article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
the @xmath0 spin dependent total cross sections /(column 7 in table 1 ) and / ( column 8) as a function of @xmath40 , the maximum pion center - of - mass momentum in units of the pion mass .
the curves represent the predictions by the jlich meson exchange model [ 6 ] .
measured , relative ps strength , @xmath42/ , of the reaction @xmath0 as a function of @xmath40 .
the total cross section @xmath42 for @xmath51 as a function of @xmath40 .
the dashed line represents the best fit with @xmath42 proportional to @xmath43 .
the solid line is the ps total cross section calculated from the jlich meson - exchange model [ 6 ] .
the total cross section @xmath52 for @xmath0 as a function of @xmath40 .
the dashed line represents the best fit with @xmath52 proportional to @xmath53 .
the solid line is the prediction of the jlich meson exchange model [ 6 ] . | we report a measurement of the spin - dependent total cross section ratios /and /of
the @xmath0 reaction between 325and 400 .
the experiment was carried out with a polarized internal target in a storage ring .
non - vertical beam polarization was obtained by the use of solenoidal spin rotators . near threshold ,
the knowledge of both spin dependent total cross sections is sufficient to deduce the strength of certain participating partial waves , free of any model .
-5 mm -5 mm -18 mm pacs numbers : 24.70.+s , 25.10.+s , 29.20.dh , 20.25.pj , 29.27.hj we present a measurement of /and /for the @xmath0 reaction .
the quantity ( or ) equals the difference between the total cross sections measured with opposite and parallel , transverse ( or longitudinal ) beam and target polarizations , while is the ( unpolarized ) total cross section . at threshold
a single partial wave ( @xmath1 ) dominates .
as the bombarding energy is increased other partial waves become significant . in the following
, we will demonstrate how the results of the present experiment can be used to gain information on these higher partial waves , making it possible to test theoretical models selectively .
the reaction amplitude for pion production in the nucleon - nucleon ( nn ) system may be separated into contributions with definite angular momentum .
these partial - wave amplitudes are labelled with the quantum numbers of the final state @xmath2l@xmath3 , where s , l and j are the spin , angular momentum and total angular momentum of the nn pair , @xmath4 is the angular momentum of the pion and j is the total angular momentum of the final state .
the initial - state quantum numbers @xmath5 , @xmath6 , @xmath7 are related to those of the final state by angular momentum , parity and isospin conservation .
it is useful to define cross sections @xmath8 which partition the total cross section according to the initial nn spin , @xmath5 , and its projection , @xmath9 , onto the direction of the incident momentum .
the three possible contributions @xmath10 , @xmath11 , and @xmath12 are related to the observable quantities by @xmath13 ^3\sigma_0\ = \
{ \mbox{\,$\sigma_{\rm tot}$\ } } \ -\ { \mbox{\,$\delta\sigma_{\rm \small t}$\ } } \!\ ! & + \
\frac{1}{2}{\mbox{\,$\delta\sigma_{\rm \small l}$\ } } \\[1.5ex ] ^3\sigma_1\ = \ { \mbox{\,$\sigma_{\rm tot}$\ } } \ & -\
\frac{1}{2}{\mbox{\,$\delta\sigma_{\rm \small l}$\ } } \ . \end{array}\ ] ] because , near threshold , the relative kinetic energies between the particles in the final state are small , the angular momenta l and @xmath4 are either 0 or 1 , giving rise to the possible final states l@xmath4 = ss , sp , ps , or pp . a list of possible @xmath0 amplitudes can easily be constructed , taking into account the negative parity of the pion , as well as conservation of angular momentum and isospin . from this list , one sees that the sp final state is not allowed , that only initial
state singlets ( @xmath10 ) contribute to ps final states , that ( @xmath12 ) contributes only to pp , and that ( @xmath11 ) may contribute to both , ss and pp . assuming that there are no final states other than ss , ps , and pp
, we require that = @xmath14 where @xmath15 is the total cross section to a certain final state , l@xmath4 . we can then summarize the above in the following way : @xmath16 ^3\sigma_0\ = \ 4 ( \sigma_{ss } + \hat{\sigma}_{pp } ) \\[1.2ex ] ^3\sigma_1\ = \ 2 ( \sigma_{pp } - \hat{\sigma}_{pp})\ .
\end{array}\ ] ] here , @xmath17 and @xmath18 represent two different combinations of pp amplitudes .
the term @xmath18 is needed in @xmath11 because @xmath19 initial states can lead to ss _ and _ pp final states .
since the term @xmath18 does not contribute to the total cross section , it must also appear in @xmath12 .
combining eqs .
1 and 2 , we arrive at the useful relation @xmath20 which states that the probability to form a ps final state can be deduced , free of any model , from the observables of this experiment .
a similar equation for @xmath21 follows easily from eqs .
1,2 .
the experiment discussed here was carried out with the indiana cooler .
protons from the cyclotron were stack - injected into the ring at 197 , reaching an orbiting current of several 100within a few minutes .
the beam was then accelerated to the energies listed in table 1 . after typically 10 minutes of data taking
, the remaining beam was discarded , and the cycle was repeated .
the target and detector used for this experiment are the same as described in ref .
[ 1 ] , and a detailed account of the apparatus can be found in ref . [ 2 ] . the internal polarized target consisted of an open - ended 25long storage cell of 12diameter and 25wall thickness .
the cell is coated with teflon to avoid depolarization of atoms colliding with the wall . during data taking ,
the target polarization @xmath22 is changed every 2pointing in sequence , up , down ( @xmath23 y ) , left , right ( @xmath23 x ) , and along , opposite to the beam direction ( @xmath23 z ) .
the magnitude of the polarization is the same within @xmath23 0.005 for all orientations [ 3,4 ] .
the detector arrangement consists of a stack of scintillators and wire chambers , covering a forward cone with a 30 opening angle .
the scintillators are capable of stopping protons from @xmath0 , thus measuring their energy . from the time of flight and the relative energy deposited in the layers of the detector ,
the outgoing charged particles are identified as protons . from the direction and energy of the two protons in the final state , the mass @xmath24 of the undetected particle is calculated . for an event of interest
, this mass has to equal the @xmath25 mass within the mass resolution of the experiment [ 1 ] .
background arises from reactions in the walls of the target cell [ 2 ] and , to a lesser degree , from a @xmath26 impurity in the target gas .
background is rejected by a condition on the relative angles of the detected protons and by a cut on the transverse distribution of the reaction vertices .
a measurement with a nitrogen target matches the shape of the @xmath24 distribution seen with the h target , except for the @xmath25 peak , and is used to subtract background which remains under the @xmath25 mass peak ( between 5 and 10 ) . within statistics ,
the background shows no spin dependence .
the validity of the background subtraction was tested by varying the range of accepted masses @xmath24 .
non - vertical beam polarization is achieved by two spin - rotating solenoids located in non - adjacent sections of the six - sided cooler .
the vertical and longitudinal components of the beam polarization @xmath27 at the target are about equal , with a small sideways component .
since the solenoid fields are fixed in strength , the exact polarization direction depends on beam energy after acceleration . only either the vertical or the longitudinal component contributes to the total cross section , depending on the orientation of the target polarization . in alternating measurement cycles , the sign of the beam polarization is reversed .
more details on the preparation of non - vertical beam polarization in a storage ring can be found in refs . [ 2,4 ] .
data are acquired for all 12 possible polarization combinations of beam ( @xmath28 ) and target ( @xmath23 x , @xmath23 y , @xmath23 z ) .
the known spin correlation coefficients of proton - proton elastic scattering [ 5 ] are used to monitor beam and target polarization , concurrently with the acquisition of @xmath0 events . to this end
, coincidences between two protons exiting near @xmath29 are detected by two pairs of scintillators placed , behind the first wire chamber at azimuthal angles @xmath30 and @xmath31 . from this measurement , the products @xmath32 and @xmath33 of beam and target polarization
are deduced .
evaluating the cross ratio of the four @xmath0 yields with @xmath34 and @xmath35 , dividing by the measured @xmath32 , results in /.
analogously , /follows from the yields with @xmath34 and @xmath36 , using the measured @xmath33 . at 375 and 400
, two data runs , separated by five months , were combined .
the detector acceptance is less than 100because of a hole in the center of the detector stack which accommodates the circulating beam , and excludes particles with @xmath37 .
the resulting loss of events ( about 30 at the lowest energy ) is taken into account by applying a correction which arises from a difference in detector acceptance for ss , ps and pp final states [ 1 ] .
this correction turns out to be small ( between 0.009 and 0.058 for / , and between 0.007 and 0.030 for / ) .
the final results are listed in table 1 .
the errors include counting statistics , and an uncertainty for background subtraction and hole correction .
the latter two were estimated by a reasonable variation ( @xmath38 ) in the amount of background and in the size of the hole correction .
the /results from this experiment were consistent with those obtained in an earlier experiment [ 1 ] .
the weighted averages of the two data sets are shown in column 7 of table 1 .
the data in the last two columns of table 1 represent all experimental information on @xmath0 spin dependent total cross sections and are shown in fig . 1 .
most of the theoretical work on @xmath0 so far deals with the lowest partial wave ( ss ) .
an exception is the meson - exchange model of the jlich group [ 6 ] which contains the higher partial waves needed to address polarization observables .
this calculation includes off - shell pion rescattering and the exchange of heavy mesons , and provides a good fit to the ss part of the total cross section close to threshold , or for @xmath39 , where @xmath40 is the maximum center - of - mass pion momentum divided by the @xmath25 mass .
the agreement is not as good for the polarization observables measured here ( fig .
1 ) which are sensitive to contributions from ps and pp partial waves . as pointed out in the first part of this paper , the present measurement allows a model - free statement about the ps and pp contributions to @xmath0 .
eq .
3 directly yields the relative ps strength , @xmath41 the result is shown in fig . 2 .
the ps strength is larger than previously thought [ 8 ] ( see below ) .
phase space arguments together with the properties of spherical bessel functions for small arguments lead to the expectation that @xmath42 is proportional to @xmath43 . in order to test this prediction
, we have to multiply our experimental @xmath44by the total cross section . at t = 325an accurate value for the total cross section exists (= 7.70 @xmath23 0.26 , [ 7 ] ) .
however ,
at higher energies data are few and of poor quality , and a good measurement of between 0.3 and 1would certainly be a much needed addition to the @xmath0 data base . for the present purpose we use a smooth approximation to the world s data , obtaining = 17 , 40 , and 86 , for t = 350 , 375 , and 400 , respectively .
the resulting values for @xmath42 are shown in fig .
3 .
the dashed line in fig .
3 represents the fit with @xmath45 .
the error of the scaling factor takes into account the errors of the data , and the uncertainty of at 325 .
one can also combine the expressions for @xmath12 in eqs .
1,2 to obtain an equation for ( @xmath46)/in terms of /.
multiplying the result by , yields the pp partial cross section , @xmath46 , shown in fig .
4 .
the dashed line in fig .
4 represents a fit with the expected energy dependence for pp waves , @xmath47 . the solid lines in figs .
3,4 are obtained with the jlich meson exchange model [ 6 ] .
one sees that the ps strength which , in this model , is dominated by the role of the @xmath48 ( a nucleon excited state ) is underestimated by about a factor of 3 ( fig .
3 ) , while for ( @xmath46 ) the discrepancy is less but the energy dependence is not correctly reproduced by the model ( fig . 4 ) . in the absence of polarization observables ,
some information on individual partial wave contributions can be extracted from a measurement of the ( unpolarized ) cross section as a function of the relative energy of the two nucleons in the final state , assuming that the corresponding energy distributions of single partial waves are known .
as our data indicate , power laws in @xmath40 seem adequate for the ps and pp partial waves , but the energy dependence of the ss partial wave is dominated by the final state interaction , and thus requires model dependent input .
such an analysis has been carried out for @xmath0 at 310 ( @xmath49 ) [ 8 ] . in that work
a significant contribution was found from an sd - ss interference term .
this conclusion was reached , however , setting @xmath50 , while from the present data , using an @xmath43 energy dependence , we estimate that , at 310 , @xmath42 contributes about 6to . in summary , we have measured both spin dependent total cross sections in @xmath0 .
these polarization observables allow a model - free determination of individual partial wave contributions .
we show in an example that this can provide a test of individual components of a model ( figs .
3,4 ) .
the availability of polarization observables is crucial when carrying out a partial wave study , since an analysis based only on the unpolarized differential cross section is subject to model dependence .
we thank dr .
ch .
hanhart for supplying us with the numerical values for the calculation shown in figs .
1,3 and 4 .
this work has been supported by the us national science foundation under grants phy96 - 02872 , phy95 - 14566 , phy97 - 22556 , and by the us department of energy under grant doe - fg02 - 88er40438 . *
table 1 * + bombarding energy t , maximum pion center - of - mass momentum @xmath40 in units of the pion mass , integrated luminosity , the products @xmath32 and @xmath33 of beam and target polarization , and the measured /and /are listed .
column 7 lists the values for /from this experiment ( column 6 ) combined with the results from an earlier measurement with purely vertical beam polarization [ 1 ] . [
cols="^,^,^,^,^,^,^,^ " , ] 1 h.o .
meyer et al . phys .
rev .
lett . * 81*:3096 , 1998 .
t. rinckel et al .
submitted to nul .
instr .
meth .
f. rathmann et al .
phys .
rev . * c58*:658 , 1998 .
b. lorentz et al .
submitted to phys .
rev .
b. v. przewoski et al .
phys .
rev .
* c58*:1897 , 1998 . c. hanhart , j. haidenbauer , o. krehl and j. speth .
phys .
lett .
* b444*:25 , 1998 .
h.o .
meyer et al . nucl .
phys .
* a539*:633 , 1992 .
j. zomaczuk et al .
phys .
lett . * b436*:251 , 1998 . |
in this contribution i discuss the possibility of production of high energy cosmic tau neutrinos ( @xmath0 gev ) in cores of active galactic nuclei ( agn ) originating from proton acceleration .
the effects of three flavour neutrino mixing on this high energy cosmic tau neutrino flux and the prospects for its observations through double shower events @xcite .
in addition to agns , high energy cosmic tau neutrinos may also be produced in several currently envisaged other cosmologically distant astrophysical sources @xcite .
for some possible effects of neutrino mixing other than the flavour one on high energy cosmic tau neutrino flux , see @xcite . for prospects of observations of high energy cosmic tau neutrinos other than the double shower technique , see @xcite .
the present study is particularly useful as several under water / ice high energy neutrino telescopes are now at their rather advanced stages of development and deployment @xcite .
i start in section 2 with a brief description of possibility of production of high energy cosmic tau neutrinos relative to electron and muon neutrinos and discuss in some detail the effect of three flavour neutrino mixing on their subsequent propagation and further discuss the prospects for their detection through double shower events . in section 3
, i summarize the results .
high energy cosmic tau neutrinos may either mainly be produced in @xmath2 or in @xmath3 collisions in a cosmologically distant environment . in @xmath2 collisions ,
high energy @xmath4 and @xmath5 are mainly produced through the resonant reaction @xmath6 .
the same collisions will give rise to a greatly suppressed high energy @xmath7 ( and @xmath8 ) flux ( @xmath9 ) mainly through the reaction @xmath10 .
in @xmath11 collisions , the @xmath7 flux may be obtained through @xmath12 . the relatively small cross - section for @xmath13 production together with the low branching ratio into @xmath7 implies that the @xmath7 flux in @xmath11 collisions is also suppressed up to 5 orders of magnitude relative to @xmath4 and/or @xmath5 fluxes .
thus , both in @xmath2 and in @xmath11 collisions , the intrinsically produced tau neutrino flux is expected to be rather quite small , typically a factor less than @xmath14 relative to electron and muon neutrino fluxes @xcite .
matter effects on vacuum neutrino flavour oscillations are relevant if @xmath15 . using @xmath16 given in ref .
@xcite , it follows that matter effects are absent for @xmath17 ev@xmath1 .
matter effects are not expected to be important in the neutrino production regions around agn and will not be further discussed here . in the framework of three flavour analysis ,
the flavour precession probability from @xmath18 to @xmath19 neutrino flavour is @xcite @xmath20 where @xmath21 or @xmath22 .
@xmath23 is the 3@xmath243 mns mixing matrix and will be used in usual notation with the standard parameterization for vanishing @xmath25 and cp violating phase @xcite . in eq .
( 1 ) , @xmath26 with @xmath27 and @xmath28 is the distance between the source and the detector .
currently , the atmospheric muon and solar electron neutrino deficits can be explained with oscillations among three active neutrinos @xcite . for this , typically , @xmath29 ev@xmath1 and @xmath30 for the explanation of atmospheric muon neutrino deficit , whereas for the explanation of solar electron neutrino deficit , we may have @xmath31 ev@xmath1 and @xmath30 [ just so solution ] or @xmath32 ev@xmath1 and @xmath33 [ sma ( msw ) ] or @xmath32 ev@xmath1 and @xmath30 [ lma ( msw ) ] . in the above explanations ,
the total range of @xmath34 is @xmath35 ev@xmath36 irrespective of neutrino flavour .
the typical energy span relevant for possible flavour identification for high energy cosmic neutrinos is @xmath37gev@xmath38 in which currently the neutrino flux from cores of agns dominate .
taking a typical distance between the agn and our galaxy as @xmath39 mpc ( where 1 pc @xmath40 m ) , the @xmath41 term in eq .
( 1 ) vanishes and so eq . ( 1 ) reduces to @xmath42 it is assumed here that no relatively dense objects exist between the agn and the earth so as to effect significantly this oscillations pattern .
the @xmath43 in eq .
( 2 ) is independent of not only @xmath34 but also @xmath44 .
let me denote by @xmath45 , the intrinsic high energy cosmic neutirno fluxes .
for simplicity , i take their ratios as @xmath46 . in order to estimate the final flux ratios on earth ,
let me introduce a 3@xmath243 matrix of vacuum flavour precession probabilities such that @xmath47 the explicit form for the matrix @xmath48 in case of just so flavour oscillations as solution to solar neutrino problem along with the solution to atmospheric neutrino deficit in terms of @xmath5 to @xmath7 oscillations with maximal depth is @xmath49 using eq . ( 4 ) and eq .
( 3 ) , we note that @xmath50 at the level of @xmath51 .
also , the unitarity conditions for @xmath52 are satisfied .
the same flux ratio is obtained in remaining two cases .
thus , essentially independent of the oscillation solution for solar neutrino problem , i have @xmath53 .
this provides some prospects for possible detection of high energy cosmic tau neutrinos .
the down ward going high energy cosmic tau neutrinos reaching close to the surface of the detector may undergo a charged current deep inelastic scattering with nuclei inside / near the detector and produce a tau lepton in addition to a hadronic shower .
this tau lepton traverses a distance , on average proportional to its energy , before it decays back into a tau neutrino and a second shower most often induced by decay hadrons .
the second shower is expected to carry about twice as much energy as the first and such double shower signals are commonly referred to as a double bangs . as tau leptons are not expected to have further relevant interactions ( with high energy loss ) in their decay time scale , the two showers should be separated by a clean @xmath54-like track @xcite .
= 2.5 in = 2.5 in the calculation of down ward going contained but separable double shower event rate for a typical km@xmath1 surface area under water / ice neutrino telescope can be carried out by replacing the muon range expression with the tauon one and then subtracting it from the linear size of a typical high energy neutrino telescope in the event rate formula while using the expected @xmath7 flux spectrum given by eq .
( 3 ) . here
, i confine myself by mentioning that the expected number of contained but separable double showers induced by down ward going high energy cosmic tau neutrinos for @xmath55 gev may be @xmath56/yr@xmath57sr , essentially irrespective of solution of solar neutrino problem , if one uses the @xmath51 from ref . @xcite as an example . at this energy ,
the two showers initiated by the down ward going high energy cosmic tau neutrinos are well separated ( see fig .
1 ) such that the amplitude of the second shower is essentially 2 times the first shower .
intrinsically , the flux of high energy cosmic tau neutrinos is quite small , typically being @xmath58 from , for instance , cosmologically distant cores of active galactic nuclei . \2 .
because of vacuum neutrino flavour oscillations , this ratio can be greatly enhanced . in the context of three flavour neutrino mixing scheme which can accomodate the oscillation solutions to solar and atmospheric neutrino deficits , the final flux ratio of high energy cosmic neutrinos on earth is @xmath59 .
this enhancement in high energy cosmic tau neutrino flux may lead to the possibility of its detection in km@xmath1 surface area under water / ice high energy neutrino telescopes . for @xmath60/gev @xmath38 , the down ward going high energy cosmic tau neutrinos may produce a double shower signature because of charged current deep inelastic scattering followed by a subsequent hadronic decay of associated tau lepton .
i thank japan society for the promotion of science ( jsps ) for financial support .
9 j. g. learned and s. pakvasa , astropart .
phys . 3 ( 1995 ) 267 .
see , e. waxman ; v. berezinsky ; p. l. biermann and g. sigl in these proceedings .
athar husain , nucl .
b ( proc . suppl . )
76 ( 1999 ) 419 and references cited therein .
for a recent discussion , see , s. iyer , m. h. reno and i. sarcevic , hep - ph/9909393 and references cited therein .
l. moscoso , these proceedings .
a some what detailed numerical study supports this order of magnitude estimate ; h. athar , r. a. vzquez and e. zas ( in preparation ) .
a. p. szabo and r. j. protheroe , astropart .
phys . 2 ( 1994 ) 375 .
ta - pei cheng and ling - fong li ( eds . ) , gauge theory of elementary particle physics , claredon press , oxford , 1984 , p. 411 . c. caso et al .
, the euro . phys .
j. c3 ( 1998 ) 103 .
m. nakahata , these proceedings .
athar , g. parente and e. zas ( to be submitted ) .
see also scanned transperacies by athar husain at url http://taup99.in2p3.fr / taup99/. | i discuss the possibility of production of high energy cosmic tau neutrinos ( @xmath0 gev ) in an astrophysical site and study some of the effects of neutrino mixing on their subsequent propagation .
i also discuss the prospects for observations of these high energy cosmic tau neutrinos through double shower events in new km@xmath1 surface area under water / ice neutrino telescopes . |
in recent work @xcite , we considered neutrino production for two leading scenarios for the sources that power long - duration grbs , namely the collapsar @xcite and supranova ( sa ) @xcite models . in the collapsar model ,
the core of a massive star collapses directly to a black hole , and the most important radiation field for photomeson neutrino production is the internal synchrotron radiation field @xcite . in the supranova ( sa ) model ,
a pulsar nebula synchrotron radiation within an expanding supernova remnant ( snr ) shell @xcite provides additional external photon target for photomeson interactions ( see also refs .
@xcite ) .
we found that the presence of the external field in the sa model can increase the number of detectable neutrinos by an order of magnitude or more over the collapsar model when the doppler factor @xmath4 .
when @xmath5 , the internal synchrotron field can become effective for photomeson interactions . in our calculations , we assumed that the energy injected in protons is equal to the energy of electrons producing the photon fluence measured at x - ray and @xmath0-ray energies . in both models ,
the likelihood of detecting a neutrino from a grb with a km - scale detector such as icecube is small except for rare grbs with fluence @xmath6 .
if grb blast waves are strongly baryon - loaded , however , as required in a grb model for high - energy cosmic rays @xcite , then we predict that 100 tev
100 pev neutrinos will be detected several times per year with icecube when @xmath5 .
here we summarize our calculations of photomeson neutrino production for the collapsar and sa models , and present new calculations for the hadronic cascade radiation from a grb .
the grb model is adapted from our photo - hadronic model for blazar jets @xcite , and takes into account the injection of nonthermal protons which lose energy through photomeson interactions .
protons are injected with a number spectrum @xmath7 at comoving proton energies @xmath8 gev up to a maximum proton energy determined by the condition that the particle larmor radius is smaller than both the size scale of the emitting region and the photomeson energy - loss length . here
@xmath9 is the lorentz factor of the blast wave . the observed synchrotron spectral flux in the prompt phase of the burst is parameterized by the expression @xmath10 , where @xmath11 kev , @xmath12 above @xmath13 and an exponential cutoff at 10 mev , and @xmath14 when @xmath15 .
at lower energies , @xmath16 . the observed total hard x - ray / soft @xmath0-ray photon fluence @xmath17 , where @xmath18 is the characteristic duration of the grb .
we consider a source at redshift @xmath19 and take the hard x - ray / soft @xmath0-ray fluence @xmath20 .
two or three grbs should occur each month above this fluence level .
a total amount of energy @xmath21 is injected in the form of accelerated proton energy into the comoving frame of the grb blast wave . here
@xmath22 is the redshift and @xmath23 is the luminosity distance .
the energy deposited into each of @xmath24 light - curve pulses ( or spikes ) is therefore @xmath25 ergs .
we assume that all the energy @xmath26 is injected in the first half of the time interval of the pulse ( to ensure variability in the grb light curve ) , which effectively corresponds to a characteristic variability time scale @xmath27 .
the proper width of the radiating region forming the pulse is @xmath28 , from which the energy density of the synchrotron radiation can be determined @xcite .
we set the grb prompt duration @xmath29s , and let @xmath30 , corresponding to @xmath31 .
the magnetic field is determined by assuming equipartition between the energy densities of the magnetic field and the electron energy . for the sa model
, we assume the existence of an external radiation field given by the expression @xmath32 , with @xmath33 kev @xcite .
the intensity of this field is determined by the assumption that the integral power @xmath34 is equal to the power of the pulsar wind @xmath35 , assuming that a total of @xmath36 of pulsar rotation energy is radiated during the time @xmath37 ( which is here set equal to 0.1 yr ) from the rotating supramassive neutron star before it collapses to a black hole .
the energy @xmath38 is the characteristic energy of synchrotron radiation emitted by electrons ( of the pulsar wind ) with lorentz factors @xmath39 in a randomly ordered magnetic field of strength @xmath40 g. the radius @xmath41 is determined by assuming that @xmath42 is the mean speed of the snr shell , and that the external photon energy density @xmath43 .
1 shows the total @xmath2 fluences expected from a model grb with @xmath44 pulses .
the thin curves show collapsar model results at @xmath45 200 , and 300 , with @xmath46 .
the expected numbers of @xmath2 that a km - scale detector such as icecube would detect are @xmath47 , @xmath48 , and @xmath49 , respectively .
there is no prospect to detect @xmath2 from grbs at these levels .
the heavy solid and dashed curves in fig . 1 give the sa model predictions of @xmath50 for both @xmath51 and @xmath52 .
the equipartition magnetic fields are 1.9 kg and 0.25 kg , respectively .
the external radiation field in the sa model makes the neutrino detection rate insensitive to the value of @xmath53 ( as well as to @xmath54 , as verified by calculations ) , but there is still little hope that a km@xmath55 detector could detect such grbs .
neutrino production efficiency would improve by at most a factor of 3 in the collapsar model if @xmath56ms ( and @xmath57 to provide the same total fluence ) .
such narrow spikes are , however , then nearly opaque to @xmath0 rays with energies @xmath58 mev @xcite . even in this case , there is little hope to detect neutrinos except from rare grbs with fluence @xmath59 . for a model grb with hard x - ray fluence @xmath60 for different doppler factors @xmath53 .
the thin curves show collapsar model results where only the internal synchrotron radiation field provides a source of target photons . ]
2 shows new calculations of photomeson neutrino production for a grb with @xmath61 and @xmath51 , as well as the accompanying electromagnetic radiation induced by pair - photon cascades from the secondary electrons and @xmath0 rays from the same photomeson interactions .
the total number of @xmath2 expected with icecube is @xmath62 .
the total fluence of cascade photons shown here is contributed by lepton synchrotron ( dot - dashed ) and compton ( dashed ) emissions . for comparison ,
the dotted curve shows the primary lepton synchrotron radiation spectrum assumed for the calculations .
the level of the fluence of the cascade photons is @xmath40% of the primary synchrotron radiation .
this means that the maximum allowed baryon loading for these parameters can not exceed a factor of @xmath63 in order not to overproduce the primary synchrotron radiation fluence .
this limits the maximum number of @xmath2 to @xmath64 even in the case of large baryon loading for rare , powerful grbs .
these numbers can not be further increased in the sa model because of the efficient extraction that is already provided by internal photons alone for these parameters . and @xmath51 .
the dotted curve shows the fluence of a grb used for calculations , and the dashed and dot - dashed curves show the compton and synchrotron contributions to the photon fluence from the electromagnetic cascade initiated by secondaries from photomeson processes , respectively . , width=415 ] in conclusion , we predict that at most a few high - energy @xmath2 can be detected with icecube even from a very bright grb at the fluence level @xmath65 , and only when the baryon loading is high @xcite .
this is because the detection of a single @xmath2 requires a @xmath2 fluence @xmath66 above 1 tev .
since the energy release in high - energy neutrinos and electromagnetic secondaries is about equal , this energy will be reprocessed in the pair - photon cascade and emerge in the form of observable radiation at hard x - ray / soft @xmath0-ray energies , and this radiation can not exceed the measured fluence in this regime .
this imposes a robust limit on the maximum number of @xmath2 even from a grb with very high baryon loading .
aa thanks the nrl high energy space environment branch for support and hospitality during visits .
the work of cd is supported by the office of naval research and nasa glast science investigation grant dpr # s-15634-y . | acceleration of high - energy hadrons in grb blast waves will be established if high - energy neutrinos are detected from grbs .
recent calculations of photomeson neutrino production are reviewed , and new calculations of high - energy neutrinos and the accompanying hadronic cascade radiation are presented . if hadrons are injected in grb blast waves with an energy corresponding to the measured hard x - ray / soft @xmath0-ray emission , then only the most powerful bursts at fluence levels @xmath1 offer a realistic prospect for detection of @xmath2 .
detection of high - energy neutrinos are likely if grb blast waves have large baryon loads and doppler factors @xmath3 .
significant limitations on the hadronic baryon loading and the number of expected neutrinos are imposed by the fluxes from pair - photon cascades initiated in the same processes that produce neutrinos .
address = centre de recherche mathmatiques , universit de montral , montral , canada h3c
3j7 address = code 7653 , naval research laboratory , washington , dc 20375 - 5352 usa |
the dynamics of spinor becs is described by a three - component gross - pitaevskii ( gp ) system of equations . in the one - dimensional approximation the gp system goes into the following multicomponent nonlinear schrdinger ( mnls ) equation in 1d @xmath6-space @xcite : @xmath7 we consider becs of alkali atoms in the @xmath0 hyperfine state , elongated in @xmath6 direction and confined in the transverse directions @xmath8 by purely optical means .
thus the assembly of atoms in the @xmath0 hyperfine state can be described by a normalized spinor wave vector @xmath9 whose components are labeled by the values of @xmath10 .
the above model is integrable by means of the inverse scattering transform method @xcite .
it also allows an exact description of the dynamics and interaction of bright solitons with spin degrees of freedom .
matter - wave solitons are expected to be useful in atom laser , atom interferometry and coherent atom transport .
it could contribute to the realization of quantum information processing or computation , as a part of new field of atom optics .
lax pairs and geometrical interpretation of the mnls models related to symmetric spaces ( including the model ( [ eq:1 ] ) ) are given in @xcite .
darboux transformation for this special integrable model is developed in @xcite . in @xcite
the authors study soliton solutions for the multicomponent gross - pitaevskii equation for @xmath1 spinor condensate by two different methods assuming single - mode amplitudes and by generalizing hirota s direct method for multicomponent systems .
they point out the importance of integrable cases , which take place for particular choices of the coupling constants .
the aim of present paper is to show that both systems mentioned above are integrable by the inverse scattering method and are related to symmetric spaces @xcite @xmath2-type : @xmath11 with @xmath4 and @xmath5 respectively . in section 2
we formulate the lax representations for the models .
section 3 is devoted to the @xmath1 bec model .
in section 4 we construct the fundamental analytic solutions of the corresponding lax operator @xmath12 and reduce the inverse scattering problem ( isp ) for @xmath12 to a riemann - hilbert problem ( rhp ) . using the special properties of the @xmath2 symmetric spaces we also obtain the minimal sets of scattering data @xmath13 each of which allow one to reconstruct both the scattering matrix @xmath14 and the corresponding potential @xmath15 .
this allows us to derive in section 5 their soliton solutions using suitable modification of the zakharov - shabat dressing method , proposed in @xcite .
mnls equations for the * bd.i . *
series of symmetric spaces ( algebras of the type @xmath16 and @xmath17 dual to @xmath18 ) have the lax representation @xmath19=0 $ ] as follows @xmath20.\end{aligned}\ ] ] where @xmath21 the @xmath22-vectors @xmath23 and @xmath24 have the form @xmath25 while the matrix @xmath26 represents the metric involved in the definition of @xmath27 , therefore it is related to the metric @xmath28 associated with @xmath16 in the following manner @xmath29 next we will use @xmath30 we will use also the `` scalar product '' @xmath31 then the generic form of the potentials @xmath15 related to these type of symmetric spaces is @xmath32 where @xmath33 are the weyl generators of the corresponding lie algebra ( see @xcite for details ) and @xmath34 is the set of all positive roots of @xmath16 such that @xmath35 . in fact @xmath36 . in terms of these notations the generic mnls type equations connected to @xmath37 acquire the form @xmath38 in the case of @xmath4 if we impose the reduction @xmath39 and introduce the new variables @xmath40 , @xmath41 , @xmath42 then we reproduce the equations ( [ eq:1 ] ) .
let us introduce hamiltonian for mnls equations ( [ eq:4.2 ] ) with @xmath43 , @xmath44 @xmath45 define the number density and the singlet - pair amplitude by @xcite @xmath46 where @xmath47 , @xmath48 , @xmath49 , @xmath50,@xmath51 .
then the singlet - pair amplitude take the form @xcite @xmath52 the physical meaning of @xmath53 is a measure of formation of spin - singlet `` pairs '' of bosons .
the assembly of atoms in the @xmath1 hyperfine state can be described by a normalized spinor wave vector @xmath54 whose components are labeled by the values of @xmath55 . here the energy functional within mean - field theory @xcite is defined by @xmath56= \int_{-\infty}^\infty d x \left(\frac{\hbar^2}{2m}|\partial_x{\bf \phi}|^2+\frac{c_0}{2}n^2 + \frac{c_2}{2}{\bf f}^2+\frac{c_4}{2}|\theta|^2\right).\end{aligned}\ ] ] the coupling constants @xmath57 are real and can be expressed in terms of a transverse confinement radius and a linear combination of the @xmath58-wave scattering lengths of atoms @xcite and @xmath59 describe spin densities @xcite . choosing @xmath60 , @xmath61 and @xmath62 we obtain integrable by the inverse scattering method model with the hamiltonian .
we set for simplicity @xmath63 without any loss of generality .
the evolution equation is described by the multi - component gross - pitaevskii equation in one dimension @xcite @xmath64}{\delta { \bf \phi^*}}.\end{aligned}\ ] ] then we have @xmath65 or in explicit form by components we have @xmath66
herein we remind some basic features of the inverse scattering theory appropriate for the special case of @xmath1 spinor bec equations . solving the direct and the inverse scattering problem ( isp ) for @xmath12 uses the jost solutions which are defined by , see @xcite and the references therein @xmath67 and the scattering matrix @xmath68 . due to the special choice of @xmath17 and to the fact that the jost solutions and the scattering matrix take values in the group @xmath69 we can use the following block - matrix structure of @xmath70 @xmath71 where @xmath72 are @xmath22-component vectors , @xmath73 is a @xmath74 block and @xmath75 , @xmath76 are scalar functions satisfying @xmath77 . important tools for reducing the isp to a riemann - hilbert problem ( rhp ) are the fundamental analytic solution ( fas ) @xmath78 .
their construction is based on the generalized gauss decomposition of @xmath70 @xmath79 here @xmath80 , @xmath81 upper- and lower- block - triangular matrices , while @xmath82 are block - diagonal matrices with the same block structure as @xmath70 above . skipping the details we give the explicit expressions of the gauss factors in terms of the matrix elements of @xmath70 @xmath83 where @xmath84 , @xmath85 and @xmath86
if @xmath87 evolves according to ( [ eq:1 ] ) then the scattering matrix and its elements satisfy the following linear evolution equations @xmath88 so the block - diagonal matrices @xmath89 can be considered as generating functionals of the integrals of motion .
the fact that all @xmath90 matrix elements of @xmath91 for @xmath92 generate integrals of motion reflect the superintegrability of the model and are due to the degeneracy of the dispersion law of ( [ eq:1 ] ) .
we remind that @xmath93 allow analytic extension for @xmath94 and that their zeroes and poles determine the discrete eigenvalues of @xmath12 .
the fas for real @xmath95 are linearly related @xmath96 one can rewrite eq .
( [ eq : rhp0 ] ) in an equivalent form for the fas @xmath97 which satisfy also the relation @xmath98 then these fas satisfy @xmath99 obviously the sewing function @xmath100 is uniquely determined by the gauss factors @xmath101 . in view of eq .
( [ eq : s_jpm ] ) we arrive to the following [ lem : ms ] let the potential @xmath15 is such that the lax operator @xmath12 has no discrete eigenvalues .
then as minimal set of scattering data which determines uniquely the scattering matrix @xmath70 and the corresponding potential @xmath15 one can consider either one of the sets @xmath13 , @xmath102 @xmath103 \i ) from the fact that @xmath104 one can derive that @xmath105 for @xmath106 .
using the analyticity properties of @xmath107 we can recover them from eq .
( [ eq:25.3 ] ) using cauchy - plemelji formulae . given @xmath13 and @xmath108 one easily recovers @xmath109 and @xmath110 . in order to recover @xmath111 one again uses their analyticity properties , only now the problem reduces to a rhp for functions on @xmath69 .
the details will be presented elsewhere .
\ii ) obviously , given @xmath13 one uniquely recovers the sewing function @xmath112 . in order to recover the corresponding potential
@xmath15 one can use the fact that the rhp ( [ eq : rhp1 ] ) with canonical normalization has unique solution .
given that solution @xmath113 one recovers @xmath15 via the formula @xmath114 which is well known .
we impose also the standard reduction , namely assume that @xmath115 , or in components @xmath116 . as a consequence we have @xmath117 and @xmath118 .
the main goal of the dressing method @xcite is , starting from a known solutions @xmath119 of @xmath120 with potential @xmath121 to construct new singular solutions @xmath122 of @xmath12 with a potential @xmath123 with two additinal singularities located at prescribed positions @xmath124 ; the reduction @xmath125 ensures that @xmath126 .
it is related to the regular one by a dressing factor @xmath127 @xmath128 note that @xmath129 is a block - diagonal matrix .
the dressing factor @xmath130 must satisfy the equation @xmath131 = 0,\ ] ] and the normalization condition @xmath132 .
besides @xmath133 , @xmath134 and @xmath130 must belong to the corresponding lie group @xmath135 ; in addition @xmath130 by construction has poles and zeroes at @xmath136 .
the construction of @xmath130 is based on an appropriate anzats specifying explicitly the form of its @xmath137-dependence @xcite and the references therein .
@xmath138 where @xmath139 and @xmath140 are projectors whose rank @xmath58 can not exceed @xmath141 and which satisfy @xmath142 .
given a set of @xmath58 linearly independent polarization vectors @xmath143 spanning the corresponding eigensubspase of @xmath12 one can define @xmath144 taking the limit @xmath145 in eq .
( [ eq : u - eq ] ) we get that @xmath146 .\ ] ] below we list the explicit expressions only for the one - soliton solutions . to this end
we assume @xmath147 and put @xmath148 . as a result
we get @xmath149 where @xmath150 . repeating the above procedure @xmath151 times
we can obtain @xmath151 soliton solutions . in this case
@xmath152 so that the generic ( arbitrary @xmath141 ) one - soliton solution reads @xmath153 and @xmath154 , @xmath155 .
the polarization vectors satisfy the following relation @xmath156 thus for @xmath4 we identify @xmath157 , @xmath158 and @xmath159 and we obtain the following solutions for the equation ( [ eq:1 ] ) @xmath160 for @xmath5 we identify @xmath161 , @xmath162 , @xmath163 , @xmath164 and @xmath165 , so that the one - soliton solution for equation ( [ eq ] ) reads @xmath166 choosing appropriately the polarization vectors @xmath167 we are able to reproduce the soliton solutions obtained by wadati et al . both for @xmath0 and @xmath1 bec . here
@xmath168 and we have two linearly independent polarization vectors @xmath169 , @xmath170 . from eq .
( [ eq:31 ] ) we get @xmath171 the corresponding expressions for the rank 2 soliton solution are obtained by inserting eq .
( [ eq : r-2 ] ) into ( [ eq : q-1s ] ) and are rather involved .
we remark here that the reduction @xmath172 may not be sufficient to ensure that @xmath173 is positive for all @xmath6 and @xmath174 , so for certain choices of @xmath175 we may have singular solitons .
these and other properties of the rank 2 soliton solutions will be analyzed elsewhere .
the main result of the present paper is that a special version of the model describing @xmath1 spinor bose - einstein condensate is integrable by the ism .
the corresponding lax representation is naturally related to the symmetric space @xmath176 , see @xcite . for a generic hyperfine spin @xmath177 , the dynamics within the mean field theory is described by the @xmath178 component gross - pitaevskii equation in one dimension . if all the spin dependent interactions vanish and only intensity interaction exists , the multi - component gross - pitaevskii equation in one dimension is equivalent to the vector nonlinear schrdinger equation with @xmath178 components @xcite
. then equations ( [ eq:4.2 ] ) with the reduction @xmath179 are natural generalization of the vector nonlinear schrdinger equation , which adequately model the spinor bose - einstein condensates for values of @xmath180 equal to 1 and 2 . we expect that for generic @xmath177 these equations may be useful in describing becs with higher hyperfine structure . here we derived only generic one - soliton solutions . following the ideas of @xcite one can classify different types of one - soliton solutions related to different possible choices of the rank of @xmath139 and its polarization vectors .
one can also derive the @xmath151-soliton solutions by either repeating @xmath151 times the dressing with @xmath181 ( see eq .
( [ eq : rank1 ] ) , or considering more general dressing factors @xmath181 with @xmath182 zeroes and poles in @xmath95 .
these and other problems will be addressed elsewhere .
this work has been supported also by the national science foundation of bulgaria , contract no . f-1410 .
gerdjikov v. s. , grahovski g. g. , ivanov r. i. and kostov n. a. , @xmath183-wave interactions related to simple lie algebras .
@xmath184- reductions and soliton solutions , _ inv . problems _ * 17 * ( 2001 ) 9991015 .
g. g. grahovski , v. s. gerdjikov , n. a. kostov , v. a. atanasov , new integrable multi - component nls type equations on symmetric spaces : @xmath185 and @xmath186 reductions , in _ `` geometry , integrability and quantization vii '' _ , eds : i. mladenov and m. de leon , softex , sofia ( 2006 ) , pp .
154175 ; ( e - print : nlin.si/0603066 ) . | a three- and five - component nonlinear schrodinger - type models , which describe spinor bose - einstein condensates ( bec s ) with hyperfine structures @xmath0 and @xmath1 respectively , are studied .
these models for particular values of the coupling constants are integrable by the inverse scattering method .
they are related to symmetric spaces of @xmath2-type @xmath3 for @xmath4 and @xmath5 .
using conveniently modified zakharov - shabat dressing procedure we obtain different types of soliton solutions .
bose - einstein condensates , integrable systems , soliton models |
financial time series are hard to model , since they are heavily influenced by unpredictable events .
natural disasters , uncertainty about public behaviour , statements from governments and central banks , etc .
, are all events that can drastically affect the market . as a consequence financial data
do not behave the same at all times , hence we can not assume any stationarity property for them .
the latter implies that classic techniques used to analyse time series are widely inadequate to model such data , therefore alternative methods have to be developed .
the family of _ markov switching models _ ( msm ) constitutes a possible solution , since these models allow us to effectively address the non - stationarity of financial data .
the main idea behind the msm is that , in order to take into account the changes in the behaviour of the data , we allow the distribution of the observations to change over time . a general msm model can be written in the following form @xmath1 where @xmath2 indicates the state of the model at time @xmath3 , @xmath4 is the vector of the parameters characterizing the model , @xmath5 is the set of all observations up to time @xmath3 , @xmath6 is the set of all observed states up to time @xmath3 , @xmath7 is the set of all possible states , and @xmath8 is the function that governs the transitions between the states .
the function @xmath9 defines how the observation at time @xmath3 depends on @xmath10 and finally , @xmath11 , where @xmath12 , @xmath13 , is the so called _ terminal time_. system clearly shows the intrinsic richness of the msm approach .
particular realizations of allow the treatment of specific problems . before getting into the details of our study ,
it is worth mentioning that in most of the dedicated literature , we can distinguish between two classes of models .
the first class consists of models that have complicated distributions for the data or a large number of states , but very simple transition laws , e.g. , a first order markov chain , see , e.g. , @xcite .
the second class is made up of models with simple assumptions and very few states , usually two , but with more complicated transition laws , see , e.g. , @xcite .
the present paper is structured as follows : in sections 1 through 4 we cover the mathematical and computational theory needed to establish the techniques that we then develop in subsequent sections ; in section 5 we introduce the jump - diffusion model , while in section 6 we present a model that uses @xmath14stable distributions ; in section 7 we explain how the models can be concretely implemented and , in section 8 , we present the related results obtained applying them to a relevant case study which concerns the s&p500 index ; conclusions and further developments are outlined in section 9 .
bayesian inference is a branch of statistical inference that assumes the parameters of a probability distribution to be randomly distributed according to a _
prior _ distribution . in particular
the idea is to exploit the observed data , along with the bayes rule , to generate the _ posterior _ distribution of the aforementioned parameters .
therefore , the posterior distribution can be interpreted as the distribution of the parameters once we have taken into account both our subjective belief about them , namely the prior , and the data .
such an approach can be rigorously represented as follows @xmath15 where @xmath16 is the prior distribution , @xmath17 is the distribution of the data depending on the parameter @xmath4 , and @xmath18 is the posterior of @xmath4 .
finally , @xmath19 is the marginal distribution of @xmath20 , namely @xmath21 clearly the choice of the prior can have a large impact on the posterior .
a particularly convenient form of prior is what is known as a _
conjugate prior_. we say that a prior distribution is conjugate if the posterior distribution derived from it belongs to the same family , as it happens , e.g. , for the beta - bernoulli pair , namely @xmath22 it follows that if we start with a @xmath23 prior and assume that the data are binomially distributed , we end up with a @xmath24 posterior .
hence , we do not have to update the distribution for each new observation , just its parameters .
we would like to note that the latter is a particularly relevant aspect from the algorithmic point of view since it translates into less computationally expensive code .
for the sake of completeness , in the following subsections we list other particularly convenient choices for distribution pairs and , in order to give clear examples , we first start by explaining how the posterior of a set of independent identically distributed ( i.i.d ) random variables is obtained .
let @xmath25 be a set of i.i.d .
random variables with density function @xmath17 .
moreover , let @xmath26 .
then @xmath27 where @xmath28 is the likelihood function of the data , and @xmath29 is the set of all possible values of @xmath4 . for the rest of this paper we will denote the vector of observations by @xmath30 .
assume that we have @xmath31 independent observations @xmath32 where @xmath33 , and the value of @xmath34 is known . in order to perform bayesian inference on the given data , we also need to place a distribution on @xmath35 .
therefore , we set @xmath36 the corresponding likelihood function is @xmath37 while for the posterior we have @xmath38 hence @xmath39 assume again that we have @xmath31 independent observations @xmath40 but , this time , @xmath33 is known , while @xmath34 is unknown .
taking @xmath41 to be inverse - gamma distributed with parameters @xmath42 and @xmath43 , and denoting the distribution by @xmath44 , we can write the density function of @xmath45 as follows @xmath46 where @xmath47 is the extension of the factorial to the set of positive real numbers , known as the _ gamma function _ , and defined by @xmath48 therefore , the associated likelihood function is @xmath49 hence , the posterior is @xmath50 unfortunately , not all distribution pairs are as convenient as the previously mentioned ones , especially from the point of view of the parameter simulation needed by concrete computational studies . when the posterior is a well known distribution , as in the _ normal - normal _ and _ inverse gamma - normal _ cases , we can simulate the parameters using , e.g. , existing r libraries .
otherwise , _ ad hoc _ sampling algorithms have to be developed .
the next section addresses these problems .
in this section , we describe two methods that will be used to sample the parameters , namely , the _ gibbs sampling method _ and the _ metropolis - hastings algorithm_. the latter will be used in situations where the posterior distribution is non - standard , while the former will be used when the distribution can be simulated using an existing software . assume that we have a model with a finite number @xmath51 of parameters , @xmath52 , and that we want to find the full posterior distribution @xmath53 .
this goal can be quite difficult to reach , since the multivariate simulation of distributions is much more tangled and computationally heavy than its univariate counterpart .
the gibbs sampling approach allows the sampling of @xmath53 , knowing only the conditional distributions @xmath54 .
let @xmath55 be the number of simulations we want to perform .
we assign arbitrary starting values @xmath56 to each of the parameters .
then , for every @xmath57 , we perform the following steps @xmath58 hence we can simulate each of the model parameters .
the first @xmath59 simulations are discarded , being part of what is called the _ burn in period _ , in order to get rid of the dependence on the arbitrary choice of the starting point @xmath56 , while the remaining @xmath60 values are assumed to be a suitable approximation of the real distribution .
it is worth mentioning that the number of iterations , as well as the length of the _ burn in period _
, should be chosen carefully , since for larger values of @xmath55 the simulations become too time consuming , while small values might not provide enough iterations for the sampler to converge .
the gibbs sampler is rather easy to implement , but its major drawback is that it requires each @xmath61 to be readily samplable , where @xmath62 is the vector @xmath63 .
the metropolis - hastings algorithm allows for a solution to such an inconvenience .
in particular , it only requires a function @xmath64 proportional to the density function @xmath65 , and a _ proposal distribution _
@xmath66 which denotes a proper probability density function defined on the space @xmath29 of all possible values of @xmath67 . in what follows
we provide the description of the general metropolis - hastings algorithm , which uses the full parameter vector @xmath67 , as it is reported in @xcite .
we underline that the algorithm remains unchanged when @xmath67 is a scalar .
let @xmath55 be the number of simulations we want to perform .
we assign an arbitrary starting value @xmath68 to the parameter vector .
then , for every @xmath57 , we perform the following steps @xmath69 having defined a method to sample the parameters , we now have the task of simulating the states of the models we will be using .
this problem is the subject of the next section .
in this section our goal is to simulate the state vector @xmath70 . in order to accomplish this ,
we first need to obtain the values @xmath71 .
we start by setting arbitrary values for the parameters , and then we use the following expression @xmath72 notice that , @xmath73 , we can sample from @xmath70 if we have @xmath74 , which is nothing more than the transition probability from one state to another , and @xmath75 .
the latter can be obtained , @xmath76 , exploiting the hamilton filter , see below .
the basic hamilton filter , see @xcite , can be described as input - output - byproduct .
* input : * @xmath77 * output : * @xmath78 * byproduct : * @xmath79 running the hamilton filter for @xmath80 , we get the desired values @xmath81 , which can be used to generate @xmath70 , as described in what follows @xmath82 such a probability is used to draw a sample of @xmath83 , i.e. @xmath84 then , the above probability together with the previously simulated @xmath83 , are both used to simulate @xmath85 , and , proceeding iteratively , we have @xmath86 therefore , we can simulate @xmath87 obtaining the last component of @xmath70 .
the latter implies that , for every @xmath80 , we know what the distribution of @xmath88 is , because we know what the state we are in is . in the next two sections we present the models that will be used later
in the paper _ the variation of certain speculative prices _ , see @xcite , benoit mandelbrot draws attention to the fact that the normal distribution is inadequate when it comes to describing economic and financial data .
he argues that although the histograms of price changes seem to behave according to a gaussian distribution , a more careful analysis reveals that the large number of outliers makes the normal distribution fitted to the data much flatter than the actual data are , and with not enough density at the tails to include all the extreme values .
if one tries to manipulate the variance of the gaussian distribution to accommodate the values around the mean , then the result is a distribution that is even worse than the previous one where the extreme values are concerned . in what follows
we will show how to solve the aforementioned issue by using a gaussian distribution , to model the values around the mean , plus jumps of stochastic intensity , to include outlying values . specifically , our model is the following @xmath89 \end{cases}\ ] ]
we divide the analysis of the model defined in into two components , the _ gaussian component _ and the _ jump component_. we will use the gaussian distribution to model most of the data by means of the random variable @xmath90 , where both the mean and the variance of @xmath91 are state dependent . in particular , we define the state dependence of the mean as follows @xmath92 hence each state has its own , constant mean , without further restrictions . concerning the variance , we assume that it increases depending on the state , namely @xmath93 where , @xmath94 , @xmath95 , which gives us @xmath96 hence , by , as we go up in states we also go up in volatility . jump diffusion models , first introduced into finance by robert c. merton in @xcite ,
are currently widely accepted as an effective way to model the behaviour of financial data , see , e.g. , @xcite . in order to incorporate the jump feature in our model
, we have to deal with two major difficulties .
first , we have to find a distribution under which the sum of independent random variables behaves well , at least from the point of view of real statistical applications .
this task is not as straightforward as it may seem , since even the sum of i.i.d .
uniform random variables has a distribution that rapidly grows in complexity with the number of addends . to overcome this particular problem ,
we have chosen to exploit the exponential distribution to model the i.i.d .
jump amplitudes , as the sum of i.i.d .
exponential random variables follows a _ gamma distribution _ , namely @xmath97 where @xmath98 is the gamma distribution in the @xmath99 parameterization , while @xmath100 and @xmath55 are given natural numbers .
the aforementioned choice leads us to the second problem , which concerns the sign of the jumps . obviously , financial shocks can have both positive and negative values , while the gamma distribution allows only for the positive ones .
the issue can be solved by multiplying the sum in by a random variable @xmath101 taking values in @xmath102 , with equal probability .
we refer to the resulting distribution as the _ symmetric gamma distribution _ and we denote it by @xmath103 .
assuming now that @xmath104 , the probability density function of @xmath105 is given by @xmath106 hence @xmath105 has mean equal to zero , and variance @xmath107 looking at eq .
one can see that @xmath108 can be used to control how much @xmath0 influences the variance of the distribution .
for example , if we take two random variables @xmath109 , we have @xmath110 and @xmath111 which is a drastic increase in variance .
taking @xmath112 on the other hand gives us the variances @xmath113 and @xmath114 , which is a much smaller increase , for the same change in @xmath0 .
the previous , rather straightforward observation , will be useful later since in our model @xmath0 will represent the number of jumps at a certain point in time . taking
a large @xmath108 means that every extra jump only slightly increases the variance of the model , hence allowing for a finer analysis of the data .
the next step consists in determining the length @xmath55 of the sum in eq . .
in particular we assume that such a sum has a state dependent length represented by a state - dependent poisson random variable @xmath115 , hence we have to determine the values of @xmath116 . in keeping with the interpretation of the states , see eq . , we want the number of jumps to increase as the state the data are in increases .
this can be done by ordering the parameters @xmath116 . moreover , in order to also allow the parameters to be sufficiently flexible for our purposes , we assume them to be distributed as follows @xmath117 which clearly guarantees that @xmath118 .
summing up the definitions stated in subsections [ gaussianelement ] and [ jumpelement ] , we can write the full model as follows @xmath119 there is no analytic expression for the distribution of @xmath88 , but we can obtain an integral form of it using the following well known fact .
let @xmath105 and @xmath120 be two independent random variables with density functions @xmath121 and @xmath122 , defined for @xmath123 . then the sum @xmath124 is a random variable with density function @xmath125 given by @xmath126 therefore , by the convolution formula in , we have @xmath127 although not very useful in general , the expression in eq
. can be computationally handled with little difficulty , a crucial fact for the concrete case study we will consider in section [ casestudy ] . in the next section we consider the @xmath0-stable distribution model .
in section [ jumpdiffusionmodel ] , we pointed out that the gaussian distribution is not adequate to model financial data , mainly because of its slim tails , which we offset by adding jumps . in
what follows , we will consider a different approach , namely we will model the data using a distribution that has fatter tails than the gaussian one , but still preserves its most important characteristics .
there are multiple equivalent ways to define a stable distribution .
we will consider the two most common ones , the interested reader can refer to , e.g. , @xcite , for the others .
a random variable @xmath105 is said to have a stable distribution if , for every @xmath129 and @xmath130 positive , there exists a positive number @xmath131 and a real number @xmath132 such that @xmath133 where @xmath134 and @xmath135 are independent copies of @xmath105 and @xmath136 stands for _ equal in distribution_. this implies that the sum of two stable independent identically distributed random variables is still a stable random variable , with the same distribution , up to a _ scale factor _ @xmath131 , and a shift component @xmath132 . as an example
, we can consider two gaussian random variables @xmath134 and @xmath135 , assumed to be independent copies of @xmath137 .
then , @xmath138 , which means that @xmath139 .
alternatively , we can define the stable distribution using characteristic functions , namely a random variable @xmath105 is said to have a stable distribution if there exist parameters @xmath140 , such that its characteristic function has the form @xmath141 } , & \alpha \neq 1 \\
e^{\left [ - \sigma | \theta | ( 1 + i \beta \frac{2}{\pi } { \text{sgn}}(\theta ) \ln |\theta| ) + i \mu \theta \right ] } , & \alpha = 1 \end{cases } \:.\ ] ] we call @xmath0 the _ stability _ parameter , @xmath108 the _ skewness _ parameter , @xmath142 the _ scale _ parameter and @xmath35 the _ location _ parameter . for @xmath143 we obtain the normal distribution , which is the only member of the stable distribution family that has finite variance .
for @xmath144 we have infinite variance and mean @xmath35 , while , for @xmath145 $ ] , both the mean and the variance are undefined .
we note that in general there is not a solution in closed form for the probability density function of a stable distribution .
the stable distribution will be denoted by @xmath146 for the remainder of the paper . in the model
we propose , the data are assumed to follow a symmetric @xmath0-stable distribution , more precisely @xmath147 .
the full model is presented in the following @xmath148 \end{cases } \:.\ ] ] the motivation behind the choice of the model represented by , mainly relies on empirical observations of financial data which exhibit fat tails that can not be well described using the gaussian approach . in particular
we believe that such phenomenon can be suitably addressed by exploiting @xmath0-stable distributions with @xmath144 .
moreover , financial data often exhibit structural breaks because of abrupt changes in the market , e.g. as during the sub - prime mortgage credit crisis of 2008 , which is the reason why we consider both the scale and the location parameters , to be state - dependent .
as we previously mentioned , in general there is no closed form for the density of an @xmath0-stable distribution .
nevertheless , this problem can be circumvented using the fact that @xmath88 can be conditionally represented as a gaussian random variable , see , e.g. , @xcite , by introducing a random variable @xmath149 and using the property @xmath150 which allows us to have an analytic likelihood function which significantly speeds up the sampling process .
analogously to what we considered in section [ jumpdiffusionmodel ] , we have one mean for each state , without further restrictions , namely @xmath151 we also want the scale parameter to be increasing with respect to the state , namely @xmath152 where @xmath153 , which leads to the property @xmath154 so that an increase in the state number indicates an increase in volatility .
in this section we get into the specifics of our two models .
in particular we provide the details regarding the likelihood functions , the priors and the posteriors , for both the _ jump diffusion model _
, described in section [ jumpdiffusionmodel ] , and for the _ @xmath14stable model _ , defined in section [ alfastablemodel ] .
the concept of duration analysis is also explained , along with its importance .
both of the models will be characterized by four states , with the states being interpreted as _ low , medium , high _ and _ very high _ volatility _
regime_. let us start by defining the following quantities @xmath155 for the rest of this section we will suppress unneeded parameters .
therefore , e.g. , the conditional posterior @xmath156 , will be denoted by @xmath157 , the general rule being that the parameters that are not being inferred on are considered known .
the description of the implementation is divided into three parts , namely : the first part deals with the form of the likelihood function , the second with the priors while the third part provides a detailed analysis of the different types of obtained posteriors .
we have to take into account whether there are jumps in the model or not , as well as the state of each observation . hence ,
if @xmath158 , we define @xmath159 while , if @xmath160 , we define @xmath161 as @xmath162 then , the full likelihood function is @xmath163 which has a standard form only if @xmath164 , for every @xmath165 .
as this very rarely happens , we will use the metropolis - hastings algorithm in this model .
* mean : * we take the mean to be normally distributed .
moreover , we give the same prior to the means of all the states , namely @xmath166 * variance : * the variance @xmath167 will have an inverse - gamma prior .
@xmath168 * h parameters : * we previously saw that in order for to hold , we need @xmath169 , hence we define @xmath170 , for all @xmath171 , and make these parameters frchet distributed , namely @xmath172 then the density function of @xmath173 reads as follows @xmath174 where @xmath175 , and @xmath9 is defined for @xmath176 . * poisson parameters : * for the priors of the poisson parameters we refer to .
* transition probabilities : * for the transition probabilities we will use a dirichlet prior , namely @xmath177 for every @xmath165 .
the density function of this particular dirichlet distribution is given by @xmath178 and it is defined on the simplex @xmath179 while , everywhere else , its value is zero . finally , the parameter @xmath180 is a constant .
* mean : * because the likelihood function depends on whether or not jumps have occurred , we have two different posteriors for the mean .
in particular , if @xmath164 , by , we have @xmath181 for all @xmath165 , while , if @xmath160 , we obtain @xmath182 * variance : * similarly to the previous point , we have to differentiate between the jump and no - jump cases . therefore ,
if @xmath164 , by , we have @xmath183 otherwise , we obtain @xmath184 * h parameters : * in order to obtain @xmath185 , we need to transform the data , also taking into account the different states .
in particular we have the following cases @xmath186 @xmath187 where we have used the transformed data set @xmath188 to obtain a posterior for @xmath189 when @xmath190 , namely @xmath191 @xmath192 @xmath193 which gives us the posterior @xmath194 hence , when @xmath195 , the posterior is analogous to the one for @xmath189 , with the only difference being that we use @xmath196 instead of @xmath188 .
@xmath197 @xmath198 which yields the posterior @xmath199 in the case where we have jumps , i.e. @xmath200 , there is no analytic expression , therefore @xmath201 * poisson parameters : * concerning the posterior of the theta parameters , for @xmath165 , we have @xmath202 * transition probabilities : * the transition probabilities differ from the other parameters in that they do not depend directly on the observations @xmath203 . instead , they depend on the vector of states @xmath70 . assuming that the vector @xmath70 is known , the posterior distribution of the transition probability vector @xmath204 , @xmath165 , has the dirichlet distribution @xmath205 where @xmath206 is the number of transitions from state @xmath171 to state @xmath207 . in what follows , we proceed analogously to subsection [ subsec : jump_diffusion_model ] . using the fact that , in the present setting , our data are conditionally normal , see , the likelihood function reads as follows @xmath208 and , unlike in the previous model , we do not have to worry about multiple cases .
* mean : * the prior of the means is the same as in eq . .
* scale : * the distribution of the scale is analogous to that of the variance in the previous model , namely @xmath209 * h parameters : * these parameters are exactly the same as they were in the previous model , in fact their role remains unchanged , since they allow for the volatility to increase as the states increase . *
lambda : * the lambda parameter follows a stable distribution , hence @xmath210 * mean : * the posterior of the mean is analogous to that of the one in eq .
, with the only difference being the form of the variance .
in particular , we have @xmath211 * scale : * the posterior of the scale is @xmath212 * h parameters : * in what follows we limit ourselves to listing the needed transformations , therefore we have @xmath186 @xmath213 @xmath192 @xmath214 @xmath197 @xmath215 the posteriors are obtained as in the previous case .
* lambda : * since there is no closed form for the posterior distribution of the lambda parameter , we only write @xmath216 the prior and posterior of the transition probabilities are the same in both proposed models . in fact , the transition probabilities do not depend by any of the parameters , but only by the state vector @xmath83 .
the expected duration of each state for a msm is a quantity of significant interest . having an estimate of
how long a certain data set remains in a particular state can give us useful insights into how the model will behave for a certain period of time . in this section
we are going to explain how the expected duration can be calculated exploiting the transition probabilities .
the expected duration , denoted by @xmath217 , is defined as follows @xmath218\;,\ ] ] where @xmath219 is the random variable that models the length of the time interval for which the time series is in state @xmath171 .
the first thing we have to consider is @xmath220 , the probability of the data being in state @xmath171 , meaning @xmath221 where @xmath222 .
it just so happens that @xmath217 has a very simple closed form , in particular @xmath223 we will use the previous expression later on , when we compare the state durations obtained in the present paper , with those provided in @xcite .
we will see that there is a significant difference in the state durations , showing that the models developed in this paper perform better than the one proposed in @xcite to model the time series of the chicago board options exchange volatility index , better known as vix .
our case study is concerned with the application of the above theory to developing an indicator that has a role similar to the one played by the vix .
in particular we use the set of s&p500 weekly prices , considering a time interval that runs from the 3rd of january , 2007 to 29th of december , 2014 .
we picked this interval to include the sub - prime mortgage crash of 2008 as well as the subsequent period of relative calm .
this choice allows us to analyse how our approach performs in both situations .
we will show that our techniques improve the results stated in @xcite , where the model was very effective in periods of high volatility , but also too smooth in case of low volatility .
our results are summarised below with respect to both the _ jump diffusion model _ , defined in section [ jumpdiffusionmodel ] , and the _ @xmath14stable model _ , provided in section [ alfastablemodel ] . for the jump diffusion model we model the data as a zero mean process in order to make the framework more parsimonious .
we take the exponential distribution parameter @xmath180 to be equal to 40 , in order to make the contribution of each extra jump to the variance relatively small .
this choice of b allows for a finer analysis .
we first present the histograms of the sampled variances , see fig .
[ fig : variance_comparison ] .
as we can see , the algorithm is rather accurate in sampling the variances . in particular , we recall that the theoretical posterior of the variances is an inverse - gamma distribution , which is exactly what we can observe in the histograms .
moreover , in table [ tbl : variance1 ] , we report the point estimates of each variance value . .gaussian variance point estimates [ cols="^,^",options="header " , ] one thing that stands out in the mean point estimates , is the sign of the mean of the fourth state , which is negative . the latter should not come as a surprise since it refers to highest volatility value in the time series , namely the one related to the mortgage crisis of 2008 .
we recall that , during a severe financial crisis , most price movements are downward , resulting in a negative drift .
in what follows , we present the transition probability matrix , see eq . , and the transition matrix , see eq .
, namely @xmath224 @xmath225 as in the previous case , we note the expected values of the state durations obtained using eq . , namely @xmath226 we can note how the difference between the results in and those in , is significant . in order to better explain the latter datum , let us define the volatility indicator within the present framework , and make a comparison with the vix index .
in particular we define a second volatility indicator , denoted by @xmath227 , which , analogously to the previous case , will stand for the expected standard deviation of the data at time @xmath3 , i.e. @xmath228 in fig .
[ fig : stable_comparison ] , we can see a visual comparison between the two values . using eq . , we obtain @xmath229 , which is a significant increase over @xmath230 .
this leads us to conclude that the estimate obtained from the jump diffusion model is closer to the vix than the one obtained from the @xmath0-stable distribution model .
we now briefly explain the difference between the results in and .
the expected duration of state 1 falls while at the same time there is a drastic increase in the expected durations of states 3 and 4 .
looking at the way the estimators behave in fig .
[ fig : jump_comparison ] and in fig .
[ fig : stable_comparison ] , it is to note that the estimator obtained from the jump diffusion model is much more _ jagged _ , because of the regular transition from one state to another ; while the one obtained from the stable distribution model is much smoother , seeing as the time series tends to stay in the high volatility states much longer .
furthermore , when the stable distribution model has to place observations , that should be in the low volatility states , in the high volatility ones , so that to solve the variance underestimation problem mentioned in @xcite , the jump diffusion model can simply add a few jumps to make up for the missing variance .
this is why , despite its attempts to increase the variance by staying in the higher states , we see the indicator of the stable model _
drooping _ and underestimating the low volatility , while , in this situation , the jump model stays much closer to the vix .
in the present paper we have presented two novel techniques to implement a markov switching model ( msm ) type approach to non - stationary data , namely a jump diffusion - msm and an @xmath14stable - msm . in sec .
[ jump_diffusion_case ] , we have shown that the first one is very effective in mimic the vix index , moreover its implementation can be smoothly done without sacrificing its theoretical peculiarities , see sec .
[ jumpdiffusionmodel ] . a slightly different situation concerns the implementation of the second approach ,
[ alpha_stable_case ] , since eeven if the @xmath0-stable - msm approach turns to be quite effective , we have to consider sampling problems of one of its parameters , implying that computational results do not behave exactly the way the are meant to .
we would like to underline that the achieved tractability of the jump diffusion model is a crucial point , and it witnesses how such technique can be fruitfully used to model any kind of time series presenting pronounced tails , not just financial ones .
as far as the issues of over - smoothing and excessive state duration , which have been stated in @xcite as the main deficiencies of the msm approach to financial data , we have shown that , using the models here presented , the state durations can been significantly reduced , see subsection [ jump_diffusion_case ] , and the problem of over - smoothing can be solved , see subesctions [ jump_diffusion_case ] and [ alpha_stable_case ] .
concerning future developments , we aim at improving our jump diffusion - msm model by considering , instead of a simple first - order markov transition law , a @xmath51-th order markov transition law .
other possibilities consist in dealing with a transition law that is state duration dependent , or allowing the law to depend on other observable quantities used as indicators of the economy behaviour , e.g. , real personal income , industrial production index , rate of private credit growth , etc . +
* acknowledgements : * we would like to sincerely thank matteo frigo for his insightful comments and his fundamental suggestions , which have helped us a lot in preparing the present work , especially with respect to the duration analysis . | we perform a detailed comparison between a _ markov switching jump diffusion model _ and a _ markov switching @xmath0-stable distribution model _ with respect to the analysis of non - stationary data .
we show that the jump - diffusion model is extremely robust , flexible and accurate in fitting of financial time series .
a thorough computational study involving the two models being applied to real data , namely , the s&p500 index , is provided .
the study shows that the jump - diffusion model solves the over - smoothing issue stated in @xcite , while the @xmath0-stable distribution approach is a good compromise between computational effort and performance in the estimate of implied volatility , which is a major problem widely underlined in the dedicated literature , see , e.g. , @xcite .
markov switching , @xmath0-stable distribution , jump - diffusion model , symmetric gamma distribution , regime switching , markov chain monte carlo , metropolis - hastings algorithm . |
let @xmath0 be a set of points in the plane and let @xmath4 be the set of lines between pairs of points in @xmath0 .
consider the following operation on @xmath0 : @xmath5 in other words , @xmath1 is the set of intersection points between pairs of distinct lines in @xmath6 .
if @xmath0 consists of @xmath7 collinear points ( or no points at all ) , then the union above is empty ; so to keep the notation consistent , we set @xmath8 for these cases . for a set of points @xmath0 that form a trapezoid . ] as a simple example of the operation @xmath3 , let @xmath0 consist of four black points that are the vertices of a trapezoid as in figure [ fig.trapezoid ] . then , @xmath3@xmath9 consists of the original four points along with two additional ones shown in gray . it should be clear that for a set of points not all collinear , we have @xmath10 . moreover , @xmath1 is finite for finite sets @xmath0 .
we are interested here in the iterations , @xmath11 , and specifically , the limiting behavior of such operations on arbitrary finite sets @xmath0 .
the study of such phenomenon naturally leads to the notion of the order of a set @xmath0 , which we define below . as a matter of convention
, we set @xmath12 = s. let @xmath0 be a set of points in @xmath13 .
the _ order _ of @xmath0 is the smallest positive integer @xmath7 such that @xmath14 .
if there is no such @xmath7 , then the order of @xmath0 is defined to be @xmath15 .
for example , the order for a set of points forming the vertices of a square is 2 .
if the order of a set @xmath0 is 1 , then we call @xmath0 _ fixed _ under @xmath3
. a set @xmath0 , therefore , has finite order if and only if @xmath16 is fixed for some nonnegative integer @xmath7 .
[ fixedprob ] describe the finite point - sets that have finite order . before discussing the answer to this problem ( in section [ fixedsetssec ] )
, we describe a nontrivial infinite point - set that has finite order .
let @xmath0 be the set of rational points on the unit circle , @xmath17 .
for a given @xmath18 , choose two points @xmath19 and @xmath20 in @xmath0 such that @xmath21 and @xmath22 are not tangent to the unit circle .
then , if @xmath23 and @xmath24 are the points of intersection of @xmath21 and @xmath22 ( respectively ) with the circle , it turns out @xcite that @xmath23 and @xmath24 are both rational .
it follows that @xmath25 for every @xmath26 , and thus @xmath27 excluding the sets of finite order , it follows that iteration of @xmath3 produces a strictly increasing chain of sets of points in the plane . in light of this observation
, a natural question is whether we arrive at a dense set of points by such a procedure . in other words , is @xmath28 dense in @xmath13 ?
a more difficult but related question is whether we get all of @xmath29 when @xmath0 consists of only rational points .
we address both of these questions with a complete answer to the first in section [ denseproof ] and some partial results for the second in section [ rationalsection ] .
[ densethm ] let @xmath0 be a finite set of points in the plane .
then , @xmath0 has infinite order if and only if @xmath28 is dense in @xmath13 .
the answer to problem [ fixedprob ] found in corollary [ fixedthmcor ] below , therefore , gives a complete characterization of when iterated line intersections are dense .
[ fulldensechar ] let @xmath0 be a finite set of points in the plane .
then , @xmath30 is dense in @xmath13 if and only if @xmath0 is not one of the following sets : 1 . the empty set .
2 . a set of collinear points . 3 . a set of collinear points with one additional noncollinear point .
the vertices of a parallelogram . 5 .
the vertices of a parallelogram and the intersection of its two diagonals . in the rational case , we conjecture a more exact result . [ rationalthm ]
let @xmath0 be a finite set of points in the plane with rational coordinates .
then , @xmath0 has infinite order if and only if @xmath31 as a step in the direction of this conjecture , we offer the following ; its proof can be found in section [ rationalsection ] .
[ rationalparallelthm ] let @xmath32 be rational points in the plane with @xmath33 and @xmath34 parallel and suppose that @xmath35 is not parallel to @xmath36 .
then , @xmath37 .
though we were not motivated by any other particular work , we should remark that a similar question posed by fejes - toth ( with circles replacing lines ) was addressed by bezdek and pach in @xcite , and related results can also be found in the papers @xcite .
additionally , theorem [ densethm ] has also been discovered recently ( independently ) by ismailescu and radoicic @xcite .
we begin by characterizing sets of finite order .
although one may deduce the main result of this section from lemmas [ trianglelemma ] and [ trianglelemma2 ] in section [ denseproof ] , the methods employed here are less cumbersome and might be of independent interest .
we will need the following result from elementary geometry .
[ sylthm ] for every set of @xmath7 noncollinear points in the plane , there exists a line that contains exactly two of the points . although this fact seems intuitively obvious , its proof eluded even sylvester , and it was only solved ( in published form ) some @xmath38 years after being posed by him @xcite .
we refer the reader to @xcite for more details .
we are ready to approach problem [ fixedprob ] .
[ fixedthm ] a finite set @xmath0 fixed under @xmath3 must be one of the following configurations : 1 . the empty set .
2 . a set of collinear points with one additional noncollinear point .
3 . the vertices of a parallelogram and the intersection of its two diagonals .
let @xmath0 be a set of @xmath7 noncollinear points in the plane that is fixed by @xmath3 .
using theorem [ sylthm ] , there exists a line intersecting @xmath0 in exactly two points @xmath39 and @xmath40 . by assumption
, there is some other point @xmath41 not on this line , and we can choose @xmath41 so that its altitude from @xmath42 is largest . if all other points lie on the line @xmath43 or if all of them lie on @xmath44 ,
then we are in configuration ( 2 ) above . the remaining possibilities break up into two cases . ] : there is a point @xmath45 not on @xmath43 and not on @xmath44 .
we first claim that @xmath46 must lie on the line through @xmath41 that is parallel to @xmath42 . indeed , any other position for @xmath46 would give rise to an intersection between @xmath47 and @xmath42 that is not @xmath39 or @xmath40 , contrary to our use of theorem [ sylthm ] and our assumption that @xmath48 .
relabeling if necessary , figure [ fig.fixedproof1 ] depicts the situation .
since @xmath0 is fixed , the intersection point , @xmath49 , of @xmath44 and @xmath50 is in @xmath0 .
it follows that @xmath43 and @xmath51 must be parallel ( otherwise , if @xmath52 is the intersection point of @xmath43 and @xmath51 , then @xmath53 would intersect @xmath42 ) .
finally , it is easy to see that there can be no other points in @xmath0 by our choice of @xmath39 and @xmath40 . ] : every point in @xmath0 lies on one of the lines @xmath43 or @xmath44 .
if @xmath0 is not a configuration of type ( 2 ) , then there are points @xmath54 such that @xmath55 is on the line @xmath43 , @xmath56 is on the line @xmath44 , and @xmath57 are not @xmath58 , or @xmath40 . by the assumption on @xmath41 and the line @xmath42 , only two configurations for @xmath55 and @xmath56
are possible ; these are depicted in figure [ fig.fixedproof2 ] . in both cases , two iterations of @xmath3
give rise to a point in @xmath0 on the line @xmath42 , a contradiction .
therefore , no fixed point - sets other than those of configuration ( 2 ) may take this form .
this completes the proof .
[ fixedthmcor ] the finite point - sets with finite order are 1 . the empty set .
2 . a set of collinear points . 3 . a set of collinear points with one additional noncollinear point .
the vertices of a parallelogram . 5 .
the vertices of a parallelogram and the intersection of its two diagonals .
let @xmath0 be a finite set in @xmath13 with order @xmath7 .
applying theorem [ fixedthm ] , it follows that @xmath59 must be one of three types . when @xmath55 is empty , then @xmath0 is either itself empty or a set of collinear points .
similarly , a set @xmath55 of collinear points with one additional point can only be obtained from a set @xmath0 that is the same as @xmath55 . finally , when @xmath55 forms a parallelogram with the intersection of its diagonals , the set @xmath0 must either be @xmath55 or @xmath55 without its diagonal intersection .
before proving theorem [ densethm ] , we record the following technical lemmas , the first of which provides a useful characterization of sets of infinite order . for ease of presentation , we say that a point is _ strictly contained _ in a set @xmath60 if it is located in its interior .
[ trianglelemma ] let @xmath0 be a finite set of infinite order .
then , there exists @xmath61 such that @xmath16 contains a subset of 4 points in which 3 of the points are noncollinear and the fourth point is strictly contained in the triangle determined by these 3 points .
we consider the number of vertices @xmath62 on the convex hull @xmath63 of @xmath0 .
when @xmath64 , the set @xmath0 can not have infinite order .
so suppose that @xmath65 .
if there is a point of @xmath0 strictly contained inside @xmath63 , then we are done . otherwise , since @xmath0 has infinite order , there must be two points of @xmath0 on different edges of @xmath63 .
an iteration of @xmath3 then produces our desired point .
] assume now that @xmath63 has exactly four vertices .
if these vertices do not form a parallelogram , then one iteration of @xmath3 gives us what we want ( see figure [ fig.trianglempic ] ) .
otherwise , there is a point in @xmath0 which is not a vertex of @xmath63 and not the intersection of the diagonals of the quadrilateral determined by @xmath63 . again in this case , one iteration of @xmath3 ( giving us the intersection of the two diagonals of @xmath63 ) produces the desired result . finally ,
if @xmath66 , then we proceed as follows .
pick two adjacent vertices @xmath19 and @xmath20 .
there must be two other vertices @xmath23 and @xmath24 such that the edges @xmath67 and @xmath68 are not parallel ( @xmath63 has at least @xmath69 vertices and is convex ) .
this reduces the problem to the case of @xmath70 vertices not forming a parallelogram ( encountered above ) and completes the proof of the lemma .
our next result allows one to produce a convergent , nested sequence of triangles .
[ trianglelemma2 ] let @xmath19 , @xmath20 , and @xmath23 be noncollinear points , and let @xmath39 be a point strictly inside @xmath71 .
then , there exist triangles @xmath72 ( @xmath73 ) strictly containing @xmath39 such that @xmath74 , and for each @xmath7 , @xmath75 given a triangle @xmath71 and a point @xmath39 strictly contained in it , we may construct the vertices of another triangle containing this point by intersecting the lines @xmath76 , @xmath77 , and @xmath78 with the edges of @xmath71 .
iterating this procedure produces a nested sequence of triangles strictly containing @xmath39 with vertices in @xmath79 ( see figure [ fig
. nestedtriangles ] ) .
this sequence contains two types of triangles ; we label the odd iterates @xmath80 , while even iterates are denoted by @xmath72 . here , the @xmath81 ( resp .
@xmath82 , @xmath83 ) are labeled so that they are the ones on the line @xmath76 ( resp .
@xmath77 , @xmath78 ) .
we claim the vertices of the triangles @xmath72 all converge to @xmath39 .
to verify this assertion , it suffices to show that @xmath84 , @xmath85 , and @xmath86 . without loss of generality ,
we prove that @xmath84 . reducing further
, we observe that it is enough to show that the area of @xmath87 is less than the area of @xmath88 ( drop altitudes to @xmath89 from @xmath19 , @xmath39 and compare similar triangles ) .
next , draw the line @xmath90 that is parallel to @xmath89 and passes through @xmath39 , and label the angles formed as in figure [ fig .
nestedtriangles2 ] . since @xmath91 and @xmath92 ( resp . @xmath93 and @xmath94 ) intersect at @xmath20 ( resp .
@xmath23 ) , it follows that @xmath95 and @xmath96 .
therefore , when we form the triangle @xmath97 that is congruent to @xmath87 , it must lie entirely inside @xmath88 .
this finishes the proof .
[ densetrianglelem ] let @xmath19 , @xmath20 , @xmath23 be noncollinear points in the plane . if @xmath60 is a dense set of points in @xmath71 , then @xmath98 is a dense set of points in the entire plane .
let @xmath39 be a point in the plane , and let @xmath99 and @xmath100 be points strictly inside @xmath71 such that @xmath101 and @xmath102 intersect at @xmath39 . since @xmath60 is dense in @xmath71 , there are a sequence of points @xmath103 and @xmath104 that converge to @xmath99 and @xmath100 , respectively . since the intersection of two lines formed by four points is continuous in the four points ( the intersection is a rational function in the coordinates of the four points ) , it follows that the intersections of @xmath105 and @xmath106 ( which are in @xmath98 ) converge to @xmath39 .
this completes the proof .
we are ready to prove theorem [ densethm ] .
the if - direction ( @xmath107 ) in the theorem statement is immediate
. therefore , let @xmath0 be a finite set of infinite order .
using lemma [ trianglelemma ] , there exists @xmath108 such that @xmath16 contains a triangle of vertices and a fourth point strictly contained in the triangle determined by these 3 vertices .
we claim that iteration of @xmath3 on these 4 points produces a dense set of points in the triangle .
the theorem then follows from lemma [ densetrianglelem ] .
let @xmath19 , @xmath20 , and @xmath23 be the vertices of the triangle strictly containing @xmath39 .
suppose that @xmath109 does not contain a dense set of points in @xmath71 ; we will derive a contradiction . using lemma [ trianglelemma2 ]
, we can produce a sequence of triangles , @xmath110 , with vertices in @xmath60 such these vertices converge to @xmath39 .
let @xmath111 be so large that the circle centered at @xmath39 with radius equal to twice the largest side of @xmath112 is strictly contained in @xmath71 . since @xmath60 is not dense in @xmath71 , it follows that @xmath60 can not be dense in @xmath112 ( again using lemma [ densetrianglelem ] ) .
let @xmath113 be the closure of @xmath60 and set @xmath114 .
also , let int@xmath115 denote the interior of @xmath116 . since @xmath60 is not dense in the triangle @xmath112 , the ( nonempty ) open set int@xmath117 contains an open ball centered at some point @xmath41 inside @xmath112 .
consider the set of all closed balls centered at @xmath41 that do not intersect @xmath113 , and let @xmath118 denote the supremum over all radii of such balls .
the closed ball @xmath119 of radius @xmath120 centered at @xmath41 must be strictly contained in @xmath121 since its interior can not contain @xmath122 , @xmath123 , or @xmath124 ( they are in @xmath60 ) and because of how we chose @xmath111 . by construction of @xmath119
, there exists a point @xmath125 intersecting the boundary of @xmath119 .
consider the lines @xmath126 , @xmath127 , and @xmath128 , and notice that they can not all be tangent to the ball @xmath119 ( there is only one tangent line through a point on a circle ) .
therefore , at least one of these lines through @xmath46 , say @xmath126 , must intersect the interior of @xmath119 .
let @xmath49 be the intersection of the line @xmath126 with the boundary of @xmath119 ( the point @xmath49 need not be in @xmath113 ) .
the situation is depicted in figure [ fig.balltangent ] .
the dashed line through @xmath46 is the line tangent to the boundary of @xmath119 at @xmath46 , while the dashed line through @xmath49 is parallel to it . to continue , we observe the following straightforward fact that was discussed in the proof of lemma [ densetrianglelem ] : if @xmath129 determine two nonparallel lines @xmath130 and @xmath131 , then the intersection point of @xmath130 and @xmath131 is in @xmath113 .
with this observation in mind , we may use lemma [ trianglelemma2 ] to obtain vertices of triangles @xmath132 in @xmath113 that contain @xmath46 and that also converge to @xmath46 .
none of the vertices @xmath133 , @xmath134 , or @xmath135 is in the interior of @xmath119 by our choice of @xmath120 .
finally , we claim that for large enough @xmath7 , the segment @xmath136 must intersect a side of @xmath137 in the interior of @xmath119 , a contradiction to our assumption on @xmath120 . to see this , notice that for a large @xmath7 , at least one of the vertices of @xmath137 must lie between the two parallel lines ( depicted in figure [ fig.balltangent ] ) through @xmath138 , while none of them will lie beneath the line through @xmath49 .
it follows that an edge of @xmath137 intersects the line @xmath126 inside @xmath119 .
this contradiction completes the proof .
we now turn our attention to the case of rational points as in the statement of conjecture [ rationalthm ] .
we note the following simple observation .
[ 6pointlemma ] suppose that @xmath139 or that @xmath140 .
then , @xmath31 iteration of @xmath3 on both sets above gives all of @xmath141 , and it is easily verified that @xmath141 generates all of @xmath2 . ] we next restrict our attention to a particular case involving a pair of parallel lines .
we need the following fact from plane geometry .
[ midpointlemma ] let @xmath142 be points in the plane with @xmath33 and @xmath34 parallel and suppose that @xmath35 is not parallel to @xmath36 .
let @xmath46 be the intersection of @xmath143 and @xmath42 and set @xmath41 to be the intersection of @xmath35 and @xmath36 . then , @xmath47 intersects @xmath33 and @xmath34 in their midpoints @xmath144 and @xmath145 , respectively .
since @xmath146 and @xmath147 are similar triangles , we have @xmath148 .
the same reasoning gives us that @xmath149 .
examining the large triangles @xmath150 and @xmath151 , it is also clear that @xmath152
. therefore , @xmath153 so that @xmath154 . a similar computation shows that @xmath155 .
we finally arrive at our main result in the rational case
. it will be a consequence of lemma [ midpointlemma ] , and it is the closest we come to proving conjecture [ rationalthm ] . since a ( rational ) translation does not change the problem
, we may assume that @xmath156 .
moreover , it is easy to see that if @xmath157 , then @xmath158 \ : \ ( a , b ) \in s\right\}\ ] ] gives rise to @xmath29 through iteration of @xmath3 if and only if @xmath0 does .
suppose that @xmath159 , @xmath160 , and @xmath161 with @xmath162 .
since @xmath33 and @xmath34 do not define the same line , it follows that @xmath163 . also , since @xmath33 and @xmath34 are parallel , we have @xmath164 .
consider the following matrix : @xmath165.\ ] ] a straightforward computation gives @xmath166 moreover , since @xmath35 is not parallel to @xmath36 , it follows that @xmath167 .
next , set @xmath168 in which @xmath169 and gcd@xmath170 .
suppose first that @xmath171 .
by successively applying lemma [ midpointlemma ] , iteration of @xmath3 on @xmath172 produces the points : @xmath173 it follows that if we choose @xmath174 such that @xmath175 , we will have @xmath176 therefore , letting @xmath177 $ ] , we must have @xmath178 an application of lemma [ 6pointlemma ] now concludes the proof of this case . finally , if @xmath179 , then the same examination as above reduces the situation to @xmath140 , also covered by lemma [ 6pointlemma ] .
we would like to thank the anonymous referees for introducing us to the references @xcite and for several suggestions that improved the exposition of this paper .
special thanks also go to kelli carlson for helping to simplify the proof of lemma [ trianglelemma2 ] and for giving useful comments on preliminary versions of this paper . | let @xmath0 be a finite set of points in the plane and let @xmath1 be the set of intersection points between pairs of lines passing through any two points in @xmath0 .
we characterize all configurations of points @xmath0 such that iteration of the above operation produces a dense set .
we also discuss partial results on the characterization of those finite point - sets with rational coordinates that generate all of @xmath2 through iteration of @xmath3 . |
globular clusters ( gcs ) , relics of some of the earliest and/or most violent phases of star and galaxy formation , can be analyzed to understand how various types of galaxies formed , as described in the review by @xcite .
two subpopulations of gcs , metal - rich ( red ) and metal - poor ( blue ) , exist in most early - type galaxies .
both are thought to be old , but the red metal - rich gcs are thought to be slightly younger @xcite .
large spiral galaxies , including the milky way @xcite and andromeda @xcite , also have metal - rich and metal - poor gc subpopulations . in smaller , later - type galaxies ,
however , often only a metal - poor population is seen @xcite .
the three basic models for bimodal gc system ( gcs ) formation in early type galaxies are dwarf galaxy accretion @xcite , in situ formation @xcite , and gas - rich mergers @xcite . in the last scenario , the original spiral galaxies likely late - type spirals
would provide the entire metal - poor gc population , and the metal - rich gc population would form in the merger . some evidence for gcs resulting from every one of these processes exists , but the properties of both blue and red gc populations have been found to scale with the mass of the host galaxy @xcite , indicating that the accretion model alone can not account for the majority of blue gcs in bright galaxies . in any case , at high redshift and with a hierarchical merging scenario as the main process of galaxy formation , the distinction between these models fades . the nearby sculptor group ( or filament ) is home to several late - type galaxies ( hubble types sc
im ) , including ngc 253 ( the largest ) , ngc 300 , ngc 55 , and ngc 45 . at a distance of 1.9 mpc @xcite ,
ngc 300 is one of the nearest spiral galaxies outside of the local group .
as a nearby example of a late - type , midsize spiral galaxy , it is especially useful for understanding the sparse and thus relatively poorly studied gcss of late - type galaxies .
@xcite studied the star clusters in ngc 300 photometrically and identified 17 objects as gc candidates based on their size , shape , and color . @xcite
performed photometry on fields in six sculptor group galaxies and identified gc candidates via shape and color , and then obtained spectra for 19 gcs .
six of these spectroscopically confirmed clusters had high enough signal - to - noise spectra for metallicity determination via lick / ids indices .
seven of the gcs olsen et al .
spectroscopically confirmed were in ngc 300 , and two had spectroscopically determined metallicities . in this paper
we present new metallicities derived from spectroscopy of gcs in sculptor group galaxies , primarily ngc 300 , using @xcite and @xcite as reference catalogs .
we compare the metallicities of ngc 300 gcs to those of m31 and milky way clusters , and also determine ngc 300 s place on the galaxy luminosity - gcs metallicity relation of @xcite .
we obtained spectra of 44 sculptor group cluster candidates with the boller & chivers ( b&c ) spectrograph on the baade telescope from 2002 november 6 to 2002 november 9 .
we chose to observe all 17 gc candidates from @xcite , plus 25 @xcite ngc 300 gc candidates with no previous spectra ( except for ngc 300ax ) observed in order of increasing magnitude from brightest to faintest .
we also observed the two brightest unconfirmed olsen et al .
gc candidates in ngc 253 .
nineteen of our 27 olsen et al .
cluster candidates are among the 38 sculptor group gc candidates for which we also obtained _ jhk _ photometry with the panic camera on the baade 6.5 m telescope ( * ? ? ?
* paper i ) .
figure 1 shows the locations of the ngc 300 objects we observed with respect to a schematic of the ngc 300 disk .
the advantages of using the long - slit b&c spectrograph rather than a fiber spectrograph are good coverage in blue optical wavelengths and better sky subtraction ( due to the sky background being determined at the same time and place as the object exposure ) . also , the sky at las campanas has significantly weaker telluric emission lines than less isolated telescope locations .
our spectra were taken with the 600 l mm@xmath0 grating blazed at 5000 .
the wavelength range is about 3700 - 6860 .
peak count rates occurred between 5000 and 5500 for objects with typical gc colors .
the dispersion of the spectra is @xmath31.6 pixel@xmath0 , and the resolution of the spectra is about 5 .
the readout noise is 3.1 e@xmath4 .
total exposure times were as low as 180 s for very bright objects and as high as 3600 s for faint objects , with 600 - 2400 s exposure times for most objects .
exposure times under 600 s were usually done as single exposures , while longer exposures were usually broken up into 2 - 4 exposures of 600 - 900 s each . seeing ranged from 0.6 - 1.9@xmath5 over the course of obesrvations , with 1@xmath5 being typical , and the slit width used was 1@xmath5 .
the median total signal - to - nose ratio ( all wavelengths combined , @xmath33700 - 6875 ) was about 20 per pixel for foreground stars and gc candidates , with values ranging from 4 to over 100 .
if background galaxies , easily distinguished from these objects by their radial velocities , are included , the mean s / n per pixel is about 15 .
bias subtraction ( with an averaged bias image ) , dark count correction , and flatfielding were done with iraf s ccdproc task , and spectra were extracted using iraf s apall task .
other tasks in the specred and onedspec packages were used to combine spectra , calibrate wavelengths , flux - calibrate the spectra using standard star observations taken each night , and eliminate pollutants such as cosmic rays and improperly - subtracted telluric lines .
velocities were determined with the xcsao task in the rvsao package @xcite in iraf using a list of standard sao templates useful for the expected features of gcs , background galaxies , and foreground stars . included
are four templates derived from m31 gcs ( m31_a_temp , m31_f_temp , m31_k_temp , and fglotemp ) , three galaxy absorption templates ( fm32temp , habtemp90 , and fn4486btemp ) , two composite stellar absorption templates ( fabtemp97 and fallstars ) , a synthetic h region / galaxy emission template ( hemtemp0.0 ) , and a synthetic ca h&k absorption template ( hkabstemp ) . for each object , we list a heliocentric velocity determined from the template with the highest r value .
objects labeled as clusters usually had spectra similar to g and early k giants .
mid - type k stars ( @xmath3 k3-k5 ) typically differed visibly from gc candidates in having a redder spectral energy distribution and strong mg and ca42 features . in order to provide a systematic , quantitative classification scheme to remove k dwarfs from the pool of gc candidates , we focused on the strong ca42 feature and compared it to two other spectral features , g43 and h@xmath6 ( using the h@xmath6a index definition in @xcite ) .
the difference between the ca42 and h@xmath6 indices was used by @xcite to distinguish between k dwarfs and gc candidates . in our own samples ,
the ca42 index also appeared strong in comparison to the g43 index in objects that had the spectral features of mid - k - type stars . to quantify the differences in these features between gcs and k dwarfs , we measured the ca42 , h@xmath6a , and g43 indices in the 41 @xcite milky way gc spectra and in 16 k dwarfs from the jones coudfeed spectral library described in @xcite .
the k dwarfs range in type from k0 to k8 .
figure 2 shows histograms of ca42-g43 and ca42-h@xmath6a ratios for the milky way gcs and jones coudfeed k dwarfs respectively .
there is some overlap between the distributions of ca42-g43 for gcs and k dwarfs , particularly among early k - type stars ( k0-k2 ) , but this index appears useful for screening out mid - k - type stars in our sample .
the ca42-h@xmath6a ratio seems more effective in discriminating between gcs and k dwarfs , with the only k dwarf overlapping the mw gc range of values being a k8 star , which in our ngc 300 sample would be easily distinguishable from a gc . the milky way gcs had a mean ca42-g43 value of -0.057 with a @xmath7 of 0.024 and a mean ca42-h@xmath6a value of 0.016 with a @xmath7 of 0.075 . for an object to be considered a gc candidate
, we insisted that it have both ca42-g43 @xmath8 0.05 ( the value for ngc 300 - 05 which , despite having a ca42-g43 ratio more than 3@xmath7 above the mw gcs , has been visually confirmed as a gc via hst archive images ) and ca42-h@xmath6a @xmath8 0.241 ( @xmath9 above the milky way gc mean ) .
this analysis excludes ngc 300cr and ngc 300df on the basis of ca42-h@xmath6a but not ca42-g43 ( which one might expect for k0-k2 dwarfs ) , and ngc 300ax , previously classified by @xcite as a gc , on the basis of ca42-g43 but not ca42-h@xmath6 .
radial velocity can provide limited additional information on the likelihood of an object being a star or a gc .
since the heliocentric radial velocity of ngc 300 is fairly low + 142 @xmath10 4 km s@xmath0 @xcite it is impossible to eliminate all stars and keep all gcs in the sample if a strict velocity cut is used to exclude candidates .
hence , we can not automatically label objects with clusterlike spectra but very low radial velocities as definite stars .
the velocity dispersion in ngc 300 s halo is estimated by @xcite to be 60 km s@xmath0 . if the milky way halo were non - rotating , the average expected radial velocity of a galactic star in the direction of ngc 300 would be about + 40 km s@xmath0 , and the velocity dispersion of the milky way halo is about 130 km s@xmath0 ( as viewed from the center of the milky way ) .
so while there s a 95% chance that an ngc 300 halo object will have a radial velocity greater than 20 km s@xmath0 , milky way halo stars can very easily have velocities similar to those of genuine ngc 300 gcs .
this means that most objects with negative radial velocities will be milky way stars , but those with positive radial velocities could easily be either foreground stars or ngc 300 clusters , and only spectral features or high - resolution imaging can distinguish them .
on the basis of radial velocity , we flag three objects with spectral features similar to gc candidates - ngc 300 cm , ngc 300co , and ngc 300db - as possible stars .
all have radial velocities less than 0 km s@xmath0 .
figure 3 shows the correlations between our two indicators of star vs. gc status , ca42-g43 and ca42-h@xmath6a , and the radial velocities of foreground stars and gc candidates in ngc 300 .
both trends are high in scatter , demonstrating the difficulty in measuring star vs. gc status on the basis of radial velocity .
the correlation between ca42-h@xmath6a and radial velocity is somewhat stronger , probably due to its greater sensitivity to k0-k2 dwarfs compared to ca42-g43 .
the best - fit velocity template also hints at whether an object is likely to be a star or a cluster .
the templates we used included several templates developed using intermediate - age and old stellar populations ( m31_a_temp , m31_f_temp , m31_k_temp , fglotemp , and fm32temp ) , a couple templates suited for field stars ( fallstars , fabtemp97 ) , and a couple synthetic templates ( hemtemp0.0 for h emission and hkabstemp for ca h & k absorption ) .
most gcs were best fit by the m31 f and k templates , while many foreground dwarfs were best fit by the stellar absorption templates .
figure 4 shows sample spectra of an old gc , a typical foreground k star , and a background galaxy .
the region shown in the plot covers wavelengths from 3700 to 5300 , containing key features such as cnb , ca 4227 , g4300 , h@xmath11 , mgb , and several fe features .
one can see the k star s strong ca 4227 and mgb features , compared to the relatively weak ca 4227 and mgb and strong g4300 of the gc .
one can also see that in the foreground star , the continuum on the blue side of the mg2 feature tends to be notably fainter than on the red side .
the galaxy is recognizable by its visibly redshifted ca h&k lines .
the names , heliocentric cross - correlation velocities , and object types ( background galaxy , foreground star , cluster candidate ) are shown in table 1 . the objects with number designations 1 - 17 are from kim et al .
2002 ; the rest are from olsen et al .
2004 . of the 17 kim et al .
candidates , only five objects 2 , 3 , 4 , 5 , and 12 appear to be old gcs .
note that object 3 is also ngc 300a in olsen et al.s catalog , and has been proven a gc and spectroscopically evaluated by olsen and collaborators .
most of the remaining kim et al .
objects were background galaxies , except for one h region ( object 6 ) and three probable ngc 300 stars ( ngc 300 - 09 , m - type ; ngc 300 - 10 , f - type ; and ngc 300 - 15 , a - type ) . since gcs in ngc 300
can easily be resolved with hst , we searched the hst acs and wfpc2 archives for images of our ngc 300 objects .
the images were from programs go-8584 , go-9162 , go-9492 , and go-10915 .
we found eight kim et al .
objects in the archival hst images : objects 1 , 5 , 6 , 7 , 8 , 9 , 10 , and 11 .
these eight objects , imaged in f555w for the acs images of objects 1 , 5 , 6 , 7 , and 11 and f606w for the wfpc2 images of objects 8 , 9 , and 10 , are shown in figure 5 .
generally , their appearances match what we expect based on their spectra .
objects 1 , 7 , and 11 are spiral galaxies ; object 8 is an early - type galaxy ; object 6 is an h complex ; object 5 appears to be a genuine gc ; and objects 9 and 10 , initially considered as possible open clusters , appear pointlike but non - saturated and without diffraction spikes .
their @xcite v magnitudes are 19.73 for ngc 300 - 09 ( which has an m - type spectrum ) and 19.12 for ngc 300 - 10 ( which has an f - type spectrum ) , consistent with supergiant stars in ngc 300 .
they both have many fainter stars surrounding them , so may also be open clusters or stellar blends .
kim et al . selected their gc candidates by using a color cut ( 0.3 @xmath12 @xmath13 @xmath12 2.0 ) and morphological considerations , including visual inspection of the shapes and brightness profiles .
they assigned three `` classes '' to the quality of gc candidates , class 1 being most likely to be gcs and class 3 being least likely .
interestingly , class 2 yielded the highest fraction of objects that turned out to be gcs : 3 out of 7 had gc spectra , as compared to 1 out of 4 for class 1 and 1 out of 6 for class 3 .
kim et al . reported using morphological considerations to determine the probability of an object being a gc .
if they required that the clusters be resolved from the ground to be class i objects , they may have actually shifted the balance out of favor of true gcs and in favor of galaxies , since a typical gc at the distance of ngc 300 is close to 1@xmath5 in diameter , which is the limit imposed by typical ground - based seeing . of the olsen et al .
candidates , 11 objects appeared to be stars , including ngc 300ax , which olsen et al .
had identified and analyzed as a gc .
its spectrum has wide spectral lines consistent with a dwarf of late k to early m type : a wide but shallow mgb ( 5180 ) feature and a wide and deep nad ( 5895 ) feature , as well as having a strong ca 4227 feature as compared to the g4300 feature much like other k dwarfs .
in addition to this , 3 more gc candidates with ca42-g43 @xmath8 0.05 and ca42-h@xmath6a @xmath8 0.241 have been flagged as possible stars on the basis of radial velocities less than 0 km s@xmath0 .
our gc velocities 9 highly probable clusters and 3 low - velocity possible clusters can be combined with seven previously observed gc velocities , five from @xcite and two from unpublished data contributed by one of us ( k. olsen ) , to obtain a gc velocity dispersion and simple mass estimate for ngc 300 .
the velocities of the seven previously observed clusters are listed in table 2 .
figure 6 shows the velocity histogram of all gc candidates , both previously observed by olsen et al . and
observed in this paper , including those with v@xmath120 .
if we include all highly probable and possible cluster candidates 19 objects in all we have a mean velocity of 79 km s@xmath0 and a velocity dispersion , corrected for the mean uncertainty of the velocities , of 92 km s@xmath0 .
if we exclude the three v@xmath120 objects , we have 16 objects with a mean velocity of 107 km s@xmath0 and an uncertainty - corrected velocity dispersion of 68 km s@xmath0 .
this latter dispersion is closer to the @xcite theoretical velocity dispersion estimate of 60 km s@xmath0 than is the velocity dispersion of all 19 objects .
we can also investigate whether there is evidence for rotation among our gcs .
figure 7 shows the velocity of cluster candidates as a function of position angle , along with a fit to the rotation of all candidates not labeled as possible stars and a representation of the @xcite h rotation curve ( with the velocity of 90 km s@xmath0 reduced by a factor of sin(i ) with i = 42.3 ) .
we fit a function of the form @xmath14 to the velocities and position angles of the 16 cluster candidates with v@xmath150 , presumed to be the best cluster candidates .
@xmath16 is the central velocity , @xmath17 is the projected rotational velocity ( the actual rotational velocity times sin(i ) ) , @xmath18 is the position angle of the cluster in degrees measured east from north , and @xmath19 is the position angle of the rotation axis .
using a nonlinear least - squares fit with weights equal to @xmath20 , fitted to all v@xmath120 clusters , we found @xmath21 km s@xmath0 , @xmath22 km s@xmath0 , and @xmath23 .
however , this rotation curve seems exceptionally strong and out of phase with the h rotation , and the uncertainty - corrected dispersion of the gc velocities minus their predicted rotational velocities is actually considerably worse than our previous result ( 80 km s@xmath0 ) .
an unweighted fit to the same points gives a similarly unsatisfactory result : @xmath24 km s@xmath0 , @xmath25 km s@xmath0 , and @xmath26 , and still a worse velocity dispersion than without rotation ( 81 km s@xmath0 ) .
we therefore conclude that there is no sign of rotation in our gc sample .
we use the projected mass estimator @xcite to estimate the total mass of ngc 300 , both using the means of the total and `` good '' samples and the mean velocity of ngc 300 itself .
the equation for the projected mass estimator is as follows : @xmath27 here we use @xmath28 ( for isotropic orbits , recommended by @xcite and @xmath29 .
using the published mean velocity of ngc 300 itself gives @xmath30 for the total sample , and @xmath31 for the v@xmath150 subsample .
the total sample reaches a maximum projected radius of 12.1 kpc , and the v@xmath150 sample reaches a maximum of 11.6 kpc .
uncertainties were estimated using the bootstrap method with 1000 resamplings .
