more defects are present .
_ random on - site disorder .
_ assume that the zeeman energies are given by @xmath49 , where @xmath50 are uncorrelated random numbers with a gaussian distribution : @xmath51 and @xmath52 @xcite .
the transition from integrability ( @xmath53 , @xmath8 ) , to chaos ( 0.05@xmath54 , @xmath55 ) is indicated by @xmath3 , which is computed in the sector @xmath42 and is shown on the right panel at the top of fig .
[ fig1:random ] .
notice that as @xmath56 , conservation of parity and total spin start playing a role . at @xmath57 ,
the correct evaluation of @xmath3 would need to take these symmetries into account @xcite .
( [ ham ] ) with @xmath48 and @xmath44 .
the value of @xmath58 is irrelevant for the dynamics .
left top panel : @xmath59 for @xmath47 .
right top panel : @xmath3 computed in the sector @xmath42 for gaussian random on - site disorder , @xmath51 and @xmath52 . left bottom panel : @xmath59 for @xmath60 , and 0.25 from bottom to top .
average over 20 realizations .
right bottom panel : @xmath59 for a sample of realizations with @xmath61.,width=316 ] on the left panel at the bottom of fig .
[ fig1:random ] , we show the time evolution of @xmath59 averaged over 20 realizations for five different values of @xmath62 in the chaotic region .
as we approach chaoticity , for @xmath63 and 0.1 , oscillations are still seen ; whereas for @xmath64 , 0.2 , and 0.3 , an exponential decay of @xmath59 takes place reaching final values between 1 and 2 .
after the decay , the larger probability to find spins pointing up in the first half of the chain is reflected by the positive values of @xmath59 , which are obtained with the majority of the realizations .
on the right panel at the bottom , we show the behavior of a sample of realizations with @xmath64 : for some of them @xmath65 decays to equilibrium , @xmath66 , indicating an equal probability to find up - spins in both halves of the chain , but for the majority of the realizations @xmath59 remains positive throughout ; hardly any curve reaches negative values of @xmath65 .
_ one defect .
_ a single defect in the middle of a chain , at @xmath67 ( or equivalently at @xmath68 ) , may lead to quantum chaos . for @xmath44 ,
this happens when @xmath69 , as discussed in @xcite .
the transition to chaos is shown by @xmath3 on the left panel at the top of fig .
[ fig2:single ] .
the curve is obtained in the @xmath42 sector , but , as mentioned before , for @xmath70 the correct evaluation of @xmath3 would require also the consideration of the @xmath38 symmetry and parity conservation@xcite . a way to break the @xmath38 symmetry and deal with larger subspaces even when the system is integrable consists of including defects at the edges of the chain @xcite .
( [ ham ] ) with @xmath48 and @xmath44 .
left panels : one defect on site @xmath71 , @xmath47 for @xmath72 .
right panels : two equal defects on sites @xmath71 and @xmath73 , @xmath47 for @xmath74 .
top panels : dependence of @xmath3 on the value of the defect(s ) .
bottom panels : @xmath59 for chaotic systems.,width=316 ] the transport of local magnetization is shown on the left panel at the bottom . for @xmath75 , 0.5 , and 1.0 , even though @xmath3 indicates chaoticity , an exponential decay is not observed . instead , for the two smaller values , especially for the smallest one ,
partial revivals are verified , whereas for @xmath76 there occurs localization of the up - spins in the first half of the chain . in contrast , a defect placed on site 7 , which shows exactly the same behavior for @xmath3 , does lead to an exponential decay of @xmath59 see right panel of fig . [
fig : scaling ] @xcite .
the reason for the different results is the following .
the initial state @xmath77 is directly coupled only with the state @xmath78 .
when the defect is on site 6 , the hopping of the up - spin from site 6 to site 7 is not favorable , since the state loses the extra energy from the defect site and the positive ising energy coming from the pair of parallel spins @xmath79 .
this explains why it becomes easy to localize the initial state by increasing the defect value .
contrary to that , if the defect is on site 7 , the ising energy lost by breaking the pair of parallel spins is regained by placing the up - spin on the defect . in this scenario ,
directly coupled states may be very close in energy , which favors delocalization .
_ two defects .
_ two equal defects in the middle of the chain , at @xmath80 , may also lead to quantum chaos . for @xmath44
this happens when @xmath81 .
the transition to chaos is shown by @xmath3 on the right panel at the top panel of fig .
[ fig2:single ] .
the curve is obtained by taking both symmetries into account : conservation of parity and @xmath82 .
the relaxation of the local magnetization for @xmath83 , 0.5 and 0.65 is shown on the right panel at the bottom .
the decay is not exponential and @xmath84 is never reached , most up - spins tending to localize in the first half of the chain .
this may again be understood by comparing the energies of the initial state and of the state it is directly coupled to .
the whole spectrum is required to obtain the plots for @xmath3 presented in figs .
[ fig1:random ] and [ fig2:single ] , therefore the decision to deal with relatively small systems , @xmath44 .
this choice was a good compromise leading to sufficient statistics and the possibility to run various realizations of random on - site energies .
in addition , notice that the behaviors of the local magnetization obtained with @xmath44 are also reproduced with 10 and 14 spins , as shown in the two panels of fig .
[ fig : scaling ] .
thus , for the analysis developed in this paper , a system with 12 spins is sufficiently adequate . in order to simulate the time evolution of @xmath59 in much larger systems ,
we could resort , for example , to the very efficient algorithm recently proposed in ref .
@xcite .
( [ ham ] ) with @xmath48 .
left panels : clean chains , @xmath47 for @xmath85 .
right panels : chaotic systems with one defect in site @xmath86 , @xmath47 for @xmath87 and @xmath88 . ,
width=316 ] in a clean heisenberg system with @xmath47 , chaos may be induced by adding frustrating next - nearest neighbor interactions where @xmath89 or by keeping @xmath48 and adding interchain interactions where @xmath90 @xcite . in both cases ,
total spin , total spin in the @xmath18 direction , and parity are conserved . _ next - nearest neighbor interactions .
_ hamiltonian ( [ ham ] ) with @xmath47 and @xmath91 describes a chaotic system @xcite .
however , the transport of local magnetization shows oscillations similar to those observed for the integrable system , although at a faster rate , as seen on the left panel of fig .
[ fig3:nnn ] .
therefore , if the bouncing behavior of @xmath59 in isolated systems at zero temperature is indeed a signature of ballistic transport , integrability is not a necessary condition for abnormal conductivity .
_ interchain interaction . _
hamiltonian ( [ ladder ] ) with @xmath47 and @xmath92 also describes a chaotic system @xcite . in this case , as shown on the right panel of fig .
[ fig3:nnn ] , a very fast exponential decay of the local magnetization to equilibrium is indeed verified .
( [ ham ] ) with @xmath44 , @xmath47 , and @xmath93 .
right panel : @xmath46 ( [ ladder ] ) with @xmath44 , @xmath47 , and @xmath94.,width=336 ] the distinct behaviors of the transport of local magnetization verified for different chaotic systems figs .
[ fig1:random ] , [ fig2:single ] , [ fig : scaling ] , and [ fig3:nnn ] prevent a clear correspondence between chaoticity and diffusive transport .
new light may be shed to the problem if this analysis is extended to the transport of other quantities , such as heat in open systems .
we propose to control transport behavior by applying dd methods .
dd schemes consist of sequences of external control operations that average out unwanted terms of the system hamiltonian . in the case of spin systems ,
the control operations correspond to very strong magnetic fields ( pulses ) able to rotate the spins and time - reverse the system evolution @xcite . here
, we assume the ideal scenario , where the pulses are arbitrarily strong and capable of performing instantaneous rotations , the so - called bang - bang controls @xcite .
our goal is to eliminate the effects of the terms leading to quantum chaos and approach the transport behavior verified for the integrable system shown fig .
[ fig1:random ] .
this can be achieved with the sequences described below , where each sequence handles a particular integrability - breaking term . _ on - site disorder . _
the effects of on - site disorder may be eliminated by rotating all spins after every interval of free evolution @xmath95 , @xmath96 , by 180@xmath97 around a direction perpendicular to @xmath18 , for example @xmath98 , as determined by the operator , @xmath99 the interaction terms remain undisturbed , but the sequence of these rotations leads to the cancellation of the one - body terms at every @xmath100 , @xmath101 , where @xmath102 is the cycle time ( for details about relevant frames and pulse generation see @xcite and references therein ) .
the propagator at a time @xmath103 is then given by @xmath104 where @xmath105 $ ] and @xmath106 denotes time ordering . by adopting the notation @xmath107 = \exp[-i h_1 \delta
t],\ ] ] @xmath108 \nonumber \\ & = & -\exp [ -i ( - h_z + h_{nn } ) \delta t ] = -\exp[-i h_2 \delta t ] , \nonumber\end{aligned}\ ] ] where @xmath109 is the identity operator , we rewrite the propagator as @xmath110.\ ] ] above , @xmath111 is the average hamiltonian @xcite , and the terms in the sum are obtained by using the baker - campbell - hausdorff expansion .
the first two terms are @xmath112 .\end{aligned}\ ] ] the sequence of pulses reshapes the hamiltonian . in the ideal limit of @xmath113 one recovers the hamiltonian for a clean heisenberg model with only nearest - neighbor couplings , @xmath114 , as desired . in fig .
[ fig4:dd ] , we show the transport of local magnetization as modified by the above sequence for two cases of on - site disorder : the top panels correspond to random defects and the bottom panels to two defects in the middle of the chain .
the transport behavior for both situations integrable system and disordered heisenberg chain subjected to the dd sequence closely coincide when the intervals between the pulses are smaller than the reciprocal interaction strength , @xmath115 , as seen on the left panels . for short time evolutions ,
good agreement between the curves is still verified for @xmath116 , whereas at longer times the accumulation of residual averaging errors become significant , this being more perceptible in the bottom right panel . to slow down error accumulation ,
randomized schemes as developed in @xcite may be incorporated to the pulse sequence .
( [ ham ] ) with @xmath48 and @xmath44 .
( blue ) curves showing a fast decay to @xmath117 correspond to disordered systems in the absence of pulses .
( green ) bouncing curves represent disordered systems subjected to the dd sequence ( [ dd_sequence ] ) ; they closely coincide with the ( red ) bouncing curves obtained with the clean heisenberg system .
top panels : gaussian random on - site disorder , @xmath51 and @xmath52 with @xmath118 .
average over 20 realizations .
bottom panels : @xmath119 and @xmath47 for @xmath74 .
left panels : @xmath120 .
right panels : @xmath121 .
data is acquired after every @xmath102.,width=336 ] _ interchain interactions . _ to decouple the two interacting chains in a two - leg spin ladder system , we may apply a sequence of @xmath122-pulses that rotates all spins of just one of the chains .
the pulses do not affect the intrachain interactions , but frequently change the sign of the interchain interactions . to eliminate the interchain couplings in the three directions , we need a sequence of at least four pulses with two alternating directions , such as @xmath123 where @xmath124 the chain which is subjected to the pulses is also alternated to guarantee the cancellation of the one - body terms as well . at any time @xmath125 , we then have @xmath126 .
_ next - nearest neighbor interactions .
_ a possible sequence to eliminate next - nearest neighbor interactions involves eight @xmath122-pulses .
the controls are applied to specifically chosen spins and are only viable if means exist to distinguish them .
the suggested pulse sequence is @xmath127 where @xmath128 at any time @xmath129 , we obtain @xmath130 .
apart from a rescaling factor 1/2 , we recover the hamiltonian for a clean heisenberg model with only nearest neighbor interactions up to first order in @xmath131 .
notice that the sequence cancels both next - nearest neighbor interactions and one - body terms .
the hamiltonians for the intervals of free evolution are given in the appendix .
since they no longer conserve @xmath82 , simulations involving this pulse sequence need to consider the whole hilbert space of dimension @xmath33 .
the purpose of this work was twofold : to contribute to the ongoing discussion about transport properties of quantum many - body systems and to study the possibility of controlling transport behavior by resorting to dynamical decoupling methods .
our conclusions are based on the analysis of the transport of local magnetization in isolated finite spin-1/2 systems with free boundary conditions and at zero temperature .
diffusive transport is associated with the exponential decay of the local magnetization and ballistic transport with its bouncing behavior . under these conditions
, it was shown that a one to one correspondence between quantum chaos and normal transport does not necessarily hold ; instead , different integrability - breaking terms may lead to both ballistic and diffusive behavior .
dynamical decoupling sequences capable of suppressing the terms of the hamiltonian leading to quantum chaos were proposed .
the goal was to approach the ballistic transport behavior of local magnetization obtained with the integrable system .
this was achieved by applying pulses separated by intervals smaller than the reciprocal interaction strength , as illustrated for two cases of on - site disorder : random defects and two defects in the middle of the chain .
the manipulation of transport behavior in systems subjected to dynamical decoupling methods is a topic that deserves further exploitation . in particular ,
the analysis of heat transport in systems coupled to two heat reservoirs at different temperature is one of our future goals . also important
is the identification of real systems where these ideas could be tested experimentally , a possible good candidate being a crystal of fluorapatite , as proposed in studies of information transport @xcite . -0.5
cm we thank w. g. brown , g. a. cwilich , and c. o. escobar for helpful discussions .
the pulses from eq .
( [ pulses ] ) lead to the propagator | we analyze transport of local magnetization and develop schemes to control transport behavior in finite spin-1/2 heisenberg chains and spin-1/2 heisenberg two - leg ladders at zero temperature . by adjusting parameters in the hamiltonians
, these quantum systems may show both integrable and chaotic limits .
we provide examples of chaotic systems leading to diffusive and to ballistic transport .
in addition , methods of coherent quantum control to induce a transition from diffusive to ballistic transport are proposed . |
sudakov - tsirelson and borell proved independently ( see @xcite ) that for @xmath0 endowed with a gaussian measure , half - spaces bounded by hyperplanes are isoperimetric , i.e. , minimize weighted perimeter for given weighted volume .
caete et al .
* question 6 ) , in response to a question of brancolini , conjectured that for @xmath0 endowed with a finite sum of gaussian measures centered on the @xmath1-axis , half - spaces bounded by vertical hyperplanes are isoperimetric .
we consider the case of two such gaussians in @xmath2 or @xmath3 .
our theorem [ 2.18 ] proves that on the double gaussian line , rays are isoperimetric .
section 4 provides evidence that on the double gaussian plane , half - spaces are isoperimetric .
theorem [ 2.18 ] states that the isoperimetric regions in the double gaussian line are rays .
we may assume that the two gaussians have centers at @xmath4 and @xmath5 .
for small variances , the theorem follows by comparison with the single gaussian . for larger variances ,
additional quantitative and stability arguments are needed to rule out certain non - ray cases . a conjecture , stated in this paper as conjecture [ 4.1 ] , of caete et al .
* question 6 ) states that isoperimetric regions in the double gaussian plane are half - planes bounded by vertical lines .
we use variational arguments to show that horizontal and vertical lines are the only lines that are candidates , and that vertical lines always beat horizontal lines .
formulas [ 2.3 ] and [ 2.5 ] state standard first and second variation formulas , analogous to the first and second derivative conditions for local minima of twice - differentiable real functions . [ 2.31 ] a _ density _ @xmath6 on @xmath0 is a positive , continuous function used to weight volume and hypersurface area . given a density @xmath6 , the ( weighted ) _ volume _ of a region @xmath7 is given by @xmath8 the ( weighted ) _ hypersurface area _ of its boundary @xmath9 is given by @xmath10 @xmath7 is called _ isoperimetric _ if no other region of the same weighted volume has a boundary with smaller hypersurface area .
we now assume that the density @xmath6 is smooth .
the existence and regularity of isoperimetric regions for densities of finite total volume is standard . [ 2.32 ] suppose that @xmath6 is a density in the line or plane such that the line or plane has finite measure @xmath11
. then for any @xmath12 , an isoperimetric region @xmath7 of weighted volume @xmath13 exists and is a finite union of intervals bounded by finitely many points in the line or a finite union of regions with smooth boundaries in the plane .
let @xmath6 be a smooth density on @xmath14 .
let @xmath7 be a smooth region in @xmath14 .
let @xmath15 be a smooth , one - parameter family of deformations on @xmath14 such that @xmath16 is the identity .
for a given @xmath17 , @xmath18 traces out a small path in @xmath14 beginning at @xmath1 and @xmath19 is a curve for each @xmath20 .
therefore @xmath21 where @xmath22 describes a perturbation of @xmath9 .
define @xmath23 [ 2.3 ] suppose that @xmath24 and @xmath25 are the inward unit normal and mean curvature of @xmath9 .
let @xmath26 be the vector field @xmath27 and @xmath28 .
then we have that @xmath29 the proof we give of theorem 2.3 is the same as that given in @xcite .
we have that @xmath30 .
therefore @xmath31 it has been shown that @xmath32 , so that @xmath33 therefore @xmath34 the last equality comes from applying the divergence theorem ; the negative sign appears because @xmath35 is the inward normal .
let @xmath36 denote the divergence relative to @xmath37 , and let @xmath38 denote the gradient relative to @xmath37 . for the perimeter , we have that @xmath39 therefore @xmath40 note that @xmath41 therefore @xmath42@xmath43 since any isoperimetric curve is a local minimum among all curves enclosing a certain volume @xmath13 , it satisfies @xmath44 for any @xmath15 such that @xmath45 for small @xmath46 [ 2.4 ] if a curve @xmath9 is isoperimetric , then @xmath47 is constant on @xmath9 .
if a curve @xmath9 is isoperimetric , then it satisfies @xmath44 . by formula
[ 2.3 ] , this occurs if and only if @xmath47 is constant on @xmath9 .