@xcite , using h rotation curve data , calculate a mass of @xmath32 out to 10.6 kpc ; @xcite give mass - model - based estimates for the mass of ngc 300 ranging from @xmath33 .
overall , our masses run large but are not outside the realm of possibility .
we measured 26 spectral indices , including the 12 indices calibrated in @xcite using the @xcite bandpass definitions and the diagnostic h@xmath6a index with the @xcite bandpass definition .
eight of the brodie and huchra indices are mostly similar to the lick / ids bandpass definitions , differing slightly from those of @xcite and @xcite .
all spectral indices not in @xcite were measured using the bandpass definitions of @xcite .
all bandpasses were shifted to the uncorrected , geocentric radial velocities of our objects .
our indices are not true lick indices because our spectra have not been degraded to match the resolution of the lick star spectra , and we did not observe any lick standard stars to calibrate the fluxes and indices precisely .
however , we did observe a template galaxy ( ngc 1052 ) that was also observed by @xcite , and found the overlapping indices to be similar to theirs .
all indices are given in magnitudes , as in brodie and huchra , for easy comparison with their m31 spectral indices and index - metallicity calibration methods .
converting to equivalent width requires multiplication by the bandwidth of the central index band shifted to the object s velocity .
we determined metallicities using a method based on the @xcite prescription , but with new index - vs .- metallicity calibrations determined from the 41 milky way spectra from @xcite , degraded to match the 5 resolution of our spectra .
we determined linear regressions for metallicity ( from @xcite for each milky way gc ) as a function of index strength for all our indices , as well as the `` range '' @xmath34 defined as in @xcite , and the @xmath35 representing the scatter about the regression line .
we used the @xmath34 and @xmath35 values to construct weights similar to those in @xcite : @xmath36 with @xmath37 being the photometric uncertainty in the entire index , the quadrature sum of the uncertainties in each band . to calculate @xmath38 , the measure of our ability to measure the same index repeatedly for the same object over the course of different nights and conditions , we used spectral index measurements for the standard star ltt 9239 , which we observed each night .
we modified the formula for the uncertainty in a band ( continuum or feature value for an index ) from @xcite to the following : @xmath39 where @xmath40 is the number of object counts per pixel , @xmath41 is the number of sky counts per pixel , and @xmath42 is the read noise .
the weighted average and uncertainty in the weighted average were determined as in @xcite .
our choice of indices to use in the weighted average was determined by considering the correlation coefficient in each milky way linear fit , the signal - to - noise ratio in the ngc 300 objects , and the degree of scatter in ngc 300 index - index relations .
the final set of indices - mg2 , fe5270 , fe5335 , ca4227 , g4300 , @xmath6 , ca4455 , and cn2 - all had relatively high correlation coefficients in the milky way index - metallicity relations and relatively high signal - to - noise in typical ngc 300 clusters .
table 3 lists measurements and uncertainties of the 12 indices calibrated in brodie and huchra and the h@xmath6a index used to help screen out foreground stars , and table 4 gives the measurements and uncertainties of the remaining 13 indices .
table 5 lists the linear relationships between index and metallicity for the 8 indices we chose along with @xmath38 measurements based on the standard star ltt 9239 , and table 6 lists derived metallicities for our clusters , along with five spectroscopically determined metallicities for sculptor group gcs from @xcite .
the highest @xmath38 for spectral features used to measure metallicity was 0.095 ( the @xmath6 index , in the region with the lowest signal - to - noise ratio and spanning a broad range of wavelengths ) and the lowest was 0.001 ( fe5270 , a narrow feature in the region with the highest signal - to - noise ratio ) .
figures 8 - 12 compare various spectral indices to those of andromeda galaxy gcs @xcite and the schiavon et al .
milky way gcs .
figure 8 shows that the cnr index lies below typical values for andromeda and milky way clusters , while the cnb index , traditionally considered a more reliable indicator of actual cn content , looks about the same as the other glaaxies .
figure 9 shows a close - up of the cnb region for averages of 10 milky way gcs and 10 ngc 300 gcs matched for fe52 strength .
the cnb depths look fairly similar , although a tilt due to the high reddening of many of the mw gcs is visible .
fe52 metallicities , as seen in figure 10 , look similar in all three galaxies .
the g - band ( figure 11 ) seems about 0.02 mag high compared to the milky way and andromeda , although overall more similar to the milky way . and
finally , h@xmath11 , the index that is sensitive to age and inversely correlated with metallicity , is shown in figure 12 .
ngc 300 gcs show a mean and distribution more similar to the milky way than to andromeda , suggesting old clusters similar to those of the milky way .
the deficit in cnr in our sample does not necessarily indicate a nitrogen deficiency in the gcs of ngc 300 , especially given the apparently normal cnb .
andromeda gcs had once been thought to have nitrogen excesses compared to milky way gcs @xcite , but this was later challenged when the metal - rich milky way gcs were taken into account @xcite . instead , puzia et al
. found that metal - rich gcs in both andromeda and the milky way were nitrogen - enhanced compared to the general old stellar populations of galaxies , which they propose may be due to low - mass gc stars accreting carbon and/or nitrogen from the agb winds of dying higher - mass stars in the cluster . a deficiency in cnr in our sample ,
therefore , may suggest a high rate of foreground star contamination especially among metal - rich objects , since stars should lack nitrogen compared to metal - rich gcs of similar metallicity . however , cnb is normal , and cnr is a weak feature that can easily be trumped by noise , so there may not be any cause for alarm .
two of the four objects on the lower right of the cnr vs. mg2 plot ( figure 8(b ) ) , with mg2 indices above + 0.125 and lacking strong cnr , are among the low - velocity candidates .
overall , the ngc 300 gcs seem similar to those of the milky way .
figure 13 shows the metallicity distribution of the ngc 300 gcs as compared to the milky way , andromeda , and m33 , including ngc 300r from @xcite . with
all candidates included , the distribution appears unimodal and rather metal - rich . with the three low - velocity objects excluded
, the distribution appears to still be fairly metal - rich .
the ngc 300 cluster sample is too small and possibly too contaminated to definitively determine bimodality or a lack thereof .
an updated version of the @xcite relation between mean gcs metallicity and total galaxy luminosity was created using relatively recent published galaxy distance and spectroscopic gc metallicity information from the literature on several galaxies , along with contemporary 2mass k and @xcite b magnitudes found on the nasa / ipac extragalactic database ( ned ) .
table 7 summarizes the data , gathered from a wide variety of studies of gcss using different selection criteria .
the relationship between the mean gcs metallicity and absolute b magnitude is shown in figure 14 , and the relationship between mean gcs metallicity and absolute 2mass k magnitude is shown in figure 15 .
ngc 300 , with its calculated mean gc metallicity of @xmath43 , lies above both the early - type and late - type galaxies on the plot .
the high metallicity calculated for ngc 300 could be due to most of the `` higher metallicity '' cluster candidates being foreground stars , as suggested above in the discussion of cnr weakness .
if the three clusters deemed to be possible foreground stars are eliminated , the mean metallicity of the remaining ten clusters ( our clusters plus ngc 300r ) is @xmath44 , which is still higher than expected for a galaxy the size of ngc 300 . besides possible foreground star contamination , color selection in the candidate samples
may have created a bias in favor of metal - rich , intrinsically red clusters that drastically increases their relative numbers in our tiny sample .
this is unlikely for the kim et al .
clusters , but the @xcite color cuts were found to be @xmath30.2 mag redder than initially thought due to a calibration error .
it is also worth noting that the spectroscopic metallicity estimates used in our metallicity - luminosity relations do not distinguish among the @xmath45 $ ] values that may differ among galaxy types .
the gcs of ngc 300 appears to have features similar to that of the milky way .
its place on the metallicity - luminosity relation is consistent with the positive correlation between galaxy luminosity and mean gc metallicity , although the average metallicity of our ngc 300 cluster sample is higher than expected . eliminating all ambiguous objects from our cluster sample
gives a mean metallicity slightly more consistent with the milky way and andromeda .
the gcs , excluding those deemed possible stars , have a velocity dispersion of 68 km s@xmath0 , and may be rotating , although the evidence does not strongly favor rotation .
the high mean metallicity of our cluster sample may be primarily a result of the difficulty in screening out stars without the ability to use radial velocity cuts .
the problem of mistaking foreground stars ( and giant stars within the target galaxy ) for gcs could easily be remedied with more space - based imaging of ngc 300 .
gcs could then be identified visually or with more accurate size and shape determinations before spectroscopic study .
we would like to thank the lco staff for helping us use the magellan telescopes and instruments .
this work has been supported by the smithsonian institution and harvard college observatory .
we also thank ricardo schiavon and collaborators for making their milky way spectra available to the public at http://www.noao.edu / ggclib/. ashman , k. m. , and zepf , s. e. 1992 , , 384 , 50 .
barmby , p. , huchra , j. p. , brodie , j. p. , forbes , d. a. , schroder , l. l. , and grillmair , c. j. , 2000 , , 119 , 727 .
beasley , m. a. , bridges , t. , peng , e. , harris , w. e. , harris , g. l. h. , forbes , d. a. , mackie , g. 2008 , , 386 , 1443 .
beasley , m. a. , forbes , d. a. , brodie , j. p. , and kissler - patig , m. 2004 , , 347 , 1150 .
bersier , d. 2000 , , 543 , l23 .
brodie , j. p. , and huchra , j. p. 1990
, , 362 , 503 .
brodie , j. p. , and huchra , j. p. 1991
, , 379 , 157 .
brodie , j. p. , and strader , j. 2006 , , 44 , 193 .
burstein , d. , faber , s. , gaskell , m. , and krumm , n. 1984 , , 287 , 586 .
cappellari , m. et al .
2006 , , 366 , 1126 .
carignan , c. and freeman , k. c. 1985 , , 294 , 494 .
chandar , r. , whitmore , b. , and lee , m. g. 2004 , , 611 , 220 .
chomiuk , l. , strader , j. , and brodie , j. 2008 , , 136 , 234 .
cohen , j. g. , blakeslee , j. p. , and ct ,
, , 592 , 866 .
ct , p. , marzke , r. o. , and west , m. j. 1998 , , 501 , 554 . de vaucouleurs , g. , de vaucouleurs , a. , corwin jr . , h. g. , buta , r. j. , paturel , g. , and fouqu , p. 1991
, third reference catalog of bright galaxies , version 3.9 ( new york : springer - verlag ) ferrarese , l. et al .
2000 , , 529 , 745 .
forbes , d. a. , brodie , j. p. , and grillmair , c. j. 1997 , , 113 , 1652 freedman , w. 1990 , , 355 , l35 freedman , w. l. , hughes , s. m. , madore , b. f. , mould , j. r. , lee , m. g. , stetson , p. , kennicutt , r. c. , turner , a. , ferrarese , l. , ford , h. , graham , j. a. , hill , r. , hoessel , j. g. , huchra , j. , and illingworth , g. d. 1994 , , 427 , 628 .
galleti , s. , bellazzini , m. , and ferraro , f. r. 2004 , a&a , 423 , 925 .
gieren , w. , pietrzynski , g. , walker , a. , bresolin , f. , minniti , d. , kudritzki , r. , udalski , a. , soszinski , i. , fouqu , p. , storm , j. , and bono , g. 2004 , , 128 , 1167 .
harris , w. e. 1996 , , 112 , 1487 .
heisler , j. , tremaine , s. , and bachall , j. n. 1985 , , 298 , 8 .
huchra , j. p. , brodie , j. p. , and kent , s. m. 1991 , , 370 , 495 .
jarrett , t. h. , chester , t. , cutri , r. , schneider , s. e. , and huchra , j. p. 2003 , aj , 125 , 525 .
jensen , j. b. , tonry , j. l. , barris , b. j. , thompson , r. i. , liu , m. c. , rieke , m. j. , ajhar , e. a. , and blakeslee , j. p. 2003
, , 583 , 712 .
kang , a. , kim , j. , shin , i. , chun , s. , kim , h. , and sohn , y. 2007 , jass , 24 , 203 .
kim , s.c . ,
sung , h. , and lee , m.g .
2002 , jkas , 35 , 9 .
kissler - patig , m. , brodie , j. p. , schroder , l. l. , forbes , d. a. , grillmair , c. j. , and huchra , j. p. 1998 , , 115 , 105 .
kundu , a. and whitmore , b. c. 1998 , , 116 , 2841 .
kurtz , m. j. and mink , d. j. 1998 , , 110 , 934 .
larsen , s. s. and brodie , j. p. 2002
, , 123 , 1488 .
larsen , s. s. , brodie , j. p. , beasley , m. a. , and forbes , d. a. 2002 , , 124 , 828 .
larsen , s. s. , brodie , j. p. , huchra , j. p. , forbes , d. a. , and grillmair , c. j. 2001 , , 121 , 2974 .
lee , m. g. , kim , e. , and geisler , d. 1998 , , 115 , 947 .
leitherer , c. et al .
1996 , pasp , 108 , 996 .
mcconnachie a.w .
, irwin m.j . , lewis g.f . ,
ibata , r.a . ,
chapman , s.c .
ferguson a.m.n .
, tanvir n. , 2004 , , 351 , l9 nantais , j. b. , huchra , j. p. , barmby , p. , olsen , k. a. g. , and jerret , t. h. 2006 , aj , 131 , 1416 .
norris , m. a. , sharples , r. m. , bridges , t. , gebhart , k. , forbes , d. a. , proctor , r. , raul faifer , f. , carlos forte , j. , beasley , m. a. , zepf , s. e. , and hanes , d. a. 2008 , , 385 , 40 .
olsen , k. a. g. , miller , b. w. , suntzeff , n. b. , schommer , r. a. , and bright , j. 2004 , , 127 , 2674 .
perelmuter , j. , brodie , j. p. , and huchra , j. p. 1995
, , 110 , 620 .
proctor , r. n. , forbes , d. a. , brodie , j. p. , and strader , j. 2008 , , 385 , 1709 .
puche , d. , carignan , c. , and bosma , a. 1990 , , 100 , 1468 .
puzia , t. h. , saglia , r. p. , kissler - patig , m. , maraston , c. , greggio , l. , renzini , a. , and ortolani , s. 2002 , a&a , 395 , 45 .
rejkuba , m. 2004 , a&a , 413 , 903 .
rhee , m. and chun , m. 1992 , jkas , 25 , 11 .
schiavon , r. p. , rose , j. a. , courteau , s. , and macarthur , l.a .
2005 , apjs , 160 , 163 .
schrder , l. l. , brodie , j. p. , kissler - patig , m. , huchra , j. p. , and phillips , a. c. 2002 , , 123 , 2473 .
sharina , m. e. , afanasiev , v. l. , and puzia , t. h. 2006 , , 372 , 1259 .
sikkema , g. , peletier , r. f. , carter , d. , valentijn , e. a. , and balcells , m. 2006 , a&a , 458 , 53 .
strader , j. , brodie , j. p. , and forbes , d. a. 2004 , , 127 , 295 .
strader , j. , brodie , j. p. , forbes , d. a. , beasley , m. a. , and huchra , j. p. 2003
, , 125 , 1291 .
tonry j.l . ,
dressler a. , blakeslee j.p . , ajhar e. a. , fletcher a.b . , luppino g.a . ,
metzger m.r .
, and moore c.b .
2001 , , 546 , 681 .
trager , s. c. , worthey , g. , faber , s. m. , burstein , d. , and gonzalez , j. j. 1998 , apjs , 116 , 1 .
worthey , g. , faber , s. , gonzalez , j. j. , and burstein , d. 1994 , apjs , 94 , 687 .
zinn , r. 1985 , , 283 , 424 .
ngc 253b & 00:46:50.70 & -25:26:38.90 & & 209 & 43 & 7.87 & m31_a_temp & gc + ngc 253ad & 00:47:31.91 & -25:26:42.53 & & 65 & 40 & 7.42 & fn4486btemp & gc + ngc 300cz & 00:53:26.93 & -37:41:55.54 & & 81 & 19 & 16.6 & fallstars & star + ngc 300cy & 00:53:27.89 & -37:55:21.48 & & 85377 & 24 & 6.73 & hemtemp0.0 & ( em.?)gal + ngc 300dt & 00:53:40.30 & -37:49:34.38 & & 127 & 40 & 8.68 & habtemp90 & gc + ngc 300db & 00:53:55.35 & -37:23:42.98 & & -123 & 46 & 5.95 & m31_f_temp & gc or star + ngc 300ds & 00:53:56.12 & -37:34:49.40 & & 169 & 21 & 13.27 & fabtemp97 & star + ngc 300dd & 00:53:56.47 & -37:58:46.59 & & 11483 & 29 & 4.26 & hemtemp0.0 & gal ? + ngc 300 cm & 00:54:05.51 & -37:42:06.11 & & -18 & 19 & 14.24 & m31_k_temp & gc or star + ngc 300 - 01 & 00:54:22.07 & -37:43:22.1 & 1 & 49374 & 34 & 8.78 & habtemp90 & gal * + ngc 300ck & 00:54:21.84 & -37:32:07.82 & & 21 & 13 & 19.27 & fallstars & gc + ngc 300cj & 00:54:24.13 & -37:25:46.17 & & 23 & 17 & 15.63 & m31_k_temp & gc + ngc 300 - 08 & 00:54:57.41 & -37:33:55.30 & 2 & 16797 & 22 & 14.43 & habtemp90 & gal * + ngc 300 - 02 & 00:54:35.80 & -37:43:05.60 & 2 & 59 & 21 & 15.11 & habtemp90 & gc + ngc 300 - 15 & 00:55:21.81 & -37:45:40.10 & 2 & 106 & 22 & 20.12 & m31_a_temp & star + ngc 300 - 03 & 00:54:40.02 & -37:49:18.71 & 3 & 78 & 28 & 10.04 & m31_f_temp & gc + ngc 300 - 04 & 00:54:45.62 & -37:43:41.00 & 2 & 161 & 28 & 9.14 & m31_f_temp & gc + ngc 300 - 14 & 00:55:17.67 & -37:36:57.90 & 2 & 16995 & 36 & 8.68 & fallstars & gal + ngc 300 - 05 & 00:54:49.35 & -37:43:30.50 & 2 & 120 & 35 & 8.08 & m31_f_temp & gc * + ngc 300 - 09 & 00:54:57.57 & -37:36:37.90 & 2 & 214 & 14 & 18.08 & fglotemp & star * + ngc 300 - 11 & 00:55:01.44 & -37:37:21.80 & 1 & 38350 & 17 & 8.83 & hemtemp0.0 & em.gal * + ngc 300cl & 00:54:53.85 & -37:23:57.44 & & 21 & 16 & 17.4 & fabtemp97 & star + ngc 300dc & 00:54:54.66 & -37:58:34.98 & & 68066 & 25 & 5.13 & hemtemp0.0 & em.gal + ngc 300df & 00:54:57.02 & -37:58:52.21 & & -16 & 48 & 5.76 & m31_k_temp & star + ngc 300 - 10 & 00:54:59.03 & -37:36:24.40 & 3 & 173 & 23 & 12.59 & m31_f_temp & star * + ngc 300ax & 00:55:01.49 & -37:47:36.40 & & -27 & 32 & 8.43 & m31_k_temp & star + ngc 300 - 07 & 00:54:53.22 & -37:43:11.1 & 1 & 43722 & 27 & 6.41 & hemtemp0.0 & em.gal * + ngc 300 - 12 & 00:55:11.03 & -37:36:46.10 & 1 & 177 & 18 & 15.19 & m31_f_temp & gc + ngc 300co & 00:55:19.49 & -37:29:38.73 & & -66 & 19 & 14.57 & m31_k_temp & gc or star + ngc 300 - 06 & 00:54:51.67 & -37:39:38.00 & 3 & 156 & 14 & 8.97 & hemtemp0.0 & hii * + ngc 300 - 17 & 00:55:38.79 & -37:46:30.70 & 3 & 16868 & 21 & 13.02 & fallstars & gal + ngc 300 - 16 & 00:55:27.56 & -37:46:09.40 & 3 & 16798 & 29 & 8.03 & m31_k_temp & gal + ngc 300 - 13 & 00:55:11.17 & -37:33:13.90 & 3 & 38182 & 20 & 6.19 & hemtemp0.0 & em.gal + ngc 300cr & 00:55:52.72 & -37:59:04.14 & & -23 & 15 & 17.05 & fallstars & star + ngc 300bd & 00:55:56.79 & -37:52:07.28 & & 68818 & 20 & 6.69 & hemtemp0.0 & gal + ngc 300cq & 00:56:07.09 & -37:57:02.64 & & -79 & 14 & 17.25 & fallstars & star + ngc 300cs & 00:56:07.62 & -37:57:23.07 & & 31 & 16 & 15.85 & fallstars & star + ngc 300da & 00:56:09.49 & -37:31:21.25 & & 49619 & 21 & 7.51 & hemtemp0.0 & em.gal + ngc 300cv & 00:56:13.97 & -37:34:16.27 & & 97 & 17 & 12.33 & fabtemp97 & star + ngc 300ct & 00:56:18.24 & -37:28:48.08 & & 4 & 15 & 17.59 & fabtemp97 & star + ngc 300cp & 00:56:19.03 & -37:29:17.23 & & 35 & 19 & 13.88 & m31_k_temp & gc + ngc 300cx & 00:56:20.30 & -37:59:11.68 & & 68618 & 53 & 5.97 & m31_f_temp & gal + ngc 300cn & 00:56:23.52 & -37:27:37.57 & & 49828 & 17 & 8.89 & hemtemp0.0 & em.gal + ngc 300cu & 00:56:24.62 & -37:52:56.29 & & 18 & 20 & 12.33 & fallstars & star + ngc 300ac & 00:54:02.48 & -37:44:31.64 & 82 & 45 & gc & olsen et al .
2004 + ngc 300ag & 00:54:23.22 & -37:59:17.67 & 120 & 38 & gc & olsen et al .
2004 + ngc 300am & 00:54:07.54 & -37:42:09.23 & 73 & 17 & gc & olsen ( unpublished ) + ngc 300ba & 00:54:05.49 & -37:23:11.24 & 118 & 60 & gc & olsen et al .
2004 + ngc 300 m & 00:54:07.04 & -37:41:10.35 & 18 & 13 & gc & olsen ( unpublished ) + ngc 300r & 00:56:23.13 & -37:33:26.98 & 248 & 15 & gc & olsen et al .
2004 + ngc 300s & 00:53:55.66 & -37:32:37.38 & 249 & 29 & gc & olsen et al .
2004 + ngc 253ad & 0.017 & 0.262 & 0.039 & 0.048 & 0.177 & 0.152 & 0.027 & 0.046 & 0.294 & 0.312 & -0.086 & 0.477 & -0.83 + @xmath7 & 0.060 & 0.071 & 0.037 & 0.022 & 0.025 & 0.043 & 0.035 & 0.041 & 0.120 & 0.121 & 0.069 & 0.031 & 0.066 + ngc 253b & -0.132 & 0.037 & 0.058 & 0.009 & 0.017 & 0.005 & -0.005 & 0.049 & 0.216 & 0.234 & -0.088 & 0.252 & 0.101 + @xmath7 & 0.042 & 0.053 & 0.033 & 0.020 & 0.022 & 0.037 & 0.033 & 0.042 & 0.071 & 0.067 & 0.050 & 0.019 & 0.045 + ngc 300 - 02 & -0.044 & 0.144 & 0.025 & 0.001 & 0.120 & 0.110 & 0.046 & 0.039 & 0.033 & 0.216 & 0.001 & 0.244 & 0.001 + @xmath7 & 0.011 & 0.015 & 0.010 & 0.006 & 0.007 & 0.012 & 0.010 & 0.013 & 0.020 & 0.019 & 0.015 & 0.005 & 0.012 + ngc 300 - 03 & -0.045 & 0.103 & 0.079 & -0.008 & 0.018 & 0.045 & -0.007 & 0.035 & 0.128 & 0.312 & -0.188 & 0.298 & 0.048 + @xmath7 & 0.026 & 0.031 & 0.020 & 0.013 & 0.014 & 0.023 & 0.021 & 0.029 & 0.047 & 0.044 & 0.035 & 0.012 & 0.028 + ngc 300 - 04 & -0.015 & 0.120 & 0.063 & 0.010 & 0.037 & 0.064 & 0.066 & 0.008 & 0.095 & 0.328 & -0.165 & 0.354 & 0.013 + @xmath7 & 0.031 & 0.037 & 0.023 & 0.014 & 0.015 & 0.026 & 0.023 & 0.030 & 0.058 & 0.057 & 0.040 & 0.015 & 0.034 + ngc 300 - 05 & -0.051 & 0.031 & 0.045 & 0.007 & 0.062 & 0.086 & 0.039 & -0.047 & -0.057 & 0.184 & -0.134 & 0.243 & 0.090 + @xmath7 & 0.036 & 0.044 & 0.029 & 0.017 & 0.019 & 0.033 & 0.029 & 0.043 & 0.068 & 0.062 & 0.051 & 0.016 & 0.038 + ngc 300 - 12 & -0.053 & 0.138 & 0.071 & 0.030 & 0.101 & 0.088 & 0.056 & 0.026 & 0.069 & 0.362 & -0.245 & 0.446 & 0.024 + @xmath7 & 0.018 & 0.023 & 0.013 & 0.008 & 0.009 & 0.015 & 0.013 & 0.016 & 0.036 & 0.034 & 0.026 & 0.009 & 0.020 + ngc 300cj & -0.058 & 0.225 & 0.054 & 0.012 & 0.136 & 0.136 & 0.051 & 0.049 & 0.175 & 0.399 & -0.187 & 0.536 & -0.043 + @xmath7 & 0.014 & 0.018 & 0.010 & 0.006 & 0.007 & 0.012 & 0.011 & 0.012 & 0.026 & 0.025 & 0.018 & 0.007 & 0.014 + ngc 300ck & -0.029 & 0.231 & 0.060 & 0.030 & 0.203 & 0.180 & 0.084 & 0.070 & 0.317 & 0.378 & -0.200 & 0.664 & -0.62 + @xmath7 & 0.016 & 0.021 & 0.012 & 0.007 & 0.008 & 0.014 & 0.012 & 0.013 & 0.032 & 0.029 & 0.023 & 0.008 & 0.017 + ngc 300 cm & -0.052 & 0.182 & 0.072 & 0.022 & 0.157 & 0.160 & 0.059 & 0.036 & 0.203 & 0.350 & -0.108 & 0.483 & -0.023 + @xmath7 & 0.018 & 0.023 & 0.014 & 0.008 & 0.009 & 0.016 & 0.014 & 0.016 & 0.034 & 0.033 & 0.023 & 0.009 & 0.019 + ngc 300co & 0.013 & 0.145 & 0.031 & 0.013 & 0.125 & 0.121 & 0.059 & 0.060 & 0.300 & 0.434 & -0.181 & 0.561 & -0.049 + @xmath7 & 0.019 & 0.024 & 0.014 & 0.009 & 0.010 & 0.017 & 0.014 & 0.016 & 0.037 & 0.035 & 0.025 & 0.009 & 0.020 + ngc 300cp & -0.047 & 0.202 & 0.090 & -0.001 & 0.152 & 0.166 & 0.071 & 0.027 & 0.142 & 0.375 & -0.082 & 0.460 & 0.005 + @xmath7 & 0.021 & 0.028 & 0.017 & 0.010 & 0.011 & 0.019 & 0.017 & 0.020 & 0.040 & 0.039 & 0.028 & 0.010 & 0.022 + ngc 300db & -0.071 & 0.262 & 0.092 & 0.002 & 0.078 & 0.114 & -0.007 & -0.026 & -0.010 & 0.268 & -0.022 & 0.226 & 0.002 + @xmath7 & 0.043 & 0.055 & 0.033 & 0.020 & 0.022 & 0.037 & 0.033 & 0.042 & 0.079 & 0.076 & 0.056 & 0.019 & 0.046 + ngc 300dt & -0.006 & 0.200 & 0.059 & 0.008 & 0.117 & 0.177 & 0.037 & 0.087 & 0.065 & 0.268 & -0.024 & 0.269 & -0.052 + @xmath7 & 0.034 & 0.044 & 0.023 & 0.013 & 0.015 & 0.026 & 0.022 & 0.028 & 0.069 & 0.065 & 0.046 & 0.017 & 0.037 + @xmath12ngc 300@xmath15 & -0.038 & 0.165 & 0.062 & 0.011 & 0.109 & 0.120 & 0.046 & 0.030 & 0.122 & 0.323 & -0.128 & 0.399 & -0.004 + @xmath7 & 0.007 & 0.009 & 0.005 & 0.003 & 0.004 & 0.006 & 0.005 & 0.007 & 0.013 & 0.012 & 0.009 & 0.003 & 0.026 + ngc 253ad & 0.041 & 0.102 & 0.088 & 0.048 & 0.056 & 0.060 & 0.052 & 0.036 & 0.036 & 0.058 & -0.014 & 0.008 & 0.028 + @xmath7 & 0.076 & 0.106 & 0.066 & 0.067 & 0.055 & 0.042 & 0.037 & 0.044 & 0.045 & 0.038 & 0.045 & 0.024 & 0.023 + ngc 253b & -0.104 & -0.048 & 0.001 & 0.049 & 0.075 & 0.067 & 0.001 & -0.029 & -0.029 & 0.001 & -0.021 & -0.010 & 0.008 + @xmath7 & 0.055 & 0.076 & 0.052 & 0.053 & 0.046 & 0.036 & 0.033 & 0.042 & 0.042 & 0.037 & 0.043 & 0.024 & 0.023 + ngc 300 - 02 & -0.030 & 0.066 & 0.067 & 0.032 & 0.046 & 0.013 & 0.032 & 0.047 & 0.047 & 0.018 & 0.022 & 0.009 & 0.012 + @xmath7 & 0.015 & 0.020 & 0.014 & 0.015 & 0.013 & 0.011 & 0.010 & 0.013 & 0.014 & 0.012 & 0.015 & 0.008 & 0.008 + ngc 300 - 03 & -0.021 & 0.072 & -0.020 & 0.023 & 0.081 & 0.024 & 0.046 & -0.008 & -0.008 & 0.019 & -0.019 & -0.032 & 0.007 + @xmath7 & 0.034 & 0.045 & 0.032 & 0.032 & 0.028 & 0.022 & 0.021 & 0.027 & 0.027 & 0.024 & 0.029 & 0.016 & 0.015 + ngc 300 - 04 & -0.026 & 0.072 & 0.060 & 0.028 & 0.069 & 0.044 & 0.065 & 0.052 & 0.052 & 0.038 & 0.024 & 0.007 & 0.011 + @xmath7 & 0.042 & 0.055 & 0.038 & 0.038 & 0.032 & 0.025 & 0.023 & 0.028 & 0.029 & 0.027 & 0.032 & 0.018 & 0.017 + ngc 300 - 05 & -0.052 & 0.081 & -0.004 & -0.045 & 0.003 & 0.012 & -0.017 & 0.031 & 0.031 & -0.025 & -0.047 & 0.005 & 0.016 + @xmath7 & 0.049 & 0.062 & 0.045 & 0.047 & 0.040 & 0.031 & 0.030 & 0.036 & 0.037 & 0.036 & 0.043 & 0.024 & 0.023 + ngc 300 - 12 & -0.023 & 0.102 & 0.064 & 0.030 & 0.069 & 0.027 & 0.036 & 0.040 & 0.040 & 0.023 & 0.022 & 0.042 & 0.064 + @xmath7 & 0.024 & 0.032 & 0.021 & 0.023 & 0.018 & 0.015 & 0.014 & 0.017 & 0.017 & 0.015 & 0.018 & 0.010 & 0.010 + ngc 300cj & -0.036 & 0.069 & 0.096 & 0.050 & 0.067 & 0.039 & 0.050 & 0.041 & 0.041 & 0.027 & 0.013 & -0.009 & 0.009 + @xmath7 & 0.018 & 0.024 & 0.017 & 0.017 & 0.014 & 0.012 & 0.011 & 0.013 & 0.014 & 0.012 & 0.014 & 0.008 & 0.008 + ngc 300ck & 0.011 & 0.147 & 0.120 & 0.069 & 0.072 & 0.050 & 0.068 & 0.077 & 0.077 & 0.016 & 0.030 & 0.001 & 0.017 + @xmath7 & 0.020 & 0.028 & 0.018 & 0.019 & 0.016 & 0.013 & 0.012 & 0.015 & 0.016 & 0.013 & 0.016 & 0.009 & 0.009 + ngc 300 cm & -0.026 & 0.144 & 0.083 & 0.071 & 0.053 & 0.004 & 0.026 & 0.052 & 0.052 & 0.001 & 0.030 & 0.020 & 0.005 + @xmath7 & 0.023 & 0.032 & 0.021 & 0.022 & 0.019 & 0.015 & 0.014 & 0.017 & 0.018 & 0.016 & 0.019 & 0.010 & 0.010 + ngc 300co & 0.039 & 0.124 & 0.039 & 0.041 & 0.049 & 0.012 & 0.044 & 0.039 & 0.039 & 0.023 & 0.019 & -0.001 & 0.008 + @xmath7 & 0.024 & 0.034 & 0.022 & 0.024 & 0.020 & 0.016 & 0.015 & 0.018 & 0.019 & 0.016 & 0.019 & 0.010 & 0.010 + ngc 300cp & -0.009 & 0.110 & 0.039 & 0.045 & 0.055 & 0.038 & 0.037 & 0.044 & 0.044 & 0.001 & 0.023 & -0.008 & 0.015 + @xmath7 & 0.027 & 0.038 & 0.026 & 0.027 & 0.022 & 0.018 & 0.017 & 0.021 & 0.022 & 0.019 & 0.023 & 0.012 & 0.012 + ngc 300db & -0.043 & 0.091 & -0.023 & -0.004 & -0.061 & 0.016 & 0.022 & 0.043 & 0.043 & -0.025 & -0.080 & -0.058 & -0.021 + @xmath7 & 0.056 & 0.074 & 0.055 & 0.057 & 0.048 & 0.037 & 0.034 & 0.041 & 0.042 & 0.038 & 0.046 & 0.024 & 0.025 + ngc 300dt & 0.000 & 0.131 & 0.134 & 0.088 & 0.063 & -0.006 & 0.037 & 0.009 & 0.009 & 0.025 & -0.048 & -0.021 & 0.002 + @xmath7 & 0.044 & 0.056 & 0.036 & 0.037 & 0.032 & 0.025 & 0.022 & 0.028 & 0.028 & 0.025 & 0.030 & 0.016 & 0.016 + @xmath12ngc 300@xmath15 & -0.018 & 0.101 & 0.055 & 0.036 & 0.047 & 0.023 & 0.037 & 0.039 & 0.037 & 0.012 & -0.001 & -0.004 & 0.012 + @xmath7 & 0.009 & 0.012 & 0.008 & 0.009 & 0.007 & 0.006 & 0.005 & 0.007 & 0.007 & 0.006 & 0.007 & 0.004 & 0.004 + ca4455 & 25.85 & -2.00 & 0.899 & 0.0387 & 0.006 & 0.003 + fe5270 & 26.50 & -2.31 & 0.888 & 0.0377 & 0.006 & 0.001 + mg2 & 9.83 & -1.89 & 0.886 & 0.1017 & 0.017 & 0.006 + @xmath6 & 4.12 & -2.68 & 0.884 & 0.2428 & 0.041 & 0.095 + fe5335 & 30.36 & -2.22 & 0.879 & 0.0329 & 0.006 & 0.005 + ca4227 & 18.65 & -2.09 & 0.878 & 0.0536 & 0.010 & 0.014 + cn2 & 9.45 & -1.27 & 0.852 & 0.1058 & 0.020 & 0.021 + g4300 & 10.37 & -2.24 & 0.844 & 0.0964 & 0.019 & 0.008 + ngc 55az & -1.76 & 0.43 & ols & 27 + ngc 253q & -1.07 & 0.30 & ols & 5 + ngc 253b & -1.96 & 0.54 & new & 26 + ngc 247a & -1.04 & 0.29 & ols & 64 + ngc 253a & -0.88 & 0.25 & ols & 46 + ngc 253ad & -0.52 & 0.56 & new & 57 + ngc 300dt & -0.89 & 0.66 & new & 13 + ngc 300db & -1.23 & 0.81 & new & 22 + ngc 300 cm & -0.54 & 0.36 & new & 33 + ngc 300ck & -0.07 & 0.25 & new & 42 + ngc 300cj & -0.71 & 0.35 & new & 45 + ngc 300 - 02 & -1.03 & 0.29 & new & n / a + ngc 300 - 03 & -1.61 & 0.47 & new & 48 + ngc 300 - 04 & -1.18 & 0.33 & new & n / a + ngc 300 - 05 & -1.57 & 0.52 & new & n / a + ngc 300 - 12 & -0.93 & 0.23 & new & n / a + ngc 300co & -0.67 & 0.29 & new & 69 + ngc 300cp & -0.59 & 0.29 & new & 116 + ngc 300r & -1.25 & 0.35 & ols & 121 + fornax & -1.71 & 0.08 & 9.28 & 0.136 & -11.38 & & & 32 & 5 & + m87 & -0.89 & 0.16 & 9.56 & 15.631 & -21.38 & 5.81 & -25.16 & 6 & 11 & 17 + m31 & -1.22 & 0.04 & 4.36 & 0.77 & -20.07 & 0.98 & -23.45 & 2 & 12 & 17 + mw & -1.31 & 0.05 & 5.52 & 0.76 & -18.9 & & & 16 & & + m81 & -1.19 & 0.14 & 7.89 & 3.6 & -19.89 & 3.83 & -23.95 & 29 , 26 & 13 & 17 + m33 & -1.55 & 0.37 & 6.27 & 0.847 & -18.37 & 4.1 & -20.54 & 6 & 14 & 17 + m49 & -0.56 & 0.3 & 9.37 & 16.368 & -21.7 & 5.4 & -25.67 & 9 & 11 & 17 + n 147 & -2.05 & 0.4 & 10.47 & 0.594 & -13.4 & 7.2 & -16.67 & 6 & 19 & 17 + n 205 & -1.40 & 0.15 & 8.92 & 0.752 & -15.46 & 5.59 & -18.79 & 6 & 30 & 17 + n 185 & -1.65 & 0.25 & 10.1 & 0.676 & -14.05 & 6.56 & -17.59 & 6 & 23 & 17 + m104 & -1.01 & 0.13 & 8.98 & 9.333 & -20.87 & 4.96 & -24.89 & 22 & 11 & 17 + n 5128 & -0.96 & 0.04 & 7.84 & 3.8 & -20.06 & 3.94 & -23.96 & 3 & 28 & 17 + n
2683 & -1.23 & 0.19 & 10.64 & 7.178 & -18.64 & 6.33 & -22.95 & 27 & 33 , 18 & 17 + n 7457 & -0.97 & 0.11 & 12.09 & 13.996 & -18.64 & 8.19 & -22.54 & 8 & 11 , 7 & 1 + n 1023 & -1.29 & 0.11 & 10.35 & 11.117 & -19.88 & 6.24 & -23.99 & 21 & 11 & 17 + n 1399 & -0.82 & 0.12 & 10 & 19.055 & -21.4 & 6.31 & -25.09 & 20 & 11 & 17 + n 3923 & -0.94 & 0.11 & 10.8 & 22.909 & -21.0 & 6.5 & -25.3 & 24 & 31 & 17 + n 524 & -0.97 & 0.12 & 11.3 & 28.2 & -20.95 & 7.16 & -25.09 & 4 & 10 & 1 + n 300 & -0.94 & 0.12 & 8.95 & 1.93 & -17.48 & 6.38 & -20.05 & 25 , this paper & 15 & 17 + | we present new metallicity estimates for globular cluster ( gc ) candidates in the sd spiral ngc 300 , one of the nearest spiral galaxies outside the local group .
we have obtained optical spectroscopy for 44 sculptor group gc candidates with the boller and chivens ( b&c ) spectrograph on the baade telescope at las campanas observatory .
there are 2 gcs in ngc 253 and 12 objects in ngc 300 with globular - cluster - like spectral features , 9 of which have radial velocities above 0 km s@xmath0 . the remaining three , due to their radial velocities being below the expected 95% confidence limit for velocities of ngc 300 halo objects , are flagged as possible foreground stars .
the non - clusterlike candidates included 13 stars , 15 galaxies , and an h region .
one gc , four galaxies , two stars , and the h region from our sample were identified in archival hubble space telescope images . for the gcs
, we measure spectral indices and estimate metallicities using an empirical calibration based on milky way gcs .
the gcs of ngc 300 appear similar to those of the milky way . excluding possible stars and including clusters from the literature
, the gc system ( gcs ) has a velocity dispersion of 68 km s@xmath0 , and has no clear evidence of rotation .
the mean metallicity for our full cluster sample plus one literature object is [ fe / h ] = @xmath1 , lying above the relationship between mean gc metallicity and overall galaxy luminosity . excluding the three low - velocity candidates , we obtain a mean [ fe / h ] = @xmath2 , still higher than expected , raising the possibility of significant foreground star contamination even in this sample .
visual confirmation of genuine gcs using high - resolution space - based imagery could greatly reduce the potential problem of interlopers in small samples of gcss in low - radial - velocity galaxies . |
according to the currently popular @xmath4cdm model of structure formation , galaxies should form hierarchically .
numerical simulations demonstrate that galaxy - sized objects form through numerous interactions and mergers ( e.g. toomre & toomre 1972 ; hernquist 1992 , 1993 ; naab & burkert 2003 ; robertson et al .
2006 ; maller et al .
. therefore , the outcomes of galaxy - galaxy interactions and mergers are of key importance in understanding the hierarchical picture of galaxy formation . however , the impact of interactions and mergers on galaxy properties is still not known well even though there have been many studies .
one of the earliest works on this problem is larson & tinsley ( 1978 ) who showed that galaxies selected from the atlas of peculiar galaxies ( arp 1966 ) have enhanced star formation ( sf ) rate compared to typical galaxies . since that time
, many studies have shown observational evidence of significant changes in galaxy properties due to interactions and/or mergers , for instance , in sf activity ( kennicutt & keel 1984 ; kennicutt et al .
1987 ; bushouse , werner & lamb 1988 ; barton et al .
2000 , 2003 , 2007 ; lambas et al .
2003 ; sanchez & gonzalez - serrano 2003 ; nikolic , cullen & alexander 2004 ; hernandez - toledo et al .
2005 ; geller et al .
2006 ; li et al .
2008 ) , galaxy structure parameters ( nikolic et al .
2004 ; patton et al . 2005 ; hernandez - toledo et al . 2005 ; coziol & plauchu - frayn 2007 ; kacprzak et al .
2007 ) , luminosity ratio ( woods et al . 2006
; woods & geller 2007 ) , and stellar mass ratio ( ellison et al . 2008 ) .
recent @xmath5-body simulations have also shown that the general features of the observations can be modeled ( e.g. barnes & hernquist 1996 ; mihos & hernquist 1996 ; tissera et al . 2002 ; perez et al .
2006a , b ; di matteo et al .
2007 ) .
however , previous studies have not always been in agreement with one another .
yee & ellingson ( 1995 ) and patton et al . ( 1997 ) found no significant difference between the mean properties of isolated and paired galaxies .
bergvall et al . ( 2003 ) found no clear difference between isolated galaxies and interacting / merging systems in their global sf rate in optical / near - ir bands .
allam et al . (
2004 ) identified a set of merging galaxies in the sdss data and found only a weak positive correlation in color for the merging pairs .
hernandez - toledo et al .
( 2005 ) analyzed the @xmath6 images of 42 elliptical / lenticular galaxies , and claimed that the structural effects of interactions on e / s0s are minor , in contrast to disk galaxies involved in interactions .
de propris et al .
( 2005 ) analyzed galaxies in the millennium galaxy catalog and found merging galaxies are only marginally bluer than noninteracting galaxies , showing an excess of both early and late types but a deficiency of intermediate type spirals .
smith et al .
( 2007 ) did not find any enhancement in spitzer mid - infrared color depending on pair separation by using pre - merger interacting galaxy pairs selected from the arp atlas .
in addition , in terms of the degree and scale of the effects of interactions on galaxy properties , the results also have not always agreed .
lambas et al .
( 2003 ) studied the galaxy pairs in the two degree field galaxy redshift survey ( 2dfgrs ) data and reported that sf in galaxy pairs is significantly enhanced over that of isolated galaxies only when the projected separation @xmath7 kpc and radial velocity difference @xmath8 km s@xmath9 .
nikolic et al . (
2004 ) reported using the sdss data that the mean sf rate is significantly enhanced for galaxy pairs with @xmath10 kpc .
but for late types , the enhancement extended out to 300 kpc .
they also noted that the sf rate slightly decreased with increasing @xmath11 , and the light concentration was lowest at @xmath12kpc and then increased rapidly inward .
alonso et al .
( 2006 ) measured sf rate of the galaxies in the 2dfgrs and sdss data and found that the sf rate was strongly enhanced when @xmath13 kpc and @xmath14 km s@xmath9 , which was more effective in low and intermediate density environments .
these discrepancies as listed above are expected mainly because the effects of interactions / mergers between different types of galaxies on their final products are different .
woods & geller ( 2007 ) found that for blue - blue major pair sample in the sdss , there exists clear correlation between specific sfr and pair separation , and for red - red pair , there is none ( cf .
tran et al . 2005 ; van dokkum 2005 ; bell et al . 2006 ) .
( 2008 ) found that for the most strongly star - forming systems , tidal interactions are the dominant trigger of enhanced star formation and the enhancement is a strong function of separation less than 100 kpc .
in addition to such small - scale environment , the large - scale environment has been known to be one of the determining factors of galaxy properties .
it is now well - known that galaxy properties correlate with the large - scale background density at low redshift ( hogg et al .
2003 ; goto et al . 2003 ; balogh et al .
2004a , b ; tanaka et al . 2004 ; kauffmann et al . 2004 ; blanton et al . 2005a ; croton
2005 ; weinmann et al . 2006 ; park et al . 2007 ) .
deep redshift surveys have extended these studies to high redshift ( cucciati et al .
2006 ; elbaz et al .
2007 ; cooper et al . 2007 , 2008 ; poggianti et al . 2008 ) . a number of papers ( kauffmann et al . 2004 ; blanton et al . 2005a ; quintero et al .
2006 ; ball , loveday , & brunner 2008 ) claimed from an analysis of the sdss data that structural properties of galaxies are less closely related to the ( large - scale ) environment than are their masses and sf related parameters such as color and sf rate .
however , the structural parameters such as concentration and sersic index are not true measure of morphology ( bamford et al . 2008 ) and indeed , van der wel ( 2008 ) pointed that morphology and structure are intrinsically different galaxy properties and structure mainly depends on galaxy mass whereas morphology mainly depends on environment .
several papers tried to use sophisticated automated morphology classifications for the study of the relationship between morphology and environment ( goto et al .
2003 ; park & choi 2005 ; allen et al .
2006 ; ball et al .
many studies reported that the sf rate of galaxies is a strongly decreasing function of the large - scale density and that there is a critical density for sf activity ( gomez et al .
2003 ; tanaka et al .
2004 ) .
park et al .
( 2007 ) , however , found that this trend was mostly because morphology and luminosity are strong functions of the large - scale density .
they made an extensive study of the environmental dependence of various physical properties of galaxies on large- and small - scale densities , and concluded that morphology and luminosity are major fundamental parameters .
once morphology and luminosity are fixed , other galaxy properties are almost independent of the large - scale density .
the large - scale density dependence of these properties reported by many previous and current studies merely reflects the correlations of the properties with morphology and luminosity rather than independent correlations with environment .
park et al . ( 2007 ) also found that galaxy morphology changes sensitively across the nearest neighbor distance of a few hundred kpc .
this work has been extended by park , gott , & choi ( 2008 ) who studied galaxy morphology as a function of the nearest neighbor separation at fixed large - scale background density and found that this characteristic scale corresponds to the virial radius of the neighbor galaxy .
park et al . ( 2008 ) s findings and claims can be summarized as follows .
1 . the effects of galaxy interactions reach farther than the distance previously thought .
they reach at least out to the galaxy virial radius , namely a few hundred kpc for bright galaxies .
the result of galaxy interactions can be very different depending on the morphological type of the nearest neighbor galaxy when the pair separation is less than the virial radius . without separating the neighbor galaxies into different morphological types
, one will find the effects of galaxy interactions are diverse but negligible on average .
the dependence on neighbor s morphology disappears at separations farther than the virial radius .
the fact that , at fixed large - scale density , the morphology of a galaxy is more likely to be a late type as it approaches a late - type neighbor , suggests that galaxies can transform their morphology from early types to late types through close encounters with cold gas - rich neighbors . \4 . in most places of the universe , except for the regions within massive clusters of galaxies , the well - known morphology - density relation originated largely due to the effects of galaxy - galaxy interactions .
galaxy morphology appears to depend on the large - scale density mainly because the mean separation between galaxies is statistically correlated with the large - scale density .
5 . a series of close interactions and mergers transform galaxy morphology and luminosity classes in such a way that galaxies on average evolve to become bright early types .
the transformation speed depends on the large - scale density .
the fourth result is supplemented by the recent work of park & hwang ( 2008 ) who found , in the special case of galaxies located within the virial radius of massive clusters , galaxy properties depend on both the distance to the nearest neighbor and clustercentric radius .
the purpose of this paper is to extend park et al . (
2008 ) s work by looking at the dependence of various other properties of galaxies , as well as morphology and luminosity , on small and large scale environmental factors .
we use a volume - limited sample of the sdss galaxies whose morphology is accurately classified .
the environment is specified by the small - scale ( distance to the nearest neighbor galaxy and morphology of the nearest neighbor galaxy ) and large - scale ( the background density ) factors .
it is hoped that the effects of galaxy interactions can be understood in fuller detail in this three - dimensional environmental parameter space .
our observational sample is one of the subsamples , d3 , used by choi et al .
( 2007 ) and park et al .
it is a volume - limited sample of galaxies extracted from a large - scale structure sample , dr4plus ( lss - dr4plus ) , of the sdss data ( york et al .
2000 ) from the new york university value - added galaxy catalog ( nyu - vagc ; blanton et al .
this sample is a subset of the spectroscopic main galaxy sample of the sdss data release 5 ( adelman - mccarthy et al .
2007 ) .
the sample d3 , together with other volume - limited samples , has been described in great detail by choi et al . to summarize
it is a sample of galaxies with the @xmath15-band absolute magnitude @xmath16 ( hereafter we drop the @xmath17 term in the absolute magnitude ) and redshifts @xmath18 or comoving distance of @xmath19mpc @xmath20mpc .
the sdss spectroscopic sample has a bright apparent magnitude limit of @xmath21 , but our sample is supplemented by brighter galaxies whose redshifts are obtained from various literature .
d3 includes 49,571 galaxies .
the rest - frame absolute magnitudes of individual galaxies are computed in fixed bandpasses , shifted to @xmath22 , using galactic reddening correction ( schlegel et al .
1998 ) and @xmath23-corrections as described by blanton et al .
the mean evolution correction given by tegmark et al .
( 2004 ) , @xmath24 , is also applied .
we adopt a flat @xmath4cdm cosmology with @xmath25 and @xmath26 .
the useful survey area of this sample , having nonzero angular selection function , is 1.362 sr .
all galaxies in d3 are plotted in figure 1 .
accurate morphology classification is critical in this work since the effects of interaction depend strongly on morphology of the target and neighbor galaxies .
we first classify morphological types of galaxies using the prescription of park & choi ( 2005 ) .
galaxies are divided into early ( ellipticals and lenticulars ) and late ( spirals and irregulars ) morphological types based on their locations in the @xmath27 color versus @xmath28 color gradient space and also in the @xmath29-band concentration index space .
the resulting morphological classification has completeness and reliability reaching 90% .
our automatic classification scheme does not perform well when an early - type galaxy starts to overlap with other galaxy .
this is because the scheme excludes galaxies with very low concentration from the early - type class and blended images often erroneously give low concentration .
since we are investigating the effects of close interaction on galaxy properties , this problem in the automatic classification has to be remedied .
we perform an additional visual check of the color images of galaxies to correct misclassifications by the automated scheme for about @xmath30 galaxies having close neighbors . in this procedure
we changed the types of the blended or merging galaxies , blue but elliptical - shaped galaxies , and dusty edge - on spirals .
some non - sense objects like blank fields and substructures of large galaxies , are removed from the samples , and some wrong central positions of merging galaxies are corrected .
after all these procedures our final sample is composed of 19,248 early - type galaxies , and 30,283 late - type galaxies with @xmath31 .
our main target galaxies for which we study the dependence of various physical parameters on environment , are those with @xmath32 .
there are 9,434 early types and 14,270 late types satisfying this condition within the sample volume .
this subset is marked by a rectangular box in figure 1 . in our analysis
we often limit the late - type galaxy sample to those with isophotal axis ratio @xmath33 greater than 0.6 .
this is to reduce the effects of internal extinction on our results .
the absolute magnitude and color of late - type galaxies with @xmath34 are very inaccurate ( see fig . 5 and 12 of choi et al .
2007 ) , and including them in the analysis introduces a large dispersion in luminosity of the volume - limited sample . since it is essential to fix luminosity in many of our analyses , it is very important to reduce the internal extinction effects by using nearly face - on late - type galaxies .
when we calculate the median values of galaxy parameters in section 3 , we will take into account the fact that only a subset of late - type galaxies are being used .
there are 8,344 late types with @xmath35 and @xmath32 in our sample .
we also often divide our sample into four subsamples : early types having early - type nearest neighbor ( the e - e galaxies ) , early types having late - type nearest neighbor ( e - l ) , late types having early - type nearest neighbor ( l - e ) , and late types having late - type nearest neighbor ( l - l ) .
there are 4675 , 4759 , 3423 , and 4921 galaxies in these subsets , respectively , when only those with @xmath36 are counted in the case of late - type target galaxies .
we consider three kinds of environmental factors .
one is the mass density described by many neighboring galaxies over a few mpc scale .
this is called the large - scale background density .
another is the small - scale mass density attributed to the closest neighbor galaxy .
the third is the morphology of the closest neighbor galaxy .
the background density at a given location of a galaxy is measured by @xmath37 using the @xmath15-band luminosity @xmath38 of the closest twenty galaxies in the sample . here
@xmath39 is the mean density of the universe , @xmath40 is the mass - to - light ratio of a galaxy , and @xmath41 is a smoothing filter function . here the mass associated with a galaxy plus dark halo system is assumed to be proportional to the @xmath15-band luminosity of the galaxy .
the mean mass density within a sample of the total volume @xmath42 is obtained by @xmath43 where the summation is over all galaxies brighter than @xmath44 in the sample . only the relative mass to light ratios @xmath40 for early and late types
are needed in equation ( 1 ) since @xmath40 s appear both in the numerator and denominator .
we assume @xmath45 at the same @xmath15-band luminosity .
this is our choice of the connection of luminosity and morphology with the halo mass .
it is based on the results given in section 2.5 .
we assume @xmath40 is constant with galaxy luminosity for a given morphological type .
the mass - to - light ratio of galaxies is actually expected to be a monotonically increasing function of galaxy halo mass over the luminosity range of our sample ( @xmath46 ; see figures 3 and 4 of kim , park , & choi 2008 ) .
then the overdensity of the high density regions will be underestimated .
however , the relation between our mass density estimate and the true one is still monotonic , and only the labels of @xmath47 and @xmath2 will change .
we find the mean mass density @xmath48 where @xmath49 is the mass of a late - type galaxy with @xmath50 .
we use the spline - kernel weight @xmath51 for the background density estimation .
we vary the size of the spline kernel to include twenty galaxies with @xmath52 within the kernel weighting .
the spline kernel is adopted because it is centrally weighted , unlike the tophat or cylindrical kernel , and has a finite tail , unlike the gaussian .
our kernel is also adaptive .
an adaptive kernel constrained to include a fixed number of objects , allows more uniform smoothing in the ` initial ' conditions compared to the method adopting a fixed - scale at the present epoch since the high density regions collapse while the under - dense regions expand as the universe evolves .
the methods of calculating @xmath47 and the examination of the results are described in full detail by park et al .
interested readers should refer to section 2.3 of the paper .
our background density estimate @xmath47 spans the large - scale environment from voids to clusters .
but the smoothing scale determined by 20 galaxies with @xmath52 is larger than the typical cluster virial radius , which is 12 @xmath0mpcs .
therefore , in our calculation @xmath47 never exceeds the virialization density @xmath53 , where @xmath54 is the critical density of the universe .
our sample includes massive clusters , and @xmath55 at the locations of cluster member galaxies ranges roughly from 50 to 400 . at these densities
cluster member galaxies are mixed with those outside clusters . in our forthcoming paper park & hwang ( 2008 )
we will resolve the virialized regions of abell clusters , and study the dependence of galaxies within 10 times the cluster virial radius on the clustercentric radius and the nearest neighbor distance .
our background density estimation is made using a spherically symmetric smoothing kernel and using the redshift space distribution of galaxies . in redshift space
some of cluster member galaxies are stretched along the line of sight , appearing as fingers - of - god .
this causes smearing of cluster galaxies into low density regions . even though the fraction of galaxies dislocated by the redshift - space distortion more than our smoothing scale is small , a caution must be given to cases showing a very weak dependence on @xmath47 in our results .
the small - scale density experienced by a target galaxy attributed to its neighbor is estimated by @xmath56 where @xmath2 is the projected separation of the nearest neighbor galaxy from the target galaxy . the density due to
the nearest neighbor used in our work does not represent the galaxy number density at small - scales , but rather the local mass density given by the nearest neighbor itself .
the method to find the nearest neighbor is described in the next section .
we will study the dependence of galaxy properties on the nearest neighbor distance normalized by the virial radius of the nearest neighbor .
we define the virial radius of a galaxy as the projected radius where the mean mass density @xmath57 within the sphere with radius of @xmath2 is 200 times the critical density or 740 times the mean density of the universe , namely , @xmath58 since we adopt @xmath59 , @xmath60 .
this is almost equal to the virialized density @xmath61 in the case of our @xmath4cdm universe ( gott & rees 1975 ) .
this is what park et al .
( 2008 ) used to define the virial radius . according to our formulae
the virial radii of galaxies with @xmath62 and @xmath63 are 260 , 300 , and 350 @xmath0 kpc for early types , and 210 , 240 , and 280 @xmath0 kpc for late types , respectively .
these sizes are much larger than the visible part of galaxies .
therefore , when we mention a galaxy , we actually mean the galaxy plus dark halo system
. in the next section we will see the importance of the galaxy - component of the galaxy - halo systems during interactions .
besides the gravitational effects , the halo determines the size of the virial radius of a system and the domain of hydrodynamic influence .
the galaxy determines which kind of hydrodynamic effects to occur .
we define the nearest neighbor galaxy of a target galaxy as the one which is located closest to the galaxy on the sky and satisfies magnitude and radial velocity conditions .
suppose we are looking for the neighbors of a target galaxy with @xmath64 and with a certain morphological type .
its neighbors are defined as those with the absolute magnitudes brighter than @xmath65 and the radial velocity difference less than @xmath11 . we adopt @xmath66 and @xmath67600 km s@xmath9 ( early - type target ) or 400 km s@xmath9 ( late - type target ) .
our results below are insensitive to the choice of @xmath68 .
but if @xmath68 is too large , the size of target galaxy sample becomes too small because the absolute magnitude limit of the full sample has the limit of @xmath31 and the target sample must have the limit of @xmath69 to be complete in neighbor sampling .
we choose @xmath70 as the optimum case making the influential neighbors included and yet yielding good statistics .
the galaxies within the rectangular box of figure 1 is our major target galaxy sample .
the reason that we use different limits to @xmath11 for early and late - type targets is explained in the next section . instead of the nearest one we have also tried to use the most - influential neighbor that produces the highest local density @xmath57 at the location of the target galaxy .
our results in the following sections are qualitatively the same when we use the most - influential neighbor instead of the nearest neighbor because the majority of the most - influential neighbors are actually the nearest ones .
the choice of @xmath11 is based on the pairwise velocity difference between target galaxies and their neighbors ( see also section 2.4 of park et al .
2008 ) . for a given target galaxy with a given morphological type and with @xmath32 we searched for all neighbors with the projected separation of @xmath71kpc .
figure 2 shows the distributions of the radial velocity difference between the target and neighbor for four cases ; early - type target and early - type neighbor ( circles ) , early - type target and late - type neighbor ( squares ) , late - type target and early - type neighbor ( crosses ) , and late - type target and late - type neighbor ( stars ) .
it is important to note that the velocity difference distributions for late - type target galaxies are quite different from those for early - type target galaxies . at fixed luminosity late types
have smaller velocity difference with their neighbors than early types .
a sample of galaxy pairs with small @xmath11 ( say , @xmath72 km s@xmath9 ) will be dominated by late - type galaxies , and that with large @xmath11 ( say , @xmath73 km s@xmath9 ) by early - type galaxies ( see the discussion section for its systematic effects ) . as can be seen in figure 2 , the velocity dispersion of neighboring galaxies around a dark halo is not simply defined since it depends on the morphology of the neighbors .
the distributions can be fit well by the gaussian function ( solid and long - dashed curves ) for early - type target galaxies , but by the exponential function ( short dashed and dotted curves ) for late - type targets . to fit the distributions we used the gaussian plus constant model for early - type targets , and the exponential plus constant model for late - type targets , namely , @xmath74 and @xmath75 respectively .
we obtained the best - fit values @xmath76 ( e - e ) , @xmath77 ( e - l ) , @xmath78 ( l - e ) , and @xmath79 km s@xmath9 ( l - l ) .
our choices of the velocity limits @xmath80 km s@xmath9 for early - type targets and 400 km s@xmath9 for late types , take into account this dependence of pairwise velocity on morphology .
@xmath81 should also depend on luminosity , but our choice is a conservative one for galaxies brighter than @xmath82 and it is expected that the dependence is not strong .
for example , if we extend the faber - jackson ( faber & jackson 1976 ) or tully - fisher relations ( tully & fisher 1977 ) to the kinematics of galaxy pairs and adopt a constant mass - to - light ratio , we obtain @xmath83 .
so there will be 26% difference in @xmath84 on average for galaxies with one magnitude difference . for target galaxies brighter than @xmath85
the gaussian fit gives @xmath86 326 ( e - e ) , 219 ( e - l ) , 224 ( l - e ) , and 152 ( l - l ) km s@xmath9 , for the four cases , and the ratios of @xmath84 between early and late - type target galaxies are 326/224 and 219/152 , both close to 1.4 . besides the morphology and luminosity dependence the velocity difference is also a function of @xmath2 .
for the target galaxies with @xmath32 the neighbor galaxies located at @xmath87 kpc have velocity dispersions of 257 , 153 , 205 , and 147 km s@xmath9 for the four cases , respectively . the velocity dispersion becomes lower for early - type targets , but higher for late - type targets , and correspondingly , the ratio becomes smaller ( 1.01.2 ) .
the relation between the velocity dispersion and the virial mass , @xmath88 , is needed to convert the measured velocity difference dispersion to the halo mass .
dark matter halo virial relation estimated from simulated @xmath4cdm models yields @xmath89 ( evrard et al .
an nfw halo with no or slightly anisotropic velocities have @xmath90 when the dispersion is measured at @xmath91kpc ( conroy et al .
2007 ) , and @xmath92 when the dispersion is measured at @xmath93kpc ( prada et al . 2003 ) . figure 2 tells that the ratio of @xmath84 for early and late - type targets is about 1.4 for @xmath13kpc .
if we adopt @xmath94 , the ratio of dark halo virial mass for early and late - type targets will be about 2 .
this is why we adopted @xmath95 .
our main conclusions are essentially independent of this approximation .
in this section we study the dependence of galaxy properties on three environmental factors ; the nearest neighbor distance @xmath2 , nearest neighbor s morphology and the large - scale background density @xmath47 .
we are going to consider the physical parameters like morphology , @xmath15-band luminosity , @xmath27 color , @xmath28 color gradient , equivalent width of the @xmath96 line , @xmath29-band concentration index , central velocity dispersion , and @xmath29-band petrosian radius .
these parameters reflect most of major physical properties of galaxies from morphology and mass to internal kinematics and sf activity .
in most cases we fix the absolute magnitude of target galaxies in a narrow range between @xmath97 and @xmath63 to examine the pure environmental effects with the effects due to the coupling of a parameter with luminosity taken out .
we have also studied a set of brighter target galaxies with @xmath98 , and obtained qualitatively the same results but with much worse statistics .
all late - type target galaxies with the @xmath29-band isophotal axis ratio less than 0.6 are discarded in the following analysis whenever necessary in order to reduce the wrong trends that can be produced by the internal extinction and by the corresponding dispersion in luminosity ( choi et al .
2007 ) . .
contours show constant early - type galaxy fraction @xmath99 .
contours are limited to regions with statistical significance above @xmath100 .
( lower ) same , but for the late - type neighbor case .
arrows at @xmath101 are drawn to guide the eye . ]
park et al .
( 2008 ) made an extensive study of the dependence of galaxy morphology on the density @xmath57 attributed to the nearest neighbor and also on neighbor s morphology .
dependence on the large - scale background density @xmath47 was also studied by analyzing subsets of galaxies located in three different bins of @xmath47 .
it was found that , when luminosity is fixed , the probability for a randomly chosen target galaxy to have an early - type , @xmath99 , depends mostly on the projected separation @xmath2 when @xmath102 , the virial radius of the nearest neighbor galaxy .
but when @xmath103 , @xmath99 depends on all three environmental factors in the sense that @xmath99 is an increasing function of both @xmath57 and @xmath47 when the neighbor is an early - type galaxy , but that @xmath99 first increases and then decreases as the galaxy approaches a late - type neighbor while it is still an increasing function of @xmath47 ( see figure 3 of park et al .
2008 ) . to clarify the morphology dependence on those environmental factors we inspect the behaviour of @xmath99 in the @xmath2-@xmath47 parameter space in figure 3 .
the absolute magnitude of target galaxies is limited to a narrower bin , @xmath104 , in order to reduce the effects of the correlation of morphology and density with luminosity .
figure 20 of park et al .
( 2007 ) showed the same kind of plot , but neighbor s morphology was not differentiated there .
the upper panel of figure 3 shows the dependence of @xmath99 on @xmath2 and @xmath47 when the nearest neighbor is an early - type galaxy .
the lower panel is the case when the neighbor is a late - type galaxy .
since the mean separation between galaxies decreases as the background density increases , there is a correlation between @xmath2 and @xmath47 .
because of this statistical correlation galaxies are distributed along the diagonal in the figure .
but there is a large dispersion in @xmath2 at a given @xmath47 , particularly when @xmath105 or @xmath106 .
for example , at a fixed large - scale background density of @xmath107 the separation to the nearest neighbor of an early - type galaxy with @xmath108 can be as small as 0.003 @xmath0mpc or as large as 2 @xmath0mpc .
we will see a large variation in galaxy properties as @xmath2 changes while @xmath47 is fixed . when @xmath47 is very small , however , the correlation is tight and the information in @xmath47 and @xmath2 is rather redundant . when @xmath47 is higher than @xmath109 , most tight pairs with @xmath110 are early - type galaxies .
smooth distributions of @xmath111 are obtained from the ratio of the sum of weighted number of early types to the sum of weighted number of all galaxies within the smoothing kernel at each point of the parameter space .
a fixed - size spline - kernel is used to give the weights .
contours denote the constant levels of @xmath99 and are limited to regions with statistical significance above @xmath100 .
a striking difference between two contour plots is seen where @xmath112 , namely when the target galaxy is located inside the virial radius of its neighbor .
if the neighbor is an early - type galaxy , the contours have positive slopes there , meaning that @xmath99 is increasing for decreasing @xmath2 and increasing @xmath47 .
but if the neighbor is a late - type galaxy , the contours have negative slopes , telling that @xmath99 decreases for decreasing @xmath2 even though it is still an increasing function of @xmath47 . at a fixed @xmath47
, @xmath99 has a maximum at @xmath113 for the late - type neighbor case .
this scale corresponds to @xmath114kpc for the late types in our sample .
galaxies having their late - type neighbors within this critical distance , start to have significant hydrodynamic effects from neighbor .
sensitivity of @xmath99 to @xmath47 exists mainly within the virialized region with @xmath115 .
when @xmath102 both panels show that galaxy morphology depends mostly on @xmath2 and is nearly independent of background density @xmath47 as can be noticed from nearly horizontal contours . contours at @xmath116 are slightly contaminated by the trend that the galaxies at the upper edge of the distribution , having the largest @xmath2 at a given @xmath47 , are relatively brighter and tend to be earlier in type ( see fig . 5 below ) .
as we decrease the absolute magnitude bin size , the contours become flatter , and show weaker dependences on @xmath47 in this pair separation range . but a slight dependence of @xmath111 on @xmath47 seems persisting at @xmath102 for the early - type neighbor case .
figure 3 confirms park et al.(2008 ) s findings that the effects of the nearest neighbors are critically important to galaxy morphology and that the large - scale density matters only when pairs are closer than the virial radius .
the early - type fraction can be very high in very high large - scale density regions even when the neighbor is a late type .
on the other hand , for the isolated galaxies with @xmath117 located in very low density regions ( @xmath118 ) , the early - type fraction @xmath119asymptotically approaches about 0.2 , which might be the inborn morphology fraction because the galaxy interaction and merger rates are low there .
gott & thuan ( 1976 ) proposed that galaxy morphology is determined by the amount of gas left over at maximum collapse of the protogalaxy .
primordial elliptical galaxies are expected to form if star formation is finished by the time of maximum collapse .
this can happen if the star formation time scale is shorter than the collapse time , which is more likely in high density regions .
in reality , according to simulations galaxies continously accrete gas and other objects till the present time , and it is difficult for sf to end in the course of formation .
it is also conceivable that an isolated system of late - type galaxies is formed in low density regions and transforms to an isolate early - type galaxy by consuming all cold gas in the system through a series of close interactions and mergers ( park et al .
these isolated early types can not be formed by the mechanisms like strangulation because they have no nearby larger halo .
this kind of early types is analogous to the central dominant elliptical galaxies in the ` fossil groups ' which is the end - result of galaxy mergers ( jones et al .
2003 ; ulmer et al . 2005 ; mendes de oliveira et al . 2006 ; donghia et al . 2005 ; adami , susseil , & durret 2007 ; von benda - beckmann et al . 2008 and references therein ) . on the other hand
, a merged object can become an early - type galaxy due to the agn heating ( croton & farrar 2008 ) and remain isolated .
are plotted .
the median @xmath64 relations are drawn for early - type neighbor ( red dots , magenta line ) and late - type neighbor ( blue dots , green line ) cases . ]
figure 4 shows the @xmath15-band absolute magnitude of galaxies as a function of @xmath2 .
the lines are the median relations .
we use all target galaxies with @xmath120 in this section .
late types with the axis ratio @xmath121 are still discarded because their absolute magnitudes are not reliable .
as figure 6 of park et al .
( 2008 ) showed , more isolated galaxies are brighter on average and such trend is greater for early - type galaxies .
figure 4 also shows that galaxies with an early - type neighbor are much brighter than those with a late - type neighbor when they have relatively large separations from neighbor ( @xmath102 ) .
figure 5 shows that the dependence of @xmath64 on @xmath2 and @xmath47 .
smooth distributions of @xmath64 are found by the following method ( it is applied to all panels in figs . 5 , 7 , and 9 ) .
at each location of the @xmath122-@xmath123 space , where a smoothed median value is to be estimated , we first sort the parameter values of the galaxies contained within a certain radius from the location . in the case of the tophat smoothing ,
these galaxies get a uniform weight and the median would be the value of the @xmath124-th galaxy when the total number of galaxies within the smoothing radius is @xmath5 .
we adopt a spline kernel smoothing for more accurate results .
we assign a spline kernel weight @xmath125 to each galaxy within the smoothing radius .
the median is given by the parameter value of the @xmath126-th galaxy in the sorted list , where @xmath127 is the sum of all weights given to the galaxies within the smoothing kernel .
the median represents the typical value of a physical parameter more reliably than the mean , when the distribution is very skewed as in the case of @xmath128 in particular .
figure 5 shows that there is strong environmental dependence of @xmath64 at separations @xmath129 while such dependence fades away at smaller separations . at @xmath106
the absolute magnitude fluctuates within 0.1 magnitude in all panels without showing significant dependence on @xmath2 or @xmath47 .
one should note that a single horizontal or vertical contour does not show any environmental dependence .
there should be a significant gradient in @xmath64 revealed by a series of parallel contours with different levels in order to claim a dependence . at a given large - scale density @xmath47 the brightest galaxies are those having the largest @xmath130 , namely the most isolated ones .
the @xmath2-dependence of luminosity persists from very high density regions to well - inside the voids .
the void galaxies are also participating in luminosity evolution , but the speed of the evolution is slow because of fewer neighbors . a fact noted from figure 5 is that such a trend exists in all density environments , all the way from voids to very high density regions with @xmath131 ( but note that we can not resolve the cluster regions where @xmath47 itself is higher than the virialized density ) .
figure 5 also shows that the luminosity - density relation ( the horizontal direction in fig .
5 ) is a very strong function of neighbor distance .
for example , in the case of the e - e galaxies , luminosity of very isolated galaxies with @xmath132 rises quickly as @xmath47 increases and reaches @xmath133 at the density as low as @xmath134 .
but for galaxies with @xmath135 it rises much slowly for increasing @xmath47 and reaches @xmath133 only at @xmath136 .
the typical luminosity of the galaxies with neighbors at much closer distances never reaches this magnitude .
it can be also noted from figure 5 that the early - type galaxies tend not to have late - type neighbors at very small distances . there is almost no e - l galaxy at @xmath137 while many e - e galaxies have close neighbors at these separations ( compare the scatter plots in the left column of fig .
this may be because the early - type galaxies having their late - type neighbors at @xmath137 can acquire enough cold gas and transform their own morphology to late type .
in fact , there are more such tight l - l galaxies than e - l galaxies as can be seen in the two bottom panels of figure 5 .
evidence for the morphology transformation from an early to a late type has been presented by park et al .
( 2008 ) .
figure 6 shows variations of three physical parameters of our target galaxies with @xmath32 as a function of pair separation normalized to the virial radius of their nearest neighbor .
left panels are for the early - type target galaxies , and right panels for the late types . in each panel
red points are galaxies with early - type neighbors , and blue points are those with late - type neighbors .
their median values are shown by magenta and green lines , respectively .
top left panel shows that the @xmath27 color of early - type galaxies hardly changes as they approach neighbors .
even when the neighbor is a late - type galaxy , the change is very small , which is surprising because morphology of target galaxies is strongly affected by the neighbor distance and tends to become a late type when the neighbor is a late type .
the main reason for this is , of course , because the subsample is already restricted to early types whose color shows a small dispersion .
but our subsamples are not simply divided by color , but divided according to morphology as much as possible ( park & choi 2005 ) and correspondingly the early - type subsample contains many blue galaxies .
for example , among the early types plotted in figure 6 more than 7% have @xmath27 color bluer than 2.4 .
one could expect to see some blueing trend for early types interacting with a late type . even though the total color does not change much
, we will see below that the sf activity of early - type galaxies having late - type neighbors is actually enhanced when the separation is much smaller than the virial radius , suggesting cold gas transfer from their neighbors .
the upper right panel of figure 6 shows that , when the target galaxy is a late type , its @xmath27 color clearly becomes redder as it approaches an early - type neighbor , but does not change much if the neighbor is a late type .
the bifurcation occurs at @xmath138 . from figure 6
one can understand why it has been so difficult to detect the change in color for interacting pairs ( de propris et al .
early types do not show any significant changes , and the changes in late types depend on neighbor s morphology . without dividing the sample according to target s and neighbor s morphology one would find little change in color . in the left column of figure 7
the behaviour of the median @xmath27 color in the @xmath2-@xmath47 space is inspected .
we again find that the color of early types is very insensitive to both @xmath2 abd @xmath47 .
the median color of the early - type galaxies having an early - type neighbor ( the e - e galaxies ) or a late - type neighbor ( the e - l galaxies ) varies only about 0.02 in the @xmath2-@xmath47 space .
the bottom two panels of the left column of figure 7 show the median @xmath27 color contours for late - type target galaxies .
it can be seen that the @xmath27 color of the l - e galaxies depends on @xmath2 weakly when @xmath139 but strongly for @xmath140 .
the dependence is negligible at @xmath102 .
the l - l galaxies tend to be slightly bluer as @xmath2 decreases .
this is why the l - e and l - l cases start to separate from each other at @xmath141 and then diverge at @xmath140 .
figure 7 also shows that , when @xmath142 , late - type galaxies become redder as their background density becomes higher .
this weak residual dependence of color of late - type galaxies on @xmath47 after fixing luminosity has been also reported by park et al .
the residual dependence is weak ; the total variation in @xmath27 of late types is about 0.2 magnitude as @xmath55 varies from 0.1 to 100 .
if luminosity or morphology are not fixed , the color variation is much larger ( see fig .
11 of park et al .
the large - scale density dependence of @xmath27 is probably a result of the accumulated effects of galaxy - galaxy interactions and mergers whose frequency is higher in higher density regions .
an interesting fact from these plots is that the dependence of color on @xmath47 becomes negligible when @xmath143 , which can be noticed from widening of contour separation and serpentine contours at small @xmath47 .
we will also see in the next section that the sf activity as measured by @xmath128 does not depend on the background density @xmath47 either when @xmath144 and @xmath129 .
on the other hand , when the background density is very high ( @xmath145 ) , the color of late - type galaxies appears to depend only on @xmath47 .
the density range corresponds to the cluster environment , and a cluster acts like a giant early - type galaxy transforming member galaxies into redder ones .
recently , park & hwang ( 2008 ) have studied properties of galaxies near and within abell clusters , and found a characteristic scale of @xmath146 times the cluster virial radius across which various galaxy properties suddenly start to show dependence on the clustercentric radius . since the cluster galaxies are smeared along the line of sight due to the finger - of - god effect in our analysis , reddening of late types due to clusters
is expected to appear rather smoothly as a function of @xmath47 .
kauffmann et al . ( 2004 ) claimed that the sf history - density correlation is sensitive to small - scale density , but that there is no evidence for the sf history to depend on large - scale ( @xmath147 mpc ) density .
we find in figure 7 that the @xmath27 color as a measure of sf history is nearly independent of small and large - scale environments for early - type galaxies , but depends on both for late - type galaxies even at fixed luminosity .
figure 6b ( middle panels ) shows variations of the equivalent width of the @xmath96 line , a measure of the sf activity , as a function of @xmath2 for the early - type ( left ) and late - type ( right ) target galaxies with @xmath32 .
galaxies are again distinguished between those having an early - type neighbor ( red dots , magenta line ) and those having a late - type neighbor ( blue dots , green line ) .
even though the total color of early - type galaxies is not much affected by interactions , the sf activity measured by @xmath128 clearly shows dependence on @xmath2 and neighbor s morphology .
this tells us that the color is simply not a very sensitive sf indicator , especially for massive galaxies .
early - type galaxies show slightly reduced sf activity when they approach an early - type neighbor .
but the e - l galaxies show relatively stronger @xmath96 line emission at @xmath115 , and the gap between two cases becomes wider at @xmath149 .
@xmath128 of late - type galaxies shows more dramatic variations as @xmath2 changes . in the right panel @xmath128 of the l - e galaxies
is nearly constant down to @xmath150 and then starts to drop below the separation . but @xmath128 of the l - l galaxies starts to rise at @xmath151 and increases rapidly below @xmath152 .
the two cases start to bifurcate at @xmath151 and to diverge at @xmath153 as in the @xmath27 case .
we note there are two characteristic scales in the effects of galaxy - galaxy interactions on the sf activity .
the first scale is the virial radius of the nearest neighbor galaxy where difference in morphology of the nearest neighbor starts to make the sf activity bifurcate .
the second one is the merger scale , about 0.05 times the virial radius , where the effects diverge depending on the neighbor morphology .
since the virial radius of an early or late - type galaxy with @xmath50 is 300 or 240 @xmath0 kpc , at the separation @xmath154 kpc , a pair of such galaxies should start to physically contact to each other .
the middle column of figure 7 shows dependence of @xmath128 on @xmath2 and @xmath47 for four cases .
the two top panels indicate that the sf activity of early types depends on neighbor s morphology .
the e - l galaxies shows overall higher @xmath128 compared to the e - e galaxies , and the direction of dependence of @xmath128 on @xmath2 is opposite for the two cases .
we also find that @xmath128 of early - type galaxies decreases slightly as @xmath47 increases .
the @xmath47 dependence of @xmath128 , however , does not always exist , but exists only over a certain @xmath2 range that depends on the neighbor morphology . for late - type galaxies ( two bottom panels ) the effects of early- and late - type neighbors are clearly distinguished .
the sf activity of late types is enhanced significantly as they approach late - type neighbors , but quenched slightly as they approach early - type neighbors . in the l - l case
the dependence on @xmath47 at small @xmath2 is qualitatively similar to the e - l case .
the scale of occurrence of the @xmath2 dependence depends on @xmath47 .
enhancement of sf activity by late - type neighbors occurs at smaller neighbor separation in higher @xmath47 regions .
this can happen if late - type neighbor galaxies tend to be less gas - rich , and the cold gas flow from late - type neighbors is less efficient in higher density regions .
in all four cases the sf activity does not depend on both @xmath2 and @xmath47 when galaxies are isolated ( @xmath129 ) and when the background density is small @xmath155 .
balogh et al . ( 2004b ) analyzed the galaxies and groups in the 2df galaxy redshift survey data and the sdss data , and reported that the fraction of star forming galaxies varied strongly with the background density ( see also balogh et al .
. the signal they found must be mostly due to the correlation of the background density with luminosity and morphology .
they looked at faint late - type galaxies in low density regions , but bright early types at high densities .
once luminosity and morphology are fixed , the sf activity in galaxies depends very weakly on the background density as shown in figure 13c of park et al .
( 2007 ) and in figure 7 of this work .
balogh et al . ( 2004b ) also presented evidence that the fraction of @xmath148 emitting galaxies is mostly dependent on the small - scale environment at high densities , but on the large - scale environment at low densities .
we found different results .
the second column of figure 7 shows that the sf activity of late - type galaxies having a late - type neighbor depends differently on @xmath2 at different @xmath47 in such a way that in higher density regions the sf activity rises at smaller @xmath2 .
but there is only a weak trend that is opposite to this for the l - e galaxies .
nikolic et al . (
2004 ) claimed that the mean sf activity is significantly enhanced for @xmath10 kpc .
for late - type targets the enhancement is found out to 300 kpc regardless of neighbor s morphology .
we find their results are true only when the neighbor ( or target ) galaxy is a late type . for the ` l - l ' galaxies we find the enhancement of sf activity extends out to @xmath156 kpc .
but when the neighbor is an early type , the sf activity drops at @xmath157kpc or at @xmath110 .
this shows us again that the effects of interaction become manifest when the sample is split according to the morphology of target and neighbor galaxies .
they also reported that there is no dependence of the sf activity on neighbor morphology nor mass .
this is certainly not supported by our results .
alonso et al . ( 2006 ) reported that there is a threshold for the sf activity induced by interactions at @xmath158kpc .
we do not find evidence for a threshold at that scale .
our results suggest thresholds only at the virial radius and merger scale , which are roughly @xmath159 and @xmath160kpc for galaxies in our sample , respectively .
they also found that interactions are more effective at triggering sf activity in low and moderate density environments .
this is consistent with our results only when the neighbor galaxy is a late type . in the second and bottom panels ( the e - l and l - l cases ) of figure 7 one
can see @xmath161 starts to increase at @xmath162 in very low density regions but at @xmath163 in high density regions .
the influence of early - type neighbors is weaker in low density regions , and the sf activity is less suppressed .
we use the @xmath28 color difference between the central region with @xmath164 and the annulus with @xmath165 as a measure of color gradient . here
@xmath166 is the petrosian radius ( petrosian 1976 ; blanton et al .
2001 ) in the @xmath29-band .
the difference is made in such a way that positive @xmath167 means a bluer central region relative to the outer region .
we corrected @xmath167 for the inclination and seeing effects as described by park & choi ( 2005 ) and choi et al .
we use the gradient in @xmath28 color rather than @xmath168 color because the @xmath169-band surface photometry is noisy for some galaxies .
but the surface photometry of sdss galaxies in the @xmath170 , and @xmath29-bands can be done relatively accurately for those in the spectroscopic sample ( apparent @xmath15-band magnitude @xmath171 ) .
the bottom panels of figure 6 clearly show that the central region of galaxies undergoing interactions and mergers becomes bluer relative to the outside . for early - type target galaxies
the effects are manifest only well within the virial radius .
but for late types the effects start to appear at separations larger than @xmath1 .
difference due to different neighbor morphology is small , which might seem inconsistent with the trends seen for @xmath27 and @xmath161 .
if the dependence of color and sf activity on neighbor s morphology is because of difference in the influence of cold and hot gases of the neighbor galaxy , one might think the color gradient depends on @xmath2 also differently for different neighbor morphology . but
figure 6 shows the color gradient always increases for decreasing @xmath2 independently of the morphology of target and neighbor galaxies .
it means that , as early - type galaxies approach their neighbors , the central part becomes bluer but the outer part becomes redder , making the color gradient increase but their total color remain almost the same .
for late - type galaxies the center becomes bluer for both early and late - type neighbor cases , but the outer part becomes much redder for the early - type neighbor case , making the total color redder .
early - type galaxies must be significantly reducing the sf activity in the outer part of their neighboring late - type galaxies .
it is known that the sf activity of late - type galaxies in clusters is severely reduced in the outer disk , with normal or enhanced activity in the inner disk ( boselli & gavazzi 2006 ) . in other words ,
the color gradient of late types becomes more positive ( redder outside ) in the cluster environment .
quenched sf activity in cluster spirals is often explained by the gas depletion through hydrodynamic interactions with the hot intracluster medium such as the ram pressure stripping , viscous stripping , thermal evaporation , and strangulation ( cf .
boselli & gavazzi 2006 and references therein ) .
we now find in figure 6 that late - type galaxies in general environment experience very similar changes in the sf activity .
in particular , when they approach an early - type galaxy ( instead of a cluster ) , they show redder total color , reduced sf activity , and more positive color gradient .
it is important to note that the sf quenching phenomenon is now found for late - type galaxies outside the cluster environment ( see the discussion section below for possible mechanisms ) .
figure 7 shows that an approach to an early - type neighbor makes a late - type galaxy redder in color , weaker in @xmath128 and more positive in color gradient in any background density environment ( see the panels in the third row of fig .
7 ) .
we adopt the inverse concentration index @xmath172 to quantify the radial surface brightness profile of galaxies .
it is defined by @xmath173 where @xmath174 and @xmath175 are the semi - major axis lengths of ellipses containing 50% and 90% of the petrosian flux in the @xmath29-band , respectively , and is corrected for the seeing effects ( park & choi 2005 ) .
top panels of figure 8 show the dependence of @xmath172 on the pair separation .
early - type galaxies show smaller change in concentration as they approach neighbors probably because they are tidally more stable due to smaller size and compactness , and also because the tidal energy deposit is relatively smaller as the relative velocity with the neighbor is higher for early types than for late types ( see fig .
2 ) . figure 8 tells that galaxies become more concentrated as they approach their neighbors .
but we note a very slight tendency that @xmath172 of the e - e galaxies first increases and then decreases as @xmath2 decreases .
when the galaxy undergoes a merger at @xmath176 , the dispersion in @xmath172 becomes large . when galaxies are merging , some fraction of mass escapes from them and form tidal tails and bridges .
this makes the petrosian radius usually larger and uncertain .
likewise , @xmath177 becomes uncertain too .
the tendency that galaxies become more concentrated as they approach their neighbors inside the virial radius is much stronger for late types .
this may be because late types are tidally more vulnerable due to their larger size and lower concentration than early types and because the velocity difference of late - type target galaxies with their neighbors is relatively smaller and correspondingly the tidal energy deposit is larger .
the effects are even stronger for the late - type neighbor case because the velocity difference between the target and neighbor is even lower in this case as shown in figure 2 .
more discussion is given in section 4 .
a closer look at the variation of @xmath172 reveals that @xmath172 of late types actually first increases and then decreases as @xmath2 decreases , but the increase is very small and the scale of the maximum @xmath172 ( least concentration ) differs for different neighbor morphology ( @xmath178 and @xmath179 or @xmath180 and @xmath181mpc for early and late - type neighbor cases , respectively ) .
this compares with nikolic et al .
( 2004 ) s result that @xmath172 peaks at @xmath182kpc when morphological types of interacting galaxies are ignored .
the typical value of @xmath172 is quite different for early and late - type galaxies , and the mean morphology of interacting galaxies varies depending on the pair separation .
therefore , if the morphological type of the target galaxies is not distinguished , in addition to the genuine trend caused by interaction , one will also see an apparent trend in @xmath172 caused by the change in the average morphology of the target galaxy as we change @xmath2 .
the distribution of @xmath172 in the @xmath2-@xmath47 space shown in the left column of figure 9 , confirms the very weak decline of @xmath172 for decreasing @xmath2 in the case of early - type galaxies ( with a weak maximum for the e - e case ) , and a strong decrease for decreasing @xmath2 in the case of late - type galaxies .
most contours in figure 7 and 9 below look noisy particularly when @xmath139 .
this is mainly because galaxy properties are almost independent of both @xmath2 and @xmath47 when galaxy luminosity and morphology are fixed , and not because statistical uncertainties are large . when @xmath183 , @xmath172 varies within 0.005 in all cases .
when @xmath184 , @xmath172 increases weakly with @xmath47 for the e - e galaxies , but does not show a dependence on @xmath47 for the e - l galaxies . the fact that @xmath47 dependence of early - type galaxies depends on neighbor morphology and the fact that the @xmath47 dependence of the e - e galaxies depends on the neighbor distance @xmath2 , together
imply that @xmath47 is not directly affecting @xmath172 .
the bottom two panels of figure 9 shows a strong increases of concentration for late - type galaxies as @xmath2 decreases .
but the dependence of @xmath172 on @xmath47 is not clear .
according to park & hwang ( 2008 ) , @xmath172 of late types decreases as the clustercentric radius becomes less than the cluster virial radius .
figure 11c of park et al .
( 2007 ) showed that @xmath172 is nearly constant of @xmath47 for bright early - type galaxies but increases very slightly with @xmath47 for the galaxies much fainter than @xmath185 .
note that our target galaxies are basically @xmath186 galaxies for which the background density dependence of @xmath172 is expected to be small .
our result is consistent with that of blanton et al .
( 2005a ) who claimed that structural properties of galaxies are less closely related to galaxy ` environment ' than are their masses and sf histories .
however , this is true only for the large - scale background density environment .
we found a significant dependence of galaxy structural parameters ( @xmath172 in this section and the central velocity dispersion @xmath187 in the next section ) on the environmental factors like neighbor distance and morphology .
van der wel ( 2008 ) has studied the dependence of galaxy morphology and structure on environment and stellar mass , and concluded that galaxy structure mainly depends on galaxy mass but morphology mainly depends on environment .
even when both galaxy luminosity and morphology are fixed , we still find galaxy structure depends sensitively on the neighbor environment .
we use the velocity dispersion measured by an automated spectroscopic pipeline called idlspec2d version 5 ( d. j. schlegel et al . 2008 , in preparation ) .
galaxy spectra of sdss galaxies are obtained by optical fibers with angular radius of @xmath188 .
the central velocity dispersion measurement has been corrected for the smoothing effects due to the finite size of the optical fiber ( see section 3.1 of choi et al .
taking into account the finite resolution of the spectrographs , we discarded galaxies with @xmath189 km s@xmath9 . the middle panels of figure 8 show variations of @xmath187 as a function of @xmath2 for the four cases .
early types hardly change their central velocity dispersion , which may be again because early types are fast , compact , tightly bound , and correspondingly are tidally more stable .
the central velocity dispersion of late - type galaxies monotonically increases as they approach their neighbors within @xmath1 .
the increase is stronger when the neighbor is a late type .
it is very likely that this is because the late - type neighbor on average has a smaller relative velocity ( see fig .
2 ) , and therefore produce more tidal energy deposit than an early - type neighbor .
coziol & plauchu - frayn ( 2007 ) have recently inspected asymmetries in galaxies pairs , and concluded that the features are consistent with tidal effects produced by companions .
the middle column of figure 9 shows the dependence of @xmath187 on both @xmath2 and @xmath47 for the four cases . in all cases we notice that @xmath187 slightly increases as @xmath47 increases which was also shown in figure 13b of park et al .
( 2007 ) for galaxies with @xmath190 or @xmath191 .
interestingly , the increase mainly occurs in low and intermediate density regions .
the top two panels show that @xmath187 increases at the smallest @xmath2 at fixed @xmath47 for the e - e galaxies , but is nearly a constant as @xmath2 decreases for the e - l galaxies . in the case of the l - e galaxies ,
the @xmath2-dependence of @xmath187 starts to appear within @xmath141 .
the neighbor dependence is strong for the l - l galaxies , in particular .
we use the petrosian radius ( graham et al .
2005 ) as a measure of galaxy size .
it is measured from the @xmath29-band images taking into account inclination and seeing ( choi et al .
2007 ) . in the case
of late - type target galaxies , we are using only those with the @xmath29-band isophotal @xmath33 ratio greater than 0.6 .
the bottom panels of figure 8 show variations of @xmath166 as a function of the pair separation .
a prominent feature in this plot is that the early - type galaxies appear much larger at @xmath192 while such increase in size is not noticeable for late - type galaxies .
the size of early - type galaxies in very close pairs may have been systematically overestimated because of blending . when undergoing mergers , early types are expected to survive longer than late types because they are relatively more compact and faster .
furthermore , the neighbor undergoing a merger with an early - type galaxy is very likely to be an early type too ( park et al .
. it would be difficult to define a boundary for a tight pair of early - type galaxies with smooth distribution of stars .
some of early types can be still identified as separate objects even at @xmath193kpc when their outer extended envelopes are already merged with those of their neighbors , and the size of such galaxies can be easily overestimated . on the other hand ,
since the pairwise velocity is smaller for late types , it is expected that their cores merge relatively quickly and that there are relatively fewer late - type pairs with very small separations .
size of late - type galaxies is determined by the light from disk which has the boundary relatively abrupt compared to those of early - type galaxies .
this may be why there is no very tight late - type pair and why an increase in @xmath166 at small @xmath2 is not observed for late types .
the top two panels of the right column of figure 9 show @xmath166 in the @xmath2-@xmath47 plane for the early - type galaxies .
they show that @xmath166 first slightly decrease between @xmath194 and then increase at shorter separations .
the size of the e - e galaxies increases rapidly at @xmath195 , and the actual scale the galaxy size starts to rise depends on @xmath47 .
the e - e galaxies in high density regions appear larger than those in low density regions when they merge . except for this
merger scales the size of early types hardly depends on @xmath2 or @xmath47 .
the size of late - type target galaxies , shown in the bottom two panels , is nearly independent of both @xmath2 and @xmath47 .
galaxy size depends strongly on luminosity and morphology as shown by figure 4 of choi et al .
for example , the typical value of @xmath166 of early types varies from 3 to @xmath196kpc , and that of late types varies from 4.5 to @xmath197kpc as @xmath64 changes from @xmath198 to @xmath199 .
correspondingly , @xmath166 of a random galaxy varies significantly as @xmath2 or @xmath47 varies because the average morphology and luminosity change too .
but once we fix morphology and luminosity within one magnitude , @xmath166 is effectively fixed showing variation less than 0.5 @xmath0kpc except for early types undergoing mergers .
in the previous section various physical parameters of galaxies are studied as a function of three environmental factors ; the nearest neighbor distance , the nearest neighbor s morphology , and the large - scale background density .
an important finding was that the virial radius of the galaxy plus dark halo systems acts as a landmark where most of the galaxy properties start to be sensitive to both nearest neighbor s distance and morphology .
if the sf activity of galaxies in pairs , for example , is enhanced due to the internal mass perturbed by the tidal force of the neighbor , it will be always enhanced by the galaxy - galaxy interactions .
contrary to this expectation , the middle panels of figure 6 and the second column of figure 7 show that it can be enhanced or suppressed depending on neighbor galaxy s morphology unless the background density is very high .
the fact suggests that it is neighbor galaxy s cold or hot gas that enhances or suppresses the sf activity , respectively , when their separation is less than the virial radius . on the other hand ,
the structural parameters changes with @xmath2 in the same direction independently of neighbor s morphology .
this tells that galaxies also evolve through gravitational effects as they approach each other .
which physical mechanism is responsible for the correlation in morphology and sf activity between neighboring galaxies ?
why do they suddenly care about neighbor s morphology when they are closer than the virial radius ?
one possible explanation is the primordial origin .
galaxies close from each other form to have similar morphology and sf activity since they are located in nearly the same large - scale environment . but this scenario can not explain why galaxy properties suddenly start to change according to the neighbor morphology at the virial radius .
the crossing time of galaxies across the virial radius is of order of @xmath200yrs , much shorter than the age of the universe .
even if there existed a primordial correlation among galaxies in physical properties over the scale of the dark halo virial radius , such a sharp transition in correlation will be wiped out due to the infall of new neighbors in the course of time . a physical process that can explain such an onset of conformity in morphology and sf activity at a characteristic separation , is the direct hydrodynamic interactions between approaching galaxies .
a galaxy plus dark halo system contains hot halo gas and/or cold disk gas which are confined within the virial radius of the system .
when a late - type galaxy enters within the virial radius of its early - type neighbor , it will start to experience the hot gas pressure in the neighbor system s halo .
the physical processes acting in this situation can be ram pressure effects due to the collision with the hot gas ball of the neighbor .
the thermal evaporation and viscous stripping of the late type s disk gas during the passage through the hot halo gas , are less able to account for the sharp transition because there is a time - delay for these processes to change galaxy properties significantly .
likewise the quenching of sf after a shutoff of cold gas supply from the halo gas ( strangulation ) is less likely too because , even if the galaxy really loses its halo gas , there will be a significant time delay for the disk gas to be consumed . if two galaxies are gravitationally bound , however , they will orbit each other within the virial radius .
then intense hydrodynamic interactions can occur repeatly many times or continuously before they merge , and all above processes can contribute to change the properties of the orbiting galaxies . since they will remain within the virial radius as they orbit , the onset of correlation in galaxy properties at the virial radius can be observed . according to numerical simulations
, interacting galaxies can start to transfer their mass after they pass the closest approach point even though the actual results depend critically on the interaction parameters ( toomre & toomre 1972 ; mihos & hernquist 1994 ; wallin & stuart 1992 ) .
then , it is conceivable that an early - type galaxy enters a late - type galaxy s virial radius , acquires some cold gas with angular momentum enough to form a disk , and transforms itself to a late type ( park et al .
. this mass transfer may be the reason why a galaxy becomes more likely to be a late type as it approaches a late - type neighbor galaxy within the virial radius even though it tends to be an early type as it approaches the late - type neighbor outside the virial radius ( see fig . 3 in section 3 and fig .
3 of park et al .
2008 ) .
in addition to the dependence on the nearest neighbor separation , galaxy properties also show dependence on the large - scale background density .
the dependence is strong only for morphology and luminosity . very interestingly
, the background density dependence of morphology appears clearly only when a galaxy is located inside the virial radius of its neighbor .
the reverse is true for luminosity ; the background density dependence of luminosity can be seen only when a galaxy is outside the neighbor s virial radius .
if the background density gives direct impacts on galaxy morphology , both isolated galaxies and galaxies in pairs should also respect the background density .
but they do not .
let us consider the reason for the relation between the large - scale density and morphology .
figure 3 shows that the early - type fraction @xmath99 monotonically increases as @xmath47 increases for both early and late - type neighbor cases when @xmath106 .
the fact that the background density dependence is manifest only when a galaxy is sitting inside its neighbor s virial radius , rules out the simple primordial origin scenario like that the relation is caused by the elliptical galaxies that preferentially formed in higher density regions .
this is because figure 3 shows isolated galaxies with @xmath102 are almost ignorant of the background density .
one might also consider the effects of other neighbor galaxies , like the second and third nearest ones , which are monotonically increasing as the background density increases .
but it again can not explain why the virial radius of the nearest neighbor should be the critical boundary for the onset of the background density dependence .
then , one would naturally suspect that some of the internal properties of the nearest neighbor galaxies depend on the background density in such a way that the strength of the neighbor influence is different at different background densities .
park et al .
( 2008 ) proposed that the background density dependence occurs due to the variation of hot halo gas property of galaxies . in higher density regions ,
the halo gas of both early and late - type galaxies seems on average hotter and denser .
galaxies with the same morphology , luminosity , and pair separation but , located in higher density regions are redder and less active than those in lower density regions ( fig . 7 )
this may be because the halo gas is maintained hotter and denser for the galaxies in higher density regions due to some internal heating mechanisms ( like supernovae and active galactic nuclei ) and external confining material .
this suggestion is yet to confirm observationally .
it is well - known that the x - ray luminosities @xmath201 of early types correlate with optical luminosity @xmath202 ( see review by mathews & brighenti 2003 ) .
more luminous early - type galaxies have larger @xmath203 , but with large scatter , than those of low luminosity .
the large scatter in the @xmath204 correlation has been explained in terms of some specific environmental effect on x - ray emission from galaxies ( brighenti & mathew 1998 ; brown & bregman 1998 ; brown & bregman 2000 ; helsdon et al .
2001 ; osullivan et al . 2001 ; ellis & osullivan 2006 ; jeltema et al .
brown & bregman ( 2000 ) found that early types in denser environments have larger @xmath204 and suggested that the x - ray luminosity is enhanced through accretion of the intergalactic gas or supression of galactic winds by the ambient medium , whereas other authors found that galaxies in regions of high local galaxy density have lower @xmath205 ( white & sarazin 1991 ; henriksen & cousineau 1999 ) .
helsdon et al .
( 2001 ) and matsushita ( 2001 ) suggested that the result of brown & bregman is thought to be due to the large fraction of luminous central - dominant group galaxies in the sample that are enhanced in @xmath201 due to an additional contribution from the intragroup or circumgalactic hot gas ( e.g. group cooling flows ) and isolated and non - central group galaxies show no significant correlation between @xmath205 and environment .
let us now consider the second major finding that more isolated galaxies are relatively brighter at fixed background density .
it has been interpreted by park et al .
( 2008 ) as evidence for transformation of galaxy luminosity class through the merger process .
the transformation rate through mergers should be higher in higher background density regions and the dependence of luminosity on the pair separation be stronger in high density regions ( but note that we are not resolving massive clusters where mergers between ordinary galaxies are expected to be difficult to happen ) . at a given background density morphology and luminosity transformations can occur through galaxy interactions and mergers . and the background density will statistically control the speed of the coupled evolution of morphology and luminosity . considering the definite dependence of morphology and luminosity on the neighbor distance at the present epoch and high redshifts ( hwang & park 2008 ) , one can draw a conclusion that these processes are the key galaxy evolution mechanisms in addition to those like cold gas accretion and internal passive evolution which happen without resorting to neighboring galaxies .
why is the correlation between luminosity and pair separation strong only when the pair separation is larger than the virial radius of the neighbor ? it can be explained if recently merged galaxies have @xmath2 larger than @xmath1 from the new nearest neighbor ( i.e. a pair of galaxies having vanishing @xmath2 merge each other and jump to large @xmath102 after merger ) .
this is possible because a pair of galaxies would merge more easily if they are located outside the virial radius of another larger galaxy .
park et al .
( 2008 ) found , in a search for evidence for this interpretation , that at fixed background density the isolated galaxies with @xmath102 show the post - merger features more frequently than those with @xmath115 , particularly in high density regions .
therefore , among the galaxies located at the same background density , more - isolated ones are more likely to be recent merger products than less - isolated ones and are likely to be brighter .
this does not mean isolated galaxies in general have experienced recent merger events .
isolated galaxies are preferentially located in low density regions where the merger rate is low , are on average expected to be passively evolving with less environmental influence . those who want to analyze passively evolving galaxies must sample isolated galaxies located in low density regions only .
our third major finding is that , once morphology and luminosity are fixed , the remaining properties of galaxies are quite insensitive to the background density , particularly when @xmath102 .
one noticeable exception is color .
the color of late - type galaxies shows a weak residual dependence on the background density above the mean density even if both morphology and luminosity are fixed and even when they are isolated . if galaxies maintain hotter and denser halo gas in higher density regions as we propose , it is possible for the sf activity of galaxies to drop for a sufficiently long time and for galaxy color to become redder relative to those in lower density regions .
enhancement of sf activity by late - type neighbor galaxies occurs at smaller neighbor separations as the background density increases ( see the e - l and l - l cases of the @xmath128 parameter in fig . 7 ) .
this observation can be also understood by the background density - dependent halo gas properties or mass transfer efficiency .
an interesting question regarding galaxy color is which is more fundamental physical property of galaxies between morphology and color .
one way to address this question is whether or not galaxy morphology shows any dependence on environment beyond its correlation to color ( ball , loveday & brunner 2008 ; van den bergh 2007 ) . to answer this question we selected two local density subsets containing galaxies with @xmath104 and @xmath206 located at the background densities @xmath143 ( low - density subset ) or @xmath207 ( high - density subset ) .
since we fixed both luminosity and color , the stellar mass of galaxies is effectively fixed ( yang et al . 2007 ) .
the fraction of early types in the low - density subset is found to be @xmath208 , but that in the high - density subset is @xmath209 .
therefore , the morphology - density relation becomes much weaker when we severely limit both luminosity and color .
but there is still some residual dependence of galaxy morphology on the background density .
this demonstrates morphology contains independent information on the environmental dependence of galaxies that color does not have .
according to the tidal interaction theory the energy deposit in a galaxy is inversely proportional to the square of the velocity difference between the interacting pairs ( binney & tremaine 1987 ) .
several previous works have reported a detection of such inverse correlation between the strength of interaction effects and the velocity difference between pairs ( barton et al .
2000 ; lambas et al . 2003 ; nikolic et al . 2004 ; alonso et al .
2006 ; woods et al .
we also find results consistent with the tidal picture from our accurate morphology subsamples as shown in figure 8 ; the l - l galaxies , having smaller velocity difference than the l - e galaxies ( see fig .