[ 4.4 ] let @xmath48 be a boundary in the line or plane with unit inward normal @xmath49 and let @xmath50 denote the standard curvature . for a density @xmath6 ,
we call @xmath51 the _ generalized curvature _ of @xmath48 . by corollary
[ 2.4 ] , all isoperimetric curves have constant generalized curvature . in the real line ,
@xmath52 , so that isoperimetric curves have @xmath53 constant . for the interval @xmath54 $ ] , the generalized curvature evaluated at @xmath55 is equal to @xmath56 while the generalized curvature evaluated at @xmath57 is equal to @xmath58 . [ 2.5 ] let the real line be with smooth density @xmath6 . if a one - dimensional boundary @xmath59 satisfies @xmath44 for any volume - preserving @xmath21 , then @xmath60 this formula comes from proposition 3.6 in @xcite , where the second variation is stated for arbitrary dimensions .
some terms from the general formula cancel in the one - dimensional case .
[ 2.7 ] let @xmath61 be a subset of the real line such that @xmath62 for all @xmath63 with equality holding at no more than one point . if @xmath64 is an isoperimetric boundary contained in @xmath61 , then @xmath64 is connected and thus a single point .
if @xmath64 has at least two connected components , then since by proposition [ 2.32 ] @xmath64 consists of a finite union of points , there is a nontrivial volume - preserving flow on @xmath64 given by moving one component so as to increase the volume and the other so as to decrease it . by the second variation formula [ 2.5 ] ,
the second variation satisfies @xmath65 this contradicts that @xmath64 is isoperimetric .
theorem [ 2.18 ] states that for the real line with density given by the sum of two gaussians with the same variance @xmath66 , isoperimetric regions are rays bounded by single points .
this theorem is a necessary condition for conjecture [ 4.1 ] , which states that isoperimetric regions in the double gaussian plane are half - planes bounded by vertical lines .
propositions [ 2.8 ] , [ 2.22 ] , and [ 2.10 ] treat the cases @xmath67 , @xmath68 , and @xmath69 .
lemma [ 2.9 ] shows that the gaussians having the same variance allows us to reduce the problem to ruling out a few non - interval but still symmetrical cases .
when the gaussians have different variances , the problem is harder and not treated by our results .
let @xmath70 denote the gaussian density with mean @xmath71 and variance @xmath66 , and let @xmath72 let @xmath73 in one dimension , the regions are unions of intervals and their boundaries are points . since the total measure is finite , isoperimetric regions exist by proposition [ 2.32 ] . for a given weighted length @xmath13 , we seek to find the set of points with the smallest total density which bounds a region of weighted length @xmath13 .
since the complement of a region of weighted length @xmath13 has weighted length @xmath74 , we can assume that our regions have weighted length @xmath75 .
the following proposition shows that it suffices to consider the density @xmath76 .
[ 2.1 ] suppose that @xmath64 is an isoperimetric boundary enclosing @xmath57 region @xmath77 of weighted length @xmath13 for the density @xmath78 .
then for any @xmath79 , @xmath80 is an isoperimetric boundary enclosing region @xmath81 of weighted length @xmath13 for the density @xmath82 let @xmath83 denote the standard gaussian density .
first , we show that for any boundary @xmath84 enclosing a region @xmath85 , the weighted length of @xmath86 for the density @xmath87 is the same as the weighted length of @xmath85 for the density @xmath78 .
we have that @xmath88 @xmath89 @xmath90 where the @xmath91 denotes the weighted length in the appropriate densities .
second , for any two boundaries @xmath92 and @xmath93 , we have that @xmath94 for @xmath95 thus , @xmath96 in the density @xmath78 exactly when @xmath97 in the density @xmath87
. therefore @xmath98 in the density @xmath87 , and if any other boundary @xmath84 enclosing region @xmath85 satisfies @xmath99 in the density @xmath87 , then since @xmath64 is isoperimetric , we have that @xmath100 in the density @xmath78
. therefore @xmath101 in the density @xmath87 , so that @xmath102 is isoperimetric . as a result of proposition
[ 2.1 ] , it suffices to consider the density @xmath103 . 0.45 and @xmath104 .
the purple curves are for @xmath105 the blue curves for @xmath106 , and the green curves for @xmath107.,title="fig : " ] 0.45 and @xmath104 .
the purple curves are for @xmath105 the blue curves for @xmath106 , and the green curves for @xmath107.,title="fig : " ] [ 2.2 ] let @xmath26 be the disjoint union of two real - lines @xmath108 and @xmath109 , each with a standard gaussian density scaled so that it has weighted length @xmath110 .
for any given length @xmath111 , the isoperimetric region in @xmath26 of length @xmath13 is a ray contained entirely in @xmath108 or @xmath112 let @xmath64 be an isoperimetric boundary and @xmath113 its intersection with @xmath114 . if either @xmath115 or @xmath116 are nonempty , then they each must be single points since the isoperimetric boundaries for the single gaussian are always single points .
assume , in contradiction to the proposition , that for @xmath117 , @xmath118 is the @xmath119-th component on the @xmath119-th gaussian bounding a ray @xmath120 of weighted length @xmath121 . since @xmath122 , it is possible to put a point @xmath123 on the 1st gaussian at the same height as that of @xmath124 bounding a ray @xmath125 disjoint from @xmath126 and with weighted length @xmath127 .
consider the boundary @xmath128 , which has the same weighted perimeter as that of @xmath64 .
there exists a single point on @xmath115 bounding a ray of area @xmath13 and with weighted density smaller than @xmath129 .
this contradicts the fact that @xmath64 is isoperimetric . [ 2.6 ] when the line is endowed with the double gaussian density @xmath76 , isoperimetric boundaries @xmath64 always consist of a single point .
thus isoperimetric regions are always rays .
let @xmath130 we first show that if @xmath64 is contained entirely outside of @xmath131 $ ] , then @xmath64 must be connected .
suppose that @xmath64 has two connected components
. then we can define an area - preserving veloctity @xmath21 .
we have that @xmath132 therefore we have that have that @xmath133 are the only zeros of @xmath134 .
in addition , we have that @xmath135 when @xmath136 we have that @xmath137 .
therefore @xmath138 for @xmath139 .
similarly , we have that @xmath140 for @xmath141 , so that @xmath138 for @xmath142 .
since @xmath138 outside of @xmath131 $ ] , by the second variation formula , @xmath143 this is a contradiction .
suppose that @xmath64 contains at least two points @xmath144 .
since @xmath64 must then contain a point in @xmath131 $ ] , we can assume @xmath145 $ ] .
we have that @xmath146 since @xmath64 has constant generalized curvature , it follows that @xmath147 .
recall that we have shown that the only zeros of @xmath134 are @xmath148 by applying the first derivative test , it is shown that the maximum and minimum of @xmath149 on @xmath150 $ ] are achieved at @xmath151 and @xmath152 , respectively . since we previously showed that @xmath138 for @xmath153 and @xmath154 for @xmath155
, it follows that @xmath156 $ ] . since @xmath157 for all @xmath158 $ ] , we have that @xmath159 for all @xmath160 .
this means that the that ray enclosing the same weighted length as @xmath64 has a smaller perimeter than that of @xmath64 , contradicting the fact that @xmath64 is an isoperimetric boundary .
we now begin analyzing the problem when the centers of the gaussians are still @xmath5 and @xmath4 but when the variances are different .
[ 2.33 ] for the double gaussian density @xmath76 , the log derivative @xmath161 is given by @xmath162 we have that @xmath163 by using the substitution @xmath164 we get that @xmath162 [ 2.8 ] for the double gaussian density @xmath76 , if @xmath165 , isoperimetric boundaries are single points .
for any given @xmath57 , we have that @xmath166@xmath167 and @xmath168 0.45 and @xmath169 .
the purple curves are for @xmath105 the blue curves for @xmath106 , and the green curves for @xmath107.,title="fig : " ] 0.45 and @xmath169 .
the purple curves are for @xmath105 the blue curves for @xmath106 , and the green curves for @xmath107.,title="fig : " ] 0.45 and @xmath169 .
the purple curves are for @xmath105 the blue curves for @xmath106 , and the green curves for @xmath107.,title="fig : " ] as shown in figure [ fg2.32 ] , @xmath170 is positive for any @xmath171 and negative for @xmath141 , so that @xmath172 achieves its unique maximum at @xmath173 for any given @xmath57 .
we have that @xmath174 , so that @xmath175 is greater than @xmath176 for @xmath177 , and less than or equal to @xmath176 for @xmath165 . if @xmath178 , by corollary [ 2.7 ] , isoperimetric boundaries are always connected .
since isoperimetric boundaries consist of finite unions of points , they must be single points . [ 2.9 ] let @xmath179 and @xmath180 be two real functions with @xmath181 .
suppose @xmath179 and @xmath180 satisfy 1 .
@xmath182 , 2 .
@xmath183 , 3 .
@xmath184 , and 4 .
@xmath185 and @xmath186 on @xmath187 .
for any @xmath188 , if @xmath189 , then @xmath190 . is blue while @xmath191 is purple .
when the areas are equal as in the picture , @xmath192 is higher .
, height=288 ] as in figure [ fg2.3 ] , for all @xmath193 by ( 2 ) and ( 4 ) @xmath194 and by ( 3 ) and ( 4 ) @xmath195 . if we choose @xmath196 so that @xmath197 , we will have that @xmath198 .
since by ( 4 ) @xmath191 is concave and @xmath192 is convex , @xmath199
@xmath200 therefore @xmath201 , so that @xmath190 , as asserted . [ 2.12 ] suppose that @xmath202 $ ] is an interval of @xmath76-weighted length @xmath111 with @xmath203 .
then there exists a union of rays @xmath204 \cup [ d,\infty]$ ] of @xmath205-weighted length @xmath13 such that @xmath206 and @xmath207 0.5 0.5 since @xmath208 , we have that @xmath209 .
the union of rays @xmath210 \cup [ 2+s,\infty)$ ] has greater @xmath205-weighted length than the @xmath76-weighted length of @xmath202 $ ] . therefore there exists @xmath211 and @xmath212 such that @xmath213 \cup [ d,\infty)$ ] has @xmath205-weighted length @xmath13 , and @xmath214 [ 2.19 ] suppose that @xmath54 $ ] is an interval of @xmath76-weighted length @xmath111 with @xmath215 .
then there exists an interval @xmath216 $ ] of @xmath205-weighted length @xmath13 such that @xmath217 and @xmath218 0.4 0.4 the interval @xmath54 $ ] has @xmath205-weighted length less than @xmath13.therefore there exists @xmath219 and @xmath220 such that @xmath216 $ ] has @xmath205-weighted length @xmath13 . since @xmath221
, we have that @xmath222 and @xmath223 .
[ 2.13 ] if @xmath224 has @xmath225-weighted length @xmath226 , then there exists @xmath227 such that @xmath228 has @xmath76-weighted length @xmath13
. 0.5 0.5 if @xmath224 has @xmath225-weighted length @xmath226 , then @xmath229 the interval @xmath230 has @xmath76-weighted length greater than @xmath13 .
therefore there exists @xmath227 such that @xmath228 has @xmath76-weighted length @xmath13 .
now we begin analyzing the case where the variance satisfies @xmath231 . [
3.36 ] if @xmath66 satisfies @xmath232 , then @xmath233 exactly when @xmath234 .
this follows from the formula for @xmath235 given in proposition [ 2.8 ] .
suppose that @xmath66 is a variance .
in the proof of the following proposition , we will use the quantity @xmath236 [ 2.14 ] suppose that @xmath232 and @xmath64 is an isoperimetric boundary with at least one point @xmath237 in @xmath238 $ ] where @xmath239 enclosing a region of weighted - length @xmath111 .
then the boundary @xmath64 is one of the following : 1 .
a single point @xmath237 enclosing the ray @xmath230 , 2 .
\{s , t } where @xmath227 enclosing the interval @xmath202 $ ] , 3 .
\{s , t } where @xmath240 enclosing the interval @xmath241 $ ] , 4 .
\{s , -s , t } enclosing @xmath242 \cup ( -\infty , t]$ ] , @xmath242 \cup [ t,\infty)$ ] or @xmath202 \cup ( -\infty , -s]$ ] .
the analogous claims apply if @xmath243 $ ] .
, there are at most three points with @xmath141 with the same value for @xmath244.,height=192 ] since @xmath64 is isoperimetric , it can contain at most one point @xmath1 at which @xmath245 if it contained two such points , then by slightly shifting the two points we could create a new region with the same weighted length . by the second variation formula , the boundary of this region would have a smaller total density .
therefore @xmath64 can contain at most one point outside of @xmath246 $ ] .
in addition , @xmath64 has constant curvature , so that @xmath247 is constant on @xmath64 ( see figure [ fg2.10 ] ) .
since @xmath248 is positive on @xmath249 and negative on @xmath250 $ ] , there exists one point @xmath251 such that @xmath252 and one point @xmath253 such that @xmath254 therefore @xmath64 is a subset of @xmath255 suppose @xmath64 is not @xmath256 . if @xmath64 contains no points outside of @xmath246 $ ] , then @xmath64 is @xmath257 .
suppose @xmath64 contains one point @xmath258 outside of @xmath246 $ ] . if @xmath259 , then the only possibilities are @xmath260 or @xmath261 . if @xmath262 , then the only possibilities are @xmath257 or @xmath261 .
the regions enclosed follow from the fact that we assume @xmath111 . [ 2.16 ] suppose that @xmath64 is an isoperimetric boundary with at least one point @xmath263.$ ] if @xmath64 is of type [ 2.14]_(3 ) _ and @xmath264 , then the region @xmath7 enclosed by @xmath64 has @xmath76-weighted length no more than @xmath265 .
we have that @xmath266 for @xmath267 so that @xmath268 is increasing on @xmath269 .
if @xmath270 then the function @xmath271 decreases on @xmath272 . since @xmath273 , we have that @xmath274 on @xmath272
. therefore @xmath275 consider the function @xmath276 which sends @xmath1 to the weighted length of @xmath277.$ ] then @xmath278 + [ f_1(x ) - f_2(x-1)].\ ] ] for @xmath279 , both the bracketed quantities are negative , so that @xmath280 is decreasing on @xmath238 $ ] .
we have that @xmath281 therefore if we can show that @xmath282 , we will have that the @xmath76-weighted length of @xmath202 $ ] is less than @xmath283 and be done .
this follows immediately when @xmath284 , since @xmath285 when @xmath286 , we observe that @xmath287 is to the left of @xmath288 , so it suffices to show that @xmath289 .
thus we want to show that @xmath290@xmath291 + [ \tanh(s / a^2 ) - \tanh((1-s)/a^2)])/(a^2 ) \geq 0\ ] ] on @xmath292 $ ] .
this is equivalent to showing that @xmath293 + [ \tanh(s / a^2 ) - \tanh((1-s)/a^2 ) ] ) \geq 0\ ] ] on @xmath238 $ ] . since @xmath294 , @xmath295 .
in addition , @xmath296 therefore it suffices to show that @xmath297 achieves its minimum value on @xmath292 $ ] at @xmath298 we will do this by using the first derivative test to show that there is only one other local extremum in the interval and further demonstrating that this local extremum is not the minimum point .
we have that @xmath299 since @xmath300 , we have that @xmath301 . in addition , @xmath302 by using the substitution @xmath303 we get that @xmath304 therefore @xmath305 we have that @xmath306 when @xmath307 , we have that @xmath308 therefore @xmath309 attains a minimum value of @xmath310 on @xmath311 $ ] .
this shows that @xmath312 on @xmath313 $ ] , so that @xmath314 by the intermediate value theorem , there exists @xmath315 such that @xmath316 it follows that @xmath317 is also a zero of @xmath318 .
now @xmath319 tends to @xmath176 as @xmath1 tends to @xmath320 , so that @xmath321 for some @xmath322 .
therefore there exists @xmath323 in @xmath324 such that @xmath325 , and @xmath326 is also a zero of @xmath318 . again using the substitution
@xmath303 we see that @xmath327 is a rational function of @xmath328 whose numerator is quartic
. therefore @xmath318 has at most @xmath329 zeros , so that @xmath330 is the only zero of @xmath318 in @xmath331 since @xmath301 , @xmath332 so that @xmath333 for @xmath334.$ ] , height=192 ] [ 2.17 ] if the variance @xmath264 , then the isoperimetric boundaries @xmath64 with one point @xmath55 in @xmath238 $ ] can not be of type [ 2.14]@xmath257 .