2 ) , show more variation in @xmath172 and @xmath187 with the neighbor separation than the l - e galaxies . to address this issue directly we divided the sample of the l - l galaxies into three subsets according to the velocity difference with neighbors @xmath11 , and measured the equivalent width of the @xmath96 line as a measure of the sf activity , and the central velocity dispersion as a measure of galaxy structure , as a function of the neighbor separation for each subset .
figure 10 shows the median relations and 68% ranges for these parameters .
the solid line is for the subset with @xmath210 km s@xmath9 , the dashed line for 70 km s@xmath211 km s@xmath9 , and the dotted line for 120 km s@xmath212 km s@xmath9 . due to the small size of each subset the errors are large ,
but one can still clearly see the relation is more sensitive to the neighbor separation for pairs with smaller velocity difference .
our result can be compared with those of lambas et al .
( 2003 ) and alonso et al .
( 2006 ) who reported that the onset of interaction - induced sf is seen for @xmath213 km s@xmath9 .
nikolic et al . (
2004 ) reported the onset even when @xmath214 km s@xmath9 .
we find no significant enhancement for pairs with @xmath215 km s@xmath9 , only a small enhancement for those with @xmath216 km s@xmath9 , and a significant enhancement for @xmath210 km s@xmath9 .
figure 2 implies that the average morphology of galaxies with small @xmath11 is more likely to be late type and the fraction of early types will increase as @xmath11 increases until it reaches the field fraction .
so the average sf activity can appear to be higher for pairs with smaller @xmath11 not because of the interaction effects but because of higher fraction of late - type galaxies .
most of the previous works did not carefully distinguish among different morphological types of galaxies in pairs .
therefore , it is likely that the results of the previous works are contaminated by the average morphology variation with @xmath11 in addition to the genuine effects of interaction .
the dependence on @xmath11 is less strong for @xmath187 , but still shows up at the smallest separation bin .
however , it is not clear whether or not the increase of @xmath187 is entirely due to the matter perturbation within galaxies or due to the additional contribution by the mass flow from the neighbor . to check
if our findings are robust against our choice of neighbor selection parameters , we redid our analyses using a sample of galaxies that are constrained to have neighbors brighter than themselves ( i.e. the limiting magnitude difference @xmath217 instead of 0.5 ) .
it was found that our findings remain true for these galaxy pairs .
we have inspected the dependence of eight physical parameters of galaxies on the small - scale ( the nearest neighbor distance , the nearest neighbor s morphology ) and the large - scale ( background density smoothed over 20 nearest galaxies with @xmath46 ) environments .
we have also studied the kinematic properties of the galaxies in pairs .
we found the impact of interaction on galaxy properties are detectable at least out to the pair separation corresponding to the virial radius of ( the neighbor ) galaxies in our sample , which is mostly between 200 and 400 @xmath0 kpc .
it was crucial to divide galaxy interactions into four cases depending on morphology of target and neighbor galaxies in order to detect these long - range interaction effects .
there are two characteristic pair - separation scales where the breaks in the dependence of galaxy properties on @xmath2 are observed .
the first scale is the virial radius of the nearest neighbor galaxy @xmath1 .
all parameters studied , except for @xmath166 , start to deviate from those of extremely isolated galaxies at @xmath218 in the case of late - type galaxies , in particular .
the second scale is the merger scale which is about @xmath219 .
this corresponds to @xmath220 kpc for the galaxies in our sample .
the sf activity of galaxies is enhanced when the nearest neighbor is a late type , but reduced when the neighbor is an early type .
these effects occur within the virial radius of the neighbor galaxy .
these are strong evidence for hydrodynamic interactions within the virial radius of the galaxy plus dark halo system during encounters .
the dependence of galaxy properties on @xmath47 is strong only for luminosity ( fig .
5 ) and morphology ( fig .
all other parameters show weak or negligible dependence on @xmath47 once luminosity and morphology are fixed .
we have inspected the small subtle dependence on @xmath47 whenever detectable .
for example , the @xmath27 color of late - type galaxies has a weak residual dependence on @xmath47 at fixed @xmath2 .
we suggest that galaxies in higher density environment maintain hotter and denser halo gas due to some internal heating mechanisms and external confining material , which can be the reason for the large - scale density dependence of morphology and sf activity parameters .
\4 . at fixed large - scale background density ,
galaxies with larger pair separations have higher luminosity .
such dependence exists mainly at the nearest neighbor distance larger than the virial radius of the neighbor .
this is interpreted as evidence for the on - going process of luminosity transformation through mergers . in a forthcoming paper ( park & hwang 2008 )
we will examine the dependence of the sdss galaxies associated with the abell clusters on the nearest neighbor separation and the clustercentric radius .
the latter is the large - scale environmental parameter replacing @xmath47 here .
this work will extend the present work to the extreme situation where the large - scale background density itself exceeds the virialized density .
we are also studying the effects of the nearest neighbor on galaxy properties using higher redshift samples like the goods and deep2 samples ( hwang & park 2008 ) to understand galaxy evolution due to galaxy - galaxy interactions in the high redshift universe .
the authors acknowledge the support of the korea science and engineering foundation ( kosef ) through the astrophysical research center for the structure and evolution of the cosmos ( arcsec ) .
funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the u.s .
department of energy , the national aeronautics and space administration , the japanese monbukagakusho , the max planck society , and the higher education funding council for england .
the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium for the participating institutions .
the participating institutions are the american museum of natural history , astrophysical institute potsdam , university of basel , cambridge university , case western reserve university , university of chicago , drexel university , fermilab , the institute for advanced study , the japan participation group , johns hopkins university , the joint institute for nuclear astrophysics , the kavli institute for particle astrophysics and cosmology , the korean scientist group , the chinese academy of sciences ( lamost ) , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , ohio state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington . , adelman - mccarthy , j. k. , et al .
2007 , , 172 , 634 , adami , c. , durret , f. , mazure , a. , et al .
2007 , a&a , 462 , 411 , allam , s. s. et al .
2004 , , 127 , 1883 , allen p. d. , driver s. p. , graham a.w . ,
cameron e. , liske , j. , de propris r. , 2006 , , 371 , 2 , alonso , m. s. , lambas , d. g. , patricia , t. & coldwell , g. 2006 , , 367 , 1029 , anderson , s. f. , et al .
2003 , , 126 , 2209 , ball , n. m. , loveday , j , & brunner , r. j. 2008 , , 383 , 907 , balogh , m. l. , baldry , i. k. , nichol , r. , miller , c. , bower , r. , glazebrook , k. 2004a , apj , 615l , 101 , balogh , m. l. , et al . 2004b , mnras , 348 , 1355 , bamford , s. p. et al .
2008 , , submitted ( arxiv:0805.2612 ) , barnes , j. e. , & hernquist , l. 1996 , , 471 , 115 , barton , e , j. , geller , m. j. , & kenyon , s. j. 2000 , , 530 , 660 , barton , e , j. , geller , m. j. , & kenyon , s. j. 2003 , , 582 , 668 , barton , e. j. , arnold , j. a. , zentner , a. r. , bullock , j. s. , & wechsler , r. h. 2007 , , 671 , 1538 , bell , e. f. , et al .
2006 , , 640 , 241 , bergvall , n. , laurikainen , e. , & aalto , s. 2003 , a&a , 405 , 31 , binney , j. , & tremaine , s. 1987 , galactic dynamics ( princeton : princeton university press ) , 438 , blanton , m. r. , et al . , 2001 , , 121 , 2358 , blanton , m. r. , eisenstein , d. , hogg , d. w. , schlegel , d. j. , & brinkmann , j. 2005a , , 629 , 143 , blanton , m. r. , et al .
2003 , , 125 , 2348 , blanton , m. r. , et al .
2005b , , 129 , 2562 , boselli , a. , & gavazzi , g. 2006 , , 118 , 517 , brighenti , f. , & mathews , w. g. 1998 , , 495 , 239 , brown b. a. , bregman j. n. , 2000 , , 539 , 592 , brown b. a. , bregman j. n. , 1998 , , 495 , l752 , bushouse , h. a. , werner , m. w. , & lamb , s. a. 1988 , , 335 , 74 , choi , y .- y . , park , c. , & vogeley , m. s. 2007 , , 658 , 884 , conroy , c. , et al .
2007 , , 654 , 153 , cooper , m. c. , newman , j. a.,weiner , b. j. , et al . 2007 , , 376 , 1445 , cooper , m. c. , et al .
2008 , , 383 , 1058 , croton , d. j. , et al .
2005 , , 356 , 1155 , croton , d. j. , & farrar , g. r. , 2008 , .386 , 2285 , coziol , r. , & plauchu - frayn , i. 2007 , , 133 , 2630 , cucciati , o. , iovino , a. , marinoni , c. , et al . 2006 , a&a , 458 , 39 , de propris , r , liske , j. , driver , s. p. , allen , p. d. , & cross , n. j. g. 2005 , , 130 , 1516 , di matteo , combes , f , melchior , a .-
l . , & semelin , b. 2007 , a&a , 468 , 61 , donghia , e. , sommer - larsen , j. , romeo , a. d. , et al .
2005 , , 630 , l109 , elbaz , d. , et al .
2007 , a&a , 468 , 33 , ellis , s. c. , & osullivan , e. 2006 , , 367 , 627 , ellison , s. l. , patton , d. r. , simard , l. , & mcconnachie , a. w. 2008 , , 135 , 1877 , evrard , a. e. , et al .
2008 , , 672 , 122 , faber , s. m. , & jackson , r. e. 1976 , , 204 , 668 , geller , m. j. , kenyon , s. j. , barton , e. j. , jarrett , t. h. & kewley , l. j. 2006 , , 132 , 2243 , gomez , p. l. , et al .
2003 , , 584 , 210 , goto , t. , et al .
2003 , , 346 , 601 , gott , j. r. , & thuan , t. x. 1976 , , 204 , 649 , graham , a. w. 2005 , , 130 , 1535 , helsdon s. f. , ponman t. j. , osullivan e. , forbes d. a. 2001 , , 325 , 693 , henriksen , m. , & cousineau , s. 1999 , , 511 , 595 , hernandez - toledo , h .m .
, avila - reese , v. , conselice , c. j. & puerari , i. 2005 , , 129 , 682 , hernquist l. , 1992 , , 400 , 460 , hernquist l. , 1993 , , 409 , 548 , hogg , d. w. , et al .
2003 , , 585 , l5 , hwang , h. s. , & park , c. 2008 , in preparation , jeltema , t. e. , binder , b. , & mulchaey , j. s. 2008 , , 679 , 1162 , jones , l. r. , ponman , t. j. , horton , a. , et al .
2003 , mnras , 343 , 627 , kacprzak , g. g. , churchill , c. w. , steidel , c. c. , murphy , m. t. , & evans , j. l. 2007 , , 662 , 909 , kauffmann , g. , white , s. d. m. , heckman , t. m. , menard , b. , brinchmann , j. , charlot , s. , tremonti , c. , & brinkmann , j. 2004 , , 353 , 713 , kennicutt , r. c. , & keel , w. c. 1984 , , 279 , 5 , kennicutt , r. c. , roettiger , k. a. , keel , w. c. , van der hulst , j. m. , & hummel , e. 1987 , 93 , 1011 , lambas , d. g. , patricia , t. , alonso , m. s. , & coldwell , g. 2003 , , 346 , 1189 , larson , r. b. , & tinsley , b. m. 1978 , , 219 , 46l , li , g. , kauffmann , g. , heckman , t. m. , jing , y. p. , & white , s. d. m. 2008 , , 385 , 190 , maller , a. h. , katz , n. , keres , d. , dave , r , & weinberg , d. h. 2006 , , 647 , 763 , mathews , w. g. , & brighenti , f. 2003 , , 41 , 191 , matsushita k. , 2001 , apj , 547 , 693 , mendes de oliveira , c. , cypriano , e. s. , & sodr jr .
, l. 2006 , , 131 , 158 .
mihos , j. c. , & hernquist , l. 1996 , , 464 , 641 , naab , t. , & burkert , a. 2003 , , 597 , 893 , nikolic , b. , cullen , h. , & alexander , p. 2004
, , 355 , 874 , osullivan , e. , forbes , d. a. , & ponman t , j. 2001 , , 328 , 461 , park , c. , & choi , y .- y .
2005 , , 635 , l29 , park , c. , choi , y .- y . , vogeley , m. s. , gott , j. r. , & blanton , m. r. 2007 , , 658 , 898 , park , c. , gott , j. r. & choi , y .- y .
2008 , , 674 , 784 , park , c. , & hwang , h. s. 2008 , in preparation , patton , d. r. , pritchet , c. j. , yee , h. k. c. , ellingson , e. , & carlberg , r. g. 1997 , , 475 , 29 , patton , d. r. et al .
2005 , , 130 , 2043 , perez , m. j. , tissera , p. b. , scannapieco , c. , lambas , d. g. , & de rossi , m. e. 2006 , a&a , 459 , 361 , perez , m. j. , tissera , p. b. , lambas , d .g . , & scannapieco , c. 2006 , a&a , 449 , 23 .
, petrosian , v. 1976 , , 209 , l1 , poggianti , b. m. , et al .
2008 , , accepted ( arxiv:0805.1145 ) , quintero , a. d. , berlind , a. a. , blanton , m. r. , & hogg , d. w. 2006 , , submitted ( arxiv : astro - ph/0611361 ) , robertson , b. et al .
2006 , , 645 , 896 , sanchez , s. f. , & gonzalez - serrano , j. i. 2003 , a&a , 406 , 435 , schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , , 500 , 525 , smith , b. j. et al .
2007 , , 133 , 791 , stoughton , c. , et al .
2002 , , 123 , 485 , tanaka , m. , et al .
2004 , , 128 , 2677 , tegmark , m. , et al .
2004 , , 606,702 , tissera , p. b. , dominguez - tenreiro , r. , scannapieco , c. , & saiz , a. 2002 , , 333 , 327 , toomre , a. , & toomre , j. 1972 , , 178 , 623 , tran , k - v . h. , et al .
2005 , , 627 , l25 , tully , r. b. , & fisher , j. r. 1977 , a&a , 54 , 661 , ulmer , m. p. , adami , c. , covone , g. , et al .
2005 , apj , 624 , 124 , van den bergh , s. 2007 , , 134 , 1508 , van der wel , a. , 2008 , , 675 , 13 , van dokkum , p. g. , 2005 , , 130 , 2647 , voges , w. , et al .
1999 , , 349 , 389 , von benda - beckmann , a. m. et al .
2008 , , 386 , 2345 , weinmann , s. m. , van den bosch , f. c. , yang , x. , & mo , h. j. 2006 , , 366 , 2 , white , r. e. , iii , & sarazin , c. l. 1991 , , 367 , 476 , woods , d. f. , & geller , m. j. 2007 , , 134 , 527 , woods , d. f. , geller , m. j. , & barton , e. j. 2006 , , 132 , 197 , yang , x. , mo , h. j. , van den bosch , f. c. , pasquali , a. , li , c. , & bardeen , m. 2007 , , 671 , 153 , yee , h. k. c. , & ellingson , e. 1995 , , 445 , 37 , york , d. , et al .
2000 , , 120 , 1579 | we inspect the coupled dependence of physical parameters of the sloan digital sky survey galaxies on the small - scale ( distance to and morphology of the nearest neighbor galaxy ) and the large - scale ( background density smoothed over 20 nearby galaxies ) environments .
the impacts of interaction on galaxy properties are detected at least out to the neighbor separation corresponding to the virial radius of galaxies , which is typically between 200 and 400 @xmath0 kpc for the galaxies in our sample . to detect these long - range interaction effects
it is crucial to divide galaxy interactions into four cases dividing the morphology of target and neighbor galaxies into early and late types .
we show that there are two characteristic neighbor - separation scales where the galaxy interactions cause abrupt changes in the properties of galaxies .
the first scale is the virial radius of the nearest neighbor galaxy @xmath1 .
many physical parameters start to deviate from those of extremely isolated galaxies at the projected neighbor separation @xmath2 of about @xmath1 .
the second scale is at @xmath3 kpc , and is the scale at which the galaxies in pairs start to merge .
we find that late - type neighbors enhance the star formation activity of galaxies while early - type neighbors reduce it , and that these effects occur within @xmath1 .
the hot halo gas and cold disk gas must be participating in the interactions at separations less than the virial radius of the galaxy plus dark halo system .
our results also show that the role of the large - scale density in determining galaxy properties is minimal once luminosity and morphology are fixed .
we propose that the weak residual dependence of galaxy properties on the large - scale density is due to the dependence of the halo gas property on the large - scale density . |
in adaptive filtering @xcite , there is a class of algorithms specifically designed for sparse system identification , where the unknown system only has a few large coefficients while the remaining ones have a very small amplitude so that they can be ignored without significant effect on the overall performance of the system .
a good example of them is the zero - attracting least mean square ( za - lms ) algorithm proposed in @xcite .
this algorithm can achieve a higher convergence speed , and meanwhile , reduce the steady state excess mean square error ( mse ) .
compared to the classic lms algorithm @xcite , the za - lms algorithm introduces an @xmath0 norm in its cost function , which modifies the weight vector update equation with a zero attractor term .
recently , the hypercomplex concepts have been introduced to solve problems related to three or four - dimensional signals @xcite , such as vector - sensor array signal processing @xcite , color image processing @xcite and wind profile prediction @xcite .
as quaternion - valued algorithms can be regarded as an extension of the complex - valued ones , the adaptive filtering algorithms in complex domain could be extended to the quaternion domain as well , such as the quaternion - valued lms ( qlms ) algorithm in @xcite . in this paper
, we propose a novel quaternion - valued adaptive algorithm with a sparsity constraint , which is called zero - attracting qlms ( za - qlms ) algorithm .
the additional constraint is formulated based on the @xmath0 norm .
both the qlms and za - qlms algorithms can identify an unknown sparse system effectively .
however , a better performance in terms of convergence speed is achieved by the latter one .
this paper is organized as follows .
a review of basic operations in the quaternion domain is provided in section [ sec : quaternion_basic ] to facilitate the following derivation of the za - qlms algorithm .
the proposed za - qlms algorithm is derived in section [ sec : za - qlms ] .
simulation results are given in section [ sec : simulations ] , and conclusions are drawn in section [ sec : conclusions ] .
quaternion is a non - commutative extension of the complex number , and normally a quaternion consists of one real part and three imaginary parts , denoted by subscripts @xmath1 , @xmath2 , @xmath3 and @xmath4 , respectively . for a quaternion number @xmath5
, it can be described as @xmath6 where @xmath7 , @xmath8 , @xmath9 , and @xmath10 are real - valued @xcite . for a quaternion , when its real part is zero , it becomes a pure quaternion . in this paper
, we consider the conjugate operator of @xmath5 as @xmath11 .
the three imaginary units @xmath12 , @xmath13 , and @xmath14 satisfy @xmath15 as a quaternion has the noncommutativity property , in multiplication , the exchange of any two elements in their order will give a different result .
for example , we have @xmath16 rather than @xmath17 . to derive the quaternion - valued adaptive algorithm , the starting point is the general operation of differentiation with respect to a quaternion - valued vector . at first , we need to give the definition of differentiation with respect to a quaternion @xmath5 and its conjugate @xmath18 .
assume that @xmath19 is a function of the quaternion variable @xmath5 , which is expressed as @xmath20 where @xmath19 is in general quaternion - valued .
the definition of @xmath21 can be expressed as @xcite @xmath22 the derivative of @xmath19 with respect to @xmath18 can be defined in a similar way @xmath23 with this definition , we can easily obtain @xmath24 some product rules can be obtained from above formulations , such as the differentiation of quaternion - valued functions to real variables .
suppose @xmath19 and @xmath25 are two quaternion - valued functions of the quaternion variable @xmath5 , and @xmath26 is the real variable .
then we can have the following result @xmath27 when the quaternion variable @xmath5 is replaced by a quaternion - valued vector @xmath28 , given by @xmath29^{t}\ ] ] where @xmath30 , @xmath31 , the differentiation of the function @xmath32 with respect to the vector @xmath28 can be derived using a combination of ( [ eq : general_definition ] ) straightforwardly in the following @xmath33 \label{eq : vector_definition}\end{aligned}\ ] ] similarly , we define @xmath34 as @xmath35 \label{eq : conj_vector_definition}\end{aligned}\ ] ] obviously , when @xmath36 , ( [ eq : vector_definition ] ) and ( [ eq : conj_vector_definition ] ) are reduced to ( [ eq : general_definition ] ) and ( [ eq : conj_general_definition ] ) , respectively .
to improve the performance of the lms algorithm for sparse system identification , the za - qlms algorithm is derived in this section . to achieve this , similar to @xcite , in the cost function , we add an @xmath0 norm penalty term for the quaternion - valued weight vector @xmath37 $ ] . for a standard adaptive filter , the output @xmath38 $ ] and error @xmath39
$ ] can be expressed as @xmath40&=&{\textbf{w}^{t}[n]}{\textbf{x}[n]}\\ e[n]&=&d[n]-{\textbf{w}^{t}[n]}{\textbf{x}[n]},\end{aligned}\ ] ] where @xmath37 $ ] is the adaptive weight vector with a length of @xmath41 , @xmath42 $ ] is the reference signal , @xmath43=[x[n-1 ] , \cdots , x[n - l]]^{t}$ ] is the input sample vector , and @xmath44 denotes the transpose operation . moreover ,
the conjugate form @xmath45 $ ] of the error signal @xmath39 $ ] is given by @xmath46=d^{*}[n]-{\textbf{x}^{h}[n]}{\textbf{w}^{*}[n]},\ ] ] our proposed cost function with a zero attractor term is given by @xmath47=e[n]e^{*}[n]+\gamma{\|\textbf{w}[n]\|}_1\;,\ ] ] where @xmath48 is a small constant .
the gradient of the above cost function with respect to @xmath49 $ ] and @xmath37 $ ] can be respectively expressed as @xmath50=\frac{\partial { j_0[n]}}{\partial \textbf{w}^ { * } } \label{eq : conj_gradient_cost_function}\ ] ] and @xmath51=\frac{\partial { j_0[n]}}{\partial \textbf{w } } \label{eq : gradient_cost_function}\ ] ] from @xcite , we know that the conjugate gradient gives the maximum steepness direction for the optimization surface . therefore , the conjugate gradient @xmath52 $ ] will be used to derive the update of the coefficient weight vector .
expanding the cost function , we obtain @xmath53&=&e[n]e^{*}[n]+\gamma{\|\textbf{w}[n]\|}_1\nonumber\\ & = & d[n]d^{*}[n]-d[n]{\textbf{x}^{h}[n]}{\textbf{w}^{*}[n]}-{\textbf{w}^{t}[n ] } { \textbf{x}[n]}d^{*}[n]\nonumber\\ & ~&+{\textbf{w}^{t}[n ] } { \textbf{x}[n]}{\textbf{x}^{h}[n]}{\textbf{w}^{*}[n]}+\gamma{\|\textbf{w}[n]\|}_1\;. \label{eq : extended_cost_function}\end{aligned}\ ] ] furthermore , @xmath54}}{\partial \textbf{w}^{*}}&=&\frac{\partial { ( e[n]e^{*}[n]+\gamma{\|\textbf{w}[n]\|}_1)}}{\partial \textbf{w}^{*}}\nonumber\\ & = & \frac{\partial}{\partial \textbf{w}^{*}}(d[n]d^{*}[n]-d[n]{\textbf{x}^{h}[n]}{\textbf{w}^{*}[n]}\nonumber\\ & & -{\textbf{w}^{t}[n ] } { \textbf{x}[n]}d^{*}[n]+{\textbf{w}^{t}[n ] } { \textbf{x}[n]}{\textbf{x}^{h}[n]}{\textbf{w}^{*}[n]})\nonumber\\ & & + \frac{\partial ( { \gamma{\|\textbf{w}[n]\|}_1})}{\partial \textbf{w}^{*}}\;. \label{eq : extended_gradient_cost_function}\end{aligned}\ ] ] details of the derivation process for the gradient are shown in the following @xmath55d^{*}[n])}{\partial { \textbf{w}^{*}[n ] } } = 0 \label{eq : part_1}\ ] ] @xmath55{\textbf{x}^{h}[n]}{\textbf{w}^{*}[n]})}{\partial { \textbf{w}^{*}[n ] } } = d[n]\textbf{x}^{*}[n ] \label{eq : part_2}\ ] ] @xmath56}{\textbf{x}[n]}d^{*}[n])}{\partial { \textbf{w}^{*}[n ] } } = -\frac{1}{2}d[n]\textbf{x}^{*}[n ] \label{eq : part_3}\ ] ] @xmath57 } { \textbf{x}[n]}{\textbf{x}^{h}[n]}{\textbf{w}^{*}[n]})}{\partial { \textbf{w}^{*}[n]}}=\frac{1}{2}{\textbf{w}^{t}[n]}{\textbf{x}[n]}{\textbf{x}^{*}[n]}. \label{eq : part_4}\ ] ] moreover , the last part of the gradient of cost function is given by @xmath58\|}_1})}{\partial \textbf{w}^ { * } } = \frac{1}{4 } \gamma\cdot{sgn(\textbf{w}[n])}\ ; , \label{eq : part_5}\ ] ] where the symbol @xmath59 is a component - wise sign function that is defined as @xcite @xmath60 combining the above results , the final gradient can be obtained as follows @xmath50=-\frac{1}{2}e[n]\textbf{x}^{*}[n]+\frac{1}{4}\gamma\cdot{sgn(\textbf{w}[n])}\;.\ ] ] with the general update equation for the weight vector @xmath61 = \textbf{w}[n]-\mu \nabla_{\textbf{w}^{*}}j_0[n],\ ] ] where @xmath62 is the step size , we arrive at the following update equation for the proposed za - qlms algorithm @xmath61 = \textbf{w}[n]+\mu(e[n]\textbf{x}^{*}[n])-\rho\cdot{sgn(\textbf{w}[n])}\ ; , \label{eq : update_weight_vector}\ ] ] where @xmath63 . the last term represents the zero attractor , which enforces the near - zero coefficients to zero and therefore accelerates the convergence process when majority of the system coefficients are nearly zero in a sparse system . note
that equation will be reduced to the normal qlms algorithm without the zero attractor term , given by @xcite @xmath61=\textbf{w}[n]+\mu(e[n]{\textbf{x}}^{*}[n])\;.\ ] ]
in this part , simulations are performed for sparse system identification using the proposed algorithm in comparison with the qlms algorithm .
two different sparse systems are considered corresponding to scenario one and scenario two in the following .
the input signal to the adaptive filter is colored and generated by passing a quaternion - valued white gaussian signal through a randomly generated filter .
the noise part is quaternion - valued white gaussian and added to the output of the unknown sparse system , with a 30db signal to noise ratio ( snr ) for both scenarios .
for the first scenario , the parameters are : the step size @xmath62 is @xmath64 ; the unknown sparse fir filter length @xmath41 is @xmath65 , with @xmath66 non - zero coefficients at the 2nd , 8th , 16th and 31st taps , and its magnitude of the impulse response is shown in fig .
[ fig : w ] ; the coefficient of the zero attractor @xmath67 is @xmath68 .
the learning curve obtained by averaging @xmath69 runs of the corresponding algorithm is given in fig .
[ fig : learning_curve_1 ] , where we can see that the za - qlms algorithm has achieved a faster convergence speed than the qlms algorithm when they both reach a similar steady state . ] ] for this case , length of the unknown fir filter is reduced to @xmath70 , still with @xmath66 active taps .
the parameters are : step size @xmath62 is @xmath71 and the value of @xmath67 is @xmath71 .
the results are shown in fig .
[ fig : learning_curve_2 ] .
again we see that the za - qlms algorithm has a faster convergence speed and has even converged to a lower steady state error in this specific scenario . ]
in this paper , a quaternion - valued adaptive algorithm has been proposed for more efficient identification of unknown sparse systems .
it is derived by introducing an @xmath0 penalty term in the original cost function and the resultant zero - attracting quaternion - valued lms algorithm can achieve a faster convergence rate by incorporating the sparsity information of the system into the update process .
simulation results have been provided to show the effectiveness of the new algorithm .
yilun chen , yuantao gu , and alfred o hero , `` sparse for system identification , '' in _ acoustics , speech and signal processing , 2009 .
icassp 2009 .
ieee international conference on_. ieee , 2009 , pp .
31253128 .
zhang , w. liu , y.g .
xu , and z.w .
liu , `` quaternion - based worst case constrained beamformer based on electromagnetic vectoe - sensor arrays , '' in _ proc .
ieee international conference on acoustics , speech , and signal processing _ , vancouver , canada , may 2013 , pp .
41496153 . x. r. zhang , w. liu , y. g. xu , and z. w. liu , `` quaternion - valued robust adaptive beamformer for electromagnetic vector - sensor arrays with worst - case constraint , '' , vol .
104 , pp . 274283 ,
november 2014 .
m. b. hawes and w. liu , `` a quaternion - valued reweighted minimisation approach to sparse vector sensor array design , '' in _ proc . of the international conference on digital signal processing _
, hong kong , august 2014 .
m. d. jiang , w. liu , y. li , and x. r. zhang , `` frequency - domain quaternion - valued adaptive filtering and its application to wind profile prediction , '' in _ proc .
of the ieee tencon conference _ , xian , china , october 2013 .
m. d. jiang , w. liu , and y. li , `` a general quaternion - valued gradient operator and its applications to computational fluid dynamics and adaptive beamforming , '' in _ proc .
of the international conference on digital signal processing _
, hong kong , august 2014 . | recently , quaternion - valued signal processing has received more and more attention . in this paper ,
the quaternion - valued sparse system identification problem is studied for the first time and a zero - attracting quaternion - valued least mean square ( lms ) algorithm is derived by considering the @xmath0 norm of the quaternion - valued adaptive weight vector . by incorporating the sparsity information of the system into the update process , a faster convergence speed is achieved , as verified by simulation results . * _ keywords : _ * quaternion ; sparsity ; system identification ; adaptive filtering ; lms algorithm . |
this paper links themes evoked in two classic papers , one in mathematics @xcite and the other in physics @xcite .
the first of these poses the question as to whether the spectral content of the radiation from a body can reveal its shape .
the second shows that the use of negatively - refracting metamaterials in a plane slab can lead to a super - resolving perfect lens , also known as a superlens .
we will consider a two - dimensional composite material , composed of polygonal inclusions made of a metamaterial ( by which we mean an artificial material with a dielectric contant which has a negative real part and a very small imaginary part ) and placed in a positive dielectric matrix material .
we will show that , in the spirit of pendry , the metamaterial makes possible resolution of an important structural feature of the inclusions , irrespective of how much smaller than the wavelength they are .
we will also show that , in the spirit of kac , this feature relates to the shape of the inclusion , being in fact the corner angle of the polygon , and that it is deduced from spectral measurements on the composite .
the fact that a spectral feature could be determined by corner shape , independent of ( say ) the area fraction of inclusions , was first suggested by hetherington and thorpe @xcite , on the basis of an elegant argument and numerical evidence for dilute composites .
we will base our demonstration firstly on analytic results relating to the spectrum of the effective dielectric permittivity function @xmath0 of the composite material , and secondly on remarkably accurate numerical results for this spectrum obtained using a new technique .
note here that we are using the word `` spectrum '' in two related , but slightly different senses . in the previous paragraph
, its usage meant that the absorption of electromagnetic waves by the composite was being determined as a function of wavelength , ranging over an appropriately - wide band . in the first sentence of this paragraph we referred to the distribution of singularities of the function @xmath0 , giving the effective permittivity of a composite having a specified geometry as a function of the ratio @xmath1 of the permittivities of inclusions and matrix .
the relation between these usages is that , as wavelength varies so does the ratio @xmath2 , so that measurements of ( say ) optical absorption by a composite over a suitable wavelength interval can reveal details of the singularities of the function @xmath3 . the numerical results for the singularity spectrum of @xmath3 reveal that it has a continuous part which runs between upper and lower limits of @xmath2 which do not vary at all with the area fraction of the inclusions .
it is complemented by a discrete spectrum of poles which does evolve with area fraction .
this evolution is in fact necessary , since the continuous spectrum for touching square inclusions in a checkerboard arrangement occupies the entire negative real axis of @xmath2 , but for non - touching square inclusions is confined to the interval @xmath4 .
the animations we give show how this transformation is achieved : the discrete spectrum becomes more and more dense as the touching configuration is approached , to supply the required spectral extension , as anticipated by one of the authors @xcite .
the results we give here are interesting in the insights they give into the connection between metamaterials and super - resolution .
they are also important in furthering our understanding of the connection between inclusion shape , geometrical arrangement and spectral properties of the effective permittivity function .
this connection helps in the design of structures having enhanced absorption over a wide wavelength range for applications in photothermal or photovoltaic captors @xcite , or offering strongly enhanced local fields for applications like sensing or nonlinear optical elements . we give a brief overview in section [ sec : over ] of some of the important properties of the function @xmath5 , including analytic results relating to the continuous part of the spectrum and some numerical investigations of both the discrete and continuous parts of the spectrum . in section [ sec : numer ] , we describe the method which enables accurate calculation of the spectrum of @xmath0 , and give numerical results illustrating its convergence , even on the negative real axis of @xmath2 . in section [ sec : anim ] , we discuss the animations which are given in the supplementary material to this paper , and the physical consequences of the behaviour they show .
we give a discussion and concluding remarks in section [ sec : disc ] .
we will now give a concise review of what is known about the properties of the effective dielectric permittivity function @xmath0 , for composites made of two materials with dielectric permittivities @xmath6 and @xmath7 , with the former corresponding to a disconnected inclusion phase and the latter to a continuous matrix phase .
we assume the geometry has cubic symmetry ( in three dimensions ) or square symmetry ( in two dimensions ) so that @xmath0 is scalar valued , i.e. the effective dielectric tensor equals @xmath8 .
this review builds on that given in perrins and mcphedran @xcite .
the calculation of @xmath0 for a given geometry is homogeneous of degree 1 in the variables @xmath6 and @xmath7 , so we can rescale to make @xmath0 a function of a single complex variable , the permittivity ratio @xmath1 : @xmath9 from now on we let @xmath0 be an abbreviated notation for the effective dielectric permittivity function @xmath10 and we let @xmath5 denote its scaled counterpart , the effective relative dielectric permittivity function , as defined via ( [ eq1 ] ) . to determine @xmath5 for a given geometry , we need to solve an electrostatic transport problem repeatedly for various @xmath2 .
more precisely , we need to solve laplace s equation for the potential @xmath11 on a periodic domain with a periodic electric field @xmath12 having a prescribed average value , and boundary conditions of continuity of @xmath11 and its normal flux @xmath13 at interfaces between materials . the theory of the function @xmath5 becomes particularly elegant when we deal with two - dimensional problems , in which @xmath14 becomes a function in the plane .
we may then apply the apparatus of complex - variable theory to the calculation of @xmath11 , and thus to @xmath5 . for the case of a doubly - periodic array of inclusions @xmath15 with unit cell @xmath16 , and square symmetry ,
the effective permittivity may be defined as @xcite @xmath17 where the integral in the numerator includes contributions @xmath18 from the inclusion region and @xmath19 from the matrix region . except for occasional comments on effective permittivity in three - dimensions , we will concentrate on two dimensions , which corresponds to arrays of cylinders of arbitrary cross - section , with the average field aligned in the @xmath20 plane
. the area fractions of the two phases will be denoted @xmath21 and @xmath22 .
since the geometry has square symmetry , keller s theorem @xcite gives @xmath23 this equation then pairs zeros of @xmath5 at values @xmath24 with poles at values @xmath25 .
of course , from ( [ eq2 ] ) , zeros of @xmath5 require that the contributions @xmath18 and @xmath19 add up to zero . since , with @xmath26 , @xmath19 is real and positive , this means @xmath18 must be real and negative , and so @xmath27 must be real and negative at any zero of @xmath5 , and thus at any pole as well .
bergman @xcite proved this property for both two - dimensional and three - dimensional composites .
bergman also recognized that even though the transport problem when the average electric field @xmath12 is prescribed does not have a solution at a pole , it should have a solution at a pole when instead the average displacement field @xmath28 is prescribed .
milton @xcite proved that for the value @xmath29 the electrostatic problem of an array of circular cylinders , with either a prescribed value of the average electric field @xmath12 or a prescribed average value of the displacement field @xmath28 , does not have a solution , compare with the discussion of ( [ eq : int1 ] ) below .
this suggests that @xmath29 is either an essential singularity or lies on a branch - cut of the function @xmath5 for arrays of circular cylinders .
the fact that branch - cuts can not be in the upper or lower half - planes , but must lie exactly on the negative real axis of @xmath2 was proved by one of the authors @xcite , using the relationship between composite materials and resistor networks .
a rigorous justification of the spectral representation for @xmath5 was given by golden and papanicolaou @xcite .
in general the function @xmath5 has the representation @xmath30 where @xmath31 and the spectral measure @xmath32 are non - negative .
the support of @xmath32 is the spectrum .
the spectral measure can be recovered from the values that the imaginary part of @xmath5 takes near the negative real @xmath2-axis since the integral of any smooth test function @xmath33 with respect to the measure @xmath32 is given by @xmath34 the discrete spectrum of @xmath5 is readily exhibited numerically .
this has been done for arrays of spheres by bergman @xcite and for arrays of circular cylinders by mcphedran and mckenzie @xcite , with both studies showing that the poles and zeros of @xmath5 converge to an essential singularity at @xmath29 .
we focus now on what can be said about the continuous spectrum of @xmath5 .
one simple geometry for which a result is immediately apparent is the checkerboard , for which dykhne @xcite obtained from keller s theorem ( [ eq2 ] ) the exact result @xmath35 this then exhibits a branch cut along the entire negative real axis of @xmath2 .
an exact result can also be obtained for the polarizability @xmath36 of a pair of touching cylinders @xcite .
using an inversion of coordinates about the contact point , the touching cylinders may be transformed into a slab of matrix material with permittivity @xmath26 surrounded by two half - planes filled with material with permittivity @xmath37 .
introducing the parameter @xmath38 it is easy to show using the method of images that the polarizability for a pair of touching cylinders of radius @xmath39 for the case of the applied field parallel to the line connecting cylinder centres is @xmath40 we see that the series in ( [ dilog1 ] ) converges provided @xmath41 , i.e. for real @xmath2 , @xmath42 .
however , we can obtain a meaningful result even when this is not the case by the technique of analytic continuation , since the series in ( [ dilog1 ] ) is a known transcendental function , called the dilogarithm , and denoted @xmath43 .
thus , we can replace ( [ dilog1 ] ) by @xmath44 the properties of the dilogarithm function are that it has a branch cut running from @xmath45 to @xmath46 , across which the discontinuity in the imaginary part of @xmath47 is @xmath48 $ ] . the branch cut in the plane of relative permittivity runs from @xmath49 to @xmath29 .
if the direction of the applied field is perpendicular to the line of centres , the branch cut runs from @xmath29 to @xmath50 .
we next consider arrays of square inclusions , for which we have already mentioned the dykhne result ( [ dyk ] ) .
a generalization of this for an array in which the square unit cell was divided into four equal squares with dielectric permittivities @xmath6 , @xmath7 , @xmath51 , and @xmath52 , was conjectured by mortola and steff @xcite : @xmath53^{1/2 } \left(\frac{\epsilon_1\epsilon_2\epsilon_3+\epsilon_1\epsilon_2\epsilon_4 + \epsilon_1\epsilon_3\epsilon_4+\epsilon_2\epsilon_3\epsilon_4 } { \epsilon_1+\epsilon_2+\epsilon_3+\epsilon_4}\right)^{1/2}\ , .
\label{eq3a}\ ] ] this conjecture was proved independently by milton @xcite and craster and obnosov @xcite .
we will consider a particular sub - case of this result , due to obnosov @xcite : an array of square cylinders with area fractions @xmath54 and @xmath55 , for which ( [ eq3a ] ) gives @xmath56 this formula yields a spectrum consisting solely of a branch - cut running from @xmath57 to @xmath58 .
we compare the result given by this formula for the real part of @xmath5 with the result of a numerical mode matching procedure in fig .
this comparison reveals the difficulty of evincing details of the spectrum using numerical methods : the mode matching technique approximates the branch cut by a discrete set of poles , which becomes more dense as the number of modes increases .
however , it is difficult to distinguish between branch cuts and sets of poles concentrating around an essential singularity by such methods .
furthermore , the mode matching method failed to give clear indications of the spectrum for area fractions of cylinders distinct from @xmath54 . ) for the effective relative dielectric permittivity of the square array of square cylinders , while the red curve gives the result of a mode - matching method @xcite.,width=453 ] the behaviour of fields near corners of inclusions in a matrix of differing dielectric permittivity and magnetic permeability was treated in electromagnetism by j. meixner @xcite . as well as obtaining the exponents which characterize the singularity of the field components near the edge , meixner pointed out that exactly the same formulae could be applied in electrostatics and magnetostatics .
hetherington and thorpe @xcite analysed the behaviour of fields near corners in electrostatics and magnetostatics , apparently without knowledge of meixner s paper .
they however pointed out the link between field behaviour near corners and the nature of the singularity spectrum in composites containing inclusions with corners .
let us suppose that the electrostatic potential varies with distance @xmath59 from a corner with an included angle @xmath60 as @xmath61 , and that the permittivity ratio inside the inclusion to that outside is @xmath2
. then @xcite @xmath62 is found by solving a transcendental equation : @xmath63}{\tan ( \beta \psi)}=-\sigma\ , , \label{transeq}\ ] ] or the same equation with @xmath2 replaced by @xmath64 . for @xmath2 real ,
the solution @xmath62 of ( [ transeq ] ) is either pure real or pure imaginary . to clarify the behaviour of fields ,
consider the case of @xmath65 , corresponding to a @xmath66 corner .
then ( [ transeq ] ) gives @xmath67 from this , we see that the right - hand side exceeds one in magnitude for @xmath2 lying between @xmath68 and @xmath69 , and that @xmath62 is then pure imaginary , corresponding to a solution for the potential which oscillates with @xmath59 , and the oscillations become ever more rapid as @xmath59 tends to zero . the electric field is given by the spatial derivative of the potential , and it diverges like @xmath70 multiplied by the oscillating term as @xmath71 .
the branch cut region here is that where @xmath62 is imaginary , i.e. , between @xmath57 and @xmath58 .
hetherington and thorpe @xcite postulate that for an @xmath72 sided regular polygon , there will be a branch cut running between @xmath73 and @xmath74 .
they made this assertion after recognizing that in this interval of @xmath2 the surface charge ( rather that just the surface charge density ) near the corner is infinite , which is unphysical .
another argument for the position of this branch cut was put forward by one of the authors @xcite . in order that @xmath0 have a significant imaginary part when @xmath75 has a very small imaginary part we see from ( [ epsdef ] ) that the electric field @xmath76 must be close to losing its square integrability . from the asymptotic form of @xmath11 near the corner
one finds that this happens when the imaginary part of @xmath2 is very small and real part of @xmath2 is between @xmath77 and @xmath78 . to elucidate this further for the case of a @xmath66 corner we solve ( [ squcorn ] ) with @xmath79 and @xmath80 where @xmath81 and @xmath82 are real , @xmath83 , and @xmath82 and @xmath84 are both very small and positive
this gives @xmath85 using polar coordinates @xmath86 near the corner , the potential @xmath11 scales as @xmath87 as @xmath88 , and so @xmath89 will be close to @xmath90 for some function @xmath91 .
it follows that with @xmath37 and @xmath26 the imaginary part of @xmath18 ( which is a measure of the power dissipation in the composite ) has a contribution near the corner from inside the radius @xmath92 of @xmath93 where the integral over @xmath94 is only over those angles in the inclusion .
thus , provided @xmath91 is non - zero , the contribution of the corner to the imaginary part of @xmath18 remains non - zero even in the limit @xmath95 ( and goes to zero when @xmath96 or @xmath97 since then @xmath98 ) .
a corner is not the only geometric feature which can act as a significant energy absorber when the imaginary part of the dielectric constant goes to zero .
the center of a sphere with a dielectric constant @xmath99 in the radial direction and dielectric constant @xmath100 in the tangential direction acts as an absorber when @xmath101 approaches real values less than @xmath102 : see figure 4 in the paper of qui and lukyanchuk @xcite ( which shows that this energy absorbing feature extends beyond the quasistatic limit ) and see also the related discussion on page 239 of @xcite ( where there is an error as the limit @xmath103 should have been taken rather than the limit @xmath104 ) .
we now describe a numerical method stable and accurate enough to verify the conjecture of hetherington and thorpe for the case @xmath105 , and to show the spectral evolution as a function of area fraction for composites with square inclusions .
laplace s equation is to be solved on a doubly periodic domain @xmath15 .
the boundary conditions on the positively oriented interface @xmath106 between the inclusion phase and the matrix phase are given in section [ sec : over ] .
an average electric field @xmath107 of unit strength is applied .
the permittivity of the matrix phase is set to @xmath26 so that the effective permittivity is equal to the effective relative permittivity . from the repeated solution to this problem for various @xmath37 we obtain @xmath5 .
two types of domains are investigated : the `` square array of square cylinders '' and the `` staggered array of square cylinders '' ( a square array of cylinders with diamond - shaped cross - sections , as studied for example in @xcite ) , see fig . [
fig : boards ] . of a square array of square cylinders with area fraction @xmath108 and a unit cell @xmath109 with a 16-panel coarse mesh on @xmath110 .
right : the same thing for a staggered array of square cylinders , but with @xmath111 and a 32-panel coarse mesh on @xmath110.,title="fig:",width=132 ] of a square array of square cylinders with area fraction @xmath108 and a unit cell @xmath109 with a 16-panel coarse mesh on @xmath110 .
right : the same thing for a staggered array of square cylinders , but with @xmath111 and a 32-panel coarse mesh on @xmath110.,title="fig:",width=132 ] of a square array of square cylinders with area fraction @xmath108 and a unit cell @xmath109 with a 16-panel coarse mesh on @xmath110 .
right : the same thing for a staggered array of square cylinders , but with @xmath111 and a 32-panel coarse mesh on @xmath110.,title="fig:",width=132 ] of a square array of square cylinders with area fraction @xmath108 and a unit cell @xmath109 with a 16-panel coarse mesh on @xmath110 .
right : the same thing for a staggered array of square cylinders , but with @xmath111 and a 32-panel coarse mesh on @xmath110.,title="fig:",width=132 ] boundary value problems on domains involving sharp corners may require extreme resolution close to corner vertices , even when the demands for overall accuracy are moderate .
one has to be selective with the choice of numerical method .
we chose an integral equation based scheme .
such schemes have the advantage that they can retain stability also in very difficult situations .
our particular choice of integral equation is standard a single - layer equation @xcite . for its solution
we use a novel numerical method called _ recursive compressed inverse preconditioning_. conceptually this is a local multilevel technique which makes a change of basis and expresses the non - smooth solution to the single - layer equation in terms of a piecewise smooth transformed layer density which can be cheaply resolved by polynomials .
discretization leads to a block diagonal transformation matrix @xmath112 ( an inverse preconditioner ) where the columns of a particular block can be interpreted as special basis functions for the original density in the vicinity of a corner vertex multiplied with suitable quadrature weights .
the blocks of @xmath113 are constructed in a fast recursion , @xmath114 , where step @xmath115 inverts and compresses contributions to @xmath113 involving the outermost quadrature panels on level @xmath115 of a locally @xmath116-ply refined mesh .
we emphasize that the method is strictly numerical and fully automatic .
there is no separation of variables or eigenvalue analysis involved .
the recursive compressed inverse preconditioning method was originally described in ref .
@xcite and further developed in refs .
[ sec : appa ] below highlights some of the method s features , relevant to the domain @xmath15 . a fuller description will be included in a forthcoming paper @xcite . .
the curves are supported by 826 adaptively spaced data points ( not all values shown due to the setting of the axes ) .
right : the relative error with ( [ eq4 ] ) as reference value.,title="fig:",height=226 ] . the curves are supported by 826 adaptively spaced data points ( not all values shown due to the setting of the axes ) .
right : the relative error with ( [ eq4 ] ) as reference value.,title="fig:",height=226 ] we first compute @xmath5 for the square array of squares at @xmath54 in the limit of @xmath2 approaching the negative real axis from the upper half - plane @xmath117 , as in the example of fig . 1 . the exact result ( [ eq4 ] ) is used as a benchmark .
[ fig : sqsq25 ] shows that the relative error is close to machine epsilon ( the upper bound due to rounding in floating point arithmetic ) except for in a neighbourhood of three points where it is higher : the ends of the branch cuts at @xmath57 and @xmath58 , and at the singularity of the integral equation at @xmath29 .
this demonstrates that the problem of computing @xmath5 for arrays of square cylinders is well conditioned in general and that our scheme is stable . .
the curves are supported by 2006 adaptively spaced data points .
right : the absolute difference between the left- and the right hand side of ( [ eq2b]).,title="fig:",height=226 ] .
the curves are supported by 2006 adaptively spaced data points .
right : the absolute difference between the left- and the right hand side of ( [ eq2b]).,title="fig:",height=226 ] the staggered array of square cylinders at @xmath118 is a more challenging geometry than the square array of square cylinders at @xmath54 : * there are more length scales involved , * @xmath5 varies more rapidly and has more poles and zeros , * there is no exact result to compare with .
the first problem is the least difficult
. the multilevel property of our numerical method should enable the resolution of almost arbitrarily small separation distances between corner vertices .
the second problem is more severe .
for @xmath5 close to zero , one can expect cancellation in ( [ eq : eff1 ] ) and the relative accuracy should suffer .
furthermore , it is harder to resolve wildly varying functions in floating point arithmetic than slowly varying ones .
the third problem is solved by using the extent to which ( [ eq2 ] ) is satisfied as an indicator of the relative error . for this , since @xmath2 and @xmath64 lie on different sides of the real axis and our numerical method takes limits from @xmath117 , we use ( [ eq2 ] ) in the equivalent form @xmath119 where the ` @xmath120 ' symbol to denotes complex conjugation .
[ fig : stsqe ] suggests that despite the difficulties , typically , only a few digits are lost compared to the square array at @xmath54 .
the numerics seem to give a relative precision of at least @xmath121 even for the most extreme values of @xmath122 . , but
@xmath2 is multiplied with a complex constant @xmath123 .
the curves are supported by 3435 data points .
the red curve is drawn on top of the blue curve.,title="fig:",height=226 ] , but @xmath2 is multiplied with a complex constant @xmath123 . the curves are supported by 3435 data points .
the red curve is drawn on top of the blue curve.,title="fig:",height=226 ] for the interpretation of various limits it might also be of interest to study @xmath5 for @xmath2 some finite distance into @xmath117 . fig . [
fig : stsqi ] shows again the staggered array of square cylinders at @xmath118 , but unlike in fig .
[ fig : stsqe ] we have here interrupted the limit process at @xmath2 a relative distance of @xmath124 away from the real axis .
this section discusses the evolution of the variation of @xmath5 with permittivity ratio @xmath2 for the square array of square cylinders and the staggered array of square cylinders as the area fraction of squares , @xmath21 , varies .
the discussion relates to two animations , called _ animation 1 _ and _ animation 2 _ , which can be viewed at http://www.maths.lth.se/na/staff/helsing/animations.html . to facilitate viewing ,
each animation is available in four versions , denoted a , b , c , and d. the versions have the same content , but differ in image format and pixel resolution .
we first consider the evolution of the variation of @xmath122 as @xmath21 in the square array ranges from zero to unity .
this evolution is shown in animation 1 , from which a typical frame is given in fig .
[ fig4 - 1 ]
. the most important feature of animation 1 is quite clear : for @xmath2 real , and for all values of @xmath21 , and except at the poles , non - zero values of @xmath125 are confined to the interval @xmath126 , in accord with the suggestion of hetherington and thorpe @xcite .
of course , the value of this imaginary part is always positive , if we restrict ourselves to composites without gain ( for which the imaginary part would be always negative ) . below the area fraction of 0.25 , the real part of @xmath127 is positive , and it develops its first pole at this area fraction .
it is interesting to compare frames from animation 1 and the left image of fig .
[ fig : sqsq25 ] with fig .
[ fig1 ] ; the results of the mode - matching method clearly correspond to those of the new method , but are capable of a resolution limited by the number of terms employed in field expansions . for @xmath128 , the real part is negative between @xmath57 and @xmath58 , while the pole migrates to more negative values of @xmath2 . for area fractions near @xmath129
( see fig .
[ fig4 - 1 ] ) , `` features '' which we call quasipoles develop from near @xmath29 , and one moves towards @xmath57 , while the other moves towards @xmath58 . when they reach these values , and then are not muted by the absorbing nature of the corners , they transform into actual poles , which move towards @xmath49 and @xmath50 respectively . at higher values of area fraction , more quasipoles evolve from @xmath29 and give rise to additional actual poles when they emerge from the branch - cut region .
@xmath130 becomes small as @xmath131 , while @xmath132 tends towards @xmath2 , apart from the increasingly numerous but increasingly narrow pole regions . as a function of permittivity ratio @xmath2.,width=453 ] for the staggered array of square prisms ,
animation 2 illustrates the behaviour of @xmath5 as a function of @xmath2 for area fractions ranging from zero to 0.5 ( with the behaviour for @xmath21 in the range 0.5 to 1.0 following from that in the lower range using keller s theorem ( [ eq2 ] ) . as has been commented in section [ sec : over ] ,
the interesting question is how the branch - cut location ( from @xmath49 to @xmath50 ) for @xmath133 of equation ( [ dyk ] ) can be reconciled with that ( from @xmath57 to @xmath58 ) predicted by hetherington and thorpe @xcite for @xmath21 arbitrarily near @xmath134 .
a mechanism for this reconciliation was provided by one of the present authors @xcite : discrete sets of poles in @xmath135 and @xmath136 were predicted to become denser and denser as @xmath21 approached @xmath134 , thus extending the branch cut in the limit to that required by ( [ dyk ] ) .
the accuracy of this prediction is evident in animation 2 : poles develop from the quasipoles generated at @xmath29 , and move left and right into the embryonic branch - cut regions @xmath135 and @xmath136 . the left frame in fig .
4 shows a stage in this evolution where @xmath21 is very close to @xmath134 .
animation 2 makes the `` nursery role '' of the region around @xmath29 in the development of the spectrum much more evident ( due to larger amplitudes of the quasipoles ) than does animation 1 .
the sensitivity of the spectral details for the staggered array near the checkerboard configuration are very evident in animation 2 . as we have commented in section [ sec : over ] , the asymptotics of fields near corners are the same in electromagnetism as in electrostatics @xcite .
thus , attempts such as that in @xcite to model the transition from electromagnetically reflecting structures to electromagnetically transmitting structures as @xmath21 moves through 1/2 would require an adaptive and recursive method like that described in section [ sec : numer ] to be able to achieve sufficient accuracy .
in this paper we have brought together rigorous mathematical results with numerical investigations of unprecedented accuracy .
the latter have revealed the generality of the former , and have substantiated a conjecture of hetherington and thorpe @xcite in a striking and conclusive way .
we conclude by commenting further on how the arguments and results we have presented can be implemented in a practical demonstration of morphological super - resolution , uniting the ideas of kac @xcite and pendry @xcite .
such a demonstration would require the fabrication of a set of parallel cylinders with a square cross section ( or polygonal cross section ) .
the cylinders do not have to be arranged in a geometrically - perfect array , and they do not have to be densely packed .
they have to be made of a material which is essentially non - absorbing and with a negative permittivity or permeability .
these requirements suggest the set of cylinders be made of metal , of size 10 @xmath137 m or larger , and be probed with wavelengths far greater than the cylinder size in the far infrared or longer .
such cylinders are large by today s lithographic standards , and so it should be possible to accurately form their corners to achieve sub - wavelength accuracy . going into the far infrared region diminishes metallic loss from its value in the visible and near - infrared @xcite .
it is crucial that the metallic loss be very low , since the experimental signature we suggest be probed is enhanced absorption by a set of such cylinders over a wavelength interval in which the metal s permittivity ranges from say -1/3 to -3 ( scaled relative to the permittivity of a host dielectric in which the cylinders are embedded ) .
note that the enhanced absorption of incident radiation detected will increase as more lines of cylinders are added to the set . the experimental result which would indicate morphological super - resolution is an enhanced absorption for wavelengths far in excess of the cylinder size , switching on and off at geometrically - determined limits described above , independent of the arrangement and area fraction of the cylinders .
we stress however that such a demonstration would indicate the physical relevance of the ideas we have described for a particular system .
the mathematical results we have described are of course rigorous , and the numerical examples of them we have given are highly accurate , so our demonstration of super - resolution for wavelengths arbitrarily larger than the size of the particles probed does not rely for its validity on experimental support
. they may be applicable to governing equations other than the helmholtz equation , for which the ideas of metamaterials and their applications are currently being explored @xcite .
ross mcphedran acknowledges support from the australian research council s discovery projects and centre of excellence schemes .
graeme milton acknowledges support from the national science foundation through grant dms-0707978 .
to keep the notation short we make no distinction between points or vectors in a real plane @xmath138 and points in a complex plane @xmath139 .
all points will be denoted @xmath140 or @xmath141 .
the potential function @xmath142 in @xmath15 is represented as a sum of a driving term and a single - layer potential with density @xmath143 @xcite .
enforcement of the boundary conditions on @xmath106 leads to the fredholm second kind integral equation @xmath144 here @xmath145 is the outward unit normal of @xmath106 at @xmath140 , @xmath110 denotes the restriction of @xmath106 to @xmath146 , and @xmath147 is as in ( [ eq : lmb ] ) . note that , as @xmath148 we have @xmath149 and ( [ eq : int1 ] ) is no longer a second kind equation , but a first kind equation whose ( unique ) solvability is by no means guaranteed .
therefore one can say that @xmath29 is a _
singularity _ of ( [ eq : int1 ] ) .
once ( [ eq : int1 ] ) is solved for @xmath143 and under the assumption that the inclusions do not overlap the unit cell boundary , the effective relative permittivity in the direction of the applied electric field can be computed from @xmath150 depending on how the unit cell is chosen , the squares in the staggered array may overlap the unit cell boundary . with the choice in fig .
[ fig : boards ] , they certainly do .
but since @xmath143 is a periodic function and identical on all squares one can circumvent this problem by modifying ( [ eq : eff1 ] ) so that it integrates @xmath143 twice on the square at the center of the unit cell and ignores @xmath143 on the other squares . we discretize ( [ eq : int1 ] ) and ( [ eq : eff1 ] ) using a nystrm method based on composite 16-point polynomial interpolatory quadrature and a parametrization @xmath151 of @xmath106 .
the parameter @xmath152 is real .
see ref .
@xcite for a review of nystrm methods including error analysis .
an initial coarse mesh that resolves the kernel of the integral operator in ( [ eq : int1 ] ) away from the corner vertices is constructed on @xmath106 , see fig .
[ fig : boards ] .
the coarse mesh is refined by subdividing those panels that neighbour corner vertices .
the subdivision is done @xmath116 times in a direction towards the vertices .
on quadrature panels which neighbour corner vertices we choose quadrature nodes according to the zeros of certain jacobi polynomials . on remaining panels we choose quadrature nodes according to the zeros of legendre polynomials .
upon discretization on the refined mesh ( [ eq : int1 ] ) assumes the form @xmath153 where @xmath154 and @xmath155 are square matrices and @xmath156 and @xmath157 are column vectors .
the vector @xmath157 corresponds to the discretization of the piecewise smooth right hand side .
now the kernel @xmath158 of the integral operator in ( [ eq : int1 ] ) is split into two functions @xmath159 where @xmath160 takes care of corner interaction and @xmath161 can be viewed as the kernel of a compact integral operator .
the kernel split ( [ eq : split ] ) corresponds to an operator split and the change of variables @xmath162 makes ( [ eq : gen1 ] ) read @xmath163 this right - preconditioned equation corresponds to the discretization of a fredholm second kind equation with compact operators .
the solution @xmath164 is the discretization of a piecewise smooth function .
the matrix @xmath165 , the density @xmath164 , and the right hand side @xmath166 in ( [ eq : gen2 ] ) can be evaluated on the coarse mesh without the loss of precision .
only @xmath167 needs the refined mesh for its accurate evaluation .
this enables a compression of ( [ eq : gen2 ] ) .
we introduce the compressed weighted inverse @xmath168 here @xmath169 is an unweighted prolongation operator that performs panelwise 15th - degree polynomial interpolation in the parameter @xmath152 ( which as we recall parameterizes @xmath106 through @xmath151 ) from points on the coarse mesh to points on the fine mesh when acting on column vectors from the left .
@xmath170 is a weighted prolongation operator .
see section 5 of ref .
@xcite .
substitution of ( [ eq : r0 ] ) into ( [ eq : gen2 ] ) and the use of some relations between prolongation operators make ( [ eq : gen2 ] ) assume the form @xmath171 this equation , defined solely on the coarse mesh , will be used in our computations .
the construction of @xmath113 from its definition ( [ eq : r0 ] ) may be a costly and unstable operation when the refined mesh has many panels .
the number of subdivisions @xmath116 needed to reach a given accuracy may grow without bounds due to the singularities in @xmath143 that arise as @xmath2 approaches certain values .
fortunately , the construction of each block of @xmath113 , associated with a corner of the square array or with a corner - meet of the staggered array , can be greatly sped up and also stabilized via a recursion .
this recursion uses matrices @xmath172 on local meshes centered around corners or corner - meets .
it would be going too far to describe all the fine details of this procedure , but sections 3.2 and 3.3 of ref .
@xcite give a fairly good idea of how the recursion is set up in the present context .
a key step is the ( partial ) conversion of the recursion into a non - linear matrix equation .
this equation is solved using a variant of newton s method relying on numerical homotopy to approach purely negative @xmath2 from the upper half - plane @xmath117 .
the ratio @xmath2 , which enters into @xmath172 , is initially multiplied with a constant @xmath173 .
the imaginary part of @xmath174 is reduced with a factor of ten after each of the first 14 newton iterations .
then @xmath174 is set to unity and the iterations are continued until either a sharp convergence criterion is met or a total of 30 iterations is reached .
a full description is given in ref .
99 born m and wolf e 1980 principles of optics pergamon press , oxford pp .
620 - 627 .
guenneau s mcphedran r c enoch s movchan a b farhat m and nicorovici n a p the colours of cloaks _ journal of optics _ , * 13 * , 024014 ( 18 pp . ) | we investigate the optical properties of periodic composites containing metamaterial inclusions in a normal material matrix .
we consider the case where these inclusions have sharp corners , and following hetherington and thorpe , use analytic results to argue that it is then possible to deduce the shape of the corner ( its included angle ) by measurements of the absorptance of such composites when the scale size of the inclusions and period cell is much finer than the wavelength .
these analytic arguments are supported by highly accurate numerical results for the effective permittivity function of such composites as a function of the permittivity ratio of inclusions to matrix .
the results show that this function has a continuous spectral component with limits independent of the area fraction of inclusions , and with the same limits for both square and staggered square arrays . |
_ hill s equation _ @xmath0 is closely related with 1-dimensional projective and 2-dimensional centro - affine differential geometry .
if @xmath1 and @xmath2 are two linearly independent solutions of the hill equation then the ratio @xmath3 gives a map @xmath4 , a non - degenerate parametric curve in the projective line , and a different choice of solutions gives a projectively - equivalent curve .
this provides a one - to - one correspondence between projective equivalence classes of non - degenerate curves in @xmath5 and second order differential operators @xmath6 .
see @xcite for basics of 1-dimensional projective differential geometry and hill s equation .
one may lift this solution curve from @xmath5 to a star - shaped curve @xmath7 in the plane satisfying the same equation @xmath8 the lift is determined by the unit wronskian condition @xmath9=1,\ ] ] where @xmath10 $ ] is the area form ( that is , the determinant of two vectors ) . the curve @xmath11 is defined uniquely , up to linear area - preserving transformations .
assume that all solutions of the hill equation are @xmath12-periodic .
then the curve ( [ hill ] ) is centrally symmetric , @xmath13 , and @xmath14-periodic ; in particular , the potential @xmath15 is also @xmath12-periodic .
the quantity @xmath16 is called the _ lyapunov integral _ , it plays an important role in the study of hill s equation . for everywhere positive @xmath15
, the following inequality was proved in @xcite : @xmath17 with equality only for constant @xmath15 .
( let us mention in this regard a series of papers by guggenheimer @xcite@xcite on geometric theory of second - order differential equations , in particular , on hill s equations with coexisting periodic solutions . )
inequality ( [ petty ] ) is deduced in @xcite from the 2-dimensional _ blaschke - santalo inequality_. let @xmath18 be a smooth convex plane curve containing the origin in its interior .
fix an area form in the plane ; then the dual plane also acquires an area form .
the polar dual curve @xmath19 lies in the dual plane and consists of the covectors satisfying the two conditions @xmath20 where @xmath21 is the pairing between vectors and covectors .
the dual curve @xmath22 is also convex and star - shaped .
let @xmath23 and @xmath24 be the areas bounded by @xmath11 and @xmath22 , and assume that @xmath11 is centrally symmetric with respect to the origin . in this case , the 2-dimensional blaschke - santalo inequality states that @xmath25 with equality only when @xmath11 is a central ellipse ( the blaschke - santalo inequality holds for not necessarily origin - symmetric convex curves ; then one considers polar duality with respect to a special , santalo , point , the point that minimizes @xmath26 ) .
the product @xmath26 is a centro - affine invariant of @xmath11 .
see @xcite concerning the blaschke - santalo and related affine geometric inequalities .
the relation between inequalities ( [ petty ] ) and ( [ bssm ] ) is as follows . if @xmath27 then the curve @xmath11 is convex .
use the area form @xmath10 $ ] to identify the plane with its dual plane . under this identification , ( [ duality ] )
holds for @xmath28 .
we have @xmath29 since @xmath30=1 $ ] , and since @xmath31=k$ ] , one has : @xmath32 .
thus ( [ petty ] ) follows from ( [ bssm ] ) .
independently of @xcite , r. schwartz @xcite , in his study of a projectively natural flow on the space of diffeomorphisms of a circle , considered a diffeomorphism @xmath33 and proved the following _ average lemma _ : @xmath34 where @xmath35 is the schwarzian derivative of @xmath36 ( the schwarzian is real for @xmath37 ) .
see also @xcite where a similar inequality for a convex curve in @xmath38 is proved ; we do not dwell on this other inequality of r. schwartz here .
a stereographic projection from a point of a circle identifies the circle in the complex plane with the real projective line , and @xmath39 can be considered as a @xmath40-periodic curve in @xmath5 ( a different choice of the center of stereographic projection gives a projectively equivalent curve ) .
this curve corresponds to hill s equation whose potential , @xmath15 , can be reconstructed as the schwarzian derivative of the ratio of its solutions , see @xcite .
a computation reveals that inequality ( [ average ] ) has the same form as ( [ petty ] ) , but without the positivity assumption @xmath41 .
figure [ fronts ] , left , depicts a non - convex star - shaped curve @xmath11 .
the polar dual curve @xmath22 is still star - shaped in that no tangent line passes through the origin ; however it has cusps , corresponding to inflections of @xmath11 ( the points at which @xmath42 ) , and self - intersections , corresponding to double tangents of @xmath11 .
such singular curves are called _ wave fronts_. the area @xmath24 is defined as the integral of the 1-form @xmath43 over the wave front @xmath22 , oriented so that the tangent line turns in the positive sense .
the average lemma of schwartz can be interpreted as a 2-dimensional blaschke - santalo inequality for such star - shaped curves .
see @xcite for a version of blaschke - santalo inequality for not necessarily convex plane curves in terms of the support function , @xcite for a version of the blaschke - santalo inequality for compact sets , and @xcite for a functional blaschke - santalo inequality .
we provide a discretization of schwartz s inequality .
namely , we prove a version of inequalities ( [ petty ] ) , ( [ bssm ] ) and ( [ average ] ) for star - shaped , but not necessarily convex , polygons .
consider an origin - symmetric star - shaped @xmath44-gon in the plane with vertices @xmath45 in their cyclic order about the origin , such that @xmath46 and @xmath47=1 $ ] for all @xmath48 .
let @xmath49 $ ] ; the sequence @xmath50 is @xmath51-periodic .
set @xmath52 .
obviously , each @xmath50 , and hence @xmath53 , is invariant under the action of @xmath54 on polygons .
[ discr1 ] one has : @xmath55 with equality only for the @xmath54-equivalence class of regular polygons .
consider an origin - symmetric @xmath44-gon @xmath56 with vertices @xmath57 .
this polygon may be self - intersecting , see figure [ polygons ] .
we shall see that @xmath56 is polar dual to @xmath58 .
using the same notation for areas as in ( [ bssm ] ) , the following polygonal blaschke - santalo inequality holds . and its dual @xmath56,width=384 ] [ bs ]
one has : @xmath59 with equality only for centro - affine regular polygons .
note that the limit @xmath60 of the right hand side of ( [ bsineq ] ) is @xmath61 , the right hand side of ( [ bssm ] ) .
theorem [ discr1 ] is reminiscent of another extremal property of regular polygons , in terms of their diagonal lengths , discovered by g. lk @xcite .
let @xmath45 be an @xmath51-gon and @xmath62 be fixed .
assume that @xmath63 for some constant @xmath64 and @xmath39 is an increasing concave function .
then @xmath65 with equality only for regular @xmath51-gons .
in particular , one has an upper bound on the average length of @xmath66-diagonals of a polygon : @xmath67 in the limit @xmath60 , one has a similar upper bound on the average chord length for smooth curves , see @xcite : @xmath68 with equality only for the round unit circle ; here @xmath69 is arc length parameter and the total length of @xmath11 is normalized to @xmath40 .
this inequality was used in @xcite to prove that many knot energies are uniquely minimized by round circles . in the spirit of theorem [ discr1 ]
, we propose to consider a centro - affine version of inequality ( [ chord ] ) .
let @xmath18 be an origin - symmetric star - shaped @xmath40-periodic curve such that @xmath70 , satisfying the unit wronskian condition ( [ wron ] ) , the centro - affine analog of arc length parameter . for @xmath71 , set @xmath72\ dt.\ ] ] [ conj ] for every @xmath73 , one has : @xmath74 with equality only for central ellipses . for infinitesimal @xmath73 , the taylor expansion up
to third order shows that ( [ conjineq ] ) implies inequality ( [ petty ] ) ( with @xmath75 ) .
thus conjecture [ conj ] is indeed a generalization of schwartz s average lemma ( [ average ] ) . in a sense
, one can solve the centro - affine parameterization equation ( [ wron ] ) .
let @xmath69 be the angular coordinate in @xmath76 ; then @xmath69 gives a parameterization of @xmath5 such that @xmath69 and @xmath77 correspond to the same point .
let @xmath36 be an orientation preserving diffeomorphism of @xmath5 which we consider as a diffeomorphism @xmath78 satisfying @xmath79
. then the curve @xmath80 satisfies ( [ wron ] ) , and all solutions are obtained this way ; see , e.g. , @xcite . conjecture [ conj ] can be reformulated as follows : @xmath81 for all diffeomorphisms @xmath39 as above and every @xmath71 , with equality only for projective diffeomorphisms of @xmath5 .
we prove a weak version of conjecture [ conj ] .
[ local ] for every @xmath71 , the central ellipses are local minima of the functional @xmath82 .
namely we shall show that the central ellipses form a critical 3-dimensional manifold of @xmath82 with a hessian , positive definite in the normal direction . in the spirit of @xcite
, one may consider the _
areal energy _ of a centro - affine parameterized curve @xmath18 , as considered above : @xmath83=\int_0^{\pi } \int_0^{\pi } g([\gamma(t),\gamma(t+\alpha)],\alpha)\ dt\ d\alpha,\ ] ] where @xmath84 is a function of two variables .
one conjectures that , for a broad class of functions @xmath84 , this areal energy @xmath85 $ ] is uniquely minimized by the central ellipses .
the content of the paper is as follows . in section [ proofs12 ]
we prove theorems [ discr1 ] and [ bs ] .
our proof of theorem [ discr1 ] is by way of morse theory on the space of equivalence classes of relevant polygons . describing this space
, we use some combinatorial formulas known in the theory of frieze patterns . in section [ proof4 ] , using fourier expansions of periodic functions , we prove theorem [ local ] .
the proof reduces to an infinite series of trigonometric inequalities .
section [ outerbill ] is an appendix devoted to a somewhat unexpected appearance of the blaschke - santalo and mahler inequalities , as well as the isoperimetric inequality in plane minkowski geometry , in the study of outer billiards , a geometrically natural dynamical system akin to the more familiar , inner , billiards . to avoid expanding this introduction any further
, we postpone the discussion of outer billiards until section [ outerbill ] .
denote by @xmath86 the space of origin - symmetric star - shaped @xmath44-gons @xmath87 satisfying @xmath46 and @xmath47=1 $ ] for all @xmath48 ; and let @xmath88 be its quotient space by @xmath54 .
[ mfld ] the spaces @xmath86 and @xmath88 are smooth @xmath51- and @xmath89-dimensional manifolds , respectively
. consider @xmath86 as a subvariety in @xmath90 defined by the conditions @xmath91 where @xmath92 $ ] .
let @xmath93 be a polygon in @xmath86 .
we want to show that our condition define a smooth submanifold : if @xmath94 at @xmath58 then all @xmath95 . consider a test tangent vector @xmath96 where @xmath97 is at @xmath66-th position
. then @xmath98.\ ] ] hence @xmath99 .
however , the vectors @xmath100 and @xmath101 are linearly independent , so @xmath102 .
this holds for all @xmath66 establishing the claim .
since @xmath54 acts freely on @xmath86 , the quotient space @xmath88 is an @xmath89-dimensional manifold .
@xmath103 [ conf ]
polygons in the projective line and in the affine plane .
one has a natural map from @xmath86 to @xmath104 , the configuration space of @xmath51 points in @xmath5 .
if @xmath51 is odd , this projection is a bijection on the connected component consisting of @xmath51-gons with winding number 1 .
indeed , let @xmath105 be such that the segments @xmath106 ( not containing other points @xmath107 ) cover the projective line once .
lift points @xmath108 to vectors @xmath109 so that @xmath110>0 $ ] for @xmath111 .
we want to rescale these vectors , @xmath112 , so that @xmath47=1 $ ] for @xmath113 , and @xmath114=1 $ ] .
this gives the system of equations @xmath115,\
\ i=0,\dots , n-2;\ \ t_{n-1}t_0=1/[u_0,u_{n-1}],\ ] ] that has a unique solution for odd @xmath51 .
this provides an inverse map @xmath116 .
however , if @xmath51 is even , the projection @xmath117 has a 1-dimensional fiber given by the scaling : @xmath118 the image of the projection @xmath117 has codimension 1 ; it is given by the condition @xmath119
= \prod_{i\ { \rm odd } } [ u_i , u_{i+1}],\ ] ] that does not depend on the lifting .
we interpret the cross - products @xmath50 as follows .
one can express each next vector @xmath120 as a linear combination the previous two , and the conditions @xmath121=[v_i , v_{i+1}]=1 , [ v_{i-1},v_{i+1}]=c_i$ ] imply that the coefficients are as follows : @xmath122 set : @xmath123 $ ] .
[ matrixlemma ] for @xmath124 , one has : @xmath125 one also has : @xmath126 equation ( [ matrix ] ) is proved by induction on @xmath127 .
the determinants satisfy the recurrence @xmath128 but , due to ( [ rec ] ) , the same recurrence holds for the cross - products : @xmath129=c_j[v_i , v_j]-[v_i , v_{j-1}],\ ] ] which makes it possible to use induction .
equation ( [ vform ] ) follows from the fact that both sides have the same cross - products with @xmath130 and @xmath131 .
@xmath103 [ closed ] one has : @xmath132 since @xmath133 and @xmath134 , equation ( [ mono ] ) follows from ( [ vform ] ) for @xmath135 and @xmath136 .
( there is a fourth condition , @xmath137 , but it follows from the fact that the monodromy map @xmath138 is area - preserving ) .
@xmath103 [ frieze ] frieze patterns .
the cross - products @xmath139 with fixed @xmath127 form the rows of a frieze pattern ; this means that the following equality holds : @xmath140 ( see @xcite concerning frieze patterns ) .
indeed , if @xmath141 then , using @xmath142=1 $ ] , one has : @xmath143 . on the other hand , the left hand side of ( [ fs ] ) is equal to @xmath144 , and ( [ fs ] ) follows .
thus the numbers @xmath50 form the first non - trivial row of a frieze pattern ( after a row of 0s and a row of 1s ) .
formula ( [ matrix ] ) can be found in @xcite too .
the functions @xmath50 serve as coordinates in @xmath88 .
they are not independent : they satisfy the three relations of corollary [ closed ] .
one can use formula ( [ vform ] ) to reconstruct the equivalence class of a polygon from @xmath50 ; this is used in the next lemma . but first consider the examples of @xmath145 and @xmath146 .
[ n45 ] @xmath145 : applying a transformation from @xmath54 , we may assume that @xmath147
. then @xmath148 .
the condition @xmath149=1 $ ] yields @xmath150 .
thus @xmath151 is the hyperbola @xmath152 .
the cross - products @xmath50 are as follows : @xmath153 and @xmath154 , which has the minimum for @xmath155 .
@xmath146 : once again , assume that @xmath156 .
let @xmath157 , and @xmath158 .
then the conditions @xmath149=[v_2,v_3]=1 $ ] yield @xmath159 .
the cross - products @xmath50 are as follows : @xmath160 and @xmath161 the only critical point of this function is @xmath162 , the golden ratio ; then @xmath163 for all @xmath48 .
consider the function @xmath164 .
[ proper ] @xmath53 is a proper function , that is , the set @xmath165 is compact for every positive constant @xmath166 .
consider a sequence of polygons @xmath167 in @xmath168 and let @xmath169 be the respective sequence of the cross - products @xmath170 $ ] .
since @xmath171 , we have @xmath172 , and hence , considering a subsequence if necessary , we may assume that @xmath173 as @xmath174 .
since @xmath175 , it follows from ( [ matrix ] ) that @xmath176 .
therefore , all @xmath177 are separated from 0 .
we can use the numbers @xmath177 to construct a polygon in @xmath88 which is the limit of the ( sub)sequence @xmath178 .
namely , choose two vectors , @xmath179 and @xmath180 with @xmath181=1 $ ] , and use the barred " version of recurrence ( [ vform ] ) to construct a polygon . the periodicity condition @xmath182 follows from the fact that equations ( [ mono ] ) still hold in the limit .
the resulting polygon is star - shaped because , in the limit , @xmath183 for @xmath111 as well .
@xmath103 now we describe the critical points of the function @xmath53 in @xmath88 .
[ crit ] a polygon @xmath93 is a critical point of the function @xmath53 if and only if all the cross - ratios are equal , @xmath184 , and the polygon is affine - regular .
assume that @xmath185 ( otherwise , see example [ n45 ] ) .
consider six consecutive vertices @xmath186 .
consider an infinitesimal deformation @xmath187 this deformation does not change the cross - products @xmath121 $ ] and @xmath188 $ ] . for
@xmath47 $ ] to remain the same in the linear approximation , one needs to have @xmath189 + \delta [ v_i , v_{i+2}]=0 $ ] .
hence @xmath190 where @xmath69 is an infinitesimal .
in particular , @xmath191 .
next one computes the rate of change of @xmath53 under the deformation ( [ deform ] ) .
this equals @xmath192+\delta [ v_{i-1},v_{i+2 } ] + \varepsilon [ v_{i-1},v_{i+2 } ] + \delta [ v_{i+2},v_{i+3 } ] = ( \varepsilon+\delta)(f_{i-1,i+2}+1).\ ] ] by ( [ matrix ] ) , @xmath193 , hence ( [ change ] ) equals @xmath194 this is zero if and only if @xmath195 .
therefore , if a point is critical then all @xmath50 are equal .
this is the case of the ( affine ) regular polygon .
it remains to check that ( an equivalence class of ) the regular polygon @xmath93 is a critical point of @xmath53 .
consider an infinitesimal deformation @xmath196 where @xmath197 .
since the deformation does not change @xmath47 $ ] , one has : @xmath198+[v_i , u_{i+1}]=0 $ ] , that is , @xmath199 . since all @xmath50 are equal , @xmath200 for all @xmath48 , one has @xmath201 , and in particular , @xmath202 .
finally , the rate of change of @xmath53 equals @xmath203&+[v_{i-1},u_{i+1}]=\\ a_{i-1 } [ v_{i-2},v_{i+1}]+b_{i-1}[v_{i},v_{i+1}]&+a_{i+1}[v_{i-1},v_{i}]+b_{i+1}[v_{i-1},v_{i+2 } ] = \\ a_{i-1 } ( c_{i-1}c_i-1)+b_{i-1}+a_{i+1}&+b_{i+1}(c_ic_{i+1}-1 ) = c^2 \sum(a_i+b_i ) = 0 . \end{split}\ ] ] thus @xmath204 .
@xmath103 now we can prove theorem [ discr1 ] .
fix a generic and sufficiently large constant @xmath166 and consider the manifold with boundary @xmath205 . by lemma
[ proper ] , @xmath168 is compact .
hence @xmath53 assumes minimum on it , and by lemma [ crit ] , this minimum corresponds to the affine - regular polygon .
the respective value of the function @xmath53 is @xmath206 .
this finishes the proof of theorem [ discr1 ] .
let @xmath93 be a star - shaped polygon in the plane .
the dual polygon @xmath207 in the dual plane is characterized by the equalities : @xmath208 for all @xmath48 . as in section
[ intro ] , we use the area form to identify the plane with its dual .
[ dual ] the dual polygon @xmath56 is given by @xmath209 , and its signed area satisfies @xmath210 .
one has : @xmath211=1\quad { \rm and } \quad [ v_{i+1}-v_i , v_{i+1}-v_i]=0,\ ] ] as needed .
next , @xmath212=\sum_1^{n } ( 2-c_i)=2n - f_n,\ ] ] as claimed .
@xmath103 now we prove theorem [ bs ] .
consider a centrally symmetric star - shaped @xmath44-gon @xmath58 satisfying the assumptions of the theorem .
one has @xmath213 , hence @xmath214 where the inequality in the middle follows from theorem [ discr1 ] .
first of all , we show that inequality ( [ intineq ] ) is indeed equivalent to ( [ conjineq ] ) .
[ eq ] for a curve @xmath11 as in ( [ param ] ) , one has : @xmath215 using complex notation , one has : @xmath216 hence @xmath217=(f'(t)f'(t+\alpha))^{-1/2 } [ e^{if(t ) } , e^{if(t+\alpha)}],\ ] ] and it remains to use the fact that @xmath218=\sin ( \psi-\phi)$ ] .
@xmath103 we use the formula of lemma [ eq ] to investigate the functional @xmath82 .
[ critical ] for each @xmath73 , the central ellipses are critical points of the functional @xmath82 . without loss of generality , assume that we are given a unit circle parameterized by the angle parameter @xmath69 , that is , @xmath219 .
consider an infinitesimal perturbation @xmath220 where @xmath84 is a @xmath221-periodic function . then @xmath222 denoting @xmath223 and @xmath224 by @xmath225 and @xmath226 respectively , one has @xmath227 and hence @xmath228 the last integral vanishes because @xmath229 as needed .
@xmath103 next we compute the hessian of the functional @xmath82 at function @xmath219 .
write : @xmath230 , where @xmath84 and @xmath231 are @xmath221-periodic , and use the same notation as in the proof of the preceding lemma .
[ hess ] one has : @xmath232 where terms of order 3 and higher in @xmath233 are suppressed .
the computation is similar to the previous proof , but this time , one considers taylor expansions in @xmath233 up to second order . from lemma [ critical ] we know that the linear term in @xmath233 in the expansion of @xmath82 vanishes .
the quadratic term in the integrand is as follows : @xmath234 @xmath235 as before , we simplify the integrals using equations ( [ zeros ] ) for function @xmath231 , similar equations for functions @xmath236 and @xmath237 , and integration by parts @xmath238 to obtain the stated result . @xmath103 next we consider the fourier expansion of the @xmath221-periodic function @xmath239 and express the hessian in terms of the fourier coefficients .
let @xmath240 [ fourier ] up to a multiplicative positive constant , the quadratic part of @xmath82 , as given in lemma [ hess ] , is as follows : @xmath241.\ ] ] first of all , we notice that the quadratic part of @xmath82 vanishes if @xmath84 is a constant .
hence we may assume that @xmath242 .
next , we have : @xmath243 now we use lemma [ hess ] and the fact that @xmath244 unless @xmath245 , in which case the integral equals @xmath40 ( we deal with even harmonics , hence we may take the limits in the integrals to be @xmath246 and @xmath221 ) . using this fact ,
each integral from lemma [ hess ] can be expressed in terms of the coefficients @xmath247 .
let us illustrate this for the term @xmath248 ; other cases are similar : @xmath249 @xmath250 collecting terms and canceling a common positive factor yields the result .
@xmath103 it remains to consider the function @xmath251 we observe that @xmath252 and @xmath253 .
the space of even harmonics of order @xmath254 is 3-dimensional .
this corresponds to the fact that the space of central ellipses is 3-dimensional : they constitute the @xmath54-orbit of the unit circle .
thus theorem [ local ] will be proved once we establish the following fact . for every even @xmath255 and @xmath256
, one has : @xmath257 ( see figure [ graphs ] ) . for @xmath258 and @xmath259,title="fig:",width=240 ] for @xmath258 and @xmath259,title="fig:",width=240 ] first , we show that @xmath260 if @xmath73 is sufficiently separated from @xmath246 and @xmath221 .
in fact , by symmetry , we may assume that @xmath261 .
one has an obvious inequality : @xmath262 consider @xmath263 using ( [ trig ] ) with @xmath264 , we have : @xmath265 thus , @xmath266 if @xmath267 , or @xmath268 increases when @xmath73 is sufficiently close to @xmath246 .
namely , let @xmath269 using some trigonometry , we see that @xmath270 if @xmath271 , or @xmath272 we want to show that ( [ trigineq ] ) holds for @xmath273 . to this end
, we use the inequality @xmath274 that can be easily proved by induction on @xmath51 .
hence , the left hand side of ( [ trigineq ] ) is greater than 1 . on the other hand ,
the right hand side of ( [ trigineq ] ) is less than 1 for all @xmath255 and @xmath273 .
thus ( [ trigineq ] ) holds .
finally , we need to show that the two above considered cases cover the whole interval of values of @xmath73 , that is , in view of ( [ cot ] ) , that @xmath275 for all @xmath255 .
indeed , the ratio of the right and left side of ( [ last ] ) increases with @xmath51 , and for @xmath145 , this ratio equals @xmath276 @xmath103 [ crcond ] critical curves of @xmath82 .
it is interesting to describe critical curves of the functional @xmath82 .
let @xmath18 be a @xmath14-periodic centrally symmetric curve , parameterized so that the unit wronskian condition ( [ wron ] ) holds .
let @xmath277 .
then @xmath11 is critical for @xmath82 in the class of curves satisfying ( [ wron ] ) if and only if @xmath278+[\gamma(t),\gamma_+'(t)-\gamma_-'(t)]=0\ ] ] for all @xmath69 .
we do not dwell on the proof , but let us mention that the infinitesimal perturbations of a curve , preserving the unit wronskian condition , are given by vector fields of the form @xmath279 where @xmath36 is an arbitrary smooth function satisfying @xmath280 .
equation ( [ cond ] ) holds for central ellipses ( each of the two cross - products vanishes ) , but we do not know whether central ellipses are the only curves satisfying this equation .
_ outer billiards _ ( a.k.a .
dual billiards ) is a discrete time dynamical system in the exterior of a planar convex domain ( outer billiard table ) defined by the following geometric construction .
let @xmath11 be the oriented outer billiard curve , the boundary of the outer billiard table , and let @xmath281 be a point in its exterior .
draw the tangent ray to @xmath11 from @xmath281 , whose orientation agrees with that of @xmath11 , and reflect @xmath281 in the tangency point to obtain a new point @xmath282 . the map @xmath283 is the outer billiard transformation , see figure [ outer ] . the map @xmath284 can be defined for convex polygons as well ( its domain is then an open dense subset of the exterior of the polygon ) .
see the article @xcite or the respective chapters of the books @xcite for a survey of outer billiards .
the monograph @xcite provides a profound study of outer billiards on a class of quadrilaterals called kites .
it was observed a long time ago that , after rescaling , the dynamics of the second iteration of the outer billiard map very far away from the outer billiard table is approximated by a continuous motion whose trajectories are closed centrally symmetric curves and which satisfies the second kepler law : the area swept by the position vector of a point depends linearly on time . without going into details that can be found in @xcite , here is an explanation of this phenomenon .
let @xmath18 be the outer billiard curve which we assume to be smooth and strictly convex .
consider the tangent line to @xmath18 .
there is another tangent line , parallel to that at @xmath18 ; let @xmath285 be the vector that connects the tangency points of the former and the latter . for points @xmath281 at great distance from @xmath11 and seen in the direction of @xmath286 from @xmath11 , the vector @xmath287 is almost equal to @xmath288 , see figure [ faraway ] .
we construct a homogeneous field of directions in the plane : along the ray generated by the vector @xmath286 , the direction of the field is that of the vector @xmath285 .
the trajectories of the second iteration of the outer billiard map at infinity " follow the integral curves of this field of directions .
these integral curves are all similar ; we denote them by @xmath289 .
a similar analysis can be performed when the outer billiard curve if a convex polygon , see , e.g. , @xcite .
for example , if @xmath11 is a triangle then @xmath289 is an affine - regular hexagon .
another example : if @xmath11 is a curve of constant width then @xmath289 is a circle .
if @xmath11 is a semi - circle then @xmath289 is a curve made of two symmetric arcs of orthogonal parabolas . clearly , @xmath292 has the direction of @xmath286 , and we need to check that @xmath293 is collinear with @xmath18 .
indeed , @xmath294 } - \frac{\gamma ' [ \gamma,\gamma'']}{[\gamma,\gamma']^2},\ ] ] hence @xmath295=0 $ ] .
@xmath103 the outer billiard motion at infinity " goes along the curve ( [ infcurve ] ) with the velocity vector at point @xmath292 being equal to @xmath296 .
this implies that @xmath297=2,\ ] ] which explains kepler s law .
furthermore , the curve ( [ infcurve ] ) is polar dual to @xmath11 : equations ( [ duality ] ) hold with @xmath298 ( providing another proof to lemma [ dualcurve ] ) . if @xmath11 is not centrally symmetric then @xmath289 is polar dual to the central symmetrization of @xmath11 , see @xcite .
the later curve , which we denote by @xmath299 , is the minkowski half - sum of @xmath11 and @xmath300 , its reflection in the origin . in other words ,
the support function of @xmath299 is given by the formula @xmath301 where @xmath302 is the support function of @xmath11 .
the curve @xmath299 is centrally symmetric and its width in every direction coincides with that of @xmath11 .
of course , if @xmath11 is centrally symmetric then @xmath303 .
the trajectories at infinity @xmath289 are defined only up to dilation .
fix one such curve , and let @xmath12 be the time it takes to traverse the curve moving with the velocity @xmath304 . scaling the curve @xmath289 by some factor , results in scaling @xmath12 by the same factor .
one can also scale the outer billiard curve @xmath11 : this results in scaling the speed by the same factor and the time by its reciprocal . to make the time scaling - independent , one multiplies @xmath12 by @xmath305 where , as before
, @xmath306 denotes the area bounded by a curve .
let us call the result of this scaling of @xmath12 the _ absolute time _
, and denote it by @xmath307 .
[ abstime ] for any outer billiard curve @xmath11 , the absolute time satisfies @xmath308 the upper bound is attained only for curves of constant width and their affine images ; the lower bound is attained only for parallelograms . if @xmath11 is a centrally symmetric @xmath44-gon then @xmath309 with equality only for affine - regular @xmath44-gons ; the same inequality holds for arbitrary @xmath51-gons . by the blaschke - santalo inequality , @xmath312 , with equality only if @xmath299 is a central ellipse , that is , if @xmath299 is affine equivalent to a circle .
but @xmath299 is a circle if and only if @xmath11 has constant width .
if @xmath11 is a centrally symmetric @xmath44-gon , the upper bound follows from theorem [ bs ] .
finally , if @xmath11 is an @xmath51-gon then @xmath299 is a centrally symmetric @xmath44-gon ( it is possible that @xmath299 has fewer than @xmath44 sides but this does not affect the inequality ) .
@xmath103 it is interesting to mention that outer billiards also
solve " the isoperimetric problem in minkoswki geometry .
let a centrally symmetric outer billiard curve @xmath11 be the unit circle of planar minkowki geometry .
then the trajectory at infinity @xmath289 is the unique solution to the isoperimetric problem in this minkowski geometry : according to busemann s theorem @xcite , the minkowski length of ( a homothetic copy of ) @xmath289 is minimal among the curves bounding a fixed area .
* acknowledgments*. it is a pleasure to thank j. c. alvarez , yu .
burago , m. ghomi , m. levi , e. lutwak , v. ovsienko , i. pak and especially r. schwartz for comments and suggestions .
the paper was written during my visit at brown university ; i am grateful to the department of mathematics for its hospitality . | r. schwartz s inequality provides an upper bound for the schwarzian derivative of a parameterization of a circle in the complex plane and on the potential of hill s equation with coexisting periodic solutions .
we prove a discrete version of this inequality and obtain a version of the planar blaschke - santalo inequality for not necessarily convex polygons .
we consider a centro - affine analog of lk s inequality for the average squared length of a chord subtending a fixed arc length of a curve the role of the squared length played by the area and prove that the central ellipses are local minima of the respective functionals on the space of star - shaped centrally symmetric curves .
we conjecture that the central ellipses are global minima . in an appendix
, we relate the blaschke - santalo and mahler inequalities with the asymptotic dynamics of outer billiards at infinity . |
triply excited states of atomic lithium represent the prototypical case of a highly correlated , multi - excited three - electron system in the presence of a nucleus , and thus typify the four - body coulombic problem .
since all the three electrons reside in higher shells leaving the k shell empty , these are often referred to as the _ hollow _ states .
as the one - step photogeneration of such a state requires coherent excitation of all the three electrons , these are usually more difficult to produce from the ground state by single photon absorption or electron impact excitation .
these , in addition to their proximity to more than one thresholds coupled with the presence of an infinite number of open channels associated with these resonances , have offered considerable challenges to both experimentalists as well as theoreticians . the recent advances in third - generation ,
extreme - uv synchrotron radiation sources as well as the availability of sophisticated and powerful quantum mechanical methodologies have stimulated an overwhelming amount of works in the last decade that shed significant light on the understanding of such systems with greater accuracy .
ever since the first observation of @xmath8 states of li in a beam - foil experiment [ 1 ] , numerous attempts [ 2 - 4 ] have been made over the years to use this technique which identified bound states such as the 2p@xmath9 @xmath6s@xmath5 besides many autoionizing states .
later , photoabsorption spectroscopy [ 5 ] using a dual laser plasma technique reported measurement of the lowest 2s@xmath32p @xmath3p@xmath5 resonance in li .
subsequently a large number of higher resonances were found and tentatively classified in a wider energy range ( 140165 ev ) [ 6 - 8 ] .
first high - precision photoelectron spectroscopic determination [ 9 ] of the partial photoionization cross sections generated tremendous interest to measure the resonance positions , widths of both even- and odd - parity hollow states having varied symmetries [ 10 - 14 ] with better resolution over wider energy regions .
some rydberg series [ 15 ] as well as the states with both k and l shell vacancies [ 8,13 ] have also been identified lately .
parallel to the experimental progresses , a substantial amount of theoretical works have also been done over the years for reliable , accurate prediction of energies , widths and lifetimes of these multi - excited states , varying in their complexity and accuracy .
one of the earliest such calculations for li - isoelectronic series was due to [ 16 ] employing a truncated diagonalization method ( tdm ) , based on the hydrogenic orbitals .
later other formalisms such as the 1/z expansion method [ 17 ] , many - body perturbation theory [ 18 ] were also used to investigate these states .
some other recent theoretical works include the state - specific theory [ 19 - 21 ] , the configuration interaction ( ci ) type approach [ 22 ] , joint saddle point ( sp ) and complex coordinate rotation ( ccr ) method [ 23 - 25 ] , the space partition and stabilization procedure [ 26 ] , a density functional formalism [ 27 ] , several variants of the r - matrix theory [ 28 - 31 ] as well as the tdm [ 32,33 ] , the hyperspherical coordinate approach [ 34,35 ] , etc .
although the most natural and commonest choice in these works was neutral atom ( li ) , other atomic systems such as the negative ions ( e.g. , he@xmath10 ) , positive ions like li - isoelectronic series or n@xmath11 [ 22 ] have also received considerable attention .
while majority of the available formalisms involve traditional wave - function - based calculations requiring large basis sets and often involving mixing of continuum states , a density - based formalism for triply excited states was proposed in [ 27 ] . in this approach
a kohn - sham ( ks ) type differential equation was solved to obtain the self - consistent set of orbitals and densities using the nonvariational local work - function based exchange potential [ 36 ] in conjunction with the nonlinear gradient- and laplacian - included correlation energy functional of lee , yang and parr ( lyp ) [ 37 ] .
fairly good quality results were reported for various closed and open shell many - electron atomic excited states including the single , double , triple as well as low and moderately high excitations ; core and valence states ; bound and autoionizing states besides the satellite states [ 38 - 43,27 ] .
this single - determinantal formalism was able to reproduce the excited state energies , radial densities , various expectation values , excitation energies ( 102000 ev ) as well as the small energy differences ( 0.0323.5 ev ) quite satisfactorily . however , all the above density functional theory ( dft)-based calculations of atomic excited states employed a numerov - type finite difference ( fd ) scheme to numerically solve the ks type equation . recently a generalized pseudospectral ( gps ) approach has been utilized to solve it in a more accurate and efficient manner [ 44 ] , as demonstrated for the singly and doubly excited high - lying rydberg states of he as well as the singly excited states of li and be .
the singly excited state energies were within 0.01% of the best literature data for he ( for other atoms it was less than 0.02% ) , while for the doubly excited states of he it was well within 3.6% .
the discrepancy in the calculated single and double excitation energies for 31 selected states were within 0.0090.632% and 0.0851.600% respectively .
the gps method has been shown to be a very powerful tool for accurate and physically meaningful results of both static and dynamic properties of coulombic singular systems such as atoms and molecules as well as other stronger singularities like the generalized spiked harmonic oscillators [ 44 - 48 ] . while in [ 27 ]
, results were presented for all the eight n=2 intrashell triply excited states of li isoelectronic series ( z=2,3,4,6,8,10 ) with reasonable accuracy , resonance series including the higher members were not considered and to the best of our knowledge , no other dft - based attempts are known so far for these systems .
the purpose of this work is therefore to explore and extend the regions of validity of the density - based approach to these challenging hollow states and also to study the spectra of these resonances in a detailed and elaborate way .
we thus focus on the energies and the excitation energies of twelve @xmath0n@xmath1 ( n@xmath22 ) odd- and even - parity doublet and quartet resonance series of li leading nearly to 270 states ( 2s@xmath3ns @xmath3s@xmath4 , 2s@xmath3np @xmath3p@xmath5 , 2s@xmath3nd @xmath3d@xmath4 , 2p@xmath3ns @xmath3d@xmath4,@xmath6p@xmath4 , 2s2pns @xmath6p@xmath5 , 2s2pnp @xmath6d@xmath4 , 2p@xmath3np @xmath3f@xmath5 , @xmath6d@xmath5 , 2p@xmath3nd @xmath3g@xmath4 , @xmath6f@xmath4 , 2s2pnd @xmath6f@xmath5 ) , up to a maximum of n=25 . as a further test , additional results
are also reported for 37 @xmath7n@xmath1 ( n@xmath23 ) doubly hollow states of li having vacancies in both the k and l shells up to n=6 , out of which seven are n=3 intrashell type .
comparisons are made with the experimental and other theoretical results , wherever possible .
the article is organized as follows : section ii presents a brief summary of the methodology as well as the computational aspects .
section iii makes a detailed discussion on the obtained results and a few conclusions are drawn in section iv .
since the work - function formalism for atomic excited states [ 38 - 43,27 ] as well as its gps implementation [ 44 ] has been discussed in detail previously , here we give only the essential steps in the calculation ( atomic units employed throughout the article , unless otherwise mentioned ) .
the local exchange potential with which the electrons move is interpreted [ 36,49 ] physically as the work done to move an electron from infinity to its position @xmath12 against the electric field @xmath13 arising out of the fermi - hole charge distribution , @xmath14 , and as such is given by the following line integral , @xmath15 this field @xmath13 , being representative of the pauli correlation , as its quantum mechanical source charge distribution is the pair correlation density , has the following form , @xmath16 this work against @xmath13 can be determined exactly as the fermi hole is known explicitly in terms of the single - particle orbitals , @xmath17 where @xmath18 denotes the single - particle density matrix spherically averaged over coordinates of the electrons of a given orbital angular momentum quantum number .
now within the central - field approximation , the orbitals expressed as @xmath19 ( @xmath20 denoting the usual spherical harmonics ) , give the total electron density as the sum of the occupied orbitals , @xmath21 and the spherically averaged radial component of the electric field is simplified as , @xmath22 now assuming that a unique local exchange potential exists for a given excited state , the following ks - type equation is solved , @xmath23 \phi_i({\mathrm{\mathbf{r } } } ) = \varepsilon_{\mathit{i } } \phi_{\mathit{i}}({\mathrm{\mathbf{r}}}),\ ] ] to obtain the self - consistent set of orbitals @xmath24 and the electron density .
here @xmath25 denotes the usual hartree electrostatic potential consisting of the electron - nuclear attraction and the interelectronic coulomb repulsion , @xmath26 and @xmath27 , the total exchange - correlation ( xc ) potential is partitioned as , @xmath28 while @xmath29 can be accurately determined as outlined above , the accurate form of @xmath30 valid for a general excited state is unknown as yet and therefore requires to be approximated .
there are several prescriptions available in the literature and the present work uses the widely used lyp functional [ 37 ] .
now we present a brief overview of the computational procedure used to solve the pertinent ks equation through the gps formalism .
one disquieting feature of the commonly used equal - spacing fd methods is that because of the existence of coulomb singularity at the origin as well as the long range nature of the coulomb potential , one requires a large number of grid points to achieve reasonably good accuracy even for the ground states .
however the gps method allows nonuniform and optimal spatial discretization maintaining similar accuracy at both small and large @xmath31 regions . thus one can work with a much lesser grid points having a denser mesh at small @xmath31 while a coarser mesh at large @xmath31 .
moreover it also has the attractive feature of possessing both the simplicity of direct fd or finite element methods and the fast convergence of the finite basis set methods .
one of the principal features of this scheme lies in the fact that a function @xmath32 defined in the interval @xmath33 $ ] can be approximated by the polynomial @xmath34 of order n so that , @xmath35 and the approximation is _ exact _ at the _ collocation points _ @xmath36 , i.e. , @xmath37 here we employ the legendre pseudospectral method using @xmath38 , @xmath39 , where @xmath40 are obtainable from the roots of the first derivative of the legendre polynomial @xmath41 with respect to @xmath42 , i.e. , @xmath43 the cardinal functions , @xmath44 in eq .
( 9 ) are given by the following expression , @xmath45 obeying the unique property @xmath46 .
now the semi - infinite domain @xmath47 $ ] is mapped into the finite domain @xmath33 $ ] by the transformation @xmath48 .
one can make use of the following algebraic nonlinear mapping [ 50,51 ] , @xmath49 where l and @xmath50 may be termed as the mapping parameters .
now , introducing the following relation , @xmath51 coupled with the symmetrization procedure [ 50,51 ] leads to the following coupled set of equations , @xmath52 a_j = ea_{j'},\ \ \ \ j=1,\ldots , n-1,\ ] ] where @xmath53^{1/2 } \psi(r(x_j))\ \left [ p_n(x_j)\right]^{-1}.\ ] ] and the symmetrized second derivative of the cardinal function , @xmath54 is given by , @xmath55^{-1 } d^{(2)}_{j'j } \left[r'(x_j)\right]^{-1},\ ] ] with @xmath56 note the advantage that this leads to a _
matrix eigenvalue problem which can be readily solved to give accurate eigenvalues and eigenfunctions . within the central - field approximation , taking the exchange potential in eq .
( 1 ) and the lyp correlation functional [ 37 ] , ks equation ( 6 ) is solved numerically in a self - consistent manner to obtain the orbitals .
these are then employed to construct the various slater determinants derived from a particular electronic configuration . in general , denoting the electronic energies of determinants and multiplets associated with a given electronic configuration by @xmath57 and @xmath58 respectively , the slater s diagonal sum rule [ 52,38 - 44,27 ] is used to calculate the @xmath58 as , @xmath59 for all the calculations reported in this work , a consistent convergence criteria of @xmath60 and @xmath61 a.u . , are used for the potential and energy respectively , whereas a maximum of 500 radial grid points sufficed to achieve convergence .
the calculated nonrelativistic term and excitation energies for various hollow states of li along with the available experimental and theoretical results are given in tables i through vii . in the literature , usually the excitation energies are reported , while the individual state energies are given in only few occasions . in a recent study of the low - lying singly excited states of several open - shell atoms ( b , c , o , f , na , mg , al ,
si , p and cl ) within the work - function formalism [ 42 ] , it was observed that the excitation energies from the exchange - only and the numerical hartree - fock ( hf ) calculations showed good agreement with each other .
interestingly however , the two correlation energy functionals ( relatively simpler local wigner and the nonlocal lyp ) did not show any considerable improvements in excitation energies , although the excited - state energies were improved significantly .
therefore we decided to present both these quantities for all the states considered in this work .
the state energies are given in atomic units , while the excitation energies in ev ( 1 a.u.=27.2076544 ev ) are given relative to the accurate ground state of li calculated from a full core plus correlation using the multiconfiguration interaction wave function , i.e. , @xmath62 a.u .