0.45 and @xmath335,title="fig : " ] 0.45 and @xmath335,title="fig : " ] + 0.45 and @xmath335,title="fig : " ] 0.45 and @xmath335,title="fig : " ] let @xmath13 be the weighted length of @xmath64 .
if , in contradiction to the proposition , @xmath64 is of type [ 2.14]@xmath257 , then @xmath64 is of the form @xmath54 $ ] where @xmath336 , as shown in figure [ fg2.11]a . by proposition
[ 2.12 ] , there exists a union of rays @xmath213 \cup [ d,\infty)$ ] with @xmath205-weighted length @xmath13 such that @xmath337 .
this is shown in figure [ fg2.11]b . by the solution to the single gaussian isoperimetric problem
, there exists a ray @xmath230 , as shown in figure [ fg2.11]c , with @xmath205-weighted length @xmath13 such that @xmath338 . by proposition
[ 2.16 ] , @xmath339 , so that @xmath229 by proposition [ 2.13 ] , there exists a ray @xmath228 , as shown in figure [ fg2.11]d , with @xmath76-weighted length @xmath13 such that @xmath340 to get a contradiction to the fact that @xmath64 is isoperimetric , we show that @xmath341 write @xmath342\]]@xmath343 + [ f_1(s ) - f(t)].\ ] ] since @xmath344 > 0 $ ] , it suffices to show that @xmath345 > [ f(t ) - f_1(s)]$ ] . since @xmath346 , we have that @xmath347 > f(b ) - f_1(c ) > f(b ) - f_1(b ) = f_2(b).\ ] ] since @xmath348 , we have that @xmath349 < f(s ) - f_1(s ) = f_2(s).\ ] ] since @xmath350 , we have that @xmath351 , and this proves the claim . [ 2.21 ] if the variance @xmath352 , then the isoperimetric boundaries @xmath64 with one point @xmath353 in @xmath246 $ ] can not be of type [ 2.14]_(2)_. 0.5 and @xmath335,title="fig : " ] 0.5 and @xmath335,title="fig : " ] we know that @xmath354 [ recall the concavity / convexity argument ] , and since @xmath355 , we must have that @xmath356 pick @xmath357 such that @xmath358 and @xmath359 .
in other words , we get @xmath216 $ ] by reflecting @xmath54 $ ] over the line @xmath360 . since @xmath177 , either have @xmath361 or @xmath362 . in the first case
, we have that @xmath216 $ ] has the same @xmath205-length as @xmath54 $ ] , and since @xmath363 and @xmath220 , we have that @xmath364 and @xmath365 . therefore @xmath366 . at the same time , the @xmath367-length of @xmath216 $ ] is less than that of @xmath54 $ ] .
this difference is at most the @xmath367 length of @xmath368 . since @xmath369
, we can find @xmath370 such that @xmath371 $ ] has @xmath76-length @xmath13 .
in addition , @xmath372 , so that @xmath54 $ ] is not isoperimetric . in the second case
, we have that @xmath373 $ ] has the same @xmath205-length as @xmath54 $ ] , and since @xmath374 , we have that @xmath375 and @xmath376 . therefore @xmath366 . at the same time , the @xmath367-length of @xmath373 $ ] is less than that of @xmath54 $ ] .
this difference is at most the @xmath367 length of @xmath368 . since @xmath377
, we can find @xmath378 such that @xmath379 $ ] has the @xmath76-length @xmath13 .
in addition , @xmath372 , so that @xmath54 $ ] is not isoperimetric .
[ 2.20 ] if the variance @xmath352 , then the isoperimetric boundaries @xmath64 with one point @xmath55 in @xmath246 $ ] can not be of type [ 2.14]_(4)_. we may assume without loss of generality that @xmath380 .
suppose that @xmath64 is of type [ 2.14]@xmath261 .
then the region @xmath77 enclosed by @xmath64 consists of the union of an interval of type [ 2.14]@xmath260 or [ 2.14]@xmath257 and a ray .
apply propositions [ 2.17 ] and [ 2.21 ] to get a new region @xmath381 that beats the interval .
since @xmath382 , @xmath381 may be chosen to not intersect the ray
. then the union of @xmath381 and the ray beats @xmath77 .
[ 2.22 ] if @xmath64 is an isoperimetric boundary and the variance @xmath352 , then @xmath64 is a single point .
if @xmath64 does not contain a point @xmath263 $ ] , then by proposition [ 2.7 ] , then @xmath64 is a single point .
otherwise , apply propositions [ 2.17]-[2.20 ] to complete the proof .
[ 2.10 ] for the line endowed with density @xmath383 , if the variance @xmath66 is such that @xmath384 , then isoperimetric regions @xmath7 are always rays with boundary @xmath64 consisting of a single point . by proposition
[ 3.36 ] , we have that @xmath385 exactly when @xmath1 is @xmath386 .
since @xmath387 for @xmath388 and @xmath389 for @xmath141 , we have that @xmath134 is negative outside of @xmath246 $ ] and is positive in @xmath390 . ,
height=192 ] , height=192 ] suppose that @xmath64 is an isoperimetric boundary containing more than two points . by corollary
[ 2.7 ] , @xmath64 does not lie entirely outside @xmath246 $ ] . since @xmath154 on @xmath390 and @xmath391 , the maximum and minimum of @xmath392 on @xmath246 $ ]
are achieved at @xmath71 and @xmath288 with @xmath393 negative and @xmath394 positive .
since @xmath392 tends to @xmath395 as @xmath1 approaches @xmath320 , there exists a unique point @xmath396 such that @xmath397 . since @xmath398 , @xmath138 outside of @xmath399 $ ] .
we claim that @xmath64 must lie in @xmath399 $ ] . since @xmath400 is constant on @xmath64 , to show that @xmath401 $ ] it suffices to show that the maximum and minimum of @xmath392 on @xmath399 $ ] are achieved at @xmath402 and @xmath55 .
since @xmath176 is a local minimum for @xmath383 , it suffices to show that @xmath403 since @xmath404 is postive and @xmath405 for @xmath406 there exists a unique point @xmath357 where @xmath407 and @xmath161 changes from positive to negative at @xmath408 . to apply lemma [ 2.9 ] , consider functions @xmath179 and
@xmath180 denoting the increase in @xmath104 moving left of @xmath408 and the decrease in @xmath104 moving right of @xmath408 : @xmath409 @xmath410 which satisfy the hypotheses of lemma [ 2.9 ] . since @xmath411 , we have that @xmath412 there are five candidates for the minimum points of @xmath383 on @xmath399 $ ] : @xmath413 , and @xmath414 . since @xmath415 , we have that @xmath416 so that @xmath414 is not a candidate .
since , also by the preceding paragraph , @xmath392 is positive between @xmath176 and @xmath71 , we have that @xmath417 .
therefore the minimum on this interval is @xmath418 .
we have that @xmath419 for all @xmath420 .
therefore we have that @xmath421 for various values of @xmath57,height=192 ] to finish the proof , we must show that @xmath422 for all @xmath1 and all @xmath423 .
consider the numerator @xmath424 of @xmath76 given by @xmath425 we have that for a given @xmath1 , @xmath424 increases when @xmath57 increases , so that @xmath426 since @xmath427 which has the same sign as @xmath428 , we see that @xmath429 is maximized at @xmath176 .
therefore @xmath430 so that @xmath431 this means that there is a ray which beats @xmath64 , contradicting the fact that @xmath64 is isoperimetric . [ 2.18 ] the isoperimetric boundaries for the double gaussian density @xmath76 are always single points enclosing rays . to cover the three cases , apply propositions [ 2.8 ] , [ 2.22 ] , and [ 2.10 ] .
[ 2.11 ] the isoperimetric boundary @xmath64 corresponding to area @xmath110 is @xmath432 .
applying propositions 2.8 and 2.9 proves the result for @xmath433 . when @xmath434 , if @xmath64 is outside of @xmath246 $ ] where @xmath435 then by corollary [ 2.7 ] @xmath64 is a single point and must be @xmath432 .
if @xmath64 contains a point @xmath1 in @xmath246 $ ] , then since [ if the proof for 1/2 turns out to be different than for other areas , the following fact will be made into a lemma ; it is proved in the proof of 2.7 ] @xmath176 is the minimum of @xmath76 in this interval , @xmath436 since @xmath64 is isoperimetric , @xmath437 .
this section describes evidence for the conjecture of caete et al . , stated here as conjecture [ 4.1 ] , which states that double gaussian isoperimetric boundaries in the plane are vertical lines .
proposition [ 4.5 ] proves that horizontal and vertical lines are the only stationary lines .
proposition [ 4.6 ] proves that vertical lines are better than horizontal lines .
first we prove some incidental symmetry results ( propositions [ 4.2 ] and [ 4.3 ] ) . [ 4.1 ] let @xmath438 be the normalized sum of two gaussian densities with the same variance and different centers .
isoperimetric regions are half - planes enclosed by lines perpendicular to the line connecting the two centers . by the planar analogue of proposition
[ 2.1 ] , it suffices to prove this conjecture in the case where the centers are @xmath439 and @xmath440 .
then we have that @xmath441 the next two propositions describe some symmetry properties of isoperimetric curves . for a curve @xmath48 ,
let @xmath442 denote the weighted area enclosed by @xmath48 .
[ 4.2 ] consider a density @xmath83 symmetric about the @xmath1-axis .
if a closed , embedded curve @xmath48 encloses the same weighted area above and below the @xmath1-axis , then there is a curve @xmath443 which is symmetric about the @xmath1-axis , encloses the same weighted area , and has weighted perimeter no greater than that of @xmath48 .
let @xmath444 and @xmath445 be the parts of @xmath48 in the open upper and lower half - planes chosen so that the weighted perimeter of @xmath444 is no bigger than that of @xmath445 .
consider the curve @xmath443 formed by joining @xmath444 with its reflection over the @xmath1-axis and taking the closure .
let @xmath446 denote the part of @xmath48 on the @xmath1-axis and @xmath447 denote the part of @xmath443 on the @xmath1-axis . since @xmath83 is symmetric about the @xmath1-axis , @xmath448 .
in addition , @xmath449 we have that @xmath450 by assumption , and since the part of @xmath48 which intersects the @xmath1-axis must include @xmath447 , @xmath451 .
therefore @xmath452 [ 4.3 ] consider a density symmetric about the x - axis .
if @xmath48 is a closed embedded planar curve symmetric about the x - axis , then the part @xmath443 of @xmath48 in the open upper half - plane encloses half as much weighted area with half the weighted length .
suppose that @xmath48 is a curve that is symmetric about the @xmath1-axis and encloses area @xmath13 . since @xmath48 is symmetric about the @xmath1-axis , @xmath48 can not have non - zero perimeter on the @xmath1-axis
. then @xmath443 encloses area @xmath453 in the upper half - plane and has weighted perimeter @xmath454 . [ 4.5 ] if the plane is endowed with density @xmath76 , then horizontal and vertical lines have generalized curvature 0 and are the only lines which have constant generalized curvature .
let @xmath455 then @xmath456 in addition , the normal to the line @xmath457 is @xmath458 at all points of the line .
therefore the generalized curvature of such a line evaluated at @xmath459 is @xmath460 and by an analogous computation the generalized curvature evaluted at @xmath461 is @xmath462 thus the generalized curvature at @xmath459 and @xmath463 are equal exactly when @xmath464 .
this shows that only non - vertical lines that could possibly have constant curvature are the horizontal lines @xmath465 . such lines have normal @xmath466 , and this , combined for our formula with the gradient , shows that horizontal lines have constant curvature @xmath467 .
an explicit computation of the same variety shows that the vertical line @xmath468 has constant curvature @xmath469 [ 4.6 ] in the plane with double gaussian density @xmath76 , vertical lines enclose given area with less perimeter than horizontal lines .
0.5 0.5 we now compare the perimeters of and areas enclosed by the horizontal line @xmath468 and the vertical line @xmath470 . by symmetry and the fact that we may assume the areas are less than @xmath110
, we can assume that @xmath55 and @xmath71 are positive and consider the areas of the regions @xmath471 and @xmath472 .
the area enclosed by the vertical line is @xmath473 which is the same as the weighted length of the ray @xmath474 on the double gaussian line .
the perimeter of the vertical line is @xmath475 which is exactly the cost of @xmath476 on the double gaussian line .
the area enclosed by the horizontal line is @xmath477 which is the same as the weighted length of the ray @xmath478 on the single gaussian ( of total weighted - length 1 ) line .
the perimeter of the horizontal line is @xmath479 which is exactly the cost of @xmath480 on the single gaussian line .
therefore it suffices to show that a ray on the double gaussian line of length @xmath13 costs less than a ray on the single gaussian line of the same weighted length .
consider the line with density @xmath83 given by a single gaussian of total length @xmath110 .
the ray on the single gaussian is equivalent to the union of two disjoint , symmetric rays on the @xmath83-line .
the ray on the double gaussian is equivalent to the union of two disjoint , non - symmetric rays on the @xmath83-line . by applying the first and second variation arguments to a single gaussian density
, we see that two non - symmetric rays are always better than two symmetric rays of the same total weighted - length .
therefore if the isoperimetric curve corresponding to area @xmath13 is a line , then it is a vertical line .
this paper is the work of the williams college nsf `` small '' 2015 geometry group .
we thank our advisor professor morgan for his support .
we would like to thank the national science foundation , williams college , the university of chicago , and the maa for supporting the `` small '' reu and our travel to mathfest 2015 .
john berry department of mathematics and statistics williams college [email protected] matthew dannenberg department of mathematics harvey mudd college [email protected] jason liang department of mathematics university of chicago [email protected] yingyi zeng department of mathematics and computer science st .
mary s college of maryland [email protected] | we consider the isoperimetric problem for the sum of two gaussian densities in the line and the plane .
we prove that the double gaussian isoperimetric regions in the line are rays and that if the double gaussian isoperimetric regions in the plane are half - spaces , then they must be bounded by vertical lines . |
the chiral multiplet of @xmath0 supersymmetry@xcite can be formulated on the lattice so that the supersymmetry is preserved and the vacuum energy is cancelled in the free limit , thanks to the bilinear nature of the free action . by using overlap ( majorana ) fermion @xcite for the fermionic component
, species doublers@xcite are successfully removed and u(1)@xmath3 symmetry can be maintained at the same time@xcite . with this chiral multiplet
, one may formulate lattice @xmath0 wess - zumino model with exact u(1)@xmath3 symmetry@xcite .
a numerical study of this lattice @xmath0 wess - zumino model has recently been reported in @xcite .
the purpose of this short article is , however , to show that in this formulation of the chiral multiplet , the reflection positivity@xcite seems to be violated in the bosonic part of the action , although it is satisfied in the fermionic part , as shown recently in @xcite .
we will also examine the spectral density of the bosonic two - point correlation function ( cf .
it is found that the positivity of the spectral density is ensured only for the momenta @xmath1 @xmath2 , and the mode with a negative density appears at the energy as low as @xmath4 for the momenta @xmath5 .
we will then argue that in formulating the lattice @xmath0 wess - zumino model with the overlap ( majorana ) fermion , one may adopt the simpler nearest - neighbor bosonic action , discarding the free limit manifest supersymmetry .
the model so constructed still preserves the u(1)@xmath3 symmetry and satisfies the reflection positivity .
this paper is organized as follows . in section
[ sec : overlap - chiral - multiplet ] , we review briefly the @xmath0 chiral multiple on the lattice formulated with overlap majorana fermion . in section
[ sec : rf - violation - bosonic - part ] , we show that the standard way to prove the reflection positivity does not work in the bosonic part of the action . the spectral density of the bosonic two - point correlation function is also examined . in section [ sec : rf - lattice - wz - model ] , we show that it is possible to formulate lattice wess - zumino model which possesses both the reflection positivity and the exact u(1)@xmath3 symmetry , by adopting the simpler nearest - neighbor bosonic action .
section [ sec : discussion ] is devoted to discussion .
the action of the free @xmath0 chiral multiplet is given by @xmath6 in this expression , we have used a decomposition of the overlap dirac operator @xcite@xcite , @xmath7 , where @xmath8 and @xmath9 note that @xmath10 and @xmath11 have different spin structures with respect to spinor space . in particular
, we have @xmath12 and @xmath13=0 $ ] . in terms of this decomposition ,
the ginsparg - wilson relation @xmath14 @xcite is expressed as @xmath15 and as a consequence , we have relations @xmath16 it is also understood that the @xmath17 identity matrix in operators @xmath18 and @xmath11 is omitted when these operators are acting on bosonic fields .
it is straightforward to see that the above free action @xmath19 is invariant under lattice @xmath0 supersymmetry " : @xmath20 where @xmath21 is a 4 component grassmann parameter .
we also note that the free action @xmath19 possesses three types of @xmath22 symmetry @xcite .
the first is a rather trivial one acting only on bosonic fields and is defined by the transformation : @xmath23 where @xmath24 is an infinitesimal real parameter .
the second one is nothing but the chiral symmetry introduced by lscher , @xmath25 thirdly , somewhat surprisingly , the _ bosonic _ sector of @xmath19 possesses a @xmath22 symmetry analogous to eq .