[ 54 ] , for a consistent comparison with the literature data .
here we note that the present calculated ground state energy of li is @xmath63 a.u . the independent particle model classification [ 32,33 ] is used throughout the article , where the six core li@xmath64 n=2 intrashell doubly excited states , _ viz . , _
2s@xmath3 @xmath65s@xmath4 , 2s2p @xmath9p@xmath5 , 2p@xmath3 @xmath9p@xmath4 , 2p@xmath3 @xmath65d@xmath4 , 2s2p @xmath65p@xmath5 and 2p@xmath3 @xmath65s@xmath4 are denoted by a , b , c , d , e and f respectively .
only the state energies were reported in [ 33 ] , so their excitation energies quoted in this work for comparison were calculated from the same li ground state as mentioned above .
the excitation energies in [ 31 ] were given with respect to the first excited 1s@xmath3 @xmath3p@xmath5 state of li and thus to put things in proper perspective , we have used e[1s@xmath32p @xmath3p@xmath5@xmath661s@xmath32s @xmath3s@xmath4]=1.848 ev to obtain the respective resonance positions relative to the li ground state . now table i tabulates the results for even - parity @xmath68a , ns@xmath69 @xmath3s@xmath4 and odd - parity @xmath68a , np@xmath69 @xmath3p@xmath5 , n=222 hollow states of li .
the experimental results for @xmath3s@xmath4 resonances are yet to be obtained .
the lower members of the series have been treated theoretically in considerable detail by several workers employing a multitude of techniques , e.g. , the n=3 state by a hyperspherical coordinate method [ 35 ] , the n=3,4 resonances by a combination of sp and ccr method [ 25 ] .
the n=39 resonances were also calculated by the r - matrix method including a set of core base states for the li@xmath64 target along with an optimization of the collisional representation [ 28 ] .
recently , an eigenphase derivative technique in conjunction with a quantum defect theory [ 31 ] reported the low and high resonances up to n=22 , whereas the same up to n=12 have been studied through the tdm [ 33 ] .
the agreement of the present density functional results for both the quantities are seen to be fairly good ; the current state energies lie about 0.360.88% above the tdm results [ 33 ] , whereas the excitation energies are higher by 0.410.51% from those of [ 31 ] .
it is noted that the small differences in excitation energies are reproduced quite nicely within the whole series ; a feature which holds true for all the resonance series dealt in this work .
coming to the @xmath3p@xmath5 states , these are the most widely studied hollow resonance series in li , both experimentally and theoretically , chiefly due to the fact that the lowest n=2 intrashell li state is the 2s@xmath32p @xmath3p@xmath5 .
photoion yield spectroscopic measurements determined its position at 142.33 ev [ 7 ] and 142.35 ev [ 6 ] respectively , which has been supported by various theoretical calculations , e.g. , the combined sp ccr method [ 23 ] , the complex scaling method having correlated basis functions constructed from b - splines [ 53 ] , the r - matrix method [ 28 ] , as well as the tdm method [ 33 ] .
our calculation gives an excitation energy of 142.385 ev which is only 0.02% above the experimental results .
it may be noted that the hf energy of the lowest resonance is @xmath662.16954 a.u .
[ 18 ] , which matches almost exactly with the present exchange - only result of @xmath662.1694 a.u . , ( 0.00014 a.u . , deviation ) illustrating the fact that the exchange contribution is included quite accurately within this formalism .
other members of the series with n=37 show good agreement with the ccr calculation [ 24 ] . for n=811 ,
the term energies are available also through the tdm calculation [ 33 ] and fairly good agreement is noticed in our results with those .
the present term energies are underestimated by 0.240.98% with respect to the ccr results [ 23,24 ] , leading to higher excitation energies ( the deviations with respect to [ 28 ] , [ 29 ] being 0.05 - 0.32% and 0.430.46% for n=29 and n=310 respectively ) .
no results could be found for the resonances with n@xmath7012 .
.[tab : table1]comparison of the state ( in a.u . ) and excitation energies ( in ev ) relative to the li ground state [ 54 ] , of 2s@xmath3ns @xmath3s@xmath4 and 2s@xmath3np @xmath3p@xmath5 resonances of li .
pw signifies present work .
[ cols= " < , < , < , < , < , < , < , < , < " , ] 3s @xmath3s@xmath4 , 2s@xmath32p @xmath3p@xmath5 , 2s2p@xmath3 @xmath6p@xmath4 , 2s2p3d @xmath6f@xmath5 and ( b ) 3s@xmath34s @xmath3s@xmath4 , 3s@xmath33d @xmath3d@xmath4 , 3s3p3d @xmath6f@xmath5 , 3p@xmath34s @xmath6p@xmath4 states respectively . ] 3s @xmath3s@xmath4 , 2s@xmath32p @xmath3p@xmath5 , 2s2p@xmath3 @xmath6p@xmath4 , 2s2p3d @xmath6f@xmath5 and ( b ) 3s@xmath34s @xmath3s@xmath4 , 3s@xmath33d @xmath3d@xmath4 , 3s3p3d @xmath6f@xmath5 , 3p@xmath34s @xmath6p@xmath4 states respectively . ] at this stage it is worthwhile to make a few pertinent remarks on the status of excited state calculations within dft .
although founded in the 1920s and later rejuvenated in the 1960s in the works of hohenberg , kohn and sham [ 58,59 ] , dft has been a very powerful and successful tool for the electronic structure calculation of atoms , molecules , solids in their ground states [ 60,61 ] , inherent difficulties were encountered for the excited states and consequently in areas such as spectroscopy , its success has been relatively less conspicuous .
the well - known fundamental problems such as the lack of a unique exact formal relationship valid for general excited states parallel to the ks method for ground states , as well as the unavailability of hitherto unfound universal xc energy density functional etc . , have been well documented in the literature in considerable length ( see , for example , [ 43 ] for a review ) . nevertheless , numerous attractive and elegant formalisms have been suggested by many authors over the years ; e.g. , the subspace formulation of dft [ 62 ] and its application to atomic excited states [ 63 ] , ensemble formalism for the unequally weighted states using a rayleigh - ritz variational principle [ 64 ] , a perturbative treatment [ 65,66 ] , calculation of multiplet energies within the hartree - fock - slater method utilizing the spin - polarized form of the density matrix and exploiting slater s sum rule [ 52,67,68 ] , ensemble approach using several appropriate functionals [ 69,70 ] , the time - dependent ( td ) dft formulation [ 71 - 73 ] making use of the td df response theory , etc .
recently correlation energies of several atomic excited states were calculated using the mcscf wave functions [ 74 ] .
while some of these methods offer good quality results , others produce large errors and yet others are computationally difficult to implement . moreover most of these methods have dealt with the lower and singly excited states ; multiple and higher excitations , especially the rydberg series such as the ones studied in this work , have not been reported so far using any other dft approach except the work - function formalism . besides , while some of these methods , such as tddft , provides a way to obtain excitation energies in an efficient and accurate manner ( calculates the linear response of the system to a td perturbation , leading to frequency dependent dynamic dipole polarizability , whose poles and residues yield the excitation energies and oscillator strengths respectively ) , the extraction of individual state energies as well as the densities , are not straightforward . in the present method however ,
the energies , excitation energies as well the densities and expectation values are obtained easily with reasonable accuracy . to the best of our knowledge ,
this is the first report of triply excited rydberg resonances of many - electron systems within a dft - based formalism .
we also note that the spin polarized version of the work - function exchange potential can be obtained as a further approximation to the accurate spin polarized x - only ks potential ( obtained from a consideration of the optimized effective potential method ) [ 75 ] .
various interesting features of the method may be found in the references [ 27,38 - 44 ] .
twelve triply excited @xmath0n@xmath1 hollow rydberg series of li ( covering a total of about 270 states ) have been studied using a density functional approach .
nonrelativistic term energies , excitation energies and radial densities are reported for both low as well as higher members ( up to n=25 ) of these odd- and even - parity resonances .
the gps solution of the ks type equation yields results in fairly good agreement with the existing literature data .
additionally 37 @xmath7n@xmath1 hollow states are also presented in the photon energy range 175.63180.51 ev , out of which only two ( 3s@xmath33p @xmath3p@xmath5 and 3s3p@xmath3 @xmath6p@xmath4 ) have been reported so far in the literature .
the resonance series interact with each other resulting in very complicated behavior in their positions .
many new resonances are reported here for the first time .
the combination of work - function exchange and the lyp correlation functional within an essentially single determinantal framework offers results quite comparable in accuracy with those from other sophisticated and elaborate quantum mechanical methods .
the exchange - only results are practically of hf - quality ; a feature of the work - function formalism observed in a number of physical systems including atoms , ions , metals , etc .
, ( see , for example [ 76,77 ] ) . consequently , since these are strongly correlated systems , one of the main sources of error in this calculation could be due to the inefficiency of the lyp functional to incorporate the subtle and intricate correlation effects , which may be either improved or replaced by more accurate density functionals .
however , this may not be misconstrued as a drawback of the methodology ; in fact the results are rather quite encouraging , especially in the light of dft s apparent weaknesses and lack of any density - based attempts as yet for these rydberg series .
the assumption of spherical symmetry in calculating the exchange potential , might be the another possible cause of inaccuracy .
thus the rotational component of the electric field may not have negligible contribution compared to the irrotational component for these states , although for atoms this usually holds true [ 49 ] .
the extension of this prescription to even higher photon - energy hollow states such as the case where all three electrons remain in shells with n@xmath704 , i.e. , the klm vacancy states etc .
, are straightforward , as well as its application to positive and negative ionic systems and extension to the relativistic domain
. it may also be interesting to employ some of the other dft - based approaches to treat these and other similar systems , so that the nature of the intricate electron correlation may be understood better .
some works in these directions are under progress .
99 r. bruch , g. paul , j. andr and l. lipsky , phys .
a * 12 * , 1808 ( 1975 ) .
m. rodbro , r. bruch and p. bisgaard , j. phys .
b * 12 * , 2413 ( 1979 ) .
m. agentoft , t. andersen and k. t. chung , j. phys .
b * 17 * , l433 ( 1984 ) .
s. mannervik , r. t. short , d. sonnek , e. trbert , g. mller , v. lodwig , p.
h. heckmann , j. h. blanke and k. brand , phys .
a * 39 * , 3964 ( 1989 ) .
l. m. kiernan , e. t. kennedy , j .-
mosnier , j. t. costello and b. f. sonntag , phys .
lett . * 72 * , 2359 ( 1994 ) .
y. azuma , h. hasegawa , f. koike , g. kutluk , t. nakata , e. shigemasa , a. yagishita and i. a. sellin , phys .
lett . * 74 * , 3768 ( 1995 ) .
l. m. kiernan , m .- k .
lee , b. f. sonntag , p. sladeczek , p. zimmermann , e. t. kennedy , j .-
mosnier and j. t. costello , j. phys .
b * 28 * , l161 ( 1995 ) .
y. azuma , f. koike , j. w. cooper , t. nagata , g. kutluk , e. shigemasa , r. wehlitz and i. a. sellin , phys . rev .
* 79 * , 2419 ( 1997 ) .
l. journel , d. cubaynes , j .- m .
bizau , s. al moussalami , b. rouvellou , f. j. wuilleumier , l. vo ky , p. faucher and a. hibbert , phys .
lett . * 76 * , 30 ( 1996 ) .
s. diehl , d. cubaynes , j .- m .
bizau , l. journel , b. rouvellou , s. al moussalami , f. j. wuilleumier , e. t. kennedy , n. berrah , c. blancard , t. j. morgan , j. bozek , a. s. schlachter , l. vo ky , p. faucher and a. hibbert , phys .
lett . * 76 * , 3915 ( 1996 ) .
d. cubaynes , s. diehl , l. journel , b. rouvellou , j .- m . bizau , s. al moussalami , f. j. wuilleumier , n. berrah , l. vo ky , p. faucher , a. hibbert , c. blancard , e. t. kennedy , t. j. morgan , j. bozek and a. s. schlachter , phys .
lett . * 77 * , 2194 ( 1996 ) .
s. diehl , d. cubaynes , f. j. wuilleumier , j .- m .
bizau , l. journel , e. t. kennedy , c. blancard , l. vo ky , p. faucher , a. hibbert , n. berrah , t. j. morgan , j. bozek and a. s. schlachter , phys .
lett . * 79 * , 1241 ( 1997 ) .
s. diehl , d. cubaynes , k. t. chung , f. j. wuilleumier , e. t. kennedy , j .- m .
bizau , l. journel , c. blancard , l. vo ky , p. faucher , a. hibbert , n. berrah , t. j. morgan , j. bozek and a. s. schlachter , phys.rev .
a * 56 * , r1071 ( 1997 ) .
s. diehl , d. cubaynes , h. l. zhou , l. vo ky , f. j. wuilleumier , e. t. kennedy , j .-
bizau , s. t. manson , t. j. morgan , c. blancard , n.berrah and j. bozek , phys . rev .
lett . * 84 * , 1677 ( 2000 ) .
s. diehl , d. cubaynes , e. t. kennedy , f. j. wuilleumier , j .- m .
bizau , l. journel , l. vo ky , p. faucher , a. hibbert , c. blancard , n. berrah , t. j. morgan , j. bozek and a. s. schlachter , j. phys .
b * 30 * , l595 ( 1997 ) .
m. ahmed and l. lipsky , phys .
a * 12 * , 1176 ( 1975 ) .
u. i. safronova and v. s. senashenko , j. phys .
b * 11 * , 2623 ( 1978 ) .
r. l. simons , h. p.
kelly and r. bruch , phys . rev .
a * 19 * , 682 ( 1979 ) . c. a. nicolaides , n. a. piangos and y. komninos , phys .
a * 48 * , 3578 ( 1993 ) .
n. a. piangos and c. a. nicolaides , phys .
a * 48 * , 4142 ( 1993 ) . c. a. nicolaides and n. a. piangos , j. phys .
b * 34 * , 99 ( 2001 ) .
n. vacek and j. e. hansen , j. phys .
b * 25 * , 883 ( 1992 ) .
k. t. chung and b. c. gou , phys .
a * 52 * , 3669 ( 1995 ) .
k. t. chung and b. c. gou , phys .
a * 53 * , 2189 ( 1996 ) .
y. zhang and k. t. chung , phys .
a * 58 * , 1098 ( 1998 ) .
h. bachau , j. phys .
b * 29 * , 4365 ( 1996 ) .
a. k. roy , r. singh and b. m. deb , int .
j. quant .
chem . * 65 * , 317 ( 1997 ) .
k. berrington and s. nakazaki , j. phys .
b * 31 * , 313 ( 1998 ) .
l. vo ky , p. faucher , h. l. zhou , a. hibbert , y .- z .
li and f. bely - dubau , phys .
a * 58 * , 3688 ( 1998 ) .
h. l. zhou , s. t. manson , l. vo ky , p. faucher , f. bely - dubau , a. hibbert , s. diehl , d. cubaynes , j .-
bizau , l. journel and f. j. wuilleumier , phys .
a * 59 * , 462 ( 1999 ) .
h. l. zhou , s. t. manson , p. faucher and l. vo ky , phys .
a * 62 * , 012707 ( 2000 )
. m. j. conneely and l. lipsky , phys .
a * 61 * , 032506 ( 2000 ) . m. j. conneely and l. lipsky , at .
data nucl .
data tables * 82 * , 115 ( 2002 ) .
t. morishita and c. d. lin , phys .
a * 59 * , 1835 ( 1999 ) .
t. morishita and c. d. lin , phys .
a * 67 * , 022511 ( 2003 ) .
m. k. harbola and v. sahni , phys .
lett . * 62 * , 489 ( 1989 ) . c. lee , w. yang and r. g. parr , phys .
b * 37 * , 785 ( 1988 ) .
r. singh and b. m. deb , j. chem .
phys . * 104 * , 5892 ( 1996 ) .
a. k. roy , r. singh and b. m. deb , j. phys .
b * 30 * , 4763 ( 1997 ) .
a. k. roy and b. m. deb , phys .
a * 234 * , 465 ( 1998 ) .
a. k. roy and b. m. deb , chem .
lett . * 292 * , 461 ( 1998 ) .
r. singh , a. k. roy and b. m. deb , chem .
* 296 * , 530 ( 1998 ) .
r. singh and b. m. deb , phys .
rep . * 311 * , 47 ( 1999 ) .
a. k. roy and s. i. chu , phys .
rev . a * 65 * , 052508 ( 2002 ) .
x. m. tong and s. i. chu , phys .
a * 64 * , 013417 ( 2001 ) .
d. telnov and s. i. chu , phys .
a * 66 * , 043417 ( 2002 ) .
a. k. roy and s. i. chu , phys .
a * 66 * , 043402 ( 2002 ) .
a. k. roy , phys .
a * 321 * , 231 ( 2004 ) .
v. sahni , phys .
a * 55 * , 1846 ( 1997 ) .
g. yao and s. i. chu , chem .
* 204 * , 381 ( 1993 ) . j. wang , s. i. chu and c. laughlin , phys . rev . a * 50 * , 3208 ( 1994 ) .
t. ziegler , a. rauk and e. j. baerends , theor .
acta * 43 * , 261 ( 1977 ) .
l. b. madsen , p. schlagheck and p. lambropoulos , phys .
a * 62 * , 062719 ( 2000 ) .
k. t. chung , phys .
a * 44 * , 5421 ( 1991 ) .
n. a. piangos and c. a. nicolaides , phys .
a * 67 * , 052501 ( 2003 ) .
k. t. chung , phys .
a * 64 * , 052503 ( 2001 ) .
t. morishita and c. d. lin , phys .
a * 64 * , 052502 ( 2001 ) .
p. hohenberg and w. kohn , phys .
rev . * 136 * , b864 ( 1964 ) .
w. kohn and l. j. sham , phys .
rev . * 140 * , a1133 ( 1965 ) .
r. g. parr and w. yang , _ density functional theory of atoms and molecules _
( oxford university press , new york , 1989 ) , edited by e. k. u. gross and r. m. dreizler ( plenum , new york , 1995 ) . n. hadjisavvas and a. k. theophilou , phys . rev .
a * 30 * , 2183 ( 1984 ) ; _ ibid . _
* 32 * , 720 ( 1985 ) .
f. tasndi and .
nagy , int .
j. quant .
chem . * 92 * , 234 ( 2003 ) .
e. k. u. gross , l. n. oliveira and w. kohn , phys .
a * 37 * , 2805 ( 1988 ) ; _ ibid . _ * 37 * , 2809 ( 1988 ) .
a. grling , phys .
a * 54 * , 3912 ( 1996 ) . c. fillippi , c. umrigar and x. gonze , j. chem .
phys . * 107 * , 9994 ( 1997 ) .
u. von barth , phys .
a * 20 * , 1693 ( 1979 ) . c. daul , int .
j. quant .
chem . * 52 * , 867 ( 1994 ) .
nagy , adv .
* 29 * , 159 ( 1997 ) .
nagy , int .
j. quant .
chem . * 69 * , 247 ( 1998 ) .
m. petersilka , u. j. gossmann and e. k. u. gross , phys .
lett . * 76 * , 1212 ( 1996 ) . c. jamorski , m. e. casida and d. r. salahub , j. chem .
phys . * 104 * , 5134 ( 1996 ) .
f. della sala and a. grling , int .
j. quant .
chem . * 91 * , 131 ( 2002 ) .
e. san fabin and l. pastor - abia , int . j. quant .
chem . * 91 * , 451 ( 2003 ) .
j. b. krieger , y. li and g. j. iafrate , phys .
a * 45 * , 101 ( 1992 ) .
v. sahni , in _ structure and dynamics of atoms and molecules : conceptual trends _ , j. l. calais and e. s. kryachko , eds .
( kluwer academic , dordrecht , 1995 ) .
v. sahni and a. solomatin , adv .
chem . * 33 * , 241 ( 1999 ) . | density functional calculations are performed for twelve @xmath0n@xmath1 ( n@xmath22 ) triply excited hollow resonance series of li , _ viz . , _
2s@xmath3ns @xmath3s@xmath4 , 2s@xmath3np @xmath3p@xmath5 , 2s@xmath3nd @xmath3d@xmath4 , 2p@xmath3ns @xmath3d@xmath4,@xmath6p@xmath4 , 2s2pns @xmath6p@xmath5 , 2s2pnp @xmath6d@xmath4 , 2p@xmath3np @xmath3f@xmath5,@xmath6d@xmath5 , 2p@xmath3nd @xmath3g@xmath4 , @xmath6f@xmath4 and 2s2pnd @xmath6f@xmath5 , covering a total of about 270 low- , moderately high- and high - lying states , with n as high as up to 25 . the work - function - based exchange potential and the nonlinear gradient plus laplacian included lee - yang - parr correlation energy functional is used .
the relevant kohn - sham - type equation is solved numerically using the generalized pseudospectral method offering nonuniform , optimal spatial discretization to obtain the orbitals and densities .
the present single determinantal approach yields fairly accurate results for the nonrelativistic energies , excitation energies as well as the radial densities and other expectation values . except for the one state ,
the discrepancy in the calculated state energies remains well within 0.98% , whereas the excitation energies deviate by 0.020.58% compared to the available experimental and other theoretical results .
additionally companion calculations are also presented for the 37 @xmath7n@xmath1 ( n@xmath23 ) doubly hollow states ( seven are n=3 intrashell type ) of li with both k and l shells empty ( up to n=6 ) in the photon energy range 175.63180.51 ev , with varying symmetries and multiplicities .
our calculation shows good agreement with the recent literature data for the only two such doubly hollow states reported so far , _ viz . , _ 3s@xmath33p @xmath3p@xmath5 and 3s3p@xmath3 @xmath6p@xmath4 .
the resonance series are found to be inextricably entangled to each other , leading to complicated behavior in their positions .
many new states are reported here for the first time .
this provides a simple , efficient and general scheme for the accurate calculation of these and other multiply excited rydberg series of many - electron atomic systems within density functional theory . |
perovskite derived ruthenates , ( sr , ca)@xmath11ru@xmath12o@xmath13 with nominally ru@xmath14 in an octahedral o environment display an extraordinarily varied set of physical properties . in the sr series ,
the end members , srruo@xmath0 and sr@xmath1ruo@xmath7 are , respectively , an itinerant ferromagnet @xcite and an unconventional superconductor , widely thought to have triplet pariing symmetry , @xcite possibly of magnetic origin .
@xcite the fermi surface and momentum - dependent spin fluctuations play key roles in the pairing in such scenarios for superconductivity .
the corresponding ca end - points , caruo@xmath0 and ca@xmath1ruo@xmath7 , which differ structurally from the corresponding sr compounds only by moderate lattice distortions , are a highly renormalized paramagnetic metal and an antiferromagnetic mott insulator , respectively .
nakatsuji1,braden , puchkov1,cao1,cao2,cao2b , nakatsuji2 the multilayer ruddlesden - popper compounds in the series show a wide variety of magnetic orderings , metal insulator transitions and unusual transport properties , with sample dependence . @xcite
not surprisingly , strong magnetoelastic effects are found and have been emphasized both theoretically @xcite and experimentally .
braden , nakatsuji2 density functional calculations of the band structure and fermi surfaces of superconducting sr@xmath1ruo@xmath7 have been reported @xcite and subsequently confirmed in detail by quantum oscillation experiments .
@xcite angle resolved photoemission ( arpes ) experiments @xcite showed features very much like those in the band calculations and quantum oscillation measurements but with small shifts . however , because of van hove singularities near the fermi energy , these shifts drastically affect the topology .
the fermi surfaces predicted by band calculations arise from three bands , one each from the three ru @xmath15 orbitals hybridized with the in - plane and , to a lesser extent , apical o @xmath4 states .
the fermi surface structure , which is highly two dimensional , thus consists to zeroth order of planar sections perpendicular to @xmath16 and @xmath17 , coming from the @xmath5 and @xmath6 orbitals , and a round cylinder centered at @xmath18 , derived from the @xmath8 orbital .
these hybridize and reconnect to form three sections , a circular cylinder around @xmath18 and square cylinders around @xmath18 and x. strong nesting features remain from the @xmath19 and @xmath5 derived sheets , which lead to soft spin - fluctuations around @xmath20 , as observed in neutron scattering .
@xcite we have pointed out earlier@xcite that the interplay of these antiferromagnetic spin fluctuations and a broad stoner background , related to the proximity to ferromagnetism , may lead to competing order parameter symmetries .
however , the situation may be complicated by magnetoelastic effects , that is , mixing between spin and charge degrees of freedom , _
i.e. _ spin - fluctuations and phonons , which could modify the pairing interaction , perhaps favoring triplet symmetries though the physics of this is not well understood .
perhaps the most similar compound to sr@xmath1ruo@xmath7 is the bilayer compound sr@xmath0ru@xmath1o@xmath2 .
@xcite it is closely related in structure , highly two dimensional and metallic .
cao and co - workers reported ferromagnetism at 104 k with additional lower temperature transitions , @xcite while other samples were not ferromagnetic at these temperatures but exhibited transitions below 20 k. @xcite meanwhile , huang and co - workers found no evidence for any long range magnetic order down to 1.6 k @xcite and perry , and ikeda and co - workers report that single crystals grown by a floating zone technique are paramagnetic , with strongly enhanced susceptibility and high wilson ratio , leading to the conclusion that the material is on the verge of ferromagnetism .
@xcite initially , sr@xmath0ru@xmath1o@xmath2 was reported to be tetragonal , @xmath21 , like sr@xmath1ruo@xmath7 .
@xcite however , subsequent measurements revealed at least disordered rotations of the octahedra .
@xcite recently , shaked and co - workers @xcite have shown that the distortions are ordered and refined the crystal structure into spacegroup @xmath22 with a rotation of approximately 7 degrees .
in contrast , superconducting sr@xmath1ruo@xmath7 is stable in the tetragonal @xmath21 structure , but has a rotational phonon mode that softens steeply from the @xmath18 point to the zone boundary interacting with and crossing the acoustic branch ( @xmath23 corresponding at @xmath24 to the main instability that produces the orthorhombic structure in sr@xmath0ru@xmath1o@xmath2 ) .
@xcite the electronic structure of sr@xmath0ru@xmath1o@xmath2 has been studied using arpes @xcite and density functional calculations @xcite and reasonable agreement was obtained , although not all the predicted sheets were resolved experimentally .
however , perry and co - workers recently reported low temperature hall measurements on high quality ( @xmath25 cm ) single crystals that seem inconsistent with the calculated tetragonal fermi surfaces.@xcite the purpose of the present paper is to report density functional calculations of the electronic structure , fermiology and magnetic properties using the orthorhombic @xmath22 crystal structure .
the calculations were done using the general potential linearized augmented planewave ( lapw ) method including local orbital extensions to treat the upper core levels and relax linearization errors .
@xcite the calculations were done within the local density approximation using the hedin - lundqvist exchange correlation function .
spin polarized calculations used the same functional and the von barth - hedin spin scaling .
well converged basis sets and brillouin zone samplings were used , similar to those described previously .
@xcite calculations for the ideal tetragonal structure were based on the experimental structure of ref . which differs slightly from that used by hase and nishihara .
@xcite however , the calculated electronic structure is practically the same as obtained by them .
the calculations for orthorhombic sr@xmath0ru@xmath1o@xmath2 are based on the neutron data of shaked and co - workers obtained on samples that , like floating - zone grown single crystals , show a susceptibility peak around 20 k presumably due to magnetism ( but see below ) .
the calculated band structure and fermi surfaces of tetragonal @xmath264/mmm sr@xmath27ru@xmath1o@xmath2 are shown in figs . [ bands ] and [ tet - fermi ] , respectively . as may be seen , they are very similar to those reported previously by hase and nishihara . @xcite
as mentioned , the three fermi surfaces of sr@xmath1ruo@xmath7 , which has one ruo@xmath1 layer per cell , may be regarded as arising from the three ru @xmath15 orbitals .
the the @xmath28 orbital gives rise to a round cylindrical electron - like sheet centered at @xmath29 ( z ) and the @xmath30 and @xmath31 orbitals provide flat sheetlike sections perpendicular to @xmath32 and @xmath33 , respectively , that after reconnection become square cylindrical sections around @xmath34 and @xmath18 along with strong nesting . to a first approximation
, the fermi surfaces of tetragonal sr@xmath0ru@xmath1o@xmath2 may be thought of as deriving from the six same bands ( three from each ruo@xmath1 layer ) with bonding antibonding ( odd even ) splittings@xcite due to the interaction between the ruo@xmath35 sheets comprising the bi - layer .
however , as may be seen from the lack of four - fold symmetry around the @xmath34 point in fig .
[ tet - fermi ] , there is more @xmath33 dispersion in tetragonal sr@xmath0ru@xmath1o@xmath2 than in sr@xmath36ruo@xmath7 , particularly for the even bands .
there is an approximately square cylindrical hole pocket around @xmath34 , deriving from the odd combinations of @xmath30 and @xmath31 orbitals .
another , cross - shaped , hole pocket arises from reconnected sections of the fermi surfaces of the even and odd parity with the same orbital character .
the even combination also provides the innermost cylinder around @xmath18 , which like the other three @xmath18 centered cylinders is electron - like .
the second @xmath18 centered electron - like cylinder is also mainly @xmath37 derived near the ( 10 ) directions , but is mixed with the @xmath38 character near ( 11).the two remaining @xmath39 centered cylinders are derived from the @xmath28 orbitals , and have very little even - odd splitting , in accord with their strongly in - plane orbital character . the three outer @xmath18 centered cylinders change topology as the electron count is increased as they touch along the basal plane @xmath40 lines giving rise to the van hove singularities seen above @xmath41 in the density of states , as in sr@xmath1ruo@xmath7 .
it is noteworthy that even before considering the effects of orthorhombicity , the fermi surface nesting is considerably reduced relative to sr@xmath1ruo@xmath7 .
this is due to ( i ) the even odd splittings , ( ii ) the @xmath33 dispersions , and ( iii ) , the strong anticrossings of the _ even _ @xmath42 and the _ odd _ @xmath38 bands .
the striking difference from sr@xmath1ruo@xmath7 stems from the fact that nearest neighbor hopping between @xmath42 and @xmath28 orbitals is forbidden within a single layer , but allowed across the layers in a bi - layer .
a @xmath43 orbital hybridizes with the @xmath28 of the _ opposite parity _ , and this is the strongest at @xmath44 ( @xmath45 correspondingly , the strongest distortion of the @xmath46 band occurs near the ( 1,1 ) directions .
note that additional hybridization between the @xmath46 band and the @xmath8 band ( and the only one existing in sr@xmath1ruo@xmath47 happens due to hopping across the sro layer .
depending on @xmath33 , this hopping is of the same @xmath48 or of the opposite sign with respect to the intra - bi - layer hopping .
this brings about an additional @xmath33 dispersion .
given the reduced nesting in sr@xmath0ru@xmath1o@xmath49 one may conjecture that the incommensurate antiferromagnetic spin - fluctuations seen in neutron scattering experiments @xcite on sr@xmath1ruo@xmath7 may be considerably less prominent in sr@xmath0ru@xmath1o@xmath2 . on the other hand
, tetragonal sr@xmath27ru@xmath1o@xmath2 is much closer to a ferromagnetic instability .
this is not surprising considering the high value of the calculated density of states ( dos ) at the fermi energy ( @xmath41 ) , @xmath50 = 4.5 states / ev ru , compared with 4.1 states / ev ru for sr@xmath1ruo@xmath7 .
@xcite non - spin - polarized local spin density approximation ( lsda ) calculations for the orthorhombic structure yield a still slightly larger @xmath51 ) = 5.0 states / ev ru and the structure of the density of states around @xmath41 is different ( with larger magnitudes corresponding to slightly narrower @xmath15 bands , as shown in fig . [ dos ] ) .
the result is a magnetic ground state .
however , before turning to the magnetic properties , we discuss the electronic properties of the non - spin - polarized system .
as the magnetism has substantially itinerant character ( see below ) the non - spin - polarized electronic structure is expected to be a reasonable approximation for the paramagnetic phase well above any ordering temperature . as a reference point
, we show in fig .
[ fold ] the fermi surfaces of the tetragonal structure , but folded into the @xmath22 orthorhombic zone .
this zone is half the area of the tetragonal zone due to the cell doubling and is rotated by 45@xmath52 .
the most striking feature of the folding is the development of cylindrical lens shaped fermi surfaces centered at the mid - points of the @xmath40 lines ( the mid - points of the @xmath10 and @xmath53 lines in the tetragonal zone ) .
these come from the two @xmath28 circular cylinders around @xmath18 and the outer ( odd ) @xmath42 cylinder .
one may expect that such folded fermi surfaces and particularly the lenses should become observable in quantum oscillation experiments in samples with ordered orthorhombic structures , and possibly in photoemission as shadow fermi surfaces with strength related to the matrix elements produced by the orthorhombic distortion even in samples where the distortion is not fully long range ordered , as long as the in - plane coherence length is large enough ( in analogy with some of the cuprate superconductors .
bisco1,bisco2,bisco3,bisco4 ) however , the actual distortion is a rotation of approximately 7@xmath52 , which is not so small , and considerably changes the fermi surfaces beyond the naive zone folding picture . in fact
, significant differences between the folded tetragonal and calculated orthorhombic fermi surfaces are found , as shown in fig .
ort - fermi .
hybridization gaps open where folded bands cross . as a result , the square sections around the center of fig .
[ fold ] , which came from the overlap of the @xmath28 derived bands , entirely disappear , while the larger square section derived from the odd @xmath42 bands reconnects with the the outer lens derived from the @xmath28 bands .
one can see remnants of this square in the dents near the tips of the outer lenses in the orthorhombic structure . also two small circular electron cylinders are introduced around the @xmath40 lines , with hardly any @xmath54 dispersion ; these are the bottoms of the @xmath55 bands , which in the tetragonal structure were located a few mry above the fermi level ( fig.bands ) .
the lenses are now considerably distorted , because of the above - mentioned interaction with the odd combination of @xmath42 states , and because of the increased the @xmath33 dispersion ( note the lack of reflection symmetry about the center line in fig .
ort - fermi ) .
interestingly , this dispersion is concentrated mostly in small and/or heavy sections , so the average ( rms ) fermi velocity remains quite two dimensional , _
i.e. _ @xmath56
cm / s and @xmath57 cm / s . within the constant scattering time approximation this implies a resistivity anisotropy @xmath58 as compared to the lda value for sr@xmath1ruo@xmath7 of @xmath59 , which is known to be an underestimate relative to experiment .
clearly , the bilayer compound is much more three - dimensional than the single layer one , and this , presumably , moves it still further from an antiferromagnetic instability in - plane .
finally , it should be mentioned that , because of the hybridization gaps near the center of the brillouin zone of fig .
[ ort - fermi ] , the fermi surface in this region is highly sensitive to the band filling .
this is illustrated in fig .
[ shift - fermi ] , where fermi surfaces calculated with 5 mev upwards and downwards shifts of @xmath41 are shown , so that the thickness of the lines is inversely proportional to the local fermi velocity .
considering the fact that lda calculations are sensitive to structural details and in any case should not be relied on at the 5 mev level , the topology in this region can not be decided based on the present calculations . because of this uncertainty , we can not unambiguously match our results with the hall measurements of perry and co - workers @xcite , but the possibility , suggested by them , that orthorhombicity provides a way of reconciling their measurements with band calculations seems plausible . if we restrict ourselves to the the `` stable '' sheets of the fermi surface , we observe two electronic @xmath55 derived cylinders , the hole pseudo - square and the hole pseudo - cross formed by the @xmath42 bands , and several sets of lenses , all electronic . the complicated shape and changing curvature of these surfaces prevents us from associating the volumes of these fermi surface pockets with hall numbers .
the only conclusion we can draw is that there is no clear way to guess the hall conductivity from the electronic structure of the corresponding tetragonal compound .
furthermore , as seen from fig .
[ shift - fermi ] , shifting the chemical potential by as little as 5 mev ( corresponding to 55 k ) drastically changes the topology of the paramagnetic fermi surfaces .
the message of this observation is that substantial and nontrivial temperature dependencies of the hall coefficient may occur independent of any anomalous hall contribution .
as mentioned , in the idealized tetragonal structure , sr@xmath0ru@xmath1o@xmath2 is found to be paramagnetic in the lsda , but with a very high susceptibility . in fixed spin moment calculations , moments up to 0.6 @xmath60/ru
are induced with an energy cost of less than 4 mev / ru a number almost indistinguishable from zero considering the accuracy of the present calculations .
the orthorhombic distortion narrows the @xmath15 manifold and increases the density of states leading to a magnetic ground state via the stoner mechanism .
furthermore , the paramagnetic orthorhombic band structure shows extremely flat bands near the x point ( cf .
shift - fermi ) , which create a low - weight but extremely sharp peak within a few k from the fermi level .
such a band structure is intrinsically unstable against any symmetry lowering splitting of this peak , including a magnetic instability .
however , since the orthorhombic unit cell is already doubled , an antiferromagnetic instability would work just as well .
there is nevertheless a strong reason for the system to prefer the ferromagnetic coupling in plane . generally speaking
, the electronic structure of ru based perovskites shows strong ru @xmath61 o @xmath4 hybridization and this favors ferromagnetic couplings@xcite ( o @xmath4 states have rather large hund rule coupling ) . within a local picture
, this tendency may be viewed as resulting from the fact that the density of states has substantial o weight around @xmath41 and all things being equal , this favors magnetic configurations where the o can polarize , _
i.e. _ ferromagnetism . within a simple stoner picture , this o contribution to the magnetic energy is proportional to the square of the o contribution to the density of states around @xmath41 .
decomposing @xmath50 onto the lapw spheres , @xcite we find 56 % ru @xmath61 , 20 % plane o @xmath4 , 2 % apical ( rocksalt ) o @xmath4 and 3 % bridging ( interlayer ) o @xmath62 with the remainder divided between other angular momentum characters ( 2 % ) , and the interstitial ( 17 % ) . considering the ionic radii ,
the interstitial contribution is likely more o @xmath63 than ru @xmath61 derived , but in any case , there is substantial o @xmath4 involvement in @xmath50 and this contribution is dominated by the in - plane oxygen .
not surprisingly then , ferromagnetic in - plane configurations are favored .
we find a self - consistent ferromagnetic ( fm ) solution with a spin magnetization of 0.80 @xmath64/ru and an energy of -23 mev / ru relative to the non - spin - polarized case .
calculations were also performed for antiferromagnetic configurations with a ( 2 @xmath65 2 ) in - plane ordering , and having adjacent ru ions in the two planes of the bilayer polarized parallel . however , no self - consistent magnetic configuration was found .
thus it may be concluded that the magnetic character _ within each plane _ is itinerant .
however , calculations in which the ru ions in a layer were ferromagnetically aligned , but the layers were stacked antiferromagnetically ( so each bi - layer had one spin up and one spin down ruo@xmath1 layer ) did yield a stable magnetic solution ( denoted af - a in the following ) , in this case with an energy of -20 mev / ru and a ru moment ( as measured by the moment in a ru sphere ) only 14% smaller than the ferromagnetic solution . the energy difference between the fm and af - a solutions contains two parts , within the most simple model .
the first is the interaction between the two planes comprising the bi - layer .
this , in turn , has a ferromagnetic part originating from the hund rule energy on the bridging oxygen , and an antiferromagnetic superexchange .
based on the calculation , we conclude that the former is slightly stronger . in any case
, it is likely larger than the inter bi - layer coupling through the rock - salt layers .
this expectation is based on the geometry ( hopping via two oxygens with unfavorable bond angles ) and the fact that the @xmath33 dispersion is considerably less than the anti - symmetric , symmetric band splittings due to the interaction between the planes comprising the bi - layer .
supposing that the inter - bilayer interaction is anti - ferromagnetic ( as expected for a superexchange coupling ) and much weaker than the intra bi - layer coupling , one may conclude that the latter is ferromagnetic with a strength of order 3 mev / ru . in this scenario ,
the ground state is antiferromagnetic , consisting of ferromagnetic bi - layers , stacked antiferromagnetically .
this represents a conceptually interesting case for the kosterlitz - thouless theory of quasi-2d magnetic phase transitions .
@xcite the difference from the textbook case is that the in - plane magnetism is itinerant ; the same qualitative picture applies , still , and so the 3d long range order ( lro ) transition temperature should be logarithmically suppressed compared to the ferromagnetic - paramagnetic energy difference .
we can safely assume that the inter - bilayer coupling is less than 30 k ( the interplane coupling within a bilayer ) , probably very much less , thus bringing about a substantial kosterlitz - thouless suppression .
one would , however , expect strong itinerant spin fluctuations of ferromagnetic character in planes for temperatures well above the lro transition temperature ; these should manifest themselves , for instance , in specific heat and magnetic susceptibility . very recently
, ikeda and co - workers reported a detailed study of single crystal floating - zone sr@xmath66ru@xmath67o@xmath68 including magnetic susceptibility , specific heat and resistivity plus magnetic measurements under pressure .
@xcite they report a susceptibility maximum at @xmath69k accompanied by structure in the resistivity , but conclude that long range antiferromagnetic order does not set in , based on the high nearly isotropic @xmath70 below @xmath71 and thus that the ground state is a paramagnetic fermi liquid on the verge of ferromagnetism .
in fact , the measured @xmath70 is @xmath72 15 times larger than that of sr@xmath67ruo@xmath73 ( ref . ) and at least as isotropic at low @xmath74 .
certainly , this is not expected with simple local moment ordered magnetism , especially in a material with strong magnetocrystalline anisotropy as is the case for magnetic ruthenates .
@xcite however , things are less clear cut in the itinerant metallic scenario above .
first of all , in the ordered itinerant case , there is a stoner continuum that contributes to @xmath70 along the direction of the moments ( the low @xmath70 direction ) ; normally this is a very small effect , but here the magnetocrystalline anisotropy is expected to be very large @xcite and the fermi liquid very soft as evidenced by the strong stoner renormalizations .
secondly , it should be noted that as long as the anisotropy is not in the trivial @xmath9-axis direction , the twinning of orthorhombic samples will mean that there will be no orientation where the applied field is aligned with the moments , again presumably lowering the anisotropy of @xmath75 in the magnetic state .
however , both of these mechanisms for lowering the anisotropy in @xmath75 would require fortuitous strong numerical coincidences to explain the observed isotropy of the measured @xmath75 in a magnetically ordered state .
on the other hand , the measured temperature dependence of the two components of @xmath75 is notably different , which is not simply anticipated for a standard enhanced paramagnet either . another interesting possibility is to associate the peak in @xmath70 not with lro antiferromagnetic ordering , but with a kosterlitz - thouless type transition associated with the 2d bilayers . in this case
, there would be no long range magnetic order below the peak , though the weak inter - bilayer interactions could still yield a 3d ordered magnetic state at very low temperatures .
this scenario would also provide a convenient explanation of the so called magnetic metallic phase in the ca@xmath76sr@xmath77ruo@xmath73 phase diagram between @xmath78 and @xmath79 where an unusual phase with a peak in @xmath70 is entered just as the system reaches an incipient ferromagnetic instability with @xmath80 decreasing through 0.5 .
@xcite finally , under applied pressures of 1 gpa ikeda and co - workers @xcite report evidence for ferromagnetism though with small moments ( @xmath81/ru ) starting at 70 k based on magnetization measurements .
while low moment ferromagnetism with this curie temperature can not be excluded , such a high ratio of @xmath82 is unusual in traditional materials .
weak ferromagnetism due to canting seems to be more likely and also fits better with our calculations .
we note that the ordering temperature in such a scenario would be sensitive to the out - of - plane superexchange coupling , which in turn is expected to increase strongly under pressure .
density functional studies of sr@xmath66ru@xmath67o@xmath68 show a substantial coupling of the electronic and magnetic properties with the octahedral rotations associated with the orthorhombic distortion .
in particular , the fermi surface nesting ( already lower than in sr@xmath67ruo@xmath73 ) is further reduced and the @xmath15 manifold is narrowed leading to magnetic tendencies . using the experimental crystal structure @xcite
we find that the 2d ruo@xmath67 planes form an itinerant system with ferromagnetic ordering tendencies , and additionally that the coupling between the two such planes comprising a bilayer is weak , but also ferromagnetic .
we argue that the @xmath9-axis coupling between bilayers is weaker still and based on speculate that the magnetic character should be highly two dimensional , perhaps exhibiting kosterlitz - thouless physics .
we are grateful for helpful discussions with c.s .
hellberg , j.d .
jorgensen , j.w .
lynn , a.p .
mackenzie , s. nakatsuji and m. sigrist .
we thank j.d .
jorgensen for a preprint of ref . .
computations were performed using facilities of the dod hpcmo asc center .
work at the naval research laboratory is supported by the office of the naval research .
this may seem counterintuitive ; however , the positions of the ru bands are mainly defined by the crystal field , which , in turn , is mainly due to ru - o hybridization
. the odd combination of ru d orbitals does not hybridize with the bridging oxygen , and thus is less pushed up than the even one .
lapw sphere radii of 2.1 , 2.0 and 1.65 a@xmath83 were used for sr , ru and o , respectively . an lapw sector basis
set cut - off of 4.55 a@xmath84 was used with the addition of local orbitals for the sr @xmath85 and @xmath4 , ru @xmath85 , @xmath4 and @xmath61 and o @xmath85 channels . | the electronic structure of layered sr@xmath0ru@xmath1o@xmath2 in its orthorhombic structure is investigated using density functional calculations .
the band structure near the fermi energy , consists of ru @xmath3 states hybridized with o @xmath4 orbitals .
the @xmath5 and @xmath6 bands , which are largely responsible for the nesting related antiferromagnetic spin fluctuations in sr@xmath1ruo@xmath7 split pairwise into even and odd combinations due to the interlayer coupling reducing the strength of the nesting .
the @xmath8 bands show much less interplanar coupling as expected , and also less @xmath9-axis dispersion , so that the barrel like sections are largely intact compared to the single layer material .
the zone folding due to orthorhombicity yields small cylindrical lens shaped fermi surfaces centered at the midpoints of the former tetragonal @xmath10 lines .
fixed spin moment calculations indicate that tetragonal sr@xmath0ru@xmath1o@xmath2 is borderline ferromagnetic but that orthorhombicity favors magnetism via a substantial magnetoelastic coupling .
these results are related to experimental observations particularly in regard to magnetic properties . 2 |
neutrino - photon interactions are of interest in astrophysics and cosmology .
it is well known that in the standard model , neutrino - photon interactions appear at the one - loop level .
it is also well known that in a vacuum the reaction @xmath1 is forbidden .
the processes with two photons , for example , neutrino - photon scattering @xmath2 , turn out to be highly suppressed . in @xcite gell - mann showed that in the four - fermi limit of the standard model the amplitude is exactly zero to order @xmath3 because , according to yang s theorem @xcite , two photons can not couple to the @xmath4 state .
therefore the amplitude is suppressed by the additional factors of @xmath5 , where @xmath6 is the photon energy and @xmath7 is the @xmath8 mass @xcite .
for example , in the case of massless neutrinos , the amplitude for @xmath9 in the standard model is suppressed by the factor @xmath10 @xcite . as a result ,
the cross section is exceedingly small .
so we see that neutrino - photon processes with one or two photons do not occur or are suppressed in the vacuum .
the presence of a medium or external electromagnetic field drastically changes the situation .
it induces an effective coupling between photons and neutrinos , which contributes to the @xmath1 process and cross - related reactions .
furthermore , it was shown in @xcite that in the presence of an external magnetic field , cross sections for neutrino - photon processes such as @xmath11 and @xmath12 are enhanced by the factor @xmath13 for @xmath14 and @xmath15 .
later this result was extended to very strong magnetic fields @xcite and arbitrary @xmath16 @xcite . in this paper
we will deal with @xmath1 and @xmath17 reactions in the presence of an external electromagnetic field .
the photon decay process @xmath18 in presence of a magnetic field was studied by several authors @xcite .
the cherenkov process @xmath19 is the cross process to photon decay and was studied in a crossed field @xcite and in a magnetic field in @xcite .
the aim of this paper is to consider the case of an arbitrary homogeneous electromagnetic field .
this case was also considered recently in @xcite but , from our point of view , the expression obtained is unfinished insofar as the very short tensorial structure of the v - a two - point amplitude , which we derive in ( [ ha9 ] ) , is nowhere visible in @xcite .
therefore it was not possible to compare our expressions to schubert s formula ( 5.15 ) .
since this topic abounds in errors and controversies ( see @xcite for a thorough discussion of the literature and its critique ) , we will devote this article mainly to a careful analysis of this expression and will postpone physical applications to another forthcoming paper .
let us begin with the effective lagrangian . at energies very much smaller than the w- and z - boson masses ,
both processes are described by an effective four - fermion interaction , @xmath21 here , @xmath22 stands for the electron field , @xmath23 , @xmath24 and @xmath25 for @xmath26 , where the first terms in @xmath27 and @xmath28 are the contributions from the w exchange diagram and the second one from the z exchange diagram .
also , @xmath29 and @xmath30 for @xmath31 .
then the amplitude for the diagram in fig.1 is given by @xmath32 where @xmath33 is the well - known polarisation operator of qed .
it was calculated earlier in @xcite .
our aim is to calculate @xmath34 . for this purpose we will use the technique developed in @xcite . in this approach
we begin with @xmath35 where @xmath36 with @xmath37 , and we have to calculate the mean value over the states x=0 : @xmath38 . for the analysis it is convenient to use a special representation for the field tensor : @xmath39 with the tensors @xmath40 and @xmath41 defined by @xmath42 where @xmath43 then . ] @xmath44 these tensors satisfy very useful identities : @xmath45 now the result of our calculation is given by @xmath46e^{-ism^{2 } } & \nonumber \\ & [ ( ck)^{\mu } ( c^{2}k)^{\nu } m_{1}+(bk)^{\mu } ( b^{2}k)^{\nu } m_{2}+ & \\ & [ ( c^{2}k)^{\mu } ( bk)^{\nu } + ( bk)^{\mu } ( c^{2}k)^{\nu } + ( kc^{2}k)b^{\mu \nu } ] m_{3 } & \nonumber \\ & [ ( b^{2}k)^{\mu } ( ck)^{\nu } + ( ck)^{\mu } ( b^{2}k)^{\nu } + ( kb^{2}k)c^{\mu \nu } ] m_{4}]\ } , & \nonumber \\\nonumber\end{aligned}\ ] ] where @xmath47 with @xmath48 as in the previous calculations in presence of a magnetic field @xcite , this result is not gauge invariant .
we also see that our amplitude is ultraviolet convergent . nevertheless , in order to preserve gauge invariance , we must regularize it using a gauge invariant regularization .
this phenomenon should be familiar from quantum electrodynamics .
for example , the photon - photon scattering amplitude , though formally converging , must be regularized to get a gauge invariant result @xcite . a convenient way of restoring gauge invariance in diagrams with pseudovector coupling is to use the pauli - villars regularization .
then it is easy to show that to obtain the final gauge invariant expression for the amplitude we must just omit the first two terms in ( [ ha7 ] ) , because they are cancelled by regulator terms .
finally , we obtain @xmath49e^{-ism^{2 } } \left\ { \right .
( ck)^{\mu } ( c^{2}k)^{\nu } m_{1}+ \nonumber \\ & & ( bk)^{\mu } ( b^{2}k)^{\nu } m_{2}+[(c^{2}k)^{\mu } ( bk)^{\nu } + ( bk)^{\mu } ( c^{2}k)^{\nu } + ( kc^{2}k)b^{\mu \nu } ] m_{3}+\\ & & [ ( b^{2}k)^{\mu } ( ck)^{\nu } + ( ck)^{\mu } ( b^{2}k)^{\nu } + ( kb^{2}k)c^{\mu \nu } ] m_{4}\left . \right\ } .\nonumber \end{aligned}\ ] ] this method of getting a gauge invariant expression allows us , unlike the authors of @xcite , to do without an anomaly cancellation mechanism at this stage of calculation .
let us note that the whole contribution of the triangle diagram is still present in ( [ ha9 ] ) .
this fact can be easily confirmed by an independent calculation of the triangle graph contribution . to deal with unwanted anomaly terms we must ,
as usual , take into account all fermions in a generation . in the following this
will always be assumed .
equation ( [ ha9 ] ) is the main result of this article .
let us now consider various checks of this expression .
first , in the limit of zero electric field it reproduces the results of @xcite .
a second nontrivial check of formula ( [ ha9 ] ) is to consider the following transformation : @xmath50 because of @xmath51 , @xmath52 , the tensor @xmath53 must be invariant under this transformation . from ( [ ha9 ] )
we see that it is indeed invariant and that under this transformation , the first term is interchanged with the second , and the third term is interchanged with the fourth , thus demonstrating the nontriviality of this check .
the next interesting limit is the case of a crossed field , where @xmath54 and @xmath55 are not only orthogonal but are also of equal modulus .
this causes the two invariants @xmath56 and @xmath57 to vanish .
the field dependence is therefore completely described by the invariant @xmath58 .
the correct procedure of finding the amplitude is to set @xmath59 first , and then take the limit @xmath60 .
using this procedure we find following expression : @xmath61 \right .
\nonumber \\ & & \int _ { 0}^{1}dv\left ( 1-v^{2}\right ) u\int _ { 0}^{\infty } dz \exp ( -i(zu+\frac{z^{3}}{3}))+\\ & & \left .
} ( f^{2}k)^{\nu } \left\ { \int _ { 0}^{1}dvu\int _ { 0}^{\infty } dz \exp ( -i(zu+\frac{z^{3}}{3}))+i\right\ } \right\}\ , , \nonumber\end{aligned}\ ] ] where @xmath62 earlier the amplitudes in crossed fields were considered in @xcite .
we can compare only parts of ( [ m3 ] ) with those existing in the literature , because all previous results were presented in contracted form and with photons on the mass shell .
taking this into account we can say that our result contradicts that of @xcite and confirms that of @xcite .
we have presented the @xmath20 amplitude in the presence of an external homogeneous constant electromagnetic field .
the result is gauge invariant and reproduces the known results for an external magnetic field .
we further identified the crossed - field limit .
let us also emphasize that the general case of electric and magnetic fields is in need when considering ( low - frequency ) multiphoton one - loop processes with and without external fields .
the situation is similar to processes where the heisenberg - euler lagrangian turns out to be useful as e.g. , in the study of photon - splitting @xcite or in photon - neutrino processes as discussed in the recent article by dicus and repko @xcite .
hence we conclude that our results allow for a variety of applications that will also be discussed in a future publication .
i would like to thank professor w. dittrich for helpful discussions and for carefully reading the manuscript .
i am also grateful to dr h. gies for valuable comments and suggestions .
this work was supported by deutsche forschungsgemeinschaft under dfg di 200/5 - 1 .
r.shaisultanov , phys.rev.lett . * 80 * , 1586 ( 1998 ) m.v .
chistyakov , a.v .
kuznetsov and n.v .
mikheev , hep - ph/9808246 tzuu - kang chyi at all , hep - ph/9907384 d. v. galtsov and n. s. nikitina , zh .
theor . fiz .
62 , 2008 ( 1972 ) [ sov
jetp 35 , 1047 ( 1972 ) ] l. l. deraad jr .
, k. a. milton , and n. d. hari dass , phys .
d * 14 * , 3326 ( 1976 ) v. v. skobelev , zh .
71 , 1263 ( 1976 ) [ sov .
jetp 44 , 660 ( 1976 ) ] a.n .
ioannisian and g.g .
raffelt , phys .
d * 55 * , 7038 ( 1997 ) c.schubert , hep - ph/9905525 i.a.batalin and a.e.shabad , zh .
60 , 894 ( 1971 ) [ sov .
jetp 33 , 483 ( 1971 ) ] w.tsai , phys .
d * 10 * , 2699(1974 ) w.dittrich and m.reuter , effective lagrangians in quantum electrodynamics ; lecture notes in physics 220 ( springer - verlag , berlin , 1985 ) v.n.baier , v.m.katkov and v.strakhovenko , zh . eksp .
68 , 405 ( 1975 ) [ sov .
jetp 41 , 198 ( 1975 ) ] ; j. schwinger , particles , sources , and fields .
( perseus books , 1998 . 3v . ) w.tsai and a.yildiz , phys .
d * 8 * , 3446(1973 ) ; w.tsai , phys .
d * 8 * , 3460(1973 ) v.n.baier , v.m.katkov and v.strakhovenko , zh .
67 , 453 ( 1974 ) [ sov .
jetp 40 , n2 ( 1975 ) ] ; v. b. berestetskii , e. m. lifshitz , l. p. pitaevskii , quantum electrodynamics .
( pergamon , 1982 ) s.treiman , r. jackiw , b. zumino , e. witten , current algebra and anomalies .
( world scientific , 1985 ) a.a .
gvozdev , n.v .
mikheev and l.a .
vassilevskaya , phys .
d * 54 * , 5674 ( 1996 ) s. l. adler , j. n. bahcall , c. g. callan and m. n. rosenbluth , phys . rev .
* 25 * ( 1970 ) 1061 ; s.l.adler , ann.phys . *
67 * , 599 ( 1971 ) d.a.dicus and w.w.repko , phys.rev.lett . *
79 * , 569 ( 1997 ) | neutrino - photon interactions in the presence of an external homogeneous constant electromagnetic field are studied .
the @xmath0 amplitude is calculated in an electromagnetic field of the general type , when the two field invariants are nonzero . |
classical charged particles , confined in a two - dimensional ( 2d ) layer and interacting via the usual three - dimensional coulomb potential , exhibit a crystallization into a wigner hexagonal structure , when kinetic energy is small compared to potential energy.@xcite we shall be interested here in bilayer systems , that describe several properties of real physical systems in condensed and soft matter , such as semiconductors,@xcite quantum dots,@xcite boron nitride,@xcite laser - cooled trapped ion plasmas,@xcite dusty plasmas@xcite and colloids.@xcite for a recent review of numerical methods for quasi-2d systems with long - range interactions , see ref.@xcite in addition , the creation of a bilayer wigner crystal on two charged plates at some distance is of primary importance in the study of `` anomalous '' strong - coupling effects such as like - charge attraction or overcharging.@xcite in this paper , we study the ground - state properties of a classical one - component plasma of identical coulombic particles of the charge @xmath6 , evenly distributed between two plates of the same homogeneous fixed charge density @xmath7 which are at distance @xmath0 .
the total surface density of the particles is @xmath8 , the particle density in each layer is @xmath9 .
the overall electroneutrality of the system is ensured by the condition @xmath10 .
the phase diagram of the system at temperature @xmath11 is determined by a single dimensionless parameter @xmath12 . by comparing the static energy of various lattices
, five distinct phases were detected to be stable ( providing global minimum of the energy ) in different ranges of @xmath13.@xcite in order of increasing @xmath13 , these phases are : a hexagonal lattice ( i ) for @xmath14 $ ] , a staggered rectangular lattice ( ii ) for @xmath15 $ ] , a staggered square lattice ( iii ) for @xmath16 $ ] , a staggered rhombic lattice ( iv ) for @xmath17 $ ] and a staggered hexagonal lattice ( v ) for @xmath18 $ ] ; although we use an index @xmath19 in @xmath20 , the transition point @xmath20 from one structure to the other is not necessarily a critical point .
the structures are pictured in fig .
[ fig : structures ] .
the different symbols correspond to particle positions on the opposite surfaces .
the primitive translation vectors of the bravais lattice on one of the surfaces are denoted by @xmath21 and @xmath22 .
the ground - state structures i , iii and v are `` rigid '' , i.e. they have fixed ( @xmath13-independent ) primary cells within their region of stability .
the structures ii and iv are `` soft '' , i.e. the shape of their primary cells is varying with increasing @xmath13 , within their region of stability .
we now outline the basic characteristics of the structures.@xcite * * structures i , ii and iii : * within one single layer , the structure corresponds to a rectangular lattice with the aspect ratio @xmath23 .
the equivalent structures on the two layers are shifted with respect to one another by a half period , i.e. by @xmath24 .
+ structure i with @xmath25 arises naturally in the simple limit @xmath26 , where the bilayer structure reduces to a single layer , which is known to crystallize in a hexagonal ( equilateral triangular ) lattice.@xcite an open question is whether phase i ( with the fixed aspect ratio @xmath25 ) exists only at @xmath27 or is stable also in a finite interval @xmath28 $ ] with some @xmath29 .
some numerical calculations indicate very small , but nonzero values of @xmath30 ( ewald technique@xcite ) and @xmath31 ( monte carlo simulations@xcite ) . on the other hand ,
another study for yukawa bilayers in the limit of infinite screening length indicates that @xmath32 , so that a buckled phase of type ii preempts structure i when @xmath13 is small but non vanishing.@xcite + structure ii continuously interpolates between the rigid structures i and iii .
the value of the aspect ratio @xmath33 then changes smoothly from @xmath34 at @xmath35 ( phase i ) to @xmath36 at the transition point @xmath37 to phase iii .
it is not clear whether or not @xmath35 , zero or nonzero , is a standard transition point between phases i and ii . the transition between phases ii and iii is continuous ( of second order ) . *
the structure iv is characterized by an angle @xmath38 between primitive cell vectors @xmath21 and @xmath22 of the same length @xmath39 .
increasing @xmath13 , the angle @xmath40 changes continuously from @xmath41 at @xmath42 ( continuous transition between phases iii and iv ) up to @xmath43 , where it drops to @xmath44 .
together with an additional shift between the sub - lattices on the two layers , this corresponds to a discontinuous ( first order ) transition to phase v. * the presence of structure v is expected for large enough @xmath45 : at large separation between the layers , the two particle sub - lattices are only weakly coupled and so two staggered hexagonal lattices form the stable structure . in this paper , we derive new series representations of the energy for all five bilayer structures .
the derivation is based on a sequence of transformations for lattice sums of coulomb potentials plus the neutralizing background , having their origin in the general theory of jacobi theta functions.@xcite the series has excellent convergence properties , which is convenient for numerical investigations , but is also conducive to analytical progress .
it will be used to improve the specification of transition points between the phases and to solve the aforementioned controversy concerning the stability of phase i. in particular , it is shown both analytically and numerically that phase i is realized only at @xmath27 , i.e. @xmath32 . the expansions of the structure energies around second - order transition points can be done analytically which enables us to derive the critical exponents at phase transition points .
the critical behavior is of the ginzburg - landau type,@xcite with the mean - field critical index @xmath5 for the growth of the order parameters . a preliminary account of this work has appeared in ref.@xcite the paper is organized as follows .
[ sec:123 ] is devoted to a detailed derivation of the series representation of the energy for structures i - iii .
the existence of phase i at @xmath27 only is established analytically and illustrated numerically .
the second - order phase transition between phases ii and iii is then described . the energy of structure iv and the second - order phase transition between phases iii and iv are treated in sec .
[ sec:4 ] . sec .
[ sec:5 ] deals with structure v and the first - order phase transition between phases iv and v , while our conclusions are finally presented in sec .
[ sec : conclusion ] .
structures i , ii and iii are treated on equal footings by considering the general case of structure ii , see fig . [ fig : structures ] .
for one single layer , the 2d lattice points are indexed by @xmath46 , where @xmath47 are any two integers ( positive , negative or zero ) and @xmath48 are the primitive translation vectors of the bravais lattice .
the lattice spacing @xmath39 is determined by the electroneutrality condition @xmath10 with the one - layer particle density @xmath49 .
the aspect ratio @xmath33 is a continuous parameter in the interval @xmath50 $ ] ; as was already mentioned , the limiting cases @xmath34 and @xmath36 correspond to the phases i and iii , respectively .
the dielectric constant of the medium is set to unity for simplicity , and we start by a preliminary remark , valid for all phases . our goal is to compute the total electrostatic energy , including particle - particle , particle - plate , and plate - plate interactions .
the latter two contributions per unit surface can be derived straightforwardly , and respectively read @xmath51 and @xmath52 .
the sum of both , @xmath53 , thus gives @xmath54 of the particle - plate energy , and this is why in the subsequent analysis , we shall add to the non trivial particle - particle energy one half of the particle - plate energy ( also referred to as the particle - background term ) .
the resulting sum provides the full energy of the system .
the energy per particle @xmath55 of the bilayer system consists of the intralayer and interlayer contributions , @xmath56 we first consider the intralayer contribution .
it is well known that lattice sums involving the pair coulomb interactions exhibit infinities which are canceled exactly by the neutralizing background term.@xcite to maintain mathematical rigor , we first restrict ourselves to a disk of finite radius @xmath57 around a reference particle localized at the origin @xmath58 .
the interaction energy due to the discrete wigner crystal is given by @xmath59 hereinafter , the omission of the lower and upper values for integer indices @xmath47 automatically means a summation from @xmath60 to @xmath2 .
the interaction of the reference particle with the 2d charge background in the disk is expressed as @xmath61 @xmath62 is the sum of ( [ contr1 ] ) plus ( [ contr2 ] ) .
we intend to rewrite @xmath62 by using the gamma identity @xmath63 a common procedure in the field.@xcite each term @xmath64 in ( [ contr1 ] ) can consequently be written as @xmath65 as concerns the background contribution ( [ contr2 ] ) , the application of the identity ( [ gamma ] ) to the term @xmath66 under integration leads to @xmath67 .\ ] ] altogether , we get @xmath68 \bigg\ } , \qquad j^2 + k^2 \delta^2 \le \left ( \frac{r}{a } \right)^2 .\ ] ] here , the subtraction of unity is due to the absence of the term @xmath69 in the sum ( [ contr1 ] ) . having all contributions under the same integration
, we are allowed to take the limit @xmath70 , which removes the exponentially small term @xmath71 $ ] and the disk constraint for lattice indices .
using the definition of the jacobi theta function with zero argument@xcite @xmath72 and making the substitution @xmath73 , we end up with the result @xmath74 .\ ] ] we shall repeatedly use the poisson summation formula @xmath75 the asymptotic behaviors @xmath76 follow immediately .
we see that the background charge contribution @xmath77 correctly cancels the @xmath78 singularity of the product of two @xmath79 functions inside the square bracket in ( [ theta3 ] ) and the integral converges .
the wigner lattices on the opposite layers are shifted with respect to one another by the vector @xmath24 , see fig .
[ fig : structures ] . to obtain the interlayer contribution to the energy ,
we first consider the disk of radius @xmath57 around the ( perpendicular ) image of the reference particle on the opposite layer .
the interaction energy of the wigner crystal is given by @xmath80 the interaction with the background charge is described by @xmath81 proceeding as in the previous case and taking into account that @xmath82 , we find @xmath83 \label{eq:14}\ ] ] with the jacobi theta function @xmath84 .
it follows from eq .
( [ poisson ] ) that @xmath85 so that the integral in eq .
( [ eq:14 ] ) converges , as it should .
the total energy per particle @xmath55 reads @xmath86 + e^{-\eta^2 t } \left [ \theta_2(e^{-t\delta } ) \theta_2(e^{-t/\delta } ) - \frac{\pi}{t } \right ] \bigg\ } .\end{aligned}\ ] ] note the invariance of @xmath55 with respect to the transformation @xmath87 , which is physically clear from the configuration sketched in fig [ fig : structures ] ( label exchange of the two bravais vectors @xmath21 and @xmath22 ) . from a numerical point of view
, there are two `` dangerous '' limits : @xmath78 and @xmath88 , that jeopardize accuracy . to simplify the integral representation of @xmath55 , we split the range of integration into two parts , @xmath89 $ ] and @xmath90 $ ] , and transform the integral over @xmath90 $ ] to the one over @xmath89 $ ] by using the poisson formula ( [ poisson ] ) .
for the term containing the product of two @xmath79 functions , the procedure reads @xmath91 & \equiv & \int_{\pi}^{\infty } \frac{dt}{\sqrt{t } } \left [ \sum_j e^{-t j^2\delta } \sum_k e^{-t k^2/\delta } - 1 - \frac{\pi}{t } \right ] \nonumber \\ & = & \int_{\pi}^{\infty } \frac{dt}{\sqrt{t } } \left [ \frac{\pi}{t } \sum_j e^{-(\pi j)^2/(\delta t ) } \sum_k
e^{-(\pi k)^2\delta / t } - 1 - \frac{\pi}{t } \right ] \nonumber \\ & = & \int_0^{\pi } \frac{\pi dt'}{(t')^{3/2 } } \left [ \frac{t'}{\pi } \sum_j e^{-t ' j^2/\delta } \sum_k e^{-t ' k^2\delta } - 1 - \frac{t'}{\pi } \right ] \nonumber \\ & \equiv & \int_0^{\pi } \frac{dt}{\sqrt{t } } \left [ \theta_3(e^{-t\delta } ) \theta_3(e^{-t/\delta } ) - 1 - \frac{\pi}{t } \right ] .\end{aligned}\ ] ] here , going from the third integral to the fourth one , we applied the substitution @xmath92 .
similarly , we have @xmath93 & \equiv & \int_{\pi}^{\infty } \frac{dt}{\sqrt{t } } e^{-\eta^2 t } \left [ \sum_j e^{-t \left ( j-\frac{1}{2}\right)^2\delta } \sum_k e^{-t \left ( k-\frac{1}{2}\right)^2/\delta } - \frac{\pi}{t } \right ] \nonumber \\ & = & \int_{\pi}^{\infty } \frac{dt}{\sqrt{t } } e^{-\eta^2 t } \left [ \frac{\pi}{t } \sum_j ( -1)^j e^{-(\pi j)^2/(\delta t ) } \sum_k ( -1)^k e^{-(\pi k)^2\delta / t } - \frac{\pi}{t } \right ] \nonumber \\ & = & \int_0^{\pi } \frac{\pi dt'}{(t')^{3/2 } } e^{-(\pi\eta)^2/t ' } \left [ \frac{t'}{\pi } \sum_j ( -1)^j e^{-t ' j^2/\delta } \sum_k ( -1)^k e^{-t ' k^2\delta } - \frac{t'}{\pi } \right ] \nonumber \\ & \equiv & \int_0^{\pi } \frac{dt}{\sqrt{t } } e^{-(\pi\eta)^2/t } \left [ \theta_4(e^{-t\delta } ) \theta_4(e^{-t/\delta } ) - 1 \right ] , \end{aligned}\ ] ] where the jacobi theta function @xmath94 .
the asymptotic behaviors @xmath95 ensure the convergence of the resulting integral . to summarize this paragraph , the total energy ( [ e ] ) can be rewritten as an integral over the finite interval @xmath89 $ ] as follows @xmath96 \nonumber \\ & & + e^{-\eta^2 t } \left [ \theta_2(e^{-t\delta } ) \theta_2(e^{-t/\delta } ) - \frac{\pi}{t } \right ] + e^{-(\pi\eta)^2/t } \left [ \theta_4(e^{-t\delta } ) \theta_4(e^{-t/\delta } ) - 1 \right ] \bigg\ } .
\label{ee}\end{aligned}\ ] ] the exact cancellation of singular terms near @xmath97 in this expression for @xmath55 represents a numerical obstacle that should be circumvented . to accomplish the cancellation analytically
, we shall consider the series representations of jacobi theta functions and apply to them the poisson transformation formula ( [ poisson ] ) ; after subtracting explicitly the singular term , the result appears as a series of special functions . in particular , for the first term in the integral ( [ ee ] ) we obtain @xmath98 & = & \int_0^{\pi } \frac{dt}{\sqrt{t } } \left [ \sum_{j , k } e^{-t j^2\delta } e^{-t k^2/\delta } - \frac{\pi}{t } \right ] - 2\sqrt{\pi } \nonumber \\ & = & \int_0^{\pi } dt \frac{\pi}{t^{3/2 } } \left [ \sum_{j , k } e^{-(\pi j)^2/(\delta t ) } e^{-(\pi k)^2\delta / t } - 1 \right ] - 2\sqrt{\pi } .
\label{rep}\end{aligned}\ ] ] the subtraction of the singularity is equivalent to the omission of the term @xmath69 from the summation . using the substitution @xmath99 and introducing the function
@xmath100 the last expression in eq .
( [ rep ] ) can be written as @xmath101 + 4 \sum_{j , k=1}^{\infty } z_{3/2}(0,j^2/\delta+k^2\delta ) - 2 \sqrt{\pi } .\ ] ] finally , performing the above procedure for all terms under integration in ( [ ee ] ) , we end up with the series representation @xmath102 + 8 \sum_{j , k=1}^{\infty } z_{3/2}(0,j^2/\delta+k^2\delta ) \nonumber \\ & & + 2 \sum_{j=1}^{\infty } ( -1)^j \left [ z_{3/2}((\pi\eta)^2,j^2/\delta ) + z_{3/2}((\pi\eta)^2,j^2\delta ) \right ] + 4 \sum_{j , k=1}^{\infty } ( -1)^j ( -1)^k z_{3/2}((\pi\eta)^2,j^2/\delta+k^2\delta ) \nonumber \\ & & + 4 \sum_{j , k=1}^{\infty } z_{3/2}(0,\eta^2+(j-1/2)^2/\delta+(k-1/2)^2\delta ) - 4 \sqrt{\pi } - \pi z_{1/2}(0,\eta^2 ) \bigg\ } . \label{series1}\end{aligned}\ ] ] the function @xmath103 with @xmath104
is related to the so - called misra function@xcite which was extensively used in single - layer lattice summations.@xcite for our bilayer system with positive @xmath13 we need the more general function ( [ z ] ) with @xmath105 .
the convergence properties of our series can be anticipated from the asymptotic relation @xmath106 in numerical calculations , for a given @xmath13 we have to find such @xmath107 which provides the minimum value of the energy ( [ series1 ] ) . in practice , the series ( [ series1 ] ) must be cut at some @xmath108 .
we document excellent convergence properties of the series ( [ series1 ] ) by considering the single - layer case @xmath109 for which the exact@xcite @xmath110 the cut at @xmath111 reproduces this exact value up to @xmath112 decimal digits , respectively .
a similar accuracy is reached in all considered cases . to be extremely accurate
, we apply the @xmath113 cut everywhere , and use the _ mathematica _ software .
another advantage of the series representation ( [ series1 ] ) is the possibility of an explicit expansion of the function @xmath114 around the controversial point @xmath27 and around the critical point @xmath37 . as will be shown , this requires an analogous taylor expansion of our @xmath115-functions .
we know@xcite that for the single layer , i.e. @xmath27 , the structure providing the minimum of the energy is the hexagonal lattice with @xmath25 . in what follows
, we shall investigate the minimum of the energy ( [ series1 ] ) in the neighborhood of the point @xmath109 .
we set @xmath116 and consider @xmath117 to be infinitesimally small . to derive a small-@xmath117 expansion of the energy ( [ series1 ] ) , we first perform this task for its series components . from the integral definition ( [ z ] )
it is easy to show that the @xmath115-functions under consideration exhibit an analytic ( taylor ) expansion in @xmath117 of the form @xmath118 @xmath119 + { \cal o}(\epsilon^3 ) , \ ] ] @xmath120 + { \cal o}(\epsilon^3 ) .
\end{aligned}\ ] ] similar expansions can be derived for @xmath121 , @xmath122 , etc . inserting these expansions into ( [ series1 ] )
, we obtain @xmath123 where the explicit form of the prefactor functions @xmath124 and @xmath125 is written in the appendix . [
tb ] /(e^2\sqrt{n})$ ] versus @xmath126 , calculated numerically by using ( [ series1 ] ) for two small values of @xmath13 : a ) @xmath127 and b ) @xmath128 . for the range of aspect ratios chosen , these curves are indistinguishable from the analytical prediction @xmath129 , stemming from eqs .
( [ epsilon1 ] ) and ( [ eq:29 ] ) .
the values of @xmath130 , which provide the energy minimum in the asymptotic limit @xmath26 according to ( [ transitionitoii ] ) , are depicted by the vertical dashed lines for comparison.,title="fig:",scaledwidth=43.0% ] /(e^2\sqrt{n})$ ] versus @xmath126 , calculated numerically by using ( [ series1 ] ) for two small values of @xmath13 : a ) @xmath127 and b ) @xmath128 . for the range of aspect ratios chosen ,
these curves are indistinguishable from the analytical prediction @xmath129 , stemming from eqs .
( [ epsilon1 ] ) and ( [ eq:29 ] ) .
the values of @xmath130 , which provide the energy minimum in the asymptotic limit @xmath26 according to ( [ transitionitoii ] ) , are depicted by the vertical dashed lines for comparison.,title="fig:",scaledwidth=43.0% ] being close to the point @xmath109 , we are interested in the small-@xmath13 behavior of the functions @xmath124 and @xmath125 .
the corresponding taylor expansions in powers of @xmath131 can be performed explicitly , too .
the explicit form of @xmath132 is here immaterial .
we find that @xmath133 , and , with a high precision , @xmath134 for a fixed @xmath13 , the extremum of the energy ( [ epsilon1 ] ) appears at @xmath135 given by the stationarity condition @xmath136 namely , @xmath137 since @xmath138 , the extremum is the minimum .
the result ( [ transitionitoii ] ) tells us that howsoever small the dimensionless distance @xmath13 is , the buckled structure ii with @xmath139 takes place . in other words , the structure i exists only strictly at @xmath32 .
the fact that in the previous works@xcite the structure i was detected also for very small positive values of @xmath13 is probably related to extremely small values of the deviation @xmath140 for these @xmath13 s , which are `` invisible '' by standard numerical methods .
like for instance , the structure i border reported in@xcite @xmath30 corresponds to @xmath141 . in fig .
[ fig : smalleta ] , we present the plots of the difference between the dimensionless energies @xmath142/(e^2\sqrt{n})$ ] versus @xmath126 , calculated numerically by using ( [ series1 ] ) for two very small values of @xmath13 : a ) @xmath127 and b ) @xmath128 which are well below / above the previous estimate of the phase i threshold,@xcite respectively .
alternatively , using the analytical expressions ( [ epsilon1 ] ) and ( [ eq:29 ] ) leads to the very same data .
the nonzero values of @xmath130 , which provide the energy minima in the asymptotic limit @xmath26 according to formula ( [ transitionitoii ] ) , are depicted by the dashed lines for comparison .
we see that the energy minima fit well with the expected @xmath130 which is clear evidence for the phase i instability .
note extremely small values @xmath143 of the energy difference , in fig .
[ fig : smalleta ] , which justifies the derivation of an accurate formula for the coulombic energy . in fig .
[ fig : zoom12 ] , the asymptotic relation ( [ transitionitoii ] ) ( dashed line ) is tested against numerical minimization of the energy ( [ series1 ] ) ( solid curve ) for small and intermediary values of @xmath13 , in the logarithmic scale .
the agreement is very good , not only for small @xmath13 , but in the whole range of stability of phase ii ( it will be shown in the next subsection that phase ii border is given by @xmath144 ) .
going from phase i to phase ii is not a phase transition in the usual sense .
however , the symmetry of the energy @xmath55 with respect to the transformation @xmath87 has the fixed ( self dual ) point at @xmath145 which is the critical point of the phase transition from phase ii to iii .
let us parameterize @xmath33 as follows @xmath146 .
the symmetry @xmath87 is now equivalent to @xmath147 , i.e. the energy @xmath55 is an even function of @xmath117 .
the expansion of @xmath55 around the critical point @xmath145 @xmath148 in small deviation @xmath117 follows from the representation ( [ series1 ] ) : @xmath149 the explicit form of @xmath150 is written in the appendix and @xmath151 is not presented due to lack of space , but has been derived .
the energy ( [ phase23 ] ) has the ginzburg - landau form , @xmath117 being the order parameter .
in contrast to that mean field theory , the expression for our energy is exact .
[ t ] ) ( dashed line ) against numerical minimization of the energy ( [ series1 ] ) ( solid curve ) , in the logarithmic scale.,scaledwidth=100.0% ] ) ( dashed line ) against numerical minimization of the energy ( [ series1 ] ) ( solid curve ) , in the logarithmic scale.,scaledwidth=100.0% ] the critical point is associated with the vanishing of the prefactor to @xmath152 , @xmath153 the values of @xmath37 obtained in the previous studies were @xmath154,@xcitewhich is remarkably precise , @xmath155@xcite and @xmath156.@xcite the functions @xmath150 and @xmath151 exhibit the following expansions around the critical @xmath37 : @xmath157 the extremum ( minimum ) of the energy ( [ phase23 ] ) appears at @xmath158 given by the condition @xmath159 for @xmath160 ( the `` ordered '' phase ii ) , we have one trivial solution @xmath161 which however provides the local maximum of the energy .
there exist two conjugate nontrivial solutions which yield the needed energy minimum ; considering one of these solutions , we arrive at @xmath162 the critical index @xmath163 , describing the growth of the order parameter from its zero critical value via @xmath164 , has the mean field value @xmath54 . in fig .
[ fig : zoom23 ] , in the logarithmic scale , the asymptotic relation ( [ transitioniitoiii ] ) ( dashed line ) is compared with the numerical minimization of the energy ( [ series1 ] ) ( solid curve ) . for @xmath165 ( the `` disordered '' phase iii )
, we have the only solution to ( [ eqn ] ) @xmath161 ( or equivalently @xmath166 ) , i.e. the rigid phase iii is stable , up to a transition to phase iv discussed in the next section .
the plot of the lattice aspect ratio @xmath107 versus @xmath13 , obtained by the numerical minimization of the energy ( [ series1 ] ) in the whole stability range of the phase ii , is pictured by the solid curve in fig .
[ fig : deltavseta ] . @xmath107 changes from @xmath34 at @xmath27 to @xmath36 at @xmath167 .
numerical data of goldoni and peeters@xcite ( open circles ) are also presented for comparison .
the asymptotic relations ( [ transitionitoii ] ) for @xmath26 and ( [ transitioniitoiii ] ) for @xmath168 are also provided , for completeness .
[ t ] versus @xmath13 , obtained by the numerical minimization of the energy ( [ series1 ] ) , is pictured by the solid curve .
numerical data of ref.@xcite are presented by open circles .
the asymptotic relations ( [ transitionitoii ] ) for @xmath26 and ( [ transitioniitoiii ] ) for @xmath168 are depicted by dashed curves.,scaledwidth=50.0% ]
in each of the two layers of phase iv ( see fig . [
fig : structures ] ) , the elementary cell is the rhombus with angle @xmath40 between the primitive translation vectors @xmath169 the lattice spacing @xmath39 is determined by the electroneutrality condition @xmath10 ; there is just one particle per rhombus of the surface @xmath170 and so @xmath171 .
the special case of @xmath172 corresponds to phase iii . as before , the energy per particle @xmath55 of the bilayer structure consists of the intralayer and interlayer contributions , @xmath173 . as concerns
the intralayer part , the 2d lattice vectors on one layer are indexed with respect to a reference particle on the same layer by @xmath174 , where @xmath175 are any two integers except for @xmath58 . the square of the lattice vector can be written as @xmath176 .\ ] ] this formula represents a kind of `` diagonalization '' of @xmath177 in indices .
if @xmath178 is an even integer , we introduce new indices @xmath179 and @xmath180 covering all integers except for @xmath181 .
if @xmath178 is an odd integer , we introduce indices @xmath182 and @xmath183 covering all integers
. thus the interaction energy due to the wigner crystal can be expressed as @xmath184 .\end{aligned}\ ] ] adding to this expression the interaction with the neutralizing background and using the gamma identity ( [ gamma ] ) in close analogy with the previous section , we obtain @xmath185 + \left [ \theta_2(e^{-t\delta } ) \theta_2(e^{-t/\delta } ) - \frac{\pi}{t } \right ] \right\ } , \ ] ] where @xmath186 .
the wigner lattices on the opposite layers are shifted with respect to one another by the vector @xmath24 . to determine the interlayer contribution to the energy
, we first consider the square of the vector between the reference particle on one layer and the vertices of the wigner crystal on the other layer at distance @xmath0 : @xmath187 + d^2 \nonumber \\ & = & a^2 \left [ ( j+k-1)^2 \cos^2(\varphi/2 ) + ( j - k)^2 \sin^2(\varphi/2 ) + ( d / a)^2 \right ] .\end{aligned}\ ] ] thus the interaction energy with the wigner crystal reads @xmath188 .\end{aligned}\ ] ] adding the background term and using the gamma identity , we find that @xmath189 + \left [ \theta_2(e^{-t\delta } ) \theta_3(e^{-t/\delta } ) - \frac{\pi}{t } \right ] \right\ } .\ ] ] the total energy per particle @xmath55 is the sum of ( [ 4intra ] ) and ( [ 4inter ] ) . note the invariance of @xmath55 with respect to the transformation @xmath190 . with respect to the definition of @xmath186 , this symmetry is equivalent to the obvious one @xmath191 . following subsequently similar lines as in previous sections
, the integral range @xmath192 $ ] can be reduced to @xmath89 $ ] by using the poisson formula ( [ poisson ] ) , @xmath193 + \left [ \theta_2(e^{-t\delta } ) \theta_2(e^{-t/\delta } ) - \frac{\pi}{t } \right ] + \left [ \theta_4(e^{-t\delta } ) \theta_4(e^{-t/\delta } ) - 1 \right ] \nonumber \\ & & + e^{-\eta^2 t/2 } \left [ \theta_3(e^{-t\delta } ) \theta_2(e^{-t/\delta } ) - \frac{\pi}{t } \right ] + e^{-(\pi\eta)^2/(2 t ) } \left [ \theta_3(e^{-t\delta } ) \theta_4(e^{-t/\delta } ) - 1 \right ] \nonumber \\ & & + e^{-\eta^2 t/2 } \left [ \theta_2(e^{-t\delta } ) \theta_3(e^{-t/\delta } ) - \frac{\pi}{t } \right ] + e^{-(\pi\eta)^2/(2 t ) } \left [ \theta_4(e^{-t\delta } ) \theta_3(e^{-t/\delta } ) - 1 \right ] \bigg\ } .\end{aligned}\ ] ] applying again the poisson transformation formula ( [ poisson ] ) to the series representations of jacobi theta functions , the singular @xmath78 terms are canceled explicitly and we end up with the representation of the energy per particle @xmath55 in terms of @xmath115-functions defined in ( [ z ] ) : @xmath194 \left [ z_{3/2}(0,j^2/\delta ) + z_{3/2}(0,j^2\delta ) \right ] + 2 \sum_{j , k=1}^{\infty } [ 2+(-1)^j ( -1)^k ] z_{3/2}(0,j^2/\delta+k^2\delta ) \nonumber \\ & & + 2 \sum_{j , k=1}^{\infty } z_{3/2}(0,(j-1/2)^2/\delta+(k-1/2)^2\delta ) + \sum_{j=1}^{\infty } [ 1+(-1)^j ] \left [ z_{3/2}((\pi\eta)^2/2,j^2/\delta ) + z_{3/2}((\pi\eta)^2/2,j^2\delta ) \right ] \nonumber \\ & & + 2 \sum_{j , k=1}^{\infty } [ ( -1)^j+(-1)^k ] z_{3/2}((\pi\eta)^2/2,j^2/\delta+k^2\delta ) \nonumber \\ & & + 2 \sum_{j , k=1}^{\infty } \left [ z_{3/2}(0,\eta^2/2+(j-1/2)^2/\delta+k^2\delta ) + z_{3/2}(0,\eta^2/2+(j-1/2)^2\delta+k^2/\delta ) \right ] \nonumber \\ & & + \sum_{j=1}^{\infty } \left [ z_{3/2}(0,\eta^2/2+(j-1/2)^2/\delta ) + z_{3/2}(0,\eta^2/2+(j-1/2)^2\delta ) \right ] - 3 \sqrt{\pi } - \pi z_{1/2}(0,\eta^2/2 ) \bigg\ } .
\label{series2}\end{aligned}\ ] ] the symmetry of the energy @xmath55 with respect to the transformation @xmath190 has the fixed point at @xmath195 which is the critical point of the phase transition between the phases iii and iv .
parameterizing @xmath196 as @xmath197 , the symmetry takes form @xmath147 and the energy @xmath55 is an even function of @xmath117 .
the expansion of @xmath55 around the critical point @xmath195 ( equivalent to @xmath198 or @xmath199 ) in small @xmath117 follows from the representation ( [ series2 ] ) : @xmath200 the explicit form of @xmath201 is presented in the appendix ; the expression for @xmath202 is too lengthy to be given , but is at our disposal .
the critical point is associated with the vanishing of the prefactor of @xmath152 , @xmath203 the values of @xmath42 obtained in the previous studies are @xmath204,@xcite @xmath205@xcite and @xmath206.@xcite the functions @xmath201 and @xmath202 exhibit the following expansions around the critical @xmath42 : @xmath207 the extremum ( minimum ) of the energy ( [ phase34 ] ) appears at @xmath208 given by the condition @xmath209 for @xmath210 ( `` disordered '' phase iii ) , we have the only solution @xmath161 ( or equivalently @xmath211 ) which provides the energy minimum , i.e. the rigid phase iii is stable . for @xmath212 ( `` ordered '' phase iv ) , the trivial solution @xmath161 becomes unstable .
the couple of conjugate nontrivial solutions , which provide the energy minimum , implies @xmath213 the critical index @xmath163 has again the mean field value @xmath54 . in fig .
[ fig : zoom34 ] , in a log - log scale , the asymptotic relation ( [ transitioniiitoiv ] ) ( dashed line ) is tested against numerical minimization of the energy ( [ series2 ] ) ( solid curve ) . in the upper inset
, we show the dependence of the energy on the logarithm of the angle parameter @xmath214 for @xmath215 , where the phase iii with @xmath195 is stable . in the lower inset ,
the analogous plot is presented for @xmath216 , where the phase iv with @xmath217 is stable ; phase iii corresponds in fact to a local maximum of the energy .
note the symmetry of the energy with respect to the transformation @xmath190 or , equivalently , @xmath218 .
the plot of the angle parameter @xmath219 versus @xmath13 , obtained by numerical minimization of the energy ( [ series1 ] ) in the whole stability range of the phase iv , is displayed by the solid curve in fig .
[ fig : deltavseta ] .
@xmath220 changes from @xmath36 at @xmath221 ( transition point from phase iii to iv ) to @xmath222 at @xmath223 ( transition point from phase iv to v , see the next section ) .
numerical data of goldoni and peeters@xcite ( open circles ) are presented for comparison . the asymptotic relation ( [ transitioniiitoiv ] ) for @xmath224 is depicted by the dashed curve .
[ t ] versus @xmath13 , obtained by the numerical minimization of the energy ( [ series2 ] ) , is shown by the solid curve .
numerical data of ref.@xcite are presented by open circles . the asymptotic relation ( [ transitioniiitoiv ] ) for @xmath224 is depicted by the dashed curve.,scaledwidth=100.0% ] versus @xmath13 , obtained by the numerical minimization of the energy ( [ series2 ] ) , is shown by the solid curve .
numerical data of ref.@xcite are presented by open circles . the asymptotic relation ( [ transitioniiitoiv ] ) for @xmath224 is depicted by the dashed curve.,scaledwidth=100.0% ]
in a single layer of the phase v ( see fig . [
fig : structures ] ) , the elementary cell of the hexagonal lattice is the rhombus with the angle @xmath44 between the primitive translation vectors @xmath225 the lattice spacing @xmath39 follows from the electroneutrality condition @xmath10 ; there is just one particle per rhombus of surface @xmath226 , so that @xmath227 .
note that the images of vertices on the opposite layer are localized in the center of triangles and not rhombuses , as was the case of phase iv .
there is no continuous way to pass from phase iv to phase v. to study the intralayer contribution to the energy of the reference particle at the origin , we first consider the wigner crystal of lattice vectors @xmath228 , where integers @xmath229 .
the square of the lattice vector is expressible as @xmath230 .\ ] ] in analogy with phase iv , we introduce new @xmath231 indices for each of the cases @xmath178 being an even and odd integer .
the interaction energy due to the wigner crystal then reads @xmath232 .\ ] ] adding to this expression the interaction with the neutralizing background and using the gamma identity , we find @xmath233 + \left [ \theta_2(e^{-t/\sqrt{3 } } ) \theta_2(e^{-t\sqrt{3 } } ) - \frac{\pi}{t } \right ] \right\ } .\ ] ] the hexagonal lattices on the opposite layers are shifted with respect to one another by the vector @xmath234 ; note that the factor @xmath235 differs from @xmath54 of the previous phases i - iv . to determine the interlayer contribution to the energy , we first consider the square of the vector between the reference particle on one layer and the vertices of the wigner crystal on the other layer at distance @xmath0 : @xmath236 + d^2 = \frac{a^2}{4 } \left [ 3 ( j+k+2/3)^2 + ( j - k)^2 \right ] + d^2 .\ ] ] going from @xmath237 to integers @xmath238 , the interaction energy with the wigner crystal on the opposite side takes the form @xmath239 .\ ] ] adding the background term , using the gamma identity and the readily derivable relations @xmath240 , \qquad \sum_j e^{-3t(j-1/6)^2 } = \frac{1}{2 } \left [ \theta_2(e^{-t/3 } ) - \theta_2(e^{-3 t } ) \right ] , \ ] ] we find that @xmath241 + \left [ \theta_2(e^{-t/\sqrt{3 } } ) \theta_2(e^{-t\sqrt{3 } } ) - \frac{\pi}{t } \right ] \right\ } .
\label{5inter}\end{aligned}\ ] ] the total energy per particle @xmath55 is given by the sum of ( [ 5intra ] ) and ( [ 5inter ] ) .
the poisson formula ( [ poisson ] ) enables us to reduce the integral range to @xmath89 $ ] , @xmath242 \nonumber \\ & & + \left ( 1 - \frac{1}{2 } e^{-(\pi\eta)^2/(2 t ) } + \frac{\sqrt{3}}{2 } e^{-3(\pi\eta)^2/(2 t ) } \right ) \left [ \theta_3(e^{-t/\sqrt{3 } } ) \theta_3(e^{-t\sqrt{3 } } ) - 1 - \frac{\pi}{t } \right ] \nonumber \\ & & + \left ( 1 - \frac{1}{2 } e^{-\eta^2 t/2 } + \frac{\sqrt{3}}{2 } e^{-3\eta^2 t/2 } \right ) \left [ \theta_2(e^{-t/\sqrt{3 } } ) \theta_2(e^{-t\sqrt{3 } } ) - \frac{\pi}{t } \right ] \nonumber \\ & & + \left ( 1 - \frac{1}{2 } e^{-(\pi\eta)^2/(2 t ) } + \frac{\sqrt{3}}{2 } e^{-3(\pi\eta)^2/(2 t ) } \right ) \left [ \theta_4(e^{-t/\sqrt{3 } } ) \theta_4(e^{-t\sqrt{3 } } ) - 1 \right ] \bigg\ } .
\label{eq:59}\end{aligned}\ ] ] in terms of the functions @xmath243
\nonumber \\ & = & 2 \sum_{j=1}^{\infty } ( -1)^j \left [ z_{3/2}(x , y+j^2/\sqrt{3 } ) + z_{3/2}(x , y+j^2\sqrt{3 } ) \right ] + 4 \sum_{j , k=1}^{\infty } ( -1)^j ( -1)^k z_{3/2}(x , y+j^2/\sqrt{3}+k^2\sqrt{3 } ) , \phantom{a}\end{aligned}\ ] ] @xmath244 \nonumber \\ & = & 2 \sum_{j=1}^{\infty } \left [ z_{3/2}(x , y+j^2/\sqrt{3 } ) + z_{3/2}(x , y+j^2\sqrt{3 } ) \right ] + 4 \sum_{j , k=1}^{\infty } z_{3/2}(x , y+j^2/\sqrt{3}+k^2\sqrt{3 } ) - \pi z_{1/2}(x , y ) , \end{aligned}\ ] ] @xmath245 \nonumber \\ & = & 4 \sum_{j , k=1}^{\infty } z_{3/2}(x , y+(j-1/2)^2/\sqrt{3}+(k-1/2)^2\sqrt{3 } ) - \pi z_{1/2}(x , y ) , \end{aligned}\ ] ]
the energy per particle @xmath55 is expressible as @xmath246 \nonumber \\ & & + \left [ i_2(0,0 ) - \frac{1}{2 } i_2((\pi\eta)^2/2,0 ) + \frac{\sqrt{3}}{2 } i_2(3(\pi\eta)^2/2,0 ) \right ] + \left [ i_4(0,0 ) - \frac{1}{2 } i_4(0,\eta^2/2 ) + \frac{\sqrt{3}}{2 } i_4(0,3\eta^2/2 ) \right ] \bigg\ } .\end{aligned}\ ] ] increasing @xmath13 from @xmath42 , phase iv is stable up to the point @xmath43 at which the energy of phase iv ( [ energy4 ] ) , evaluated at @xmath220 which minimizes this energy , equals to the energy of phase v ( [ energy5 ] ) .
our result is @xmath247 the values of @xmath43 obtained in the previous studies were relatively dispersed : @xmath248,@xcite @xmath249@xcite and @xmath250.@xcite the phase transition is of first order since the energies of phases iv and v have as functions of @xmath13 different slopes which causes the discontinuity of the first derivative of the energy with respect to @xmath13 at the transition point @xmath43 .
the angle parameter @xmath196 , which minimizes the energy of the phase iv at the critical point @xmath43 , is found to be @xmath222 . since @xmath186 , the corresponding angle @xmath251
; this angle is very close to the estimate @xmath252 of the work.@xcite going from phase iv to v , the angle skips to @xmath253 as is indicated in fig .
[ fig : deltavseta ] . two different `` sum - rules '' can be derived , that allow for a critical assessment of the results obtained .
the simplest one relies on the geometrical proximity between structures i ( a single hexagonal crystal ) and v ( two hexagonal crystals at half density ) .
for large distances , the two crystals decouple and we have , making use of straightforward notations , @xmath254 with our series representations ( [ series1 ] ) for @xmath255 and ( [ energy5 ] ) for @xmath256 , this identity holds .
another more subtle constraint follows from a combination of elementary geometric considerations,@xcite which impose that @xmath257 we have also checked that this identity holds with the expressions provided above ( note though that @xmath27 lies outside the stability range of structure v ) .
finally , it is interesting to consider both the large small and large distance behavior of the energy . for small @xmath13
, it can be shown that both structures i and ii share the same energy expansion , up to order @xmath258 included : @xmath259 where @xmath107 is the previously introduced optimal aspect ratio that minimizes @xmath260 for a given @xmath13 .
explicit calculation up to order @xmath131 shows that @xmath261 where the precise value of @xmath262 has been given in eq .
( [ eq : e0 ] ) .
we note that the linear term in ( [ eq : smalletaexp ] ) generates a contact pressure @xmath263 for @xmath26 .
a similar term was reported in,@xcite where however the term in @xmath131 differs from ours by a large factor ( 0.2122 instead of 2.5937 ) at large distances , the relevant phase is structure v , from which the inter - plate pressure follows .
the large @xmath13 case is encoded in the small @xmath264 limit of eq .
( [ eq:59 ] ) , or equivalently , eq . ( [ 5inter ] ) .
a saddle point argument leads to @xmath265 we recover an expression already obtained in,@xcite at variance with other approaches.@xcite taking the @xmath13 derivative and remembering that @xmath266 yields the inter - plate pressure @xmath267 the @xmath13 dependence is well known , since it can be written @xmath268 , with @xmath269 the modulus of the first reciprocal lattice vector .
it should be noted though that the prefactor differs from the often reported @xmath270 ( see e.g. @xcite ) , by a factor 3 .
the system of classical charged particles , forming a sequence of bilayer wigner structures at zero temperature as the distance between the plates is increasing , has a rather long history .
we have presented here a new method to calculate the coulomb ground - state energy of each wigner structure . based on a series of transformations and using general properties of the jacobi theta functions
, we expressed the energies in terms of quickly converging series of the functions @xmath103 defined in ( [ z ] ) .
the presence of the neutralizing background manifests itself simply as the subtraction of singularities of the jacobi theta functions under an auxiliary integration .
[ htb ] only ( vanishing inter - plate separation).,title="fig:",scaledwidth=52.0% ] numerical evaluation of the series requires modest computer and programming facilities , and at the same time provides extremely accurate estimates of the energy . we took advantage of this feature , supplemented by analytical work , to improve and complete previous studies in three aspects : * there was a relatively large dispersion in the determination of the transition points between phases ; this concerns especially the first - order phase transition between phases iv and v. our method improves significantly the location of all transition points , that can be worked out with arbitrary precision .
figure [ fig : summary ] gives an overview of the sequence of phases together with the corresponding thresholds .
* we resolved , analytically and numerically , a previous controversy about the stability phase i , thereby corroborating the findings of ref.@xcite we found that this phase is stable only at zero distance between the plates , @xmath271 . to confirm numerically this result
, we worked with extremely small values of the energy differences @xmath143 for distances @xmath127 and @xmath272 ( see fig .
[ fig : smalleta ] ) , which are `` invisible '' by standard numerical methods .
the agreement between the @xmath26 asymptotic relation ( [ transitionitoii ] ) , calculated analytically by using the taylor expansions of the functions @xmath103 , and the numerical minimization of the energy , presented in fig .
[ fig : zoom12 ] , is excellent . *
the expansions of the structure energies around second - order transition points can be done analytically which enables us to specify the critical phenomena at the phase transition points ; see the expansions pertaining to the transitions from phase ii to iii ( [ transitioniitoiii ] ) and from phase iii to iv ( [ transitioniiitoiv ] ) .
the agreement between these analytic formulas and numerical minimization of the ground - state energy is very good ; it can be appreciated in figs .
[ fig : zoom23 ] and [ fig : zoom34 ] .
quite expectedly for a zero temperature situation , the critical behavior is always of the ginzburg - landau type , with the mean - field critical index @xmath5 for the growth of the order parameters in the `` ordered '' phases . it is clear that our method can be directly generalized to other problems concerning the lattice summations over pair interactions , not only the coulomb ones .
the bilayers with repulsive yukawa interactions , extensively studied in the past,@xcite or with inverse power laws,@xcite deserve our attention .
we additionally emphasize that the ground states under consideration here are such that the ions are distributed evenly ( 50% on each plate ) : in other words , the ionic surface density of one wigner crystal on a given plate is @xmath273 , and coincides with the plate homogeneous surface density .
when dealing with asymmetric plates , this local neutrality assumption should be relaxed,@xcite still enforcing global neutrality .
this makes the asymmetric problem significantly more complex , and an interesting perspective for future work . finally , consideration of dielectric jumps between the walls and the interstitial slab is also a relevant venue for forthcoming investigations .
we would like to thank m. mazars and c. texier for useful discussions .
l. . is grateful to lptms for their kind hospitality .
the support received from the grants vega no .
2/0049/2012 and ce - sas qute is acknowledged .
the prefactor functions @xmath124 and @xmath125 of the expansion ( [ epsilon1 ] ) read @xmath274 + 8 \sum_{j , k=1}^{\infty } \left ( k^2 - \frac{j^3}{3 } \right ) z_{5/2}(0,j^2/\sqrt{3}+k^2\sqrt{3 } ) \nonumber \\ & & + 2 \sum_{j=1}^{\infty } ( -1)^j j^2 \left [ z_{5/2}((\pi\eta)^2,j^2\sqrt{3 } ) - \frac{1}{3 } z_{5/2}((\pi\eta)^2,j^2/\sqrt{3 } ) \right ] \nonumber \\ & & + 4 \sum_{j , k=1}^{\infty } ( -1)^j ( -1)^k \left ( k^2 - \frac{j^2}{3 } \right ) z_{5/2}((\pi\eta)^2,j^2/\sqrt{3}+k^2\sqrt{3 } ) \nonumber \\ & & + 4 \sum_{j , k=1}^{\infty } \left [ \left ( k-\frac{1}{2 } \right)^2 - \frac{1}{3 } \left ( j-\frac{1}{2 } \right)^2 \right ] z_{5/2}(0,\eta^2+(j-1/2)^2/\sqrt{3}+(k-1/2)^2\sqrt{3 } ) \bigg\ } , \end{aligned}\ ] ] @xmath275 \nonumber \\ & & + 8 \sum_{j , k=1}^{\infty } \left [ \frac{1}{2 } \left ( k^2 - \frac{j^3}{3 } \right)^2 z_{7/2}(0,j^2/\sqrt{3}+k^2\sqrt{3 } ) - \frac{j^2}{3^{3/2 } } z_{5/2}(0,j^2/\sqrt{3}+k^2\sqrt{3 } ) \right ] \nonumber \\ & & + 2 \sum_{j=1}^{\infty } ( -1)^j \left [ \frac{j^4}{2 } z_{7/2}((\pi\eta)^2,j^2\sqrt{3 } ) + \frac{j^4}{18 } z_{7/2}((\pi\eta)^2,j^2/\sqrt{3 } ) - \frac{j^2}{3^{3/2 } } z_{5/2}((\pi\eta)^2,j^2/\sqrt{3 } ) \right ] \nonumber \\ & & + 4 \sum_{j , k=1}^{\infty } ( -1)^j ( -1)^k \left [ \frac{1}{2 } \left ( k^2 - \frac{j^3}{3 } \right)^2 z_{7/2}((\pi\eta)^2,j^2/\sqrt{3}+k^2\sqrt{3 } ) - \frac{j^2}{3^{3/2 } } z_{5/2}((\pi\eta)^2,j^2/\sqrt{3}+k^2\sqrt{3 } ) \right ] \nonumber \\ & & + 4 \sum_{j , k=1}^{\infty } \bigg [ \frac{1}{2 } \left [ \left ( k-\frac{1}{2 } \right)^2 - \frac{1}{3 } \left ( j-\frac{1}{2 } \right)^2 \right]^2 z_{7/2}(0,\eta^2+(j-1/2)^2/\sqrt{3}+(k-1/2)^2\sqrt{3 } ) \nonumber \\ & & - \frac{1}{3^{3/2 } } \left ( j - \frac{1}{2 } \right)^2 z_{5/2}(0,\eta^2+(j-1/2)^2/\sqrt{3}+(k-1/2)^2\sqrt{3 } ) \bigg ] \bigg\ } .\end{aligned}\ ] ] the prefactor function @xmath150 of the expansion ( [ phase23 ] ) takes the form @xmath276 \nonumber \\ & & + 2 \sum_{j , k=1}^{\infty } \left [ ( j^2-k^2)^2 z_{7/2}(0,j^2+k^2 ) - ( j^2+k^2 ) z_{5/2}(0,j^2+k^2 ) \right ] \nonumber \\ & & + \sum_{j=1}^{\infty } ( -1)^j \left [ j^4 z_{7/2}((\pi\eta)^2,j^2 ) - j^2 z_{5/2}((\pi\eta)^2,j^2 ) \right ] \nonumber \\ & & + \sum_{j , k=1}^{\infty } ( -1)^j ( -1)^k \left [ ( j^2-k^2)^2 z_{7/2}((\pi\eta)^2,j^2+k^2 ) - ( j^2+k^2 ) z_{5/2}((\pi\eta)^2,j^2+k^2 ) \right ] \nonumber \\ & & + \sum_{j , k=1}^{\infty } \left ( \left [ ( j-1/2)^2-(k-1/2)^2 \right]^2 z_{7/2}(0,\eta^2+(j-1/2)^2+(k-1/2)^2 ) \right .