( [ twoxseven ] ) : @xmath26 due to the ginsparg - wilson relation . the lattice action @xmath19 is not invariant under a uniform rotation of the complex phase of bosonic fields , @xmath27 , @xmath28 , due to the presence of terms @xmath29 and @xmath30 .
the above provides a lattice counterpart of this uniform phase rotation of bosonic fields under which the free action @xmath19 is invariant . using a linear combination of the above three @xmath22 symmetries , it is possible to define the @xmath31 symmetry @xcite in the interacting system .
in this subsection , we will formulate the reflection positivity condition .
it has been rigorously shown that the lattice theory satisfying the reflection positivity condition corresponds to the quantum theory with unitary time evolution @xcite . here
we consider the generic case in which there are both a bosonic field @xmath27 and a fermionic field @xmath33 .
let us assume that @xmath34 is the action of a lattice model instead of @xmath35 for the notational simplicity .
this notation never means that @xmath36 is an analytic function of @xmath27 . ] and its partition function @xmath37 is given by the path integration @xmath38[{\cal d\psi}{\cal d\bar{\psi } } ] \,{\rm e}^{-s(\phi,\psi,\bar\psi)}.\end{aligned}\ ] ] we set the lattice spacing @xmath39 to be unity , and assume the finite volume hypercubic lattice @xmath40 .
we impose the _ anti_-periodic boundary condition in the time direction for the fermionic field @xmath33 , while the periodic boundary condition for the bosonic field @xmath27 . in the spacial directions ,
periodic boundary conditions are imposed for both fields . to formulate the reflection positivity condition
, we first introduce the time reflection operator @xmath41 as follows .
for each site @xmath42 , we denote @xmath43 .
this is the time reflection with respect to the @xmath44 plane .
we define the operation of @xmath41 for bosonic fields as @xmath45 and for functions of bosonic fields @xmath46 as @xmath47 where * means complex conjugation . for fermionic fields ,
the @xmath41 reflection is defined as @xmath48 we extend this @xmath41 operation to the whole field algebra @xmath49 .
we define the field algebra @xmath49 , the algebra of observables , as the grassmann algebra generated by the fermionic fields with the coefficients of the continuous functions of bosonic fields which are integrable with respect to the bosonic gaussian functional measure . for @xmath50 ,
the @xmath41 operation is defined by the relations @xmath51 for instance , if @xmath52 has the form of @xmath53 its @xmath41 reflection should be @xmath54 let @xmath55 ( resp .
@xmath56 ) be the set of lattice sites with positive ( resp .
non - positive ) time components , and @xmath57 be the subalgebras of @xmath49 , which depends only upon fields on @xmath58 . in this notation , @xmath41 is a map from @xmath58 into @xmath59 and from @xmath57 into @xmath60 .
reflection positivity condition is defined through this @xmath41 map . for a lattice theory with the expectation functional @xmath61 defined for @xmath62 as @xmath63[{\cal d\psi}{\cal d\bar{\psi } } ] \,{\rm e}^{-s(\phi,\psi,\bar\psi)}\mathcal{f}(\phi,\psi,\bar\psi ) , \end{aligned}\ ] ] we say the theory is reflection positive with respect to @xmath41 if any function @xmath64 fulfills the inequality @xmath65 in this subsection , we investigate the reflection positivity of the bosonic sector of the free chiral multiplet .
it will be shown in the following that the bosonic sector does not seem to satisfy the reflection positivity condition . after integrating out the auxiliary field @xmath28
, we have the overlap boson system which is defined through the lattice action on @xmath66 @xmath67 where we have defined @xmath68 and @xmath69 with @xmath70 being the unit spinor matrix .
the operator @xmath71 given above is the bosonic overlap operator on @xmath72 and @xmath73 is that on @xmath66 with periodic boundary conditions .
the field algebra @xmath74 of this overlap boson system is defined as the set of all continuous functions of bosonic field configulations @xmath75 , which are integrable with respect to the bosonic gaussian measure @xmath76[{\cal d\phi^*}]\,{\rm e}^{-s_b(\phi)}.\end{aligned}\ ] ] the expectation of this theory is defined by the bosonic path integration @xmath77[{\cal d\phi^*}]\,{\rm e}^{-s_b(\phi)}\mathcal{f}(\phi ) , \qquad \mathcal{f}\in{\cal a}.\end{aligned}\ ] ] the standard way of investigating the reflection positivity of lattice field theory is to prove that the action can be written in the form of @xmath78 with @xmath79 , where in the third term @xmath80 are elements of @xmath81 parametrized by some discrete parameter @xmath82 @xcite . to see that the equation indeed implies the reflection positivity
, we first note that for an arbitrary @xmath64 , @xmath83[{\cal d\phi^*}]\theta({\cal f}_+){\cal f}_+{\nonumber}\\ & = \int\prod_{x\in\lambda_+}d\phi(x)d\phi^*(x)\,{\cal f}_+(\phi ) \int\prod_{x\in\lambda_-}d\phi(x)d\phi^*(x)\,\theta({\cal f}_+)(\phi ) { \nonumber}\\ & = \left| \int\prod_{x\in\lambda_+}d\phi(x)d\phi^*(x)\,{\cal f}_+(\phi ) \right|^2 \ge 0.\label{zero positive}\end{aligned}\ ] ] if the action is given in the form of , we obtain for all @xmath84 , @xmath85 this last expression is clearly positive from .
this immediately implies the reflection positivity because @xmath86 we note that the third term in may be given by an integration over a continuous parameter @xmath82 as @xmath87 this type of the action appears in the case of overlap fermions .
see ref.@xcite for detail . therefore , to prove the reflection positivity of the ` overlap boson ' system reduces to find the decomposition of the action into .
we first note that @xmath88 can be written as @xmath89 where @xmath90 is the kernel of the operator @xmath71 on @xmath66 . to establish the decomposition
, we should find that ( i ) the second term is the @xmath41 reflection of the first term , and that ( ii ) the last term is written in the form of @xmath91 for some @xmath92 parametrized by some parameter @xmath82 .
note that this second condition is equivalent to say that @xmath93 for some _ non - negative _ function @xmath94 . in this bosonic system , while ( i ) holds true , the property ( ii ) breaks down , as will be shown below . to show this
, we will derive the spectral representation of the kernel @xmath90 .
first , the fourier transformation @xmath95 is given by : @xmath96 ^ 2}}{\nonumber}\\ & = 1+\frac{b({\boldsymbol{p}})-\cos p_0}{\sqrt{a({\boldsymbol{p}})-2b({\boldsymbol{p}})\cos p_0}},\end{aligned}\ ] ] where @xmath97 from this formula we obtain the following three - space representation of the kernel @xmath90 , @xmath98 ^ 3 } \frac{d^3{\boldsymbol{p}}}{(2\pi)^3}\,{\rm e}^{i{\boldsymbol{p}}\cdot ( { \boldsymbol{x}}-{\boldsymbol{y}})}i(x_0-y_0;{\boldsymbol{p } } ) , \ ] ] where we have defined @xmath99 we can transform @xmath100 into a spectral representation by applying cauchy s integration theorem to the contour in the complex @xmath101 plane drawn in the fig.1 . in the fig.1 , the original integration contour for @xmath101 is the interval @xmath102 $ ] on the real axis , and @xmath103 is defined as the positive solution of @xmath104 in the case of @xmath105 , we use the left contour and , in the case of @xmath106 , we use the right . note that the contributions coming from integrations along the edge with infinite ( or minus infinite ) imaginary part ( upper edge in the case of @xmath105 , lower in the case of @xmath106 ) vanishes . furthermore , because of the periodicity of the integrand , the contributions coming from integration along the edge whose real part is @xmath107 ( the right edge of the contour ) and from integration along the edge @xmath108 ( left edge of the contour ) cancel each other . in this way
, the original @xmath101 integration on the interval @xmath102 $ ] can be expressed as an integration on the interval @xmath109 $ ] ( or @xmath110 $ ] ) along the imaginary axis : @xmath111 since the kernel of the operator on the finite lattice @xmath112 is defined as , it is straightforward to derive the following spectral representation of @xmath113 , @xmath114 \frac{b({\boldsymbol{p}})-\cos p_0}{\sqrt{a({\boldsymbol{p}})-2b({\boldsymbol{p}})\cos p_0}}.\label{spectral representation of ddaggerd}\end{aligned}\ ] ] the second term represents a finite volume effect .
now we can see the property does not hold . from eq .
, we obtain @xmath115 where we define @xmath116 in this case , @xmath117 plays a role of the parameter @xmath82 in . for the condition to be satisfied
, the coefficient factor @xmath118 should be non - negative for any @xmath117 satisfying @xmath119 , but this is _ not _ the case .
in fact , @xmath118 can become _ both _ positive and negative in general , which prevents us from proving the reflection positivity . in the previous section
, we have shown that the standard way of proving the reflection positivity does not work for the overlap boson . in this section
, we will investigate the klln - lehmann representation of the propagator . here
, we will observe that the spectral density function , which is expected to be positive ( non - negative ) for unitary quantum theories , is not a positive function .
this implies that the overlap boson system has pathological spectrum of energy momentum operators .
the spectral density @xmath120 is geven in the the euclidean version of klln - lehmann representation by @xmath121 in the present case , one can explicitly estimate the propagator and the spectral density @xmath120 by using the fourier transformation , @xmath122 the result is @xmath123 this formula is derived in appendix [ app ] .
the second term , continuous spectrum , is not positive ( non - negative ) because of the factor @xmath118 , which is exactly the same one as appeared in .
the fact that the positivity of the spectral density breaks down is an indirect but strong circumstantial evidence that the overlap boson system does not define a quantum mechanical system with physically satisfactory energy - momentum spectrum .
further , it is very probable that the overlap boson system breaks the reflection positivity condition .
in fact , one can prove in mathematically rigorous manner @xcite that a lattice theory satisfying the reflection positivity condition , in addition to some technical assumptions which are satisfied by the overlap boson system , must have the non - negative spectral density .
the proof goes as follows . for a detail , see @xcite .
suppose that the reflection positivity is satisfied , and take the infinite volume limit @xmath124 .
then , from the reflection positivity condition , the hilbert space of state vectors is constructed and for each @xmath125 , momentum operator @xmath126 acting on this hilbert space is defined as an infinitesimal self adjoint generator of translation in each direction . we denote @xmath127 and @xmath128 .
since translations in different directions commute , @xmath126 and @xmath129 commute with each other if @xmath130 . therefore @xmath131 possesses complete orthonormal set of simultaneous eigenvectors @xmath132 with @xmath133 , @xmath134 .
let @xmath135 be @xmath136 .
through momentum operators @xmath137 , field operator at @xmath138 , @xmath139 is related to the field operator at the origin @xmath140 by the relation @xmath141 the two point function is expressed in terms of these ingredients of hilbert space , @xmath142 by the translational invariance and , and by inserting the identity in the form of @xmath143 we obtain @xmath144 where the integration measure is formally given by @xmath145 even though the above computation is rather formal , the measure has mathematically rigorous meaning as a @xmath146-dimensional borel measure supported on @xmath147 .
in fact , one can derive in mathematically rigorous manner .
it is well known in measure theory that any measure @xmath148 on @xmath149 can be uniquely decomposed into two parts @xmath150 where @xmath151 is singular and @xmath152 is absolutely continuous with respect to the lebesgue measure on @xmath149 ( lebesgue decomposition theorem ) . in more ordinary expression in physics , this theorem states that if we write @xmath153 @xmath120 can be written corresponding to as @xmath154 where @xmath155 is _ non - negative _ integrable function .
this is not the case for the overlap boson . from the explicit form of the spectral density
, we can find where in the brillouin zone the reflection positivity is violated .
one notes that there is the region in the spacial brillouin zone where the spectral density @xmath120 can not become negative .
let us call this region @xmath156 .
the region @xmath156 is characterized by the condition that the negative value of @xmath118 be avoided .
the necessary and sufficient condition on spacial momenta @xmath157 to avoid negative @xmath118 is that @xmath158 which is equivalent to the condition @xmath159 @xmath160 or , equivalently , @xmath161 then , define @xmath162 where @xmath146 is the spacetime dimension .
now let us estimate the size of @xmath156 to investigate whether we can ignore the violation of the reflection positivity or not .
in the case of @xmath163 , there are three spacial momentum components .
first , we consider the case in which @xmath164 . in this direction , the safe momentum region has the extent @xmath165 second , we consider another direction @xmath166 . in this case , in the safe region @xmath156 , @xmath167 is restricted by @xmath168 these regions are a little bit lager than @xmath169^{d-1}$ ] .
when the spacial momenta @xmath157 does not belong to @xmath156 , the spectral density @xmath120 has to become negative on the energy interval @xmath170 , where @xmath103 and @xmath171 are determined by @xmath172 since @xmath173 is equivalent to @xmath174 when @xmath175 .
we will numerically estimate @xmath103 and @xmath171 , the lower and upper bound of the energy interval on which the spectral density become negative .
for instance , if @xmath163 , these energy values are computed as shown in the following table : [ cols="^,^,^,^,^,^",options="header " , ] whether these values are large enough or not should depend on the physics one wants to see through the overlap boson .
to remedy the violation of the reflection positivity , one may adopt the simpler nearest - neighbor action for the boson fields , @xmath27 and @xmath28 as follows .
of course this does not change any physical results .
it is simply because this convention has been used in the proof of the reflection positivity for the overlap fermions in our previous work @xcite . ] : @xmath176 this action still possesses three types of @xmath22 symmetry , eq .
( [ twoxfifteen ] ) , ( [ twoxseven ] ) and @xmath177 instead of eq .