\nonumber \\ & & \left .
- \left [ ( j-1/2)^2+(k-1/2)^2 \right ]
z_{5/2}(0,\eta^2+(j-1/2)^2+(k-1/2)^2 ) \right ) \bigg\ } .\end{aligned}\ ] ] finally , the prefactor function @xmath201 of the expansion ( [ phase34 ] ) can be written @xmath277 + \sum_{j=1}^{\infty } ( -1)^j \left [ j^4 z_{7/2}(0,j^2 ) - j^2 z_{5/2}(0,j^2 ) \right ] \nonumber \\ & & + 2 \sum_{j , k=1}^{\infty } \left [ ( j^2-k^2)^2 z_{7/2}(0,j^2+k^2 ) - ( j^2+k^2 ) z_{5/2}(0,j^2+k^2 ) \right ] \nonumber \\ & & + \sum_{j , k=1}^{\infty } ( -1)^j ( -1)^k \left [ ( j^2-k^2)^2 z_{7/2}(0,j^2+k^2 ) - ( j^2+k^2 ) z_{5/2}(0,j^2+k^2 ) \right ] \nonumber \\ & & + \sum_{j , k=1}^{\infty } \big ( \left [ ( j-1/2)^2-(k-1/2)^2 \right]^2 z_{7/2}(0,(j-1/2)^2+(k-1/2)^2 ) \nonumber \\ & & - \left [ ( j-1/2)^2+(k-1/2)^2 \right ] z_{5/2}(0,(j-1/2)^2+(k-1/2)^2 ) \big ) \nonumber \\ & & + \sum_{j=1}^{\infty } \left [ j^4 z_{7/2}((\pi\eta)^2/2,j^2 ) - j^2 z_{5/2}((\pi\eta)^2/2,j^2 ) \right ] + \sum_{j=1}^{\infty } ( -1)^j \left [ j^4 z_{7/2}((\pi\eta)^2/2,j^2 ) - j^2 z_{5/2}((\pi\eta)^2/2,j^2 ) \right ] \nonumber \\ & & + 2 \sum_{j , k=1}^{\infty } ( -1)^j \left [ ( j^2-k^2)^2 z_{7/2}((\pi\eta)^2/2,j^2+k^2 ) - ( j^2+k^2 ) z_{5/2}((\pi\eta)^2/2,j^2+k^2 ) \right ] \nonumber \\ & & + \sum_{j=1}^{\infty } \left [ ( j-1/2)^4 z_{7/2}(0,\eta^2/2+(j-1/2)^2 ) - ( j-1/2)^2 z_{5/2}(0,\eta^2/2+(j-1/2)^2 ) \right ] \nonumber \\ & & + 2 \sum_{j , k=1}^{\infty } \big ( \left [ ( j-1/2)^2-k^2 \right]^2 z_{7/2}(0,\eta^2/2+(j-1/2)^2+k^2 ) \nonumber \\ & & - \left [ ( j-1/2)^2 + k^2 \right ] z_{5/2}(0,\eta^2/2+(j-1/2)^2+k^2 ) \big ) \bigg\ } .\end{aligned}\ ] ] this value is fully consistent with that reported in refs.@xcite in addition , it also provides us with the madelung constant of the hexagonal wigner crystal on a single charged plate @xmath278 , as can be found in refs . | we study the ground - state properties of a system of identical classical coulombic point particles , evenly distributed between two equivalently charged parallel plates at distance @xmath0 ; the system as a whole is electroneutral .
it was previously shown that upon increasing @xmath0 from @xmath1 to @xmath2 , five different structures of the bilayer wigner crystal become energetically favored , starting from a hexagonal lattice ( phase i , @xmath3 ) and ending at a staggered hexagonal lattice ( phase v , @xmath4 ) . in this paper
, we derive new series representations of the ground - state energy for all five bilayer structures .
the derivation is based on a sequence of transformations for lattice sums of coulomb two - particle potentials plus the neutralizing background , having their origin in the general theory of jacobi theta functions .
the new series provide convenient starting points for both analytical and numerical progress .
its convergence properties are indeed excellent : truncation at the fourth term determines in general the energy correctly up to 17 decimal digits .
the accurate series representations are used to improve the specification of transition points between the phases and to solve a controversy in previous studies . in particular , it is shown both analytically and numerically that the hexagonal phase i is stable only at @xmath3 , and not in a finite interval of small distances between the plates as was anticipated before .
the expansions of the structure energies around second - order transition points can be done analytically , which enables us to show that the critical behavior is of the ginzburg - landau type , with a mean - field critical index @xmath5 for the growth of the order parameters . |
understanding the flow inside sessile liquid structures of different shapes is important in a variety of solution - based material deposition and patterning processes .
solvent evaporation inherent in these processes is already known to lead to a rich variety of flows @xcite .
the small length scale and general lack of shape symmetry implies the potential for capillary pressure gradients and corresponding flow phenomena .
finally , the non - instantaneous nature of the formation of these liquid structures adds another element to the flow complexity . in the linked videos , we investigated the internal flow of inkjet - printed water rivulets of finite length using optical microscopy .
six millimeter - long rivulets were formed by printing a pre - determined number of drops ( @xmath0 m ) at controlled frequency ( @xmath1hz ) and spatial overlap ( @xmath2 m ) .
microscope cover slips made of borosilicate glass were surface coated and used as substrates .
two surface coatings were investigated - s1805 photoresist with and without koh etching .
rivulets were printed inside a controlled humidity chamber at two relative humidity levels - @xmath3 and @xmath4 .
the ambient , ink , and substrate were kept at room temperature ( @xmath5c ) . for fluorescent microscopy videos ,
the water was seeded with nile red polystyrene spheres ( @xmath6 m ) at @xmath7 volume fraction .
we report the experimental details and results of a wider range of both printing parameters , ink properties , and substrate surfaces elsewhere @xcite .
the two surface coatings resulted in different values of the static advancing and receding contact angles ( @xmath8 and @xmath9 ) for water and in the rivulet formation process during printing .
water droplets on the photoresist - coated surface exhibited @xmath10 and @xmath11while those on the etched surface @xmath12 and @xmath13 .
complete rivulet formation was inhibited on the first surface type due to a non - zero @xmath9 , which allowed capillary - driven contact line recession and rivulet breakup into individual droplets @xcite .
the zero @xmath9 on the second surface type prevented rivulet breakup @xcite but resulted in the formation of a distinctive bulge at the starting end of the water rivulet .
the bulge grows immediately after the coalescence of the first few drops during rivulet formation .
we investigated the flow that causes bulge growth by using optical fluorescent microscopy . during rivulet formation ,
a distinctly pulsatile axial flow drives fluid away from the terminal end of the growing rivulet where printed drops are landing .
the frequency of the flow is the same as the drop frequency .
it has been hypothesized that the large local mean curvature in the region where drops land causes a capillary pressure gradient along the length of the rivulet @xcite .
using @xmath14piv , the height- and width - averaged axial velocity was measured and showed that the amplitude of the pulsatile flow decreases with reduced relative humidity in the ambient , all other conditions being fixed .
the decreased flow magnitude results in the inhibition of bulge formation .
the video entry to the 2013 gallery of fluid motion is shown in * a schematic of the flow observation approach * visualization of rivulet break up during printing onto a substrate with non - zero receding contact angle ( surface type 1 ) * visualization of successful rivulet formation on surface type 2 and concequent bulge growth * fluorescence microscopy video showing pulsatile axial flow and corresponding mean axial speed * the effect of liquid evaporation on axial flow and bulge growth inhibition | we used optical microscopy to investigate flows inside water rivulets that were inkjet - printed onto different surfaces and under different ambient conditions .
the acquired fluid dynamics videos were submitted to the 2013 gallery of fluid motion . |
the modern technique of nanocrystal growth has created high optical quality nanocrystals with narrow size distribution@xcite that can be also chemically doped.@xcite this allows one to study the level structure of the nanocrystals using interband optical spectroscopy,@xcite tunneling spectroscopy,@xcite and the far infrared intraband spectroscopy.@xcite important additional information about the origin of these levels and their symmetry could be obtained by using optical magnetospectroscopy of the intraband and interband transitions and tunnelling magnetospectroscopy , as was done many years ago in the case of atoms , hydrogenlike shallow defects and excitons .
there is the other group of experiments , such as magneto circular dichroism,@xcite spin
flip raman scattering,@xcite time dependent faraday rotation,@xcite epr and odmr measurements,@xcite and magneto - photoluminescence,@xcite where the magnetic field is an essential component of the experimental technique .
inrepretation of all these experiments requires the knowledge of electron and hole energy spectra in magnetic fields .
a wide class of semiconductor materials can be prepared in nanocrystal form . in most cases
they have a spherical shape and can be bare semiconductor nanocrystals or onion type heterostructure nanocrystals composed of an inner semiconductor core coated with several spherical shells of different semiconductors.@xcite technology allows one to vary the radius of the nanocrystal core from 1040 in a controlled manner .
the energy spectra of the ncs are formed by the discrete quantum size levels ( qsls ) .
the qsls of the confined electrons in the conduction band are characterized by the value of the orbital momentum @xmath2 and the total angular momentum @xmath3 , similar to the bound electron states in atoms@xcite . in a zero magnetic field
, the levels are degenerate with respect to the projection of the total momentum @xmath4 .
the spectra of the nanocrystals in an external magnetic field @xmath5 is well described as an atomic zeeman effect .
a diamagnetic @xmath6 contribution to the spectra could be neglected because the nanocrystal radius , @xmath7 , is considerably smaller than the magnetic length , @xmath8 , where @xmath9 is the absolute value of a free electron charge , @xmath10 is the speed of light . in a reasonable magnetic field
@xmath11 t the diamagnetic contribution to the qsls @xmath12 . however , the effect of the external magnetic field on the electron qsls depends on the zero field spin - orbit splitting , @xmath13 , of the levels with different @xmath14 ( one can find a description of the similar effect for atoms in ref . ) .
if @xmath13 is larger than the zeeman energy , the weak magnetic field splits the electron levels : @xmath15 where @xmath16 is the bohr magneton , @xmath17 is the free electron mass , @xmath18 is the projection of the total momentum on the magnetic filed direction and @xmath19 is the effective @xmath1-factor of the corresponding state , that is an analog of the _ land_@xmath20 _ factor _ for atoms . if @xmath13 is small , the zeeman splitting can be described as the sum of the spin and orbital contributions : @xmath21 where @xmath22 and @xmath23 are the electron spin and angular momentum projections on a magnetic field direction correspondingly , and @xmath24 and @xmath25 are the spin and orbital g - factors of the electron in the corresponding state . in bare nanocrystals made of semiconductors with
a simple parabolic conduction band @xmath24 is equal to the bulk electron effective @xmath1factor @xmath26 and @xmath27 , where @xmath28 is the bulk electron effective mass .
spin flip raman scattering studies of cds nanocrystals,@xcite however , have demonstrated the dependence of the ground 1s electron state @xmath24 on the excitation energy , and thus on the energy of the electron state .
the variation of @xmath24 with energy can be connected with the energy dependence of the bulk electron @xmath1factor.@xcite the eight band kane model , that is used to describe the electron energy spectra in most direct gap semiconductors , gives the following dependence of the electron @xmath1factor @xmath29 on its energy , @xmath30:@xcite @xmath31 here @xmath32 is the band gap energy , @xmath33 is the spin
orbit splitting of the top of the valence band , @xmath34 is the energetic kane parameter , and @xmath35 , where @xmath36 is the free electron @xmath1factor and @xmath37 is the contribution of remote bands to the bulk electron @xmath1 factor .
the second term in eq . ( [ bulkg ] ) describes the negative contribution of the valence band ( with @xmath38 ) into the electron @xmath1-factor .
it decreases with increasing the energy of qsls caused by the reduction of the nanocrystals size .
thus , the energy dependence of @xmath29 in eq .
( [ bulkg ] ) is conducive to the size dependence of @xmath24 in eq .
( [ delsl ] ) , that has been measured recently in bare and core - shell cdse nanocrystals.@xcite the size dependence of @xmath24 for the ground 1s electron state in the spherical heterostructure has been calculated in ref .
within the eight band kane model .
it has been shown that in a spherical heterostructure formed by two semiconductors , @xmath39 and @xmath40 , @xmath24 is the sum of the weighted volume contributions of each material and the interface contributions : @xmath41 .
the interface term , @xmath42 , is proportional to the square of the conduction band component of the wave function , @xmath43 , at the @xmath44 heterointerface .
this calculation shows clearly that @xmath24 is very sensitive to the value of the wave function at the heterointerface and to its leakage under the barrier and thus to the boundary conditions ( bcs ) imposed on the wave function at the heterointerface . in ref .
the size dependence of the electron g - factor is calculated using standard bcs that assumes a continuity of @xmath43 at the interface .
this leads to the vanishing of the interface term , @xmath42 , in the bare semiconductor nanocrystals , that are modeled by the infinite potential barrier in the layer @xmath40 because the standard bc at the surface in this case is @xmath45 .
however , the standard bcs are not always justified even for infinite potential barriers ( see for example refs . ) . in general
, the wave function @xmath43 does not vanish at the surface of bare semiconductor nanocrystals and satisfies the general boundary conditions that are the characteristics of the particular surface.@xcite the effect of a semiconductor surface on the light absorption in indirect band semiconductor nanocrystals has been studied in ref . and was found to be significant .
in this paper we study the effect of the general boundary conditions on the qsls of an electron in zero and weak external magnetic fields .
we show that there is an additional surface induced spin orbit term for the electron states whose magnitude is proportional to @xmath46 .
this surface induced spin
orbit interaction leads the additional magnetic moment of the electrons in an external magnetic field in analogy with the additional relativistic magnetic moment of the electrons in atoms.@xcite this is conducive to the nonzero surface contribution to the electron @xmath1 factor that controls the linear zeeman splitting of the qsls .
the fitting of the experimental size dependence of the electron @xmath1 factor in the ground state@xcite has allowed us to determine the appropriate parameter of the general boundary conditions for bare cdse nanocrystals .
the spectra of excited electron states has been considered in the two limits of the relation between the zero field spin
orbit splitting and the electron energy in magnetic field similar to consideration conducted for atoms.@xcite the paper is organized in the following way : in sec .
[ theory ] we describe the eight band hamiltonians that include the external magnetic field effects . in sec .
[ spher ] we derive the general boundary conditions for the conduction band component of the envelope function at the abrupt surface ( see sec .
[ spherbc ] ) and analyze the surface effect on the electron qsls in a zero magnetic field ( see sec .
[ spherso ] ) .
the effect of the weak external magnetic field on the electron qsls is considered in sec .
[ magnetic ] .
the size dependence of the electron ground state @xmath1 factor in bare cdse nanocrystals has been calculated and the value of the surface boundary parameter has been determined by fitting experimental data in sec .
[ ground ] .
the symmetry of the electron excited states in an external magnetic field is studied in the low and strong magnetic field regimes ( see sec .
[ excited ] ) .
the results are summarized and discussed in sec .
the energy band structure of the cubic semiconductors near the center of the first brillouin zone can be well described within the eight
band @xmath47 model.@xcite in the homogenous semiconductor , the full wave function can be presented as the expansion:@xcite @xmath48 where @xmath49 and @xmath50 are the eigenfunctions of the spin operator @xmath51 , where @xmath52 are the pauli matrices , @xmath53 is the bloch function of the conduction band edge at the @xmath54point of brilluoin zone representing the eigenfunction of internal momentum @xmath55 , and @xmath56 , @xmath57 , @xmath58 are the bloch functions of the valence band edge at the @xmath54point of brillouin zone . the combination of these functions : @xmath59 and @xmath60 are the eigenfunctions of the internal momentum @xmath61 with projections @xmath62 and @xmath63 on the @xmath64 axis respectively ( see refs . ) .
the smooth functions @xmath65 are the components of the conduction band spinor envelope function @xmath66 , and @xmath67 , @xmath68 , @xmath69 are the @xmath70 components of the valence band spinor envelope vector @xmath71 . in the presence of external magnetic field the eight component envelope function @xmath72 is the solution of the schrdinger equation@xcite @xmath73 here , the energy @xmath30 is measured from the bottom of the conduction band , @xmath74 , where @xmath75 is the momentum operator and @xmath76 $ ] is the vector potential of the magnetic field . the @xmath77 unit matrix , @xmath78 , and the pauli matrices , @xmath79 , @xmath80 and @xmath81 , are acting on the spinor components of the wave functions ( @xmath82 ) .
@xmath83 is the kane matrix element describing the coupling of the conduction and valence bands ( the kane energy parameter : @xmath84 , was introduced in eq .
( [ bulkg ] ) , respectively ) . the off - diagonal matrix element in eq .
( [ hfull ] ) acts on the @xmath85 as a scalar product @xmath86 .
the conduction band part of the hamiltonian , @xmath87 , acting on the spinor function @xmath88 has the form : @xmath89 where @xmath90 is defined in eq .
( [ bulkg ] ) , and @xmath91 takes into account the contribution of remote bands to the energy
dependent electron effective mass , @xmath92 . in cubic and zinc - blende semiconductors
this dependence has the following form : @xmath93 \ , .
\label{bulkm}\end{aligned}\ ] ] in this paper we focus only on the electron qsls with energies @xmath94 .
this allows us to neglect hereafter the @xmath95 terms in the valence band part of the hamiltonian , @xmath96 , and to present it in the spherical approximation as:@xcite @xmath97 here , the hamiltonian @xmath98 should be considered as the @xmath99 matrix acting on the spinor vector @xmath85 rather than the @xmath100 matrix hamiltonian acting on the six component wave function as in ref . .
correspondingly , @xmath101 in eq .
( [ h6 ] ) is the vector operator .
it is straightforward to show that in this representation @xmath102 $ ] , where @xmath103 is an arbitrary vector .
the terms proportional to @xmath33 in hamiltonian ( [ h6 ] ) describe the effect of the spin orbit coupling and @xmath104,@xcite where @xmath105 is the magnetic luttinger parameter.@xcite one can find that @xmath106 from the relation:@xcite @xmath107 , where the effective mass luttinger parameters,@xcite @xmath108 , can be presented as : @xmath109 and @xmath110 in the approximation used above.@xcite following the perturbation approach developed in ref . , we present the hamiltonian ( [ hfull ] ) as : @xmath111 where the hamiltonian @xmath112 describes the electron states in zero magnetic field : @xmath113 and the hamiltonians @xmath114 describes the effect of a weak external magnetic field .
linear in the magnetic field terms of @xmath115 can be written : @xmath116 where @xmath117 denotes an anticommutator and @xmath118 is the envelope velocity operator .
the operators @xmath119 and @xmath120 are defined as : @xmath121 @xmath122\\ -i p \dfrac{m_0}{\hbar } [ \mbox{\boldmath $ n$ } \times \mbox{\boldmath $ r$}]\ , & 0 \end{array } \right)\hat u_2\ , , \end{aligned}\ ] ] where @xmath123 is the unit vector directed along the external magnetic field , and @xmath124 $ ] is the envelope angular momentum operator .
generally , the calculation of the linear zeeman effect for the carriers confined in a @xmath125 external potential of arbitrary shape is a straightforward theoretical problem .
the same procedure is used for an impurity potential and for a zero dimensional quantum dot of an arbitrary shape . using the unperturbed hamiltonian @xmath112 one has to find first the energy levels @xmath126 and the corresponding eigen functions @xmath127 of the carrier in the external confining potential .
then , the effect of the magnetic field can be found perturbatively : @xmath128 .\ ] ] where the wave functions should be normalized @xmath129 .
substituting @xmath130 into this expression we obtain : @xmath131 where the contribution of the valence and conduction bands to zeeman effect @xmath132 and @xmath133 correspondingly are : @xmath134 and the coupling correction is given by @xmath135 | { \bf \psi}^v \right >
\nonumber \\ & + & \left < { \psi}^c | i \frac{m_0p}{\hbar}\hat u_2 [ \mbox{\boldmath $ n$}\times \mbox{\boldmath $ r$ } ] | { \bf \psi}^v \right>^ * \ , . \label{dcoup}\end{aligned}\ ] ] in principle , this straightforward approach allows us to find the linear zeeman splitting of the energy levels in any arbitrarily shaped hetero - nanostructure .
however , the calculations of the zero field wave functions @xmath130 for arbitrary shape heterostructure is a rather cumbersome procedure and , consequently , the zeeman splitting calculation that uses these wave functions is very complicated . in the present paper , from hereafter , only spherical semiconductor nanocrystals are considered .
the high symmetry of these quantum dots allow us to calculate the zeeman splitting of the qsls .
we will consider bare semiconductor nanocrystals which surface can be modeled by an impenetrable barrier . to find unperturbed envelope wave functions that are described by the hamiltonians of eq .
( [ h0 ] ) in the bulk region one should impose an appropriate boundary conditions on these functions at the nanocrystal surface . for the impenetrable barrier the general boundary conditions ( gbcs ) have been shown in ref . to guarantee the vanishing of the normal to the surface component of the envelope flux density matrix . in a spherical nanocrystal
the normal envelope flux density matrix , @xmath136 , should vanish at any point of the nanocrystals surface : @xmath137 \right|_{r = a } = 0 \ , , \label{generalflux}\end{aligned}\ ] ] where @xmath7 is the nanocrystal radius and @xmath138 ( in spherical coordinates @xmath139 is defined as @xmath140 ) .
this general requirement of eq .
( [ generalflux ] ) should be satisfied for two arbitrarily chosen eigenfunctions @xmath141 and @xmath142 of the hamiltonian @xmath112 defined in eq .
( [ h0 ] ) with energies @xmath126 and @xmath143 , respectively . substituting the explicit expressions for the normal components of the envelope
velocities @xmath144 into eq .
( [ generalflux ] ) one obtains : @xmath145 \right|_{r = a } = 0 \ , .
\label{fluxelec}\end{aligned}\ ] ] using equation @xmath146 one can express the components of the spin vector @xmath85 through @xmath147 : @xmath148\ , , \label{psiv}\end{aligned}\ ] ] where the energy dependent electron effective mass @xmath92 and @xmath1factor @xmath29 are defined in eqs .
( [ bulkm ] ) and ( [ bulkg ] ) respectively .
the normal projection @xmath149 can then be written : @xmath150 where @xmath151 .
substituting eq .
( [ psivtau ] ) into eq .
( [ fluxelec ] ) one obtains the general requirement for the conduction band component of the envelope wave function at the nc surface : @xmath152 \right|_{r = a } = 0 .
\label{fluxelecso}\end{aligned}\ ] ] equation ( [ fluxelecso ] ) must be fulfilled at each point of the surface .
it is satisfied for all arbitrary chosen energy states @xmath153 and @xmath154 if and only if the boundary conditions for @xmath155 as well as for @xmath156 are given by : @xmath157 where @xmath158 is a real number constant independent of @xmath30 and @xmath159 is the lattice constant .
we assume that the surface of the nanocrystal also possesses the spherical symmetry and is characterized by the same bcs at any point of the surface . in this case
parameter @xmath158 in the gbcs of eq .
( [ sphbc ] ) is independent of angles @xmath160 and @xmath161 .
let us consider the effect of the gbcs of eq .
( [ sphbc ] ) on the electron qsls in the absence of a magnetic field .
( [ psiv ] ) that expresses the valence band component @xmath162 of the wave function through its conduction band component @xmath88 one can derive the bulk schrdinger equation describing @xmath88 inside a semiconductor nanocrystal : @xmath163 one can see that each spin component of @xmath88 is a solution of the standard bulk schrdinger equation with the energy dependent effective mass @xmath164 .
equation ( [ shr ] ) does not contain a spin - orbit term and its solution can always be presented as @xmath165 , where @xmath2 and @xmath166 are the angular momentum and angular momentum projection , correspondingly , @xmath167 are the spherical harmonics , and the radial function @xmath168 satisfies to the following equation : @xmath169 at the same time the gbcs of eq .
( [ sphbc ] ) mixes two electron spin sublevels , and all electron states are characterized now by the total angular momentum : @xmath170 . as a result @xmath88 is an eigen spinor of the operator of the total angular momentum : @xmath171 where @xmath172 are the spherical spinors , @xmath4 and @xmath18 are the total angular momentum and its projection respectively , and @xmath2 is the angular momentum of the electron state .
the two radial functions @xmath173 describe the two electron states with @xmath174 ( note that the states described by @xmath175 exist only for @xmath176 ) .
we use the definition of the spherical spinors as in ref .
@xmath177 where @xmath174 and @xmath178 are the clebsch - gordan coefficients . substituting eq .
( [ psic ] ) into eq .
( [ sphbc ] ) one can write the general boundary condition for the radial wave functions @xmath179 : @xmath180 \ , , \label{bcfso}\end{aligned}\ ] ] where @xmath181 for the states with @xmath182 and @xmath183 for the states with @xmath184 .
the conventionally used standard bcs for the impenetrable barrier assumes vanishing of the wave function @xmath88 at the nanocrystal surface , and thus correspond to @xmath185 .
in general , however , the conduction band component of the wave function does not vanish at the surface .
furthermore , it has been shown in ref . that the boundary condition with @xmath185 is incorrect in cases where interband coupling is important .
the appropriate value of @xmath186 should be determined from the fitting to the experimental data .
the solutions of eq .
( [ kaneel ] ) can be written as @xmath187 , where @xmath188 are spherical bessel functions , @xmath189 is the normalization constant determined by the condition @xmath190 , and @xmath191 is connected with the energy , @xmath30 , of the electron level : @xmath192 .
an equation that determined @xmath191 is obtained by substitution of @xmath179 into the bcs of eq .
( [ bcfso ] ) : @xmath193 = \label{bcl } \\ \frac{t a_0}{a } \frac{m_0}{m_c(e ) } \phi_{l , n}^\pm & & \left [ \frac{l}{2l+1 } j_{l-1}(\phi_{l , n } ) - \frac{l+1}{2l+1 } j_{l+1}(\phi_{l , n } ) \right ] \ , \ , .
\nonumber\end{aligned}\ ] ] the @xmath194 solution of this equation defines the energy of the @xmath194 electron level with the angular momentum @xmath2 and the total angular momentum @xmath174 . in the case of the standard bcs @xmath185 , and the solution
@xmath191 is the @xmath195th zero @xmath196 of the spherical bessel functions @xmath197 .
the general bcs with @xmath198 takes into account the effect of the surface on the electron energy levels that is important in small nanocrystals . in large nanocrystals satisfying the condition @xmath199 , the solutions @xmath191 are close to @xmath196 and the effect of the surface on the confined electron states is negligible .
the electron states with @xmath182 and @xmath184 have different energies for @xmath176 as a result of the general bcs of eq .
( [ bcfso ] ) .
this difference describes the surface induced spin
orbit splitting of the excited states : @xmath200 .
it can be shown that @xmath13 is positive if the spin orbit splitting of the valence band @xmath201 .
figure [ levels ] shows the effect of the general bcs on the energy of the ground @xmath0 ( @xmath202 ) and the first excited @xmath203 ( @xmath204 ) electron states in bare cdse nanocrystals .
the surface induced spin - orbit interaction splits the last state into two states with the total angular momentum @xmath205 ( @xmath206 state ) and @xmath207 ( @xmath208 state ) and the size dependence of this splitting @xmath209 is shown in fig .
[ levels](b ) .
the calculations have been performed for the following bulk parameters of cdse : @xmath210 ev and @xmath211 @xmath17 from ref . , @xmath212 ev , @xmath213 mev , and @xmath214 from ref .
, that result in @xmath215 and @xmath216 .
one can see that even a small surface parameter introduced by the gbcs , @xmath217 @xmath218 , significantly affects both ground and excited states in small nanocrystals .
the positive ( negative ) value of the surface parameter @xmath219 ( @xmath220 ) increases ( decreases ) the energy of all states in comparison with @xmath185 .
the relative order of the @xmath206 and @xmath208 states with the total momentum @xmath205 and @xmath207 always coincides with the relative order of the valence band subbands characterized by the same total momentum .
the value of the spin - orbit splitting , @xmath13 , however , remains for all excited states much smaller than the averaged energy @xmath221 of the level with the orbital momentum @xmath2 .
if @xmath222 , the averaged energies @xmath223 and corresponding wave numbers @xmath224 can be found from the simplified bcs given by eqs .
( [ bcfso ] , [ bcl ] ) with @xmath225 .
the energy corrections @xmath226 coming from the small @xmath227 can be found perturbatively . in the appendix
we obtain the following energy corrections : @xmath228 where @xmath229 .
note that although @xmath230 , the normalization constants @xmath231 .
however , in the considered approximation they can be replaced with high accuracy by the averaged constant @xmath232 that is determined by the approximate normalization condition : @xmath233 that neglects the second term in eq .
( [ psiv ] ) .
thus , if @xmath234 , @xmath235 , the energy levels of the excited states @xmath236 with @xmath14 are the eigen energies of the effective hamiltonian @xmath237 : @xmath238 where the energy dependent spin
orbit splitting @xmath239 is given by @xmath240 the approximate expressions of eqs .
( [ deltac],[sigmasur ] ) describe the exact splitting @xmath241 of the @xmath203 states in cdse nanocrystals shown in fig .
[ levels](b ) with high accuracy ( the largest error in the smallest nanocrystals is about @xmath242 for @xmath243 @xmath218 ) .
it is important to note here that while the matrix elements of the spin orbit operator @xmath244 are nonzero only for @xmath176 , eq . ( [ sigmasur ] ) defines also the surface parameter @xmath245 for the @xmath202 @xmath246 symmetry states .
the dimensionless parameter @xmath245 is proportional to the spin
orbit splitting in the valence band @xmath33 and to the square of the wave function at the surface .
it describes the surface induced spin orbit coupling of the conduction electron qsls in bare semiconductor nanocrystals . in bulk semiconductors the integration in eq .
( [ sosur ] ) results in zero spin orbit splitting of the conduction band states because it must be carried out at the remote bounding surface where the wave function is vanishing .
@xmath247 if one assumes the vanishing of the conduction band component wave function at the surface of spherical nanocrystals .
this assumption is never justified if the coupling between conduction band and valence band components is significant .
thus , the spin
orbit splitting of the electron qsls is caused by the admixture of the valence band states near the surface .
let us consider now the energy of the electron qsls , @xmath248 , in a weak magnetic field . substituting the zero field wave functions @xmath249 given by eq .
( [ psic ] ) and ( [ psiv ] ) into eqs .
( [ dec ] ) and ( [ dcoup ] ) , we obtain for the energy correction @xmath250 : latexmath:[\ ] ] this magnetic moment arises from the surface induced spin orbital term of eq .
( [ hlso ] ) which is modified by the larmor precession of the electron in an external magnetic field .
indeed , the surface energy term @xmath264 can be obtained directly by replacing @xmath265 in the spin
orbit perturbation term @xmath266 \mbox{\boldmath $ \hat \sigma$ } ) $ ] of eq .
( [ hlso ] ) .
the origin of the surface induced electron magnetic moment is similar to those of an additional relativistic magnetic moment of the electron in atoms , and their values are derived in a similar way ( see for example refs . ) .
it is necessary to note that the surface - induced magnetic moment @xmath267 does not vanish for the states with @xmath202 even when it arises from the spin
orbit coupling term @xmath268 .
all electron states that have @xmath246 symmetry are kramers doublets that are degenerate with respect to its spin projection .
the external magnetic field lifts this degeneracy and splits these states into two levels with energy : @xmath269 where the @xmath270 signs correspond to the electron states with spins parallel and antiparallel to the magnetic field direction .
the spin effective @xmath1 factor can be obtained from eq .
( [ zeeman ] ) as : @xmath271 the effective electron @xmath1 factor for the @xmath246 electron states in spherical heterostructures , @xmath272 , was first obtained in ref . where the standard bcs at the heterointerface were used to describe the size dependence of the electron @xmath1 factor . in the case of the bare semiconductor
nanocrystals the standard bc at the surface ( @xmath273 ) would lead to zero surface contribution to the effective electron @xmath1 factor .
we have examined here the effect of the general bc of eq .
( [ bcfso ] ) on the electron effective @xmath1 factor in cdse nanocrystals
. the weighted @xmath1 factor of the states with @xmath246 symmetry in eq .
( [ gs ] ) can be written : @xmath274 where the corrections @xmath133 of eq .
( [ dev ] ) to the @xmath275 of eq .
( [ bargs ] ) have been added .
figure [ sizegfact ] shows the size dependence of the ground state electron @xmath1 factor , @xmath272 , calculated using the standard bc with @xmath276 and the general bc with @xmath217 @xmath218 .
one can see that the experimental size dependence of the electron @xmath1-factor taken from ref .
is described very well by the general bc with the negative parameter @xmath277 , while the use of the positive parameter brings the theoretical curve far away from the data . the size dependence of the bulk like energy dependent @xmath1 factor , @xmath278 , calculated with the help of eq .
( [ bulkg ] ) for @xmath185 is also shown in fig .
[ sizegfact ] .
one can see that the bulk @xmath1factor @xmath26 fails to describe the experimental data even in the largest dots . at the same time
, the difference between two @xmath24 curves calculated with @xmath277 @xmath218 and @xmath279 is not very large .
this is because the general bcs affects the electron g - factor in two ways : indirectly , through the change of the zero field energy @xmath280 decreasing / increasing the electron g - factor , and directly , through the surface contribution @xmath281 , that is always positive . in the case of the negative @xmath220
the gbcs decreases the energy of the ground electron state decreasing the electron g - factor , and these two effects partly compensate for each other .
however , the surface contribution @xmath281 calculated with @xmath282 @xmath218 is significant in small nanocrystals ( see fig .
[ sizesurf ] ) .
the fine structure of the excited electron states in an external magnetic field is very sensitive to the relation between @xmath239 and the energy of the orbital magnetic moment @xmath283 . in the _ low field regime _
, when the magnetic energy is smaller than the separation between the electron states with @xmath284 , the external magnetic field satisfying the condition @xmath285 splits the levels according to projection @xmath18 of the full momentum @xmath4 on the magnetic field as : @xmath286 for @xmath182 and @xmath287 for @xmath184 .
the effective electron @xmath1factors @xmath288 are the analogs of the lande factors for electrons in atoms in the case of the `` anomalous '' zeeman effect.@xcite the surface contributions to the electron @xmath1factors ( @xmath289 ) are similar to the relativistic corrections in refs . . when the zero field fine structure splitting becomes smaller than the energy of the orbital magnetic moment , the magnetic field zeeman term mixes the states with different @xmath284 . in this case the projections of the spin momentum @xmath290 and orbital momentum @xmath166 on the direction of the magnetic field are a more convenient notation for describing the electron state fine structure . the spin
orbit term @xmath291 and the additional surface moment @xmath292 may , however , mix the states with different values of the @xmath293 product .
let us consider the matrix elements of the surface magnetic moment operator @xmath267 on the eigen - functions of one angular momentum @xmath2 .
assuming that the states with different @xmath2 are not mixed , one can find : @xmath294 where @xmath295 and @xmath296 is the kronecker delta symbol .
substituting eq .
( [ rij ] ) into @xmath297 ) = ( \mbox{\boldmath $ h \hat \sigma$ } ) - ( \mbox{\boldmath $ h \tau$})(\mbox{\boldmath $ \tau \hat \sigma$})$ ] , we rewrite the magnetic field hamiltonian @xmath298 of eq .
( [ zeeman ] ) as : @xmath299 here the spin electron @xmath1 factor is given by @xmath300 the matrix elements of the last term in eq .
( [ zeemanpb ] ) are zero for @xmath202 . if @xmath301 one can neglect the off diagonal elements of the operator @xmath302 and replace them with @xmath303 .
this is the case for the first excited @xmath203 state in cdse nanocrystals ( for comparison , see the corresponding curves in figs .
[ sizesurf ] and [ sizegl ] ) . in this case
the last term of eq .
( [ zeemanpb ] ) describes the surface contribution to the spin splitting of the electron levels that depends on the projection of the orbital momentum on the magnetic field.@xcite in the _ strong field regime _
( similar to the case of the `` quasi
normal '' zeeman effect or `` complete '' paschen - back effect for the electrons in atoms@xcite ) one can additionally neglect the off diagonal elements of the spin
orbit operator @xmath304 .
the fine structure of the electron level with orbital momentum @xmath2 is described by : @xmath305 where @xmath306 and @xmath22 are the projection of the electron angular momentum and spin on the magnetic field direction and @xmath307 thus the structures of the excited qsls with @xmath176 may be different in weak and strong magnetic fields and undertake the anticrossing in the intermediate magnetic field ( similar to those known for atoms@xcite ) . in nanocrystals
the weak , intermediate and strong field regimes depend strongly on the nanocrystal size as well as on its surface conditions .
we have studied the effect of general boundary conditions ( the surface effect ) on the electron qsls and their zeeman splitting in spherical bare semiconductor nanocrystals . the above consideration that has been carried on in the eight band effective mass model can be easily extended to describe spherically layered heterostructures and wide gap semiconductor nanocrystals , which are better described by the fourteen band model ( see for example ref . ) . comparing the results of our theoretical calculation with the experimental size dependence of the electron @xmath1-factor we have determined the surface parameter @xmath308 @xmath218 in bare cdse nanocrystals .
fitting the experimental data , we used @xmath210 ev , that has been obtained independently from bulk measurements@xcite and that describes better the experimental data than @xmath309 ev , previously used for cdse in refs . .
the surface parameter @xmath186 characterizes the electronic properties of the surface and should be considered as the additional one to the set of parameters that describes the bulk properties of semiconductors . in cdse nanocrystals prepared by a different technique
@xmath186 can be different .
the extracted absolute value and the negative sign of @xmath277 @xmath218 is consistent with our theoretical expectations for @xmath186 in studied cdse nanocrystals .
its value is very close to the theoretical value of the surface parameter for semiconductors with a symmetrical band structure@xcite @xmath310 @xmath218 .
it can also be shown within the eight band effective mass model , that the negative sign of the surface parameter does not allow the existence of the surface localized states with @xmath311 ( gap states ) ( one can find similar consideration in ref . ) . indeed , the cdse samples studied in refs .
show very high pl quantum efficiency and do not show deep gap transitions .
the negative parameter @xmath186 leads also to an additional nonparabolicity , bowing the size dependence of the electron energy levels ( see fig .
[ levels ] ) and therefore may describe the unexplained experimental size dependence of the 1s electron level in small cdse nanocrystals.@xcite using the gbcs we have found also the direct surface contribution to the spin orbit effects in zero and weak external magnetic fields . the surface contribution to the zero field spin
orbit splitting is similar to the interface contribution obtained in refs . for @xmath312 electrons in planar quantum wells by using the spin dependent boundary conditions .
it has been pointed out in ref . that the interface contribution to the rashba spin orbit term in the @xmath312 hamiltonian@xcite is related to discontinuity of the band parameters at the semiconductor heterointerface and that this contribution is an additional one to those connected with the space charge and/or the external electric field .
the importance of the effects described by the rashba term for the 2d electrons confined near the curved surface@xcite and cylindrical semiconductor quantum dots@xcite has been emphasized recently .
the rashba spin orbit term in our spherical dots is a direct consequence of the gbcs for the envelope function .
the same consideration can be made for cylindrical dots or any other nanostructure geometry ( the results will be published elsewhere ) .
the gbcs provide an important connection between the constant that describes the magnitude of the spin orbit term and surface conditions in nanostructures .
_ in conclusion _
, we have shown an important influence of the semiconductor surface on the electron energy structure in bare spherical nanocrystals .
the effect of the surface has been modeled through the choice of the boundary condition parameter that describes the nonzero value of the envelope function at the nanocrystal surface .
the additional nonparabolicity of the quantum size energy levels , the spin orbit splitting of the electron quantum size levels , and the additional magnetic moment of the electrons have been shown to be induced by the surface .
the effects are significant in small nanocrystals and their considerations require a multiband effective mass approach because interband coupling is important there .
the analysis of the experimental data allows us to determine the appropriate parameter of the boundary conditions that characterize the surface in studied bare cdse semiconductor nanocrystals .
the authors thank i. a. merkulov and b. k. meyer for helpful discussions and j. tischler for a critical reading of the manuscript .
the work of a. v. rodina was supported in part by the alexander von humboldt foundation , deutsche forschungsgemeinschaft ( dfg ) and the swiss national science foundation .
al . l. efros thanks the u.s .
office of naval research ( onr ) , u.s .
department of energy ( doe ) , and the darpa / quist program for financial support .
alekseev acknowledges support from the swiss national science foundation and of the grant of intas 99 - 1705 .
we are interested in deriving a variation of the electron level energy @xmath30 caused by a small variation of the parameter @xmath313 that characterizes the surface bc for the bulk wave function @xmath314 .
the gbcs at the spherical surface of the nanocrystal with radius @xmath315 can be written as @xmath316 where @xmath317 is the real number constant .
the function @xmath314 is the solution of bulk schrdinger equation @xmath318 where the constant @xmath319 is independent of @xmath313 .
taking a derivative of the schrdinger equation on @xmath320 , multiplying both parts of the resulting equation by @xmath321 and integrating it over the sphere volume , one can obtain : @xmath322 taking a derivative of eq .
( [ bcsig ] ) and substituting @xmath323 into eq .
( [ fluxsig ] ) one arrives at the final expression for the energy variation : @xmath324 substituting @xmath325 and @xmath326 $ ] with @xmath327 into this equation we obtain eq .
( [ sosur ] ) for the energy correction @xmath328 to the averaged energy of the electron level with the orbital momentum @xmath2 caused by the small perturbation of the boundary conditions of eq .
( [ bcfso ] ) .
a. hasselbarth , a. eychmller , r. eichberger , m. giersig , a. mews , h. weller , j. phys . chem . * * 9**7 , 5333 , 1993 ; h.weller , adv.matter . * 5 * , 88 ( 1993 ) ; d. schooss , a. mews , a. eychmller , h. weller , phys
b * 49 , 17072 ( 1994 ) ; a. mews , a. v. kadavanich , u. banin , and a. p. alivisatos , phys .
b * * 5**3 , r13242 ( 1996 ) . | the structure of the electron quantum size levels in spherical nanocrystals is studied in the framework of an eight band effective mass model at zero and weak magnetic fields .
the effect of the nanocrystal surface is modeled through the boundary condition imposed on the envelope wave function at the surface .
we show that the spin
orbit splitting of the valence band leads to the surface induced spin
orbit splitting of the excited conduction band states and to the additional surface induced magnetic moment for electrons in bare nanocrystals .
this additional magnetic moment manifests itself in a nonzero surface contribution to the linear zeeman splitting of all quantum size energy levels including the ground @xmath0 electron state .
the fitting of the size dependence of the ground state electron @xmath1 factor in cdse nanocrystals has allowed us to determine the appropriate surface parameter of the boundary conditions .
the structure of the excited electron states is considered in the limits of weak and strong magnetic fields . |
the hugoniot curves ( see [ 6 ] ) correspond to the linkage between the velocity @xmath2 and the pressure @xmath3 in a shock wave travelling through a material after an impact . by starting from a position at rest , the algebraic equation of such curves
has the form @xmath4 where @xmath5 is the mass density of the material , @xmath6 is its speed of sound , and @xmath7 is a dimensionless constant retated to this material , obtained from experiments . for water , @xmath8 and @xmath9 by solving with respect to @xmath2 , we get @xmath10 where @xmath11 and @xmath12 whose unit is the inverse of an energy .
we notice the relation @xmath13 the values of these parameters come from [ 5 ] .
1.8ema general one dimension velocity - pressure model , with a constant mass density @xmath5 , is made of a dynamics equation of the form @xmath14 where @xmath15 is the time and @xmath16 is the position along a vertical upwards oriented axis ( with @xmath17 at sea level ) , and a hooke law of the form @xmath18 where @xmath19stands for the pressure depending wave velocity , to be identified .
these two equations compose a hyperbolic system whose riemann invariants have the form@xmath20 by derivating the expression of @xmath21 from the hugoniot curves , one gets @xmath22 identifying the two expressions and using @xmath23 lead to the formula @xmath24 this reads like a state law for our velocity pressure model .
1.8emup to now we were only concerned with homogeneous equations since riemann invariants only exists in this case .
we also have to take in account the gravity effects , since the pressure is increasing with the depth of water , and some friction effect since stillness is a stable configuration .
we use a friction term of the strickler type as usual in hydraulics , but other choices are possible and will lead to similar results .
that way , the dynamics equation is replaced by @xmath25 where @xmath26 is the gravity constant and @xmath27 the friction parameter .
the hooke law and the state law are unchanged , and we take @xmath28 in the numerical experiments . the size of the friction parameter is a priori unknown and will be discussed later .
we look for a quantity @xmath29 satisfying the transport equation @xmath30 since @xmath31 depends on @xmath3 we get @xmath32 to be compared with the hooke law .
we get @xmath33 which is easily solved and gives @xmath34 for the choice @xmath35 the wave velocity can be written as a function of the strain density @xmath31 as @xmath36 we check now the conservation of the momentum @xmath37 we get @xmath38 where @xmath39 we introduce a pressure term , standing as a strain pressure , @xmath40 and we get the conservative equation for the momentum @xmath41 the system made of ( [ tsunami9 ] ) ( [ tsunami8 ] ) has the form of the well known euler equations for gas dynamics with a larger adiabatic coefficient @xmath42 instead of the usual value @xmath1 for gases , and can be handled in the same way , especially for the shock waves .
the same rankine hugoniot condition is valid , connecting the velocity of a shock wave to the two states @xmath43 and @xmath44 by @xmath45
we first compute the state at rest . in case of stillness the equations ( [ tsunami9 ] ) and ( [ tsunami8 ] ) reduce to @xmath46 since @xmath47 by denoting @xmath48 the strain density at rest
, the integration gives @xmath49 recalling that @xmath50 we can use the atmospheric pressue @xmath51 at the surface ( @xmath17 ) and get @xmath52 which corresponds to the geostrophic equilibrium state .
1.8emto compute the strain density profile we use the deviation variable @xmath53 and look for linkage of the form @xmath54 as in any @xmath55system with a source term ( see [ 4 ] , or annex below in section 5 ) , where @xmath56 and @xmath57 are constant .
since stillness is reached for @xmath58 we have @xmath59 . besides , since @xmath60 does not depend on @xmath15 , we have @xmath61 which means that @xmath56 corresponds to the wave velocity , that we name the reference velocity , corresponding to a reference state @xmath62 such that @xmath63 .
we compute , for @xmath64 which is always expected , @xmath65 to be introduce into ( [ tsunami8 ] ) which becomes @xmath66 \eta _ { z}\ + gq-\frac{gq_{0}}{c(q_{0})^{2}}\left({c^{2}-a^{2}(\frac{q - q_{0}}{q})^{2}}\right ) \ + \ \frac{kq}{\rho _ { 0}}\frac{a^{2}}{q^{2}}\left({q - q_{0}}\right ) ^{2}=0\ .\ ] ] this equation reduces to @xmath67 \ \eta _ { z}\ + \
gq\ + \ \frac{gq_{0}}{c(q_{0})^{2}}\left[{a^{2}\left({\frac{q - q_{0}}{q}}\right ) ^{2}-c^{2}}\right ] \ + \ \frac{k\
a^{2}}{\rho _ { 0}\ q}\left({q - q_{0}}\right ) ^{2}\ = \ 0\ .\ ] ] by multiplying by @xmath68 and using @xmath69 we get @xmath70 since @xmath71 this is a differential equation which can be integrated by using standard numerical methods . an increasing profile is expected .
since the reference velocity @xmath72is far larger than @xmath73 the coefficient @xmath74 of @xmath75 is always negative in practice .
the two last terms are always positive , and the friction term is the predominant one .
the term @xmath76 is always negative and is always balanced by the friction term when the friction coefficent @xmath27 is not too small .
1.8emthe value of @xmath56 is determined by the strength of the sismic wave at the bottom of the ocean , whose depth is denoted @xmath77 this corresponds to a reference state @xmath78 we have @xmath79 we get @xmath80 whichs gives @xmath81 1.8emnow we can compute the profile of the stain wave as the solution of ( [ tsunami5 ] ) .
figure 1 presents a series of numerical computation tests using the reference value @xmath82 , for a depth @xmath83 meters , that is an increasing of about 1500% above the natural pressure on the bottom of the ocean .
2.1emthe velocity field @xmath84 increases from a few meters per second near the bottom to more than 300 meters per second near the front shock wave drawn here , which is here an hypothetic one .
the tests performed with too small friction coefficients ( observed here for @xmath85 ) yield decreasing profiles , as expected from the remark above about the size of the friction term .
the real front shock wave will progress more slowly than the strain wave , and its amplitude will decrease rapidly , and the effective values of the velocity field @xmath84 will be strongly reduced near the front schock wave . 1.8emthe front shock wave connects the geostrophic equilibrium state @xmath60 , with the velocity @xmath86 , to a value @xmath31 on the strain wave , with the velocity @xmath87 the rankine hugoniot condition ( [ tsunami4 ] ) gives the velocity of this shock wave , which reads here : @xmath88 we have the following result : the shock wave propagates slower than the stain wave , and faster than the local wave speed @xmath89 , that is @xmath90 * proof : * we fix @xmath15 and set @xmath91 and @xmath92 , as the left and right values of the shock wave
. then the shock velocity reads @xmath93 we shall use some @xmath94q_{0},q_{s}[$ ] such that @xmath95 we have @xmath96 where @xmath97 we get @xmath98 which reduces to @xmath99 and is equivalent to @xmath100 which is true since @xmath101 and @xmath102 1.8emon the other hand , since the expression @xmath103 is an increasing function of @xmath104 for @xmath105 , we get obviously the other inequality.(end of proof ) 1.8emnow we can construct the whole wave , made of a regular part corresponding to a part of the strain wave and a front shock whose position is determined by the rankine hugoniot condition ( [ tsunami6 ] ) intertpreted as a differential equation whose solution @xmath106 gives the position of the shock .
figure 2 shows the different positions of the shock wave for different values of the friction coefficient .
1.8emthe velocity @xmath56 is equal to @xmath107 that is @xmath108 .
the shock on the bottom of the ocean , at the depth @xmath109 corresponds to 15 times the value of the usual geostrophic pressure , thai is @xmath110 this corresponds to strain densities @xmath111 we notice that the variation from @xmath112 to @xmath113 corresponds to an increasing of 11.5 % which is an increasing of about 1400 % of the pressure .
the velocity of the wave is computed from the values @xmath114 we notice that the velocities @xmath115 and @xmath116are for more important than @xmath6 .
the real profiles are drawn in red , and are shaped as a part of the strain wave profile cut by the slower shock wave .
the eliminated parts are drawn in blue .
we observe an important difference between the velocites of the different fronts , for several values of the friction coefficient .
we denote by @xmath117 the mean depth of the ocean .
the wavelength @xmath118 of the saint - venant waves must satisfy a condition of the form @xmath119 where @xmath120 the number of sonic interactions ( back - and - forth ) between the bottom and the surface of the ocean .
this condition means that along a horizontal distance of a wavelength @xmath118 , there are at least @xmath120 such sonic interactions .
the use of the saint - venant model is as more appropriate as @xmath120 is great . as in [ 2 ] for rogue waves ,
we propose to require @xmath121 which implies for example a wavelength greater than @xmath122 for an ocean depth of @xmath123 .
a tsunami wave is expected when the condition ( [ tsunami7 ] ) is fullfilled . a linear model for the surface elevation propagation
was proposed in the historical paper [ 1 ] by k.kajiura .
1.8emthe numerical tests in section 3 show the effect of the friction , whose effect is to shape up the amplitude of the wave near the front shock . at the same time , this front shock magnitude is eroded progressively as the wave propagates upwards to the surface . since the amplitude of the front shock decreases , the velocity of this front shock decreases too .
by following the propagation of a wave with an incidence of angle @xmath124 ( that is only changing @xmath26 into @xmath125 and @xmath16 into @xmath126 ) , we get a longer path to travel with a weaker gravity constant ) and a delayed wave compare to the case withou incidence .
this is show on figure 3 .
1.8emthe value of the initial amplitude has a capital effect .
it must be large enough to get , after erosion by the friction , a remaining wave near the surface which is sufficient to raise up the sea surface and provoke a wave .
the physical wave starts as a sperical wave , and propagates according to the incidences .
the part with a small incidence will reach the surface later and will help the formation of the water wave .
the part with a larger incidence will disappear because of the friction effect .
the question of the value of the friction coefficient stays open , since the use of the strain density was never done before .
it seems from the numerical tests that the correct values lay between 1 and 10 .
too large values provoke a sharp front wave which erodes rapidly and will never reach the surface with a sufficient amplitude to make a wave , which is not expected , since sometimes , tsunamis really occur .
we consider a general 2x2 hyperbolic system whose first equation has the form @xmath127 we denote by @xmath128 and @xmath129 the eigenvalues of the flux matrix , which depend on @xmath31 and @xmath130 only .
then the general form of the second equation is @xmath131 or @xmath132 by using the notations @xmath133 and @xmath134 is a source term , assumed to be not identically zero .
we have the following result : the nonlinear non homogeneous system ( [ tsunami21 ] ) , ( [ tsunami22 ] ) admits non constant local solutions which are also solutions to the linear homogeneous system @xmath135 where @xmath56 is a real constant , with the linkage @xmath136 another real constant .
* proof : * we look for local solution with a linkage of the form @xmath137 . then the system becomes @xmath138 since @xmath139and @xmath140 , the second equation becomes @xmath141 which has the form @xmath142 by introducing a real function @xmath143 whose derivative is@xmath144 now , integrating with respect to @xmath145gives @xmath146 where @xmath147is an integrating constant which may depend on @xmath15 . next , derivating with respect to @xmath15 gives @xmath148 where we set @xmath149hence we get @xmath150 and recalling that @xmath151it remains @xmath152 a new derivation with respect to @xmath153 leads to @xmath154 and since the solution is not constant , we get @xmath155that is @xmath156with some constants @xmath56 and @xmath157 next , @xmath158 and we have got @xmath159 for some constant @xmath160 or , locally , @xmath161 , which satisfies to ( [ tsunami23 ] ) , since @xmath162.(end of proof ) 1.5emthis result is a very general one , since no special hypotheses were needed on the second equation ( [ tsunami22 ] ) .
such waves are very common in the nature : water waves such as roll waves , rogue waves , tidal bore waves or also many other waves as reported in [ 2 ] or [ 4 ] .
for example the double property of being either a solution to a non linear , non homogeneous systems and a linear homogeneous system provides the linkage between acoustics and gas dynamics in a wind instrument ( see [ 3 ] ) .
1.8emin the case of a conservative system , invariant by galilean transform , the only choice of the function @xmath163 is reduced to @xmath164 we easily construct this way the usual saint venant system in hydraulics or the euler equations in gas dynamics .
1.8emin section 3 , the state at rest is not @xmath165 , but @xmath166 so we look for a solution with the linkage @xmath167 since @xmath168 at rest .
we find a differentiel equation which is more complex than ( [ tsunami24 ] ) .
we have used the idea of source term linearization effects on waves as in [ 2],[3 ] and [ 4 ] .
other models are developped in [ 1 ] , [ 5 ] and [ 7 ] .
the data [ 6 ] were used to valuate the parameters . a .- y.leroux , m .- n.leroux , a mathematical model for rogue waves , using saint venant equations with friction .
conservation laws pr .
server : www.math.ntnu.no@xmath169conservation@[email protected] | by using the hugoniot curve in detonics as a riemann invariant of a velocity - pressure model , we get a conservative hyperbolic system similar to the euler equations .
the only differences are the larger value of the adiabatic constant ( @xmath0 instead of @xmath1 for gas dynamics ) and the mass density replaced by a strain density depending on the pressure .
the model is not homogeneous since it involves a gravity and a friction term .
after the seismic wave reaches up the bottom of the ocean , one gets a pressure wave propagating toward the surface , which is made of a frontal shock wave followed by a regular decreasing profile . since this regular profile propagates faster than the frontal shock waves , the amplitude of the pressure wave is strongly reduced when reaching the surface . only in the case of a strong earth tremor
the residual pressure wave is still sufficient to generate a water elevation with a sufficient wavelengths enable to propagate as a saintvenant water wave and to become a tsunami when reaching the shore .
we describe the construction of the model and the computation of the wave profile and discuss about the formation or not of a wave .
1.5em we propose a model using a constant mass density because variable mass density models are often unstable , since a tiny variation of the density allways causes a large variation of the pressure .
in the model presented here , the transported variable is not the mass density but a new variable , called the strain density , which has the same properties of conservation without this drawback of numerical unstability .
1.8emthe first section deals with the construction of the velocity - pressure model and the new conservative variable of strain density is designed in the second section .
the new non homogeneous euler - like model is then studied in the third section and one dimension numerical computations of the profile of the wave are reported in the fourth section , for different values of the friction coefficient .
we conclude by some discussions about the emergence of a tsunami wave or not . |
the dynamics of a particle that interacts with its environment constitutes a basic problem in physics .
classically , upon elimination of the environmental degrees of freedom , the reduced dynamics is most simply described in terms of langevin equation .
solution of this equation , by utilizing fokker planck equation , is well known @xcite . in the absence of external potential
, it yields spreading and diffusion .
the latter effect is due to the interplay of noise and dissipation .
however , diffusion may arise also from the interaction with disordered environment .
this kind of non - dissipative `` random - walk '' diffusion is encountered , for example , in solid state physics , while analyzing electrical conductivity .
it is well known @xcite that this latter type of diffusion may be suppressed quantum mechanically due to localization effect .
still , diffusive - like behavior is recovered if noise and dissipation are taken into account .
the unified modeling of the environment in terms of noise , dissipation and disorder is the first stage of the present study .
one may take the notion of particle literally , and identify the environment as either external or internal bath that consists of infinitely many degrees of freedom .
the bath may be a large collection of other particles or field modes ( photons , phonons ) .
else , the internal degrees of freedom of the particle itself are considered to be the bath .
the latter point of view has been suggested by gross @xcite in order to analyze inelastic scattering of heavy ions . a totally different point of view , promoted by caldeira and leggett ( cl ) @xcite ,
consider the notion of particle as a token for some macroscopic degree of freedom .
a linear interaction with a speculated bath that consists of infinitely many uncoupled harmonic oscillators , is assumed ( zwanzig @xcite ) .
the known classical limit , namely langevin - type equation , serves as a guide for the construction of the proper hamiltonian .
( phenomenological rather than microscopic considerations are used , hence the usage of the term ` speculated bath ' ) .
the power inherent in this approach is the capability to introduce an explicit path integral expression for the propagator , using the feynman vernon ( fv ) @xcite formalism .
this fv - cl propagator constitutes a quantized description of brownian motion ( bm ) .
the term `` bm - model '' will be associated from now on with this propagator .
the first question which should be asked concerning the applicability of cl approach is obviously whether either the coupling , the bath , or both are `` too simple '' in order to account for the variety of physical phenomena that are associated with generalized bm .
the term `` generalized bm '' is used in order to describe any dynamical behavior that corresponds in the classical limit to langevin - like equation . in the simplest case langevin equation is @xmath0 where @xmath1 and @xmath2 are the mass and the friction coefficient , respectively .
one should specify the stochastic force @xmath3 .
this force may arise from interaction with some fluctuating homogeneous field ( @xmath4 ) , which is the usual formulation .
however , _ more generally _ , this force may arise from the interaction with _ disordered _ potential ( @xmath5 ) .
in the latter case the spatial auto - correlations of the force are significant .
to avoid misunderstanding it should be emphasized that there are other aspects in which bm can be generalized ( for review see @xcite ) . in the present paper
we construct a unified model for the study of diffusion , localization and dissipation ( dld ) .
this model describes generalized bm in the sense specified above .
the disordered environment may be either static ( quenched ) , noisy or dynamical .
the model is treated within the framework of the fv formalism .
the resultant path integral expression for the propagator contains a functional @xmath6 with kernel @xmath7 that corresponds to friction , and a functional @xmath8 with kernel @xmath9 that corresponds to the noise .
both functionals depend also on a suitably defined auto - correlation function @xmath10 that characterizes the disorder .
bm model constitutes ( formally ) a special case where the disorder auto - correlation length is taken to be infinite .
optional derivations of the resultant path integral expression are presented for the particular cases of either classical or quantal systems with non - dynamical disordered potential .
obviously , in the latter case the friction functional @xmath6 is not generated by the derivation .
localization problem is obtained if the noise auto - correlation time is taken to be infinite ( @xmath11 ) .
an explicit computation of both the classical and the quantal propagators @xmath12 will be carried out .
this propagator generates the time evolution of @xmath13 , which is either the wigner function or the corresponding classical phase - space distribution . spreading and diffusion profiles are found for either noisy or ohmic environment .
quantal corrections to the classical result are discussed . a new strategy is developed in order to handle the integration over paths .
this strategy is utilized in order to study the anomalous `` diffusion '' profiles due to colored noise .
later it is extended in order to demonstrate how localization comes out from the path integral expression .
again , one may ask , whether the dld model is the `` ultimate '' model for the description of bm in the most generalized way ( as far as generic effects are concerned ) . in case of 2-d generalized bm
one should consider also the effect of `` geometric magnetism '' @xcite , which is not covered by the 1-d dld model . here
we limit the discussion to 1-d bm . in order to answer this question one should consider a general nonlinear coupling to a thermal , possibly chaotic bath . in the limit of weak coupling one
may demonstrate ( see sec.vi(b ) ) that indeed the bath can be replaced by an equivalent `` effective bath '' that consists of harmonic oscillators , yielding the dld model .
a further reduction to bm model is achieved if the coupling is linear .
the derivation also demonstrates why the so called `` ohmic '' bath is generic .
however , we can not prove that the path - integral expression that corresponds to the dld model is the most generalized description of bm ( in the sense of this paper ) .
gefen and thouless @xcite , wilkinson @xcite and shimshoni and gefen @xcite have emphasized the significance of landau - zener transitions as a mechanism for dissipation .
the weak coupling approximation misses this effect . still , there is a possibility that some future , more sophisticated derivation , will demonstrate that an equivalent `` oscillators bath '' can be defined also in the case of strong coupling .
the existence of such derivation is most significant , since it implies that no `` new effects '' ( such as `` geometric magnetism '' in case of 2-d generalized bm ) can be found in the context of 1-d generalized bm .
referring again to the landau - zener mechanism , wilkinson has demonstrated that anomalous friction , which is not proportional to velocity , may arise @xcite .
the bm model can not generate such anomalous effect .
however , we shall demonstrate that the _ non - ohmic _ dld model can be used in order to generate this effect .
we turn to review previous works that are related to the present study .
traditional approach to the study of dissipation is based on the master equation formalism @xcite , which is the quantal analog of the classical fokker planck treatment @xcite .
systematic derivations are usually based , in some stage , on the markovian approximation .
an alternative route is to apply the fv formalism @xcite .
this has been done by mhring and smilansky @xcite , following gross @xcite , in order to study deep inelastic collisions of heavy ions .
later , the fv formalism has been applied by cl @xcite and followers @xcite in particular to the study of macroscopic quantum tunneling .
hakim and ambegaokar @xcite has applied fv model in order to compute the spreading and the diffusion of quantum brownian particle .
cohen and fishman @xcite have computed the full wigner propagator in case of general quadratic hamiltonian , possibly time dependent .
the latter study has demonstrated more clearly the significance of _ noise time - autocorrelations_. non - trivial auto - correlations may lead to non - markovian effects due to the non - local ( in time ) nature of the noise functional @xmath8 .
the simplest example is the suppression of diffusion at low temperatures @xcite due to negative power - law correlations of the noise @xcite .
a less trivial manifestation of non - markovian effect has been found by cohen @xcite while analyzing diffusion due to the destruction of localization by colored noise .
non markovian features are usually speculated to be less relevant if disorder is taken into account .
one may expect that for generalized bm , disorder will lead , at all temperatures , to normal diffusion .
we shall find later in this paper that non - markovian effects are quite effective also in case of the dld model , and lead to anomalous spreading profiles .
however , it will be clear that these effects , though counter intuitive at first sight , are of classical nature . non markovian effects can also arise from the non - locality of the friction functional @xmath6 , this is the case for non - ohmic bath .
the retarded response of the bath has then long memory for the particle s dynamics , see for example the review papers by grabert et al . and by hnggi et al .
in particular , for unbounded motion , in the absence of disorder ( bm model ) , it will be demonstrated that one encounters infinities in computations of the friction and of the effective mass .
these infinities are avoided if disorder is taken into account ( sec.iv(b ) ) .
the suppression of quantal interference due to dephasing process is an important issue for both semiclassical @xcite and mesoscopic @xcite physics .
the dephasing of interference in metals due to electromagnetic fluctuations has been discussed by altshuler , aronov and khmelnitskii @xcite .
general considerations has been presented by stern , aharonov and imry @xcite .
the strong dimensionality dependence of the dephasing process has been emphasized . in this paper , the dld model
is used in order to study the suppression of quantal interference due to the local interaction with either noisy or dynamical _ disordered _ environment .
the dephasing is determined by the noise functional @xmath14 $ ] . due to the local nature of the dephasing process
, the decay of coherence is exponential in time , and no longer depends on geometrical considerations . in case of quenched disordered environment
the dld model reduces to a localization problem that its solution is well known @xcite . in particular , for @xmath15 correlated potential the result for the localization length has been pointed out by thouless @xcite . as far as we know ,
the real - time feynman path integral formalism has not been utilized so far in order to re - derive this result , though functional integration is frequently employed in closely related computations .
the case of noisy disordered environment , where the potential is @xmath15 correlated both in time and in space has been considered by jayannavar and kumar @xcite . in the latter reference
the spatial spreading has been computed , and a classical - like result has been obtained .
quantal corrections to the dispersion profile , has not been discussed in the latter reference .
no solution exists for a model that `` interpolates '' the crossover from quenched noise , via colored noise , to white noise disordered potential .
marianer and deutsch @xcite have considered the problem of white noisy disordered potential with added dissipation .
using bm model , they have demonstrated that a classical - like results is obtained for the spatial spreading .
again , neither the spreading profile nor quantal corrections have been considered .
the outline of this paper is as follows : in sec.ii the unified model for the study of diffusion , localization and dissipation ( dld ) is constructed .
derivation of the reduced classical dynamics is presented in sec.iii-iv .
it is shown that a well defined langevin equation is obtained for both subohmic and superohmic bath , as well as for ohmic bath .
this is contrasted with bm model , where only the ohmic case is well defined . in sec.v - vi ,
four derivations of the fv path integral expression for the propagator are presented : ( a ) a classical derivation that is based on langevin equation ; ( b ) a quantum mechanical derivation for non - dynamical noisy or quenched disordered potential ; ( c ) a quantum mechanical derivation for the full dld model ; ( d ) a quantal derivation for the general case of weak nonlinear coupling to a thermal , possibly chaotic bath .
app.a clarifies the relation of bm - model and dld - model to the case of interaction with external bath that consists of extended field modes . in sec.vi(c ) the explicit expression for the influence functional allows concrete predictions concerning the loss of interference due to dephasing .
app.b introduce a gedanken experiment that clarifies the manifestation of interference and dephasing in the presence of dynamical disordered environment . in sec.vii-viii three strategies for the computation of the quantal propagator
are discussed . spreading and diffusion profiles are found for either noisy or ohmic environment .
the dld model is compared with the bm model . in the latter case
the result is classical - like , while in case of the dld model , a singular `` quantal correction '' should be included .
the significance of this `` correction '' is further clarified in app.b .
the strategy of sec.viii , which enables computation of diffusion profiles in the presence of _ colored _ noise , is extended in sec.ix .
we use this strategy in order to demonstrate how , in the case of _ quenched _ noise , localization comes out from the general fv path integral expression .
summary and conclusions are presented in sec.x .
as a starting point for later generalization we consider the classical langevin equation .
this equation describes the time evolution of brownian particle under the the influence of so called `` ohmic '' friction and stochastic force . @xmath16 } \ \ \ \ .\end{aligned}\ ] ] in the _ standard _ langevin equation the stochastic force represents stationary `` noise '' which is zero upon ensemble average , and whose autocorrelation function is @xmath17}. \end{aligned}\ ] ] usually white noise , which is @xmath15 correlated in time , is considered .
the standard langevin equation can be _
generalized _ by assuming that the stochastic force is due to some noisy potential , namely @xmath18 where @xmath19 denotes spatial derivative .
one may assume that @xmath20 is zero upon ensemble averaging , and satisfies @xmath21}\end{aligned}\ ] ] the autocorrelation function of the stochastic force @xmath3 at some specified point @xmath22 is @xmath23 in practice @xmath24 will be assumed to have the factorized form @xmath25 where , without loss of generality , @xmath26 is normalized so that @xmath27 , the normalization constant being absorbed into @xmath28 . for distribution of particles , the _ standard _
langevin equation predicts rigid motion with no diffusion . this feature is eliminated if an average over realizations of @xmath3 is performed .
the _ standard _ langevin equation can be viewed as a special case of the _ generalized _ version ( [ e2_3 ] ) . for
this one should take @xmath29 .
alternatively , @xmath30}.\end{aligned}\ ] ] another choice for spatial autocorrelations that corresponds to disordered environment is @xmath31 } .\end{aligned}\ ] ] the parameter @xmath32 constitutes a measure for the microscopic scale of the disorder . for distribution of particles ,
the resultant motion will be diffusive - like rather than rigid , even without averaging over realizations .
it is crucial to do this important observation if one wishes to introduce a _ quantized _ version of langevin equation . disregarding the friction term ,
the langevin equation may be derived from the hamiltonian @xmath33 where average over realizations of @xmath20 is implicit .
if @xmath20 is time independent , we shall use the notion _ quenched _ disordered environment .
if @xmath20 is uncorrelated also in time , we shall use the notion _ noisy _ environment .
if @xmath2 is non - zero , energy may dissipate , and we should consider `` dynamical '' environment rather than `` noisy '' or `` quenched '' one . in order to model the environment we generalize the fv - cl approach .
namely , to facilitate later mathematical treatment , the interaction potential is assumed to be linear in the bath coordinates : @xmath34 where @xmath35 denotes the dynamical coordinate of the @xmath36 scatterer or bath mode .
@xmath37 is the normalized interaction potential , and @xmath38 are coupling constants . in case of dynamical environment ,
the @xmath39 are assumed to be oscillators coordinates .
the bath hamiltonian is : @xmath40 the total hamiltonian of the system plus the heat bath is @xmath41 in order to further specify the dld model , we should characterize the spectral distribution of the bath - oscillators , as well as their interaction with the particle @xcite . in what follows
we consider the case of localized scatterers .
the scatterers are assumed to be uniformly distributed in space with @xmath42 .
the joint distribution of the bath oscillators with respect to @xmath43 and @xmath44 is assumed to be factorized .
consequently we can write @xmath45 the bath is characterized by the spectral function @xmath46 .
if we consider a partition of the bath oscillators into subsets of oscillators whose positions @xmath44 are the same , then locally , within each subset , the @xmath43 will have the same distribution .
the interaction is characterized by well defined spatial autocorrelation function @xmath47 , namely @xmath48 where `` density '' refers to the uniform spatial distribution @xmath49 .
the scattering potential @xmath50 will be normalized so that @xmath27 , the normalization constant being absorbed in the coefficients @xmath38 . disregarding for a moment the dynamical nature of the bath , thus considering again the case of either `` noisy '' or `` quenched '' environment , one obtains @xmath51 which upon recalling previous definitions , leads to ( [ e2_4 ] ) with @xmath52 . thus , if the dynamical nature of the bath is ignored , the problem reduces to solving langevin equation with the appropriate @xmath47 and @xmath53 .
the spectral function @xmath46 and the spatial autocorrelation function @xmath26 constitute a complete specification of the model hamiltonian ( [ e3_4 ] ) .
this hamiltonian will be used in order to study diffusion localization and dissipation ( dld ) within the framework of a unified formalism .
the bm model , where @xmath54 up to normalization , constitutes _ formally _ a special case of the dld model .
it will be possible to apply the `` unified '' formalism that will be developed for the dld model , also for the analysis of the bm model , merely by substitution of the appropriate @xmath26 .
namely , one may use @xmath26 of ( [ e2_5 ] ) , or alternatively one may take the limit @xmath55 in ( [ e2_6 ] ) . however , there are some minor , but significant differences that will be noted in due time .
this is because our derivations rely on the _ local _ nature of @xmath26 .
therefore we consider the notion `` bm model '' distinctively from the notion `` dld model '' .
also the results will be both qualitatively and quantitatively different , in - spite of being derived , formally , from look - alike formulas .
fig.[models ] illustrates the bm hamiltonian and the dld hamiltonian .
looking at the figure it is apparent that the natural application of the bm model is for the description of the dynamics of a composite particle , with many internal degrees of freedom .
this interpretation is due to gross ref.@xcite , who suggested to apply such models for the analysis of heavy - ion collisions .
the bath according to this interpretation is `` internal '' , carried by the particle , rather than `` external '' .
the interaction hamiltonian ( [ e3_2 ] ) may also describe the interaction with some `` external '' bath that consists of extended field modes .
for example , the electron - photon interaction and the electron - phonon interaction may be cast into the form of ( [ e3_2 ] ) , with @xmath56 , where @xmath57 is either the speed of light or the speed of sound . in the long wavelength limit
this interaction resembles the bm - model rather than the dld - model . with minor modifications ,
also these models may be treated within the framework of our `` unified '' treatment ( appendix a ) .
finally , we re - emphasize that in general , the dld model as well as the bm model may be used on a phenomenological basis for the description of dissipation in mesoscopic quantum devices .
this point of view will be discussed further later on ( sec.vi(b ) in particular ) .
the quantal state of the particle may be represented by wigner function @xmath58 .
the time evolution of wigner function corresponds to that of classical distribution in phase - space . in this section
we consider a ` classical treatment of the dynamics ' .
the latter term implies that the system is considered to be classical , ( @xmath59 ) , while the bath gets full quantum mechanical treatment .
the limit @xmath60 is not taken .
the equations of motion of classical points that form a distribution in phase - space are and @xmath61 the variables @xmath62 satisfy the equation @xmath63 which can be solved explicitly , namely , @xmath64 substitution of the latter expression into ( [ e4_1 ] ) yields @xmath65 with @xmath66 the response kernel @xmath67 is defined for positive times ( @xmath68 ) as follows : @xmath69 in order to make further progress , a specification of the initial state of the system plus the bath is needed .
we shall assume that initially ( at time @xmath70 ) the system is prepared in some arbitrary quantal state while the bath oscillators are in thermal canonical equilibrium with some reciprocal temperature @xmath71 .
the wigner function representation of the probability density matrix is then ( see app .
a of ref.@xcite ) : @xmath72 where @xmath73 \ \ .
\end{aligned}\ ] ] using ( [ e4_6 ] ) one obtains the expectation values @xmath74 hence it is easily found that @xmath75 while @xmath76 thus , one identifies that the noise term is characterized by the autocorrelation function ( [ e2_4 ] ) with @xmath77 \end{aligned}\ ] ] the langevin equation ( [ e4_1 ] ) together with ( [ e4_2])-([e4_4 ] ) and ( [ e4_7 ] ) constitutes an exact and complete description of the reduced dynamical behavior of the system , as long as the system is considered to be classical in nature . for the bm model @xmath54 .
the derivation of langevin equation ( [ e2_1 ] ) leads to @xmath78 that satisfies ( [ e2_2 ] ) .
the latter may be interpreted as a special case of ( [ e2_3 ] ) , provided equation ( [ e2_5 ] ) for @xmath26 is used . however , in the expression for the retarded force ( [ e4_3 ] ) , @xmath79 is replaced by @xmath80 , rather than by @xmath81 . the physical significance of this minor difference is discussed below . in the equation for the retarded force ( [ e4_3 ] ) one may extend the integration over @xmath82 to infinity , provided @xmath67 is replaced by @xmath83 , where @xmath84 is the step function . in turn
, this kernel may be written as a sum of three terms , namely @xmath85 .
the kernel @xmath67 is the asymmetric continuation of @xmath86 to the domain @xmath87 , while @xmath88 is the symmetric continuation .
the latter is split into @xmath89 which is a delta function , and @xmath90 which satisfies @xmath91 .
consequently , the retarded force , for the bm model , may be written as the sum : @xmath92 where @xmath93 , and @xmath94 , and @xmath95 , and @xmath96 , with @xmath97 , and @xmath98 , and @xmath99 .
it has been assumed that @xmath67 has short range duration @xmath100 , much shorter than the physically relevant time scales of the dynamics .
_ this assumption is not true in general _ , and the consequences will be discussed later in the next section . the switching impulse act on the particle if it starts its trajectory at a point @xmath101 .
this term originates due to the fact that the initial preparation is such that the bath is in thermal equilibrium provided @xmath102 .
the @xmath103 force may be avoided if we care to include in the bm hamiltonian the proper `` counter term '' .
namely : @xmath104 this expression is manifestly invariant under space translations .
sanchez - canizares and sols have further considered this issue @xcite .
we turn back to the dld model with @xmath105 as found in equation ( [ e4_3 ] ) . here , performing similar treatment , neither @xmath106 nor @xmath107 are encountered .
formally , this is due to the fact that the difference @xmath81 appears in ( [ e4_3 ] ) , rather than @xmath80 by itself .
physically , this is due to the inherent homogeneity of the interaction with the environment ( [ e3_2 ] ) . as for @xmath108 and @xmath109 , here
one obtains @xmath110 if @xmath67 has short range duration @xmath100 , and if furthermore the velocity of the particle is not too high ( @xmath111 ) , then , using @xmath112 and , the above expressions reduce to those that follow equation ( [ e4_8 ] ) . if @xmath67 is not short range , to be discussed below , then the retarded force in the bm model is characterized by long range memory for the dynamics .
the approximation that has been used in the preceding subsection is no longer valid , and infinities are encountered in the computations of the constants there . in general , these annoying features are not shared by the dld model ( ( [ e4_11 ] ) with ( [ e2_6 ] ) ) , since a finite cutoff @xmath113 exists , where the microscopic length scale @xmath32 characterizes the interaction range with the scatterers .
in order to make further progress , a specification of the spectral function @xmath46 is required .
this function , defined in ( [ e3_5 ] ) , characterizes the distribution of the bath - oscillators with respect to their frequencies .
following cl we assume it to be of the form @xmath114 here the exponent @xmath115 characterize the singular behavior of @xmath46 in the vicinity of @xmath116 , while @xmath117 denotes a smooth cutoff function . the latter satisfies @xmath118 .
moreover , it is assumed that @xmath119 is analytic for real frequencies .
for example , @xmath117 may be chosen to be either lorentzian or gaussian .
the latter possibility will be adopted from now on .
the commonly used exponential cutoff @xmath120 will not be used since it does not satisfies the mentioned requirements .
exponential cutoff results in singular behavior that corresponds to the exponents @xmath121 rather than `` pure '' singular behavior that corresponds to @xmath115 alone .
the asymptotic behavior of both @xmath67 and @xmath28 is dictated by the singularity of @xmath46 at @xmath116 .
the physical significance of @xmath115 is further discussed in app.a and in sec.vi(b ) . in order to display explicit expressions for both @xmath67 and @xmath28 , it is useful to define the following function : @xmath122 for @xmath123 this function is a normalized gaussian . for odd
@xmath115 one obtains @xmath124 \ \ \ \ \ \mbox{[n is integer]}\end{aligned}\ ] ] for general @xmath115 it starts at @xmath125 positive @xmath126 and cross the @xmath127 value @xmath128 $ ] times ( now we consider @xmath129 odd number , possibly non - integer ) . the short range oscillatory behavior dies out on a time scale of the order @xmath130 . for larger times a power - law decay
is found : @xmath131}\end{aligned}\ ] ] the total algebraic `` area '' under @xmath132 is infinite for @xmath133 , finite for @xmath123 , and zero for @xmath134 .
representative plots of @xmath132 are displayed in fig.[g_s_plot ] .
for the spectral function @xmath46 as in ( [ e5_1 ] ) , with gaussian cutoff , the kernel @xmath67 is @xmath135 also @xmath28 can be expressed in terms of @xmath132 in both cases of `` high '' and `` low '' temperatures . in the first case
it is assumed that @xmath136 ( that has dimensions of time ) is much shorter than any dynamical time scale .
one may use then the approximation @xmath137}.\end{aligned}\ ] ] else , if @xmath136 is long , one obtains @xmath138},\end{aligned}\ ] ] where @xmath139 is riemann zeta function ( sum over @xmath140 ) . the notion `` low temperatures '' means here @xmath141 .
the temperature - dependent constant term results from the integration @xmath142 $ ] . from now on we shall use the notions of `` high '' and `` low '' temperatures in the sense of the above approximations .
we turn back to the computation of the terms in ( [ e4_8 ] ) . for the @xmath103 term
the result is always finite @xmath143 } \
\ .\end{aligned}\ ] ] for the @xmath144 one obtains a finite , non - zero result , ( @xmath145 ) only in the case of an ohmic bath ( @xmath123 ) . for subohmic bath ( @xmath133 )
the friction is _ infinite _ , while for superohmic bath ( @xmath134 ) the friction is _ zero_. in case of the dld model finite results are obtained for all cases . using ( [ e4_11 ] ) one finds out @xmath146 \
\nonumber \\ \ \ = \ -\frac{2}{\pi}\gamma(1{+}\frac{s}{2 } ) \ 2^{s/2 } \cdot \frac{\tau_{\ell}^3 } { ( \tau_c^2+\tau_{\ell}^2)^{1+\frac{s}{2 } } } \cdot \eta v \ \
\ \ -\frac{2}{\pi}\gamma(1{+}\frac{s}{2 } ) \
2^{s/2 } \cdot \ell^{1-s } \cdot \eta v^s \\end{aligned}\ ] ] above , the notation @xmath147 has been used . in the last step the limit
@xmath148 has been taken since it leads to a finite , non - zero result .
note that the cutoff frequency @xmath149 is not relevant physically as long as @xmath150 .
the above computation clarifies the significance of the various time scales .
formally , the bm model constitutes a special case with @xmath151 .
for @xmath123 , the so called `` ohmic '' case , one obtains @xmath152 .
this result holds in the case of the dld model as well as in the case of the bm model . in the latter case , for @xmath153 or for @xmath134 ,
the friction force is either infinite or zero respectively .
this is due to a long range memory effect .
the cutoff frequency @xmath149 , by itself , does not prevent this feature .
finite results for the components of retarded force in the bm model ( [ e4_8 ] ) may be obtained only for bounded systems , where @xmath154 explores only a finite portion of the space .
the calculation above ( [ e5_12 ] ) manifests the fact that a finite result , in case of unbounded system , may be obtained if a physical cutoff @xmath155 is introduced .
such arise naturally in case of the dld model .
similar picture emerges upon calculation of the effective mass . for the bm model , a finite result
is obtained in case of an ohmic bath ( @xmath123 ) .
@xmath156 \cdot \tau^2 \cdot d\tau \ = \ -\sqrt{\frac{2}{\pi}}\ \frac{\eta}{\omega_c } \ \ \ \ \ \ \ \ \ \ \ \ \ \mbox{for $ s=1$}\end{aligned}\ ] ] also for @xmath157 , the power law tail of @xmath67 decays sufficiently fast to guarantee a finite result : @xmath158 for @xmath159 ( @xmath160 ) , the effective mass is infinite .
in order to obtain a finite result we should turn to the dld model ( [ e4_11 ] ) . here , as in the case of friction calculation ,
the spectral cutoff @xmath149 is not important , and will be taken to infinity .
thus the integration may be performed using the asymptotic expression for @xmath67 .
substitution of ( [ e2_6 ] ) , ( [ e5_8 ] ) and ( [ e5_7 ] ) yields @xmath161 \ \tau^2 \ d\tau \nonumber \\ \ \ & = & -\frac{s(s-1)}{\pi}\cos(\frac{\pi}{2}s ) \gamma(s)\gamma(1{-}\frac{s}{2 } ) \ 2^{-s/2 } \cdot \eta\tau_{\ell}^{2-s } \
-\frac{\eta}{v^{s{-}2 } } \ \ \ \ \mbox{[for $ 0{<}s{<}2$]}\end{aligned}\ ] ] for the special case @xmath123 one obtains @xmath162 which is consistent with ( [ e5_13 ] ) in the limit @xmath148 . for @xmath163 a negative infinite value
is obtained , while from ( [ e5_14 ] ) a positive infinite value is obtained . in order to describe correctly the crossover at @xmath164
, one should introduce finite cutoff @xmath149 as well as finite @xmath155 .
the calculation will not be carried here .
in this section we shall develop a path - integral expression for the propagator of a particle that interacts either with static ( quenched ) or noisy environment .
first we develop a classical expression , and then we generalize to the quantal regime .
we refer to the dynamics generated by the hamiltonian ( [ e3_1 ] ) .
the classical liouville propagator , over infinitesimal time @xmath165 , and for definite realization of @xmath20 is : @xmath166 it is more convenient to write an expression for the propagator of the fourier transformed probability function ( ) , namely , @xmath167 \nonumber\end{aligned}\ ] ] here a dummy parameter @xmath168 has been inserted .
its value does not have any effect here .
however , later comparison to the quantum mechanical version will be more transparent . for finite time
, the convolved propagator may be written as a functional integral : @xmath169 \nonumber\end{aligned}\ ] ] here the measure is @xmath170 and the restrictions at the endpoints are @xmath171 , @xmath172 and @xmath173 , @xmath174 .
it is now possible to average over realizations of @xmath20 , using the well known gaussian identity @xmath175 one obtains @xmath176 } \ \ \mbox{e}^{-\frac{1}{\hbar^2}s_n[r , r]}\end{aligned}\ ] ] where @xmath177= s_{free}[r , r ] - \int_0^t d\tau \frac{\partial v}{\partial r } \cdot r(\tau ) \ \ \ \ \ \ \mbox{[classical]}\end{aligned}\ ] ] and @xmath178=\frac{1}{2}\int_0^t\int_0^t d\tau_1 d\tau_2 \ [ -\phi''(r(\tau_2){-}r(\tau_1),\tau_2{-}\tau_1 ) ] \ r(\tau_2)r(\tau_1)\end{aligned}\ ] ] the free part of the effective action , for use in ( [ e6_3 ] ) , is @xmath179=m \int_0^t d\tau\dot{r}\dot{r}$ ]
. it may be more convenient to write down the path - integral expression for @xmath12 .
this expression is obtained by double fourier transform of @xmath180 .
the result is @xmath181 } \ \ \mbox{e}^{-\frac{1}{\hbar^2}s_n[r , r]}\end{aligned}\ ] ] here , @xmath182=-m
\int_0^t d\tau\ddot{r}{r } \ \ \ \ \ \ \ \mbox{[for use in ( \ref{e6_7 } ) ] } \ \ \ .\end{aligned}\ ] ] note that the integration @xmath183 is not restricted at the end - points , whereas the integration @xmath184 is restricted at the end - points both in @xmath185 and in @xmath186 .
the restriction on @xmath186 at the endpoints is implicit , through the dispersion relation @xmath187 .
a similar expression may be obtained for the quantal propagator .
again , we refer to the dynamics generated by the hamiltonian ( [ e3_1 ] ) . the environment may be either `` noisy '' or `` quenched '' , where the latter case constitute formally a special case of the former .
the expression that will be obtained is a generalization of a result that has been obtained in ref.@xcite for white noise potential .
the feynman path - integral expression for the propagator of the quantal wave - function is , @xmath188 the path integral expression for the propagator of the density probability function constitutes summation @xmath189 over the pairs of paths @xmath190 and @xmath191 .
alternatively , we may use also the coordinates @xmath192 and @xmath193 , thus the summation will be @xmath194 , namely @xmath195 - i\frac{1}{\hbar}\int_0^t d\tau ( { \cal u}(x'',\tau)-{\cal u}(x',\tau ) ) } \nonumber\end{aligned}\ ] ] where @xmath196=s_{free}[r , r ] - \int_0^t d\tau ( v(x'',\tau){-}v(x',\tau ) ) \end{aligned}\ ] ] it is important to notice that the quantal definition of the measure is identical with the classical one ( [ e6_1 ] ) . in order to perform the average over realizations of @xmath20 using the gaussian identity ( [ e6_2 ] ) , one may write the last expression as @xmath197 -i\frac{1}{\hbar}\int_0^t d\tau \int_{-\infty}^{\infty } dz ( \delta(z{-}x'')-\delta(z{-}x')){\cal u}(z,\tau ) } \nonumber\end{aligned}\ ] ] one easily find that the final result may be cast to the form of equation ( [ e6_3 ] ) or ( [ e6_7 ] ) with @xmath198=\frac{1}{2}\int_0^t\int_0^t d\tau_1 d\tau_2 \ ( \phi(x_2''{-}x_1'',\tau_2{-}\tau_1 ) + \phi(x_2'{-}x_1',\tau_2{-}\tau_1 ) -2\phi(x_2''{-}x_1',\tau_2{-}\tau_1))\end{aligned}\ ] ] where @xmath199 is a short notation for @xmath200 .
we are now in position to compare the classical propagator ( ( [ e6_7 ] ) with ( [ e6_4])-([e6_5 ] ) ) , with the quantal one ( ( [ e6_7 ] ) with ( [ e6_4])-([e6_5 ] ) ) . in the latter case
@xmath168 is , in general , no longer a `` dummy variable '' .
the exception being the case where the actions are quadratic in the path variables , which is the case with bm model provided @xmath201 is quadratic .
the `` classical feature '' may be characterized as arising from invariance under the scaling transformation of the auxiliary integration - variable @xmath202 . in the quantal regime
the replacement @xmath203 can not be compensated by the scaling @xmath204 .
note however that the limit @xmath205 is equivalent to taking leading behavior of the actions in the limit @xmath206 .
here we follow closely the notations in ref.@xcite .
the path - integral expression for the reduced propagator of the probability density function is of the general form ( [ e6_7 ] ) , with @xmath207=s_{free}[r , r ] - \int_0^t d\tau ( v(x'',\tau){-}v(x',\tau ) ) + s_f[x'',x ' ] \end{aligned}\ ] ] the expressions for the reduced - action - functionals @xmath208 $ ] and @xmath209 $ ] in case of bm model , are given in equations ( 2.13)-(2.14 ) of the latter reference . in the bm model
the interaction is via the dynamical variable @xmath22 , while , in dld model , the interaction is via @xmath210 .
thus , in the expressions for the friction functional @xmath208 $ ] and for the noise functional @xmath209 $ ] one should make the replacements @xmath211 , and sum over @xmath36 .
thus one obtains @xmath212 \ = \ \frac{1}{\mbox{density } } \times \int_0^t\int_0^t d\tau_1 d\tau_2
\ 2\tilde{\alpha}(\tau_2{-}\tau_1 ) \ \frac{1}{2 } \sum_{\alpha } ( u_{\alpha}(x_2''){-}u_{\alpha}(x_2 ' ) ) ( u_{\alpha}(x_1''){+}u_{\alpha}(x_1 ' ) ) \nonumber\end{aligned}\ ] ] and @xmath213 \ = \ \frac{1}{\mbox{density } } \times \frac{1}{2 } \int_0^t\int_0^t d\tau_1 d\tau_2 \
\phi(\tau_2{-}\tau_1 ) \ \sum_{\alpha } ( u_{\alpha}(x_2''){-}u_{\alpha}(x_2 ' ) ) ( u_{\alpha}(x_1''){-}u_{\alpha}(x_1 ' ) ) \ \ \ \ \ \ .\nonumber\end{aligned}\ ] ] above @xmath214 .
these expressions may be simplified using ( [ e3_6 ] ) .
the results are @xmath215=\int_0^t\int_0^t d\tau_1 d\tau_2 \
2\tilde{\alpha}(\tau_2{-}\tau_1 ) \cdot \frac{1}{2 } [ w(x_2''{-}x_1'')-w(x_2'{-}x_1')+w(x_2''{-}x_1')-w(x_2'{-}x_1'')]\end{aligned}\ ] ] and @xmath216=\frac{1}{2 } \int_0^t\int_0^t d\tau_1 d\tau_2 \
\phi(\tau_2{-}\tau_1 ) \ [ w(x_2''{-}x_1'')+w(x_2'{-}x_1')-2w(x_2''{-}x_1')]\end{aligned}\ ] ] in the next paragraph we discuss further the physical significance of the functional @xmath208 $ ] . for the bm model
it has been noted before that in practice one could substitute ( [ e2_5 ] ) .
thus @xmath212=\int_0^t\int_0^t d\tau_1 d\tau_2 \
2\tilde{\alpha}(\tau_2{-}\tau_1 ) \ r(\tau_2)\cdot ( r(\tau_1)-r(\tau_2 ) ) \ \ \ \ . \nonumber\end{aligned}\
] ] however , the formally `` correct '' expression is somewhat different ( ref.@xcite equation ( 2.13 ) ) .
namely , the integrand in the above equation is with @xmath217 , rather than @xmath218 .
one may say that the bm reduced - action functional includes an additional term .
it is not difficult to demonstrate that the latter can be split into two terms that corresponds exactly to @xmath107 and to @xmath106 discussed in sec.iii(b ) .
the @xmath103 term may be absorbed in the definition of @xmath219 , while the @xmath220 term , which is @xmath221 , may be factored out of the path integral expression ( see ref.@xcite equation ( 2.34 ) ) .
it has the effect of operating on the initial probability function with an impulse that acts on the particle if it starts its trajectory in a point @xmath222 .
this term originates due to the fact that the initial preparation is such that the bath is in thermal equilibrium provided @xmath102 .
both the `` switching '' term and the additional effective potential , are absent in dld model .
this obvious results stems from the assumed inherent homogeneity of the environment .
we turn now to the general expression ( [ e7_2 ] ) . again , as in sec.iii(b ) , it is convenient to express @xmath86 as a sum of its symmetric , and antisymmetric continuations .
here we focus on the resultant friction functional , which is @xmath223 for ohmic bath one obtains @xmath224 where in the last stage we indicated the classical limit .
note that the classical expression for @xmath6 can be easily derived .
for this one should include the friction force in the derivation of sec.v(a ) . a totally different derivation of the path integral expression for the propagator is possible in the general case of weak coupling to thermal , possibly chaotic bath .
this derivation , in case of _ linear _ coupling , has been introduce already by fv ref.@xcite .
the case of general , nonlinear coupling , has been considered by mhring and smilansky ref.@xcite . here
we shall take a step further , and demonstrate that under `` normal '' circumstances , it will reduce to an ohmic dld model .
we consider a bath hamiltonian of the general form : @xmath225 where @xmath226 and @xmath227 are the eigenstates and eigenenergies of the bath hamiltonian in the absence of the coupling .
this hamiltonian depends on @xmath22 , the system variable , as a parameter .
however , once the full hamiltonian ( [ e3_4 ] ) is considered , @xmath22 becomes a dynamical variable .
the so called _ influence functional _ , in our notations is ( by definition ) : @xmath228 \ u_{mn}^{\star}[x'(\tau ) ] \ \ \ \ , \end{aligned}\ ] ] units with @xmath229 are used here .
@xmath230 $ ] is the evolution operator of the bath , in the presence of the `` driving force '' @xmath231 .
the bath is assumed to be in canonical thermal equilibrium .
the probability of the @xmath232-th eigenstate is @xmath233 . using leading order perturbation theory one obtains @xmath234 \ \approx
\ -i\int_0^t d\tau \ { \cal u}_{mn}(x(\tau ) ) \
\mbox{e}^{i(e_m - e_n)\tau } \ \ \ \ \ \ \mbox{for $ m\ne n$ } \
\ .\end{aligned}\ ] ] similar expression holds for @xmath235 ( see ref.@xcite ) .
substitution into ( [ e7_8 ] ) yields : @xmath236 \end{aligned}\ ] ] now we take a further assumption which will reduce the resultant expression for @xmath6 and @xmath8 to the form of ( [ e7_2 ] ) and ( [ e7_3 ] ) respectively .
the matrix elements @xmath237 are assumed to be real , while their dependence on @xmath22 is assumed to be characterized by the function @xmath238 the reduction of ( [ e7_10 ] ) to ( [ e7_2 ] ) and ( [ e7_3 ] ) is easily verified via algebraic manipulation , using @xmath239{=}\sum_{nm}(p_n{-}p_m)\{ .. \}$ ] and @xmath240{=}\sum_{nm } \coth(\frac{1}{2}\beta\omega)(p_n{-}p_m)\{ .. \}$ ] where @xmath241 .
thus , for general nonlinear coupling , dld model constitutes an equivalent representation for the bath , as far as the reduced dynamics of the system is concerned . for the particular case of the weak
_ linear _ coupling , there is further reduction to the bm model , with @xmath242 in the above expression @xmath243 denotes the collective bath degree of freedom via which the interaction takes place , namely , in ( [ e7_7 ] ) one should substitute @xmath244 .
the bath variable @xmath243 may be some complicated nonlinear combination of many elementary bath degrees of freedom . for both , the dld model and the bm model , the expression for @xmath46 ,
may be cast into the form @xmath245 where @xmath246 is the partition function , @xmath247 is the standard deviation of the off - diagonal matrix elements ( @xmath248 being the offset ) , and @xmath249 is the density - density autocorrelation function of the spectrum @xmath250 , appropriately averaged over the relevant energy scale .
the latter is related , for small energy differences , to the level spacing distribution . however , for any practical use , one should ignore the effect of level spacing statistics on @xmath46 , since it corresponds to non - physically very long times .
thus , the generic behavior of @xmath46 , for physically - relevant small @xmath248 , is expected to be @xmath251 .
this leads to the conclusion that under `` normal '' circumstances the ohmic dld model is good representation for the dissipation process .
this conclusion does not hold if _ strong _ coupling to a chaotic bath is considered . in the latter case zener transitions may dominate the dissipation process @xcite .
we do not know whether the influence functional ( [ e7_8 ] ) for that case can be reduced to a form that resembles that of the dld model .
the suppression of quantal interference is an important issue in both semiclassical and mesoscopic physics .
it may arise that the quantum mechanical propagator may be expressed as a sum of probabilities to go either via one classical trajectory @xmath252 or via a different classical trajectory @xmath253 , plus an interference term .
the expression for the _ influence functional _ may be used in order to compute the suppression of the interference due to the interaction with the environment .
the interference term is multiplied by a `` dephasing '' factor @xmath254 , where we follow a notation due to stern , aharonov and imry @xcite . in our notations ,
the dephasing factor is identified with @xmath255)$ ] .
we defer further discussion of interference within the framework of fv formalism to the last paragraph of this subsection .
it is enlightening to consider the case of white noise , with @xmath256 .
for the bm model one obtains @xmath257 \ \ \ \ . \end{aligned}\ ] ] thus , interference is suppressed more effectively if the two interfering paths are better separated .
a totally different result is obtained in case of the dld model .
here we assume that the two interfering paths are well separated with respect to the microscopic scale @xmath32 , namely @xmath258 most of the time .
it follows that @xmath259 \ \ \ \ . \end{aligned}\ ] ] here , the interference decays exponentially in time , and the actual spatial separation of the paths play no role . due to the disorder ,
dephasing events are as effective for small separations as for large separations .
the dephasing of interference in metals due to electromagnetic fluctuations has been discussed by altshuler , aronov and khmelnitskii in ref.@xcite .
their results have been re - derived by stern , aharonov and imry @xcite .
a somewhat simplified derivation is reconstructed in app.a .
the strong dimensionality dependence of the dephasing process has been emphasized .
this dependence is due to the separation - dependence as in ( [ e_supp ] ) , and see also ( [ ea_supp1 ] ) . in case of the dld model
, the local nature of the dephasing process will eliminate this feature . in order to understand how interference arise from the fv path integral expression ( [ e6_7 ] ) , it is convenient to rewrite it in the following form @xmath260 \ \ \ \ , \end{aligned}\ ] ] where @xmath261 $ ] is a real functional , which is defined by the expression : @xmath262 = \int_{unrestricted } \!\!\!\ !
{ \cal d}r \ \ \mbox{e}^{i\frac{1}{\hbar}s_{eff}[r , r ] } \ \ \mbox{e}^{-\frac{1}{\hbar^2}s_n[r , r ] } \ \ \ \ \ .\end{aligned}\ ] ] if the path integral expression for the evolution operator ( [ e6_9 ] ) is dominated by a _
single _ classical path @xmath252 , then the @xmath184 integration in ( [ e_krp ] ) will be dominated by @xmath263 .
obviously , in order to obtain a non - vanishing result , the endpoint conditions should be compatible . turning to the computation of @xmath264 $ ] via ( [ e_kr ] )
, one observes that the @xmath183 integration is dominated by the trivial trajectory @xmath265 .
we use the subscript @xmath185 in order to suggest that in general this trajectory should be @xmath185-dependent . indeed , this is the case if two classical trajectories @xmath252 and @xmath253 dominate .
one should consider then the `` interference path '' @xmath266 , for which the @xmath183 integration is dominated by the non - trivial paths @xmath267 .
the existence of non - trivial path @xmath268 is the fingerprint of interference phenomena .
a classical trajectory @xmath269 for which @xmath270 will not be damped , since @xmath271=0 $ ] then .
in contrast , an interference path , for which @xmath272 , is damped , since @xmath273 $ ] in general .
however , in sec.ix , where localization effect is discussed , we shall encounter a vast family of interference trajectories that are not damped by the noise functional . in the latter case
, the interference paths are found to be dominant in the computation of the propagator .
another issue that deserves attention is the interplay of friction and interference . by inspection of ( [ e7_5 ] )
it is clear that for _ disordered _ environment , quantal interference is unaffected by friction .
this is true as long as @xmath274 and @xmath275 are well separated in space . in sec.vii(b ) we shall encounter a related quantal manifestation of this observation .
the main results of the two last sections are the path integral expression ( [ e6_7 ] ) for the propagator @xmath12 , with the appropriate action functionals @xmath276 ( [ e6_8 ] ) , @xmath277 ( [ e7_5 ] ) , and @xmath8 ( [ e7_3 ] ) . the classical limit of @xmath8 is presented in ( [ e6_5 ] ) . for ohmic friction , both the quantal version and its classical limit ( [ e7_6 ] ) will be further considered and compared .
friction in case of non - ohmic bath has been discussed in sec.iv , and its quantal analog will not be considered further . in order to get preliminary insight into the path - integral expression , consider first the case of free particle in `` white '' noisy environment .
namely , @xmath278 , with @xmath279 as in [ e5_9 ] . in the classical case ( [ e6_5 ] ) , one obtains @xmath280 = \frac{1}{2 } \nu \int_0^t \
r(\tau)^2 d\tau \ \ \ .\end{aligned}\ ] ] independent of the spatial auto - correlation function @xmath26 .
the observation that spatial correlations are of no importance , as long as the noise is uncorrelated in time , is trivial from classical point of view . in the quantum mechanical case ,
the corresponding expression is @xmath281 = \nu \int_0^t \ ( w(0){-}w(r(\tau ) ) \ d\tau \ \ \ .\end{aligned}\ ] ] in contrary to classical intuition , spatial correlations may be of importance .
however , for the bm model , ( [ e2_5 ] ) , one recovers the classical result . in the path
integral expression ( [ e6_7 ] ) , one may perform the integration . in the absence of @xmath8 , each integration over @xmath282 results in delta function of velocities , namely @xmath283 . in the presence of @xmath8 ,
the integration @xmath282 is weighted , and as a result , each @xmath15 function is smeared .
the propagator constitutes a convolution of these smeared @xmath15 functions .
in particular , both in the classical case and in the bm model , each smeared @xmath15 function is a gaussian .
it is obvious , that both in the classical and in the quantal case , @xmath8 leads to stochastic - like spreading . in
what follow we want to estimate this spreading .
we turn now to estimate the spreading _ in the absence of disorder _ , which is the standard bm model .
the noise functional is quadratic in the path - variable @xmath284 , and is independent of the path - variable @xmath185 , namely @xmath285=\frac{1}{2 } \int_0^t\int_0^t d\tau_1 d\tau_2 \
\phi(\tau_2{-}\tau_1 ) \ r(\tau)r(\tau')\end{aligned}\ ] ] here , an exact treatment is available @xcite .
one may expand both @xmath276 and @xmath8 around the so - called classical paths , that are determined by the variation @xmath286 , with the constraints @xmath287 at the endpoints .
the gaussian integration is performed exactly .
general expressions may be found in ref.@xcite .
the phase space propagator @xmath12 is obtained by double fourier transform of @xmath180 , and obviously results in gaussian function . in particular , one is interested in the spatial spreading . setting @xmath288 and integrating over the final momentum @xmath289 , one obtains @xmath290\end{aligned}\ ] ] a general expression for the spatial spreading , that applies to any @xmath28 , may be obtained @xcite : @xmath291 where @xmath292 solves the linearized classical equation of motion ( @xmath293 ) with initial conditions @xmath294 and @xmath295 .
note that ohmic friction is considered , for which the bm model is well defined .
for the simplest case of white noise without friction one obtains @xmath296 .
friction leads to damping and diffusion . in the latter case , considering long time @xmath297 , one may disregard a short transient and substitute @xmath298 .
consequently @xmath299 .
however , at low temperatures @xmath300 while @xmath301 . thus @xmath302 diffusion is suppressed due to the negative autocorrelations of the noise .
this effect is classical in nature .
intentionally we did not use the explicit expression for the constant , namely @xmath303 .
the presence of @xmath304 in the formula , rather than @xmath305 , without explicit subscript , may miss - lead the reader .
the particle can be treated as a classical object , within the framework of e.g. langevin equation , and still the suppression of diffusion will occur .
both the classical limit and the bm model generate `` classical '' dissipation effects .
we turn now to the dld model , ( ( [ e8_2 ] ) with ( [ e2_6 ] ) ) . here
the situation is quite different . in order to compute the propagator , a different , more powerful strategy is required .
we shall exploit the fact that for white noise @xmath306 $ ] is still independent of @xmath269 , while @xmath307 $ ] is linear in @xmath269 .
it is most convenient to fourier transform @xmath180 to @xmath308 .
the path integral expression may be written for this representation as follows _ ( we suppress from now on printing @xmath168 , but shall restore it later ) _ : @xmath309d\tau } \nonumber\end{aligned}\ ] ] the @xmath184 integration can be performed now , yielding a delta function @xmath310 at each point along the trajectory . the trajectory for which the integrand does not vanish will be denoted by @xmath311 , where @xmath312 .
this trajectory is required to satisfy the endpoints conditions @xmath313 and @xmath174 .
hence , the result of the path - integration is @xmath314d\tau } \nonumber\end{aligned}\ ] ] the inverse fourier transform yields @xmath315 } \
\ \mbox{e}^{-\nu\int_0^t[w(0){-}w(r(r_0,r,\tau)]d\tau } \nonumber\end{aligned}\ ] ] to simplify the latter expression we note that @xmath316 is a constant of the motion .
the propagator @xmath12 is the fourier transform in the variables @xmath284 and @xmath317 . in order to get insight
we restrict ourselves to the reduced kernel @xmath318 , namely @xmath319 } \ \ \mbox{e}^{-\nu\int_0^t(w(0){-}w(r(r_0,0,\tau))d\tau}\end{aligned}\ ] ] as before we distinguish the case of frictionless propagation , for which @xmath320 , from the case of damped particle with @xmath321 . in the latter case
the trajectory is modified for @xmath322 , where @xmath323 .
consequently the `` phase '' in ( [ e9_1 ] ) is @xmath324 and @xmath325 for the two corresponding types of trajectories . as for the noise argument in ( [ e9_1 ] ) , it is @xmath326 for @xmath327 , and @xmath328 or @xmath329 for @xmath322 , depending whether friction is absent or present respectively .
the separation of scales both in @xmath330 and in @xmath331 enables splitting the integral in a convenient way , namely @xmath332 where @xmath333 is a smooth , symmetric cutoff function that equals @xmath334 for @xmath335 and equals @xmath336 for @xmath337 . the integration in ( [ e9_1 ] )
is performed , and the following expression is obtained for the propagator .