( [ twoxsixteen ] ) . in this formulation of the chiral multiplet
, the action of the lattice @xmath0 wess - zumino model may be given as follows : @xmath178 where @xmath179 is an auxiliary majorana fermion field and @xmath180 .
then one may define the @xmath31 symmetry as follows : @xmath181 the reflection positivity is now satisfied in this formulation of the wess - zumino model .
the @xmath41-reflection is defined for the bosonic fields @xmath182 in the same way as in the generic case , @xmath183 and for the fermionic fields @xmath184 as in , @xmath185 note that this definition of @xmath41 reflection does not contradict to the majorana conditions @xmath186 and @xmath187 .
our field algebra @xmath74 here is that of the polynomial algebra of fermionic fields whose coefficients are the well - behaved functions of the bosonic fields .
we extend @xmath41 operation to whole algebra @xmath74 , by the relations and . to prove the reflection positivity of the wess - zumino model , it is sufficient to show that the action can be rewritten in the form of @xmath188 @xcite
. let us first consider the free part of .
the first term in , the overlap majorana fermion , can be written in the form of as is shown in ref @xcite . on the other hand ,
the second term can be written in the form of , as is well - known .
furthermore , the third and fourth terms in is @xmath41 reflection of themselves and do not contain any ` time hopping terms ' . therefore , these terms can be written in the form of @xmath189.\end{aligned}\ ] ] the rest of the terms in are interaction terms , @xmath190 which are all strictly local .
they are equal to theta - reflection of themselves again , and do not contain any nonlocal ` time hopping ' terms either .
this means that @xmath191 can also be written as @xmath192 with @xmath193 which obviously belongs to @xmath194 .
therefore , one concludes that this lattice wess - zumino model satisfies the reflection positivity condition .
preserving r symmetry exactly is a useful way in formulating supersymmetric field theories on the lattice .
this point has been emphasized by elliot , giedt and moore @xcite in their formulation of four - dimensional @xmath195 super yang - mills theory .
the discrete r symmetry in the two - dimensional @xmath196 wess - zumino model @xcite has played an important role in the numerical study of the correspondence to @xmath196 conformal field theories@xcite . in formulating the exact r symmetry on the lattice
, however , there is a freedom in the choice of the bosonic part of the action .
when one can preserve some part of the extended supersymmetries in the theories with @xmath197 @xcite , it seems useful to adopt the bosonic actions to preserve the supersymmetries , although one should take into care a possible effect of the violation of the reflection positivity .
but , for the theories of @xmath198 , it seems difficult to preserve the supersymmetry in general@xcite , and the free limit supersymmetry does not necessarily help in taking the supersymmetric limit in the interacting models . in such situations , thought , if one can preserve the fundamental requirement of the reflection positivity condition , it may serve as a possible guideline to choose a bosonic action
. it would be interesting to examine further the inter - relation among the reflection positivity , the vacuum energy cancellation(the exact supersymmetry ) and the exact u(1)@xmath3 symmetry of free chiral multiplet on the lattice .
if one adopts the majorana wilson fermion for the fermionic component of the chiral multiplet , one can show that the bosonic part of the supersymmetric action now fulfills the reflection positivity condition . in this case , the u(1)@xmath3 symmetry is not manifest .
but , through the block spin transformation , it is recovered in the fixed point action @xcite . in this course of the renormalization group transformatons
, it seems possible to maintain the vacuum energy cancellation by adjusting the parameters in the block - spin kernels and the normalization factors .
then , if the reflection positivity could also be maintained through the block - spin transformation , all the three conditions could be fulfills in the fixed point approach@xcite .
k.u . would like to thank tsutomu t. yanagida for continuous encouragement .
he also thanks kenji maeda for valuable discussions .
is supported by global coe program the physical science frontier " , mext , japan .
this work was supported by world premier international center initiative ( wpi program ) , mext , japan .
y.k . is supported in part by grant - in - aid for scientific research no . 21540258 , 21105503 .
in this appendix , we derive the formula for @xmath163 . to find the explicit form of the spectral density @xmath199
, we express the propagator by its fourier transformation : @xmath200 in the following analysis , we will apply cauchy s theorem to the @xmath101 integration @xmath201 note that though the integration is well defined , the integrand function of @xmath157 in : @xmath202 is not defined at @xmath203 , because at @xmath204 this @xmath101 integration does not exist . since the value of the integrand on a set of zero measure @xmath205 does not contribute the integration , we may assume @xmath206 ( mod @xmath207 ) in when applying cauchy s theorem .
define a function @xmath208 as @xmath209 from eq .
, one finds @xmath210 where , for notational simplicity , we have written @xmath211 and @xmath212 . here , we have to clarify the meaning of the square root of complex variables .
we define the square root of @xmath213 as @xmath214 namely , we choose the branch where @xmath215 .
first , since the square root of negative real number can not be defined , @xmath216 can not be defined where @xmath217 to find the more explicit condition which is equivalent to , put @xmath218 . since @xmath219 , is equivalent to @xmath220 which holds true when and only when @xmath221 the second condition is equivalent to @xmath222 but the second choice @xmath223 is impossible because in this case the first condition of becomes @xmath224 which is never true for real @xmath225 .
therefore , is equivalent to @xmath226 , @xmath227 , and @xmath228 hence the condition occurs when and only when @xmath229 and @xmath230 next , we investigate the pole type singularity of @xmath216 which may appear where the denominator @xmath231 vanishes . to find the _ necessary _ condition of @xmath232 ,
let us assume @xmath232 .
then , by taking the square of the both sides of @xmath233 one finds @xmath234 using the identity @xmath235 and putting @xmath236 again , one arrives at @xmath237 and @xmath238 eq . implies @xmath239 and we consider both cases respectively . in the first case , @xmath223 , becomes @xmath240 which is possible only when @xmath241 , equivalently , @xmath242 therefore , this case @xmath223 occurs only when the spacial momentum satisfies @xmath243 as we noted at the beginning of this appendix , we may assume at least one of the @xmath244 s ( @xmath245 ) is odd ( otherwise the @xmath101 integration is ill defined ) .
if is satisfied , the right hand side of becomes @xmath246 and implies @xmath247 but , this condition , @xmath223 and @xmath248 , is not sufficient for @xmath232 .
in fact , for @xmath227 , @xmath249 because @xmath250 due to the fact that at least one of @xmath251 s is equal to @xmath252 . in the second case of , @xmath253 , becomes @xmath254 then , this implies that @xmath255 is even and @xmath256 define @xmath257 as @xmath258 and we obtain as a necessary condition for @xmath232 , @xmath259 . to find a sufficient condition for @xmath232 , let us assume , conversely , when @xmath260 , then , @xmath261 hence , the necessary and sufficient condition for @xmath262 is @xmath255 is even and @xmath263 which is equivalent to @xmath264 namely , @xmath265 . for a moment ,
let us assume spacial momentum @xmath157 satisfies @xmath268 in this case , @xmath269 is a simple pole of @xmath216 , as will be seen . from the above argument
, @xmath216 is analytic on @xmath270 expand @xmath271 in taylor series around @xmath272 : @xmath273 on @xmath274 for sufficiently small @xmath275 , and we obtain @xmath276 on @xmath274 . then , we find @xmath277 is analytic on @xmath274 , and then @xmath278 are simple poles of @xmath216 with residues @xmath279 applying cauchy s theorem on the contour drawn in fig.2 . , we obtain , for @xmath105 , @xmath280}\frac{dp_0}{2\pi}\,{\rm e}^{ip_0 x_0 } \frac{1}{\box}(p_0,{\boldsymbol{p}}){\nonumber}\\ = & \int_{[-\pi,\pi]}\frac{dz}{2\pi}\,{\rm e}^{izx_0}f(z){\nonumber}\\ = & 2\pi i \,\text{res}(f , z_0)-\bigg(\int_{i\infty+0}^{ie_1 + 0}+\int_{i\infty-0}^{ie_1 - 0}\bigg)\frac{dz}{2\pi}\,{\rm e}^{iz x_0}f(z).\end{aligned}\ ] ] recalling our definition of the square root , one finds @xmath281 then , the integrations of the second term in is computed by putting @xmath282 to become @xmath283 by substituting and into , we arrive at @xmath284}\frac{dp_0}{2\pi}\,{\rm e}^{ip_0 x_0 } \frac{1}{\box}(p_0,{\boldsymbol{p}}){\nonumber}\\ = & 2\pi \frac{a-2b\sqrt{a - b^2}}{\sqrt{a - b^2}\sqrt{a - b^2 - 1}}{\nonumber}\\ & \qquad+\int_{e_1}^\infty\frac{de}{\pi}{\rm e}^{-ex_0}\frac{(b-\cosh e)\sqrt{2b\cosh e - a}}{\cosh^2 e - a+b^2},\end{aligned}\ ] ] in agreement with . considering the case where spacial momentum @xmath157 satisfies @xmath285
we find that there is no pole term and only the second term of survives .
note that , in the case of equality , even though @xmath232 , @xmath216 has no isolated pole . in this case
, the numerator of @xmath208 also vanishes and @xmath286 .
j. wess and b. zumino , phys .
b * 49 * , 52 ( 1974 ) .
h. neuberger , phys .
b * 417 * , 141 ( 1998 ) .
h. neuberger , phys .
b * 427 * , 353 ( 1998 ) .
y. kikukawa and h. neuberger , nucl .
b * 513 * , 735 ( 1998 ) [ arxiv : hep - lat/9707016 ] .
l. h. karsten and j. smit , nucl .
b * 183 * , 103 ( 1981 ) .
h. b. nielsen and m. ninomiya , nucl .
b * 185 * , 20 ( 1981 ) [ erratum - ibid .
b * 195 * , 541 ( 1982 ) ] .
h. b. nielsen and m. ninomiya , nucl .
b * 193 * , 173 ( 1981 ) .
t. aoyama and y. kikukawa , phys .
d * 59 * , 054507 ( 1999 ) [ arxiv : hep - lat/9803016 ] .
k. fujikawa and m. ishibashi , nucl .
b * 622 * , 115 ( 2002 ) [ arxiv : hep - th/0109156 ] .
k. fujikawa and m. ishibashi , phys .
b * 528 * , 295 ( 2002 ) [ arxiv : hep - lat/0112050 ] .
k. fujikawa , nucl .
b * 636 * , 80 ( 2002 ) [ arxiv : hep - th/0205095 ] .
m. bonini and a. feo , jhep * 0409 * , 011 ( 2004 ) [ arxiv : hep - lat/0402034 ] .
y. kikukawa and h. suzuki , jhep * 0502 * , 012 ( 2005 ) [ arxiv : hep - lat/0412042 ] . m. bonini and a. feo , phys .
d * 71 * , 114512 ( 2005 ) [ arxiv : hep - lat/0504010 ] .
c. chen , e. dzienkowski and j. giedt , arxiv:1005.3276 [ hep - lat ] .
k. osterwalder and r. schrader , commun .
phys . *
31 * , 83 ( 1973 ) .
k. osterwalder and r. schrader , commun .
phys . * 42 * , 281 ( 1975 ) .
k. osterwalder and e. seiler , annals phys .
* 110 * , 440 ( 1978 ) .
m. luscher , commun .
phys . * 54 * , 283 ( 1977 ) .
p. menotti and a. pelissetto , nucl .
suppl . * 4 * , 644 ( 1988 ) .
y. kikukawa and k. usui , phys .
d * 82 * , 114503 ( 2010 ) [ arxiv:1005.3751 [ hep - lat ] ] .
m. luscher , arxiv : hep - th/0102028 .
k. usui , in preparation .
j. w. elliott , j. giedt and g. d. moore , phys .
d * 78 * , 081701 ( 2008 ) [ arxiv:0806.0013 [ hep - lat ] ] .
y. kikukawa and y. nakayama , phys .
d * 66 * , 094508 ( 2002 ) .
h. kawai and y. kikukawa , phys .
d * 83 * , 074502 ( 2011 ) [ arxiv:1005.4671 [ hep - lat ] ] .
y. kikukawa and f. sugino , nucl .
b * 819 * , 76 ( 2009 ) [ arxiv:0811.0916 [ hep - lat ] ] .
m. kato , m. sakamoto and h. so , jhep * 0805 * , 057 ( 2008 ) .
p. hasenfratz , v. laliena and f. niedermayer , phys .
b * 427 * , 125 ( 1998 ) [ arxiv : hep - lat/9801021 ] .
p. hasenfratz , nucl .
b * 525 * , 401 ( 1998 ) [ arxiv : hep - lat/9802007 ] . | by using overlap majorana fermions , the @xmath0 chiral multiple can be formulated so that the supersymmetry is manifest and the vacuum energy is cancelled in the free limit , thanks to the bilinear nature of the free action .
it is pointed out , however , that in this formulation the reflection positivity seems to be violated in the bosonic part of the action , although it is satisfied in the fermionic part
. it is found that the positivity of the spectral density of the bosonic two - point correlation function is ensured only for the spacial momenta @xmath1 @xmath2 .
it is then argued that in formulating @xmath0 wess - zumino model with the overlap majorana fermion , one may adopt a simpler nearest - neighbor bosonic action , discarding the free limit manifest supersymmetry .
the model still preserves the would - be u(1)@xmath3 symmetry and satisfies the reflection positivity . |
the detailed circuit diagram we use for our experiments is sketched in fig.s1 .
all the electronic components for controlling the coupling are placed on a pcb ( printed circuit board ) that is shielded with a metal box .
the output of the pcb goes to the electrical feedthroughs of the vacuum chamber in which the membrane - capacitor system is placed .
the decay rate of the lc circuit is a crucial parameter for reaching the strong coupling regime since the coupling rate should exceed this rate . in our setup
, we use a high - q ferrite inductor which is made of ferrite 61 material ( including a mixture of zinc and nickel ) that is known for minimized core and eddy current losses .
the inductor is wound with litz wires of optimal wire diameter for low loss in the frequency range of interest ( @xmath137 khz ) . even though the q of our inductor reaches 500 around this frequency range , our combined circuit exhibits lower q due to several elements .
we have observed that connection of the pcb ( dielectric losses ) and nearby magnetic elements ( optical table , shielding box etc ) contribute to the reduction of the q - factor .
a significant reduction occurs when the whole lc circuit is connected to the membrane - capacitor chip inside the chamber .
this comes mainly from the contact resistance ( @xmath138 @xmath139 ) of the gold electrode connection lines deposited on the capacitor chip . in the mhz range ,
antennas pick up radiation that permeate the earth s atmosphere , caused mainly by lightnings , but also man - made , and at higher frequencies , galactic sources @xcite .
litz wires shielded with aluminium foil are used to reduce pick - up in the cables ( @xmath140 ) connecting the box containing the printed circuit board control unit with the inductor .
the inductor / antenna is placed in a metal box to reduce the amplitude of the large atmospheric pickup in order to better demonstrate the sensitivity of the transducer .
still , the radiation picked up generates voltages largely exceeding the johnson noise in the circuit .
we have measured the scaling of the mechanically transduced pickup signals as a function of bias voltage ( fig .
[ fig : excessnoise ] ) .
the noise amplitude shows scaling which would be expected for a signal coupled through the inductor port according to our model .
the nonlinear function used for the fit is described by @xmath141 where @xmath142 is a proportionality constant for @xmath143 .
the function is derived from eq .
( [ e : displacement ] ) for the case of resonant driving ( @xmath144 ) with the assumption that white noise ( @xmath85 ) is the dominant source throughout the relevant frequency range .
the capacitive transducer in our setup consists of a 4-segment electrode and a metal coated membrane .
microscope images of both the chip and the membrane are shown in fig.s3 .
the chip has an area roughly four times larger than that of the membrane .
the two pairs of segments are connected via extra electrodes such that the diagonal parts form ( + ) and ( - ) electrodes .
we note that the overlap between the membrane area and the 4-segment electrodes after the assembly is not perfectly symmetric .
the initial distance between the membrane and the capacitor chip is paramount in order to achieve high coupling rates at small dc bias voltages . for our setup ,
the distance is set by the pillar height ( 600 nm and @xmath145 ) deposited on the glass chip . however , most membrane - chip samples showed distances of a few @xmath146 .
this is attributed to the unevenness of the membrane and chip surface , as well as residues ( silicon , resist , dust ) acquired during the assembly process . to overcome this problem
, we used an extra etching process for a batch of our membranes to remove a 50 @xmath146 thick layer of a large area within the membrane frame surface .
this technique helped us reduce the distance closer to the pre - determined value .
as shown in fig .
[ fig : whitelight ] , the distance was experimentally determined using white light interferometry .
after assembly of the membrane - electrode system , we typically measured the linewidth of the thermally driven mechanical mode using the optical readout .
typically values of @xmath147 were observed . for some membranes
, we used ringdown measurements for a more accurate determination of the mechanical quality factor .
to that end , the membrane was excited through the capacitor using a small applied voltage .
the membrane motion was again monitored optically , and its decay recorded after the excitation was switched off .
figure [ fig : ringdown ] shows an example of such a measurement , with a fitted ( amplitude ) ringdown time of @xmath148 , corresponding to a linewidth of @xmath149 and @xmath150 .
when arranging the membrane very close to the electrode surface ( @xmath151 ) , we typically found a reduction of the membrane quality factor , with mechanical linewidths often on the order of @xmath152 .
the precise origin of this increased damping sill has to be better understood , but could possibly arise from surface charge - membrane interaction or , for very small distances , residues getting in contact with the membrane .
the increased damping limits the cooperativity in measurements with the highest coupling , such as the one shown in fig . 1 of the main manuscript . in this section ,
we explore the limits of our device based on the maximum achievable cooperativity .
cooperativity ( @xmath153 ) is in principle limited by the maximum applicable dc voltage that yields a static displacement of the membrane towards the electrode direction .
this voltage , known as the pull - in voltage for capacitive systems@xcite , can be calculated by using the scaling of the capacitive force ( @xmath154 ) and the equation @xmath155 at the equilibrium point , where @xmath89 and @xmath156 refer to the intrinsic spring constant and membrane - electrode distance , respectively .
this expression for the electrostatically modified spring constant goes to zero if the membrane is displaced by @xmath157 due to the applied dc voltage , causing instability and eventual collapse of the membrane on the capacitor chip . considering this limit , we implement the following scaling argument for our experimental data in fig.4 . assuming a homogenous force and neglecting the deflection geometry of the membrane , frequency shift and static displacement
can be related by @xmath158 which sets the maximum allowable frequency shift as @xmath159 khz where @xmath160 khz is the bare membrane frequency .
given the empirically found frequency shift of 88 khz with 21v dc bias at a distance of @xmath161 m and the frequency shift scaling @xmath162 , the applied voltage can be increased by a factor of @xmath163 which leads to a 3 times improved cooperativity of @xmath164 . consequently , this would correspond to a sensitivity limit of 2.9 pv@xmath165 .
the hamiltonian @xcite @xmath166 describes the electromechanical system in terms of the flux @xmath25 in the inductor , the charge @xmath26 on the capacitors , as well as the displacement @xmath27 and momentum @xmath28 of the considered fundamental membrane mode . in this scalar description
the coupling arises due to the displacement dependence of the capacitance @xmath167 .
the hamiltonian directly leads to the langevin equations @xmath168 in which the dissipation rate @xmath169 of the membrane , and @xmath170 of the resonance circuit have been included , as well as driving terms @xmath171 ( a force on the membrane ) and @xmath172 ( a voltage induced in the circuit ) . assuming that the only acting force is the thermal langevin force @xmath173 , and writing the applied voltage as a large d.c .
offset and a small fluctuating input @xmath174 we can linearize the langevin equations around an equilibrium @xmath175 characterised by @xmath176 and , to first order in the small fluctuations ( provided the system is stable ) , @xmath177 here , we have introduced the coupling parameter @xmath178 as well as the frequency shift @xmath179 absorbing the frequency shift into a re - defined @xmath180 and transforming to the fourier domain yields @xmath181 these algebraic equations can be used to calculate the response of the system to excitations through a force or voltage drive . for notational convenience
, we often use the bare susceptibilities @xmath182 of the mechanical and @xmath13 resonator , respectively .
the system ( [ e : eomlfx])-([e : eomlfphi ] ) represents two oscillators coupled with a coupling energy @xmath183 where @xmath184 and @xmath185 are the zero - point fluctuations of the membrane displacement and capacitor charge , respectively .
comparison of the corresponding coupling rate @xmath40 with the dissipation rate yields the electromechanical cooperativity @xcite @xmath186 the voltage across the capacitor(s ) , measured at port `` 1 '' of the circuit , is given by @xmath187 after again applying linearisation .
together with ( [ e : eomlfx])-([e : eomlfphi ] ) this yields @xmath188 if we assume the voltage is most significantly affected by a voltage @xmath189 induced in the circuit , which is the case when the circuit is driven through a sufficient strong voltage pickup ( port `` 2 '' ) .
figure [ f : mitcapacitor ] shows typical expected response curves , showing the transition from mechanically induced transparency to normal mode splitting .
, @xmath190 , @xmath191 , @xmath192 , @xmath193 , @xmath194 , and @xmath195 for @xmath196blue , violet , yellow , green@xmath197 curves , in the right panel , the transmission on resonance @xmath198 is shown as a function of @xmath60 . ] the optical field reflected from the membrane experiences a phase shift of @xmath199 where @xmath200 is the readout light s wavenumber . from ( [ e :
eomlfx])-([e : eomlfphi ] ) we have @xmath201 for the mechanical displacement , so that the total measured optical phase shift is @xmath202 + \delta \varphi_\mathrm{im}({\ensuremath{\omega } } ) , \label{e : phaseshift}\end{aligned}\ ] ] where we have taken the imprecision @xmath203 of the measurement into account , but neglected optical backaction @xcite , which would only become relevant for much greater optical powers than the ones used here ( @xmath204 ) . from eq .
( [ e : phaseshift ] ) , the spectral density of phase fluctuations is given by @xmath205 for uncorrelated inputs @xmath206 , @xmath207 and @xmath203 . a voltage signal can therefore be detected as soon as its spectral density exceeds both the optical readout noise @xmath208 as well as the noise due to thermal fluctuations of the membrane position @xmath209 the condition @xmath210 can be achieved by increasing the optical power and thus reducing the shot noise limit of @xmath211 .
underthese conditions , with the spectral density of the langevin force @xmath212 , the fundamental limit for the sensitivity can be written as @xmath213 in the context of efficient light modulation by an rf field , the coupling strength can be quantified in terms of the voltage @xmath72 , which , if applied to the lc circuit , shifts the reflected light s phase by @xmath73 .
focusing on the resonant response , we find from eq .
( [ e : phaseshift ] ) @xmath214 which is maximum at @xmath215 , yielding @xmath216 in this case .
it is straightforward to estimate the conversion efficiency of radio - frequency photons to optical sideband photons of the light reflected from the membrane , @xmath217 note that as opposed to @xmath72 where @xmath218 was found to be optimum , high conversion efficiencies can be achieved for large cooperativities as the impedance of the circuit is increased and thus less power is provided by the source for a given source voltage @xmath22 . for sufficiently high cooperativity , @xmath219 , and with @xmath220 , @xmath221 , @xmath222 , @xmath223 ,
we find @xmath224 .
the bare membrane mechanical characteristics can be extracted from a measurement of its displacement when thermally driven only , in absence of electromechanical coupling ( @xmath225 ) . to this end
, we use a doppler vibrometer which provides a calibrated displacement spectrum .
a simple lorentzian model @xmath226 is fit to this spectrum and used to extract the mechanical damping @xmath227 and the effective mass @xmath29 ( assuming @xmath228 ) .
note that this mass is an _ effective optical mass _ , that is , it may vary slightly depending on the probing point of the optical beam . the variations we observe
are on the order of @xmath229 .
the response of the electromechanical systems to a series voltage picked up in the inductor ( port `` 2 '' ) , and measured within the circuit , at the capacitor ( port `` 1 '' ) , is described by eq . ( [ e : mitcapacitor ] ) . here
, we have assumed that the signal is dominated by the induced voltage .
the corresponding model @xmath230 is fit to the response functions measured in this fashion , and used to extract the parameters @xmath180 , @xmath231 , @xmath232 , @xmath60 and @xmath189 ( the amplitude of the induced driving voltage ) .
the required total capacitance has been determined beforehand in independent measurements , and is typically at a level of @xmath233 .
note that in this approach , the electromechanical coupling parameter @xmath60 depends indirectly on the ( optically determined ) effective mass @xmath29 , as the actual electromechanical coupling physics ( and thus the shape of the spectra , e. g. splitting ) is determined by the coupling energy @xmath234 , which is independent of the optical probing point .
the response of the optical phase shift to a voltage pickup in port `` 2 '' is readily described by eq .
( [ e : phaseshift ] ) , in the limiting case where the driving voltage dominates the signal , @xmath235 by performing a calibration of the spectra into mechanical displacement units , a simplified fit model of @xmath236 can be used , with input parameters @xmath29 and @xmath237 from independent measurements , as well as @xmath232 as determined by the fits of the voltage measurements at port `` 1 '' ( see above ) .
the parameters @xmath180 , @xmath231 , @xmath60 and @xmath238 are adjusted by the fit .
the results for @xmath180 , @xmath231 , and , importantly , @xmath60 , typically agree to the fits to the voltage spectra to within @xmath61 .
as described in the previous section , fits to the spectra of the intracircuit voltage ( port `` 1 '' ) or interferometrically measured optical phase shift yield values of the coupling parameter @xmath60 .
note that determination of @xmath60 is performed by fitting the _ shape _ of the spectra with the amplitude as a free parameter .
the amplitudes can be used in an additional approach to determine @xmath60 ( see below ) .
the parameter @xmath60 can also be re - expressed into a coupling rate @xmath40 using the identity @xmath239 and thus made independent of the optical mass @xmath29 .
alternatively , the _ amplitudes _ of the voltage in the circuit @xmath242 and the mechanical displacement @xmath243 can be compared .
for example , at resonance , we find @xmath244 encoding @xmath60 as a function of the experimentally measured modulation amplitudes . in practice
it turned out to be more robust to evaluate this ratio not at one particular frequency ( here e. g. @xmath245 ) , but throughout the entire measured frequency range .
this can be conveniently accomplished by evaluating the left - hand - side of eq .
( [ e : ampratio ] ) using the models ( [ e : fitmodelv ] ) and ( [ e : fitmodelx ] ) , with the parameters fitted to the experimental spectra , including crucially the ( possibly ) different drive amplitudes @xmath189 determined in these fits , which uniquely encode the amplitudes of the measured spectra .
for the driven measurements with the lock - in ( in units of vrms ) , we implement the following approach to calibrate the readings to mechanical amplitudes ( shown in fig.2 ) .
the membrane is first excited with a single frequency sine wave of known amplitude from the rf output of the lock - in via a drive coil .
this appears as a sharp mechanical response on the displacement spectrum as measured with the vibrometer , which is by default calibrated in meters .
this value is then used to calibrate the wide membrane spectrum when we scan the rf excitation signal and record the driven response of the vibrometer output which is fed to the lock - in . a third , independent approach to estimating the coupling strength is by measuring the frequency shift when a voltage is applied .
above we found @xmath247 where we have chosen canonical coordinates such that the oscillator mass equals the physical membrane mass @xmath248 .
to relate the frequency shift and the coupling rate , a relation of the first and second derivative of the inverse capacitance at the operating point @xmath32 is necessary . without loss of generality , we can write @xmath249 with some characteristic length @xmath250
. then we obtain @xmath251 an approximation for @xmath250 can be obtained by assuming @xmath252 where @xmath156 is an equilibrium distance , e. g. between the plates of a capacitor , one of which can move .
this model yields @xmath253 for a large capacitance offset @xmath254 .
more generally , for any capacitive force for which @xmath255 holds approximately , one finds @xmath253 . for a general choice of gauge for the oscillator mass @xmath256 ( accompanied by a suitable rescaling of @xmath27 ) , the chain rule yields @xmath257 ; thus @xmath258 this expression can be used to estimate the coupling parameter @xmath60 by measuring the frequency shift @xmath259 , if the distance @xmath156 , effective and physical masses @xmath29 and @xmath248 , total capacitance @xmath260 and applied bias voltage @xmath21 are known .
the total capacitance of the membrane - electrode arrangement depends on the membrane mode shape and the electrode mask as well as their relative alignment .
this capacitance can be determined analytically in a quasi - electrostatic calculation by employing certain geometrically motivated approximations @xcite .
we proceed by assuming that the curvature of the membrane is sufficiently small that we , for purposes of calculating the capacitance , can take it to be locally flat .
if we furthermore neglect edge effects , symmetry considerations imply that we may model the membrane - electrode capacitance _ locally _ as that of parallel plates ( assumed to be perfectly conducting ) .
this local capacitance per area only depends on the local membrane - electrode separation along the direction normal to the plane defined by the electrodes . within the stated approximations
, we can now construct the total capacitance of the arrangement in terms of the local capacitance contributions .
noting that the membrane - electrode separation is significantly smaller than the inter - electrode gaps of the quarter - segment geometry ( see fig .
[ fig : microscope]a ) , we can neglect the direct capacitance between electrode segments . instead
, the capacitance between the positive and negative electrodes is mediated by the membrane , so that the integrated local capacitance between membrane and positive electrode segments , @xmath262 , acts in series with the corresponding quantity for the negative electrode segments , @xmath263 : @xmath264 where @xmath9 is the tuning capacitance acting in parallel with the membrane modulated capacitance . with this model of the total capacitance @xmath237
, we may proceed to calculate the coupling strength @xmath60 along the lines of the derivation given in subsection [ sub : theory_emcoupling ] .
the displacement configuration of the membrane @xmath265 relative to its static deflection ( induced by the dc voltage bias ) can be expanded on the set of drum modes ( where @xmath266 and @xmath267 are cartesian coordinates of the membrane plane ) : @xmath268 the coefficients @xmath269 are canonical drum mode position coordinates each of which has the free evolution of a harmonic oscillator .
their frequencies are given by the drum mode spectrum , while we have a gauge freedom of choosing their masses by means of canonical rescaling of the @xmath269 and their conjugate momenta .
this also implies a rescaling of the modes @xmath270 in order to retain the form of eq .
( [ e : drum_mode_expansion ] ) ( note that the coordinate axis for @xmath271 is fixed once and for all in specifying the membrane - electrode distance @xmath156 in absence of bias voltage ) .
that is , focusing on a single drum mode @xmath272 henceforth , the choice of gauge mass @xmath29 for the canonical variables associated with this mode entails the requirement @xmath273 where @xmath248 is the physical mass of the membrane and @xmath274 is the membrane area .
in parallel to eq .
( [ e : g_def ] ) we write @xmath275}\right)\right|_{\text{eq}},\ ] ] where eq signifies that the derivative should be evaluated at the static displacement equilibrium configuration of the membrane
. however , assuming the static displacement to be negligible compared to @xmath156 , it is a good approximation to evaluate eq .
( [ e : g_beta ] ) at the flat membrane configuration . inserting our model for @xmath237 , eq .
( [ e : c_series_model ] ) , we find @xmath276 ^ 2 } + \frac{o^{(1)}_{-}}{[o^{(0)}_{-}]^2}}{\left(\frac{1}{o^{(0)}_{+ } } + \frac{1}{o^{(0)}_{- } } \right)^2}\right],\ ] ] where @xmath277 is the membrane side length and we have introduced the overlap factors @xmath278 with @xmath279 and @xmath280 being the area of the membrane above electrodes of polarity @xmath281 ; @xmath282 is a product of heaviside functions that masks the electrode gaps as well as the hole in the membrane metalization ( see fig .
[ fig : microscope ] ) . in order to evaluate @xmath60
, we need to calculate the overlap term appearing in brackets in eq .
( [ e : g_beta_result ] ) for our geometry . for the fundamental mode we find the value @xmath283 provided perfectly central positioning of the membrane with respect to the quarter - segment electrodes . in a range of up to 25% lateral misalignment
the value varies between @xmath284 and @xmath285 with an ( unweighted ) average of @xmath286 , which is the value we will use below .
the physical mass can be estimated to @xmath287 using the mass densities @xmath288 and @xmath289 . the membrane side length @xmath277 is @xmath290 . for the specific experiment presented in fig . 2 of the main text
, we have effective mass @xmath291 and membrane - electrode separation @xmath292 . as the lc mode is retuned into resonance with the mechanical mode each time the bias voltage @xmath293 is modified , eq .
( [ e : g_beta_result ] ) acquires an additional @xmath293 dependence through @xmath294 , where @xmath9 is the tuning capacitance .
however , this amounts to a correction of a few percent over the voltage range considered in the aforementioned figure ; for simplicity then , we may take @xmath295 independently of voltage . evaluating eq .
( [ e : g_beta_result ] ) for these values , we find @xmath296 . for @xmath297
we find a value of @xmath298 . | * low - loss transmission and sensitive recovery of weak radio - frequency ( rf ) and microwave signals is an ubiquitous technological challenge , crucial in fields as diverse as radio astronomy , medical imaging , navigation and communication , including those of quantum states .
efficient upconversion of rf - signals to an optical carrier would allow transmitting them via optical fibers instead of copper wires dramatically reducing losses , and give access to the mature toolbox of quantum optical techniques , routinely enabling quantum - limited signal detection .
research in the field of cavity optomechanics @xcite has shown that nanomechanical oscillators can couple very strongly to either microwave @xcite or optical fields @xcite .
an oscillator accommodating both these functionalities would bear great promise as the intermediate platform in a radio - to - optical transduction cascade . here
, we demonstrate such an opto - electro - mechanical transducer following a recent proposal @xcite utilizing a high - q nanomembrane .
a moderate voltage bias ( @xmath0 ) is sufficient to induce strong coupling @xcite between the voltage fluctuations in a radio - frequency resonance circuit and the membrane s displacement , which is _ simultaneously _ coupled to light reflected off its metallized surface .
the circuit acts as an antenna ; the voltage signals it induces are detected as an optical phase shift with quantum - limited sensitivity .
the corresponding half - wave voltage is in the microvolt range , orders of magnitude below that of standard optical modulators .
the noise added by the mechanical interface is suppressed by the electro - mechanical cooperativity @xmath1 and has a temperature of @xmath2 , where @xmath3 is the room temperature at which the entire device is operated .
this corresponds to a sensitivity limit as low as @xmath4 , or @xmath5 in a narrow frequency band around @xmath6 .
our work introduces an entirely new approach to all - optical , ultralow - noise detection of classical electronic signals , and sets the stage for coherent upconversion of low - frequency quantum signals to the optical domain @xcite . * opto- and electromechanical systems
@xcite have gained considerable attention recently for their potential as hybrid transducers between otherwise incompatible ( quantum ) systems , such as photonic , electronic , and spin degrees of freedom @xcite .
coupling of radio - frequency or microwave signals to optical fields via mechanics is particularly attractive for today s , and future quantum technologies .
photon - phonon transfer protocols viable all the way to the quantum regime have already been implemented in both radio- and optical - frequency domains separately @xcite . among the optomechanical systems that have been considered for radio - to - optical transduction @xcite
, we choose an approach @xcite based on a very high @xmath7 nanomembrane @xcite which is coupled capacitively @xcite to a radio - frequency ( rf ) resonance circuit , see figure [ fig : setup ] . together with a four - segment gold electrode , the membrane forms a capacitor , whose capacitance depends on the membrane - electrode distance @xmath8 . with a tuning capacitor @xmath9 , the total capacitance @xmath10 forms a resonance circuit with a typical quality factor @xmath11 of 130 when
an inductor wired on a low - loss ferrite rod ( @xmath12 ) is used .
the inductor serves as an antenna feeding rf signals into the series @xmath13-circuit .
the circuit s resonance frequency @xmath14 is tuned to the frequency @xmath15 of the fundamental drum mode of the membrane .
the membrane - circuit system is coupled to a propagating optical mode reflected off the membrane .