@xmath338 \ \ \ .\end{aligned}\ ] ] the symbol @xmath339 stands for convolution , and results in smearing of the classical propagator on scale @xmath340 . for frictionless propagation one should use the replacement @xmath341 .
we have used the notation @xmath342 .
the above result holds also if @xmath343 , in this case @xmath344 will propagate as if friction is absent , while @xmath345 will propagate as in the classical limit .
obviously , if friction is indeed absent , then again @xmath344 and @xmath345 will coincide .
the kernel @xmath346 denotes the classical result , ( [ e8_5 ] ) .
the expression for the quantal propagator demonstrates that a piece of the wavepacket is frozen due to the disorder .
this is a non trivial quantal effect that indeed can be entitled `` quantum dissipation '' .
we emphasize again that such quantal effect is absent in the bm model .
however , the expression for the propagator also demonstrates that the `` quantal correction '' goes to zero exponentially in time , as in the case of interference discussed in sec.vi(c ) .
the disadvantage of the treatments that have been presented in the preceding section , is the difficulty to extend them to the _ general _ case of disordered environment , namely if the disorder ( the noise ) is correlated in time .
we therefore turn to a somewhat more heuristic approach , that will enable approximated treatment . for the computation of @xmath12 we shall use the classical limit ( [ e6_5 ] ) of @xmath306 $ ] .
the `` quantal correction '' in ( [ e9_2 ] ) , is not considered again in the present section . in the classical limit ( [ e_krp ] )
constitutes a formal solution of langevin equation .
the real functional @xmath261 $ ] has a simple probabilistic interpretation . for ohmic friction it takes the form @xmath347 \
\ = \ \ \ \int { \cal d}r \ \
\mbox{e}^{-i\int_0^t d\tau \
\delta_r(\tau ) \
r(\tau ) } \ \ \mbox{e}^{-\frac{1}{2}\int_0^t\int_0^t
d\tau d\tau ' \ \phi_r ( \tau,\tau ' ) \
r(\tau ) \ r(\tau')}\end{aligned}\ ] ] where @xmath348 and @xmath349 formally , the unrestricted @xmath183 integration may be performed exactly , yielding the result @xmath350 \ = \ \sqrt{\det[\phi_r]}\cdot \exp\left(-\frac{1}{2}\int_0^t\int_0^t d\tau d\tau ' \
\phi_r(\tau,\tau ' ) \
\delta_r(\tau ) \ \delta_r(\tau')\right)\end{aligned}\ ] ] where @xmath351 is the reciprocal of @xmath352 . in order to compute @xmath12
one should identify the most contributing paths , for which @xmath261 $ ] is maximal . in subsection
a , where we discuss short - time correlated colored noise , we assume that _ one _ optimal path dominates the computation . in subsection b , where we discuss static or almost - static noisy potential
, we shall identify a whole family of optimal paths . in the next section , where quantum localization is discussed
, we shall use a similar strategy , and a family of interference paths will be identified . in this section
we shall analyze the diffusive behavior which is encountered in the absence of disorder .
our assumption will be that the @xmath184 integration is dominated by _
one _ smooth `` optimal path '' .
we shall substantiate this assumption by demonstrating consistency with the exact result that has been presented in sec.vii(a ) .
furthermore , it will be argued that the `` optimal path '' for short - time correlated noise is the same as for white noise . in order to find the path @xmath353 that maximize @xmath261 $ ]
, we consider first the case of white noise , where @xmath354 .
hence @xmath355 , with @xmath356 ( the subscript @xmath185 is reserved for later use ) . as in sec.vii(a ) we focus the attention on the computation of the reduced propagator @xmath318 .
formally , the path integral expression for @xmath318 is identical with ( [ e_krp ] ) , except for the restriction at the endpoints . for @xmath318 ,
the relaxed constraints are @xmath357 , @xmath358 , and @xmath359 . denoting @xmath360 ,
the variational equation for @xmath353 , including lagrange multiplier , is @xmath361 d\tau = 0 \nonumber\end{aligned}\ ] ] one obtains @xmath362 the last two equations are the constraints .
the solution for damped propagation ( @xmath363 ) is easily found .
for sake of comparison also the solution for frictionless propagation ( @xmath364 ) is displayed .
@xmath365 in the first formula a constant prefactor that equals @xmath334 has been dropped , since we assume here @xmath321 .
the optimal paths of ( [ e10_6 ] ) are illustrated in fig.[traj_dif ] .
computation of @xmath366 for the optimal path is straight forward . for damped propagation
the computation is trivial since @xmath367 is constant .
substitution into ( [ e10_5 ] ) yields : @xmath368 \ = \ \
const\cdot\exp\left(-\frac{1}{2}\cdot \frac{(r{-}r_0)^2}{\frac{1}{\eta^2}\ \nu_r \ t}\right ) \ \ \ , \end{aligned}\ ] ] which is in consistency with the exact result of sec.vii(a ) .
it is easily verified that also for frictionless propagation , consistency with the exact result is maintained . for short - time
correlated colored noise ( @xmath369 ) it is natural to replace @xmath9 by @xmath370 , with the effective white noise intensity @xmath371 .
as long as @xmath372 ( finite temperatures ) , the long time behavior is diffusive , and consistency with ( [ e8_7 ] ) is easily verified .
still , a more elaborated argument is required in order to substantiate the `` white noise approximation '' .
this argument will be discussed now . by inspection of ( [ e10_5 ] )
it is clear that the most contributing paths , for which @xmath261 $ ] is large , must satisfy @xmath373 . for white noise @xmath374 , but still the `` optimal path '' is smooth . by `` smooth ''
one means that @xmath375 is concentrated within the interval @xmath376 where @xmath377 is the relevant time scale for the system s dynamics .
consider now a noise auto - correlation function of the form @xmath378 .
its fourier transform satisfies @xmath379 for @xmath380 .
thus , the first requirement for the `` white noise approximation '' to hold should be @xmath381 . for @xmath382 the spectral function @xmath383 drops to zero , which implies that such high frequency components are not favored .
so far there is consistency with our assumption that the most contributing paths are smooth .
however , in the vicinity of @xmath149 the spectral function @xmath383 is peaked .
it implies , that unlike the case of white noise , an oscillatory component with time period @xmath100 is favored . obviously , such component arise from the strong accelerations that the particle experiences within short periods whose duration is @xmath100 . over these short periods
the maximum displacement is @xmath384 . using ( [ e5_6 ] ) it
is found that this amplitude is proportional to @xmath385 .
for @xmath386 the amplitude @xmath387 goes to zero as @xmath388 .
therefore , in this restricted regime ( @xmath389 ) , and in particular for @xmath164 ( low temperature ohmic noise ) the `` white noise approximation '' should be adequate . encouraged by the consistency of the heuristic approach with exact results
, we turn now to analyze the diffusion due to short - time correlated noise _ in the presence of disorder_. we shall use the `` white noise approximation '' whose validity has been discussed in the preceding subsection .
namely , for short range correlated noise , the most contributing paths are concentrated around the same @xmath353 that has been found for white noise . from ( [ e10_6 ] )
it follows that @xmath353 is , up to end point transients , a free - like propagation .
consequently , the effective noise autocorrelation function is @xmath390 and we define @xmath391 \ d\tau\end{aligned}\ ] ] in general @xmath392 , also in the limit of zero temperature . in the presence of disorder , the effective white noise intensity is ( in general ) a function of the endpoint conditions , rather than a constant . for typical noise autocorrelation function of the form
@xmath393 one obtains @xmath394 the computation has been carried out by taking in ( [ e10_8 ] ) the fourier transform of both @xmath395 and @xmath396 , and performing @xmath397 integration rather than @xmath165 integration .
substitution into ( [ e10_7 ] ) suggests that : @xmath398 the `` tail '' of the dispersion profile is universal .
it depends on @xmath149 , but it is independent of the nature of the noise . in contrast , the short range profile is determined by the low frequency bath - oscillators , and thus it is sensitive to the exact value of @xmath115 . for @xmath123 , ( ohmic model , high temperatures ) , normal diffusive behavior prevails . for @xmath164 ( ohmic model , low temperatures ) the diffusion freezes
the dispersion profile is exponential rather than gaussian , namely @xmath399 } \right ) \ \ \ \ \ \ \ \mbox{[ohmic , low temperatures]}\end{aligned}\ ] ] the dispersion here , in the dld model , is of the order @xmath400 rather than @xmath401 .
( bm model ( [ e8_9 ] ) ) .
one should observe that on for @xmath402 both models predict consistently dispersion on spatial scale @xmath32 . for larger noise intensity
, the bm model is not valid , and dld predicts always a larger dispersion , which is intuitively expected . for weak noise ( @xmath403 ) the spatial spreading
is on scale less than @xmath32 .
our treatment of the dld model is not valid on this microscopic scale .
however , in this regime the bm model can be trusted . for @xmath404 equation ( [ e10_10 ] ) implies that the particle is evacuates from the vicinity of @xmath405 .
the `` body '' of the `` diffusion '' profile ( [ e10_10]a ) is determined by a low velocity paths for which @xmath406 .
it is easily verified that a sufficient condition for the validity of the `` white noise approximation '' is @xmath407 .
this condition is satisfied in the relevant spatial range ( @xmath408 ) , except for a relatively small interval around @xmath405 , which is determined by the large ratio @xmath409 .
the `` tail '' of the `` diffusion '' profile ( [ e10_10]b ) is determined by a high velocity paths for which @xmath410 . here
the the validity argument should be modified .
the spectral function @xmath383 is peaked around @xmath411 rather than around @xmath412 .
thus , oscillatory component which is characterized by period @xmath155 is favored .
the maximal spatial amplitude of this component is @xmath413 .
it is convenient to use a special notation for the standard deviation of the disordered potential , namely @xmath414 .
using this notation the amplitude is @xmath415 .
this amplitude is required to be much less than @xmath32 , or alternatively @xmath416 or alternatively @xmath417 , with @xmath418 . in this subsection
we have limited the discussion to the case of short - time correlated noise for which @xmath100 is small in some sense . indeed ,
if it is assumed that @xmath419 , then the validity condition will be satisfied automatically .
in this subsection we shall discuss the case of static or almost - static disordered non - dynamical environment ( @xmath364 ) .
the `` noise '' is assumed to possess long - time correlations . specifically :
@xmath420 where @xmath421 is the standard deviation of the disordered potential . here
@xmath100 is assumed to be large , much larger than @xmath418 .
in general the `` white noise approximation '' breaks down .
the path integral expression is no - longer dominated by `` smooth '' trajectories .
it is difficult to use the explicit formula ( [ e10_5 ] ) in order to identify the family of `` optimal paths '' since @xmath422 is no longer diagonal .
we therefore prefer to use heuristic considerations .
no new insight is gained if one insists using ( [ e10_5 ] ) .
we take first the limit @xmath423 .
the classical mean free path of the particle is @xmath424 the corresponding time is @xmath425 . after that time
the probability of being backscattered is of order 1 . in _
one dimension _ this backscattering will lead to ( classical ) localization of the particle .
the localization length is exponentially large for high energies .
if @xmath100 is finite , rather than infinite , classical localization will manifest itself only if @xmath426 . within the time scale @xmath100 , the particle spreads over spatial range of the order @xmath427 , while its velocity is randomized .
it follows that @xmath318 may be used as a stochastic kernel .
hence the markovian property is recovered over time scales that are much larger than @xmath100 , and a diffusive behavior follows with coefficient @xmath428 .
this diffusion is non - dissipative `` random walk '' like .
as in the last subsection , we shall discuss the case of quenched disorder .
however , here the _ quantal _
analysis will be carried out .
the variance of the disordered potential is @xmath429 , with auto - correlation length @xmath32 .
we approximate the gaussian correlation ( [ e2_6 ] ) by a delta function . defining @xmath430 , the path - integral expression for
the propagator is ( [ e6_7 ] ) with ( [ e6_10 ] ) and @xmath431 = \frac{1}{2 } aw^2 \int_0^t\int_0^t ( \delta(x_2''{-}x_1'')+\delta(x_2'{-}x_1')-2\delta(x_2''{-}x_1 ' ) ) d\tau_1 d\tau_2 \end{aligned}\ ] ] the essential feature of this functional , is its non - local nature .
our first step will be to get some insight into @xmath8 .
let us consider segment @xmath432 and segment @xmath433 that belong both to the same path , either @xmath434 or @xmath435 , within the same spatial interval @xmath436 .
this includes the possibility @xmath437 .
the contribution to @xmath8 is @xmath438 where @xmath439 is the velocity @xmath440 within @xmath436 .
however , if @xmath432 belongs to @xmath434 , while @xmath433 belongs to @xmath435 , or vise versa , then the contribution to @xmath8 is @xmath441 one easily convince oneself that the contribution of each spatial interval @xmath436 is _ non - negative_. a zero value may be obtained if to each segment @xmath432 , that belongs to @xmath434 , corresponds segment @xmath433 that belongs to @xmath435 , with @xmath442 .
for example , a zero value for @xmath8 may be obtained if @xmath191 is shifted in time with respect to @xmath190 . referring to the endpoints
, one should identify @xmath443 with @xmath125 .
it follows that @xmath444 implies that the two paths @xmath434 and @xmath435 are either _ identical _ , or satisfy the constraint @xmath445 . if @xmath446 , the value @xmath444 may be obtained only if @xmath434 and @xmath435 are identical . more generally ,
if @xmath447 , one may prove that the following inequality holds for any pair of smooth paths @xmath434 and
@xmath435 , @xmath448 } \ \ \le \ \ e^{-\frac{aw^2}{(\hbar v_0)^2 } \min(|r_t{-}r_0|,\frac{1}{2}(|r_t|{+}|r_0| ) ) } \ \ \ .\end{aligned}\ ] ] for particular @xmath269 one may ask what is the @xmath202 for which @xmath306 $ ] is minimal . the trivial minimum , which is also the absolute minimum , is @xmath449 , for which @xmath306=0 $ ]
. however , any small perturbation on @xmath202 will make @xmath306 $ ] much larger .
therefore , we are tempted to assume that there may be some other , more stable ( local ) minimum @xmath450 .
a non - trivial local minimum @xmath450 does not exist for any @xmath269 .
however , one can prove that there is a large family of @xmath185-s for which such minimum exists , by actually constructing them .
this is done by following the considerations that were presented at the beginning of this paragraph , an example for such construction is presented in fig.[traj_loc ] .
the situation here should be contrasted with that encountered in sec.viii . in subsection a ( there ) we could have defined _ one _ optimal path @xmath353 . here , there is a whole family of `` optimal paths '' , as in the case of subsection b ( there ) .
however , in the present case these paths are `` interference paths '' rather than `` classical paths '' .
the following observations concerning the relevant optimal paths are important : ( a ) they consist of many straight segments , and have zigzag character ; ( b ) the final velocity @xmath451 is favored to be equal in absolute value to the initial velocity ; ( c ) turning points impose significant restrictions ; ( d ) the non trivial minimum @xmath450 is isolated ; ( e ) the non - trivial minimum is relatively stable .
the last point is the most difficult to observe .
first it should be noted that ( e ) must be true a - priori .
else , if the trivial minimum dominates the path - integral expression , then the result would be that the particle has roughly the same probability to go from any initial conditions to any final conditions , irrespective of proximity considerations or even energy conservation .
still , a reasonable argument is required why the non trivial minimum is relatively stable . for this
consider a straight segment @xmath432 for which @xmath452 and @xmath453 .
one observes that if @xmath268 is perturbed by a fluctuation of time period @xmath454 , or by some higher harmony , then the contribution to @xmath306 $ ] is negligible .
this is to be contrasted with the case of the trivial minimum @xmath449 , where any fluctuation has high cost .
we turn now to the formal extension of the procedure that has been presented in the previous section .
we expand @xmath306 $ ] around the non - trivial minimum : @xmath455 \approx s_n[r , r_r ] + \int_0^{\infty}\int_0^{\infty}d\tau d\tau ' \
\phi_r(\tau{-}\tau ' ) \cdot ( r(\tau){-}r_r(\tau ) ) \cdot ( r(\tau'){-}r_r(\tau ' ) ) \ + \ ... \end{aligned}\ ] ] here we are not able to write an explicit expression for the highly complicated kernel @xmath456 . however , we proceed , and write down the result of the gaussian integration : @xmath457 \ = \ \ const \cdot \cos\left(\int_0^{\infty}d\tau \ \delta_r(\tau)r_r(\tau)\right ) \
\mbox{e}^{-s_n[r , r_r ] } \ \
\\ \cdot \ \
\exp\left(-\frac{1}{2}\int_0^t\int_0^t d\tau d\tau ' \ \phi_r(\tau-\tau ' )
\ \delta_r(\tau ) \ \delta_r(\tau')\right)\end{aligned}\ ] ] now , we should perform the @xmath184 integration .
this integration will be dominated by the family of `` optimal paths '' .
note that the @xmath458 term in ( [ e11_4 ] ) equals unity since for @xmath268 the trajectories @xmath434 and @xmath435 are in a sense `` shifted '' one with respect to the other , hence @xmath459 . within the family of optimal paths ,
not all have the same contribution .
one should expand around those that have the largest contribution .
for these paths @xmath306=\frac{aw^2}{v_0 ^ 2}|r - r_0|$ ] .
here we consider endpoints conditions @xmath460 at @xmath70 and @xmath461 at time @xmath462 .
the time @xmath462 is assumed to be sufficiently large to guaranty steady state distribution .
it follows that @xmath463 where @xmath464 for convenience @xmath168 has been restored in the latter formula . in the application of the inequality ( [ e11_2 ] ) a sub - family of `` optimal paths '' has been ignored , for which @xmath465 .
it is justified provided this sub - family constitutes a zero fraction of the whole family .
for this we should assume sufficiently long time ( @xmath462 ) , for which a steady state distribution is attained .
both the _ validity _ and the _ applicability _ of the inequality ( [ e11_2 ] ) demonstrates the vulnerability of our localization argument .
it is important to consider circumstances in which either of these conditions is not satisfied .
the inequality ( [ e11_2 ] ) will not be _ valid _ if the noise is not infinitely correlated in time , i.e. the disordered potential is not static when viewed on large time scales .
the inequality ( [ e11_2 ] ) will not be _ applicable _ if additional white noise is added to the hamiltonian . in the latter case
, large @xmath450 will be suppressed by the corresponding additional term in the noise functional .
consequently , the sub - family of paths for which @xmath465 will not constitute zero fraction of the whole family .
expression ( [ e11_6 ] ) for the localization length agrees with the well know result for 1d localization of `` free '' particle ( thouless ref.@xcite ) .
note that thouless uses some scaled units , resulting in the expression @xmath466 .
we shall show now that the above expression is somewhat more general , and applies also to cases where the dispersion law is different from @xmath467 .
for this one should replace in ( [ e6_9 ] ) the kinetic term @xmath468 by some general function @xmath469 , resulting in @xmath470 where @xmath471 is the dispersion law .
the subsequent formalism is easily generalized .
the expression ( [ e11_6 ] ) for @xmath472 is unaffected .
actually , the general expression for @xmath472 could have been guessed .
we use naively the born approximation , calculate the mean free path @xmath32 , and rely on the fact that @xmath472 is twice the mean free path ( thouless ref.@xcite ) . by the golden rule ,
the probability of being backscattered is @xmath473 where @xmath474 is the length of the available space , and @xmath475 is the kinetic energy .
the matrix element is @xmath476|^2 } = \frac{1}{l } aw^2\end{aligned}\ ] ] where in the last equality ft [ ] denotes fourier transform , and @xmath20 is assumed to be uncorrelated in space ( `` white spatial noise '' ) .
the result ( [ e11_6 ] ) is easily recovered .
one should use @xmath477 , and @xmath478 , while @xmath32 is evaluated via the relation @xmath479 . in order to further demonstrate the generality of ( [ e11_6 ] ) ,
let us consider anderson tight binding model .
the spacing will be denoted by @xmath480 .
the hamiltonian is where @xmath481 is uniformly distributed in @xmath482 $ ] .
the transition amplitude is @xmath483 .
the kinetic energy in the center of the band is @xmath484 , where @xmath485 is the momentum .
hence @xmath486 there .
the dispersion of the on - site energies is @xmath487 .
the `` area '' under the triangular autocorrelation function is @xmath488 .
formula ( [ e11_6 ] ) suggests that @xmath489 .
this results agrees with that of ref.@xcite ( equation ( 66 ) there ) , including the prefactor .
a unified treatment of diffusion localization and dissipation ( dld ) has been presented in this work .
all these phenomena may be derived from the general path integral expression ( [ e6_7 ] ) , @xmath490 } \ \ \mbox{e}^{-\frac{1}{\hbar^2}s_n[r , r ] } \ \ \ \ , \end{aligned}\ ] ] upon inclusion of the appropriate functionals @xmath491 and @xmath8 .
general expressions for these functionals are available and various limits may be considered : ( a ) quantal versus classical expressions ; ( b ) disordered versus non - disordered environment ; ( c ) dissipative versus non - dissipative environment ; ( d ) quenched versus noisy environment . in the classical limit
the dld model constitutes a formal solution of langevin equation .
the classical limit may be obtained by linearization of the quantal @xmath307 $ ] with respect to @xmath284 , while expanding @xmath306 $ ] to be quadratic in this path - variable .
the disorder or its absence depends on the choice of the spatial auto - correlation function @xmath26 . the dissipation is turned on if the friction functional @xmath6 is included in @xmath491 .
the nature of the noise , whether it is `` quenched '' , `` colored '' or `` white '' , is determined by the noise kernel @xmath28 . in the latter case any combination may be considered as well ( see further discussion at the end of this section ) .
the classical bm model is well defined in terms of an appropriate langevin equation only in the case of an ohmic bath .
this is not the case with the classical dld model .
the classical dynamics in the latter case is well defined in terms of appropriate langevin equation also for non - ohmic bath .
explicit expressions for the friction force , and for the effective mass have been derived .
another nice feature of the dld model is the absence of `` switching impulse '' . once the noise auto - correlation function @xmath28 is specified , the bm model is in - distinguishable from its classical limit .
as long as the external potential @xmath219 is quadratic ( at most ) , the quantal propagator is identical with the classical one , and langevin equation can be used in order to describe correctly the time evolution of wigner function .
all the quantal effects that are associated with the standard zwanzig - caldeira - leggett bm - model are ( formally ) reproduced by solving the classical langevin equation with an appropriate noise term . _
the dld model is different_. the non - stochastic , genuine quantal features of the dld model have been discussed .
these features constitute a manifestation of either the disorder or the chaotic nature of the bath . for either noisy or ohmic environment both the bm model and
the dld model leads to either spreading or diffusion . for the bm model ,
the spreading and the diffusion profiles are described by a gaussian distribution . for
the dld model one should include a quantal correction .
the disorder freezes a piece of the wavepacket , letting it to propagate as if it were a free particle . both , the `` quantal correction '' to the propagator , as well as any other interference phenomena , die out exponentially in time .
this exponential decay , due to dephasing , is independent of geometry .
_ it should be contrasted _ with the results for loss of interference in the presence of bm - like environment @xcite .
another important observation is that for _ disordered _ environment , quantal interference is unaffected by friction . on the classical level
it is fascinating to analyze the diffusion profile in the presence of disorder . for the low temperature ohmic bm - model , it is found that diffusion is suppressed , though its gaussian profile is maintained .
the dld model , in the same circumstances , leads to an exponential profile that does not change with time .
this new effect is due to the interplay of the temporal ( negative ) autocorrelations of the noise with the spatial disorder .
even more fascinating `` diffusion '' profiles are found for other types of noise autocorrelations .
quenched disorder in _ one dimension _
leads to classical localization , as well as to quantal localization .
the former is characterized by a localization length which is exponentially large at high energies .
quantal localization on the other hand is dominated by interference phenomenon .
we have identified the interference paths within the framework of fv formalism , and demonstrated how the well known exponential profile emerges .
the localization length in the quantal case is proportional to the square of the velocity .
the applicability of this result to other dispersion relations has been pointed out .
it is obvious from the derivation , that localization can not be argued if the noise is not strictly static ( e.g. slowly modulated ) .
diffusive behavior is recovered if the noise possess long but _
finite _ auto - correlation time . also
the case of white noise `` on top '' of the static disorder will evidently lead to diffusion @xcite@xcite . in this latter case , which has not been considered in this paper , non
perturbative effects may manifest themselves @xcite as in the case of the `` quantum kicked particle '' @xcite .
the dld model should also account for diffusion in the presence of both quenched disorder , noisy potential and friction ( all together ) .
there should be a way to derive systematically well known heuristic results that corresponds to hopping , or variable range hopping @xcite .
classical non - dissipative `` random walk '' diffusion has been discussed in the restricted case of long - time correlated noise .
it is essential to generalize the dld model to _ more than one dimension _ in order to account for this phenomenon in the case of strictly static disorder .
auto - correlations of the disordered potential , both in time and in space , should be considered , in order to generate the dynamics which is described by boltzmann transport equation . in the limit of quenched disorder , localization effect should be encountered .
+ acknowledgment i thank shmuel fishman from the technion for supporting and encouraging the present study in its initial phases .
i also thank harel primack and uzy smilansky for interesting discussions and suggestions , and yuval gefen for his comments .
the research reported here was supported in part by the minerva center for nonlinear physics of complex systems .
this appendix illustrates how the unified formalism that has been presented in this paper should be modified in order to deal with external bath that consists of extended field modes . unlike the case of the bm - model ,
this modification is not as immediate as the mere substitution of appropriate @xmath26 . in the present derivation
the bath hamiltonian is not specified , and also the assumption @xmath492 is altered .
for the standard derivation of the dld model it has been assumed that the interaction hamiltonian is ( [ e3_2 ] ) with @xmath492 , leading to the factorized noise auto - correlation function ( [ e2_4 ] ) . in the more general case
the classical derivation as well as the quantal derivations lead to a stochastic force that satisfies ( [ e2_3 ] ) or to the noise functional ( [ e6_11 ] ) respectively , with noise auto - correlation function which is @xmath493 here @xmath494 , and the interaction hamiltonian ( [ e3_2 ] ) is used with @xmath495 without loss of generality .
the fields mode are still assumed to be decoupled , but the bath hamiltonian is not specified . instead , one relys on fluctuation - dissipation theorem as in ref.@xcite , writing the general expression : @xmath496 \ \ \ \ .\end{aligned}\ ] ] note that if the field modes are simple harmonic oscillators , then one should substitute @xmath497 .
alternatively , one may speculate @xmath498 using known response characteristics of the bath . to make further progress the interaction hamiltonian ( [ e3_2 ] ) should be further specified .
using `` standing waves '' decomposition it is assumed to be @xmath499 substitution of the appropriate @xmath500 into ( [ ea_phi ] ) , and converting the summation to an integral over directions and over @xmath501 , the result can be cast into the following form @xmath502 the average over directions can be performed leading to @xmath503 in 1-d , bessel function @xmath504 in 2-d , and @xmath505 in 3-d . in ref.@xcite the interaction with electromagnetic fluctuations in metal has been considered .
it has been assumed that each mode is characterized by an ohmic response . in our notations
it corresponds to @xmath506 . due to this assumption
, the noise auto - correlation function becomes factorized at high temperatures , as in the dld model .
in particular , for the noise functional ( [ e6_11 ] ) one obtains @xmath507 \ =
\ \int_0^t d\tau \
\frac{1}{\beta } \int_0^{\infty}\frac{k^{d{-}1}dk}{\pi } \
\eta(k ) \ \langle\sin^2(k\hat{\omega}\cdot r(\tau ) ) \rangle_{\omega } \ \ \ \ . \end{aligned}\ ] ] for the discussion of dephasing , we write again this result using the notations of ref.@xcite .
it is assumed that @xmath508 , where @xmath509 is the charge of the electron and @xmath510 is the conductivity .
integration over @xmath501 leads to the result @xmath511 \ \ \ \ \mbox{for $ d \le 2 $ } \ .
\end{aligned}\ ] ] this result , that has been obtained in ref.@xcite ( equation ( 7.8 ) there ) , is very similar to the corresponding result ( [ e_supp ] ) for the bm - model , the difference being the power @xmath512 which is not universal .
an apparently simpler example for an interaction with external bath that consists of extended field modes , is the electron - phonon interaction .
here we want to question the applicability of the bm model as an approximated description .
the electron is assumed to be confined to a one dimensional `` quantum wire '' , while the phonons dwell in the 3-d bulk .
considering longitudinal modes , the coupling of the @xmath513 `` oscillator '' with the electron is @xmath514 , with a coupling constant @xmath515 . summing over @xmath36 , as defined in ( [ e3_5 ] ) , one obtains effectively @xmath516 where @xmath517 is the dimensionality of the space where the phonons dwell .
thus a phonons - bath is similar to a superohmic bm - model .
however , this derivation is somewhat miss - leading , since a cutoff @xmath149 should be introduced , while for the original electron - phonon interaction a natural cutoff @xmath518 exists .
in this appendix we consider the interference phenomenon from two different points of view , and demonstrate their consistency .
first we consider the free propagation of two - wavepackets superposition .
the decay of the interference pattern will be dictated by the propagator ( [ e9_2 ] ) .
then we consider the scattering of a simple wavepacket from a double - barrier .
the suppression of interference paths will be dictated by the noise functional ( sec.vi(c ) ) .
finally , we argue that both points of view are physically equivalent , and must lead to the _ same _ result , which is indeed the case . consider a superposition of two gaussian wavepackets which have the same momentum @xmath519 , the same initial spatial spreading @xmath520 , while their initial locations satisfies @xmath521 .
the wigner function for this preparation is easily computed , and is of the form @xmath522 where @xmath523 denotes `` minimum - uncertainty '' gaussian distribution , and @xmath524 . for free propagation
@xmath525 will develop fringes on the spatial scale @xmath526 .
note that this gedanken experiment is formally equivalent to the usual `` two slit '' diffraction experiment upon the definition @xmath527 .
for propagation in noisy non - disordered environment the interference pattern is smeared on scale @xmath528 , due to the diffusive momentum spreading .
the smearing factor is @xmath529 $ ] leading to an exponential decay @xmath530 $ ] , that depends on the separation @xmath531 .
on the other hand , for propagation in noisy disordered environment , using ( [ e9_2 ] ) , the exponential decay is @xmath532 $ ] , independent of geometry .
consider now scattering problem in one dimension .
the potential is assumed to be @xmath533 for @xmath534 , and @xmath535 for @xmath536 .
a simple gaussian wavepacket with momentum @xmath537 and spatial spreading @xmath538 is launched from @xmath405 .
the scattered particle is detected at the range @xmath185 which is assumed to be much larger than @xmath405 . in the absence of noise and dissipation , it is not difficult to work out the explicit solution of this scattering problem .
a `` train '' of gaussian - like wavepackets will emerge from the scattering region .
the spatial separation between each two wavepackets is @xmath531 , and the probability density function will contain interference pattern in between . from the fv path - integral point of view ( sec.vi(c ) ) , the interference pattern is due to the existence of `` interference paths '' . in the presence of noisy disordered environment the contribution of these interference paths to the propagator
is suppressed exponentially in time .
assuming that the dephasing during the time interval @xmath539 is negligible , one may use the results of the first gedanken experiment , leading to the same exponential decay .
thus , the result ( [ e9_2 ] ) for the propagator is consistent with the analysis of interference in sec.vi(c ) .
our gedanken experiment can be used also in order to illustrate and clarify the observation that friction does not affect the interference phenomenon .
this observation holds for disordered dynamical environment .
mott and e.a .
davis , _ electronic processes in non - crystalline materials _
, + ( clarendon press , oxford , 1971 ) .
lifshits , s.a .
gredekul and l.a .
pastur , _ introduction to the theory of disordered systems _ , + ( john wiley and sons , 1988 ) . a different type of generalized bm model , using ( [ e3_2 ] ) with e.g. quadratic @xmath50 has been introduced by lindenberg et al .
( physica * a 126 * , 489 ( 1984 ) ) .
the latter type of generalization has been used by straus et .
( j. chem . phys . * 98 * , 4082 ( 1993 ) ) in order to study position dependent friction . | a new model that generalizes the study of quantum brownian motion ( bm ) is constructed .
we consider disordered environment that may be either static ( quenched ) , noisy or dynamical .
the zwanzig - caldeira - leggett bm - model constitutes formally a special case where the disorder auto - correlation length is taken to be infinite .
alternatively , localization problem is obtained if the noise auto - correlation time is taken to be infinite . also the general case of weak nonlinear coupling to thermal , possibly chaotic bath is handled by the same formalism .
a general , feynman - vernon type path - integral expression for the propagator is introduced .
wigner transformed version of this expression is utilized in order to facilitate comparison with the classical limit .
it is demonstrated that non - stochastic genuine quantal manifestations are associated with the new model .
it is clarified that such effects are absent in the standard bm model , either the disorder or the chaotic nature of the bath are essential .
quantal correction to the classical diffusive behavior is found even in the limit of high temperatures .
the suppression of interference due to dephasing is discussed , leading to the observation that due to the disorder the decay of coherence is exponential in time , and no longer depends on geometrical considerations .
fascinating non - markovian effects due to time - correlated ( colored ) noise are explored . for this , a new strategy is developed in order to handle the integration over paths .
this strategy is extended in order to demonstrate how localization comes out from the path integral expression . |
in many spiral galaxies the outer regions of the disk curve away from the symmetry plane of the inner disk , resembling an integral sign . this is called a _
warp_. the cause of this phenomenon is still uncertain , but a dark halo may play an important role .
sparke & casertano ( 1988 ) found some long - lived warping normal modes inside an oblate halo potential , but they used a halo that was unperturbed by the disk .
later , dubinski & kuijken ( 1995 ) and nelson & tremaine ( 1995 ) , following a comment by toomre ( 1983 ) , showed that when the response of the halo is taken into account the sparke & casertano modes are no longer long - lived .
cosmic infall of matter onto galaxies has been proposed as a possible explanation for the warp phenomenon ( ostriker & binney 1989 ) .
this infall likely causes the net angular momentum vectors of disk galaxies to reorient each hubble time . in this work ,
we simulate and analyze the response of a disk embedded in a halo whose orientation is changed slowly .
the properties of the warps generated this way are studied later , both in shape and amplitude .
the halo and the disk have been constructed following kuijken & dubinski ( 1995 ) , using a self - consistent approach .
the halo has a lowered evans distribution function , and the disk is modeled with concentric spinning rings .
the mass profile and rotation curves of the different models can be seen in figure [ 1 ] .
the main parameters of both models are summarized in table [ 1 ] .
the number of particles used for each model was of @xmath0 particles for the halo , and 400 rings , each of them consisting of 90 particles .
.model parameters : ( 1 ) disk mass , ( 2 ) disk scale length , ( 3 ) disk truncation radius , ( 4 ) disk scale height , ( 5 ) disk truncation length , ( 6 ) halo central potential , ( 7 ) halo velocity dispersion , ( 8) halo potential flattening , ( 9 ) halo concentration , @xmath1 , ( 10 ) characteristic halo radius . [ cols="^,^,^,^,^,^,^,^,^,^,^ " , ] the n - body code use to evolve the system uses a hybrid approach : the halo is made of particles , and its effect is computed with a self - consistent field code ( hernquist & ostriker , 1992 ) , expanding the potential in terms of some basic functions . the disk , however , is modeled using pivoted spinning rings .
the dynamics on the disk are computed calculating the potential and acceleration on the `` ring - particles '' , and then these are used to calculate the torque on each ring .
this is accomplished with a tree code ( barnes & hut , 1986 ) .
the first approach to the problem was to consider a rigidly rotating halo , unaffected by the gravitational effect of the disk .
the slewing rate of the halo was chosen to be h@xmath2 ( 100 km / s / mpc for a disk scale length of 3.5 kpc ) , so that the angular momentum of the halo re - orients in the predicted timescale.the axis of rotation of the halo lies in the initial halo - symmetry plane .
the results of the simulations for models * c * and * d * are shown in figure [ 2 ] . here
the inclination of the rings and the position angle of the line of nodes is plotted for different timesteps .
these plots show that the disk reorientation process is fast enough to follow the halo in the inner parts .
the disk warps where its self - gravitation is not enough to hold it in the same plane .
this is in perfect agreement with the briggs law ( briggs , 1990 ) : straight line of nodes in the center , leading spiral further out .
if the mass of the disk is doubled , the region aligned with the symmetry plane of the halo increases ( approximately one scale length ) , as expected .
as mentioned in the introduction , the assumption of a non - responding halo is very simplistic .
the next step is to simulate the re - orientation of a responsive halo that cosmic infall would cause . to do this
we will put a rotating _ tidal field _ through the galaxy .
this will make the symmetry plane of the outer halo change , gaining angular momentum this will be done by placing 2 _ satellites _ symmetrically , so that the monopole term of the tidal field is zero , avoiding relative movements of the galaxy with respect to the _ satellites_. this will also provide information about the influence of satellites on disks .
barnes , j. , hut , p. , 1986
, nature , 324 , 446 briggs , f.,1990 , , 352 , 15 dubinski , j. , kuijken , k. , 1995 , , 442 , 492 hernquist , l. , ostriker , j. p. , 1992
, , 376 , 375 kuijken , k. , dubinski , j. , 1995 , , 277 , 1341 nelson , r. w. , tremaine , s. , 1995 , , 275 , 897 ostriker , e. c. , binney , j. j. , 1989 , , 237 , 785 sparke , l. s. , casertano , s. , 1988 , , 234 , 837 toomre , a. , 1983 , iau symp .
100 , internal kinematics & dynamics of galaxies , ed .
e. athanassoula , pp .
dordrecht : reidel | secondary infall in galaxies could cause the angular momentum of the outer halo to change its orientation .
the generation of warps due to this effect is studied using n - body simulations . |
ultraviolet observations from space have shown that old stellar populations are prolific producers of hot stars .
elliptical galaxies and large spiral bulges exhibit an unexpected `` ultraviolet - upturn '' or `` excess '' ( uvx ) , a sharp rise in flux at wavelengths shorter than 2000 , which is produced by stars with temperatures of @xmath4 k ( code 1969 , burstein et al .
1988 , greggio & renzini 1990 , brown et al . 1997 , oconnell 1999 , and references therein ) .
the uv upturn is the most variable photometric feature of old populations : uv to optical flux ratios vary by a factor of 100 , a much larger range than for optical - ir colors ( faber 1983 ; burstein et al . 1988 ; dorman , oconnell , & rood 1995 , hereafter dor95 ) .
the first large survey of the uvx in nearby galaxies , based on observations with the _ international ultraviolet explorer _ ( iue ) , appeared to show a good correlation between its strength and metal abundance as measured by the mg@xmath1 index at 5170 . a more extensive recent survey from the _ galaxy evolution explorer _
( galex , martin et al .
2003 ) by rich et al .
( 2005 ) confirms the enormous range in uv : optical flux ratios but finds no correlation between uv properties and metallicity .
this remarkable variation in uv light has led to the expectation that the uvx could be a uniquely sensitive probe of the star formation and chemical enrichment histories of elliptical galaxies .
if it can be calibrated , the uv upturn would be an important test , for instance , of the formation times of distant elliptical galaxies , since preliminary models @xcite imply that it will only appear in systems 5 gyr or older .
the _ hubble space telescope _ ( hst ) and galex have already detected a uvx at redshifts of @xmath5 @xcite and @xmath6 @xcite , respectively .
we believe we know the _ kinds _ of stars responsible for the uvx ( reviewed in oconnell 1999 ) . uv imaging and spectroscopy from the _ astro _ missions provided good evidence that old , low - mass stars on the `` extreme '' horizontal branch ( ehb ) , with @xmath7 k , and their hot descendents produce the upturns .
we also understand the basic physics of the hot evolutionary phases .
the ehb stars have very small envelope masses ( @xmath8 ) , and smaller envelopes produce hotter objects .
differences of as little as 0.01 m@xmath9 can have large effects .
hence , the net uv output of an ehb population is governed mainly by envelope mass loss processes on the red giant branch and by the helium abundance . in turn ,
mass loss efficiency is strongly linked to age ( older stars have smaller envelopes to lose ) and metal abundance ( mass loss is likely enhanced at higher metallicites by radiation pressure on grains and molecules ) .
less than 20% of the evolving population in the galaxies need to pass through the hot phase to match the observations .
what we do not yet understand is how the parent population s global characteristics ( age , helium abundance , metal abundances , dynamics ) determine the distribution of ehb temperatures and hence the strength and shape of the upturns .
observations of globular clusters ( gcs ) undoubtedly offer the best means of clarifying these global effects .
gcs are the best understood stellar systems from the standpoint of evolution , and they are ( mostly ) simple stellar populations with a small internal dispersion in age and abundance .
unfortunately , the available uv data on gcs are limited and hard to interpret , and the galactic gc sample does not overlap the galaxies in metallicity as measured by mg@xmath1 ( summarized by dor95 ) .
some globulars have total uv - star fractions ( as indicated by their @xmath10 colors ) that rival those of any elliptical galaxy .
however , the uv fractions are largest for moderately metal - poor clusters ( with [ fe / h ] @xmath11 ) , such as @xmath12 cen @xcite and ngc 6752 @xcite
. there is large scatter at all metallicities and no strong trend with metallicity ( although the uv sample is far from complete ) .
meanwhile , the recent galex observations by rey et al .
( 2005 ) show that the far - uv properties of m31 gcs are similar to those of mwgcs .
by contrast , the uv fractions among nearby galaxies in the iue sample appear to increase strongly with metallicity .
such differences in behavior could be linked to differences in age or helium abundance or to other factors influencing the ehb , such as the dynamical environment ( e.g. paresce et al .
1991 , fusi - pecci et al.1993 , bailyn 1995 , sweigart 1997 ) .
there is a critical gap in metallicity between the most metal - rich of the observed milky way globulars and the bulk of elliptical galaxies .
this corresponds to mg@xmath1 values in the range @xmath13 0.10.3 ( see figure 1 of dor95 or figure [ f : fuvv_mg2 ] below ) .
an obvious next step is to fill in this gap with clusters of higher metallicity .
such objects are rare in our galaxy and also difficult to study because they tend to be preferentially situated in regions with large uv extinction .
where data have been obtained on metal - rich galactic globulars , there is a large dispersion : 47 tuc has a definite but small ehb population @xcite while ngc 6388 and 6441 have rich ehb populations @xcite .
a small ehb component is present in the high metallicity open clusters ngc 188 and ngc 6791 @xcite .
but overall the milky way is not a fertile terrain for investigating the uv behavior of metal - rich clusters . to enlarge the sample of metal rich objects that can link globular cluster populations to those in galaxies
, we must instead turn to the large cluster systems of giant elliptical galaxies .
giant elliptical galaxies possess enormous populations of gcs , many of which have high metal abundances .
hundreds of such globulars in elliptical galaxies in the virgo cluster have @xmath14 .
these systems are essentially free of foreground extinction , with @xmath15 . among the virgo galaxies
, m87 has the largest and best studied gc system .
numerous , faint point sources around m87 were first identified as globular clusters by @xcite .
early photographic studies showed that the gcs in m87 are on average bluer than the galaxy background @xcite and that the mean color of the clusters has a radial trend such that clusters become bluer as radial distance increases @xcite .
the radial dependence of mean color was later confirmed by the ccd observations of @xcite and @xcite .
kundu et al .
( 1999 ; hereafter k99 ) studied the gc system in the inner region of m87 using images taken with the hst wfpc2 camera .
their sample , over a thousand clusters , allowed them to investigate the luminosity function , the optical color distribution , and structural information .
an important result is the presence of bimodality in the optical color distribution .
bimodality of gc systems in early - type galaxies was first ( statistically ) studied by @xcite .
recent studies of hst archival data ( kundu & whitmore 2001a , larsen et al .
2001 ) have shown that bimodal color distributions are present in at least 30% and up to 60% of early - type gc systems .
several models have been proposed to explain the observed bimodal cluster groups : major mergers of late - type galaxies @xcite , two bursts of in situ star formation @xcite , and hierarchical formation @xcite .
the goal of this study is to determine how the m87 metal - rich clusters compare to the galactic globulars and the elliptical galaxies in their uv color properties . among the observational questions of particular interest are these : ( 1 ) does the m87 cluster system overlap in uv properties with the galactic system or is it distinct ? ( 2 ) does the m87 cluster system overlap with the e galaxies ?
( 3 ) does the m87 system represent a transition between milky way - type clusters and the galaxies ? ( 4 ) are secondary dependencies found in uv properties , e.g. a dependence on galactocentric radius at a given metallicity ? ( 5 ) do the uv properties of the two m87 cluster subsystems differ ?
we have obtained near- and far - uv photometry of m87 globular clusters using the imaging mode of the hst space telescope imaging spectrograph ( stis ) to answer these questions . in the next section
we describe our observations and reductions and our far - uv detection rate .
subsequent sections discuss interpretation of the resulting photometry , comparison to the globular clusters in the milky way and m31 , the absence of radial trends in our photometry , comparison to elliptical galaxies , comparison to existing model grids for clusters , possible selection effects , and the surprisingly large sample of `` uv - only '' sources we detect .
the data used in this study consist of hst / stis far- and near - uv images of four different fields in the inner ( @xmath16 ) region of m87 . at our adopted distance modulus of @xmath17 @xcite for m87 , this region corresponds to @xmath18 kpc .
figure [ f : fuvpointings ] shows the location of the far- and near - uv fields superposed on the median - subtracted wfpc2 image of k99 .
the fields are labeled from 1 to 4 , and we use these designations throughout this paper .
fields 1 , 2 , and 3 were observed with the hst / stis far - uv camera in hst program go-8643 ( pi : r. oconnell ) .
this camera has a field of view of 25@xmath1925with 0.024pixels .
total exposure time was 10,480 s for each field .
the far - uv camera employs a multianode micro - channel array ( mama ) which has no readout noise and little sensitivity to cosmic rays . since the long wavelength cutoff is determined by the rapidly decreasing detector sensitivity longward of 2000 , red leaks are negligible .
we used the f25srf2 longpass filter ( 1270 2000 , pivot wavelength 1460 ) to extend the bandwidth of our images as far to the blue as possible while avoiding the strong geocoronal lyman-@xmath20 emission at 1216 .
the fuv - mama / f25srf2 combination admits the dayglow emission lines of ( 1302 , 1356 ) , which create a relatively bright sky background .
however , the features are strongly variable in both time and geographic latitude , and their strength can not be confidently predicted in advance .
therefore , instead of restricting our observing periods to orbital night by using the shadow time special requirement , we observed in timetag mode . in this observing mode ,
the data consist of a stream of time - tagged photon events and x / y coordinates .
periods of bright dayglow emission can be rejected during the subsequent data reduction , thereby making best use of available low dayglow periods .
we also obtained three optical band images centered on the same positions as the far - uv frames of fields 1 , 2 , and 3 , using the stis ccd and f28x50lp ( 5500 11000 , pivot wavelength 7230 ) filter .
these images were obtained mainly to verify pointings of our targets .
each ccd image has a field of view of 50@xmath1950 of which only the central 28@xmath1950 is usable due to the size of the filter .
all three far - uv fields fall on the usable area of the ccd .
the exposure time for the ccd observations was 518 seconds for each field .
observations of field 4 were taken from the stsci / hst archive .
these included a total of four far - uv and sixteen near - uv images .
the relevant observations are those of hst programs go-8048 , -8140 , -8780 , and -9461 ( pi : j. biretta ) . all downloaded images were observed in accum mode ( normal integration mode ) either using fuv - mama / f25srf2 in the far - uv or nuv - mama / f25qtz ( 1480 3500 , pivot wavelength 2360 ) in the near - uv .
each observation has four near - uv exposures with integration times of @xmath21 seconds .
far - uv images were available only in the go-8140 observation , whose field coverage is shown in figure [ f : fuvpointings ] .
four fuv - mama exposures with integration times ranging from 850 to 910 seconds were taken .
the orientations of go-8048 , -8780 , and -9461 observations slightly differ from one another , but the field coverage is not much different from that of field 4 in figure [ f : fuvpointings ] .
we denote these three fields respectively as field 4a , 4b , and 4c when we need to distinguish them .
all of these observations were taken as part of the ongoing investigation of the variability of the m87 nonthermal jet ( e.g.perlman et al .
field 4 is subject to the bright diffuse background that increases rapidly toward the m87 nucleus .
a summary of all observational material is given in table [ t : obslog ] .
the stis ccd images for fields 1 , 2 , and 3 were calibrated using the standard iraf.calstis pipeline .
each ccd image went through the steps of data quality initialization , overscanning , bias subtraction , cosmic ray rejection , dark subtraction , flat - fielding , and geometric correction .
the final calibrated ccd images still show a considerable number of remaining hot pixels not removed from the dark subtraction stage .
we used the imgclean procedure in idl . to eliminate these hot pixels .
imgclean was originally designed to remove cosmic ray hits from wfpc images , but extensive tests showed that it works very well on our stis ccd images .
the ccd images are only used for astrometry of sources in this study , so any photometric bias introduced by the imgclean routine has no impact . because the far - uv images of field 1 , 2 , and 3 were obtained in the timetag mode , our first step was to convert the timetag tables into 2-d images . for each data set we plotted the total counts in each five second bin as a function of time .
since the total count is dominated by the sky background , higher total count rates indicate higher background counts . to reject periods of bright skyglow , we can simply reject periods of high total count rates .
we note that exposures _
o6be02ceq _ of field 1 and _ o6be04z4q _ of field 2 were made entirely in dark background periods .
the time periods are input to the inttag task in the iraf / stsdas package to produce 2-d images . in practice , it was straightforward to identify intervals with serious skyglow .
once the conversion was done for all of our exposures , we proceeded with the standard stis imaging calibration procedures as described below .
the dark current in the stis far - uv camera is known to have a low but spatially and temporarily variable enhancement in the upper left quadrant of the image , which is called the _ glow_.
it is known that the glow correlates well with the detector temperature . while the temperature of the detector itself is hard to measure , the charge - amplifier temperature ( cat ) , which is recorded in the image headers , is a reliable proxy for the detector temperature .
the glow is negligible for cat lower than 35 k but above that temperature the global count rates for dark frames increase with increasing cat .
most of our far - uv frames have cat higher than 35 k so we performed subtraction of the glow pattern in the following manner .
first , a constant dark component was subtracted by the calstis routine .
next , to deal with the glow component on the upper left - hand quadrant of some images , we used the `` glow only '' image as a template which is available online . while the glow profiles of our far - uv frames are well represented by this image ,
the intensity of the glow varies image by image .
we have generated projected mean profiles along the x and y axes of those images and used them as references .
then we scaled and subtracted the glow profile from the images affected by the glow and generated profiles .
this step was iterated until the subtracted profile closely matched the reference profile .
each glow - subtracted image was visually inspected for any anomaly caused by this technique .
it turned out that this method works reasonably well except for field 3 , which has a strong stellar background gradient due to its proximity to the m87 nucleus .
since the background dominates for this field , we did not attempt to subtract the dark glow . finally , we flagged all of the hot pixels present in our science images .
the four individual images of fields 1 and 2 have slight positional offsets . by inspecting the locations of the bright sources in these images , we derived and applied the integer shifts in x and y directions for registration .
( note that a one pixel shift is a small fraction of the 70% encircled energy radius for the point spread function , which is 8 pixels . )
the images were then co - added . for field 3
, there was an interval of 4 months between the two visits and the orientations differ by approximately 170 degrees .
we performed a rotation on the earlier epoch images using the rot function in idl with the cubic parameter set to @xmath22 .
the cubic convolution interpolation method used in this function closely approximates the optimum sinc interpolation function by adopting cubic polynomials . to ensure that introducing rotation does not affect the object photometry ,
we have compared the photometry ( explained below ) done on several sources in both the pre- and post - rotated images .
the differences in all cases were negligible . after rotating the images ,
integer shifts were applied using the same approach as in fields 1 and 2 .
the mama images of field 4 were also calibrated with the standard calstis pipeline .
we did not attempt to subtract the dark glow of the field 4 far - uv images because of the high galaxy diffuse background .
the registration and co - adding for each observation of field 4 were done as described above , resulting in one combined fuv - mama image and four combined nuv - mama images . in figure 2 , we show the combined stis fuv - mama and stis ccd images for field 1 , 2 , and 3 . images and identifications in field 4
are shown and discussed in 9.2 .
we derived astrometric solutions for all of our stis images ( ccd , fuv - mama , and nuv - mama ) using the wfpc2 image as a reference to aid the identification of sources .
the iraf task tfinder was used for doing the astrometry .
the offsets between new solutions and old header records are less than 1 in most cases for all stis images , and the uncertainties for the solutions are less than 0.02 .
sources in the stis ccd ( optical - band ) frames of fields 1 , 2 , and 3 were identified using the automated daofind routine ( stetson 1987 ) .
since the k99 observations go deeper than ours , all sources found in our ccd images are also listed in the photometry of k99 .
for the far - uv and near - uv images , we did not use an automated detection algorithm since the sources have poorly defined point spread functions ( psfs ) .
instead , we centroided on each object detected by eye and cross - identified them with objects listed in k99 photometry
. when detected far-/near - uv sources did not match any of the objects in k99 photometry , we labeled them as `` uv - only sources . ''
similarly , sources not found in our far-/near - uv images but listed in k99 photometry within our field of view were labeled as `` optical - only sources . ''
the nuv - mama images of field 4 covered four different epochs separated from five months to as long as two years .
we tabulate identifications separately for each epoch in table [ t : catalog3 ] to study possible variable sources .
we note that in field 4 , all uv - only sources in the far - uv frame are also found in one or more of the near - uv frames .
aperture photometry on the uv sources was performed using a modified version of the idl aper routine , an implementation of the daophot @xcite photometry routine .
the dark current is so low in the mama detectors that the background levels are quantized to just a few values .
the median or mode is not a good estimation of the background when dealing with such signals . according to @xcite , the best way to estimate the background is to identify all hot pixels in the local background of each source , and then to use a _ mean _ for the remaining pixels .
our modified aper routine fully accommodated this approach .
the net flux of each source was measured through an aperture of 8 pixels in radius , corresponding to 0.20 arcsec and using a local sky background measured within an annulus of 20 to 30 pixels in radii . in some cases , the sources are so close to each other that the background measurements are affected by the nearby sources . for those
, we masked the nearby sources with apertures of appropriate size prior to doing photometry .
the conversion of mama counts into magnitudes in the stmag system was done using the expression : @xmath23 where photflam defines the inverse sensitivity in units of ergs s@xmath24 @xmath25 @xmath24/(counts s@xmath24 ) and exptime is the exposure time in seconds . the version of the calstis routine we used ( v2.13 , 26-april-2003 ) does not take into account the time dependent sensitivity changes for mama imaging modes .
the photflam values in the image headers all default to @xmath26 ergs s@xmath24 @xmath25 @xmath24/(counts s@xmath24 ) for the fuv - mama and @xmath27 ergs s@xmath24 @xmath25 @xmath24/(counts s@xmath24 ) for the nuv - mama , regardless of the date of observation . to calculate the correct photflam values for our observing periods
, we used the bandpar task in iraf stsdas.synphot package with the observation julian dates as inputs .
the final integration times and photflam values used are tabulated in table [ t : photflam ] . for uv - only and optical - only sources ,
we determined upper limits in @xmath28 and far-/near - uv fluxes using the nominal flux plus @xmath29 when the measured fluxes are positive and @xmath29 when the measured fluxes are negative .
in the case of the uv - only sources , we re - measured the original k99 image frames at the coordinates of the uv identifications , applying the k99 photometric procedures .
we attempted to construct far-/near - uv growth curves for a few of our bright sources to derive aperture corrections .
however , because even the brightest sources have a poorly defined psf , it is not possible to derive reliable aperture corrections .
therefore , we rely on the aperture corrections for point sources tabulated in table 8 of @xcite .
the foreground reddening towards m87 estimated from diffuse 100@xmath30 dust emission is @xmath31 @xcite . using this value and the extinction law of @xcite , we derived extinctions of @xmath32 mag and @xmath33 mag , implying @xmath34 and @xmath35 .
all photometry was corrected for extinction using these values .
foreground @xmath36 estimated from hi emission is zero ( burstein & heiles 1984 ) , but even if we had adopted the latter value , the large color offsets discussed below between the m87 and milky way clusters would still persist .
our photometric catalogs are presented in tables [ t : catalog1 ] and [ t : catalog2 ] . for each object
we list the following parameters : identification number , coordinate offsets from the center of m87 , radial distance from the center , far- and near - uv magnitudes and errors in the stmag system , and dereddened uv - optical colors .
details are given in the footnotes to the tables .
we use the wfpc2 @xmath28 photometry of k99 to examine the detection rate of clusters in our fields .
our limiting far - uv magnitude is @xmath37 , corresponding to @xmath38 .
we detect in the fuv a total of 69 objects in fields 1 , 2 , and 3 , of which 50 are globular clusters identified by k99 .
there are 127 optically - identified clusters in the k99 sample in these fields to @xmath39 .
( we discuss the non - k99 identifications later in 9 ) . in field 4 , 7 and
37 objects are detected in the far- and near - uv images , respectively . in figure
[ f : fuvdetection ] , we show our far - uv detections with respect to optical detections in fields 1 , 2 , and 3 .
( field 4 has been excluded from this plot since the short exposure time and the strong background gradient make the detection rate low and not comparable to the other fields . ) for @xmath40 , all but two k99 clusters are detected in the far - uv , resulting in a detection rate of 94% with respect to optical detections .
the mean @xmath41 color for these 29 clusters is @xmath42 .
k99 show that the completeness of their cluster selection for @xmath40 is nearly 100% .
for @xmath43 and 24.0 , our detection rate drops to 63% and 48% , respectively .
overall , we detect 39% of the k99 cluster sample .
we return to this subject and consider possible selection effects in 8 .
as seen in figure [ f : fuvdetection ] , the clusters detected in the far - uv are at the bright end of the optical luminosity function for the m87 gc system ( also see figure 5 of k99 ) .
the two peaks in the optical color distribution found for m87 clusters by k99 lie at @xmath44 and @xmath45 .
the histogram in fig .
3 shows that most of the clusters we detect in the far - uv belong to the bluer [ @xmath46 subpopulation .
in the subsequent sections , we discuss our results for the m87 system in the context of the available uv photometry samples of milky way globular clusters and early - type galaxies .
most of the existing data are from the _ international ultraviolet explorer _ ( iue ) , but several other experiments also contributed data ( see appendix a and dor95 ) .
the recently - launched galex mission will rapidly enlarge these samples .
the far - uv band is sensitive only to hot stars . in purely old populations ,
these exist in significant numbers only on the horizontal branch .
the far - uv flux therefore depends mainly on the number of hb stars hotter than about 10000 k , with a modest dependence on their temperature distribution . on the other hand ,
such stars contribute almost nothing to the optical bands ( here @xmath28 and @xmath47 ) , which are instead dominated by main sequence and giant light ( dor95 , schiavon et al .
thus , we take the @xmath48 color to be a measure of the _ fraction _ of the stellar population that appears in the form of hot hb stars .
the situation in the near - uv ( @xmath49 ) band is more complicated ( burstein et al .
1988 ) . in the absence of hb light , the near - uv band will be dominated ( over 90% contribution ) by stars near the main sequence turnoff . for a given age , there will be a strong dependence of color on metallicity , estimated by dorman , oconnell , & rood ( 2003 ) to be @xmath50 mags / dex .
but hot hb stars do often have a major impact in the nuv .
dorman et al . ( 2003 ) estimated that hot hb stars contribute 4080% of the nuv light in a sample of nearby ge galaxies , and comparable numbers should also apply to those globular clusters with bluer @xmath48 colors .
the @xmath51 color for populations with comparable ages and metallicities is in principle a measure of the temperature of the hot hb stars .
the simple models of dor95 indicate that the fuv light of objects with @xmath52 is dominated by hb populations that extend to the ehb range ( @xmath53 k ) .
figure [ f : colormag ] shows the extinction - corrected @xmath54 and @xmath55 color - absolute magnitude diagrams ( cmds ) for m87 clusters . we do not include the fuv limits for the optical - only sources . for comparison
, we also plot the milky way gcs as _
open triangles_. the milky way data were taken from dor95 but updated with the new @xmath36 values listed in @xcite ( see appendix for details ) .
the figure shows that we have been able to sample in m87 almost the full v - band luminosity range present in the milky way uv - detected sample , although the average m87 detection is more luminous than in the milky way .
there are three m87 clusters that are brighter than the most massive milky way gc , @xmath12 cen , in @xmath56 and two in @xmath57 .
( while @xmath12 cen and m54 have similar brightnesses in @xmath28 , @xmath12 cen is brighter in the far - uv by 1.2 mags . )
the m87 clusters in figure 4 extend to a significantly bluer @xmath2 color than the milky way gcs .
figure [ f : colorhist2_15v ] shows the @xmath58 color distributions for all m87 clusters plotted in figure [ f : colormag ] ( _ upper panel _ ) and for the clusters brighter than @xmath59 ( _ lower panel _ ) .
we also plot the histogram of the color distribution for milky way gcs . the mean color of the m87 clusters in figure [ f : colorhist2_15v]a is @xmath60 ( the error being the standard error of mean ) ; and for clusters in figure [ f : colorhist2_15v]b , the mean color is @xmath61 . both
m87 histograms have peaks at @xmath62 ( the error being half of the bin size ) .
the mean and peak for the milky way gcs are at @xmath63 and @xmath64 , respectively .
thus , the m87 gc sample we observed is significantly bluer than the mw sample .
the two - sided kolmogorov - smirnov ( k - s ) test shows that the probability of the milky way and m87 gc color distributions being drawn from the same parent distribution is only @xmath65 . only a few mw clusters , including @xmath12 cen and ngc 6752 ,
have @xmath58 colors comparable to the peak of the m87 distribution . to compare our m87 results to those for m31 clusters from rey et al.(2005 ) , we must first convert the galex photometric system ( defined in terms of flux per unit frequency ) to our stmag system ( defined in terms of flux per unit wavelength ) .
the conversion is @xmath66 .
the detected m31 gcs lie in the range @xmath672.2 , within the blue envelope of the milky way clusters and not as extreme as the m87 sample . in figure
[ f : colorcolor ] , we show the @xmath68 color - color diagram for all data plotted in figure [ f : colormag ] .
the @xmath69 colors corresponding to the centers of the two cluster subpopulations as defined by k99 are indicated by arrows .
most of our detections belong to the bluer group , although they tend to lie on the red side of the color centroid .
they also correspond in @xmath70 color to the average of the milky way sample .
seventeen of the uv - detected m87 clusters have @xmath71 .
these evidently belong to the red subpopulation .
they are redder in @xmath70 , and presumably more metal - rich , than any milky way sample cluster . on average , these objects have redder @xmath72 colors than the other m87 gcs , but they are considerably bluer than metal - rich milky way clusters like 47 tuc and ngc 6388 . because of the selection effects ( see 8) evident in figure [ f : fuvdetection ] , the true mean color of the metal - rich m87 clusters must be redder than suggested by the figure .
however , it is clear that m87 clusters with metallicities considerably in excess of those in the milky way are capable of producing large ehb star populations .
the m87 clusters show significant scatter in figure [ f : colorcolor ] but form a rough diagonal sequence along which clusters bluer in @xmath69 color are also bluer in @xmath2 .
the slope of this sequence is measured to be @xmath73 . since @xmath69 colors are widely used as a proxy for metallicity ( assuming a modest age dispersion ) , this trend in turn suggests that more metal - poor gcs in our m87 sample produce larger fractional hot hb populations .
the milky way sample is too small and inhomogeneous to establish such a metallicity - uv dependence , but it not inconsistent with the m87 trend . except for the two bluest milky way clusters in @xmath2 , @xmath12 cen and ngc 6752 ,
_ there is no overlap _ between the m87 and milky way clusters in the color - color diagram . at a given @xmath69 ,
the m87 gcs are significantly bluer in @xmath2 than the milky way clusters .
combining both m87 and milky way data , there is a hint in figure [ f : colorcolor ] of three parallel sequences in the diagram with offsets in @xmath10 : m87 gcs with @xmath74 , milky way gcs ( unlabeled ) plus m87 gcs with @xmath75 , and a sequence of uv - faint milky way gcs ( labeled ) .
the latter includes 47 tuc .
however , the inhomogeneous photometry and limited milky way sample preclude any definite conclusion , nor is there an obvious physical explanation for such a separation .
if we consider the @xmath69 color as a proxy for metallicity again , the pronounced scatter in @xmath2 colors at a given @xmath69 in figure [ f : colorcolor ] is an extreme manifestation of the `` second parameter effect '' ( see lee , demarque , & zinn 1994 and references therein ) .
the main conclusion from this section is that despite strong overlap in v- and i - band properties , the m87 gcs have uv / optical properties that are distinct from clusters in the milky way and in m31 .
m87 clusters produce larger hot hb populations than do their milky way analogues , and metal poor gcs are better at this than are metal rich gcs .
because we have detected a large fraction of all cluster candidates in our fields ( see 8) , we do not believe this result is strongly biased by preferential selection of an unrepresentative extreme uv - bright tail of the m87 cluster population .
m87 metal rich clusters can produce significant hot hb populations .
figure [ f : radial_mag_color ] tests for a variation of the absolute far - uv magnitude , @xmath76 , and the @xmath77 colors with radial distance @xmath78 from m87 nucleus . since field 4 has a different exposure depth than the other fields , we separated this field with dashed lines in both plots .
studies of radial gradients in the optical color distribution of gcs in m87 by @xcite , @xcite , @xcite , and @xcite show that on average , gc optical colors become bluer as radial distance increases .
@xcite found that the color distribution is flat for @xmath79 ( which encompasses 3 of our 4 fields ) and then becomes bluer with radial distance .
@xcite studied the integrated far - uv color gradient in m87 , which is dominated by the diffuse background , and found that the @xmath80 color is flat out to a radial distance of @xmath81 . beyond that , the color becomes redder with increasing radial distance . in our case , we find no evidence for radial trends in the colors of individual gcs in figure [ f : radial_mag_color ] , although our radial coverage is not large .
here we present population diagnostic diagrams similar to those in dor95 , including data for both the m87 and milky way globular cluster samples , for two galactic open clusters , and for early - type galaxies . to place the m87 clusters in the context of other old populations with uv data , we employ the lick mg@xmath82 narrow - band absorption line index as a metallicity indicator .
the mg@xmath82 index is widely used in population studies , and it is available for many milky way gcs and nearby early - type galaxies .
uv color - mg@xmath82 diagrams were presented for early type galaxies in figure 1 of @xcite and for galaxies and galactic globular clusters in figure 1 of dor95 .
unfortunately , neither mg@xmath82 nor any other spectroscopic measures are available for our m87 clusters . to place them in the diagram , we have estimated mg@xmath82 values from their @xmath41 colors using the color-[fe / h]-mg@xmath82 relations derived in appendices a and b. these are based on color and spectroscopic data for gcs in the milky way , the ge galaxy ngc 1399 , and a small sample of gcs on the outskirts of m87 ( see the appendices for details ) . in appendix b we show that the three cluster systems lie on a common @xmath41-[fe / h ] relation . the inferred mg@xmath82 values for the m87 clusters must obviously be treated with caution .
however , the color range in this calibration encompasses all of our uv - detected sample , and our basic results do not depend on the details of the conversion unless there are strong nonlinearities not evident in the data sets discussed in the appendices .
an alternative that would avoid whatever uncertainty resides in this conversion would have been to use @xmath41 colors , which are known to correlate well with metallicity , instead of mg@xmath82 .
unfortunately , most of the early type galaxies with existing far - uv observations do not presently have published @xmath41 colors .
another potential difficulty for interpretation is that it is now well established that the abundances of the light metals in old populations , including mg , do not change in lockstep with the iron peak ( e.g. worthey 1998 ; thomas , maraston , & bender 2003 ; schiavon 2005 ; and references therein ) .
trager et al .
( 2000 ) estimate that the light metals are enhanced by @xmath83 dex with respect to the fe - peak in typical luminous elliptical galaxies . however , it is the light metals , especially o , that dominate the overall metal abundance .
they therefore have a dominant influence on the evolution of galaxy spectral energy distributions , although the fe - peak is also important through its effects on opacity .
mg is probably a better indicator of metallicity than fe .
the relatively small enhancements found by trager et al .
( 2000 ) should not distort the metallicity scale much with respect to one based on the fe - peak .
in particular , they should not importantly affect the ranking of objects by metallicity , which is the main issue in the following discussion . in figure
[ f : fuvv_mg2 ] , we have added the m87 clusters to the far - uv color - mg@xmath82 diagram for milky way gcs and nearby early - type galaxies presented by dor95 .
the galaxy measurements refer to the central @xmath84 region covered by the iue aperture .
the lowest inferred mg@xmath82 indices for the m87 clusters shown in the figure coincide with the lowest values for milky way clusters .
but the m87 sample extends to considerably higher values than does the milky way sample .
the most metal - rich m87 gc is inferred to have solar metallicity . as in figure
[ f : colorcolor ] , the m87 sample is significantly bluer in @xmath2 at a given mg@xmath82 than the milky way clusters .
the metal - rich m87 clusters have filled the gap between the milky way cluster and galaxy sequences in the diagram , as we had hoped in planning the observations .
however , the clusters are still clearly separated from the galaxies in this diagram . many of the m87 gcs are @xmath85 mag bluer than the bluest ellipticals in @xmath2 .
about ten metal - rich m87 clusters overlap with the lower metallicity end of the galaxy sequence .
( the s0 galaxy m85 shows evidence for star formation and internal extinction [ kinney et al .
1993 ] , so we exclude this object from further discussion . ) the lowest metallicity galaxy in the sequence is the local group dwarf m32 .
many studies have indicated that m32 has approximately solar abundance ( worthey 2004 and references therein ) .
our uv - detected metal - rich m87 gc sample is @xmath863 mags bluer than m32 .
in fact , it has an average @xmath72 color bluer than 90% of the entire galaxy sample .
because of selection effects , we can not compare m32 to a complete sample of m87 clusters with comparable metallicity .
however , an age difference is likely to be a contributing factor to the large color discrepancy between the available metal - rich sample and m32 .
the luminosity - weighted age of m32 is known to be @xmath1335 gyr @xcite , whereas the clusters are probably 812 gyr old .
younger systems are expected to have a smaller uvx ( other things being equal ) because the envelope masses of their hb stars are larger , yielding lower hb temperatures ( e.g. yi et al . 1999 , oconnell 1999 and references therein )
. rose ( 1994 ) and rose & deng ( 1999 ) had already demonstrated significant population differences , probably involving age , between m32 and the metal rich milky way cluster 47 tuc .
they made a careful analysis of spectral features in the 2600 - 4400 region , which originate mainly on the main sequence and giant branches .
the differences on the hot hb revealed in the far - uv by figure [ f : fuvv_mg2 ] are much more conspicuous . using the simple formulation for treating the uv spectra of old populations described in dorman et al .
( 2003 ) , we find that populations like the bluest of the m87 metal rich clusters could contribute at most about 6% of the v - band light in m32 . as mentioned in the introduction , the recent galex survey of quiescent early - type galaxies ( rich et al .
2005 ) exhibits a large range of uv / optical color but no correlation of color with metallicity , unlike the iue sample shown in figure [ f : fuvv_mg2 ] .
the differences between the two samples are not understood and are still being explored . however , after taking account of the differences in definition between the galex and stmag systems ( 4 ) , we find that roughly 40% of the galex sample is bluer than ngc 1399 , the bluest elliptical in the nearby iue sample .
the bluest galex objects correspond to @xmath87 , which is 2 magnitudes bluer than ngc 1399 .
this situation is not unphysical for an old population because dor95 showed that if all horizontal branch stars evolved through the hottest ehb channel , such colors could result .
however , it does signify rather different conditions in the galex sample than in nearby systems .
whatever these are ( e.g. contamination by recent star formation , environmental effects leading to larger ages ) , they may have concealed the trends with metallicity found by burstein et al.(1988 ) . the key result from figure [ f : fuvv_mg2 ]
is that the metal rich subset of m87 globular clusters does indeed overlap the mg@xmath1 values found on the elliptical galaxy sequence but the uv - detected clusters produce much more far - uv light than do the galaxies at a given mg@xmath1 .
the metal - rich clusters of m87 do not appear to represent a transition between mw - type clusters and e galaxies .
this comparison is tentative because of incompleteness in both samples and the uncertainty in the inferred mg@xmath1 values for the m87 clusters .
however , the effect is in the correct sense if the detected clusters have significantly older light - weighted ages than the galaxies .
we have marked in figure [ f : fuvv_mg2 ] the approximate locations of two well - known galactic open clusters , ngc 188 and ngc 6791 .
these have estimated ages of 6 and 7 gyr , respectively ( dinescu et al .
1995 ; kaluzny & rucinski 1995 ) .
integrated far - uv fluxes were based on ultraviolet imaging telescope observations made by landsman et al.(1998 ) .
v - band integrated fluxes were taken from lata et al.(2002 ) . to estimate mg@xmath1
, we used the synthesis predictions for lick indices by thomas et al .
we assumed the ages quoted above , [ @xmath20/fe ] @xmath88 , and adopted [ fe / h ] values from friel et al .
( we did not use the empirical calibration for globular clusters discussed in the appendices because these objects are significantly younger than the globulars . ) the positions of these clusters are rough because of lack of completeness , uncertainties in the mg@xmath1 conversion , and stochastic effects in the ( small ) number of hot stars responsible for the uv light in these clusters .
both clusters have appreciable far - uv contributions , and , as anticipated by landsman et al .
( 1998 ) , their @xmath72 colors are comparable to the bluest elliptical galaxies .
they fall together with the metal - rich m87 clusters in figure [ f : fuvv_mg2 ] and have much bluer colors than elliptical galaxies with similar mg@xmath1 values .
ngc 6791 is as blue as the bluest milky way globulars .
theoretical models for the uv upturn ( e.g. yi et al .
1999 ) predict that its strength is a step function in age , with populations older than @xmath89 gyr having a significant upturn ( i.e. a rich ebh ) .
therefore , even if the locations of ngc 188 and 6791 in figure [ f : fuvv_mg2 ] are accurate , we can not conclude that the surrounding m87 metal rich clusters have intermediate ages like those of the open clusters .
the open cluster observations can also be taken as evidence that the transition age is younger than 6 - 7 gyr .
if so , figure [ f : fuvv_mg2 ] indicates that the light of most of the galaxies is dominated by populations younger than about 6 gyr .
the far - uv in these two open clusters is produced by extreme hb stars , just as we believe is the case in the elliptical galaxies .
blue straggler stars ( bss ) are not important in the far - uv in either case ( landsman et al . 1998 ) . however , under certain circumstances , bss can dominate the uv light .
the open cluster m67 is the best example of this .
landsman et al .
( 1998 ) showed that a single , unusually hot bss ( f81 , with @xmath90 k ) contributes over 90% of the far - uv flux in m67 .
the predicted integrated @xmath72 is @xmath91 , comparable to that of ngc 6791 .
however , the more typical bss stars , with @xmath92 k , make very little contributions to the light below 1600 ( e.g. see fig . 3 of landsman et al
. 1998 ) . therefore , it is possible for bss - rich populations to match the far - uv colors of the metal - rich m87 globulars or the bluer ge galaxies .
but we do not think bss components are important in those systems for several reasons ( see landsman et al .
first , massive bss capable of affecting the integrated far - uv light will appear only for a narrow range of population age , near that of m67 ( 4 gyr ) .
they are not important in clusters like ngc 188 and 6791 , only a few gyr older , because they become too cool . in younger populations ,
the uv influence of bss is overwhelmed by the light of main sequence turnoff stars .
second , their importance in m67 has probably been enhanced by dynamical evaporation of lower mass stars , which is not a factor in galaxies and is probably less important in massive globular clusters .
finally , it is thought that many bss are binaries produced by stellar collisions ( bailyn 1995 ) , but the dynamical environment of galaxies is much less favorable to this process than is the case in clusters . in the near - uv ( 20003200 )
the influence of bss can be considerably larger .
they probably affect the nuv integrated spectra of clusters with significant bss components , even where there is strong competition from horizontal branch stars . however , for the last two reasons discussed above , it is doubtful that bss have much effect on the nuv spectra of galaxies . among our data ,
only field 4 has near - uv observations . since the far - uv observations in field 4 are not as deep as the near - uv , we have @xmath51 colors for only seven clusters ( see table [ t : catalog2 ] ) .
the @xmath93 versus mg@xmath82 diagram is shown in figure [ f : nuvv_mg2 ] .
this is the nuv analogue to figure [ f : fuvv_mg2 ] .
we plotted only m87 gcs with @xmath94 .
because of the large contribution of main sequence stars to the nuv fluxes , we expect the @xmath93 color to become redder as the metal abundance increases ( burstein et al .
1988 , dorman et al .
2003 and 3 ) .
the m87 gcs seem to follow this general trend .
however , as in figure [ f : fuvv_mg2 ] , they are systematically offset to bluer colors than the milky way gcs for a given metallicity .
the slope of the blue envelope in figure [ f : nuvv_mg2 ] agrees roughly with expectations ( see 3 ) for a metallicity effect if vertical scatter has been introduced by a variable hot horizontal branch population .
the five most mg - rich galaxies have bluer colors than the rest of the galaxy sample , probably due to strong contamination of the nuv light by their large ehb populations ( also evident in figure [ f : fuvv_mg2 ] ) .
figure [ f : fuvnuv_mg2 ] shows the @xmath51 color versus mg@xmath1 for the three samples .
the m87 gcs have @xmath51 colors bluer than the milky way gcs and comparable to the average galaxies .
however , in this diagram the m87 clusters appear to be an extension of the milky way gcs , becoming bluer in @xmath51 with increasing metallicity . as noted in our discussion above on interpreting uv colors , colors bluer than @xmath95
indicate the presence of dominant ehb populations .
most galaxies and 6 of the m87 clusters fall in this category .
m32 is again significantly redder than metal - rich m87 gcs in this plot , by @xmath86 mags in @xmath51 .
the uv color - color diagrams are shown in figure [ f : fuv_nuv_v_colorcolor ] .
dor95 emphasized the separation between milky way clusters and galaxies in similar plots . here , the m87 gcs again extend the milky way sequence to the bluer side , and the distinction between the gc and galaxy sequences is preserved in both plots .
dor95 suggested that the galaxies have bluer @xmath51 than clusters because the more metal rich stars in the galaxies have larger nuv line blanketing .
it is interesting to note that the metal - rich milky way cluster ngc 6441 , and to a lesser extent ngc 6388 , has photometric properties in figures 711 that are comparable to those of the uv - detected m87 gcs .
ngc 6388 and 6441 are known to have rr lyrae variables with periods anomalously larger than any other milky way clusters @xcite .
further studies of these two clusters may provide hints on the origin of the bluer @xmath2 colors seen in m87 gcs compared to the milky way counterparts .
the important empirical results here ( for what is admittedly a small cluster sample ) are that the m87 clusters appear to represent an extension of the milky way gc sequence in fuv - nuv - v - mg@xmath1 space and that the galaxies are found to have stellar populations distinct from those in either cluster system .
the uv light in the galaxies is evidently not produced by globular cluster - like subpopulations .
the bimodal distribution of cluster optical broadband colors seen in many elliptical galaxies indicates the presence of two different subpopulations ( see 1.2 ) . since optical broadband colors of very old stellar populations
are more sensitive to metallicity than age ( e.g. see yi 2003 ) , the bimodality is normally assumed to reflect the presence of distinct metal - poor and metal - rich populations .
the prevailing interpretation is that the bluer , metal - poor population is older and formed during the earliest phases of galaxy assembly , whereas the redder , metal - rich population formed later as a byproduct of gas - rich mergers between systems with more advanced chemical evolution .
k99 argued that the color and luminosity distributions in m87 were consistent with the redder population being 3 - 6 gyr younger .
however , the interpretation of the red clusters as younger is controversial ( e.g. catelan et al .
1998 ) . using their keck spectra of 150 gcs and the stellar population models of @xcite
, @xcite found no significant age different between the two subpopulations .
also , @xcite used the strmgren photometry of 628 clusters and various population synthesis models @xcite to conclude that within their measurement errors , the two subpopulations are coeval .
by contrast , puzia et al . ( 2005 ) recently used lick absorption line indices to identify a group of metal - rich , younger ( 5 - 8 gyr old ) gcs in m31 . here
we investigate the extent to which our uv data can contribute to the question of m87 cluster ages .
we compare our data to the most complete fiducial model grid of lee , lee & gibson ( 2002 ; hereafter llg02 ) .
the advantage of their models lies in the fact that they include detailed hb modeling and provide isochrones for a wide range of colors .
llg02 and @xcite show that , with proper calibration of the hb phases , colors such as @xmath2 can be more sensitive to ages than any optical - band integrated color .
however , the llg02 models do not include the ehb population ( @xmath96 k ) .
our version of llg02 figures 14-(a ) and ( b ) is shown in figure [ f : iso ] .
we have plotted the m87 cluster sample using [ fe / h ] values transformed from their @xmath41 colors , as described in 6.1 and the appendices .
we superpose the isochrones used in llg02 , kindly provided by h .- c .
the models exhibit rapid bluing of @xmath2 colors for a given metal abundance as the age increases past a threshold that varies with abundance . because of their blue @xmath2 colors , the m87 clusters fall at the extreme right hand side of the llg02 isochrone grid , where the models predict large ages for a given metallicity . according to k99 ,
the boundary between the blue and red subpopulations is near @xmath97 , which corresponds to [ fe / h ] @xmath98 .
if we adopt this as a dividing line in figure [ f : iso ] , we find that the models imply a large internal age spread of 26 gyr in both the blue and red subpopulations .
it seems clear that the llg02 isochrones are not realistic in the context of these data because 15 m87 clusters lie beyond the @xmath99 gyr contour and many of the others would be 24 gyr older than milky way clusters .
these values are unphysical because the color - magnitude diagrams for milky way clusters show them to be at the maximum age of @xmath100 gyr allowed by current cosmological parameters .
the reason for the age discrepancies found above is almost certainly the lack of a well - established physical prescription for determining the hb distribution for a population of given age and abundance .
dependences on core mass and abundance are well understood , but there is no deterministic theory for the rgb mass loss that fixes the hb envelope mass .
higher mass loss near the helium flash on the rgb will yield a hotter hb for a given age or metallicity .
the effects seen in figure [ f : iso ] could be explained with larger rgb mass loss in m87 than prevails in milky way clusters . in any case
, we can not use our uv observations to constrain the ages of the m87 clusters without a better understanding of rgb mass loss .
an alternative , or possibly complementary , possibility is that the m87 clusters are super - helium - rich , as recently proposed for one of the subpopulations in the galactic globular cluster @xmath12 cen by @xcite .
we explore this possibility in a forthcoming paper ( kaviraj et al .
2005 , in preparation ) .
one of our main conclusions in the sections above is that the m87 globular clusters have systematically brighter fuv components than do their milky way counterparts .
for instance , figure [ f : colorcolor ] shows that for a given @xmath69 , m87 gcs are always bluer in @xmath72 than mwgcs . to what extent
is this result affected by observational selection ?
unfortunately , selection effects are difficult to judge .
the milky way data are from various spacecraft , and target selection was inhomogeneous , based for instance on apparent magnitude and low foreground extinction rather than metal abundance or luminosity .
our m87 selection is actually much more homogeneous and is limited only by apparent uv magnitude ( @xmath101 ) .
an attempt to evaluate selection effects as a function of @xmath69 color is shown in figure [ f : selection ] . the upper panel of figure [ f : selection ] is the color / apparent magnitude diagram for the m87 optical and uv samples in fields 1 - 3 taken from figure 3 .
our uv sample includes almost all clusters brighter than @xmath102 but none fainter than @xmath103 .
we have divided the intermediate region into two bins : ( 1 ) @xmath104 and @xmath105 , ( 2 ) @xmath104 and @xmath106 .
the vertical dividing line corresponds to the color of the reddest cluster ( 47 tuc ) in the milky way uv - detected sample .
very few of the clusters in region ( 2 ) are detected in the fuv , whereas 59% of the clusters in region ( 1 ) are detected .
overall , we have uv detections for 61% of the m87 optical sample of clusters with @xmath70 colors in the range covered by the milky way gc sample . in the lower panel ,
we plot the @xmath58 colors or upper limits for the m87 optical sample and for the uv - detected milky way sample .
the blueward half of this panel makes a good case that the difference in uv properties between the m87 and milky way clusters is genuine and not a selection effect .
if one takes all the detections and limits plotted in this region as a hypothetical parent distribution , it is clear that there is a very low probability of choosing by chance a sample with the redder @xmath58 colors of the milky way subset .
optical @xmath69 colors are known to be good , if not perfect , indicators of the main sequence and giant branch populations in clusters , and they are negligibly influenced by the hb stars that produce most of the fuv light .
the inference is that for a given pre - hb population , the m87 clusters typically produce larger hot hb populations than do milky way clusters . although the milky way sample is incomplete , it does include most of the well - studied clusters with [ fe / h ] @xmath107 , and there is no reason to suppose that a better mw sample would change this situation . recall that the galex sample of m31 clusters ( preferentially luminous ones ) has properties similar to those of the mwgcs ( rey et al .
2005 ) , so m87 is distinct from both of the other two cluster systems explored in the uv .
we doubt that cluster mass is a factor here .
although , based on their v - band luminosities , the m87 detections are on average more massive than the mw sample , the uv color distinction persists throughout nearly the full mass range present in the milky way sample ( see figure 4 ) . turning to the redder side of the lower panel in figure [ f : selection ] , we have uv detections for about 50% of the m87 clusters in the range @xmath108 but very few of the redder clusters .
we can not determine typical properties for the redder m87 clusters , but we can say that some very metal - rich m87 clusters produce uv - bright populations comparable to those of metal - poor mw clusters .
a surprising result of our photometry is that in all four fields , we find sources that are detected in the far- or near - uv frames but _ not _ in the optical frames . in fields 1 , 2 , and 3
, we find 19 `` uv - only '' sources ( the second digit of the i d number is set to 2 in table [ t : catalog1 ] ) , and in field 4 , we find 24 uv - only sources ( table [ t : catalog3 ] ) .
all of the field 4 uv - only sources are detected in the nuv exposures ( see 2.3 ) , which go deeper than the fuv exposures , so we consider only those here .
first , we discuss the far - uv - only sources in fields 1 , 2 , and 3 .
these lie in the range @xmath109 .
close structural inspections show that they are similar to other sources , but our poor psfs limit conclusions .
we photometered these sources independently on all images and found that none show significant variations beyond photon statistics . in figure
[ f : fuvonly_colorlimits ] , we plot the color limits for these sources in the cmd .
all sources have colors bluer than 0.1 in @xmath58 , compared to the two bluest optically - confirmed clusters found in figure [ f : colormag ] , which have @xmath110 and @xmath111 .
there are several alternative explanations for these sources : * first , we doubt that many are detector artifacts .
mama detectors do not suffer from cosmic rays .
other detector artifacts should persist for some period of time ; therefore they should appear in the same location in different fields , but they do not .
no other studies using the stis fuv - mama have reported artifacts similar to the effect found in our fields . finally , in 9.2 below
, we show that there are a number of `` nuv - only '' sources that are detectable at different epochs at different detector locations .
* we can not rule out the possibility that these sources are old m87 clusters with a large ehb star fraction .
however , dor95 predicts that if all rgb stars go through the ehb channel , the @xmath2 color can not be bluer than @xmath112 .
a similar limit was predicted from the models of @xcite .
most sources in figure [ f : fuvonly_colorlimits ] have colors much bluer than the theoretical blue limit .
hot pagb stars are an alternative source of uv light , but considering the short lifetime of pagb stars , it is unlikely that a sufficiently large population of such objects could be produced . * these sources could also be young star clusters with ages @xmath113 gyr . according to @xcite
, the @xmath2 colors of young stellar systems can be as blue as @xmath114 .
colors in the observed range of 0 to @xmath115 correspond to cluster ages of 30300 myr . however , there is no other evidence for recent star formation in m87 .
* some of them could be transient sources in m87 such as novae .
many classical novae in m31 are found to reach @xmath116 to @xmath117 at their maximum @xcite , which is comparable to integrated magnitudes of globular clusters .
@xcite estimate the nova eruption rate of 100 - 1000 per year for m87 , following a simple purely theoretical calculation .
our far - uv observations span periods from a few hours to several months for different fields . since the maxima of fast novae only last for a day or so , the uv - only sources detected in field 3 for which exposures were split by 4 months are not likely to be of this kind .
we conclude that only a few of the far - uv - only sources in fields 1 , 2 , and 3 are likely to be nova eruptions caught near their maximum . *
uv - bright background galaxies are perhaps the best candidates .
so far , there have been only a few deep uv searches for galaxies .
the uv ( 2000 ) survey of galaxies in sa57 by the foca balloon - borne camera @xcite showed that many uv - selected late type galaxies in the redshift range @xmath118 have @xmath119 colors as blue as @xmath120 .
they also find a number of sources showing extreme uv - optical colors ( bluer than normal hubble sequence galaxies ) , reaching up to @xmath121 . these are thought to be galaxies undergoing intense star formation with little internal extinction .
the foca sample is much brighter than the objects we detect and has too low a surface density ( @xmath122 arcmin@xmath123 down to @xmath124 ) to be important in our fields ( our combined field coverage is 0.52 arcmin@xmath125 ) , but there are analogues to these sources at fainter magnitudes .
we can use the stis observation of the hubble deep field south ( hdfs ; gardner et al .
2000 ) to estimate the density of the fainter uv background sources in our fields .
we note that the far - uv observation of the hdfs was also done using the fuv - mama but with a different filter ( f25qtz ; 1450 1900 , pivot wavelength 1590 ) .
we converted their ab magnitudes to stmag magnitudes for direct comparison with our far - uv magnitudes and selected the sources with @xmath126 colors or color limits bluer than 0.5 .
then , we counted the number of sources that have far - uv magnitudes or limits in the range @xmath127 .
we obtain a surface density of @xmath128 sources per arcmin@xmath125 .
this is in acceptable agreement with @xmath129 sources per arcmin@xmath125 , the number density of the uv - only sources in our fields .
we therefore believe that the most likely explanation is that the far - uv - only sources in fields 1 , 2 , and 3 are uv - bright background galaxies .
we believe we do detect a number of m87 novae in field 4 ( see next section ) , but the ratio of novae to background galaxies should decline with radius , so that fewer will be found in fields 1 , 2 , and 3 .
the statistical uncertainty of these observations is too large for a meaningful estimate of the uv optical depth of m87 , but it is clearly small .
interestingly , @xcite note that they do not find any objects in their uv images that do not appear on the optical image of the hdfs . in the case of the m87 fields
, the bright diffuse galaxy background seriously limits the detection depth at optical wavelengths . a deeper optical image of our fields
will possibly reveal the nature of these sources .
we now discuss the uv - only sources on the near - uv frames of field 4 .
the near - uv frames were taken in four different epochs that span more than three years in time . the long span of these observations allows us to investigate the variability of the uv - only sources . in figure 15
we plot the long - term variability of the nuv - only sources .
the @xmath130 magnitudes of each source are plotted for four different epochs . for sources not detected in specific epochs , we provide the 3@xmath131 upper limits of the @xmath130 magnitudes .
the finding charts for these sources are provided in figure 16 .
based on the variability of the sources , we can divide them into three different categories ; ( 1 ) sources showing constant brightness : nuv-01 , -03 , -04 , -05 , -10 , and -21 ; ( 2 ) sources only detected in one epoch : nuv-02 , -08 , -09 , -11 , -12 , -13 , -17 , -18 , -19 , -20 , and , -22 ; and ( 3 ) recurrent sources which appear variable : nuv-07 , and -16 .
the sources showing constant brightness are likely to be background galaxies as discussed above for fields 1 , 2 , and 3 .
sources that fall into category ( 2 ) are candidates for nova outbursts .
recently , in an attempt to search for microlensing events in m87 , @xcite have monitored in @xmath47 and broad @xmath132 bands the field around the center of m87 over a month - long interval . out of the seven variable sources ,
they identify two as obvious novae : one classical ( pc1 - 1 ) and one in a globular cluster ( wfc2 - 6 ) ; two as possible novae ( pc1 - 3 and pc1 - 3 ) ; and one as a microlensing event ( wfc2 - 5 ) .
shara et al .
( 2004 ) also recently reported a classical nova outburst ( at optical wavelengths ) in an m87 globular cluster .
two of the @xcite sources , pc1 - 1 and pc1 - 2 , lie in our near - uv fields .
we detected pc1 - 1 in only one of the epochs ( go-8780 : nuv-13 ) while pc1 - 2 was not detected in any . coincidentally , the date of the go-8780 near - uv observation is approximately 40 days after the date pc1 - 1 reached its maximum brightness ( @xmath133 ) in optical bands . from the light curve in figure 6 of @xcite
, the optical luminosity of pc1 - 1 seems to decline rather rapidly and after 40 days it is expected to be significantly fainter than it was at maximum light .
interestingly , the 53 day oao - a2 observation of nova fh serpentis 1970 @xcite showed that the nova became progressively bluer during the early decline of the visual light curve with the bolometric luminosity remaining constant .
our detections of pc1 - 1 in the go-8780 near - uv frames suggest that this behavior was also shown by pc1 - 1 .
it is likely that most of the other sources in category ( 2 ) are also novae .
on the other hand , the recurrent variable sources such as nuv-07 , and -16 are difficult to explain .
they do not fit the pattern expected for novae , and most other kinds of accretion - driven variable stellar sources are intrinsically faint in the uv . because both sources nuv-07 and -16 lie close to m87 s nonthermal jet in projection
, a more exotic alternative is that they are a new class of source induced by activity in the jet .
earlier authors have speculated that the continuing , dramatic activity in the m87 jet might , through an undetermined mechanism , trigger point - like flares in its vicinity ( e.g. sparks et al .
2000 , perlman et al .
however , if triggering occurs , the observed variability will be strongly influenced by light travel time effects .
these objects are interesting enough to justify detailed monitoring .
we have obtained deep hst / stis far- and near - uv photometry of four fields near the center of the ge galaxy m87 in order to study the hot - star populations in its globular cluster system .
we made new fuv - mama / f25srf2 imaging observations in fields 1 , 2 , and 3 ( 25square each ) using timetag data acquisition in order to maximize access to periods of low sky background .
we extracted additional fuv and nuv images taken at four different epochs from the hst archive for a nuclear field ( field 4 ) .
we performed aperture photometry on all uv sources detected in the fields and at the positions of all optically - identified clusters in the deep hst survey of kundu et al.(1999 ) .
our limiting far - uv magnitude is @xmath37 , corresponding to @xmath38 .
we detect in the fuv a total of 69 objects in fields 1 , 2 , and 3 , including 50 of the 127 globular clusters identified by k99 to @xmath39 .
most of our uv detections appear to belong to the bluer of the two cluster subpopulations identified by k99 .
we did not find any radial trends of uv luminosities or uv / optical colors of our clusters , although given its small range of radii our sample is not well suited for testing such trends .
we compare the uv / optical photometric properties of the m87 clusters to those of milky way clusters ( from dor95 ) and m31 clusters ( from rey et al .
2005 ) . despite strong overlap in v- and i - band properties ,
the m87 gcs have uv / optical properties that are distinct from clusters in the milky way and in m31 .
the color distribution of the m87 clusters has a strong peak at @xmath134 and a mean of 1.50 , compared to corresponding values for the milky way sample of 2.35 and 2.30 , respectively . in the @xmath68 color - color diagram
, there is almost no overlap between the m87 and milky way samples , with m87 clusters being significantly bluer in @xmath2 at any @xmath69 .
m31 clusters have uv properties like the milky way sample and unlike m87 . in color plots
including the nuv band ( only seven m87 clusters here ) , the m87 clusters again are distinct from milky way clusters but appear to represent an extension of the mw sequence .
ngc 6441 is the closest milky way analogue to the m87 clusters .
seventeen of the uv - detected m87 clusters are redder in @xmath70 , and presumably more metal - rich , than any milky way sample cluster . some of these have @xmath72 colors as blue as any metal - poor mw cluster .
we do not sample the more metal rich population as well as the lower metallicity clusters .
however , there is a suggestion of a trend in which lower metallicity gcs in our m87 sample have bluer @xmath2 colors .
it is difficult to assess selection effects , especially because the milky way uv sample is very inhomogeneous .
overall , in fields 1 , 2 , and 3 we have uv detections for 61% of the m87 optical sample with @xmath70 colors in the range covered by the milky way gc sample .
we therefore believe our uv detections are representative of the m87 gc population for metallicities found among local group clusters and that the difference in uv properties between them and m87 gcs is genuine and not a selection effect .
we believe the stars producing the far - uv light lie on the extreme horizontal branch .
blue stragglers are in general too cool to contribute to the far - uv , though they could be important in the near - uv light of some clusters ( if not elliptical galaxies ) .
the inference is then that for a given pre - hb population ( i.e. main sequence and red giant branch ) , m87 clusters typically produce larger hot hb populations than do milky way clusters . m87
globulars with metallicities considerably in excess of those typical of the milky way are capable of producing large hot hb populations .
the five bluest m87 clusters have a mean value of @xmath72 of 0.62 .
this implies , based on the estimate of dor95 that the bluest possible color for an old population is @xmath135 , that @xmath136% of the hb stars in these clusters evolve through the hot ehb channels .
this sounds like a dramatic concentration , but it is only a factor of @xmath132 higher than occurs in well - observed clusters like @xmath12cen .
existing model grids for cluster uv properties ( e.g. llg02 ) do not match the m87 observations well .
effects such as larger giant branch mass loss or perhaps high helium abundances evidently act to enhance the importance of hot hb stars over the predictions for a given age and abundance . in order to compare our m87 clusters to earlier compilations of elliptical galaxy data
, we have estimated mg@xmath1 line index values for them from their @xmath70 colors using a new calibration based on data for clusters in the milky way , the ge galaxy ngc 1399 , and the outer parts of m87 .
we find that the three cluster systems lie on a common @xmath41-[fe / h ] relation .
the most metal - rich m87 gcs in our sample are near solar metallicity and overlap the e galaxy sample in estimated mg@xmath1 line indices .
nonetheless , the detected clusters produce much more uv light at a given mg@xmath1 , being up to 1 mag bluer than any ge galaxy in @xmath2 color and up to 3 mag bluer than the fiducial local group elliptical m32 at the same mg@xmath1 value .
the detected metal - rich clusters have an average @xmath72 color bluer than 90% of the entire galaxy sample .
they do not appear to represent a transition between milky way - type clusters and e galaxies .
we find that two galactic open clusters , ngc 188 and 6791 , have ehb components that place them among the metal - rich m87 clusters in the mg@xmath1-@xmath72 diagram
. these have well - determined ages of 6 - 7 gyr .
if , as the theoretical models predict , there should be a transition age above which the ehb becomes well populated , then this is younger than 6 gyr .
our main conclusion is that insofar as their uv / optical properties are concerned , the milky way clusters , the m87 clusters , and the elliptical galaxies constitute three distinct types of stellar populations .
more subtle distinctions in the optical , near - ir , and near - uv among cluster systems and between clusters and e galaxies have been established for some time ( e.g. frogel , persson , & cohen 1980 ; burstein et al.1984 ; ponder et al .
cluster mass does not seem to be a factor in the differences between the milky way and m87 cluster systems because the color effects are detected at all masses .
the very different dynamical environment of m87 could influence its cluster populations , for instance by affecting concentration ( but this can not be measured at the resolution of hst ) . enhanced red giant branch mass loss or high helium abundances
might also account for the large hot hb populations in m87 clusters .
the uv color differences are in the correct sense if the detected m87 clusters have significantly older light - weighted ages than the e galaxies .
whether or not the trends of elliptical galaxy uv colors discussed in 6 are consistent with the `` old base plus younger frosting '' models discussed by trager et al.(2000 ) is something that should be explored .
the large samples from galex will help to clarify the behavior of hot populations in elliptical galaxies . the existing galex sample ( rich et al .
2005 ) does not exhibit the strong metallicity - uvx correlation of the older iue sample ( in figure 8) .
however , we find that 40% of the galex sample is bluer than any galaxy in the iue sample , which suggests that additional factors such as age or recent star formation are playing a role . surprisingly , we find that 43 of our uv detections have no optical - band counterparts .
we consider various explanations for these `` uv - only '' sources .
most appear to be uv - bright background galaxies , seen through m87 , which therefore has low uv optical depth .
a set of 11 nuv variable sources detected at only one epoch in the central field are probably classical novae .
two recurrent variable sources have no obvious explanation but could be related to activity in the relativistic jet .
this work was supported at uva in part by nasa grants go-8643 and go-9455 from the space telescope science institute and long term space astrophysics grant nas5 - 02052 .
stsci is operated by the association of universities for research in astronomy , inc .
, under nasa contract na5 - 26555 .
ak acknowledges support from nasa ltsa grant nag5 - 12975 and grant ar-09208.01 from stsci .
jaf thanks dr .
sean solomon for his hospitality at the dtm , carnegie institution of washington .
the referee , ruth peterson , made some important suggestions that significantly improved the paper .
we are grateful to robert rood , ben dorman , and ricardo schiavon for helpful conversations and to h .- c .
lee for providing his model isocrones .
dor95 compiled uv data for milky way gcs from various sources and listed them in their table 1 . because the metallicities and reddening values they used are outdated ,
we have updated them with the new [ fe / h ] and @xmath36 values in the most recent version of harris catalog @xcite .
the update was done as following . for the @xmath137 , @xmath138 , and @xmath139 colors
, we reverted to the `` uncorrected '' values using the old @xmath36 and the reddening law by @xcite , and then corrected for reddenings , this time using the new @xmath36 and the same reddening law .
the oao far - uv fluxes @xcite were only used for four clusters : 47 tuc , ngc 1851 , m79 , and m5 . for m79 , we averaged the colors from oao and uit observations listed in dor95 .
the far - uv observation for @xmath12 cen is taken from uit photometry @xcite . to these data
, we have added integrated @xmath41 colors and @xmath56 magnitudes taken from the harris catalog .
the catalog was also updated by adopting mg@xmath1 measurements from the recent literature .
specifically , the lick / ids mg@xmath1 measurements and associated errors of ( 1 ) @xcite were adopted for ngc 5024 , 5272 , 5904 , 6205 , 6341 , and 7078 , and of ( 2 ) @xcite were adopted for ngc 6388 , 6441 , and 6626 . for clusters that overlap in both studies ( ngc 6356 , 6624 , and 6637 )
, we have taken error - weighted means .
although dor95 used a linear relation for the conversion of [ fe / h ] to mg@xmath1 for clusters that do not have mg@xmath1 measurements , other authors ( e.g. burstein et al .
1984 , brodie & huchra 1990 ) found nonlinear relationships .
@xcite noted that the linear relation breaks down in the metal - rich regime and provide a parabolic relation between various line indices and [ fe / h ] .
however , we found that the coefficient for the mg@xmath1 versus [ fe / h ] listed in their table 6 is incorrect in the sense that the fitted curve in their fig 7 is not consistent with the coefficients provided . using the same clusters as @xcite , we derived the following relation : @xmath140 } + 0.041(\pm 0.008){\rm [ fe / h]^{2}}. { \rm mg_{2 } } = 0.268(\pm 0.006 ) + 0.204(\pm 0.011){\rm [ fe / h ] } + 0.042(\pm 0.004){\rm [ fe / h]^{2}}.\end{aligned}\ ] ] we used the relation above to derive predicted mg@xmath1 values from their [ fe / h ] for milky way clusters in our updated table that do not have mg@xmath1 measurements either in @xcite or @xcite .
the adopted values of all parameters for our milky way gcs are presented in table [ t : mwgc_dor95_dat ] . for consistency
, we have also updated the mg@xmath1 measurements for the galaxies discussed in 6 by adopting the lick / ids mg@xmath1 measurements of @xcite .
the mg@xmath82 index is , of course , sensitive to age as well as metallicity ( e.g. worthey 1994 ) .
the calibration here is technically correct only for systems in the age range @xmath141 - 13 gyr ; but the actual age range in the milky way sample evidently produces only a small effect on the line index .
here we investigate the transformation between @xmath142 colors and [ fe / h ] for the m87 clusters .
a number of such studies of other cluster systems have been made .