* optoelectromechanical system . *
the central part of the optoelectromechanical transducer ( * a * ) is an al - coated sin @xmath16 square membrane in vacuum ( @xmath17 mbar ) .
it forms a position - dependant capacitor @xmath18 with a planar 4-segment gold electrode in the immediate vicinity ( @xmath19 ) .
a laser beam is reflected off the membrane s al coating @xcite , converting its displacement into a phase shift of the reflected beam .
( * b * ) the membrane capacitor is part of a resonant @xmath13-circuit , tuned to the mechanical resonance frequency by means of a tuning capacitor @xmath20 ( see si for details ) .
a bias voltage @xmath21 applied to the capacitor then couples the excitations of the @xmath13-circuit to the membrane s motion .
the circuit is driven by a voltage @xmath22 in series , which can be injected through the indicated coupling port ` 2 ' or picked up by the inductor from the ambient rf radiation .
( * c * ) for a membrane - electrode distance of @xmath23 , the optically observed response of the membrane to a weak excitation of the system clearly shows a split peak ( dashed lines : fitted lorentzian resonances ) , due to hybridisation of the @xmath13-circuit mode with the mechanical resonance of the membrane . ]
the electromechanical dynamics is described most generically by the hamiltonian @xcite @xmath24 where @xmath25 and @xmath26 , the flux in the inductor and the charge on the capacitors , are conjugate variables for the @xmath13 circuit ; @xmath27 and @xmath28 denote the position and momentum of the membrane with an effective mass @xmath29 .
the last two terms represent the charging energy @xmath30 of the capacitors , which can be offset by an externally applied bias voltage @xmath21 ( fig .
[ fig : setup ] ) .
this energy corresponding to the charge @xmath31 leads to a new equilibrium position @xmath32 of the membrane .
furthermore , the position - dependent capacitive force @xmath33 causes spring softening , reducing the membrane s motional eigenfrequency by @xmath34 @xcite .
much richer dynamics than this shift may be expected from the mutually coupled system ( [ eq : ham ] ) . for small excursions
@xmath35 around the equilibrium @xmath36 , it can be described by the linearised interaction term ( ref .
@xcite and si ) @xmath37 parametrized either by the coupling parameter @xmath38 or the electromechanical coupling energy @xmath39 .
this coupling leads to an exchange of energy between the electronic and mechanical subsystems at the rate @xmath40 ; if this rate exceeds their dissipation rates @xmath41 , @xmath42 , they hybridise into a strongly coupled electromechanical system @xcite .
our system is deeply in the strong coupling regime ( @xmath43 for a distance @xmath44 and a bias voltage of @xmath45 ( fig .
[ fig : setup]c ) . here ,
for the first time , we detect the strong coupling using an independent optical probe on the mechanical system .
we have performed an experimental series , in which the bias voltage is systematically increased , with a different sample and a larger distance @xmath46 .
the system is excited inductively through port ` 2 ' ( fig .
[ fig : setup]c ) , inducing a weak radio wave signal of ( r.m.s ) amplitude @xmath47 , at a frequency @xmath48 .
the response of the coupled system can be measured both electrically as the voltage across the capacitors ( port ` 1 ' in fig .
[ fig : setup]b ) and optically by analyzing the phase shift of a light beam ( wavelength @xmath49 ) reflected off the membrane .
both signals are recorded with a lock - in amplifier , which also provides the original excitation signal . *
mechanically induced transparency .
* response of the coupled system to weak excitation ( through port ` 2 ' in fig .
[ fig : setup]b ) probed though both , ( * a * ) the voltage modulation in the @xmath13 circuit ( at port ` 1 ' ) , and ( * b * ) the optical phase shift induced by membrane displacements .
the data ( coloured dots ) measured for five different bias voltages agree excellently with model fits ( curves ) corresponding to @xmath50 .
note that each curve is offset so that its baseline corresponds to the applied bias indicated on the scale between the panels .
grey circles indicate the mechanical resonance frequency extracted for each set of data .
a shift @xmath51 can clearly be discerned ( dashed line is a fit ) .
the inset shows the effective linewidth of the mechanical resonance extracted from full model fits to the electrically ( circles ) and optically ( boxes ) measured response and simple lorentzian fits to the optical data ( diamonds ) . ]
the electrically measured response ( fig . [ fig : mit]a ) clearly shows the signature of a mechanically induced transparency @xcite indicated by the dip in the @xmath13 resonance curve .
independently , we observe the rf signal generated in the @xmath13 circuit optically via the membrane mechanical dynamics ( fig . [
fig : mit]b ) .
in particular , the electromechanical coupling leads to broadening of the mechanical resonance , an electromechanical damping effect analogous to optomechanical dynamical backaction cooling @xcite , to a new effective linewidth @xmath52 , where @xmath53 is the electromechanical cooperativity @xmath54 the width of the induced transparency dip and the mechanical linewidth grow in unison , and in agreement with our expectations as @xmath55 ( inset ) .
both these features also shift to lower frequencies as the bias voltage is increased , following the expected @xmath56 dependence @xcite .
note that in each experiment we have tuned the @xmath13 resonance frequency to @xmath57 . using the model based on the full langevin equations ( si ) , derived from the hamiltonian ( [ eq : ham ] )
, we fit the electronically and optically measured curves , and obtain fit parameters @xmath57 , @xmath58 , @xmath59 , and @xmath60 which for the two curves agree typically within @xmath61 .
together with the intrinsic damping @xmath62 determined independently from thermally driven spectra , the system s dynamics can be quantitatively predicted .
our data analysis allows us to quantify the coupling strength in three independent ways , by ( i ) analysis of the mechanical responses spectral shape , ( ii ) comparison of the voltage and displacement modulation amplitudes , and ( iii ) the frequency shift @xcite of the mechanical mode , and compare these experimental values with ( iv ) a theoretical estimate taking the geometry of the electromechanical transducer into account ( si ) . for @xmath63
we find @xmath64 following to the first method , and similar values using the three others ( cf .
si ) , corroborating our thorough understanding of the system . in another experimental run ( @xmath65 , fig .
[ fig : strong ] ) , we have characterised the strong electromechanical coupling @xcite via the normal mode splitting @xcite giving rise to an avoided crossing of the resonances of the electronic circuit and the mechanical mode , as the latter is tuned through the former using the capacitive spring effect @xcite .
in contrast to earlier observations @xcite we can simultaneously witness the strong coupling through the optical readout , in which the recorded light phase reproduces the membrane motion ( fig .
[ fig : strong]c , e ) .
again , the predictions derived from the langevin equations are in excellent agreement with our observations , yielding a cooperativity of @xmath66 for these data with @xmath67 , @xmath68 . * strong coupling regime . * ( * a * ) measured coherent coupling rate @xmath69 as a function of bias voltage ( points ) and linear fit ( line ) .
the shaded area indicates the dissipation rate @xmath70 of the lc circuit .
( * b*-*e * ) normalised response of the coupled system as measured on port ` 1 ' ( fig .
[ fig : setup]c ) ( * b , d * ) and via the optical phase shift induced by membrane displacements ( * c , e * ) .
upon tuning of the bias voltage the mechanical resonance frequency is tuned through the @xmath13 resonance , but due to the strong coupling an avoided crossing is very clearly observed .
panels ( * d , e * ) show the spectra corresponding to the orange line in ( * b , c * ) , at @xmath71 , where the electronic and mechanical resonance frequencies coincide .
points are data , orange line is the fit of the model . ]
we now turn to the performance of this interface as an rf - to - optical transducer
. a relevant figure of merit for the purpose of bringing small signals onto an optical carrier is the voltage @xmath72 required at the input of the series circuit in order to induce an optical phase shift of @xmath73 .
achieving minimal @xmath72 requires a tradeoff between strong coupling and induced mechanical damping .
for the optimal cooperativity @xmath74 we find @xmath75 at resonance @xmath76 , orders of magnitude below commercial modulators optimised for decades by the telecom industry , but also explorative microwave photonic devices @xcite based on electronic nonlinearities @xcite .
it is interesting to relate this performance to more fundamental entities , namely the electromagnetic field s quanta that constitute the signal .
indeed it is possible to show that the _ quantum _ conversion efficiency , defined here as the ratio of optical sideband photons to the rf quanta extracted from the source @xmath77 for @xmath78 , is given by ( see si ) @xmath79 this corresponds to the squared effective lamb - dicke parameter @xmath80 enhanced by the number of photons sampling the membrane during the membrane excitations lifetime .
we have tested that the membranes can support optical readout powers of more than @xmath81 without degradation of their ( intrinsic ) linewidth .
we thus project that conversion efficiencies on the order of @xmath82 are available .
so far we have measured a few percent in the laboratory , with further experiments ongoing . note
that this transducer constitutes a phase - insensitive amplifier , and can thus reach conversion efficiencies above one at the expense of added quantum noise . for the recovery of classical signals ,
the sensitivity and bandwidth of the interface is of greatest interest .
the signal at the optical output of the device is the interferometrically measured spectral density of the optical phase @xmath83 of the light reflected off the membrane , @xmath84 the voltage @xmath85 at the input of the resonance circuit ( denoted here as its spectral density @xmath86 ) is transduced to a phase shift via the circuit s susceptibility @xmath87 , the coupling @xmath60 , the effective membrane susceptibility @xmath88 and the optical wavenumber @xmath89 ( see si ) .
the sensitivity is determined by the noise added within the interface .
this includes in particular , the imprecision in the phase measurement ( @xmath90 ) , but also the random thermal motion of the membrane induced by the langevin force ( @xmath91 ) .
the former depends on the performance of the employed interferometric detector and can be quantum - limited ( @xmath92 ) .
we demonstrate this transduction scheme by measuring the ambient rf radiation background @xcite .
this radiation induces a voltage @xmath85 on the order of @xmath93 in the circuit , as we can determine through an electrical measurement on port ` 1 ' .
alternatively , we measure this signal optically , as shown in fig . [
fig : sens]a . * voltage sensitivity and noise . * ( * a * ) optical measurent of ambient rf radiation ( red ) with a signal - to - noise ratio of nearly @xmath94 . the sharp peak at @xmath95 is due to calibration with a known phase modulation .
thick lines show models for the total signal ( @xmath96 , blue ) , with contributions from the rf radiation ( @xmath97 , violet ) , the optical quantum phase noise ( @xmath98 , yellow ) , and membrane thermal noise ( @xmath99 , green ) .
( * b * ) data and models as in ( a ) , but divided by the interface s response function @xmath100 , and thus referenced to the voltage @xmath22 induced in the antenna . ] in this experiment , we use a home - built interferometer operated at @xmath101 and with a light power of @xmath102 returned from a thicker membrane ( @xmath103 , @xmath104 , @xmath105 ) . optical quantum ( shot ) noise limits the phase sensitivity to @xmath106 , corresponding to membrane displacements of @xmath107 .
this can be translated to a voltage sensitivity limit by division with the total input - output transfer function @xmath108 of the transducer .
values as low as @xmath109 are achieved within the resonant bandwidth of this proof - of - principle transducer ( yellow curve in fig .
[ fig : sens]b ) . with improved detection ( probing power @xmath110 , unity visibility ) , this number reduces to @xmath111 . evidently , yet
more sensitive optomechanical transduction could readily be achieved with known systems @xcite . the data in fig .
[ fig : sens ] corresponds to a measurement with about @xmath112 dynamic range , and we still expect a significant margin towards the onset of mechanical nonlinearities , which were observed to occur only as the displacement amplitudes are approaching nm - levels , corresponding to an expected dynamic range of @xmath113 for broadband signals . the noise added by the membrane thermal agitation is more critical in nature , as it can usually only be reduced using cryogenic cooling . remarkably , however , this noise is strongly suppressed in this setting .
if recast into an equivalent input voltage noise @xmath114 , this contribution is found as low as @xmath115 at resonance , where @xmath3 is the temperature of the _ membrane_. thus the noise added by the membrane , as seen by the optical readout , has a temperature of only @xmath116 . for the data of fig .
[ fig : sens ] , we have obtained a cooperativity of @xmath117 with a bias voltage of @xmath118 .
the corresponding noise temperature of @xmath119 should therefore allow to recover signals down to a level of @xmath4 over a bandwidth of @xmath120 ( green curve in fig .
[ fig : sens]b ) . for comparison , the ultralow - noise operational amplifier arrangement , which we use to measure on port ` 1 ' has a nominal voltage noise of @xmath121 and negligible current noise ; in practice , with a gain of @xmath122 , we achieve an input noise ( at port ` 1 ' ) of @xmath123 . referenced again to the voltage @xmath22 driving the circuit , this translates to a voltage sensitivity of @xmath124 , to be compared with the current shot - noise limited @xmath109 of our device and its @xmath4 membrane noise .
this superior performance of our transducer remains unchanged also off resonance ( @xmath125 ) as both methods benefit equally from the @xmath13-circuit s resonant enhancement of the signal voltage @xmath22 ( cf . fig .
[ fig : setup]b ) . further improved sensitivity is readily available , not least due to the still considerable margin for increasing the cooperativity : the pull - in instability of the membrane is estimated to occur only at @xmath126 , and @xmath127 in the present system ( see si ) . within the stability regime ,
stronger coupling , and correspondingly higher cooperativities , can be achieved by reducing the offset capacitance @xmath9 .
the ultimate noise floor of our opto - mechanical transducer is well below the room temperature johnson noise from the circuit s @xmath128 .
our ultralow - noise transducer / amplifier can therefore be of particular relevance in applications where this noise is suppressed .
for example , for direct electronic ( quantum ) signal transduction , the resonance circuit must be overloaded with a cold transmission line which carries the signal of interest . in radio astronomy @xcite , highly efficient antennas looking at the cold sky
have noise temperatures in the ghz range significantly below room temperature .
the usually required cryogenically cooled pre - amplifiers might be replaced by our transducer .
finally , in nuclear - magnetic resonance experiments including imaging , cooled pickup circuits can deliver a significant sensitivity improvement , yet this approach is challenging current amplifier technology @xcite . for applications with the centre radio - frequency in the ghz band strong electro - mechanical coupling to the membrane in the mhz range
can be achieved by using an oscillating coupling voltage @xcite instead of the dc voltage used in the present work .
the capacitor is fabricated by standard cleanroom microfabrication techniques .
electrodes made of gold ( thickness of 200 nm ) are deposited on a glass substrate and structured by ion - beam etching .
each segment is @xmath129 m long , with @xmath130 m gaps between the segments .
pillars of a certain height ( 600 nm , @xmath131 m ) are placed to determine the membrane - electrode distance .
the capacitor is connected in parallel to a ferrite inductor with a @xmath132 ( around 700 khz ) and @xmath133h .
the inductor is wound with litz wires to ensure high @xmath134-factor .
a variable trimming capacitor is used to tune the resonance frequency of the lc circuit .
the capacitance of the membrane - electrode system is measured to be roughly @xmath135pf .
the mechanical resonator consists of a 50 nm thick aluminum layer on top of a high - stress stoichometric sin layer with a thickness of 100 nm and 180 nm for different samples .
the al layer is deposited on top of the whole wafer after the membranes have been released .
photolithography and chemical etching are subsequently used to remove the metal from the anchoring regions and from a circle in the middle of the membrane .
the metal layer on sin causes roughly a 10 percent decrease in the eigenfrequency of the fundamental mode .
optical interferometry is carried out via a commercial doppler vibrometer ( msa-500 polytec ) and by a home made michelson interferometer ( for the data set in fig.4 ) .
the vibrometer uses optical heterodyne detection of the light returned from the membrane to recover the displacement spectrum of the membrane s surface .
the membrane - electrode distance is determined using the white light interferometry functionality of the same device .
interference fringes are collected from the partly reflective membrane layer and the electrode layer underneath , which is then used to determine their relative distance .
the home made michelson interferometer consists of two optical paths , namely the beam sent on the membrane and on a piezo - controlled mirror in the reference arm .
the beams are re - combined and the relative phase measured with a high - bandwidth ( 0 - 75mhz ) ingaas balanced - homodyne receiver .
shot noise limited measurement with an overall quantum efficiency of @xmath136 ( losses , visibility , detector efficiency ) is achieved .
the slow signals from the two dc monitoring outputs of the detector are used to generate the differential error signal , which is then fed to the piezo for locking the interferometer to the midpoint ( maximum slope ) of the interference pattern .
the rf output of the detector is high - pass filtered and fed to a spectrum analyzer in order to record the vibrations of the membrane .
absolute calibration of the mechanical amplitude is carried out via a calibration peak generated by driving the piezo with a known voltage at a frequency close to the mechanical peak .
the generated signal power is then converted to displacement by referring to the visibility equation and full - fringe voltage measurement determined by a slow piezo scan .
this work was supported by the darpa project quasar , the european union seventh framework program through siqs ( grant no .