@xcite derived an empirical relationship between [ fe / h ] and @xmath142 using the compilation of milky way clusters by @xcite and @xcite .
more recently , @xcite derived the dependence of @xmath69 on [ fe / h ] using the 1999 version of the mcmaster catalog of milky way clusters . however , there are well - known pitfalls ( see e.g. kissler - patig et al .
1998 ) in trying to extrapolate results for milky way clusters to higher metallicities such as those characteristic of many m87 clusters . @xcite
combined their data for the ge galaxy ngc 1399 with those of the milky way clusters in order to extend the relation to the higher metallicity range .
meanwhile , @xcite pointed out that the difference between the @xcite and @xcite relations are mostly due to the choice of the independent variable .
they also argue that neither of the quantities should be used as an independent variable but instead use a `` bisector '' @xcite to derive the relationship . to study the color - metallicity relation for m87 clusters , we use ten m87 clusters in table 2 of the spectroscopic study by @xcite that overlap with the @xmath28 and @xmath47 photometry of k99 .
using the i d numbers of @xcite , these are 5001 , 5002 , 5012 , 5021 , 5028 , 649 , 697 , 746 , 750 , and 892 . the metallicities of these clusters are in the range @xmath143 [ fe / h ] @xmath144 . here
, we derive our own relation for the combined sample consisting of milky way clusters with @xmath145 , ngc 1399 gcs , and m87 gcs .
we used a bisector method @xcite to derive a new empirical relation as shown in figure [ f : vifeh_conversion ] .
the figure shows that , with the exception of one outlier , the m87 clusters fall on the same linear color - metallicity relation as the milky way and ngc 1399 clusters .
the correlation is good despite what must be a finite dispersion in ages among the three systems and within each system , although it would undoubtedly be improved if ages could be assigned to individual clusters .
table [ t : vifeh_relation ] lists our fitting parameters and compares them to those of other studies mentioned above .
we adopt this relation for the discussion of m87 clusters in the age - abundance grid ( 7 ) .
we also use this relation in combination with the results of appendix a to estimate mg@xmath1 values for our m87 cluster sample ( 6 ) .
we use this approach in preference to attempting to determine a color - mg@xmath1 relation for m87 using the spectroscopy of @xcite because of reservations they expressed about their mg indices .
ashman , k. m. , & zepf , s. e. 1992 , , 384 , 50 ashman , k. m. , bird , c. m. , & zepf , s. e. 1994 , , 108 , 2348 ashman , k. m. , & zepf , s. e. 1998 , globular cluster systems ( cambridge : cambridge univ .
press ) baum , w. a. 1955 , , 67 , 328 bailyn , c. d. 1995 , , 33 , 133 baltz , e. a. , lauer , t. r. , zurek , d. r. , gondolo , p. , shara , m. m. , silk , j. , & zepf , s. e. 2004 , , 610 , 691 bohlin , r. c. , et al .
1993 , , 417 , 127 bressan , a. , chiosi , c. , & tantalo , r. 1996 , , 311 , 425 brodie , j. p. , & huchra , j. p. 1990
, , 362 , 503 brown , t. m. , ferguson , h. c. , davidsen , a. , f. , & dorman , b. 1997 , , 482 , 685 brown , t. m. , ferguson , h. c. , deharveng , j. -m . , jedrzejewski , r. i .. 1998 , , 508 , l139 brown , t. m. , bowers , c. w. , kimble , r. a. , & ferguson , h. c. 2000a , apj , 529 , l89 brown , t. m. , kimble , r. a. , ferguson , h. c. , gardnerm j. p. , collins , n. r. , & hill , r. s. , 120 , 1153 brown , t. m. , ferguson , h. c. , smith , e. , bowers , c. w. , kimble , r. a. , renzini , a. , & rich , r. m. 2003 , , 584 , l69 bruzual , a. g. , & charlot , s. 1993 , , 405,538 burstein , d. , bertola , f. , buson , l. m. , faber , s. m. , & lauer , t. r. 1988 , , 328 , 440 burstein , d. , faber , s. m. , gaskell , c. m. , & krumm , n. 1984 , 287 , 586 burstein , d. , et al .
2004 , , 614 , 158 burstein , d. , & heiles , c. 1984 , , 54 , 33 cardelli , j. a. , clayton , c. , & mathis , j. s. 1989 , , 345 , 245 catelan , m. , borissova , j. , sweigart , a. v. , & spassova , n. 1998 , , 494 , 265 code , a. d. 1969 , pasp , 81 , 475 cohen , j. g. , blakeslee , j. p. , & ryzhov , a. 1998 , , 496 , 808 ct , p. , marzke , r. o. , & west , m. j. 1998 , , 501 , 554 couture , j. , harris , w. e. , & allwright , j. w. b. 1990 , , 73 , 671 dcruz , n. l. , dorman , b. , rood , r. t. , & oconnell , r. w. 1996 , , 466 , 359 dcruz , n. l. , et al .
2000 , , 530 , 352 de boer , k. s. 1985 , , 142 , 321 della valle , m. 1991 , , 252 , l9 dinescu , d. i. , demarque , p. , guenther , d. b. , & pinsonneault , m. h. 1995 , , 109 , 2020 dorman , b. , rood , r. t. , & oconnell , r. w. 1993 , , 419 , 596 dorman , b. , oconnell , r. w. , & rood , r. t. 1995 , , 442 , 105 ( dor95 ) dorman , b. , oconnell , r. w. , & rood 2003 , apj , 591 , 878 faber , s. m. 1983 , highlights of astronomy , 6 , 165 forbes , d. a. , brodie , j. p. , & grillmair , c. j. 1997 , , 113 , 1652 friel , e. d. , janes , k. a. , tavarez , m. , scott , j. , rocio , k. , lotz , j. , hong , l. , & miller , n. 2002 , , 124 , 2693 frogel , j. a. , persson , s. e. , & cohen , j. g. 1980 , , 240 , 785 fusi - pecci , f. , ferraro , f. r. , bellazzini , m. , djorgovski , s. , piotto , g. , & buonanno , r. 1993 , aj , 105 , 1145 gallagher , j. s. , & code , a. d. 1974 , , 189 , 303 gardner , j. p. et al .
2000 , , 119 , 486 greggio , l. , & renzini , a. 1990 , , 364 , 35 harris , w. e. 1996 , , 112 , 1487 harris , w. e. , harris , g. l. h. , & mclaughlin , d. e. 1998 , 115 , 1801 harris , w. e. , kavelaars , j. j. , hanes , d. a. , hesser , j. e. , & pritchet , c. j. 2000 , , 533 , 137 isobe , t. , feigelson , e. d. , akritas , m. g. , & babu , g. j. 1990 , , 364 , 104 jordn , a. , ct , p. , west , m. j. , & marzke , r. o. 2002 , , 576 , l113 kaluzny , j. , & rucinski , s. m. 1995 , a&as , 114 , 1 kinney , a. l. , bohlin , r. c. , calzettia , d. , panagia , n. , & wyse , r. 1993 , , 86 , 5 kissler - patig , m. , brodie , j. p. , schroder , l. l. , forbes , d. a. , grillmair , c. j. , & huchra , j. p. 1998
, , 115 , 105 kundu , a. , & whitmore , b. c. 1998 , , 116 , 2841 kundu , a. , whitmore , b. c. , sparks , w. b. , macchetto , f. d. , zepf , s. e. , & ashman , k. m. 1999 , , 513 , 733 ( k99 ) kundu , a. , & whitmore , b. c. 2001a , , 121 , 2950 kundu , a. , & whitmore , b. c. 2001b , , 122 , 1251 landsman , w. b. , sweigart , a. v. , bohlin , r. c. , neff , s. g. , oconnell , r. w. , roberts , m. s. , smith , a. m. , & stecher , t. p. 1996
, , 472 , l93 landsman , w. , bohlin , r. c. , neff , s. g. , oconnell , r. w. , roberts , m. s. , smith , a. m. , & stecher , t. p. 1998 , aj 116 , 789 larsen , s. s. , brodie , j. p. , huchra , j. p. , forbes , d. a. , & grillmair , c. j. 2001 , , 121 , 2974 lata , s. pandey , a. k. , sagar , r. , & mohan , v. 2002 , a&a , 388 , 158 lee , h. -c . ,
lee , y. -w . ,
& gibson , b. , k. 2002 , , 124 , 2664 ( llg02 ) lee , m. g. , & geisler , d. 1993 , , 106 , 493 lee , y. -w . ,
demarque , p. , & zinn , r. 1994 , , 423 , 248 lee , y. -w .
et al . 2005a , , 619 , l103 lee , y. -w .
et al . 2005b , , 621 , l57 leitherer , c. , et al .
2001 , `` stis instrument handbook '' , version 5.1 , ( baltimore : stsci ) liebert , j. , saffer , r. a. , & green , e. m. 1994 , apj , 107 , 1408 ma , c. , et al . 1998 , ,
116 , 516 maraston , c. 1998 , , 300 , 872 martin , c. et al .
2003 , spie , 4854 , 336 milliard , b. , donas , j. , laget , m. , armand , c. & vuillemin , a. 1992 , , 257 , 24 oconnell , r. w. 1980 , , 236 , 430 oconnell , r. w. et al .
1997 , aj , 114 , 1982 oconnell , r. w. 1999 , , 37 , 603 ohl , r. g. , et al .
1998 , , 505 , l11 paresce , f. , meylan , g. , shara , m. , baxter , d. , & greenfield , p. 1991
, nature , 342 , 297 perlman , e. s. , harris , d. e. , biretta , j. a. , sparks , w. b. , & macchetto , f. d. 2003 , , 599 , 65 ponder , j. m. et al .
1998 , , 116 , 2297 pritzl , b. , smith , h. a. , catelan , m. , & sweigart , a. v. 2000 , , 530 , l41 pritzl , b. , smith , h. a. , catelan , m. , & sweigart , a. v. 2001 , , 122 , 2600 pritzl , b. , smith , h. a. , catelan , m. , & sweigart , a. v. 2002 , , 124 , 949 proffitt , c. r. , brown , t. m. , mobasher , b. , & davies , j. 2003 , `` instrument science report stis 2003 - 01 '' , ( baltimore : stsci ) puzia , t. h. , saglia , r. p. , kissler - patig , m. , maraston , c. , greggio , k. , renzini , a. , & ortolani , s. 2002 , , 395 , 45 puzia , t. h. , perrett , k. m. , & bridges , t. j. 2005 , , in press racine , r. 1968 , , 80 , 326 reed , b. c. , hesser , j. e. , & shawl , s. j. 1988 , , 100 , 545 rey , s. -c .
2005 , , in press rich , m. r. , et al .
1997 , , 484 , l25 rich , m. r. , et al .
2005 , , in press rose , j. a. 1984 , , 89 , 1238 rose , j. a. 1994 , aj , 107 , 206 rose , j. a. , & deng , s. 1999 , apj , 117 , 2213 schiavon , r. p. 2005 , apjs , in press schiavon , r. p. , caldwell , n. , & rose , j. a. , , 127 , 1513 schiavon , r. p. , faber , s. m. , castilho , b. v. , & rose , j. a. 2002 , , 580 , 850 schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , , 500 , 525 shara , m. m. , zurek , d. r. , baltz , e. a. , lauer , t. r. , silk , j. 2004 , , 605 , l117 sparks , w. b. , ferguson , h. , jedrzejewski , r. , macchetto , f. d. , biretta , j. , & livio , m. 2000 , baas , 197 , # 75.11 stetson , p. b. 1987 , , 99 , 191 strom , s. e. , forte , j. c. , harris , w. e. , strom , k. m. , wells , d. c. , & smith m. g. 1981 , , 245 , 416 sullivan , m. , treyer , m. a. , ellis , r. s. , bridges , t. j. , milliard , b. , & donas , j. 2000 , , 312 , 442 sweigart , a. v. 1997 , , 474 , l23 thomas , d. , maraston , c. , & bender , r. 2003 , mnras , 339 , 897 trager , s. c. , worthey , g. , faber , s. m. , burstein , d. , & gonzlez , j. j. 1998 , , 116 , 1 trager , s. c. , faber , s. m. , worthey , g. , & gonzalez , j. j. 2000 , , 120 , 188 whitmore , b. c. , sparks , w. b. , lucas , r. a. , macchetto , f. d. , & biretta , j. a. 1995 , , 454 , l73 whitney , j. h. , et al .
1994 , , 108 , 1350 whitney , j. h. , et al .
1998 , , 495 , 284 worthey , g. 1994 , , 95 , 107 worthey , g. 2004 , , 128 , 2826 yi , s. , demarque , p. , & oemler , a. jr . 1997 , , 486 , 201 yi , s. , demarque , p. , & oemler , a. jr . 1998 , , 492 , 480 yi , s. , lee , y .- w . , woo , j .- h . , park , j .- h . ,
demarque , p. , & oemler , a. 1999 , apj , 513 , 128 yi , s. k. 2003 , , 582 , 202 yong , h. , demarque , p. , & yi , s. 2000 , , 539 , 928 zepf , s. e. , & ashman , k. m. 1993 , , 264 , 611 zinn , r. 1985 , , 293 , 424 cccccc + o6be01oqq & fuv - mama & f25srf2 & timetag & 2680 & 2001 - 04 - 27 20:36:28 + o6be01osq & fuv - mama & f25srf2 & timetag & 2900 & 2001 - 04 - 27 22:08:09 + o6be02010 & ccd & f28x50lp & accum & 518 & 2001 - 04 - 23 18:39:19 + o6be02ceq & fuv - mama & f25srf2 & timetag & 2000 & 2001 - 04 - 23 18:52:00 + o6be02cgq & fuv - mama & f25srf2 & timetag & 2900 & 2001 - 04 - 23 20:12:30 + + o6be03xzq & fuv - mama & f25srf2 & timetag & 2680 & 2002 - 06 - 12 01:10:06 + o6be03ygq & fuv - mama & f25srf2 & timetag & 2900 & 2002 - 06 - 12 02:40:57 + o6be04010 & ccd & f28x50lp & accum & 518 & 2002 - 06 - 12 04:20:04 + o6be04z4q & fuv - mama & f25srf2 & timetag & 2000 & 2002 - 06 - 12 04:33:45 + o6be04znq & fuv - mama & f25srf2 & timetag & 2900 & 2002 - 06 - 12 05:53:18 + + o6be06010 & ccd & f28x50lp & accum & 518 & 2002 - 06 - 08 12:10:52 + o6be06omq & fuv - mama & f25srf2 & timetag & 2000 & 2002 - 06 - 08 12:24:33 + o6be06p3q & fuv - mama & f25srf2 & timetag & 2900 & 2002 - 06 - 08 13:43:21 + o6be15m9q & fuv - mama & f25srf2 & timetag & 2680 & 2002 - 02 - 14 16:47:22 + o6be15mbq & fuv - mama & f25srf2 & timetag & 2900 & 2002 - 02 - 14 18:19:24 + + o5dc01bpq & fuv - mama & f25srf2 & accum & 856 & 1999 - 05 - 17 10:33:00 + o5dc01bqq & fuv - mama & f25srf2 & accum & 850 & 1999 - 05 - 17 10:48:00 + o5dc01btq & fuv - mama & f25srf2 & accum & 910 & 1999 - 05 - 17 12:05:00 + o5dc01buq & fuv - mama & f25srf2 & accum & 910 & 1999 - 05 - 17 12:21:00 + o5dc01bxq & nuv - mama & f25qtz & accum & 614 & 1999 - 05 - 17 13:31:00 + o5dc01byq & nuv - mama & f25qtz & accum & 600 & 1999 - 05 - 17 13:42:00 + o5dc01c0q & nuv - mama & f25qtz & accum & 600 & 1999 - 05 - 17 13:53:00 + o5dc01c2q & nuv - mama & f25qtz & accum & 600 & 1999 - 05 - 17 14:03:00 + o6a351evq & nuv - mama & f25qtz & accum & 600 & 2001 - 07 - 30 23:18:00 + o6a351ewq & nuv - mama & f25qtz & accum & 600 & 2001 - 07 - 30 23:29:00 + o6a351eyq & nuv - mama & f25qtz & accum & 558 & 2001 - 07 - 30 23:40:00 + o6a351f0q & nuv - mama & f25qtz & accum & 600 & 2001 - 07 - 30 23:50:00 + o53c61keq & nuv - mama & f25qtz & accum & 600 & 2002 - 02 - 27 04:54:00 + o53c61kfq & nuv - mama & f25qtz & accum & 600 & 2002 - 02 - 27 05:53:00 + o53c61khq & nuv - mama & f25qtz & accum & 558 & 2002 - 02 - 27 06:04:00 + o53c61kjq & nuv - mama & f25qtz & accum & 600 & 2002 - 02 - 27 06:14:00 + o6m201w4q & nuv - mama & f25qtz & accum & 600 & 2002 - 07 - 17 08:54:00 + o6m201w5q & nuv - mama & f25qtz & accum & 600 & 2002 - 07 - 17 09:05:00 + o6m201w7q & nuv - mama & f25qtz & accum & 600 & 2002 - 07 - 17 09:15:00 + o6m201w9q & nuv - mama & f25qtz & accum & 600 & 2002 - 07 - 17 09:25:00 lcc 1 & 8150 . & @xmath146 + 2 & 7650 . & @xmath147 + 3 & 6600 . & @xmath148 + 4 fuv & 3526 . & @xmath149 + 4 nuv & 2414 .
& @xmath150 + 4a nuv & 2358 . & @xmath151 + 4b nuv & 2358 . & @xmath152 + 4c nuv & 2358 . & @xmath153 crrccccccccc 1001 & @xmath154 & @xmath155 & 38.71 & 22.830 & 0.063 & 1.101 & 0.064 & 0.954 & 0.015 + 1002 & @xmath156 & @xmath157 & 38.72 & 25.040 & 0.267 & 1.656 & 0.269 & 0.947 & 0.055 + 1003 & @xmath158 & @xmath159 & 35.14 & 24.744 & 0.218 & 1.136 & 0.224 & 0.953 & 0.082 + 1004 & @xmath160 & @xmath161 & 43.92 & 22.493 & 0.052 & 1.310 & 0.052 & 1.279 & 0.008 + 1005 & @xmath162 & @xmath163 & 36.85 & 23.121 & 0.074 & 1.628 & 0.075 & 1.010 & 0.014 + 1006 & 0.39 & @xmath164 & 32.06 & 22.477 & 0.052 & 1.347 & 0.052 & 0.928 & 0.011 + 1007 & 2.12 & @xmath165 & 47.42 & 24.511 & 0.179 & 3.083 & 0.179 & 1.188 & 0.011 + 1008 & 4.66 & @xmath166 & 40.91 & 24.505 & 0.201 & 1.477 & 0.203 & 1.138 & 0.039 + 1009 & 6.63 & @xmath167 & 43.29 & 24.220 & 0.159 & 1.908 & 0.160 & 1.145 & 0.020 + 1010 & 0.01 & @xmath168 & 31.51 & 23.110 & 0.078 & 1.371 & 0.079 & 0.929 & 0.017 + 1011 & 0.13 & @xmath169 & 32.34 & 24.948 & 0.290 & 1.827 & 0.293 & 0.887 & 0.067 + 1012 & 0.44 & @xmath170 & 36.65 & 23.869 & 0.126 & 2.155 & 0.126 & 1.311 & 0.013 + 1013 & 11.31 & @xmath171 & 27.91 & 23.852 & 0.124 & 2.335 & 0.124 & 1.283 & 0.016 + 1014 & 11.97 & @xmath172 & 36.10 & 23.138 & 0.083 & 1.259 & 0.084 & 1.111 & 0.017 + 1015 & 12.09 & @xmath173 & 36.99 & 22.509 & 0.056 & 2.095 & 0.056 & 1.197 & 0.006 + 1016 & 13.23 & @xmath174 & 43.59 & 23.286 & 0.091 & 1.563 & 0.091 & 1.032 & 0.014 + 1017 & 13.24 & @xmath175 & 47.11 & 25.078 & 0.304 & 0.935 & 0.310 & 1.041 & 0.092 + 1018 & 14.34 & @xmath176 & 35.93 & 21.675 & 0.035 & 0.963 & 0.035 & 1.018 & 0.007 + 1101 & @xmath177 & @xmath178 & 36.07 & @xmath179 & & @xmath180 & & 1.237 & 0.091 + 1102 & @xmath181 & @xmath182 & 40.47 & @xmath183 & & @xmath184 & & 1.116 & 0.049 + 1103 & @xmath185 & @xmath186 & 43.35 & @xmath187 & & @xmath188 & & 0.982 & 0.111 + 1104 & @xmath189 & @xmath190 & 49.85 & @xmath191 & & @xmath192 & & 1.186 & 0.031 + 1105 & @xmath193 & @xmath194 & 28.91 & @xmath195 & & @xmath196 & & 1.270 & 0.023 + 1106 & @xmath197 & @xmath198 & 50.19 & @xmath199 & & @xmath200 & & 1.254 & 0.088 + 1107 & 0.08 & @xmath201 & 35.11 & @xmath202 & & @xmath203 & & 1.173 & 0.038 + 1108 & 1.35 & @xmath204 & 49.51 & @xmath205 & & @xmath206 & & 1.261 & 0.021 + 1109 & 1.36 & @xmath207 & 47.03 & @xmath208 & & @xmath209 & & 1.333 & 0.040 + 1110 & 3.35 & @xmath210 & 44.84 & @xmath211 & & @xmath212 & & 1.125 & 0.093 + 1111 & 3.63 & @xmath213 & 38.28 & @xmath214 & & @xmath215 & & 0.947 & 0.287 + 1112 & 5.54 & @xmath216 & 41.68 & @xmath217 & & @xmath218 & & 1.279 & 0.035 + 1113 & 5.59 & @xmath219 & 44.85 & @xmath220 & & @xmath221 & & 1.211 & 0.031 + 1114 & 7.93 & @xmath222 & 35.70 & @xmath223 & & @xmath224 & & 1.070 & 0.168 + 1115 & 8.58 & @xmath225 & 33.39 & @xmath226 & & @xmath227 & & 1.233 & 0.048 + 1116 & 8.74 & @xmath228 & 42.11 & @xmath208 & & @xmath229 & & 1.136 & 0.030 + 1117 & 9.16 & @xmath230 & 36.48 & @xmath231 & & @xmath232 & & 0.791 & 0.134 + 1118 & 10.67 & @xmath233 & 39.24 & @xmath234 & & @xmath235 & & 1.256 & 0.030 + 1119 & 11.27 & @xmath236 & 31.93 & @xmath237 & & @xmath238 & & 1.322 & 0.035 + 1120 & 11.34 & @xmath239 & 26.52 & @xmath240 & & @xmath241 & & 1.227 & 0.081 + 1121 & 11.35 & @xmath242 & 35.16 & @xmath243 & & @xmath244 & & 1.203 & 0.282 + 1122 & 11.97 & @xmath172 & 36.10 & @xmath245 & & @xmath246 & & 1.111 & 0.017 + 1123 & 14.32 & @xmath247 & 31.36 & @xmath248 & & @xmath249 & & 1.211 & 0.062 + 1124 & 14.52 & @xmath250 & 34.26 & @xmath251 & & @xmath252 & & 1.242 & 0.047 + 1125 & 14.74 & @xmath253 & 40.16 & @xmath254 & & @xmath255 & & 1.418 & 0.091 + 1126 & 18.45 & @xmath256 & 43.19 & @xmath248 & & @xmath257 & & 0.901 & 0.102 + 1127 & 18.69 & @xmath258 & 46.21 & @xmath259 & & @xmath260 & & 1.146 & 0.155 + 1128 & 19.18 & @xmath261 & 46.07 & @xmath262 & & @xmath263 & & 1.355 & 0.069 + 1201 & @xmath264 & @xmath265 & 39.07 & 24.944 & 0.253 & @xmath266 & & & + 1202 & @xmath267 & @xmath268 & 36.93 & 25.395 & 0.336 & @xmath269 & & & + 1203 & @xmath270 & @xmath271 & 44.21 & 24.940 & 0.241 & @xmath272 & & & + 1204 & @xmath273 & @xmath274 & 41.07 & 24.773 & 0.214 & @xmath275 & & & + 1205 & 2.28 & @xmath276 & 31.98 & 23.921 & 0.124 & @xmath277 & & & + 1206 & 4.20 & @xmath278 & 49.49 & 25.388 & 0.328 & @xmath279 & & & + 1207 & 9.06 & @xmath280 & 32.67 & 24.752 & 0.246 & @xmath281 & & & + 1208 & 12.72 & @xmath282 & 36.33 & 24.729 & 0.264 & @xmath283 & & & + 1209 & 14.22 & @xmath284 & 32.10 & 24.291 & 0.170 & @xmath285 & & & + + + 2001 & 44.34 & 45.39 & 63.45 & 24.012 & 0.121 & 1.112 & 0.122 & 0.925 & 0.030 + 2002 & 50.64 & 55.81 & 75.36 & 22.011 & 0.042 & 1.906 & 0.042 & 1.133 & 0.004 + 2003 & 52.51 & 35.20 & 63.21 & 24.771 & 0.197 & 1.207 & 0.200 & 1.059 & 0.051 + 2004 & 53.77 & 48.32 & 72.29 & 23.168 & 0.081 & 1.340 & 0.081 & 0.974 & 0.013 + 2005 & 56.01 & 59.09 & 81.42 & 23.591 & 0.095 & 1.469 & 0.096 & 1.026 & 0.015 + 2006 & 57.50 & 56.72 & 80.76 & 25.405 & 0.337 & 2.067 & 0.338 & 1.098 & 0.034 + 2007 & 57.63 & 63.33 & 85.63 & 24.908 & 0.213 & 0.618 & 0.219 & 0.798 & 0.097 + 2008 & 58.52 & 38.44 & 70.01 & 24.303 & 0.178 & 2.453 & 0.178 & 1.107 & 0.012 + 2009 & 59.68 & 53.04 & 79.84 & 24.278 & 0.171 & 2.623 & 0.171 & 1.207 & 0.010 + 2010 & 60.29 & 42.14 & 73.56 & 23.521 & 0.108 & 1.378 & 0.108 & 1.044 & 0.015 + 2011 & 62.53 & 42.94 & 75.86 & 23.160 & 0.088 & 1.274 & 0.088 & 1.062 & 0.013 + 2012 & 63.68 & 46.28 & 78.72 & 23.044 & 0.080 & 1.208 & 0.080 & 1.045 & 0.011 + 2013 & 63.86 & 43.62 & 77.34 & 24.434 & 0.216 & 1.505 & 0.217 & 1.055 & 0.027 + 2014 & 65.49 & 49.50 & 82.09 & 23.117 & 0.082 & 0.928 & 0.083 & 0.986 & 0.017 + 2101 & 44.97 & 52.04 & 68.78 & @xmath286 & & @xmath287 & & 1.368 & 0.214 + 2102 & 46.17 & 41.83 & 62.30 & @xmath288 & & @xmath289 & & 1.200 & 0.104 + 2103 & 46.46 & 54.90 & 71.92 & @xmath290 & & @xmath291 & & 1.165 & 0.025 + 2104 & 48.05 & 43.02 & 64.49 & @xmath292 & & @xmath293 & & 0.925 & 0.065 + 2105 & 49.29 & 45.24 & 66.90 & @xmath294 & & @xmath295 & & 1.160 & 0.170 + 2106 & 49.45 & 37.71 & 62.19 & @xmath296 & & @xmath297 & & 1.076 & 0.133 + 2107 & 49.67 & 41.46 & 64.69 & @xmath298 & & @xmath299 & & 1.379 & 0.143 + 2108 & 49.84 & 37.69 & 62.49 & @xmath288 & & @xmath300 & & 1.008 & 0.190 + 2109 & 49.93 & 41.95 & 65.22 & @xmath301 & & @xmath302 & & 1.087 & 0.354 + 2110 & 51.23 & 39.52 & 64.71 & @xmath303 & & @xmath304 & & 1.457 & 0.150 + 2111 & 52.53 & 51.91 & 73.86 & @xmath305 & & @xmath306 & & 1.172 & 0.040 + 2112 & 53.40 & 36.46 & 64.66 & @xmath294 & & @xmath307 & & 1.355 & 0.065 + 2113 & 54.04 & 56.46 & 78.15 & @xmath308 & & @xmath309 & & 1.329 & 0.027 + 2114 & 54.39 & 52.36 & 75.50 & @xmath310 & & @xmath311 & & 1.348 & 0.034 + 2115 & 55.39 & 58.38 & 80.48 & @xmath308 & & @xmath312 & & 1.205 & 0.093 + 2116 & 55.55 & 36.30 & 66.36 & @xmath313 & & @xmath314 & & 0.765 & 0.296 + 2117 & 56.66 & 54.46 & 78.59 & @xmath315 & & @xmath316 & & 0.817 & 0.031 + 2118 & 56.93 & 43.15 & 71.43 & @xmath317 & & @xmath318 & & 1.420 & 0.102 + 2119 & 56.96 & 62.77 & 84.77 & @xmath319 & & @xmath320 & & 1.067 & 0.031 + 2120 & 57.96 & 41.65 & 71.37 & @xmath321 & & @xmath322 & & 1.338 & 0.062 + 2121 & 58.08 & 62.43 & 85.27 & @xmath323 & & @xmath324 & & 0.937 & 0.407 + 2122 & 60.14 & 58.64 & 84.00 & @xmath325 & & @xmath326 & & 1.655 & 0.437 + 2123 & 60.35 & 37.09 & 70.84 & @xmath327 & & @xmath328 & & 1.221 & 0.138 + 2124 & 64.90 & 49.68 & 81.73 & @xmath329 & & @xmath330 & & 1.366 & 0.019 + 2125 & 66.74 & 41.56 & 78.63 & @xmath331 & & @xmath332 & & 0.879 & 0.193 + 2201 & 45.12 & 52.96 & 69.58 & 25.578 & 0.347 & @xmath333 & & & + 2202 & 53.09 & 44.46 & 69.25 & 24.558 & 0.201 & @xmath334 & & & + 2203 & 55.06 & 44.59 & 70.85 & 24.968 & 0.275 & @xmath335 & & & + 2204 & 67.12 & 46.96 & 81.92 & 23.125 & 0.082 & @xmath336 & & & + + + 3001 & 6.85 & 16.78 & 18.13 & 22.765 & 0.077 & 1.407 & 0.080 & 1.054 & 0.033 + 3002 & 11.66 & 17.01 & 20.62 & 24.756 & 0.309 & 1.376 & 0.324 & 1.064 & 0.161 + 3003 & 12.06 & 21.49 & 24.65 & 23.099 & 0.087 & 1.192 & 0.089 & 1.000 & 0.035 + 3004 & 13.36 & 14.08 & 19.41 & 24.297 & 0.214 & 1.525 & 0.223 & 1.248 & 0.094 + 3005 & 13.81 & 16.00 & 21.13 & 23.125 & 0.093 & 0.858 & 0.099 & 1.056 & 0.052 + 3006 & 15.16 & 10.83 & 18.63 & 23.519 & 0.132 & 2.569 & 0.133 & 1.257 & 0.019 + 3007 & 17.44 & 5.25 & 18.21 & 23.258 & 0.108 & 1.612 & 0.111 & 1.191 & 0.042 + 3008 & 19.22 & 8.68 & 21.09 & 24.210 & 0.211 & 1.316 & 0.219 & 0.986 & 0.093 + 3009 & 19.35 & 24.44 & 31.17 & 24.579 & 0.217 & 0.827 & 0.228 & 0.833 & 0.119 + 3010 & 19.58 & 13.69 & 23.89 & 23.912 & 0.155 & 0.629 & 0.174 & 0.941 & 0.127 + 3011 & 21.28 & 26.49 & 33.98 & 23.752 & 0.122 & 0.964 & 0.125 & 0.985 & 0.041 + 3012 & 21.93 & 15.66 & 26.95 & 23.743 & 0.129 & 0.591 & 0.138 & 0.947 & 0.083 + 3013 & 22.48 & 8.25 & 23.94 & 23.634 & 0.127 & 1.489 & 0.130 & 1.037 & 0.046 + 3014 & 22.63 & 22.07 & 31.61 & 23.169 & 0.086 & 1.763 & 0.086 & 1.072 & 0.013 + 3015 & 22.76 & 4.98 & 23.30 & 24.025 & 0.170 & 2.869 & 0.170 & 1.104 & 0.019 + 3016 & 26.85 & 13.34 & 29.98 & 22.612 & 0.065 & 1.119 & 0.066 & 1.044 & 0.017 + 3017 & 27.24 & 20.19 & 33.91 & 25.041 & 0.290 & 1.308 & 0.295 & 0.945 & 0.091 + 3018 & 28.12 & 16.76 & 32.73 & 24.268 & 0.175 & 2.635 & 0.175 & 1.140 & 0.016 + 3101 & 4.02 & 18.15 & 18.59 & @xmath337 & & @xmath338 & & 1.232 & 0.223 + 3102 & 7.04 & 11.73 & 13.68 & @xmath339 & & @xmath340 & & 1.435 & 0.289 + 3103 & 8.06 & 19.93 & 21.50 & @xmath341 & & @xmath342 & & 1.332 & 0.089 + 3104 & 9.35 & 18.19 & 20.45 & @xmath343 & & @xmath344 & & 1.157 & 0.134 + 3105 & 9.91 & 17.21 & 19.86 & @xmath345 & & @xmath346 & & 1.065 & 0.200 + 3106 & 11.68 & 8.12 & 14.22 & @xmath347 & & @xmath348 & & 1.020 & 0.198 + 3107 & 14.05 & 23.59 & 27.46 & @xmath349 & & @xmath350 & & 1.314 & 0.299 + 3108 & 14.11 & 20.36 & 24.77 & @xmath292 & & @xmath351 & & 0.995 & 0.203 + 3109 & 15.38 & 5.77 & 16.43 & @xmath352 & & @xmath353 & & 1.044 & 0.088 + 3110 & 15.86 & 19.23 & 24.92 & @xmath354 & & @xmath355 & & 1.293 & 0.178 + 3111 & 18.06 & 3.84 & 18.47 & @xmath356 & & @xmath357 & & 1.053 & 0.148 + 3112 & 18.91 & 6.18 & 19.90 & @xmath358 & & @xmath359 & & 1.291 & 0.137 + 3113 & 19.08 & 20.78 & 28.21 & @xmath360 & & @xmath361 & & 1.240 & 0.210 + 3114 & 19.17 & 13.22 & 23.29 & @xmath362 & & @xmath363 & & 1.099 & 0.088 + 3115 & 20.34 & 11.55 & 23.39 & @xmath364 & & @xmath365 & & 1.321 & 0.059 + 3116 & 20.72 & 15.91 & 26.13 & @xmath366 & & @xmath367 & & 1.574 & 0.230 + 3117 & 21.77 & 17.85 & 28.15 & @xmath368 & & @xmath369 & & 1.329 & 0.038 + 3118 & 22.10 & 5.69 & 22.82 & @xmath370 & & @xmath371 & & 0.999 & 0.229 + 3119 & 22.53 & 7.63 & 23.79 & @xmath372 & & @xmath373 & & 1.478 & 0.208 + 3120 & 24.37 & 20.00 & 31.53 & @xmath374 & & @xmath375 & & 1.305 & 0.124 + 3121 & 25.02 & 21.20 & 32.79 & @xmath376 & & @xmath377 & & 1.242 & 0.065 + 3122 & 25.96 & 14.92 & 29.94 & @xmath378 & & @xmath322 & & 1.249 & 0.088 + 3123 & 26.12 & 15.91 & 30.58 & @xmath379 & & @xmath380 & & 1.268 & 0.097 + 3124 & 26.41 & 23.34 & 35.25 & @xmath381 & & @xmath382 & & 1.230 & 0.137 + 3201 & 8.69 & 20.86 & 22.60 & 23.560 & 0.124 & @xmath383 & & & + 3202 & 18.15 & 23.79 & 29.93 & 24.307 & 0.182 & @xmath384 & & & + 3203 & 21.28 & 19.63 & 28.95 & 24.754 & 0.261 & @xmath385 & & & + 3204 & 22.76 & 11.15 & 25.34 & 24.872 & 0.306 & @xmath386 & & & + 3205 & 23.89 & 8.30 & 25.29 & 24.645 & 0.262 & @xmath387 & & & + 3206 & 25.09 & 16.86 & 30.22 & 24.984 & 0.299 & @xmath388 & & & crrcccccccccccc 4001 & @xmath389 & @xmath273 & 14.95 & & & 24.531 & 0.099 & & & 1.238 & 0.106 & 1.028 & 0.065 & 62 + 4002 & @xmath390 & @xmath391 & 13.96 & & & 24.356 & 0.076 & & & 1.145 & 0.086 & 0.947 & 0.066 & 56 + 4003 & @xmath392 & 9.20 & 15.50 & 23.849 & 0.224 & 24.265 & 0.084 & 1.634 & 0.225 & 2.055 & 0.086 & 1.140 & 0.026 & 60 + 4004 & @xmath393 & @xmath394 & 11.30 & & & 24.965 & 0.122 & & & 1.493 & 0.136 & 1.064 & 0.095 & 44 + 4005 & @xmath395 & 8.23 & 13.42 & 23.671 & 0.201 & 24.295 & 0.078 & 0.442 & 0.206 & 1.071 & 0.091 & 0.921 & 0.081 & 50 + 4006 & @xmath396 & @xmath397 & 10.33 & & & 24.938 & 0.121 & & & 0.832 & 0.169 & 1.345 & 0.154 & 39 + 4007 & @xmath398 & 7.56 & 10.23 & 23.321 & 0.147 & 23.866 & 0.047 & 1.681 & 0.148 & 2.231 & 0.049 & 1.064 & 0.023 & 26 + 4008 & @xmath399 & 2.31 & 6.28 & & & 24.605 & 0.096 & & & 1.975 & 0.107 & 1.082 & 0.069 & 19 + 4009 & @xmath400 & @xmath401 & 7.84 & 22.915 & 0.114 & 23.242 & 0.035 & 1.372 & 0.115 & 1.704 & 0.038 & 1.052 & 0.023 & 24 + 4010 & @xmath402 & 13.34 & 13.85 & & & 24.992 & 0.137 & & & 0.981 & 0.171 & 0.645 & 0.201 & 47 + 4011 & @xmath403 & 7.82 & 8.04 & 23.792 & 0.275 & 24.049 & 0.069 & 1.273 & 0.277 & 1.535 & 0.078 & 1.087 & 0.061 & 20 + 4012 & @xmath404 & 11.86 & 11.99 & 23.540 & 0.201 & 23.426 & 0.037 & 1.351 & 0.202 & 1.242 & 0.043 & 0.960 & 0.036 & 30 + 4013 & @xmath405 & 4.00 & 4.02 & & & 24.094 & 0.084 & & & 2.286 & 0.090 & 1.160 & 0.048 & 3 + 4014 & @xmath406 & 12.75 & 12.76 & & & 24.193 & 0.133 & & & 0.917 & 0.143 & 0.919 & 0.095 & 35 + 4015 & 0.96 & @xmath407 & 3.14 & 23.190 & 0.160 & 23.899 & 0.060 & 1.561 & 0.162 & 2.275 & 0.064 & 1.194 & 0.033 & 8 + 4016 & 3.41 & @xmath408 & 4.66 & & & 24.011 & 0.068 & & & 1.267 & 0.084 & 0.842 & 0.086 & 11 crrrrrrrrrrrrrrcc nuv-01 & @xmath409 & @xmath410 & 25.063 & 0.351 & & 24.581 & 0.227 & & 24.590 & 0.186 & & 24.527 & 0.215 & & 24.568 & 0.120 + nuv-02 & @xmath411 & @xmath412 & @xmath413 & & & @xmath414 & & & 23.763 & 0.117 & & @xmath415 & & & 23.763 & 0.117 + nuv-03 & @xmath416 & @xmath417 & @xmath418 & & & @xmath419 & & & 24.757 & 0.240 & & 24.875 & 0.293 & & 24.804 & 0.186 + nuv-04 & @xmath420 & @xmath421 & 23.799 & 0.122 & & 23.958 & 0.146 & & 24.153 & 0.139 & & 23.926 & 0.130 & & 23.947 & 0.067 + nuv-05 & @xmath422 & @xmath423 & 25.288 & 0.356 & & 25.053 & 0.265 & & 24.874 & 0.220 & & @xmath424 & & & 25.010 & 0.153 + nuv-06 & @xmath425 & @xmath426 & 25.008 & 0.226 & & 24.775 & 0.226 & & 25.475 & 0.293 & & 25.129 & 0.229 & & 25.054 & 0.120 + nuv-07 & @xmath427 & @xmath428 & & & & @xmath429 & & & 24.399 & 0.114 & & 25.346 & 0.266 & & 24.546 & 0.105 + nuv-08 & @xmath430 & @xmath431 & 24.175 & 0.132 & & @xmath432 & & & @xmath433 & & & @xmath434 & & & 24.205 & 0.132 + nuv-09 & @xmath435 & @xmath436 & 25.059 & 0.230 & & @xmath437 & & & @xmath438 & & & @xmath439 & & & 25.059 & 0.230 + nuv-10 & @xmath440 & @xmath441 & 25.022 & 0.293 & & 24.963 & 0.279 & & @xmath442 & & & 25.182 & 0.307 & & 25.049 & 0.169 + nuv-11 & @xmath443 & @xmath444 & 24.491 & 0.168 & & @xmath445 & & & @xmath446 & & & @xmath447 & & & 24.491 & 0.168 + nuv-12 & @xmath448 & @xmath449 & 24.165 & 0.136 & & @xmath450 & & & @xmath451 & & & @xmath452 & & & 24.165 & 0.136 + nuv-13 & @xmath453 & @xmath454 & @xmath455 & & & 24.543 & 0.240 & & @xmath456 & & & @xmath457 & & & 24.543 & 0.240 + nuv-14 & @xmath458 & @xmath459 & @xmath460 & & & @xmath461 & & & @xmath462 & & & 24.861 & 0.198 & & 24.861 & 0.198 + nuv-15 & @xmath463 & @xmath464 & @xmath465 & & & @xmath466 & & & @xmath467 & & & 24.871 & 0.234 & & 24.871 & 0.234 + nuv-16 & @xmath468 & @xmath469 & 25.338 & 0.350 & & & & & & & & 23.885 & 0.125 & & 23.885 & 0.125 + nuv-17 & @xmath470 & @xmath471 & @xmath472 & & & @xmath473 & & & @xmath474 & & & 24.736 & 0.207 & & 24.736 & 0.207 + nuv-18 & @xmath475 & @xmath476 & @xmath477 & & & @xmath478 & & & @xmath479 & & & 24.772 & 0.190 & & 24.772 & 0.190 + nuv-19 & @xmath480 & @xmath481 & @xmath482 & & & @xmath262 & & & @xmath467 & & & 24.700 & 0.174 & & 24.700 & 0.174 + nuv-20 & @xmath483 & @xmath484 & @xmath485 & & & @xmath486 & & & & & & 25.022 & 0.250 & & 25.022 & 0.250 + nuv-21 & @xmath487 & @xmath488 & 25.026 & 0.307 & & @xmath489 & & & @xmath490 & & & 24.894 & 0.268 & & 24.951 & 0.202 + nuv-22 & @xmath491 & @xmath492 & @xmath493 & & & @xmath494 & & & & & & 24.452 & 0.164 & & 24.452 & 0.164 lcc this study & @xmath495 & 4.26 + couture , harris , & allwright ( 1990 ) & @xmath496 & 5.05 + kundu & whitmore ( 1998 ) & @xmath440 & 4.72 + kissler - patig et al .
( 1998 ) & @xmath497 & 3.27 + harris et al .
( 2000 ) & @xmath498 & 5.88 llcccccccrcccc 104 ...... & 47 tuc & 4.64 & & 1.50 & & 3.80 & 0.02 & 1.10 & @xmath499 & @xmath500 & 0.137 & 0.014 & 0.04 + 362 ...... & & & & & & 1.89 & 0.08 & 0.95 & @xmath501 & @xmath502 & 0.088 & 0.014 & 0.05@xmath503 + 1261 ...... & & & & & & 1.89 & 0.08 & 0.91 & @xmath504 & @xmath505 & 0.069 & 0.014 & 0.01@xmath503 + 1851 ...... & & 2.29 & & 0.17 & & 2.11 & 0.08 & 0.99 & @xmath506 & @xmath507 & 0.082 & 0.014 & 0.02@xmath503 + 1904 ...... & m79 & 1.70 & & 0.28 & & 1.42 & & 0.89 & @xmath508 & @xmath509 & 0.051 & 0.014 & 0.01@xmath503 + 2298 ...... & & 2.36 & 0.07 & 0.51 & 0.06 & 1.86 & 0.09 & 0.94 & @xmath510 & @xmath511 & 0.034 & 0.014 & 0.14@xmath503 + 5024 ...... & m53 & 3.12 & 0.07 & 0.93 & 0.06 & 2.19 & 0.09 & 0.84 & @xmath512 & @xmath513 & 0.039 & 0.010 & 0.02@xmath503 + 5139 ...... & @xmath12 cen & 1.44 & 0.06 & @xmath514 & 0.10 & 1.74 & 0.08 & 0.91 & @xmath515 & @xmath516 & 0.048 & 0.014 & 0.12@xmath503 + 5272 ...... & m3 & 3.41 & 0.04 & 1.39 & 0.02 & 2.01 & 0.05 & 0.91 & @xmath517 & @xmath509 & 0.040 & 0.008 & 0.01 + 5897 ...... & & & & & & 1.77 & 0.08 & 0.93 & @xmath518 & @xmath488 & 0.037 & 0.014 & 0.09@xmath503 + 5904 ...... & m5 & 1.96 & & @xmath519 & & 2.01 & 0.08 & 0.91 & @xmath520 & @xmath521 & 0.067 & 0.010 & 0.03 + 6093 ...... & m80 & 2.18 & 0.22 & @xmath522 & 0.18 & 2.28 & 0.29 & 0.89 & @xmath523 & @xmath524 & 0.040 & 0.014 & 0.18@xmath503 + 6121 ...... & m4 & & & & & 2.89 & 0.08 & 0.98 & @xmath525 & @xmath526 & 0.084 & 0.014 & 0.36@xmath503 + 6205 ...... & m13 & 1.63 & 0.00 & 0.03 & 0.00 & 1.59 & 0.00 & 0.84 & @xmath512 & @xmath527 & 0.039 & 0.005 & 0.02 + 6266 ...... & m62 & 2.30 & 0.07 & @xmath528 & 0.07 & 2.59 & 0.10 & 1.01 & @xmath529 & @xmath530 & 0.075 & 0.014 & 0.47@xmath503 + 6341 ...... & m92 & 2.07 & 0.04 & 0.63 & 0.02 & 1.44 & 0.05 & 0.86 & @xmath531 & @xmath532 & 0.021 & 0.006 & 0.02 + 6356 ...... & & & & & & 4.00 & 0.12 & 1.11 & @xmath533 & @xmath534 & 0.169 & 0.008 & 0.28@xmath503 + 6388 ...... & & 3.27 & 0.22 & 0.29 & 0.16 & 2.98 & 0.27 & 1.02 & @xmath499 & @xmath535 & 0.146 & 0.006 & 0.37@xmath503 + 6402 ...... & m14 & & & & & 3.06 & 0.08 & 0.91 & @xmath536 & @xmath537 & 0.066 & 0.014 & 0.60@xmath503 + 6441 ...... & & 2.20 & 0.26 & 0.05 & 0.18 & 2.14 & 0.32 & 1.05 & @xmath538 & @xmath197 & 0.176 & 0.006 & 0.47@xmath503 + 6541 ...... & & 2.06 & 0.07 & 0.56 & 0.05 & 1.51 & 0.09 & 0.84 & @xmath539 & @xmath540 & 0.035 & 0.014 & 0.14@xmath503 + 6624 ...... & & & & & & 3.78 & 0.05 & 1.08 & @xmath541 & @xmath542 & 0.156 & 0.008 & 0.28@xmath503 + 6626 ...... & m28 & 1.92 & 0.41 & 0.12 & 0.29 & 1.81 & 0.50 & 0.92 & @xmath543 & @xmath544 & 0.092 & 0.014 & 0.40@xmath503 + 6637 ...... & & & & & & 3.59 & 0.05 & 1.08 & @xmath545 & @xmath546 & 0.151 & 0.005 & 0.16@xmath503 + 6681 ...... & m70 & 1.74 & 0.04 & 0.04 & 0.04 & 1.70 & 0.05 & 0.91 & @xmath547 & @xmath548 & 0.056 & 0.014 & 0.07 + 6715 ...... & m54 & 2.37 & 0.04 & 0.27 & 0.04 & 2.10 & 0.05 & 0.92 & @xmath549 & @xmath550 & 0.051 & 0.014 & 0.15@xmath503 + 6752 ...... & & 1.34 & 0.04 & @xmath551 & 0.02 & 1.52 & 0.05 & 0.89 & @xmath552 & @xmath553 & 0.052 & 0.014 & 0.04 + 6779 ...... & m56 & & & & & 2.08 & 0.05 & 0.92 & @xmath554 & @xmath555 & 0.030 & 0.014 & 0.20@xmath503 + 6864 ...... & m75 & 2.64 & 0.22 & 0.42 & 0.16 & 1.60 & 0.08 & 0.96 & @xmath556 & @xmath502 & 0.088 & 0.014 & 0.16@xmath503 + 7078 ...... & m15 & 2.36 & 0.07 & 0.67 & 0.05 & 1.69 & 0.09 & 0.73 & @xmath557 & @xmath558 & 0.023 & 0.007 & 0.10 + 7099 ...... & m30 & 2.71 & 0.04 & 0.83 & 0.04 & 1.88 & 0.05 & 0.82 & @xmath559 & @xmath560 & 0.024 & 0.014 & 0.03@xmath503 | to explore the production of uv - bright stars in old , metal rich populations like those in elliptical galaxies , we have obtained hst / stis far- and near - uv photometry of globular clusters in four fields in the ge galaxy m87 . to a limit of @xmath0
we detect a total of 66 globular clusters ( gcs ) in common with the deep hst optical - band study of kundu et al .
( 1999 ) . despite strong overlap in v- and i - band properties ,
the m87 gcs have uv / optical properties that are distinct from clusters in the milky way and in m31 .
m87 clusters , especially metal - poor ones , produce larger hot hb populations than do milky way analogues . in color plots including the nuv band , the m87 clusters appear to represent an extension of the milky way sequence .
cluster mass is probably not a factor in these distinctions .
the most metal - rich m87 gcs in our sample are near solar metallicity and overlap the local e galaxy sample in estimated mg@xmath1 line indices .
nonetheless , the clusters produce much more uv light at a given mg@xmath1 , being up to 1 mag bluer than any ge galaxy in @xmath2 color .
the m87 gcs do not appear to represent a transition between milky way - type clusters and e galaxies .
the differences are in the correct sense if the clusters are significantly older than the e galaxies .
comparisons with galactic open clusters indicate that the hot stars lie on the extreme horizontal branch , rather than being blue stragglers , and that the ehb becomes well populated for ages @xmath3 gyr .
existing model grids for clusters do not match the observations well , due to poorly understood giant branch mass loss or perhaps high helium abundances .
we find that 43 of our uv detections have no optical - band counterparts .
most appear to be uv - bright background galaxies , seen through m87 .
eleven nuv variable sources detected at only one epoch in the central field are probably classical novae .
two recurrent variable sources have no obvious explanation but could be related to activity in the relativistic jet . |
the foreign exchange market ( fx ) , with its daily volume of over five trillions usd in 2009 is by far the world s largest financial market .
any other financial market can hardly approach such volume .
this market connects international institutions participating in currency exchange transactions all across the world and encompasses essentially everything of what is going on in the world , first of all including economic factors , political conditions and market psychology , all of them constantly changing . also , this market has direct influence on all other markets because any price is expressed in terms of a currency .
the large volume makes it virtually impossible to control from outside and there is no friction ( transactions are basically commission free ) .
due to time differences fx transactions are performed 24 h a day , 5 days a week with maximum volume between 13:00 and 16:00 gmt , when both american and european markets are open . hence , the fx time series relations represent an exceptionally complex network indeed , and they therefore constitute an especially challenging target of a detailed quantitative analysis . in connection with the almost continuous trading
, fx is also much more effective and liquid than other speculative markets .
a significance of this market ( an example of globalization ) is even more important , since it became an indicator of a condition of the world s economy . from a physicist s viewpoint fx
is a complex system with extremely convoluted time dependencies .
the fx effectiveness is intensified by the correlations between exchange rates known as the triangular arbitrage .
it is possible only on small time scales and disappears immediately after taking advantage of inconsistent crossrates by the traders . to quantify such correlations we employed the multifractal analysis measuring nonlinear features of time series , in particular their multifractal spectra .
especially interesting is a relation between the fractal properties of the exchange rates remaining in the triangular dependency . encapsulating this
relation - especially of the empirical residuals within the triangle - may shed more light on this so far poorly understood issue .
the resulting scale free statistics encodes information about complex interactions in fx .
the fx data @xcite used in the present analysis include the following six indicative exchange rate pairs : usd / eur , eur / gbp , gbp / usd , jpy / gbp , gbp / chf , and chf / jpy sampled with 1 minute frequency over the period from 21:00 on january 2 , 2004 , until 21:00 on march 30 , 2008 ( 1183 days , 169 weeks ) .
the selection of currency pairs and time interval was constrained by availability of the sufficient quality data , so not all important pairs could be included in our analysis . in a consequence , we deal with six synchronized time series of length @xmath2 that can be labeled as @xmath3 , @xmath4 , where @xmath5 , denotes a value of currency @xmath6 expressed in terms of a currency @xmath7 at time @xmath8 . consequently , the corresponding returns over the time period @xmath9 are expressed as @xmath10 let us define residual returns as @xmath11 which are expected to fulfill the following relation : @xmath12 departures from ( [ triangle ] ) generate the so - called triangular arbitrage opportunities that , whenever detected , may be exploited and in fact are commonly used for the risk free profit generation . in the contemporary markets an execution of the corresponding operation typically takes at most a few seconds @xcite and is thus far below the scales of 1 min . considered here .
of course , this does not yet imply that returns of the corresponding three exchange rates synchronously evaluated at larger time scales @xmath13 obey eq .
( [ triangle ] ) exactly .
viewed at the same instant of time some mismatch may result just from the time needed ( a few seconds ) to execute the arbitrage opportunity and this introduces some dispersion .
it may also reflect some noise component involved .
clearly , on the larger time scales such effects of departure from zero in eq .
( [ triangle ] ) become less and less relevant relative to the total return . for the two exchange rate triangles , eur - gbp - usd and gbp - chf - jpy , that operate within the six currency exchange rates considered here , the logarithmic returns defined by eq .
( [ g ] ) are shown in fig . 1 for all the three exchange rates involved in each triangle accompanied by the sum of these three returns ( the left hand side of eq . ( [ triangle ] ) ) .
these characteristics are presented for the shortest time lag here accessible of @xmath14 min ( left panels ) and for @xmath15 min ( right panels ) .
as far as the magnitude of the fluctuations is concerned one sees essentially no sizeable difference for the four series so generated for @xmath14 min .
the only visibly detectable difference is in the structure of fluctuations ; the sum of the three returns ( lowest row in each panel ) looks more uniform in each case and large fluctuations are less frequent .
the situation changes considerably for @xmath15 min .
what is natural , the magnitude of returns for the individual exchange rates significantly increases while , at the same time , their sum in each triangle decreases even relative to @xmath14 min .
what seems also worth pointing already at this stage is that the fluctuations of returns in the eur / gbp exchange rate look more quiet than for the remaining two pairs in this triangle for both time lags of 1 and 60 min .
the same applies to the chf / gbp exchange rate in the second triangle .
irrespective of such details , the relation expressed by eq .
( [ triangle ] ) definitely introduces a crucial factor that affects the dynamics of the currency exchange network in multiple ways .
it first of all sets constraints on the dynamics by effectively reducing the number of independent degrees of freedom . for @xmath16 currencies instead of @xmath17 there in fact exists @xmath18 independent exchange rates .
this crucially shapes topology of the corresponding exchange rate network structure @xcite .
furthermore , some of the exchange rates may be primarily driven by the trade needs or some speculation specific arguments while dynamics of the others may be affected more by the market adjustments towards eliminating or at least reducing the arbitrage opportunities . in the following
we analyze a few most informative statistical characteristics of time series for the exchange rates listed above .
these characteristics determine the sectional organization of the paper . in each section
some novel results not discussed previously in literature are presented . in section 2 we show that although the distribution of returns for the individual exchange rates can be approximated ( similar to the returns from other financial markets ) by the @xmath0-gaussian distributions , the residual signals @xmath19 clearly do not .
section 3 is devoted to an analysis of temporal correlations and detection of repeatable patterns of market activity .
we found that exchange rates exhibit different temporal correlation properties depending on the trading significance of a particular currency pair .
next , in section 4 we study multifractal properties of the exchange rates in terms of the singularity spectra and , for the first time , identify their anomalous structure : negative exponents @xmath20 and negative @xmath1 .
finally , section 5 deals with the cross - correlation structure of the currency triangles , documenting occurrence of the epps effect , i.e. increase of coupling strength between the exchange rates with increasing sampling time @xmath9 , observed for surprisingly long @xmath9 .
one of the most relevant quantitative characteristics of the financial dynamics is the functional form of the distribution of returns .
the related - in the past well identified stylized fact - is the so - called inverse cubic power - law @xcite which applies to developed stock markets @xcite , to some emerging stock markets like the polish market @xcite , to the commodity market @xcite , as well as to the most traded currency exchange rates @xcite in the early 1990s .
of course , this type of distribution is lvy - unstable and thus for the sufficiently long time lags @xmath9 the returns distribution is expected to converge towards a gaussian .
this convergence , and thus departures from the inverse cubic power - law , has been found to be very slow as a function of @xmath9 . in the more recent data
, however , this convergence appears @xcite significantly faster and departures from the inverse cubic power - law in the contemporary stock markets can be seen already for @xmath14 min . a formalism that appears @xcite attractively compact and economic for describing the two extremes -
the inverse cubic as well as the gaussian distributions - including all the intermediate cases is the one based on the generalized nonextensive entropy @xcite .
accordingly , optimization of the corresponding generalized entropic form under appropriate constraints @xcite yields the following @xmath0-gaussian form for the distribution of probabilities @xmath21 where the constants @xmath22 , @xmath23 , @xmath24 and the @xmath0-exponential function @xmath25 are defined in appendix . in order to attain a better stability in this kind of the analysis
, we prefer to use cumulative form of the distribution ( [ px ] ) - see appendix for the corresponding formulae . as a standard procedure that makes the distributions for different exchange rates
directly comparable , we convert @xmath26 of eq .
( [ g ] ) into the normalized returns @xmath27 defined as @xmath28 where @xmath29 is the standard deviation of returns over the period @xmath30 . the empirical cumulative distributions for all the exchange rates considered here versus their best fits in terms of the @xmath0-gaussians ( eq . ( [ pcx ] ) ) are shown in fig . 2 .
the left column corresponds to the three exchange rates form the eur - gbp - usd triangle while the right column to the exchange rates from the gbp - chf - jpy triangle .
as one can see in all the cases the @xmath0-gaussians provide a very reasonable representation over the whole span of fluctuations and for the increasing return time - lags @xmath9 of 1 min .
, 10 min . and
some asymmetry between the left and right wings in the distributions , as expressed by the slightly different values of the corresponding @xmath0-parameter can be detected .
as expected , with the increasing time - lags @xmath9 the @xmath0-values decrease which reflects an expected ( slow ) convergence to the gaussian ( @xmath31 ) distribution . among the pairs considered here one interesting difference in this respect can be detected however : the decrease of the @xmath0 values with increasing @xmath9 is slower for the intra - european exchange rates ( eur / gbp and gbp / chf ) than for the intercontinental ones .
the convergence towards a gaussian distribution is thus slower in the former case .
this result is qualitatively similar to an earlier observation based on data from 1992 - 93 ( i.e. long before the introduction of euro ) , which was documented in ref .
this indicates that the global forex market is largely stable as regards the statistical distribution of returns .
moreover , the degree of convergence towards the gaussian distribution appears to behave similarly as in the contemporary stock markets @xcite .
it also needs to be noticed that analogously as the recent s&p500 analysis shows @xcite a slight departure from the inverse cubic power law ( which in the present formalism of the @xmath0-gaussians corresponds to @xmath32 ) takes place already for @xmath14 min .
perhaps this law is approached more accurately only for the time - lags even smaller than 1 min .
it is interesting to look at the distributions of the residual returns as defined by eq .
( [ triangle ] ) , since - according to our knowledge - they have not yet been shown in literature .
3 shows such distributions for the same time scales : @xmath14 , 10 and 60 min .
the @xmath0-gaussians are now able to fit only the central part of the distributions up to about three mean standard deviations .
the corresponding @xmath0-values only weakly decrease with @xmath9 starting from @xmath33 for @xmath14 min .
the tails of the empirical distributions are disparately thicker than expected by the @xmath0-gaussian model and become even more such for larger @xmath9 .
such an effect is in fact visible already in fig . 1 . the background fluctuations in their lowest panels
sizeably decrease with increasing @xmath9 while at the same time the largest fluctuations remain of the same order as compared to the individual exchange rates shown in the upper panels .
these outliers may reflect a longer time needed to balance departures from eq .
( [ triangle ] ) resulting from sudden large returns in one of the pairs in the triangle .
these characteristics are very similar in both the triangles considered here .
the issue of the character of temporal correlations is equally important in the financial context and many related questions still remain open .
the simplest measure of the temporal correlations is in terms of the autocorrelation function @xmath34 which for a function @xmath35 is defined as @xmath36 where @xmath37 denotes an average over @xmath38 .
the most studied cases in the financial context correspond to the autocorrelation of returns , here represented by @xmath39 , and of the volatility which can be defined as the modulus @xmath40 of returns .
4 shows the return autocorrelation as a function of @xmath41 , in the upper panel for the exchange rates from the eur - gbp - usd triangle and in the lower panel for the exchange rates from the gbp - chf - jpy triangle .
similarly as for the typical stock market returns , in fact even faster because in all these six cases already for @xmath42 min .
, such an autocorrelation is seen to assume values at the null level .
this even faster disappearance of the fx return autocorrelation is probably related to a higher liquidity involved .
characteristic for all the cases considered here is appearance of the correlation hole @xcite of negative autocorrelation for @xmath43 min . which reflects some anti - persistent tendencies on such short time scales .
there is no unique explanation of this effect first mentioned in ref .
@xcite ; it can originate from the divergent opinions of traders about the direction of imminent price changes as well as from certain actions of market - makers and banks @xcite .
the depths of these holes is different for different pairs even within the same triangle . for the pairs that can be considered leading in the fx dynamics ( gbp / usd , usd / eur , and jpy / gbp ) this depth at @xmath43 min
can be seen to be smaller than for the other ( eur / gbp , gbp / chf , chf / jpy ) pairs .
the volatility autocorrelations for the same six pairs of currencies are shown in fig .
it is quite obvious , due to the log - log scale used in this figure , that their behavior can well be approximated by the power law time dependance @xmath44 with @xmath45 .
this value of the scaling index does not differ from the value which is typical for a majority of stock markets @xcite .
even more , all the volatility autocorrelations between events that are separated by more than about @xmath46 basic units ( 1 min . ) suddenly drop down and start oscillating between the positive and negative values ( not visible in this figure ) with a decreasing amplitude .
this effect has recently been found also for stock markets in @xcite .
its natural interpretation is that the so - determined time horizon of the power - law volatility autocorrelations corresponds to an average length of either low or high volatility clusters .
as far as the fx exclusively characteristics are concerned , one more effect needs to be pointed out based on fig .
5 . for the same pairs ( gbp / usd , usd / eur , jpy / gbp ) that above have been indicated as the leading ones in the fx dynamics , the volatility autocorrelation is systematically stronger than for the remaining pairs ( as the relative location of the corresponding lines shows ) . a somewhat more advanced method to quantify the character of financial temporal correlations is to use a variant of the correlation matrix . in this approach , initiated in ref .
@xcite , the entries of the corresponding matrix are the correlation coefficients between the time series of returns representing different disconnected time - intervals , like the consecutive trading days or weeks .
the structure of eigenvalues and eigenvectors of such a matrix allows then to quantify further characteristics of the temporal correlations .
suppose from the time series @xmath47 of length @xmath30 one extracts @xmath48 disconnected series @xmath49 @xmath50 of length @xmath51 .
of course , the condition @xmath52 has to be fulfilled . by using such time series
as rows one forms an @xmath53 matrix @xmath54 .
then , the correlation matrix is defined as @xmath55 , where @xmath56 is matrix transpose . by diagonalizing @xmath57 : @xmath58 one obtains the eigenvalues @xmath59 @xmath60 and the corresponding eigenvectors @xmath61 . in the limiting case of entirely random correlations ,
the density of eigenvalues @xmath62 is known explicitely @xcite , and reads : @xmath63 where @xmath64 with @xmath65 , @xmath66 , and where @xmath67 is equal to the variance of the time series . for a better visualization
, each eigenvector can be associated with the corresponding time series by the following expression : @xmath68 thus , these new time series form orthogonal components into which the original signal @xmath49 is decomposed .
they reflect distinct patterns of oscillations common to all the time intervals labeled with @xmath69 .
these time series can therefore be called the eigensignals .
the above methodology on the weekly basis is now applied to the present fx data . our original time series of returns comprise @xmath70 complete weeks counted from sunday 21:00 until friday 22:00 gmt .
the length is @xmath71 min . for each pair of currencies and for the residual signals ,
the distributions of matrix elements are displayed in fig .
as it can be seen , for a majority of the rates the empirical distributions are gaussian - like . only for the most heavily - traded pairs : usd / eur and gbp / usd @xmath72 has a significant number of non - gaussian entries .
eigenvalue density for the corresponding correlation matrices for all the six currency exchange rates discussed here , including the residual time series representing departures from the triangle rule , are shown ( histograms ) in fig .
the left panel corresponds to the eur - gbp - usd triangle and the right panel to the gbp - chf - jpy one . for comparison , the pure noise distribution - as prescribed ( eq . ( [ rho ] ) ) by the corresponding wishart ensemble of random matrices @xcite - is indicated by dashed lines .
as one can see , besides some small departures at the edges , the empirical eigenvalue distributions do not differ much from their pure noise counterparts .
this fact may indicate that the intra - week exchange rate behaviour does not involve any particularly significant repeatable patterns .
this observation applies more to the pairs within the gbp - chf - jpy triangle .
the most evident departures between the empirical and the theoretical distributions one observes for the usd / eur and gbp / usd exchange rates . in both these cases ,
the two largest eigenvalues stay visibly outside the noise range and thus may carry some system - specific information .
this is in agreement with the distributions of matrix entries shown in fig .
6 . more insight into this issue can be gained by looking at the corresponding eigensignals as defined by eq .
( [ eigensignal ] ) . for each exchange rate
four such eigensignals are shown in fig . 8 : the ones corresponding to the first three largest eigenvalues and the fourth one which corresponds to an eigenvalue that is embedded deeply @xmath73 in the spectrum .
indeed , the cases of the gbp / usd and of the usd / eur exchange rates look most spectacular . in both cases the first two eigensignals and even the third one display large outstanding fluctuations that by a factor of even about 50 surpass the neighbouring ones .
their presence documents an enhanced market activity systematically at the same instants of time during the consecutive weeks .
interestingly , such instants of time are concentrated more before the weekend than just after it .
such special hours are 13.30 gmt and others .
what is natural , the eigensignals from the bulk of the spectrum do not display such a kind of activity .
another fact that deserves special attention is the eur / gbp exchange rate .
even though belonging to the same triangle , the dynamics is equally smooth for all the eigensignals .
one may hypothesize that this indicates a different mechanism that governs dynamics of the exchange rates within this pair as compared to the gbp / usd and usd / eur .
it seems that in a world in which currency trade is dominated by gbp / usd and usd / eur , the complementary rate eur / gbp can in the first approximation be considered only as a spectator adjusting its value to the behaviour of the in - play rates .
this receives additional support from the fact , that on time scales longer than @xmath74 min .
considered here , the time series of eur / gbp has fluctuations resembling the ones corresponding to the other exchange rates . from the same perspective
, the dynamics within the gbp - chf - jpy triangle looks more tranquil . yet
, within the jpy / gbp and gbp / chf pairs one also sees the outlying fluctuations ( though relatively smaller than in the previous case ) at the well recognizable instants of time during the week period .
at present , the most compact frame to globally grasp the whole richness of structures and correlations as identified above is - if applicable - in terms of the multifractal spectra @xcite .
the presence of the long - range nonlinear power law temporal correlations , possibly accompanied by the non - gaussian character of fluctuations , constitute the necessary - likely not sufficient however - ingredients in this respect @xcite .
furthermore , by now there exists quite convincing collection of evidence @xcite that the financial dynamics often carries signatures of multifractality . in this section
we therefore examine the multifractal characteristics of all the exchange rates considered in the previous sections . the multifractal detrended fluctuation analysis ( mfdfa )
@xcite is the most efficient practical method to quantify multifractality in the financial time series @xcite . in mfdfa for an @xmath75 discrete signal
one starts with the signal profile @xmath76 , where @xmath77 denotes averaging over all @xmath78 s .
then one divides the @xmath79 into @xmath80 non - overlapping segments of length @xmath81 ( @xmath82 ) starting from both the beginning and the end of the signal ( @xmath83 segments total ) . for each segment a local trend is estimated by fitting an @xmath84th order polynomial @xmath85 , which is then subtracted from the signal profile .
for the so - detrended signal a local variance @xmath86 in each segment @xmath87 is calculated for each , from @xmath88 to @xmath89 , scale variable @xmath81 .
finally , by averaging @xmath86 over all segments @xmath87 one calculates the @xmath90th order fluctuation function : @xmath91^{r/2 } \bigg\}^{1/r } , \label{ffunction}\ ] ] where @xmath92
. the relevant power law behavior of the fluctuation function reads : @xmath93 where @xmath94 is a generalized hurst exponent . for simple fractals
if @xmath94 in addition depends on @xmath90 , the signal is multifractal
. then the singularity spectrum @xmath1 @xcite can be calculated as : @xmath96 + 1 , \label{falpha}\ ] ] where @xmath97 is the singularity strength .
equivalently , with the commonly used scaling exponent @xmath98 the singularity strength is expressed as @xmath99 in ref .
@xcite some results were shown concerning requirements to reliably determine the multifractal spectra such that potentially spurious effects are eliminated .
this in particular concerns the length of the time series and the quantitative characteristics of the temporal correlations that determine the size of the scaling intervals .
since , as shown above , the fx time series develop heavy tails , the range of the index @xmath90 needs to be appropriately narrow ; we consistently thus choose @xmath100 $ ] .
the detrending polynomial @xmath85 used is of the second order which , as usual in this kind of analysis , proves to be an optimal choice .
an example of @xmath101 for the gbp / usd returns is shown in fig .
9 . for the other exchange rate returns the overall @xmath101 picture looks qualitatively similar .
the scaling of the fluctuation functions @xmath101 is quite convincing and the scaling indices @xmath94 depend on @xmath90 .
this , as is shown in the inset to the same fig .
9 , results in a concave shaped @xmath102 which is characteristic to conventional multifractals .
the reference dashed line in this inset represents @xmath102 calculated from the randomized original time series of gbp / usd returns , i.e. , by randomly shuffling the data points - a procedure which entirely destroys the temporal correlations . from the corresponding linear dependence of @xmath102 one straightforwardly identifies a monofractal with all the strength concentrated at @xmath103 .
11.5 cm 11.5 cm singularity spectra @xmath1 calculated for all the exchange rates are presented in fig . 10 .
they are all multifractal with the widths ranging from ca .
0.15 ( usd / eur ) to ca . 0.25 ( gbp / chf , chf / jpy , eur / gbp ) with the maxima located at around @xmath104 , similarly as for the typical stock market cases .
the dispersion of the maxima of @xmath1 is however larger within the gbp - chf - jpy triangle than within eur - gbp - usd .
an even more interesting difference is seen in the shape of @xmath1 for different exchange rates .
majority of them develop an asymmetric @xmath1 with the distortion somewhat towards the shapes characteristic to bifractals @xcite . the most beautiful and symmetric shape - like the model binomial cascade @xcite
- is developed by the gbp / usd ( london - new york cable connection ) and to a lesser extent by the usd / eur exchange rates .
interestingly , these two are the leading and the most traded exchange rates and the previously seen graphs illustrating their return fluctuations look the most erratic among all the exchange rates considered here .
a trace of similar effect can be seen in the second , gbp - chf - jpy triangle . there
, the most symmetric shape of @xmath1 corresponds to the leading jpy / gbp exchange rate .
the degree of symmetry can also be seen to go in parallel with the strength of the volatility autocorrelation ( fig .
stronger volatility autocorrelation corresponds to the more symmetric shape of @xmath1 .
these effects to some extent resemble the situation encountered in the human electrocardiogram .
there the most healthy and at the same time the most erratic case generates the widest and the most symmetric singularity spectrum @xcite .
11.5 cm 11.5 cm in contrast to the proper exchange rate returns , we do not observe such conventionally interpretable multifractal characteristics for time series of the residual returns @xmath105 .
complexity of the processes underlying such signals can be assessed from the scale @xmath81-parameter dependence of the fluctuation function @xmath101 for different values of @xmath90 .
the result based on calculation within the same range of the parameters @xmath90 as before is shown in fig .
11a for the eur - gbp - usd triangle and in fig .
11b for the gbp - chf - jpy triangle . while for the individual values of @xmath90 the fluctuation functions @xmath101 clearly behave scale free to a similar accuracy as in fig . 9 , the @xmath90-dependence of the corresponding scaling indices @xmath94 is significantly different .
for the positive values of @xmath90 , with increasing their values , the slope ( in the log - log scale ) of @xmath101 systematically redirects its orientation such that @xmath94 becomes negative . for
the negative values of @xmath90 , on the other hand , the slope of @xmath101 - and thus @xmath94 - almost does not depend on @xmath90 which signals a monofractal character of the small fluctuations as these are predominantly filtered by the negative parameters @xmath90 .
the resulting @xmath102 is shown in the inset and can be seen to have a profoundly different functional form as compared to the original returns case of figs . 9 and 10 .
still , by randomly shuffling the residual returns time series one obtains the same monofractal form of @xmath102 ( dashed lines in the insets to fig .
11a and 11b ) as before when the series of returns were shuffled .
this signals that the presently observed anomalous functional form of @xmath102 for the residual returns @xmath105 is primarily encoded in the specific form of the temporal correlations in @xmath105 .
11.5 cm 11.5 cm the singularity spectra @xmath1 that correspond to the above two cases are shown in fig .
they develop essentially only the left wing which corresponds to the positive values of @xmath90 .
somewhat related `` left - sided '' multifractals have in fact already been considered in the literature @xcite in applications to diffusion limited aggregates ( dla ) and to fully developed turbulence .
this may signal further analogies between the fx dynamics and the physics of turbulence in accord and giving more arguments in favor of the conjecture put forward in ref . @xcite .
furthermore , @xmath1 extends to the negative values of the singularity exponents @xmath20 where , at the edge , @xmath1 even assumes the negative values . to our knowledge
such an anomalous form of multifractality has never been identified before in the context of the financial dynamics .
however even such a possibility appears to be implicitly involved in the mandelbrot considerations @xcite on the fluid dynamics and already explicitly in more recent statements on the issue of negative critical dimensions @xcite .
this latter extended study has been motivated by a rigorous demonstration @xcite of the presence of negative fractal quantities for the ( conformal invariant ) harmonic measure around a number of incipient percolation clusters .
a related indication is that the `` multifractal anomalies '' arise when the system under study behaves canonically - in the statistical physics sense - instead of microcanonically .
illustrated by means of the binomial cascade this extends to the situation such that the sum of partitions at each recursion is not preserved exactly but only in the average .
several quantitative characteristics seen above indicate that the dynamics associated with the constraint imposed by the fx triangle rule belongs to this category of phenomena . in the real fx dynamics
the triangle relation expressed by the eq .
[ triangle ] is obeyed also only in the average .
a complementary interpretation of the negative fractal dimensions is that they describe the missing fluctuations - therefore typically large and thus filtered by the positive parameters @xmath90 - in a studied finite size sample .
such fluctuations are thus expected to come into view in another realization of a finite size sample from the same ensemble .
it is well - known that time series of returns of different assets traded on the same market are typically cross - correlated .
this holds true also for the currency market @xcite . unlike other studies before ,
here we analyze a specific case of correlations between the exchange rates coupled by the triangle rule .
we expect that deviations from the perfect triangle relation can be observed not only by means of the residual returns distributions ( fig . 3 ) but also by means of the eigenvalue spectra of the correlation matrix calculated from the triples of time series corresponding to currency triangles .
we can exploit here the fact that our time series were recorded simultaneously .
we follow the same procedure of constructing a correlation matrix as above , but now we consider the complete time series of length @xmath2 . for a few different choices of the time scale @xmath9 , we create two matrices of size @xmath106 , each for one currency triangle . due to
( [ triangle ] ) , the triangle rule , if fulfilled , implies that @xmath72 has only two non - zero eigenvalues whose sum satisfies the condition : @xmath107 .
existence of @xmath108 would thus mean a possibility of a triangular arbitrage . 9.0 cm 9.0 cm 9.0 cm 9.0 cm in fig .
13 top panels show functional dependence of the eigenvalues of @xmath72 on @xmath9 for the two considered currency triangles . for the shortest @xmath74 min . , the data clearly does not comply with eq .
( [ triangle ] ) and @xmath72 has three non - zero eigenvalues in both cases . although possibilities of the triangular arbitrage , with the today s computer trading , are not expected to last longer than a fraction of a second , in the correlation matrix representation their trace can be seen clearly on much longer time scales .
this is because all the exchange rates were sampled precisely at the same time and , thus , the inconsistencies in exchange rates could not be consumed yet . since these inconsistencies ,
as regards their absolute magnitude , are the same no matter which time scale one considers , their relative magnitude should gradually decline with increasing @xmath9 ( and , therefore , with increasing variance of the unnormalized returns ) .
this effect should manifest itself by a declining value of @xmath109 . due to the fact that trace of @xmath72 is independent of @xmath9 , decreasing the level of @xmath110
must be associated with an increase of the two remaining eigenvalues .
this resembles the well - known epps effect observed on the stock markets @xcite .
in fact , even the time scales at which @xmath111 saturates ( 50 - 100 min . ) are roughly the same as those found for stocks .
the same effect can in general be observed for any triples of the exchange rates not neccessarily forming a triangle , as it is documented by two examples shown in bottom panels of fig . 13 , where the exchange rates are formed from 4 currencies and ,
thus , do not constitute any cycle .
the only quantitative difference between the eigenvalues of the triangles and the `` non - triangles '' is in the asymptotic magnitude of @xmath109 , which in the former case is zero and in the latter case is small but positive .
one therefore sees that the triangle rule implies that the fully - developed couplings among the involved exchange rates are associated with a zero mode of @xmath72 .
the origin of the epps effect in the forex market is likely to be similar to its counterpart for the stock market : a finite speed of information spreading among the assets .
one possible source is a lack of transaction synchronicity on different assets , which introduces noise - like effects on their correlated evolution @xcite .
however , this non - synchronicity of trading is probably not the unique cause of the epps effect in the currency market : the trading frequency on this market is much higher than its counterpart on the stock markets , yet the time scales of saturation are comparable on both market types .
this suggests that some other factor may play an important role in the development of correlations .
indeed , other sources of the epps effect have already been proposed and may be relevant here like the microstructure noise and the discretization error @xcite .
however , their influence in this case has yet to be assessed .
in this paper we analyzed time series of currency exchange rate returns for the two triples of currencies forming triangles : eur - gbp - usd and gbp - chf - jpy .
market efficiency requires that cycling through currencies in such triangles must not be profitable except very short time scales , which is reflected in the triangle rule . for the original fx time
series we find that the main statistical properties of the corresponding returns - their distributions , temporal correlations , and multifractality - are qualitatively similar to those found for other markets .
however , we found also some quantitative differences between properties of different exchange rates which may reflect their different significance in the world currency exchange system .
we also studied the residual signals consisting of short - time deflections from the perfect no - arbitrage condition .
a related interesting observation is that while the proper exchange rate returns are well modeled by the @xmath0-gaussian distributions , the residual returns develop disproportionately heavier tails . among the most illuminating views
is the one that can be obtained after diagonalizing the correlation matrices constructed from time series representing different weeks and calculating the corresponding eigensignals , i.e. independent components of dynamics associated with repeatable patterns of activity .
eigensignals carry system - specific information if one can identify large fluctuations which can be related to some periodic external perturbation of the market ( e.g. economical news releases ) .
it occurs that such fluctuations are clearly visible already on 1 min .
time scale for heavily - traded cross - rates like usd / eur and gbp / usd , but only on longer time scales for less frequently traded rates like eur / gbp .
the same refers to the second triangle , in which more popular rates : jpy / gbp and gbp / chf have more characteristic eigensignals at 1 min .
time scale , while the less popular chf / jpy rate has more universal ( noisy ) eigensignals .
we argued that this effect reflects the fact that less popular exchange rates play a passive role , tuning their values according to changes in dominant rates as demanded by the triangle rule . a parallel effect is related to the different shapes of the singularity spectra for different currency pairs .
in this respect the most symmetric @xmath1 spectrum is observed for the gbp / usd pair , while other currency pairs have spectra which are more asymmetric , especially those from the gbp - chf - jpy triangle .
even more intriguing are the signatures of negative singularity exponents and negative singularity spectra for the triangle residual returns .
this opens an exciting direction for further investigations towards perhaps establishing closer analogy between the fx dynamics and the phenomenon of turbulence .
we also found that some inefficiency of the market allowed for extremely short time scales leads to the emergence of the epps effect , i.e. an increase of couplings between different exchange rates from the same triangle , if going from shorter to longer time scales .
our result indicates that , from a point of view of returns , the influence of market inefficiency on cross - correlations among exchange rates can be neglected on time scales longer than , roughly , an hour .
an increase of coupling with similar characteristics time scales involved is in fact observed for any triples of exchange rates not necessarily forming a triangle .
the @xmath0-gaussian distribution is defined by @xcite : @xmath112 where @xmath113 } { \gamma \left [ \frac{2-q}{1-q } \right ] } \sqrt{\frac{1-q}{\pi } \mathcal{b}_{q } } & ~~{\rm for } & q<1 \\[3 mm ] \frac{\gamma \left(\frac{1}{q-1}\right)}{\gamma \left(\frac{3-q}{2 ( q-1)}\right ) \sqrt{\frac{\pi}{(q-1 ) \mathcal{b}_q } } } & ~~{\rm for } & 1<q<3 \end{array } \right . \\ \bar{\mu } _ { q}= \,\int x\frac{\ \left [ p\left ( x\right ) \right ] ^{q}}{\int \left [ p\left ( x\right ) \right ] ^{q}dx}\ dx\equiv \left\langle x\right\rangle _ { q } , \ \ \ \ \mathcal{b}_{q}=\left [ \left ( 3 - 1\right ) \,\bar{\sigma}_{q}^{2}\right ] ^{-1 } \nonumber\end{aligned}\ ] ] and
@xmath114 denotes the @xmath0-exponential function @xmath115 ^ { \frac{1}{1-q}}. \label{eq}\ ] ] for @xmath116 this distribution asymptotically @xmath117 develops a power law form @xmath118 .
in particular , for @xmath32 , on the level of the cumulative distribution , it recovers the inverse cubic power law .
this is a particularly useful aspect of the functional form expressed by eq .
( [ px ] ) because it at the same time provides a compact form for the probability distribution for any value of @xmath119 . instead of directly using eq .
( [ px ] ) it is more practical to convert it to the cumulative form by defining @xmath120 where the + and - signs correspond to the right and left wings of the distribution , respectively . by using eq .
( [ px ] ) one obtains @xmath121 where @xmath122 , @xmath123 , @xmath124 , @xmath125 and @xmath126 is the gauss hypergeometric function : @xmath127 | we present a systematic study of various statistical characteristics of high - frequency returns from the foreign exchange market .
this study is based on six exchange rates forming two triangles : eur - gbp - usd and gbp - chf - jpy .
it is shown that the exchange rate return fluctuations for all the pairs considered are well described by the nonextensive statistics in terms of @xmath0-gaussians .
there exist some small quantitative variations in the nonextensivity @xmath0-parameter values for different exchange rates ( which depend also on time scales studied ) and this can be related to the importance of a given exchange rate in the world s currency trade .
temporal correlations organize the series of returns such that they develop the multifractal characteristics for all the exchange rates with a varying degree of symmetry of the singularity spectrum @xmath1 however .
the most symmetric spectrum is identified for the gbp / usd .
we also form time series of triangular residual returns and find that the distributions of their fluctuations develop disproportionately heavier tails as compared to small fluctuations which excludes description in terms of @xmath0-gaussians .
the multifractal characteristics for these residual returns reveal such anomalous properties like negative singularity exponents and even negative singularity spectra .
such anomalous multifractal measures have so far been considered in the literature in connection with the diffusion limited aggregation and with turbulence .
studying cross - correlations among different exchange rates we find that market inefficiency on short time scales leads to the occurrence of the epps effect on much longer time scales , though comparable to the ones for the stock market .
although the currency market is much more liquid than the stock markets and it has much larger transaction frequency , the building - up of correlations takes up to several hours - time that does not differ much from what is observed in the stock markets .
this may suggest that non - synchronicity of transactions is not the unique source of the observed effect . |
situations where interference is known non - causally at the transmitter but not at the receiver model several useful multiuser communication scenarios . in @xcite , costa introduced and studied coding for such situations and called it `` writing on dirty paper '' . dirty paper coding ( dpc ) is now recognized as a powerful notion central to approaching capacity on multiuser channels .
lattice - based ideas for dpc were suggested and shown to be capacity - approaching in @xcite .
recently , many designs of lattice - based dpc schemes have been proposed in @xcite .
lattice - based schemes typically use cosets of a convolutional code for lattice - quantizing or shaping to minimize the energy of the difference of the coded symbols and the interfering symbols .
a part of the message bits is used to choose the specific coset used in the minimization .
in addition to the shaping convolutional code , an error correction code needs to be used to obtain coding gain and approach capacity .
the main source of complexity in lattice - based dpc designs is combining shaping and coding encoders / decoders at the transmitter / receiver .
simple concatenation schemes are not applicable because of the following reasons - outer shaping followed by inner coding results in unshaped parity symbols that increase transmitted energy , while outer coding followed by inner shaping results in a poor inner code that needs to be iteratively decoded at the receiver with the outer code . in @xcite , encoding is done on a combined trellis of the source code ( turbo tcq ) and a channel code ( turbo tcm ) . at the receiver , decoding is done for turbo tcm followed by syndrome computation to recover message bits .
the transmitter is complex in @xcite because of the use of the joint trellis .
the dpc method proposed in @xcite is similar to that of @xcite . in @xcite ,
multilevel coding is used , and there are different codes for different bits of the symbols . at the receiver , iterations have to be performed between decoders for some of the channel codes and the shaping decoder . in @xcite and @xcite , shaping follows channel coding and the receiver performs iterations between the shaping and channel decoders . in this work
, we propose a lattice - based method that uses a novel combination of a convolutional code for sign - bit shaping and a low density parity check ( ldpc ) code for channel coding .
as shown in specific designs and simulations , the method provides good shaping and coding gains at moderate complexity .
the main idea for reducing complexity at the receiver is the introduction of a one - codeword delay at the transmitter , and the shaping of symbols from current message bits combined with parity bits from the previous codeword .
this enables the ldpc decoder to be run first at the receiver ( with a one - codeword delay ) without any need for iterations with a shaping decoder . as an application ,
we use the proposed dpc method to design codes for superposition coding in two - user gaussian broadcast channels . by simulations , we show that rate points outside the time - sharing region are achieved . the rest of the paper is organized as follows . after a brief review of the lattice - based dpc coding method in section [ cls ] , we present the proposed dpc method in section [ ps ] .
this is followed by description and simulation of specific designs of dpc codes in section [ sr ] . in section [ bc ] , design of a superposition scheme using the proposed dpc method is described and simulation results are presented . concluding remarks are made in section [ conc ] .
in a gaussian dirty paper channel , the received symbol vector @xmath0 $ ] is modeled as @xmath1 where @xmath2 $ ] denotes the transmitted vector , @xmath3 $ ] denotes the interfering vector assumed to be known non - causally at the transmitter and @xmath4 denotes the additive gaussian noise vector .
the transmit power is assumed to be upper - bounded by @xmath5\leq p_x$ ] per symbol , and the interference power is denoted @xmath6= p_s$ ] per symbol . the noise variance per symbol is denoted @xmath7 . in @xcite , costa shows that the capacity of the dirty paper channel is @xmath8 i.e. known interference can be canceled perfectly at the transmitter .
the interfering vector @xmath9 is used as an input in the encoding process and plays an important role to determine a suitable transmit vector @xmath10 . a coding strategy for choosing
@xmath10 needs to overcome the imminent addition of @xmath9 and protect the transmitted information from the addition of the noise @xmath4 .
such coding strategies are called dirty paper coding ( dpc ) methods . in @xcite , a dirty paper coding ( dpc )
scheme based on lattice strategies was proposed and shown to achieve the capacity of the dirty paper channel .
we follow @xcite for a brief review of the transmitter and receiver structure in the lattice dpc method @xcite .
let @xmath11 denote an @xmath12-dimensional lattice with fundamental voronoi region @xmath13 having averaged second moment @xmath14 and normalized second moment @xmath15 . also let @xmath16 i.e. @xmath17 is a random variable ( dither ) uniformly distributed over @xmath13 .
the lattice transmission approach of @xcite@xcite is as follows .
* _ transmitter _ : the input alphabet @xmath18 is restricted to @xmath13 . for any @xmath19 , the encoder sends @xmath20\text{mod}\ \lambda,\ ] ] where @xmath21 is a mmse scaling factor @xcite . *
_ receiver _ : the receiver computes @xmath22\text{mod}\ \lambda\ ] ] the channel from * _ v _ * to @xmath23 defined by ( 1 ) and ( 2 ) is equivalent in distribution to @xmath24\text{mod}\ \lambda{,}\ ] ] where @xmath25\text{mod}\ \lambda.\ ] ] lower bounds on achievable rates for the above equivalent channel is shown in @xcite to be equal to @xmath26 for optimal shaping , @xmath27 and we approach capacity of the dirty paper channel .
note that the dither is assumed to be known at the transmitter and receiver ( say , through the use of a common seed in a random number generator ) .
the proposed scheme uses a convolutional code for sign - bit shaping @xcite and low density parity check ( ldpc ) codes for channel coding . we assume a @xmath28-pam signal constellation with a carefully chosen bit - to - symbol mapping that is compatible with sign - bit shaping and bit - interleaved coded modulation ( bicm ) @xcite . for @xmath29 , the constellation and mapping
are shown in fig .
[ 16pam ] .
the mapping in fig .
[ 16pam ] is suited for sign - bit shaping , since a flip of the most significant bit results in a significant change in symbol value for all possible 4-bit inputs . also , the mapping is mostly gray except for a few symbol transitions .
gray mapping is known to be the most effective mapping for bicm with ldpc codes .
this heuristic choice of mapping enables the possibility of good shaping and coding gains to be obtained simultaneously .
as expected , larger values of @xmath28 will result in larger shaping gains in our design , and we stick to the 16-pam shown in fig .
[ 16pam ] for illustration and simulation .
the encoder structure for the proposed scheme is as shown in fig .
[ encoder ] .
we describe the operations in the encoder at time step @xmath30 or in the @xmath30-th block .
a @xmath31-bit message @xmath32}$ ] is encoded into a @xmath33-symbol vector @xmath34}$ ] from the @xmath28-pam constellation @xmath35 , where @xmath36 is assumed to be an integer .
let @xmath37 and let @xmath38 denote the bit - to - symbol mapping .
the bits that map to the @xmath39-th symbol are denoted @xmath40 ; the sign - bit vector is denoted @xmath41}$ ] , and we define vectors @xmath42}$ ] for @xmath43 . finally , we have @xmath44 , where @xmath45 operates component - wise on a vector input .
let us assume that the vectors @xmath46 , @xmath43 are available at the encoder .
the sign - bit shaping convolutional code is used to determine the sign - bit vector @xmath47 as follows .
a part of the message @xmath48 $ ] with @xmath49 bits is mapped to a coset leader of the convolutional code using an inverse syndrome former @xcite .
note that we need the rate of the convolutional code to be @xmath50 .
let the coset chosen by @xmath51 be denoted @xmath52 .
the sign - bit vector @xmath47 is chosen from @xmath52 so as to minimize the squared sum ( energy ) of the vector @xmath53 , where @xmath21 is the mmse factor and @xmath9 is the interference vector .
that is , @xmath54 the minimization in ( [ eqn : min ] ) is implemented using the viterbi algorithm @xcite .
the @xmath46 , @xmath43 are determined as follows .
an @xmath55 ldpc code is used at the encoder with a systematic encoder @xmath56 .
let @xmath57 $ ] be input to the systematic ldpc encoder to obtain the codeword @xmath58 $ ] , where @xmath59 is the parity - bit vector for the @xmath30-th block .
the parity - bit vector is delayed by one time step . for the @xmath30-th block , the @xmath60 bits in @xmath61 $ ] and @xmath62 are rearranged by a permutation @xmath63 to form the vectors @xmath46 , @xmath43 .
this permutation is necessary in an implementation of bicm @xcite .
note that both the shaping and coding objectives have been met at the encoder .
the transmitted symbols @xmath64 have minimal energy in the lattice defined by sign - bit shaping using the convolutional code . selected bits in successive blocks of symbols form codewords of the ldpc code . in summary ,
the encoder structure achieves dpc shaping and ldpc coding with bit - interleaved modulation .
the decoder for the proposed scheme is as shown in fig.[decoder ] .
the demapper computes log likelihood ratios ( llrs ) for the bits from the received symbols in @xmath65 .
the llrs of the @xmath66 message bits after a delay of one time step , and the llrs of the @xmath67 parity bits are de - interleaved .
the @xmath68 llrs of the sign bits after a delay on one time step , and the @xmath69 output llrs of the de - interleaver are given as the input to the ldpc decoder .
the ldpc decoder outputs @xmath70 message bits and @xmath33 bits of the sign bit vector of the previous block .
now , the @xmath33-bit sign vector is passed through the syndrome former to recover the remaining @xmath71 message bits .
the demapper function at the receiver has to calculate llrs taking into account the modulo @xmath28 operation at the encoder @xcite .
therefore , the received constellation @xmath72 is a replicated version of the @xmath28-pam constellation @xmath73 used at the transmitter ( assuming that scaling factors have been corrected at the receiver ) .
that is , @xmath74 the number of replications @xmath75 is chosen so that the average power of @xmath72 is approximately equal to the total average power @xmath76 , and the bit mapping of the symbol @xmath77 @xmath78 is the same as that for @xmath79 .
the llr for the @xmath39-th bit in the @xmath80-th symbol @xmath81 is computed according to the constellation @xmath72 using the following formula : @xmath82 since the constellation mapping is nearly gray , iterations with the demapper do not provide significant improvements in coding gain , particularly for large @xmath28 .
for simulations , we have taken @xmath83 , @xmath84 , @xmath85 with @xmath29 ; this results in @xmath86 .
the constellation mapping is as given in fig .
[ 16pam ] .
we have chosen a rate-1/2 memory 8 ( 256 state ) non - systematic convolutional code with generator polynomials @xmath87 as the sign - bit shaping code .
a non - systematic convolutional code is used to avoid error propagation problems .
a randomly constructed irregular ldpc code ( 40000 , 35000 ) of rate @xmath88 with variable node degree distribution : @xmath89 and check node degree distribution @xmath90 is used as the channel code .
the overall rate of transmission is seen to be @xmath91 bits per channel use .
[ dpc ] shows ber plots over an awgn channel and a dpc channel with interference known at the transmitter .
the interfering vector was generated at random for different power levels .
the plot with interference did not change appreciably for all power levels of interference , and we have provided one plot for illustration .
we see that a ber of @xmath92 is achieved at a snr of 19.45db with interference , and at a snr of 19.33 db without interference .
we have simulated 1000 blocks of length 40000 to obtain sufficient statistics for a ber of @xmath92 . the awgn capacity at an snr of @xmath93 db for a rate of 3 bits per channel use .
this shows that we are 1.46 db away from ideal dirty paper channel capacity .
the granular gain @xmath94 is computed from the simulations to be 1.282db @xcite , where @xmath95 is the rate before channel coding , and @xmath96 is the transmit power ( obtained through simulations ) . from this , the shaping loss is calculated as follows : @xmath97 so , of the total gap of 1.46 db , we have a shaping gap of 0.2548db , and a coding gap of 1.2052db to capacity .
we use the proposed scheme for superposition coding in a two - user gaussian broadcast channel @xmath98 and @xmath99 with @xmath100 .
we let @xmath101 and @xmath102 , where @xmath103 is the total transmit power . here , user 2 is coded using dpc considering user 1 as interference .
user 1 is shaped using sign - bit shaping and coded using an ldpc code over @xmath28-pam .
[ hbc ] shows a block diagram of the transmitter and receivers .
the encoder structure for user 1 is as in fig .
[ encoder ] with the interference vector @xmath104 .
hence , for user 1 , the shaping coder minimizes the energy of @xmath105 .
the demapper at receiver 1 calculates llr for the @xmath39-th bit in the @xmath80-th receiver symbol @xmath106 using the following formula .
@xmath107 where @xmath108 for @xmath109 represents the _ a priori _ probability of the @xmath28-pam symbol @xmath79 . at the receiver , we approximate @xmath110 using a gaussian distribution with variance @xmath111 assuming that the distribution of @xmath28-pam symbols is approximately gaussian .
we simulated a two user degraded broadcast channel with @xmath112 and @xmath113 using the proposed scheme with parameters from section [ sr ] .
the total transmit power , power for user 1 and power for user 2 required for a bit error rate of @xmath92 ( at both receivers ) are estimated from the simulation and denoted @xmath103 , @xmath114 and @xmath115 , respectively .
the snr for receiver 1 is computed as @xmath116 db . since dpc is done for user 2 , the effective snr at receiver 2 is computed as @xmath117 db . comparison with the snr needed for a single user capacity of 3 bits per channel use ( which is 17.99 db ) shows that the total loss for both the users is about 2.4642db .
[ 256state1 ] shows the ( 3 , 3 ) rate pair in the capacity region of the two - user gaussian broadcast channel with total transmit power @xmath103 and noise power @xmath118 , @xmath119 , which is defined by @xmath120 , @xmath121 for @xmath122 .
we see that the ( 3,3 ) rate point is clearly outside the time - sharing region .
in this work , we have proposed a method for designing lattice - based schemes for dirty paper coding using sign - bit shaping and ldpc codes .
simulation results show that the proposed design performs 1.46db away from the dirty paper capacity for a block length of @xmath83 at the rate of 3 bits / channel use .
this performance is comparable to other results in the literature .
however , as discussed in this article , a novel method for combining shaping and coding results in good gains at lesser complexity in our design , when compared to other lattice - based strategies . as an application
, we have designed a superposition coding scheme for gaussian broadcast channels that is shown to perform better than time - sharing through simulations . out of the 1.46 db gap to capacity ,
about 1.2 db is gap attributed to a sub - optimal choice of the ldpc code . optimizing the ldpc code will require use of genetic algorithms and asymmetric density evolution @xcite , which are topics for future work .
t. philosof , u. erez , and r. zamir , `` combined shaping and precoding for interference cancellation at low snr , '' in _ information theory , 2003 . proceedings .
ieee international symposium on _
, june-4 july 2003 , pp .
68. y. sun , y. yang , a. liveris , v. stankovic , and z. xiong , `` near - capacity dirty - paper code design : a source - channel coding approach , '' _ information theory , ieee transactions on _ , vol .
55 , no . 7 , pp .
30133031 , july 2009 .
y. sun , a. liveris , v. stankovic , and z. xiong , `` near - capacity dirty - paper code designs based on tcq and ira codes , '' in _ information theory , 2005 .
isit 2005 . proceedings .
international symposium on _ , sept .
2005 , pp . 184188 . | dirty paper coding ( dpc ) refers to methods for pre - subtraction of known interference at the transmitter of a multiuser communication system .
there are numerous applications for dpc , including coding for broadcast channels .
recently , lattice - based coding techniques have provided several designs for dpc . in lattice - based dpc
, there are two codes - a convolutional code that defines a lattice used for shaping and an error correction code used for channel coding .
several specific designs have been reported in the recent literature using convolutional and graph - based codes for capacity - approaching shaping and coding gains . in most of the reported designs ,
either the encoder works on a joint trellis of shaping and channel codes or the decoder requires iterations between the shaping and channel decoders .
this results in high complexity of implementation . in this work
, we present a lattice - based dpc scheme that provides good shaping and coding gains with moderate complexity at both the encoder and the decoder .
we use a convolutional code for sign - bit shaping , and a low - density parity check ( ldpc ) code for channel coding .
the crucial idea is the introduction of a one - codeword delay and careful parsing of the bits at the transmitter , which enable an ldpc decoder to be run first at the receiver .
this provides gains without the need for iterations between the shaping and channel decoders .
simulation results confirm that at high rates the proposed dpc method performs close to capacity with moderate complexity . as an application of the proposed dpc method
, we show a design for superposition coding that provides rates better than time - sharing over a gaussian broadcast channel . |
a topological action of a finite group on an @xmath0-manifold is called locally linear , if the local action of each isotropy group can be conjugated to a linear action on @xmath1 . according to classical results every topological action of a finite group on a @xmath2- or @xmath3-manifold @xmath4
is locally linear ( cf .
@xcite ) and preserves a piecewise linear- and a smooth structure on @xmath4 . on @xmath5
there are wild topological actions for which none of these properties holds @xcite .
however , at least for actions on closed @xmath6-manifolds it is known that these properties are still equivalent ( * ? ? ?
2.1 , cor . 2.2 ) .
this is not true anymore in dimension four .
there are locally linear actions of finite groups on closed @xmath7-manifolds that do not preserve any piecewise linear- or smooth structure ( * ? ? ?
* thm 3.1 , p. 257 ) .
it has been conjectured though that every piecewise linear action of a finite group on a closed piecewise linear @xmath7-manifold @xmath4 preserves some smooth structure on @xmath4 @xcite .
we confirm this conjecture in the following stronger form .
a piecewise linear- and a smooth structure on a manifold @xmath4 are called _ compatible _ with each other , if there exists a triangulation of @xmath4 as a piecewise linear manifold all of whose simplices are smoothly embedded with respect to the smooth structure . due to a theorem by whitehead every smooth manifold @xmath4 admits a unique compatible piecewise linear structure @xcite .
an equivariant version of this result for smooth actions of finite groups on @xmath4 holds by a theorem of illman @xcite .
conversely , any piecewise linear manifold of dimension @xmath8 admits a _ smoothing _ ,
i.e. a compatible smooth structure , which is unique up to isotopy for @xmath9 ( cf . section [ sec : smoothing_theory ] ) . in this paper
we show that for @xmath10 piecewise linear manifolds can be equivariantly smoothed in the following sense , thereby confirming the conjecture in @xcite .
let @xmath4 be a piecewise linear manifold of dimension @xmath11 on which a finite group @xmath12 acts by piecewise linear homeomorphisms .
then there exists a compatible @xmath12-equivariant smooth structure on @xmath4 , i.e. a smoothing with respect to which @xmath12 acts smoothly . in dimension higher than four
the statement of the theorem is false , even without the compatibility condition ( cf .
@xcite or remark [ rem : counter_example ] below ) .
it is always possible to find a piecewise linear triangulation of @xmath4 with respect to which the group @xmath12 acts simplicially ( cf . section [ sub : pl_structure ] ) . for @xmath13 we can choose the lengths of the segments of such a triangulation so that @xmath4 becomes metrically flat and @xmath12 acts isometrically . in this way
we obtain a desired smooth structure .
the task of finding such in the case @xmath14 appears in thurston s book @xcite in a slightly modified version as a `` problem '' and in the case @xmath15 as a `` challenge '' .
thurston asks for so - called _ canonical smoothings _ of triangulated piecewise linear manifolds , which are , in particular , preserved by simplicial isomorphisms . as a hint for the case @xmath15
he remarks that one probably needs some `` heavy machinery such as the uniformization theorem for riemannian metrics on @xmath16 , used with ingenuity '' .
the uniformization theorem implies that smooth actions of finite groups on @xmath16 are smoothly conjugate to linear actions .
the corresponding property for @xmath5 is assumed in @xcite while formulating the above mentioned conjecture and has later been proven by dinkelbach and leeb @xcite using ricci flow techniques , thereby drawing on earlier work of perelman and others on the geometrization conjecture .
indeed , it turns out that the key ingredients for proving our main result are uniqueness of smoothings and the linearizability of finite smooth group actions on spheres up to dimension three . for our proof
we specify the notion of canonical smoothings ( cf .
definition [ dfn : can_smoothing ] and remark [ rem : thurston_can ] ) in a way that allows us to show essentially by induction on the dimension that piecewise linear manifolds can be canonically smoothed up to dimension four ( cf .
proposition [ prp : can_smoothing ] ) . here , essentially means that additional technical difficulties have to be handled in dimensions three and four in order to guarantee the compatibility of the constructed smooth structure due to the necessity of introducing simplicial subdivisions .
we recall some concepts from piecewise linear topology .
references are @xcite and @xcite .
a subset @xmath17 is called a _ polyhedron _ if , for every point @xmath18 , there exists a finite number a simplices contained in @xmath19 such that their union is a neighborhood of @xmath20 in @xmath19 .
an _ @xmath0-cell _ is a convex compact polyhedron that spans a subspace of dimension @xmath0 .
for a cell @xmath21 and a point @xmath22 the set @xmath23 is defined to be the union of all lines through @xmath20 in @xmath1 that intersect @xmath24 in an arc with @xmath20 in its interior .
if there are no such lines , then @xmath23 is defined to be @xmath25 and @xmath20 is called a vertex of @xmath24 . otherwise @xmath23 is an affine subspace . in general the cell @xmath26
is called a face of @xmath24 and we write @xmath27 .
a _ cell complex _
@xmath28 is a collection of cells in @xmath1 satisfying the following conditions 1 . if @xmath29 and @xmath30 then @xmath31 .