600645 ) , the erc grants qios ( grant no .
306576 ) and interface ( grant
no .
291038 ) .
we would like to thank jrg helge mller for valuable discussions , andy barg and andreas nsby for assistance with the interferometer , and louise jrgensen for cleanroom support .
10 url # 1`#1`urlprefix[2]#2 [ 2][]#2 & .
_ _ * * , ( ) . , & .
_ _ ( ) .
_ et al . _
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
, , , & .
_ _ * * , ( ) . , , & .
_ _ * * , ( ) . , , , & .
_ _ * * , ( ) . , , & .
_ _ * * , ( ) . & .
_ _ * * , ( ) . & . _
_ * * , ( ) . & .
_ _ * * ( ) .
, , , & .
_ _ * * ( ) .
_ et al . _ . _
_ * * , ( ) .
, , , & .
_ _ * * , ( ) . , , & .
_ _ * * , ( ) . , , , & .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) . , & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . , , , & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . & . in , & ( eds .
) _ _ , vol . ,
chap . ,
( , ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
, , , & .
_ _ * * , ( ) . , , , & .
_ _ * * , ( ) .
, , & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
, , & .
_ _ * * , ( ) . & .
_ _ * * , ( ) . .
_ _ * * , ( ) . _ _ ( , ) .
_ et al . _ . _
_ * * , ( ) . , & .
_ _ * * , ( ) . |
discovery of giant magnetoresistance ( mr ) in magnetic multilayers @xcite , which consist of alternating ferromagnetic metal ( f ) and normal metal ( n ) layers , stimulated intensive studies of spin - polarized transport through layered magnetic structures @xcite . in a paper
@xcite the semiclassical theory of mr in magnetic multilayers has been developed for geometry , in which the electric current is perpendicular to the layers ( cpp geometry ) .
it has been shown that when thickness of the metals in a stack is small compared with the spin diffusion length , the magnetoresistance of multilayers can be calculated within the two - current series - resistor model @xcite . in this case
mr is expressed via the resistance of the f / n interface ( boundary resistance ) , @xmath0r_{f / n}^{\ast } $ ] , and the bulk specific resistance of the ferromagnetic layer , @xmath1\rho _ { f}^{\ast } $ ] @xcite . in these expressions @xmath2 and @xmath3 are parameters of spin asymmetry of the boundary and bulk scattering resistances .
it follows from works @xcite that their linear combination , @xmath4 , determines sign of mr in layered @xmath5\times n$ ] structures ( @xmath6 is the thickness of the ferromagnetic layers ) . in works @xcite positive ( inverse ) mr due to negative value of @xmath2 in @xmath7 multilayers has been observed for the first time .
theoretical calculations of the boundary resistance @xcite established the strong influence of the spin - dependent band structure of ferromagnetic metals on magnetoresistance . using the approach of ref .
@xcite and transmission probabilities through the f / n interface , calculated earlier from first principles @xcite , stiles and penn @xcite obtained from numerical calculations a negative value of @xmath8 for the @xmath9 interface , and a positive @xmath2 for iron - group ferromagnet - noble metal interfaces .
however , conditions on the parameters of contacting materials resulting in negative or positive values of spin - asymmetry of the boundary resistance were not discussed in the above papers . in this paper
we derive for the first time quasiclassical boundary conditions for electrochemical potentials of diffusive ferromagnetic and normal metals , which can be used for solution of a wide class of problems in spintronics .
our expression for the boundary resistance accurately accounts for the momentum conservation law as well as essential gradients of the chemical potentials .
we establish conditions on parameters of the contacting metals , at which spin asymmetry of the boundary resistance has a positive or negative sign
. dependence of the spin asymmetry and the boundary resistance on the exchange splitting of the conduction band offers one more way to estimate spin polarization of conduction band of ferromagnetic metals .
we give an example of such an estimation .
we derive boundary conditions for electrochemical potentials of diffusive metals using the quasiclassical green functions technique .
the `` diffusive ferromagnetic metal '' is a metal in which spin splitting of the conduction band is small compared with the momentum relaxation rate for conduction electrons .
let us consider that axis @xmath10 is perpendicular to the @xmath11 boundary , and neglect reversal of the electron spin upon transmission through the interface .
then , for each of the metals equations for the green functions @xmath12 read @xcite : @xmath13 @xmath14 here @xmath15 ; @xmath16 $ ] is the single - particle quasiclassical green function symmetric ( antisymmetric ) with respect to a projection of the fermi momentum @xmath17 on the axis @xmath10 ; @xmath18 is a projection of the fermi velocity on the axis @xmath10 ; @xmath19 is a spin index , and @xmath20 is a coordinate in a plane of the contact .
the bar above @xmath21 means integration over the solid angle : @xmath22 .
the boundary conditions to eqs .
( [ eq1 ] ) and ( [ eq2 ] ) are as follows @xcite : @xmath23 @xmath24 in the equations ( [ eq3 ] ) and ( [ eq4 ] ) @xmath25 and @xmath26 are the fermi momenta in ferromagnetic and normal metals , respectively ; @xmath27 is a projection of a momentum on the plane of the contact ; @xmath28 and @xmath29 are the spin - dependent , quantum - mechanical transmission and reflection coefficients .
boundary conditions ( [ eq3 ] ) , ( [ eq4 ] ) obey the specular reflection law : @xmath30 the angles @xmath31 in ( [ eq5 ] ) are measured from the axis @xmath10 , a range of variation for the biggest one is @xmath32 $ ] .
the quasiclassical equations ( [ eq1 ] ) and ( [ eq2 ] ) , and the boundary conditions ( [ eq3 ] ) and ( [ eq4 ] ) are formulated for a single electron trajectory determined by the angles @xmath33 and @xmath34 . upon solution of the system of equations ( [ eq1 ] ) and ( [ eq2 ] ) we shall consider that the ferromagnet is located to the left of the boundary @xmath35 , and the normal metal - to the right ( @xmath36 ) , and that the functions @xmath21 are homogeneous in the plane of the contact . then , the system of equations ( [ eq1 ] ) and ( [ eq2 ] ) can be solved in the form of integral equations for the functions @xmath37 and @xmath21 in the energy representation , @xmath38 : @xmath39 @xmath40 in a dirty metal the solid - angle averaged function @xmath41 obeys the diffusion equation with a decay length which is much longer than the mean free path @xmath42 .
then , we expand @xmath43 in the r.h.s . of eqs .
( [ eq6 ] ) and ( [ eq7 ] ) near point @xmath10 and take out from the integrals independent on @xmath44 terms . substituting the resulting expansions into the boundary condition ( [ eq4 ] ) we find : @xmath45 _ { x=0}. \label{eq8}\ ] ] to formulate boundary conditions for the functions @xmath46 ( which are , in fact , chemical potentials - see below ) we use a matching procedure proposed in ref .
@xcite . from equation ( [ eq1 ] ) it follows that for distances of the order of @xmath47 from the interface @xmath48 hence , @xmath49 in each of the metals .
now we calculate , for example , @xmath50 using the expression ( [ eq8 ] ) for @xmath51 .
then , we calculate @xmath50 far from the interface using an approximate expression for @xmath52 , @xmath53 which follows from equation ( [ eq2 ] ) after expansion of @xmath54 on legendre polynomials . equating values of the constant @xmath50 calculated in the two ways , and applying relationship between the averaged green function and the electrochemical potential , @xmath55 we receive the boundary condition for the electrochemical potentials at the interface @xmath35 : @xmath56 where @xmath57 xd_{\alpha } , \label{eq13 } \\ x = \cos ( \theta _ { f,\alpha } ) , \ \
\,\,d\omega _ { f,\alpha } = \sin ( \theta _ { f,\alpha } ) \,d\theta _ { f,\alpha } \,d\varphi .
\nonumber \end{aligned}\ ] ] the limits of angular integration must satisfy specular reflection conditions at the interface , eq .
( [ eq5 ] ) . when deriving eq .
( [ eq12 ] ) we have used conservation of the current density at the interface , which follows from eq .
( [ eq3 ] ) , @xmath58 where @xmath59 and @xmath60 are the bulk , spin - channel conductivities of the metals : @xmath61 equation ( [ eq14 ] ) is actually the second , complementary to eq .
( [ eq12 ] ) , boundary condition for the semiclassical description of the spin - polarized transport in magnetic multilayers in terms of electrochemical potential .
the derivative from the electrochemical potential in eq .
( [ eq12 ] ) can be expressed in terms of the density of current ( [ eq14 ] ) , and we find the spin - dependent resistance of the interface @xmath62 : @xmath63 @xmath64 where @xmath65 is the area of the contact .
it follows from equation ( [ eq20 ] ) that , in the quasiclassical approach at specular reflection from the interface , the boundary resistance between ferromagnetic and normal metals is determined only by the fermi momenta of the contacting metals and coefficient of transmission through the interface .
experimental data are given for the spin asymmetry of boundary resistance , @xmath2 , and for the renormalized resistance of the interface , @xmath66 , determined as follows : @xmath67 to calculate dependence of @xmath2 and @xmath66 on the fermi momentum of the non - magnetic metal for various values of the ferromagnet conduction band polarization we have used the fermi - momentum - mismatch model for the transmission coefficient : @xmath68 $ ] .
results are presented in figures 1 ( @xmath69 ) , 2 ( @xmath70 ) and 3 ( @xmath71 ) . from our calculations
it follows that for a non - magnetic metal with a low density of conduction electrons ( small value of the fermi momentum @xmath26 ) the spin asymmetry of the boundary resistance @xmath2 is always negative ( fig .
1 ) . on the contrary , for a non - magnetic metal with a high density of conduction electrons the spin asymmetry of the boundary resistance is always positive ( fig .
3 ) . in an intermediate situation
@xmath8 can be negative as well as positive ( see fig .
2 ) . to attain the maximum amplitude of negative magnetoresistance in @xmath11
multilayers the spin asymmetry of the boundary resistance @xmath2 and the asymmetry of the bulk resistance @xmath3 should be both positive and close to unity . according to our calculations , the fermi momentum of the non - magnetic metal should be as far as possible close to the fermi momentum of the majority subband of ferromagnetic metal ( fig .
2 , @xmath72 and fig .
3 , @xmath73 ) .
the spin asymmetry of the bulk resistance can be adjusted by the type and concentration of impurities in the ferromagnetic metal @xcite .
clearly , similar arguments can be applied to the opposite case of negative values of @xmath2 and @xmath3 , which will result in positive magnetoresistance in multilayers of alternating ferromagnetic and non - magnetic metals . however , a negative asymmetry of the bulk resistance is met less often @xcite .
competition of opposite in a sign asymmetries of boundary and bulk resistances can result in negative or positive magnetoresistance depending on the choice of materials and thickness of the ferromagnetic layers @xcite .
experiments on cpp transport in multilayers are very complicated because the resistance of a stack of layers of nanometer thickness is very small ( order of @xmath74 m@xmath75 ) .
nevertheless , available experimental data ( see reviews @xcite and references in them , and also works @xcite ) allow to test the internal consistency of the theory .
multilayers of the iron - group ferromagnetic metals with noble metals most likely belong to case 3 ( @xmath76 , and parameter of the spin asymmetry @xmath8 is positive @xcite .
for example , @xmath77 @xcite . then , intersection of the horizontal dash line @xmath78 in the top field of fig .
3 with the curve @xmath79 , corresponding to @xmath80 , gives @xmath81 . accepting @xmath82 as a trial value for the fermi momentum of the majority subband of cobalt we receive @xmath83 , which fits fairly good the free - electron - model value for copper , @xmath84 @xcite .
there are data for silver as the non - magnetic spacer : @xmath85 @xcite . in a similar way
, we obtain from fig .
3 , @xmath86 , which fits well the free - electron - model value @xmath87 @xcite . let us look now at consistency of the theory with boundary resistance data .
for the combination @xmath88 , @xmath89 @xcite . continuing the vertical dash line for copper in fig . 3 into the bottom field till intersection with the curve @xmath90 ,
corresponding to @xmath80 , we obtain @xmath91 .
a similar procedure gives for silver @xmath92 ( compare with @xmath93 @xcite ) .
it is worthy to notice that the theory reproduces fairly good closeness of the boundary resistances of the co / cu and co / ag interfaces . for @xmath94 multilayers
the asymmetry of the boundary resistance is negative , @xmath95 @xcite .
this value is admissible for cases 1 ( fig.1 , top field ) and 2 ( fig .
2 , top field ) . without details , we conclude that the first case does not match the expected value of the conduction band polarization parameter for co @xmath96 as well as results in about two orders in magnitude higher boundary resistance . the second case ( see dash lines in fig .
2 ) results in a fermi momentum @xmath97 , and in a boundary resistance @xmath98 .
the fermi momentum is satisfactory in the frame of the free electron model @xcite .
the boundary resistance is close enough to the experimental value @xmath99 @xcite .
we expect that better matching of the band structures of cobalt and chromium , both belonging to the iron - group metals , would results in a weaker influence of the real band structure on the boundary resistance .
a discrepancy with experiment in an absolute value of boundary resistance of about 20 - 45% seems not catastrophic because of the following reasons .
first , the trial choice of @xmath82 was not optimized .
second , we used the free _
s_-electron model as a background for the theory .
one may expect that for an interface between a metal with predominantly _
d_-electron conduction band ( iron group ) and an _ s_-electron metal ( cu , ag ) , reduced overlapping and symmetry mismatch may noticeably increase the boundary resistance .
third , garca and stoll have shown ( fig . 2 of ref .
@xcite ) that the interface roughness also increases boundary resistance .
this increase is estimated below 20 - 60% for different models and magnitudes of the interface roughness , and reasonable differences in the fermi momenta of contacting metals .
the interface roughness is much less important for the case of cpp transport , considered in this paper , than for current - in - plane ( cip ) transport @xcite .
finally , spin - reversal at the interface opposes the both previously considered processes , decreasing the boundary resistance .
the reversal of the electron spin by the spin - orbit interaction is always expected upon refraction of the electron wave or scattering on roughness at the interface between two metals .
quantitative analysis of the competition between the above minor mechanisms of boundary resistance is beyond the scope of the paper .
however , it is our expectation that the key quantity , spin asymmetry of the boundary resistance @xmath8 , is only weakly dependent on band structures matching , interface roughness , spin reversal _
etc_. because of considerable cancellations in the dimensionless ratio , eq .
( [ eq21 ] ) .
our trial evaluations show that the experimental data for the spin asymmetry of boundary resistance and the absolute value of boundary resistance in @xmath88 , @xmath100 and @xmath94 multilayers can be consistently described with the use of the spin polarization parameter for the conduction band of cobalt @xmath101 . at the level of the experimental accuracy and completeness of the theory the estimated value of @xmath102 is identical to @xmath103 , which we have estimated @xcite from experiments of garca _ et al . _
@xcite on magnetoresistance of cobalt nanocontacts .
experiments on andreev spectroscopy cite similar values of @xmath102 for cobalt @xcite .
thus , the spin asymmetry of the boundary resistance in combination with the absolute value of the boundary resistance can be used for estimations of the spin - polarization parameter @xmath104 of the conduction band of ferromagnetic metals .
the work was supported by ec via the contract no nmp4-ct-2003 - 505282 , and by rfbr via projects no 03 - 02 - 17432 and no 03 - 02 - 17656 .
l r t acknowledges numerous illuminating discussions with professor nicolas garca .
99 baibich m n , broto j m , fert a , nguyen van dau f , petroff f , etienne p , creuzet g , friederich a and chazelas j 1988 _ phys . rev . lett . _ * 61 * 2472 . fig .
dependence of the spin - asymmetry of boundary resistance @xmath105 ( top field ) and the renormalized resistance of the interface @xmath66 ( bottom field ) on the fermi momentum of the non - magnetic metal for the case ( @xmath106 ) . | quasiclassical boundary conditions for electrochemical potentials at the interface between diffusive ferromagnetic and non - magnetic metals are derived for the first time .
an expression for the boundary resistance accurately accounts for the momentum conservation law as well as essential gradients of the chemical potentials .
conditions are established at which spin - asymmetry of the boundary resistance has positive or negative sign .
dependence of the spin asymmetry and the absolute value of the boundary resistance on the exchange splitting of the conduction band opens up new possibility to estimate spin polarization of the conduction band of ferromagnetic metals .
consistency of the theory is checked on existing experimental data . |
Subsets and Splits