2 . if @xmath32 then @xmath33 is a face of both @xmath34 and @xmath24 .
@xmath28 is locally finite in the sense that for every cell @xmath35 and every point @xmath36 there are only finitely many cells in @xmath28 that contain @xmath20 .
the linear image of a cell is again a cell and the intersection @xmath37 of two cell complexes @xmath28 and @xmath38 is again a cell complex .
@xmath39-skeleton _ of @xmath28 is the cell complex comprising all cells of @xmath28 of dimension smaller or equal to @xmath39 .
simplicial complex _ is a cell complex all of whose cells are a simplices .
the underlying space @xmath40 of a simplicial complex @xmath28 is a polyhedron .
conversely , every polyhedron @xmath19 in @xmath1 is the underlying space of some simplicial complex @xmath28 .
such a complex is called a _ triangulation _ of @xmath19 .
we denote the @xmath39-th _ barycentric subdivision _ of a simplicial complex @xmath28 by @xmath41 .
the _ support _ @xmath42 of a point @xmath20 in a simplicial complex @xmath28 is defined to be the smallest dimensional simplex of @xmath28 that contains @xmath20 .
the _ star _
@xmath43 of a simplex @xmath44 in a simplicial complex @xmath28 is the smallest simplicial complex that contains all simplices of @xmath28 that contain @xmath44 .
the link of @xmath44 is defined as @xmath45 we write @xmath43 and @xmath46 interchangeably for the star and the link as a simplicial complex and as their underlying space .
also we sometimes omit the complex @xmath28 in @xmath42 , @xmath43 and @xmath46 if its meaning is clear .
a _ subdivision _ of a cell complex @xmath28 is a simplicial complex @xmath47 such that @xmath48 and every simplex of @xmath47 is contained in a cell of @xmath28 .
we have [ lem : subdivision_cell_simplex ] any cell complex can be canonically subdivided into a simplicial complex .
the @xmath2-skeleton of @xmath28 is already a simplicial complex .
we successively subdivide the @xmath49-skeleton of @xmath28 .
assume that we have already subdivided the @xmath50-skeletons of @xmath28 .
then we star the @xmath51-cells of its @xmath51-skeleton at their barycenters , i.e. we replace each @xmath51-cell by the join of its boundary with its barycenter . having replaced the @xmath0-cells we obtain a simplicial complex that subdivids our initial cell complex .
a map @xmath52 between polyhedra @xmath17 , @xmath53 is called _ piecewise linear _ ( _ pl _ ) , if its graph @xmath54 is a polyhedron .
a _ pl chart _
@xmath55 for a second - countable hausdorff topological space @xmath56 is an embedding @xmath57 of a polyhedron @xmath19 .
two pl charts @xmath55 and @xmath58 are said to be _ compatible _ if @xmath59 is a polyhedron and @xmath60 is piecewise linear .
pl structure _ on @xmath56 is a maximal family of compatible pl charts covering @xmath56 .
the space @xmath56 endowed with a pl structure is called a _ pl
space_. in this way , one can define pl maps and pl homeomorphisms between pl spaces .
a pl space @xmath56 is called a _
pl manifold _ of dimension @xmath0 , if for every point @xmath61 there exists a chart @xmath62 of @xmath56 such that @xmath63 .
every pl manifold of dimension @xmath0 can be triangulated by some simplicial complex @xmath64 , i.e. it can be realised as a polyhedron ( * ? ? ?
3.5 , p. 80 ) ( * ? ? ?
7.1 , p. 53 ) . from now on all pl manifolds are regarded as polyhedra in some euclidean space .
if @xmath56 is a pl manifold , then the link of every simplex of @xmath28 is a pl standard sphere , i.e. it is pl homeomorphic to @xmath65 .
the following result is certainly known , but the author has not found a reference . [
prp : triangulation ] a piecewise linear manifold @xmath4 on which a finite group @xmath12 acts by piecewise linear homeomorphisms can be triangulated by a simplicial complex @xmath28 such that the action of @xmath12 on @xmath66 is simplicial , i.e. maps simplices linearly onto simplices .
we first triangulate @xmath4 by a simplicial complex @xmath67 . for each @xmath68
we can choose a subdivision @xmath69 of @xmath28 such that @xmath70 maps simplices of @xmath69 linearly into simplices of @xmath28 ( cf .
* thm . 3.6 , b , p. 84 ) ) .
let @xmath38 be the cell complex obtained by intersecting the cell complexes @xmath69 .
then the restriction of each element @xmath71 to each cell of @xmath38 is linear .
therefore , the translates @xmath72 are again cell complexes .
hence , their intersection @xmath73 is a cell complex on which @xmath12 acts cellularly , i.e. it maps cells linearly onto cells .
now we apply lemma [ lem : subdivision_cell_simplex ] to this complex . by construction ,
the group @xmath12 acts simplicially on the resulting simplicial complex .
the following definition is central for comparing piecewise linear and smooth spaces .
we call a map @xmath74 from a polyhedron @xmath19 to a smooth manifold @xmath4 _ piecewise differentiable _ or _ pd _ , if there exists a triangulation @xmath28 of @xmath19 such that the restriction of @xmath75 to each simplex is smooth .
we call @xmath75 a _ pd homeomorphism _ , if it is moreover a homeomorphism and each simplex is smoothly embedded , i.e. for each simplex @xmath76 and each point @xmath77 the differential @xmath78 is injective .
a smooth structure on a pl manifold @xmath4 is called _ compatible _ with the pl structure of @xmath4 , if the identity map from @xmath4 as a pl manifold to @xmath4 as a smooth manifold is a pd homeomorphism .
it is known that all pl manifolds up to dimension @xmath79 can be smoothed and that such smoothings are unique up to isotopy in dimensions lower or equal to @xmath80 .
general smoothing theory is described in @xcite and @xcite .
for the proof of our main result we only need the uniqueness statement in dimensions @xmath81 , which follows from the fact that every diffeomorphism of @xmath16 can be extended to a diffeomorphism of the unit @xmath6-ball ( cf .
* thm . 3.10.9 , thm .
3.10.11 , p. 202 ) ) . in this section
we explain how pd maps can be approximated by pl maps .
two pd maps @xmath82 are called _ @xmath83 @xmath84-close _ , if there exists a triangulation @xmath28 of @xmath19 such that both @xmath85 and @xmath86 are smooth and the values of @xmath87 and their first derivatives are bounded by @xmath84 for every simplex @xmath76 .
the following result is proven in ( * ? ? ?
8.8 , p. 84 and
8.4 , p. 81 ) .
[ thm : pd_approx_close ] let @xmath88 be a pd homeomorphism from a compact polyhedron to a smooth submanifold @xmath4 of @xmath1 . then there exist some @xmath89 such that every pd map @xmath90 that is @xmath83 @xmath84-close to @xmath75 is also a pd homeomorphism . in order to approximate pd maps by pl maps we need the following concept ( cf .
@xcite ) . let @xmath47 be a subdivision of a simplicial complex @xmath28 and let @xmath91 be a pd map .
the _ secant map _
@xmath92 is defined to be the map that is linear on the simplices of @xmath47 and coincides with @xmath75 on the vertices of @xmath47 . by definition @xmath93
is a pl map .
we would like to choose the subdivisions @xmath47 such that @xmath93 becomes close to @xmath75 in the @xmath83 sense .
to this end one needs to ensure that the thickness of the simplices in the subdivisions stays bounded from below .
the thickness of a simplex is defined to be the ratio of the minimal distance of its barycenter to its boundary and its diameter ( cf .
@xcite ) . the following result holds ( * ? ? ?
9.6 , p. 94 ) .
[ thm : pd_approx ] let @xmath28 be a finite simplicial complex and let @xmath91 be a pd map .
then for every @xmath94 there exists a subdivision @xmath47 of @xmath28 such that the secant map @xmath93 is @xmath83 @xmath84-close to @xmath75 .
using _ ricci flow _
techniques dinkelbach and leeb showed that any smooth action of a finite group on @xmath5 is smoothly conjugate to an orthogonal action @xcite .
the same statement is true for smooth actions of finite groups on @xmath16 , but in this case it follows more elementary , either by the _ uniformization theorem _ or by the _ geometrization of spherical 2-orbifolds _ ( cf .
let @xmath28 be a simplicial complex that is also a pl manifold of dimension @xmath11 .
we are going to show that the complex @xmath28 admits a smoothing with respect to which every simplicial isomorphism of @xmath28 is smooth .
this implies our main result according to proposition [ prp : triangulation ] .
parts of the proof are by induction on the dimension . however , in dimension three and four additional issues have to be dealt with .
let us first sketch the induction step before we describe the proof in detail . by induction we will obtain a certain smoothing of the link of each vertex in @xmath28 .
we use such a smoothing to construct a chart contained in the star of the corresponding vertex .
we do this in such a way that charts contained in simplicially isomorphic stars are equivalent , that the obtained charts form an atlas of @xmath28 and that the resulting smooth structure is compatible with the piecewise linear structure . in constructing such charts we are facing two disjoint problems . extending the smooth structure inwards to the cone point in a compatible way and extending it outwards such that overlapping charts are compatible .
the problem of extending the smooth structure inwards to the cone point can be treated by using the following two key lemmas .
the first lemma just formulates the uniqueness of smoothings and the linearizability of finite group actions on spheres in dimensions @xmath95 in a suitable manner ( cf . section [ sub : linearization ] and section [ sec : smoothing_theory ] ) .
[ lem : rad_linearization ] let @xmath28 be a triangulated pl @xmath0-sphere , @xmath81 , on which a finite group @xmath12 acts simplicially .
suppose that @xmath28 is equipped with a smoothing with respect to which @xmath12 acts smoothly .
then there exists a group homomorphism @xmath96 and a @xmath12-equivariant pd homeomorphism @xmath97 . for a topological space @xmath56
we denote by @xmath98 its _ open cone _ which is defined to be the quotient space @xmath99 where @xmath100 if and only if @xmath101 . for a pl space @xmath56
, the cone @xmath98 inherits a natural pl space structure from @xmath56 .
moreover , if @xmath56 is a pl @xmath0-sphere , then @xmath98 is pl homeomorphic to @xmath102 ( cf .
@xcite ) .
[ lem : rad_extension ] let @xmath28 be a triangulated pl @xmath0-sphere on which a finite group @xmath12 acts simplicially .
suppose there exists a group homomorphism @xmath96 and a @xmath12-equivariant pd homeomorphism @xmath97 .
then for every @xmath103 this pd homeomorphism can be extended to an @xmath12-equivariant pd homeomorphism @xmath104 such that @xmath105 for all @xmath106 and all @xmath107 .
we can not simply extend the map @xmath75 linearly to the origin , because in this case the restriction to a simplex could be degenerate at the cone point .
however , we can isotopy @xmath75 along the radial direction to a map that can be linearly extended to the cone point in a compatible way .
we do this in two steps .
in the first step we isotopy @xmath75 such that the wildly embedded simplices of @xmath28 become spherical simplices . the second isotopy deforms the spherical simplices into flat simplices ( cf .
figure [ fig : isotopy ] ) .
more precisely , let @xmath109 be the radial projection and let @xmath110 be a smooth cutoff function with @xmath111 , @xmath112 , @xmath113 for @xmath114 and @xmath115 for @xmath116 .
according to theorem [ thm : pd_approx ] we can subdivide the complex @xmath28 such that the map @xmath117 becomes arbitrarily close to @xmath75 in the @xmath83-sense .
hence , the same is true for the map @xmath118 .
for a sufficiently good approximation the map @xmath119 with @xmath120 is well - defined for @xmath121 .
moreover , if we choose a sequence of subdivisions such that @xmath122 converges to @xmath75 in the @xmath83-sense , then @xmath123 converges uniformly in @xmath124 to @xmath75 in the @xmath83-sense .
therefore the map @xmath125 defines a @xmath12-equivariant pd homeomorphism onto its image by theorem [ thm : pd_approx_close ] applied to the maps @xmath123 . with @xmath126 we set @xmath127 & { \rightarrow } & { \mathbb{r}}\\ & ( x , t ) & { \mapsto } & \theta_2(t ) \frac{1}{\left\| \tilde{f}(x ) \right\| } + ( 1-\theta_2(t ) ) \end{array}\ ] ] and define @xmath128 & { \rightarrow } & { \mathbb{r}}^{n+1 } \\ & ( x , t ) & { \mapsto } & t\mu(x , t)\tilde{f}(x ) \end{array}\ ] ] then the map @xmath129 descends to a @xmath12-equivariant pd homeomorphism @xmath130 .
finally , we can replace @xmath131 by @xmath132 to guarantee the last condition .
the compatibility of overlapping charts will be guaranteed by defining the outward extensions in a certain canonical way . to this end
we endow a simplicial complex @xmath28 with a _
polyhedral metric _ such that any @xmath2-simplex in @xmath28 has unit length and such that the simplices of @xmath28 are flat .
we denote the set of vertices of @xmath133 , the barycentric subdivision of @xmath28 , whose supporting simplex in @xmath28 has dimension @xmath134 by @xmath135 and set @xmath136 where @xmath137 .
each point @xmath138 has a neighborhood @xmath139 that splits isometrically as a product @xmath140 of open sets @xmath141 and @xmath142 , where @xmath143 is the set of points in @xmath144 for which the straight line between @xmath20 and @xmath145 meets @xmath42 orthogonally .
note that open sets @xmath146 and @xmath147 are disjoint for distinct @xmath148 .
[ dfn : unifrom - cover ] a neighborhood @xmath139 of @xmath149 as above is called a _ product neighborhood_. we call it a _ symmetric product neighborhood _ if @xmath150 is in addition invariant under all simplicial isomorphisms of @xmath151 that leave @xmath42 invariant .
an open cover @xmath152 of @xmath28 consisting of symmetric product neighborhoods @xmath153 of @xmath149 is called a _ symmetric product cover _ , if for all @xmath154 and all @xmath148 any simplicial isomorphism between @xmath151 and @xmath151 that maps @xmath42 onto @xmath42 , maps @xmath150 onto @xmath155 . in order to have control on the sizes of product neighborhoods of a symmetric product cover @xmath156 ( cf .
proof of lemma [ lem : canonical_smooth_n:2 ] ) we introduce its fineness @xmath157 defined as @xmath158 [ dfn : unifrom - cover2 ] we call a collection of symmetric product covers , one for each triangulated piecewise linear @xmath0-manifold , _ uniform _ , if the conditions in definition [ dfn : unifrom - cover2 ] also hold for points @xmath20 and @xmath145 in distinct triangulated piecewise linear @xmath0-manifolds and if the finenesses of the symmetric product covers are uniformly small
. the supremum of @xmath157 over covers of the collection is called the _ fineness _ of the collection of symmetric product covers .
it is clear that triangulated piecewise linear @xmath2-manifolds admit a uniform collection of symmetric product covers with arbitrarily small fineness . in higher dimensions such
will later be constructed inductively .
given a symmetric product cover , we would like to find a smoothing that behaves accordingly in the following sense .
[ dfn : prod_smoothing ] let @xmath156 be a symmetric product cover of a triangulated piecewise linear @xmath0-manifold @xmath28 .
a smoothing of @xmath28 is called a _ product smoothing with respect to @xmath156 _ , if for each product neighborhood @xmath159 of a point @xmath149 the sets @xmath160 and @xmath161 are smooth submanifolds of @xmath28 , the set @xmath160 is moreover a smooth submanifold of the simplex @xmath42 and the restriction of the smooth structure on @xmath28 to @xmath150 coincides with the product smooth structure on @xmath162 .
[ dfn : uniform_smoothings ] suppose we are given a uniform collection of symmetric product covers , one for each triangulated piecewise linear @xmath0-manifold .
we say that a collection of product smoothings , one for each triangulated piecewise linear @xmath0-manifold adapted to the corresponding symmetric product cover , is _ uniform _ , if for any two points @xmath149 and @xmath163 in any two triangulated piecewise linear @xmath0-manifolds @xmath28 and @xmath38 the following condition holds .
any simplicial isomorphism between @xmath164 and @xmath165 that maps @xmath42 onto @xmath166 and thus @xmath162 onto @xmath167 restricts to a diffeomorphism between @xmath161 and @xmath168 and hence to a diffeomorphism between @xmath150 and @xmath155 . 1 .
there exists a uniform collection of symmetric product covers for triangulated piecewise linear @xmath0-manifolds , 2 .
any triangulated piecewise linear @xmath0-manifold @xmath28 admits an atlas defining a compatible smooth structure and comprising one chart for each vertex of @xmath28 that is contained in the star neighborhood of this vertex 1 .
the resulting smoothings form a uniform collection of product smoothings with respect to the uniform collection of symmetric product covers , 2 .
any simplicial isomorphism between star neighborhoods of vertices in any two triangulated piecewise linear manifolds maps corresponding charts onto each other . in view of proving proposition [ prp : can_smoothing ]
we first prove a stronger statement in dimensions @xmath171 .
note that the proof of this statement is formulated in a more general way .
except for certain parts which require the assumption @xmath171 as will be pointed out explicitly , the proof can later again be applied in higher dimensions .
[ lem : canonical_smooth_n:2 ] for @xmath172 pl @xmath0-manifolds can be canonically smoothed .
moreover , the smoothings can be chosen in such a way that for a given triangulated pl @xmath0-manifold all of its simplices are smoothly embedded .
any triangulated pl @xmath2-manifold with its canonical metric is either isometric to a circle or to the real line and by this identification we obtain the desired smooth structure .
let @xmath14 , fix corresponding uniform collections of symmetric product covers and product smoothings for triangulated pl @xmath2-manifolds and let @xmath173 be small .
the link of any vertex @xmath174 in any triangulated pl
@xmath3-manifold @xmath28 is a triangulated circle .
let @xmath12 be the group of all simplicial isomorphism of such a link @xmath175 .
this group acts smoothly with respect to the canonical smooth structure of @xmath175 . by lemma [ lem : rad_linearization ] there exists a group homomorphism @xmath176 and a @xmath12-equivariant pd homeomorphism @xmath177 . by lemma [ lem : rad_extension ]
we can extend @xmath75 to a @xmath12-equivariant pd homeomorphism @xmath178 such that @xmath105 for all @xmath179 and all @xmath180 .
note that in this case of dimension @xmath14 we can simply identify a rescaled version of @xmath175 with respect to its canonical metric with @xmath181 to obtain the map @xmath177 instead of applying lemma [ lem : rad_linearization ] .
hence , in this case the map @xmath131 can be chosen such that its restriction to the cone over each simplex of @xmath175 is a smooth embedding ( cf .
proof of lemma [ lem : rad_extension ] ) .
this construction defines charts @xmath182 which are however in general not compatible with each other .
we want to replace them by charts @xmath183 , defined on a small open neighborhood of @xmath184 , that are compatible with each other .
the idea is to construct the map @xmath185 as indicated in figure [ fig : function_h ] .
the construction will guarantee that the smooth structures defined by two overlapping charts are product smooth structures on their overlap that fit together by induction .
the map @xmath186 is going to be composed of building blocks @xmath187 that are constructed in the following lemma . for @xmath15 the need of a modified version of this lemma ( cf .
lemma [ lem : modified_version ] ) will become clear later . recall that a map @xmath188 between subsets @xmath189 is called smooth , if it admits a smooth extension onto an open neighborhood of @xmath190 and that it is called a diffeomorphism , if it is smooth and has a smooth inverse .
[ lem : function_h_local ] let @xmath191 , @xmath192 , be a simplex together with a symmetric product cover @xmath156 .
we endow @xmath193 with a flat metric such that all sides of @xmath194)/\sim$ ] have unit length .
then for every @xmath195 there exists a diffeomorphism @xmath196 such that the following properties hold 1 .
the map @xmath188 is equivariant with respect to the group of isometries of @xmath197 .
2 . for all @xmath198 and all @xmath199
we have @xmath200 .
3 . @xmath201 } = \mathrm{id}$ ] .
4 . for all @xmath202 , all @xmath138 and all @xmath203 , @xmath204 , we have @xmath205 a proof of this lemma can be found below .
copies of @xmath188 can be put together to a homeomorphism @xmath185 that is equivariant with respect to the group of all simplicial isomorphisms of @xmath175 .
we define @xmath206 for some small @xmath207 that will be specified later and claim that the new charts are compatible with each other . before proving this
we make the following observations .
let @xmath44 be a simplex of @xmath28 with two different vertices @xmath208 and @xmath174 .
then we have @xmath209 and thus there exists a simplicial isomorphism between the corresponding stars that maps @xmath210 onto @xmath211 and is the identity on @xmath212 .
such a simplicial isomorphism restricts to a diffeomorphism between the product neighborhoods @xmath162 of @xmath213 in @xmath214 and @xmath167 of @xmath215 in @xmath216 corresponding to the uniform collection of symmetric product covers and smoothings of triangulated pl @xmath2-manifolds that we have fixed above , and further to a diffeomorphism @xmath217 ( cf .
definition [ dfn : uniform_smoothings ] ) .
note that this diffeomorphism @xmath75 is independent of the specific choice of the simplicial isomorphism between @xmath214 and @xmath216 we started with .
suppose we have @xmath218 , @xmath219 and some @xmath220 with @xmath221 .
then it is clear by the properties of the maps @xmath186 and @xmath222 and because of @xmath223 that @xmath224 holds for all @xmath225 ( cf .
@xmath226 and @xmath227 in lemma [ lem : function_h_local ] ) . by this observation and the properties of the map @xmath188 ( cf .
@xmath227 in lemma [ lem : function_h_local ] ) , we can construct a symmetric product cover of @xmath28 , and more generally of any triangulated piecewise linear @xmath3-manifold , as described in the following lemma .
[ lem : new_product_cover ] suppose we have fixed a uniform collection of symmetric product covers of triangulated piecewise linear @xmath0-manifolds , @xmath192 , and some @xmath228 . for a triangulated piecewise linear @xmath229-manifold @xmath28 and a vertex @xmath230 let @xmath231 .
for @xmath232 we set @xmath233 otherwise @xmath174 is a vertex of @xmath28 and we set @xmath234 .
then the sets @xmath235 define a uniform collection of symmetric product covers of triangulated piecewise linear @xmath229-manifolds .
moreover , the fineness of the collection of sets @xmath236 can be chosen to be arbitrarily small by choosing @xmath173 and the fineness of the uniform collection of symmetric product covers of triangulated piecewise linear @xmath0-manifolds we started with sufficiently small .
let @xmath235 be the symmetric product cover of @xmath28 given by lemma [ lem : new_product_cover ] and let @xmath208 and @xmath174 be adjacent vertices of @xmath28 . for fixed fineness
we can choose @xmath207 such that @xmath237 for all simplices @xmath44 of @xmath28 with @xmath238 .
then we have @xmath239 where the union runs over all simplices of @xmath28 that contain @xmath208 and @xmath174 as a vertex .
then , by the observation above and the regularity of the map @xmath188 from lemma [ lem : function_h_local ] , the composition @xmath240 factorises through a product of a diffeomorphism between @xmath241 and @xmath242 ( recall that @xmath160 and @xmath243 are smooth submanifolds of @xmath244 and @xmath245 , respectively , cf
. definition [ dfn : prod_smoothing ] ) and the diffeomorphism @xmath217 introduced above .
hence , the restriction of the transition map @xmath246 to @xmath247 is smooth and thus it is everywhere smooth because of ( [ eqn : inclusion ] ) .
moreover , we see that the resulting smooth structure on @xmath248 is a product of the smooth structures on @xmath249 inherited from @xmath44 and on the submanifold @xmath250 . consequently , the collection of charts @xmath251 defines an atlas and by construction any simplicial isomorphism between star neighborhoods of two vertices in any two triangulated pl 2-manifolds maps corresponding charts onto each other .
moreover , since the map @xmath188 in lemma [ lem : function_h_local ] is a diffeomorphism and since we did not have to subdivide @xmath175 in order to construct the map @xmath131 in the case @xmath14 , the inclusions of the simplices of @xmath28 are smooth embeddings . in particular , the defined smooth structure is compatible with the pl structure of @xmath28 .
it is even a product smoothing according to what has been shown above .
finally , it is clear that this construction yields a uniform collection of product smoothings corresponding to the given uniform collection of product covers .
we can assume that @xmath197 is embedded in @xmath102 such that the cone point lies in the origin and such that @xmath252 lies in the @xmath253 plane .
we first construct a vector field @xmath56 with flow @xmath254 such that the restriction of the map @xmath255 to @xmath197 satisfies all conditions of the lemma , but @xmath256 instead of the orthogonality condition in @xmath227 . here
@xmath257 are cone coordinates corresponding to @xmath258 .
afterwards we reparametrise the map @xmath259 along the flow lines of @xmath254 in the interval @xmath260 such that the reparametrised map satisfies all conditions formulated in the lemma .
more precisely , for @xmath94 we define @xmath261 where @xmath262 is the open @xmath84-neighborhood of @xmath160 in @xmath263 and @xmath264 is the open @xmath84-neighborhood of @xmath161 in @xmath265 .
we choose @xmath84 such that the collection of sets @xmath266 still defines a symmetric product cover of @xmath191 .
for @xmath267 we define @xmath268 note that for each @xmath154 the sets @xmath269 , @xmath270 , are disjoint and for @xmath271 we have @xmath107 .
let @xmath272 be the union of the @xmath269 , @xmath273 .
clearly @xmath274 is an open cover of @xmath275 .
we want to define a smooth vector field @xmath56 on @xmath276 that is equivariant with respect to the group of isometries of @xmath197 such that for some small @xmath277 that will be specified later the following conditions are satisfied 1 .
@xmath278 for all @xmath279 .
2 . @xmath280 if @xmath281 .
3 . @xmath282 for @xmath283 with @xmath284 .
4 . @xmath285 for @xmath286 with @xmath287 and @xmath288 .
@xmath289 for all @xmath290 with @xmath291 and @xmath292 . to this end , for a subset @xmath293 , we define @xmath294 and @xmath295 the sets @xmath296 form a partition of @xmath297 .
we define an equivariant vector field @xmath298 on @xmath297 by @xmath299 if @xmath300 .
the vector field @xmath298 satisfies conditions @xmath226-@xmath301 on @xmath302 , but it is neither smooth nor continuous .
however , we can mollify @xmath298 such that the resulting vector field @xmath303 is smooth and satisfies conditions @xmath226-@xmath301 on @xmath302 by a convolution with a centrosymmetric mollifier with sufficiently concentrated support .
the vector field @xmath304 is well - defined on the upper halfspace , smooth and equivariant and satisfies conditions @xmath226-@xmath301 on @xmath305 .
hence , we can interpolate between @xmath306 and @xmath303 in the interval @xmath307 to obtain a desired vector field .
let @xmath259 be the map defined above with the flow @xmath254 of this vector field . by continuity , for sufficiently small @xmath277
the @xmath191-component in cone coordinates of @xmath308 lies in @xmath309 for all @xmath310 and all @xmath311 .
hence , for such a @xmath277 the map @xmath259 satisfies all conditions of the lemma by construction except the orthogonality condition in @xmath227 .
instead we have @xmath312 by @xmath226 .
for sufficiently small @xmath277 we can construct smooth maps @xmath313 by using the implicit function theorem such that for all @xmath107 and all @xmath314 we have @xmath315 we can use a partition of unity @xmath316 with @xmath317 on @xmath309 for all @xmath318 with @xmath319 to interpolate the maps @xmath320 , @xmath138 , to a map @xmath321 , which can be assumed to be equivariant with respect to all isometries of @xmath197 , such that for all @xmath107 , all @xmath138 and all @xmath314 we have @xmath322 let @xmath110 be a smooth function with @xmath115 for @xmath323 , @xmath113 for @xmath324 and @xmath325 everywhere . suppose that @xmath326 . then the reparametrised map @xmath327 has all desired properties . for a vertex @xmath174 of a simplex @xmath191 and @xmath328 we define @xmath329 ) /\sim \subset \delta^n,\ ] ] where we have realised @xmath191 as a closed cone @xmath330 $ ] with vertex @xmath331 .
analogously we define for a vertex @xmath174 of an @xmath0-dimensional simplicial complex @xmath28 @xmath332 if the complex is understood we simply write @xmath333 .
similar as in lemma [ lem : canonical_smooth_n:2 ] , the statement in the @xmath6-dimensional case is formulated in a more general way .
the additional conclusions will be needed to treat the @xmath7-dimensional case below .
[ lem : canonical_smooth_n:3 ] pl @xmath6-manifolds can be canonically smoothed . moreover , for some @xmath334 the smoothings can be chosen in such a way that for any triangulated pl @xmath6-manifold @xmath28 with symmetric product cover @xmath156 and any @xmath6-simplex @xmath335 of @xmath28 the cells @xmath336 , @xmath337 , are smoothly embedded and for all @xmath338 and all @xmath339 , @xmath149 , @xmath340 , we have @xmath341
. we proceed as in the proof of the two - dimensional case ( cf .
lemma [ lem : canonical_smooth_n:2 ] ) , except that we may have to subdivide the link of a vertex @xmath174 of a triangulated pl @xmath6-manifold @xmath28 during the procedure of lemma [ lem : rad_extension ] .
therefore , the simplices of @xmath28 are not necessarily smoothly embedded anymore .
however , if we choose @xmath342 from the proof of lemma [ lem : canonical_smooth_n:2 ] , the cells @xmath336 are smoothly embedded in @xmath28 for any @xmath6-simplex @xmath335 of @xmath28 .
the reason is that the necessary subdivisions introduced in the proof of lemma [ lem : rad_extension ] only cause non - smooth breaks for @xmath343 .
we construct a uniform collection of symmetric product covers as in lemma [ lem : new_product_cover ] and observe that for such a cover @xmath156 of a triangulated piecewise linear @xmath6-manifold @xmath28 we have @xmath341 for all vertices
@xmath338 of @xmath28 and all @xmath339 , @xmath149 , @xmath340
. therefore everything follows as in the proof of lemma [ lem : canonical_smooth_n:2 ] except the compatibility of the piecewise linear and the smooth structure . to see this compatibility
, we subdivide the simplicial neighborhoods @xmath344 as in the proof of lemma [ lem : rad_extension ] and extend this subdivision to a subdivision of the initial triangulation of @xmath28 as in the proof of lemma [ lem : subdivision_cell_simplex ] . with respect to such a triangulation
all simplices are smoothly embedded and thus the claim follows .
the new difficulty compared to the proof of lemma [ lem : canonical_smooth_n:2 ] and lemma [ lem : canonical_smooth_n:3 ] is that in general we can not apply lemma [ lem : rad_linearization ] without subdividing the link of a vertex of a triangulated pl @xmath7-manifold @xmath28 .
if we applied lemma [ lem : function_h_local ] as before , we would still obtain a smooth structure invariant under all simplicial isomorphisms , but we could not guarantee its compatibility anymore , since we had no control about the effect of the maps @xmath345 on the simplices in our subdivision .
however , we can use the following modified version of lemma [ lem : function_h_local ] instead .
the construction of the new map @xmath188 is sketched in figure [ fig : function_h ] .
it is piecewise linear in all regions @xmath346 in which we had to introduce nonsmooth breaks in the step before .
[ lem : modified_version ] let @xmath335 be a simplex together with a symmetric product cover @xmath156 and let @xmath347 be such that @xmath348 for all @xmath349 .
we endow @xmath350 with a flat metric such that all sides of @xmath351)/\sim$ ] have unit length . then for every @xmath195 there exists a homeomorphism @xmath352 such that the following properties hold 1 .
the map @xmath188 is equivariant with respect to the group of isometries of @xmath353 .
2 . for all
@xmath198 we have @xmath354 .
3 . @xmath355 } = \mathrm{id}$ ] .
the restriction of @xmath188 to @xmath356 $ ] is pl and the restrictions of @xmath188 to @xmath357 and @xmath358 , @xmath359 , are linear .
the restrictions of @xmath188 to @xmath360 and to @xmath361 are diffeomorphisms onto their images .
6 . for all @xmath202 , any vertex @xmath174 of @xmath335 and for all
@xmath362 we have @xmath363 7 . for all @xmath364 , all @xmath138 and all @xmath365 , @xmath204
, we have @xmath205 it is easy to define the map for @xmath366 .
then the map is determined on @xmath367 by condition @xmath368 of the lemma . with the same methods as in lemma [ lem : subdivision_cell_simplex ] the map can be extended to the rest of @xmath197 such that all desired properties hold .
let us continue with the proof of lemma [ lem : four - dim ] .
the properties of the maps @xmath188 from this lemma and from lemma [ lem : function_h_local ] differ only for small @xmath124 .
therefore , using lemma [ lem : modified_version ] with @xmath369 as in lemma [ lem : canonical_smooth_n:3 ] instead of lemma [ lem : function_h_local ] to construct the maps @xmath186 in the proof of lemma [ lem : canonical_smooth_n:2 ] , the compatibility of overlapping charts follows as before .
in fact , everything else follows as in the proof of lemma [ lem : canonical_smooth_n:2 ] and lemma [ lem : canonical_smooth_n:3 ] except the compatibility of the pl and smooth structure .
we set @xmath370 @xmath371 and @xmath372 . here
the _ join _
@xmath373 is defined to be the set of points in @xmath374 that lie on a shortest path connecting a point of @xmath375 with @xmath174 .
recall that we have endowed @xmath28 with a smooth structure . by the properties of the map @xmath188 in lemma [ lem : modified_version ] , the spaces @xmath376 , @xmath377 and @xmath38 are polyhedra .
moreover , the inclusion @xmath378 is pd and the inclusion of @xmath38 endowed with the smooth structure from the simplices of @xmath28 into @xmath28 is smooth and nondegenerate ( cf .
lemma [ lem : canonical_smooth_n:3 ] ) .
we can first choose a triangulation of @xmath379 such that all its simplices are smoothly embedded in @xmath28 .
then we can extend this triangulation to a triangulation of @xmath28 all of whose simplices are smoothly embedded by again applying the method from the proof of lemma [ lem : subdivision_cell_simplex ] as indicated in figure [ fig : lj_sub ] .
hence the constructed smooth structure is compatible with the pl structure and the claim follows .
[ rem : counter_example ] there exist piecewise linear actions of @xmath380 on a @xmath7-dimensional piecewise linear sphere that can not be equivariantly smoothed .
one way to obtain such an example is as follows ( cf .
@xcite ) . the group @xmath380 admits a piecewise linear action on @xmath381 whose fixed point set is a knotted @xmath16 @xcite . by suspending this action
one obtains a piecewise linear action of @xmath380 on @xmath382 with fixed point set @xmath5 .
however , this action can not be equivariantly smoothed because its fixed point set @xmath383 is not locally flat .
counterexamples in higher dimensions can be obtained by taking suspensions of this example .
* acknowledgements . * parts of this work were written while the author stayed at the pennsylvania state university .
i am grateful to the mathematical department there and to anton petrunin for their hospitality and support .
i would like to thank fernando galaz - garcia for drawing my attention to @xcite and everybody else who answered my questions .
i thank stephan stadler and stephan wiesendorf for helpful discussions and comments on different attempts to prove this paper s result .
i am especially grateful to alexander lytchak for his constant support .
morris w. hirsch and barry mazur , _ smoothings of piecewise linear manifolds _ , princeton university press , princeton , n. j. ; university of tokyo press , tokyo , 1974 .
annals of mathematics studies , no .
robion c. kirby and laurence c. siebenmann , _ foundational essays on topological manifolds , smoothings , and triangulations_. princeton university press , princeton , n.j . ; university of tokyo press , tokyo , 1977 . with notes by john
milnor and michael atiyah ; annals of mathematics studies , no . | we prove that every piecewise linear manifold of dimension up to four on which a finite group acts by piecewise linear homeomorphisms admits a compatible smooth structure with respect to which the group acts smoothly . |
the fact that today matter and dark energy have comparable contributions to the energy density , @xmath1 ev)@xmath2 in natural units , poses one of the most important problems for models of dark energy .
thus , if dark energy is a cosmological constant , its energy density would remain constant throughout the history of the universe , whereas those of the rest of components ( matter and radiation ) grow as we go back in time .
then the question arises as to whether it is a _ coincidence _ ( or not ) that they have comparable values today when they have differed by many orders of magnitude in the past .
notice also that if @xmath0 is a fundamental constant of nature , its scale ( around @xmath3 ev ) is more than 30 orders of magnitude smaller than the natural scale of gravitation , @xmath4 with @xmath5 gev . on the other hand , alternative models in which dark energy is a dynamical component rather than a cosmological
constant also require the introduction of unnatural scales in their lagrangians or initial conditions in order to account for the present phase of accelerated expansion .
such models are usually based on new physics , either in the form of new cosmological fields or modifications of einstein s gravity @xcite and they are generically plagued by additional problems such as classical or quantum instabilities , or inconsistencies with local gravity constraints . therefore , we would like to find a description for dark energy without dimensional scales ( apart from newton s constant @xmath6 ) , with the same number of free parameters as @xmath0cdm , with natural initial conditions , with good fits to observations and no consistency problems . in this work
we will show that vector theories can do the job .
with that purpose , we present two examples of such theories which have been recently proposed @xcite ( for other vector models see references in those works ) .
the action in this case reads @xcite : @xmath7 notice that the theory contains no free parameters , the only dimensional scale being the newton s constant .
the numerical factor in front of the vector kinetic terms can be fixed by the field normalization .
also notice that the `` mass '' term @xmath8 can be written as a combination of derivative terms as @xmath9 and therefore the theory contains no potential terms .
this action resembles that of maxwell electromagnetism in the feynman gauge with a mass term .
the classical equations of motion derived from the action in ( [ scaling ] ) are the einstein s and vector field equations : @xmath10 where @xmath11 is the conserved energy - momentum tensor for matter and radiation and @xmath12 is the energy - momentum tensor coming from the vector field .
for the simplest isotropic and homogeneous flat cosmologies , we assume that the spatial components of the vector field vanish , so that @xmath13 and that the space - time geometry will be given by : @xmath14 for this metric ( [ eqa ] ) reads : @xmath15a_0=0\label{fieldeq0}\end{aligned}\ ] ] assuming that the universe has gone through radiation and matter phases in which the contribution from dark energy was negligible , we can easily solve this equation in those periods . in that case
, the above equation has a growing and a decaying solution : @xmath16 with @xmath17 constants of integration and @xmath18 in the radiation era , and @xmath19 in the matter era .
on the other hand , the @xmath20 component of einstein s equations reads : @xmath21 \label{friedmann}\end{aligned}\ ] ] where the vector energy density is given by : @xmath22 = 6.5 cm ( left ) evolution of energy densities for the best fit model . dashed ( red ) for radiation ,
dotted ( green ) for matter and solid ( blue ) for vector dark energy .
we show also for comparison the cosmological constant density in dashed - dotted line .
( right ) evolution of dark energy equation of state for the best fit model .
the lower panel shows the 1@xmath23 confidence interval . using the growing mode solution from ( [ fieldsol ] )
, we obtain @xmath24 with @xmath25 in the radiation era and @xmath26 in the matter era .
thus , the energy density of the vector field starts scaling as radiation at early times , so that @xmath27 const .
however , when the universe enters its matter era , @xmath28 starts growing relative to @xmath29 eventually overcoming it at some point , in which the dark energy vector field would become the dominant component ( see fig .
1 ) . notice that since @xmath30 is essentially constant during radiation era , solutions do not depend on the precise initial time at which we specify it .
thus , once the present value of the hubble parameter @xmath31 and the constant @xmath30 during radiation ( which fixes the total matter density @xmath32 ) are specified , the model is completely determined , i.e. this model contains the same number of parameters as @xmath0cdm , which is the minimum number of parameters of a cosmological model with dark energy .
as seen from fig.1 the evolution of the universe ends at a finite time @xmath33 where @xmath34 with @xmath35 finite , @xmath36 , @xmath37 and @xmath38 .
this corresponds to a type iii ( _ big - freeze _ ) singularity according to the classification in @xcite .
we can also calculate the effective equation of state for dark energy as : @xmath39 again , using the approximate solutions in ( [ fieldsol ] ) , we obtain : @xmath40 in the radiation era and @xmath41 in the matter era . as shown in fig .
1 , the equation of state can cross the so called phantom divide , so that we can have @xmath42 . in order to confront the predictions of the model with observations of high - redshift supernovae type ia ,
we have carried out a @xmath43 statistical analysis for two supernovae datasets , namely , the gold set @xcite , containing 157 points with @xmath44 , and the more recent snls data set @xcite , comprising 115 supernovae but with lower redshifts ( @xmath45 ) . [
cols="^,^,^,^,^ " , ] best fit parameters with @xmath46 intervals for the vector model ( vcdm ) and the cosmological constant model ( @xmath0cdm ) for the gold ( 157 sne ) and snls ( 115 sne ) data sets .
@xmath47 denotes the present equation of state of dark energy .
@xmath30 is the constant value of the vector field component during radiation .
@xmath48 is the deceleration - aceleration transition redshift .
@xmath49 is the age of the universe in units of the present hubble time .
@xmath33 is the duration of the universe in the same units . in table 1
we show the results for the best fit together with its corresponding @xmath46 intervals for the two data sets .
we also show for comparison the results for a standard @xmath0cdm model .
we see that the vector model ( vcdm ) fits the data considerably better than @xmath0cdm ( in more than @xmath50 ) in the gold set , whereas the situation is reversed in the snls set .
this is just a reflection of the well - known @xmath50 tension @xcite between the two data sets .
compared with @xmath0cdm , we see that vcdm favors a younger universe ( in @xmath51 units ) with larger matter density .
in addition , the deceleration - acceleration transition takes place at a lower redshift in the vcdm case . the present value of the equation of state with @xmath52 which clearly excludes the cosmological constant value @xmath53 .
future surveys @xcite are expected to be able to measure @xmath47 at the few percent level and therefore could discriminate between the two models .
we have also compared with other parametrizations for the dark energy equation of state @xcite .
since our one - parameter fit has a reduced chi - squared : @xmath54 , vcdm provides the best fit to date for the gold data set .
we see that unlike the cosmological constant case , throughout radiation era @xmath55 in our case .
moreover the scale of the vector field @xmath56 @xmath57 in that era is relatively close to the planck scale and could arise naturally in the early universe without the need of introducing extremely small parameters ( for instance in an inflationary epoch ) , thus avoiding the coincidence problem . in order to study the model stability we have considered the evolution of metric and vector field perturbations .
thus , we obtain the dispersion relation and the propagation speed of scalar , vector and tensor modes . for all of them we obtain @xmath58 which is real throughout the universe evolution , since the value @xmath59 exactly corresponds to that at the final singularity .
therefore the model does not exhibit exponential instabilities .
as shown in @xcite , the fact that the propagation speed is faster than @xmath60 does not necessarily implies inconsistencies with causality .
we have also considered the evolution of scalar perturbations in the vector field generated by scalar metric perturbations during matter and radiation eras , and , again , we do not find exponentially growing modes . if we are interested in extending the applicability range of the model down to solar system scales then we should study the corresponding post - newtonian parameters ( ppn ) .
we can see that for the model in ( [ action ] ) , the static ppn parameters agree with those of general relativity @xcite , i.e. @xmath61 . for the parameters associated to preferred frame effects we get : @xmath62 and @xmath63 where @xmath64 is the norm of the vector field at the solar system scale .
current limits @xmath65 ( or @xmath66 for static vector fields during solar system formation ) then impose a bound @xmath67 . in order to determine whether such bounds conflict with the model predictions or not , we should know the predicted value of the field at solar system scales , which in principle does not need to agree with the cosmological value . indeed , @xmath64 will be determined by the mechanism that generated this field in the early universe characterized by its primordial spectrum of perturbations , and the subsequent evolution in the formation of the galaxy and solar system .
another potential difficulty arising generically in vector - tensor models is the presence of negative energy modes for perturbations on sub - hubble scales .
they are known to lead to instabilities at the quantum level , but not necessarily at the classical level as we have shown previously .
work is in progress in order to determine which are the necessary conditions to avoid the presence of such states in this model .
in the previous section , we have seen that a generic vector field whose action resembles that of electromagnetism with a mass term could be a good candidate for dark energy , but what about standard electromagnetism ? . in a very recent work @xcite
, it has been shown using the covariant ( gupta - bleuler ) formalism that it is indeed possible to explain cosmic acceleration from the standard maxwell s theory .
we start by writing the standard electromagnetic action including a gauge - fixing term in the presence of gravity : @xmath68 \label{action}\end{aligned}\ ] ] the gauge - fixing term is required in order to define a consistent quantum theory for the electromagnetic field @xcite , and we will see that it plays a fundamental role on large scales .
still this action preserves a residual gauge symmetry @xmath69 with @xmath70 .
electromagnetic equations derived from this action can be written as : @xmath71 notice that since we will be using the covariant gupta - bleuler formalism , we do not a priori impose the lorentz condition .
we shall first focus on the simplest case of a homogeneous electromagnetic field @xmath72 in a flat robertson - walker background . in this space - time
, equations ( [ fieldeq ] ) read : @xmath73 we can solve ( [ feqrw ] ) during the radiation and matter dominated epochs when the hubble parameter is given by @xmath74 with @xmath75 for radiation and @xmath76 for matter .
in such a case the solutions for ( [ feqrw ] ) are : @xmath77 where @xmath17 and @xmath78 are constants of integration .
hence , the growing mode of the temporal component does not depend on the epoch being always proportional to the cosmic time @xmath79 , whereas the growing mode of the spatial component evolves as @xmath80 during radiation and as @xmath81 during matter , i.e. at late times the temporal component will dominate over the spatial ones .
the energy densities of the temporal and spatial components read : @xmath82 notice that we need @xmath83 in order to have positive energy density for @xmath30 .
in fact , it is possible to show that imposing canonical normalization for the corresponding creation and annihilation operators we get @xmath84 @xcite . besides , when inserting the growing modes of the fields into these expressions we obtain that @xmath85 , @xmath86 and @xmath87 const .
thus , the field behaves as a cosmological constant throughout the evolution of the universe since its temporal component gives rise to a constant energy density whereas the energy density corresponding to @xmath88 always decays as radiation .
moreover , this fact prevents the generation of a non - negligible anisotropy which could spoil the highly isotropic cmb radiation .
finally , when the universe is dominated by the electromagnetic field , both the hubble parameter and @xmath30 become constant ( one can straightforwardly check that this is a solution of the complete system of equations ) so the energy density is also constant and the electromagnetic field behaves once again as a cosmological constant leading therefore to a future de sitter universe . as the observed fraction of energy density associated to a cosmological constant today is @xmath89 , we obtain that the field value today must be @xmath90 .
the effects of the high electric conductivity @xmath23 can be introduced using the magneto - hydrodynamical approximation and including the current term @xmath91 on the r.h.s . of maxwell s equations . notice that
because of the universe electric neutrality , conductivity does not affect the evolution of @xmath92 .
the infinite conductivity limit simply eliminates the growing mode of @xmath93 in ( [ azsol ] ) . we still need to understand which are the appropriate initial conditions leading to the present value of @xmath30 . in order to avoid the cosmic coincidence problem , such initial conditions should have been set in a natural way in the early universe . in a very interesting work @xcite , it was suggested that the present value of the dark energy density could be related to physics at the electroweak scale since @xmath94 , where @xmath95 gev .
this relation offers a hint on the possible mechanism generating the initial amplitude of the electromagnetic fluctuations .
indeed , we see that if such amplitude is set by the size of the hubble horizon at the electroweak era , i.e. @xmath96 , then the correct scale for the dark energy density is obtained .
thus , using the friedmann equation , we find @xmath97 , but according to ( [ rhoa0 ] ) , @xmath98const .
, so that @xmath99 as commented before .
a possible implementation of this mechanism can take place during inflation .
notice that the typical scale of the dispersion of quantum field fluctuations on super - hubble scales generated in an inflationary period is precisely set by the almost constant hubble parameter during such period @xmath100 , i.e. @xmath101 @xcite .
the correct dark energy density can then be naturally obtained if initial conditions for the electromagnetic fluctuations are set during an inflationary epoch at the scale @xmath102 . despite the fact that the background evolution in the present case is the same as in @xmath0cdm , the evolution of metric perturbations could be different , thus offering an observational way of discriminating between the two models .
in fact , the evolution of the scalar perturbation @xmath103 with respect to the @xmath0cdm model gives rise to a possible discriminating contribution to the late - time integrated sachs - wolfe effect @xcite .
the propagation speeds of scalar , vector and tensor perturbations are found to be real and equal to the speed of light , so that the theory is classically stable .
we have also checked that the theory does not contain ghosts and it is therefore stable at the quantum level . on the other hand , using the explicit expressions in @xcite for the vector - tensor theory of gravity corresponding to the action in ( [ action ] ) , it is possible to see that all the parametrized post - newtonian ( ppn ) parameters agree with those of general relativity , i.e. the theory is compatible with all the local gravity constraints for any value of the homogeneous background vector field @xcite .
the presence of large scale electric fields generated by inhomogeneities in the @xmath30 field opens also the possibility for the generation of large scale currents which in turn could contribute to the presence of magnetic fields with large coherence scales .
this could shed light on the problem of explaining the origin of cosmological magnetic fields .
work is in progress in this direction .
we have shown that vector theories offer a simple and accurate description of dark energy in which the coincidence problem could be easily avoided . in our first example , the scaling behaviour during radiation and the natural initial conditions for the vector field offer a neat way around the problem . moreover , in our second example , the presence of a cosmological electromagnetic field generated during inflation provides a natural explanation for the cosmic acceleration .
this result not only offers a solution to the problem of establishing the true nature of dark energy , but also explains the value of the cosmological constant without resorting to new physics . in this scenario
the fact that matter and dark energy densities _ coincide _ today is just a consequence of inflation taking place at the electroweak scale .
present and forthcoming astrophysical and cosmological observations will be able to discriminate these proposals from the standard @xmath0cdm cosmology .
_ acknowledgments : _ this work has been supported by dgicyt ( spain ) project numbers fpa 2004 - 02602 and fpa 2005 - 02327 , ucm - santander pr34/07 - 15875 and by cam / ucm 910309 .
aknowledges support from mec grant bes-2006 - 12059 .
99 c. wetterich , _ nucl .
phys . _ * b302 * , 668 ( 1988 ) ; r.r .
caldwell , r. dave and p.j .
steinhardt , _ phys .
lett . _ * 80 * , 1582 ( 1998 ) c. armendariz - picon , t. damour and v. mukhanov , _ phys . lett . _
* b458 * , 209 ( 1999 ) s.m .
carroll , v. duvvuri , m. trodden , m.s .
turner , _ phys .
_ * d70 * : 043528 , ( 2004 ) g. dvali , g. gabadadze and m. porrati , _ phys .
lett . _ * b485 * , 208 ( 2000 ) j. beltrn jimnez and a.l .
maroto , _ phys .
* d78 * , 063005 ( 2008 ) and arxiv:0807.2528 [ astro - ph ] j. beltrn jimnez and a.l .
maroto , arxiv:0811.0566 [ astro - ph ] s. nojiri , s. d. odintsov and s. tsujikawa , _ phys .
_ * d71 * ( 2005 ) 063004 ; m. bouhmadi - lpez , p. f. gonzlez - daz and p. martn - moruno , _ phys .
lett . _ * b659 * ( 2008 ) 1 a.g .
riess at al .
_ astrophys.j .
_ * 607 * , 665 ( 2004 ) p. astier et al . , _ astron .
astrophys . _ * 447 * : 31 - 48 , ( 2006 ) .
s. nesseris and l. perivolaropoulos , _ jcap _ * 0702 * : 025 , ( 2007 ) .
r. trotta and r. bower , _ astron .
geophys . _ * 47 * : 4:20 - 4:27 , ( 2006 ) r. lazkoz , s. nesseris , l. perivolaropoulos , _ jcap _ * 0511*:010 , ( 2005 ) e. babichev , v. mukhanov and a. vikman , jhep * 0802 * , 101 ( 2008 ) c. will , _ theory and experiment in gravitational physics _ , cambridge university press , ( 1993 ) c. itzykson and j.b .
zuber , _ quantum field theory _ , mcgraw - hill ( 1980 ) n. arkani - hamed , l. j. hall , c. f. kolda and h. murayama , _ phys .
lett . _ * 85 * ( 2000 ) 4434 a. linde , _ particle physics and inflationary cosmology _ , harwood academic press ( 1996 ) r.g .
crittenden and n. turok , _ phys .
lett . _ * 76 * ( 1996 ) 575 j. beltrn jimnez and a.l .
maroto , arxiv:0811.0784 [ astro - ph ] | we show that vector theories on cosmological scales are excellent candidates for dark energy .
we consider two different examples , both are theories with no dimensional parameters nor potential terms , with natural initial conditions in the early universe and the same number of free parameters as @xmath0cdm .
the first one exhibits scaling behaviour during radiation and a strong phantom phase today , ending in a `` big - freeze '' singularity .
this model provides the best fit to date for the snia gold dataset . the second theory we consider is standard electromagnetism .
we show that a temporal electromagnetic field on cosmological scales generates an effective cosmological constant and that primordial electromagnetic quantum fluctuations produced during electroweak scale inflation could naturally explain , not only the presence of this field , but also the measured value of the dark energy density .
the theory is compatible with all the local gravity tests , and is free from classical or quantum instabilities .
thus , not only the true nature of dark energy could be established without resorting to new physics , but also the value of the cosmological constant would find a natural explanation in the context of standard inflationary cosmology .
epsf address = departamento de fsica terica , universidad complutense de madrid , 28040 madrid , spain address = departamento de fsica terica , universidad complutense de madrid , 28040 madrid , spain |
gauged supergravities in four dimensions have been the subject of a renewed interest in recent time , especially because of their connection with higher dimensional theories compactified on manifolds which allow fluxes of p - forms along the internal directions .
a particularly appealing class of such theories is given by type iia and iib compactifications on ricci - flat manifolds which , in absence of fluxes , leave @xmath4 unbroken supersymmetries in four dimensions . turning on fluxes develops a scalar potential in these theories , which usually gives moduli stabilization , reduced supersymmetry and , in certain cases , leaves a vanishing cosmological constant @xcite@xcite .
in fact , theories with vanishing cosmological constant are generalized no - scale models , which were studied long ago in the pure supergravity framework @xcite .
another interesting class of models recently studied in the literature is given by compactifications on @xmath5 and @xmath6 of iib orientifolds with three - form fluxes turned on @xcite .
these theories have @xmath7 and @xmath8 unbroken supersymmetries respectively , and the presence of fluxes may give rise to partial supersymmetry breaking @xmath9 with vanishing vacuum energy . due to the precise knowledge of the moduli space of these theories , an exact estimate of the ( tree - level ) scalar potential is possible in all these models .
the flux - compactifications share the feature that their low energy description is given in terms of an extended supergravity with mass deformations and a scalar potential @xcite . in the supergravity framework
this can be achieved through the gauging of some of the isometries of the non linear @xmath10-model spanned by the scalar sector .
indeed , the relevant isometries which are gauged in presence of fluxes are those associated to axion symmetries of those scalars coming from r - r forms @xmath11 ( @xmath12 ) and from the n - s two - form @xmath13 field .
in fact the latter only contributes in the generalized case of orientifolds of @xmath5 where the orientifold projection acts only on some directions of the @xmath5 torus @xcite .
a similar phenomenon occurs in the scherk
schwarz generalized compactification of m - theory @xcite . indeed ,
schwarz spontaneously broken supergravity can be shown to be completely equivalent to a gauged supergravity with a non semisimple gauge group given by the semi - direct product of a @xmath14 factor , gauged by the kaluza klein vector , times axionic symmetries corresponding to isometries in the scalar sector coming from five dimensional vectors @xcite ( see @xcite also for the construction of the five dimensional gauged supergravity which describes the spontaneously broken model deriving from @xmath15 scherk
schwarz dimensional reduction ) .
the @xmath14 symmetry is related to the central charge of the bps massive representations of the spontaneously broken theory @xcite . in all
the above mentioned compactifications , the underlying supergravity relies on a gauge group which is a certain subgroup of the duality group having a linear ( adjoint ) action on the vector fields of the theory .
since the most general duality transformation in @xmath16 is an element of @xmath17 ( where @xmath18 denotes the number of vectors in the theory ) , the gauge group must be a _
lower block - triangular _ symplectic matrix , in order not to mix electric with magnetic fields .
the same ungauged @xmath19-extended supergravity can therefore have several inequivalent deformations , corresponding to different choices of the gauge group .
the latter is chosen by selecting , among the duality symmetries , the ones which are realized linearly on the vector fields of the theory .
according to gaillard and zumino @xcite , the most general linear transformation between electric and magnetic field strengths @xmath20 ( @xmath21 ) , in a given four dimensional theory with vectors coupled to scalar fields , is a @xmath22 transformation , acting on the self - dual components @xmath23 is defined by @xmath24 as ^= , [ dual]where the matrix @xmath25 is symplectic . , which implies , by choosing as symplectic metric @xmath26 , that @xmath27 , @xmath28 are symmetric and @xmath29 . passing to the algebra , for @xmath30 , @xmath31 , this implies @xmath32 , @xmath33 , @xmath34 .
] if we define a complex symmetric matrix @xmath35 through the equation g^+_= _
f^+then under a duality rotation ( [ dual ] ) @xmath36 gets transformed to ^= ( c+d ) ( a+b)^-1 .
the matrix @xmath36 appears in the vector lagrangian as = 2 ( _ f^+ f^+)= 2(g^+_f^+).for infinitesimal transformations we have f^+ & = & a^
_ f^+ + b^ g^+ _ , b^= b^ ; + g^+ _ & = & c_f^+ -a^ _
g^+ _ , c_= c _ ; + & = & c - a^t - a - b .
the most general subgroup of @xmath22 leaving the lagrangian invariant up to a total derivative is obtained by setting @xmath37 . in this case
the vector fields @xmath38 ( such that @xmath39 ) transform as a^= a^ _ a^and = c - a^t - a . a duality transformation can be chosen as gauge symmetry if it can be found as a certain subalgebra of those duality symmetries which leave the action invariant , that is only if @xmath40 .
axion symmetries have @xmath41 . in this case
the gauge group can be non - abelian or abelian depending on whether the block @xmath42 is different from zero or not .
moreover , axion isometries are usually embedded in nilpotent groups , and this implies that the matrix @xmath42 itself is also lower ( or upper ) triangular .
@xmath10-models coupled to vectors have a richer structure than ordinary @xmath10-models . for coset spaces
@xmath43 , @xmath44 acts on the vector field strengths and their duals in a symplectic representation . in particular
, in this representation the coset representatives are symplectic matrices @xmath45 through which , by defining f ( ) = 12(a - ib ) , h ( ) = 12(c - id ) , one gets @xmath46 as an explicit , scalar dependent , expression for the @xmath36 matrix .
this formula allows the computation of the vector lagrangian from first principles .
a simplification here occurs by choosing a solvable parametrization of the coset in the symplectic representation . in this case @xmath47 and @xmath48 , so that = a^-1 = ca^-1 - i ( aa^t)^-1 .
let us now suppose to gauge some subalgebra of the @xmath49 generators which has an adjoint action on the vectors ( just trivial in the abelian case ) .
then , in such case a^
_ = f^ _ ^ ; c _ = c_,^ where the constants @xmath50 satisfy @xmath51 and an additional cocycle condition in the non abelian case , where @xmath52 : @xmath53\delta}+c_{\gamma\delta,[\sigma } f^\gamma_{\lambda]\omega}&=&0\end{aligned}\ ] ] these conditions appear because of the fact that , when @xmath54 generators ( axions ) are gauged , a gauge- and supersymmetry - invariant lagrangian can be found only after adding to the lagrangian a chern - simons type term , proportional to @xmath50 @xcite .
the above is the general structure of the gauge transformations in extended supergravity when the axion symmetries are gauged .
this is a feature common to all orientifold models with fluxes turned on as well as to models obtained from scherk
schwarz generalized dimensional reduction .
in the present section we shall briefly review the construction , which was accomplished in @xcite , of some new four dimensional orientifold models ( both in type iib and iia ) where the orientifold projection involves an orbifold projection with respect to the space inversion @xmath55 on @xmath56 coordinates , transverse to the @xmath57-brane world volume . since in this construction
the @xmath57brane world volume fills the non - compact space - time , the torus @xmath5 is naturally split into neumann directions @xmath58 and dirichlet directions @xmath59 .
these models therefore will be denoted by @xmath60 .
their low - energy descriptions ( in absence of fluxes and @xmath61branes ) are all given in terms of a four dimensional ungauged @xmath62 supergravity coupled to six vector multiplets from the closed string sector .
in the following discussion we shall always restrict ourselves to the bulk degrees of freedom .
the scalar fields span a manifold with the following @xmath43 form : @xmath63 the duality group being @xmath64 .
we summarize in table 1 the bosonic field content of these models in the type iib and iia settings respectively ( in our conventions the directions of @xmath5 are labeled by an index @xmath65 whereas the directions of @xmath58 and of @xmath59 are labeled respectively by indices @xmath66 and @xmath67 and the non compact space time directions by greek letters ) . 0.5 cm .massless degrees of freedom for the iib / iia orientifolds [ cols=">,^,^,^",options="header " , ] the @xmath62 orientifold models ( in the absence of fluxes ) can be consistently constructed as truncations of the unique four dimensional @xmath68 supergravity which describes the low - energy limit of toroidally compactified type ii superstrings . in the scalar - field sector
this amounts to defining the embedding of the @xmath69 scalar manifold into the @xmath68 one .
the duality group of the latter theory is @xmath70 which acts non - linearly on the @xmath71 scalar fields , and linearly , as a @xmath72 symplectic transformation , on the @xmath73 electric field strengths and their magnetic dual . within this theory an intrinsic group - theoretical characterization of the ten dimensional origin of the scalar and vector fields
can be achieved . in the so - called solvable
lie algebra representation of the scalar sector @xcite , the scalar manifold @xmath74 is expressed as the group manifold generated by the solvable lie algebra @xmath75 defined by the iwasawa decomposition of the @xmath76 algebra : @xmath77 in this framework , there is a natural one - to - one correspondence between the scalar fields and the generators of @xmath75 .
the latter consist of the @xmath76 cartan generators @xmath78 , parametrized by the @xmath5 radii @xmath79 together with the dilaton @xmath80 , and of the shift generators corresponding to the @xmath81 positive roots @xmath82 of @xmath76 , which are in one - to - one correspondence with the axionic scalars that parametrize them
. this correspondence between cartan generators and positive roots on one side and scalar fields on the other , can be pinpointed by decomposing @xmath75 with respect to some relevant groups @xcite .
the axions deriving from ten dimensional tensor fields ( i.e. the kalb
ramond form @xmath83 and the r
r fields ) transform in tensorial representations of the @xmath84 isometry group of the @xmath5 metric moduli and so do the corresponding solvable generators .
if we express the @xmath76 roots with respect to an orthonormal basis of euclidean vectors @xmath85 , @xmath86 , the precise correspondence between axions and @xmath76 nilpotent generators reads : @xmath87 where @xmath88 the torus radii @xmath89 and the ten dimensional dilaton @xmath80 enter this description in the following combination with the cartan generators : @xmath90 so that the kinetic term of the axion corresponding to the root @xmath82 contains the correct exponential factor @xmath91 .
as far as the scalar sector is concerned , the embedding of the @xmath62 orientifold models @xmath60 ( in absence of fluxes ) inside the @xmath92 theory ( in its type iia or iib versions ) is defined by the condition that the embedding of the @xmath93 duality group @xmath94 inside the @xmath70 fulfill the following condition : @xmath95 condition ( [ inter ] ) amounts to requiring that the @xmath5 metric moduli in our models are related either to the @xmath58 metric @xmath96 or to the @xmath59 metric @xmath97 .
it fixes the ten dimensional interpretation of the fields in the ungauged @xmath62 models ( except for the cases @xmath98 and @xmath99 ) , and for a given @xmath100 it is consistent with the bosonic spectrum resulting from the corresponding orientifold reductions . in the @xmath98 and @xmath99 cases ,
the two embeddings are characterized by a different interpretation of the scalar fields , related to the @xmath101 orientifold reduction in the presence of d3 or d9 branes .
the axions not deriving from the internal metric consist of @xmath102 in the external @xmath103 factor , @xmath104 moduli @xmath105 in the bi - fundamental of @xmath106 and @xmath107 r r moduli which we shall generically denote by @xmath108 and which span the maximal abelian ideal @xmath109 of @xmath110 .
the scalars @xmath111 and @xmath108 parametrize a @xmath112 dimensional subalgebra @xmath113 of @xmath110 consisting of nilpotent generators only . switching on fluxes amounts in the four dimensional theory to introducing a suitable gauge group contained inside @xmath113 .
let us first list for the various models the generator content of @xmath113 and their algebraic structure . _
@xmath114 model _
in this case @xmath115 is abelian and its generic element is @xmath116 . _
@xmath117 model _ the algebra @xmath118 is 23dimensional and has the following structure : @xmath119\,=\ , t^{iabc}\,,\,\,\,\,\,\left[t^{ia}_b,\ , t^{jbcd}\right]\,=\,\epsilon^{ij}\ , \epsilon^{abcd}\ , t.\end{aligned}\ ] ] _ @xmath120 model _ the algebra @xmath121 is 23dimensional and has the following structure : @xmath122\,=\ , t^{ia}\,,\,\,\,\,\left[t^{ia}_b,\ , t^{jb}\right]\,=\,t^{ijab}.\end{aligned}\ ] ] _ @xmath123 model _ in this case @xmath124 is abelian and its generic element is @xmath125 . _
@xmath126 model _ the algebra @xmath127 is 20dimensional and has the following structure : @xmath128\,=\ , \epsilon^{abcde}t_{e}.\end{aligned}\ ] ] _ @xmath129 model _ the algebra @xmath130 is 24dimensional and has the following structure : @xmath131\,=\ , t^{iab}\,,\,\,\,\,\left[t_b^{ia},\ , t^{jbc}\right]\,=\,\epsilon^{ijk}\,\epsilon^{abc}\ , t_{k}.\end{aligned}\ ] ] _ @xmath132 model _ the algebra @xmath133 is 20dimensional and has the following structure : @xmath134\,=\ , t^{ij6}.\end{aligned}\ ] ] 0.5 cm _ fluxes and gaugings : a preliminary analysis _ let us consider the @xmath117 model in presence of the fluxes @xmath135 and @xmath136 .
these fluxes appear as structure constants @xmath137 & = & \epsilon_{ij}\ , h_a\ , x^a\ , , \nonumber \\
\left[x_i,\,x^a\right ] & = & f_i^{ab}\ , x_b \ , , \label{algebra2}\end{aligned}\ ] ] of a gauge algebra @xmath138 with connection @xmath139 , all other commutators vanishing .
the identification @xmath140 of the gauge generators with the isometries of the solvable algebra ( in our conventions @xmath141 and @xmath142 ) , reproduces only a contracted version of the algebra ( [ algebra2 ] ) in which three of the central charges @xmath143 vanish and we are left with @xmath144 .
if we denote by @xmath145 these three central generators , we see that the subgroup @xmath146 of the isometry group which is gauged coincides with the quotient : @xmath147 that amounts to imposing the vanishing of the central terms @xmath148 on all fields . one can verify that the vectors @xmath149 , @xmath150 and @xmath151 transform in the co - adjoint representation of @xmath152 .
the non - abelian field strengths are : @xmath153 and the covariant derivatives for the scalar fields read : @xmath154 as far a the @xmath120 model is concerned let us consider the gauging which corresponds to turning on the fluxes @xmath155 .
it is useful to describe the @xmath156 and @xmath157 forms as elements of an @xmath158 doublet labeled by an index @xmath159 , the couple of indices @xmath160 span the representation @xmath161 of @xmath162 .
we shall then denote the flux
forms by @xmath163 where @xmath164 is the index of the @xmath161 in which the invariant metric is diagonal @xmath165 .
similarly the generators @xmath166 will be denoted by @xmath167 and the vectors @xmath168 by @xmath169 .
inspection of the dimensionally reduced three form kinetic term indicates for the four dimensional theory a gauge group @xmath44 with connection @xmath170 and the following structure : @xmath137&= & h_{ij}^\lambda x_\lambda .
\label{galg}\end{aligned}\ ] ] if we identify the gauge generators with isometries as follows:@xmath171 then it can be shown that they close the algebra ( [ galg ] ) only if @xmath172 which amounts to requiring that @xmath173 ( this condition is consistent with a constraint found in @xcite on the embedding matrix of a new gauge group in the @xmath174 theory , which seems to yield an @xmath175 `` lifting '' of the type iib orientifold models @xmath60 discussed here ) .
the vectors transform indeed in the co adjoint of @xmath44 and the non abelian field strengths are : @xmath176 the relevant covariant derivatives for the axions read : @xmath177\lambda}\ , , \nonumber\\ d_\mu\phi^\lambda_i & = & \partial_\mu \phi^\lambda_i -h^\lambda_{ij}\,g_\mu^j \,.\end{aligned}\ ] ]
in this section we are going to describe the basic features of the supersymmetry breaking pattern induced by the presence of non trivial 3-form fluxes in two orientifold compactifications of iib superstring .
in particular , we will focus on the bulk sector of the theories which is the one responsible for the supersymmetry breaking and moduli stabilization , without considering here the d - brane degrees of freedom . before entering into the details ,
let us briefly summarize the conditions for the vacuum to preserve some unbroken supersymmetry .
they are exhausted by the request that , in the vacuum , the supersymmetry transformation laws of all the fermions @xmath180 are zero for the corresponding supersymmetry parameter @xmath181 , that is < f > = 0 \{. where @xmath182 are the gravitini and @xmath183 generic spin 1/2 fields , while @xmath184 and @xmath185 denote the gravitino mass matrix and fermionic shifts respectively .
@xmath186 is the scalar potential defined by @xmath187 both the models considered in this section have a positive definite scalar potential , allowing vacuum configurations with partial super higgs and zero vacuum energy .
this is related to the fact that , in both cases , the gauging responsible for the presence of a scalar potential corresponds to switching on charges only for translational isometries of the scalar manifolds .
it only depends , in fact , on universal properties leading to cancellation of positive and negative contributions in the scalar potential as it occurred in @xmath188 no - scale models @xcite the bulk sector of the @xmath189 orientifold model is described by @xmath7 supergravity coupled to 6 vector multiplets , where , as discussed in section 3 , the 38 scalars parametrize the coset manifold @xmath190 $ ] ( for a derivation of this model from dimensional reduction of type iib theory see @xcite and references therein ) .
its gauged supergravity description ( first constructed in @xcite ) relies on an abelian gauge algebra which corresponds to a 12 dimensional subalgebra of the 15 shift symmetries of the scalars coming from the r - r 4-form @xmath191 ; @xmath192 .
the gauge vectors @xmath193 ( with @xmath194 ) come from the n - s and r - r two forms with one internal index . in terms of the @xmath195 the covariant derivatives are _ b^=_b^ - f^ _ a^_where @xmath196 are the real n - s and r - r three - form fluxes . in presence of the fluxes @xmath197 ,
the theory develops a scalar potential . in order for the potential to have a minimum with vanishing cosmological constant
, it is required that the fluxes are subject to the constraint f^-_1 = i f^-_2 [ constr]where @xmath82 is a complex constant ( the minus apex denotes the anti self - dual projection in the internal manifold ) .
condition ( [ constr ] ) breaks the symmetry @xmath198 to @xmath199 . in the vacuum ,
the @xmath200 variable @xmath201 is actually fixed at the value @xmath202 , that is , in terms of the n - s and r - r iib dilatons @xmath203 e^= , c = - ( > 0 ) , [ dil ] where @xmath204 is the complex component of the coset representative satisfying the @xmath205 condition @xmath206 . equation ( [ dil ] ) also implies that the n - s and r - r fluxes must both be present to have dilaton stabilization , so that the solution is non perturbative at the string level .
the basic features of the model are encoded in the two matrices f^ijk l^f_^ijk , |f^ijk |l^f_^ijk where @xmath207 and @xmath208 are the coset representatives of @xmath209 $ ] . indeed , by using the @xmath210 spin(@xmath211 ) @xmath212-matrices one can define the two complex symmetric matrices s_ab & = & - |f^-ijk(_ijk)_ab + n^ab & = & |f^+ijk(_ijk)^ab where @xmath184 is the gravitino mass matrix and @xmath213 the fermionic shift in the dilatino supersymmetry transformation law .
they have the same @xmath14 charge , but contra - gradient @xmath214 representations @xmath215 respectively . the scalar potential is actually given in terms of @xmath213 : v for generic values of the fluxes ( still constrained by equation ( [ constr ] ) ) equation @xmath216 stabilizes the dilaton @xmath217 and axion @xmath218 together , while fixing @xmath208 to the diagonal form diag@xmath219 .
there is no residual supersymmetry in this case .
some residual supersymmetry is preserved when some extra constraints among the fluxes are satisfied . by adopting complex
coordinates @xmath220 , @xmath221 ( so that @xmath222 ) , the four gravitino masses are given by |f_ijk| , |f_i||k| ; ij k there are in fact four different values of this type .
when any one of them vanishes , then one supersymmetry remains unbroken . in this case
some of the @xmath208 moduli remain unfixed . for instance , for preserving @xmath223 supersymmetry it is needed that @xmath224 , while only @xmath225 .
then the @xmath96 entries of the scalar metric are fixed to zero , while the @xmath226 remain unfixed . correspondingly , the six axions @xmath227 are eaten by six of the vectors , while the remaining nine axions @xmath228 remain massless .
the resulting @xmath223 moduli space is @xmath229 $ ] , as predicted by @xmath223 supergravity .
there is a single massive spin 3/2 multiplet , which is long ( not bps saturated ) and this is in agreement with the fact that there are not gauged central charges in this model .
this is to be contrasted to what happen with the scherk
schwarz @xmath230 spontaneously broken supergravity , where all the massive multiplets are 1/2 bps saturated .
let us now turn to discuss the @xmath231 orientifold model , which has been studied in the literature both as a compactification from ten dimensional iib theory as well as a four dimensional gauged supergravity . from a four dimensional point of view
it is an @xmath8 supergravity model coupled to 3 vector multiplets and 20 hypermultiplets .
the scalars of the vector multiplets parametrize the special khler manifold _ v = ^3 = while the scalars in the hypermultiplets parametrize the quaternionic manifold _
h = .the latter can be regarded as a fibration over @xmath232 = ^++ * 22*where the presence on @xmath233 of 22 translational isometries @xmath234 , ( @xmath235 ; @xmath236 ) corresponding to the degrees of freedom of the 10 dimensional r - r 4-form , with two indices on the @xmath237 and two on the torus , is put in evidence .
switching on three - form fluxes corresponds to gauging some of the 22 translational isometries of @xmath233 by some of the vectors @xmath238 ( @xmath239 ) in the theory , where @xmath240 , @xmath241 , @xmath242 .
the covariant derivatives are @xmath243 here , the couplings @xmath244 correspond to the n - s and r - r three - form fluxes with one index on the torus and two on @xmath237 .
the presence of fluxes allows step - wise partial super - higgs @xmath245 with zero vacuum energy .
@xmath8 and @xmath246 vacua correspond to two gaugings different in the choice of quaternionic isometries and of gauged vectors . in all cases
many of the moduli are stabilized . to obtain configurations with @xmath8 supersymmetry
one should switch on the fluxes @xmath247 , taking as would be goldstone bosons the scalars @xmath248 . in this case
the vectors which become massive are the vector partner of iib dilaton and of @xmath249 complex structure moduli .
in fact , two of the original massless hypermultiplets and the two vector multiplets of @xmath250 and @xmath251 combine into two long massive vector multiplets @xmath252 $ ] .
we see that the @xmath8 configurations are just an example of the higgs phenomenon of two vector multiplets .
the residual moduli space is .let us note that to have @xmath8 preserving vacua it is not possible to gauge more than two vectors , since it would be incompatible with the given symplectic frame , used to reproduce the orientifold configuration .
it is easy to see that the given choice is the only one stabilizing the moduli in a way compatible with the domain of definition of the scalar fields .
this same result is derived in section 5 of @xcite , with topological arguments . on the contrary , to obtain configurations with @xmath253 supersymmetry one should switch on , at least , the couplings @xmath254 corresponding to the scalar isometries @xmath255 . the massive vectors are in this case the graviphoton and vector partner of the @xmath256 volume modulus . for general values of the fluxes ,
the supersymmetry is completely broken ; the @xmath188 case is obtained by imposing further @xmath257 the massless spectrum of the @xmath188 reduced theory is the following . from the 58 scalars of @xmath258
there remain 20 scalars parametrizing @xmath259 . from the 22 axions there
remain 20 . all together
they complete the scalar content of 20 chiral multiplets .
the spectrum includes two massless vector multiplets corresponding to @xmath260 and @xmath261 and an extra chiral multiplet whose scalar field comes from the @xmath8 vector multiplet sector .
there is then one long massive gravitino multiplet @xmath262 $ ] , containing as vectors the graviphoton @xmath263 and the vector @xmath264 .
note that this @xmath8 model allows partial super higgs , evading the no - go theorem for @xmath8 supergravity @xcite .
this is possible because the symplectic frame chosen such as to reproduce the string model given uses a degenerate symplectic section for special geometry on @xmath265 , where no prepotential function @xmath266 exists .
more general vacua , preserving @xmath253 supersymmetry , can be obtained by considering an arbitrary vector coupling @xmath267 . in this more general case ,
@xmath268 and @xmath269 gauge two of the isometries @xmath270 , while @xmath271 and @xmath272 gauge two of the isometries @xmath273 .
the @xmath188 preserving vacua have 18 left over chiral multiplets .
this report is based on collaborations with c. angelantonj , r. dauria , f. gargiulo , m. a. lled and s. vaul , which we would like to thank .
would like to thank h. samtleben for useful discussions .
m.t . would like to thank the th .
division of cern , where part of this work has been done , for their kind hospitality .
the work of s.f . has been supported in part by european community s human potential program under contract hprn - ct-2000 - 00131 quantum space - time , in association with infn frascati national laboratories and by d.o.e .
grant de - fg03 - 91er40662 , task c. the work of m.t . is supported by a european community marie curie fellowship under contract hprn - ct-2001 - 01276 . | we report on the gauged supergravity interpretation of certain compactifications of superstring theories with p - form fluxes turned on . + we discuss in particular the interplay of duality symmetries in type iib orientifolds and gauged isometries in the corresponding supergravity models .
+ turning on fluxes is generally described by the gauging of some nilpotent lie group whose generators correspond to axion symmetries of r - r and n - s scalars . #
1#2#3 figurecaption#1#2 makefnmark@xmath0 1.5pt # 1#2#3#4#1 * # 2 * , # 3 ( # 4 ) 1.5pt 4.5 cm 5.5 cm * fluxes , supersymmetry breaking and gauged supergravity * + 0.5 cm _ @xmath1 cern , theory division , ch-1211 geneva 23 , switzerland , _ + e - mail : [email protected] + _ @xmath2 cern , theory division , ch-1211 geneva 23 , switzerland and _ + _ laboratori nazionali di frascati , infn , italy , _ + e - mail : [email protected] + _ @xmath3 spinoza institute , leuvenlaan 4 nl-3508 , utrecht , the netherlands , _ + e - mail : [email protected] contribution to the proceedings of `` sugra20 '' conference , department of physics , northeastern university , boston ( ma ) 02115 usa ; march 2003 |
here , we study the cdw formation in the single @xmath0-orbital hubbard model . as far as only repulsive coulomb interaction
is taken into account , the vc driven cdw can not emerge in the single - orbital hubbard model .
however , we show that the cdw is formed due to the cooperation of the vc and the @xmath198-ph interaction .
the band - dispersion is @xmath215 , where @xmath216 ev , @xmath217 and @xmath218 for ybco @xcite . in the random - phase - approximation ( rpa ) without the vc , the spin ( charge ) susceptibility is given as @xmath219 : @xmath220 is the on - site coulomb interaction , @xmath221 is the bare bubble , and @xmath222 . here and hereafter ,
@xmath57 and @xmath58 , where @xmath59 and @xmath60 .
figure [ fig : figs1 ] ( a ) shows @xmath223 for @xmath224 ev , @xmath26 and @xmath225 ev . the spin stoner factor @xmath226 is 0.99 . in contrast , @xmath227 is suppressed by @xmath220 within the rpa . in one - orbital hubbard model in the rpa for @xmath117 .
( b ) @xmath228 , and ( c ) @xmath44 and @xmath229 .
( d ) @xmath230 given by the sc - vc method with the @xmath198-ph interaction . ] next , we discuss the charge susceptibility beyond the rpa by taking the vc into account . it is given as @xmath231 where @xmath232 , and @xmath233 is the al - type vc @xcite . at @xmath234 ,
the al - term is given as @xmath235 where @xmath75 .
figure [ fig : figs1 ] ( b ) shows the obtained @xmath228 for @xmath224 ev ( @xmath117 ) , which shows the maximum at @xmath88 and @xmath89 , and the relation @xmath236 is satisfied . the cdw instability will be caused by the large @xmath237 , which is much larger than @xmath238 . here
, the @xmath46-dependence of @xmath228 is mainly given by the three - point vertex .
to show this , we approximate the al - type vc for @xmath239 as @xmath240 where @xmath92 and @xmath241 .
figure [ fig : figs1 ] ( c ) shows the @xmath46-dependences of @xmath242 and @xmath44 along the @xmath93-axis .
it is apparent that the large peak of @xmath228 at @xmath103 originates from the three - point vertex .
although the charge vc becomes very large in the presence of strong spin fluctuations , @xmath230 in eq .
( [ eqn : chic - one ] ) can not exceed @xmath243 .
in real compounds , however , strong attractive interaction @xmath244 due to the buckling mode with @xmath245 had been predicted by the first principle study @xcite . in this case , @xmath220 in eq .
( [ eqn : chic - one ] ) is replaced with @xmath246 , and then @xmath247 can be strongly enlarged if @xmath248 is negative . that is , _
the cdw instability due to the @xmath198-ph interaction is effectively enlarged by the spin - fluctuation driven al - type vc . _ in fig .
[ fig : figs1 ] ( d ) , we show the strong developed @xmath230 obtained for @xmath224 ev and @xmath249 ev . in the main text , we show that the cdw due to the vc is much easily realized in the three - orbital @xmath0-@xmath1 model with degenerate @xmath7 and @xmath8 orbitals , _ without introducing the @xmath198-ph interactions_.
in the main text , we studied the three - orbital model with on - site coulomb interactions ( @xmath27 , @xmath28 ) and inter - site coulomb interaction @xmath29 between the nearest cu - o sites . here
, we introduce the coulomb interaction between the nearest o - sites , @xmath250 , and discuss the vc - driven cdw .
-@xmath250 plane at @xmath62 , in the case of @xmath117 .
( b ) the form factor on the cdw boundary as function of @xmath251 . ]
figure [ fig : figs2 ] ( a ) shows the cdw boundary ( @xmath108 ) in the @xmath29-@xmath250 plane at @xmath62 , in the case of @xmath117 ( @xmath252 ev ) . here
, @xmath253 corresponds to the main text .
thus , the nematic cdw at @xmath122 is realized in the presence of small @xmath29 and/or @xmath250 , by taking the al - term into account .
figure [ fig : figs2 ] ( b ) shows the form factor on the cdw boundary , @xmath254 , which is given by the eigenvector of @xmath125 for the largest eigenvalue .
the charge density modulation ( @xmath255 ) at @xmath103 is proportional to @xmath256 .
we see that the form factor sensitively depends on the ratio @xmath251 .
we stress that the cdw order with @xmath257 is realized when @xmath250 is comparable or larger than @xmath258 .
recently , the cdw form factors are determined in several compounds by using the resonant @xmath259-ray measurements . by comparing the experimental cdw form factors and fig .
[ fig : figs2 ] ( b ) , we can determine the ratio @xmath251 in each real compound .
the cdw order parameter @xmath260 ( @xmath261 ) is given by @xmath262 e^{i\e_n\cdot 0 } , \end{aligned}\ ] ] where @xmath263 is the green function . by expanding the green function with respect to @xmath264
, we obtain @xmath265 where @xmath266 : here , @xmath267 is the bare susceptibility , and @xmath268 is the charge vc beyond the mean - field approximation , given by the ward identity with respect to the self - energy .
@xmath269 is the bare coulomb interaction for the charge sector introduced in the main text .
then , the linearized equation with respect to @xmath270 is @xmath271 where the eigenvalue @xmath272 reaches unity at @xmath273 .
therefore , the form factor @xmath274 at @xmath273 is given by the eigenvector of @xmath275 for the largest @xmath276 , which is the charge stoner factor .
when @xmath277 , it is easy to show that the form factor @xmath278 is approximately equal to the eigenvector of the charge susceptibility @xmath279^{-1}{\hat \phi}^c(\q)$ ] for the largest eigenvalue . in other words , under the condition @xmath280 , the maximum of the inner product @xmath281 is realized for @xmath282 in the case of @xmath283 , if the largest eigenvalue of @xmath275 is nondegenerate .
in the main text , we studied the role of the vc for the cdw formation in cuprates , by taking only the al - vc into account . here
, we show the @xmath46-dependence of the al - vc as well as that of the maki - thompson ( mt ) vc : the mt - vc is the first - order term with respect to the fluctuations .
figure [ fig : figs3 ] shows the al - vc and mt - vc for @xmath8-orbital , @xmath284 and @xmath285 respectively , in the three - orbital @xmath0-@xmath1 hubbard model studied in the main text . here
, we put @xmath39 ev , and @xmath73 ( @xmath117 ) . | we present a microscopic derivation of the nematic charge - density - wave ( cdw ) formation in cuprate superconductors based on the three - orbital @xmath0-@xmath1 hubbard model , by introducing the vertex correction ( vc ) into the charge susceptibility . the cdw instability at @xmath2 , @xmath3 appears when the spin fluctuations are strong , due to the strong charge - spin interference represented by the vc . here
, @xmath4 is the wavenumber between the neighboring hot spots .
the obtained spin - fluctuation - driven cdw is expressed as the `` intra - unit - cell orbital order '' accompanied by the charge transfer between the neighboring atomic orbitals , which is actually observed by the stm measurements .
we predict that the cuprate cdw and the nematic orbital order in fe - based superconductors are closely related spin - fluctuation - driven phenomena .
the rich phase diagram of cuprate high-@xmath5 superconductors has been actively studied in condensed matter physics .
the non - fermi - liquid - like electronic states near the optimally - doped region , including the @xmath0-wave transition temperature at @xmath6 k , are well understood in terms of the nearly antiferromagnetic fermi liquid picture @xcite , whereas strong - coupling theories were developed to describe the under - doped region @xcite . in the pseudo - gap state of
slightly under - doped cuprates , superconducting fluctuations play important roles @xcite .
however , many mysteries concerning in the pseudogap region remain unsolved , such as the fermi arc formation @xcite and the small fermi pockets detected by quantum oscillations @xcite . the recent discovery of the axial charge - density - wave ( cdw ) parallel to the nearest cu - cu direction in y- , bi- , hg- , and la - based cuprates by the stm studies @xcite and by x - ray scattering studies @xcite constituted a significant advancement in understanding the pseudogap phenomena .
this finding indicates that both spin and charge fluctuations cooperatively develop in under - doped cuprates , and the interference between charge and spin order parameters has been discussed intensively based on various effective and microscopic models @xcite .
the aim of this paper is to present a quantitative microscopic explanation for the experimentally observed axial cdw , since the diagonal cdw is derived in previous theoretical studies @xcite .
importantly , the cdw wavevector changes with doping , coinciding with the nesting vector between the neighboring hot - spots ( see fig . [
fig : fig1 ] ( b ) ) in y- , bi- and hg - based cuprates @xcite .
in addition , all @xmath7 , @xmath8 , and @xmath9 orbital electrons contributes to the cdw formation @xcite , consistently with the local lattice deformation reported in ref .
@xcite .
the latter fact indicates that _ @xmath0-@xmath1 multiorbital model _ should be analyzed to reveal the origin of the cdw .
the mean - field - level approximations , such as the random - phase - approximation ( rpa ) , are insufficient to explain these experimental facts .
thus , we study the role of the vertex correction ( vc ) in multiorbital models that describes the strong charge - spin interference @xcite .
other than cuprates , nematic states are realized in multiorbital systems with strong correlations . in fe - pnictides ,
spin - nematic mechanism @xcite and orbital - nematic one @xcite have been proposed . in both scenarios ,
spin - fluctuation - driven nematicity is discussed .
the latter scenario proposes the orbital - order due to spin - fluctuation - driven vc , and this scenario is applicable even when the spin fluctuations are incommensurate @xcite , like in ba(fe@xmath10co@xmath11)@xmath12as@xmath12 with @xmath13 ( @xmath14 k ) . in cuprates
, the vc will develop for both the @xmath0 and @xmath1 orbitals , since both orbitals largely contribute to the density - of - states ( dos ) at the fermi level .
thus , the multiorbital cdw formation in cuprates could be explained by applying the orbital - spin mode - coupling theories @xcite . in this paper , we find the significant role of the aslamazov - larkin vc ( al - vc ) , which had not been analyzed in previous studies , in the formation of the axial cdw in cuprates . by analyzing the @xmath0-@xmath1 hubbard model with realistic parameters
, we reveal that the axial cdw instability at the wavevectors @xmath2 and @xmath3 , connected by the neighboring hot - spots , is realized by the al - vc in the charge susceptibility .
the cdw emerges only in under - doped region since the al - vc increases in proportion to the spin susceptibility .
the obtained cdw with inter - orbital charge transfer is consistent with the stm measurements @xcite .
figure [ fig : fig1 ] ( a ) shows the three - orbital @xmath0-@xmath1 model for cuprates in real space .
the nearest @xmath0-@xmath1 , @xmath0-@xmath0 , and @xmath1-@xmath1 hopping integrals are shown as @xmath15 , @xmath16 , and @xmath17 , respectively .
we use the hopping integrals of the first - principles model for la@xmath12cuo@xmath18 listed in table 2 ( @xmath19 ) of ref .
@xcite , in which the 2nd - nearest ( @xmath20 , @xmath21 , @xmath22 ) and the 3rd - nearest ( @xmath23 ) hopping integrals exist . in addition , we include the 3rd - nearest @xmath0-@xmath0 hopping @xmath24 ev to make the fermi surface ( fs ) closer to y- and bi - based cuprates . the obtained hole - like fs for the electron filling @xmath25 ( hole filling is @xmath26 ) is shown in fig .
[ fig : fig1 ] ( b ) .
we also introduce the on - site coulomb interactions ( @xmath27 , @xmath28 ) and the nearest @xmath0-@xmath1 coulomb interaction ( @xmath29 ) shown in fig .
[ fig : fig1 ] ( a ) .
the interaction parameters used in the present study is @xmath30 in ev @xcite : the ratios @xmath31 and @xmath32 are consistent with the first principle study @xcite .
later , we will show that the spin ( charge ) susceptibility is mainly enlarged by @xmath27 ( @xmath27 and @xmath29 ) sensitively , whereas both susceptibilities are insensitive to @xmath28 @xcite .
-@xmath1 model for the cuo@xmath12 plane .
( b ) fs for @xmath26 .
the integrand in eq .
( [ eqn : lambda ] ) is large when three points @xmath33 , @xmath34 , @xmath35 ( @xmath36 ) are connected by the nesting vectors .
( c ) @xmath37 given by the rpa .
the unit is ev@xmath38 .
we put @xmath39 ev and @xmath40 .
( d ) diagrammatic expression of @xmath41 .
( e )
al - vcs @xmath42 and @xmath43 .
( f ) @xmath44 and @xmath45 as functions of @xmath46 . ]
first , we study the spin and charge susceptibilities by using the rpa .
we denote @xmath47 . in the rpa without the vc , the spin ( charge ) susceptibility in the @xmath48 matrix form is given as @xmath49 , where @xmath50 is the coulomb interaction for the spin ( charge ) sectors : @xmath51 , @xmath52 , @xmath53 , and @xmath54 .
@xmath55 is the bare bubble , and @xmath56 . here and hereafter ,
@xmath57 and @xmath58 , where @xmath59 and @xmath60 .
figure [ fig : fig1 ] ( c ) shows the @xmath0-orbital spin susceptibility @xmath61 for @xmath39 ev and @xmath40 in the case of @xmath62 and @xmath63 ev .
( @xmath64 is independent of @xmath29 . )
the spin stoner factor @xmath65 , defined as the maximum eigenvalue of @xmath66 , is 0.99 .
however , the rpa fails to give any cdw instability by using the present interaction parameters . to improve the rpa ,
we calculate the charge susceptibility by including the vc , given as @xmath67 where @xmath68 , and @xmath69 is the irreducible vc for the charge sector .
when the vc is large , @xmath70 is enlarged in multiorbital models @xcite .
here , we consider the al - vc , which is the second - order term with respect to the fluctuations , since it is scaled by the square of the spin correlation length @xmath71 in two - dimensional systems @xcite .
the al - vc gives the nematic orbital order in fe - pnictides @xcite .
the al - vc increases rapidly with @xmath27 in proportion to @xmath72 @xcite , whereas it is insensitive to @xmath28 and @xmath29 . with this in mind , for simplicity , we present the expression of the al - vc for @xmath73 : @xmath74 where @xmath75 , @xmath76 , and @xmath77 : the relation @xmath78 holds in the present model . its diagrammatic expression is shown in fig .
[ fig : fig1 ] ( d ) .
the dominant contribution of the al - vc has been verified by the functional rg method @xcite . in the sc - vc method @xcite
, we calculate both @xmath79 and @xmath80 self - consistently . in the present model
, however , we verified that the positive feedback effect from @xmath81 to @xmath82 , which is important in fe - pnictides @xcite , is very small .
thus , we can safely replace @xmath79 in eq .
( [ eqn : al ] ) with @xmath83 .
we verified that the maki - thompson ( mt ) vc is considerably smaller than the al - vc ; see refs @xcite and supplemental material @xcite . in cuprates , both @xmath0 orbital and @xmath1 orbital al - vcs
are strongly enhanced when @xmath84 is large , since the @xmath1-orbital dos is large at the fermi level @xcite .
figure [ fig : fig1 ] ( e ) shows the obtained @xmath85 and @xmath86 for the parameters used in fig .
[ fig : fig1 ] ( c ) .
@xmath87 shows the maximum at @xmath88 ( @xmath89 ) , and it is about one - third of @xmath90 in magnitude .
the al - vc for @xmath8-orbital is approximately given as @xmath91 where @xmath92 .
the @xmath46-dependences of these functions along the @xmath93-axis are shown in fig .
[ fig : fig1 ] ( f ) . here
, @xmath44 has a maximum at @xmath94 , and its width is about @xmath95 : a weak shoulder structure of @xmath44 at @xmath96 reflects the incommensurate peaks of @xmath97 at @xmath98 . on the other hand , @xmath45 in fig .
[ fig : fig1 ] ( f ) takes the maximum value at @xmath99 , reflecting the nesting between the hot spots .
in fact , the integrand of eq .
( [ eqn : lambda ] ) is large in magnitude when @xmath34 , @xmath33 , @xmath100 are on the fs and connected by the nesting vector .
( @xmath101 orbital has large weight around y ( x ) point . )
thus , the large peak of @xmath102 at @xmath103 originates from the strong @xmath46-dependence of the three - point vertex @xmath104 in the present parameters .
note that @xmath45 at @xmath105 is much smaller than that at @xmath103 as shown in fig .
[ fig : fig1 ] ( f ) .
thus , the axial cdw is selected by the strong @xmath46-dependence of @xmath45 , contrary to many previous theoretical studies that predicted the diagonal cdw @xcite .
thanks to large al - vc , the charge susceptibility in eq .
( [ eqn : chic ] ) are enhanced at @xmath103 , and it diverges when the charge stoner factor @xmath106 , defined as the maximum eigenvalue of @xmath107 , reaches unity .
the cdw ( @xmath108 ) is realized due to the finite off - diagonal elements of @xmath109 ; @xmath110 .
we will show later that the cdw emerges when @xmath111 , and @xmath112 scales as @xmath113 .
thus , the larger @xmath29 is , the smaller @xmath65 for realizing the cdw is . in figs .
[ fig : fig2 ] ( a)-(c ) , we show the largest three susceptibilities , @xmath114 , @xmath115 , and @xmath116 at @xmath62 , in the case of @xmath39 ev and @xmath40 ( @xmath117 ) .
we also put @xmath118 ev , at which @xmath106 reaches @xmath119 .
both @xmath120 and @xmath121 show large positive values at @xmath122 , whereas @xmath123 develops negatively .
the charge density modulation ( @xmath124 ) at @xmath103 is proportional to the form factor that is given by the eigenvector of @xmath125 for the largest eigenvalue .
the form factor for figs .
[ fig : fig2 ] ( a)-(c ) is given as @xmath126 , which means that ( @xmath0 , @xmath8 ) orbitals form the `` antiphase cdw state '' .
( note that the form factor is sensitive to the model parameters . ) a possible charge distribution patterns for @xmath127 is depicted in fig .
[ fig : fig2 ] ( d ) .
we verified that the antiphase cdw with respect to the nearest ( @xmath128 , @xmath129 ) develops if we introduce small repulsion @xmath130 ; see supplemental material @xcite . here
, we calculate the @xmath0- and @xmath1-orbital local doss in the nematic cdw shown in fig .
[ fig : fig2 ] ( d ) , under the cdw order parameter at @xmath131 predicted by the present theory @xmath132 .
figure [ fig : fig2 ] ( e ) shows the obtained local dos @xmath133 at two @xmath8-sites , and the total dos for @xmath134 .
the pseudo - gap appears due to the cdw hybridization gap .
( here , we put @xmath135 since @xmath136 is achieved at @xmath137 in the present single - layer model ; see fig .
[ fig : fig3 ] ( a ) .
this will be justified in double - layer ybco and bscco since the fs of the bonding - band is large . ) in fig .
[ fig : fig2 ] ( f ) , we show the obtained ratio @xmath138 at cu and o sites for @xmath139 ev .
the realized intra - unit - cell nematic order looks similar to the recent stm results @xcite .
moreover , the fermi arc structure found by arpes @xcite would be formed by the single - q or double - q cdw order @xcite .
the fermi arc structure similar to cuprates was recently reported in sr@xmath12iro@xmath18 @xcite . , ( b ) @xmath140 and @xmath141 , ( c ) @xmath142 and @xmath143 .
we put @xmath39 ev , @xmath40 , and @xmath118 ev .
( d ) a possible charge pattern of the cdw ( @xmath136 ) . since the charge transfer between the neighboring @xmath144 and @xmath145 occurs , ( @xmath144 , @xmath145 )
are in antiphase in the intra - unit - cell .
( e ) total dos and local doss at two @xmath8-sites in the cdw state with @xmath127 .
( f ) @xmath146 in the cdw state for @xmath139 ev . ] here , we present an analytic explanation why the nematic cdw is realized by the al - vc in the presence of small @xmath29 .
to simplify the discussion , we consider only @xmath147 and @xmath148 and put @xmath40 in eq .
( [ eqn : chic ] ) .
the obtained results at @xmath149 are @xmath150 and @xmath151 , where @xmath152 .
thus , the charge susceptibilities develop divergently when @xmath153 is greater than @xmath154 due to the al - vc .
note that @xmath155 according to the first principle study @xcite . in the rpa without the vc
, @xmath125 diverges when @xmath29 is larger than @xmath156 ev , which is much larger than the first principle value @xcite .
worse still , the divergence occurs at @xmath157 in this model .
thus , the vc is indispensable to realizing the stripe cdw state .
the rpa analysis on a @xmath0-@xmath1 model with @xmath130 was done in ref .
@xcite in detail . ,
@xmath158 and @xmath159 obtained in the @xmath0-@xmath1 model as functions of @xmath160 .
( inset ) the values of @xmath29 required to give the cdw .
( b ) the values of @xmath161 as a function of @xmath162 , which will depict a qualitative behavior of @xmath163 .
note that @xmath164 shifts to @xmath165 for @xmath166 .
( inset ) @xmath167 ( i)-(iii ) are shown .
( c ) schematic phase diagram of cuprates . ]
now , we study the hole carrier ( @xmath168 ) dependence of the cdw state .
figure [ fig : fig3 ] ( a ) shows @xmath4 , and obtained @xmath158 and @xmath159 for @xmath40 , by choosing @xmath27 and @xmath29 so as to satisfy @xmath169 . the inset of ( a ) shows the used @xmath29 , which is much smaller than the first principle value for both @xmath40 and @xmath170 ev . also , the used @xmath27 is @xmath171 ev . here
, @xmath158 decreases for @xmath172 as observed by neutron measurements .
in contrast , @xmath159 increases as @xmath172 with satisfying the relation @xmath173 , which is widely observed in y- , bi- and hg - based compounds @xcite . also , the relation @xmath174 is consistently with experiments . here
, we explain why the cdw appears only in slightly under - doped region . in fig .
[ fig : fig3 ] ( b ) , we show the inverse of @xmath29 at the cdw boundary , @xmath175 , for @xmath40 , by adjusting @xmath27 to satisfy @xmath176 . in the case of ( i )
@xmath177 , @xmath161 decreases as @xmath172 , since the al - vc at @xmath2 , which is proportional to @xmath178 in fig .
[ fig : fig1 ] ( f ) , becomes small when @xmath179 .
however , @xmath65 decreases with @xmath162 in cuprates , which is reproduced by the flex approximation using a fixed @xmath27 @xcite .
thus , the cdw should disappear in over - doped region since the al - vc is scaled by @xmath180 @xcite .
for this reason , we also set @xmath181 as ( ii ) @xmath182 and ( iii ) @xmath183 . in case ( iii ) , if we fix @xmath184 ( dotted line ) , the cdw is realized only for @xmath185 .
thus , the phase diagram in fig .
[ fig : fig3 ] ( c ) is well understood . in la - based compounds ,
the relation @xmath186 is satisfied @xcite , differently from other compounds . to understand this fact in the present theory , we study the case @xmath187 , in which @xmath188 at @xmath189 reaches @xmath190 ev@xmath38 , which is still smaller than the neutron experimental data in 60k ybco @xcite . in this case , the incommensurate peak in @xmath97 becomes sharper as observed in la - based compounds .
then , the shoulder peak in @xmath44 at @xmath191 becomes prominent as shown in fig .
[ fig : fig4 ] ( a ) . for this reason ,
the cdw wavevector @xmath159 is fixed at @xmath95 as shown in fig .
[ fig : fig4 ] ( b ) . in this case , @xmath192 ev .
thus , the relation @xmath186 can be realized when @xmath97 shows clear incommensurate peak structure .
therefore , the present cdw mechanism due to al - vc can explain both the relations @xmath193 and @xmath194 , and the latter is realized when @xmath97 shows clear incommensurate peaks .
this result would be a great hint to understand the cdw in lsco .
note that the relation @xmath195 is naturally understood since the al - vc represents the interference of two magnons . for @xmath26 and ( b ) @xmath158 and @xmath159 in the case of @xmath187 ( @xmath196 ev ) and @xmath197ev .
] in our theory , the cdw originates from the repulsive interactions , and the @xmath198-ph interaction is unnecessary . in real compounds , the coulomb - interaction - driven cdw fluctuations couple to the lattice due to finite @xmath198-ph interactions
, so the kohn anomaly will emerge .
@xcite .
finally , we discuss the close relation between the cdw in cuprates and the nematic orbital order in fe - pnictides . in both systems
, the charge - spin interference , which is given in the al - vc , causes the inter - orbital charge transfer when @xmath199 @xcite . in fe - pnictides , both @xmath157 and @xmath200 @xmath0-orbital orders
/ fluctuations have been discussed intensively @xcite , and both fluctuations will contribute to the superconductivity . in summary
, we revealed that the axial nematic cdw in under - doped cuprates originates from the al - vcs in @xmath201 , which describes the interference of two - magnons .
it is shown that both the spin fluctuations at @xmath202 and charge - orbital fluctuations at @xmath203 develop mutually .
( this vc - driven cdw can not emerge in the single - orbital hubbard model , as we discuss in supplemental material @xcite . )
we predict that charge - orbital - spin multimode fluctuations emerge ubiquitously in cuprates , fe - pnictides , and other strongly correlated electron systems , due to the significant contribution of the al - vc .
we are grateful to s. onari , m. tsuchiizu , s. borisenko , y. matsuda and t. hanaguri for fruitful discussions . this study has been supported by grants - in - aid for scientific research from mext of japan
. 99 t. moriya and k. ueda , adv .
phys . * 49 * , 555 ( 2000 ) .
k. yamada : _ electron correlation in metals _
( cambridge univ .
press 2004 ) .
d. j. scalapino , phys . rep . * 250 * , 329 ( 1995 ) .
h. kontani , rep .
prog .
phys . * 71 * , 026501 ( 2008 ) ; h. kontani , _ transport phenomena in strongly correlated fermi liquids _ ( springer - verlag berlin and heidelberg gmbh & co. k , 2013 ) .
p. a. lee , n. nagaosa , and x .-
g .
wen , rev .
mod .
phys . * 78 * , 17 ( 2006 ) ; m. ogata and h. fukuyama , rep .
. phys .
* 71 * , 036501 ( 2008 ) .
v. j. emery and s. a. kivelson , nature * 374 * , 434 ( 1995 ) .
q. chen , i. kosztin , b. jank , and k. levin , phys .
rev .
lett . * 81 * , 4708 ( 1998 ) .
m. sato , h. harashina , j. takeda , s. yoshii , y. kobayashi , and k. kakurai , j. phys .
chem .
solids * 62 * , 7 ( 2001 ) .
t. yoshida _
et al_. , phys .
rev .
lett . * 91 * , 027001 ( 2003 ) ; t. yoshida _
et al_. , j. phys .
. jpn . * 81 * , 011006 ( 2012 ) .
a. kanigel
_ et al_. , nature physics * 2 * , 447 ( 2006 ) .
a. a. kordyuk , s. v. borisenko , v. b. zabolotnyy , r. schuster , d. s. inosov , d. v. evtushinsky , a. i. plyushchay , r. follath , a. varykhalov , l. patthey , and h. berger , phys .
rev .
b * 79 * , 020504(r ) ( 2009 ) .
t. kondo et al . ,
nature physics * 7 * , 21 ( 2011 ) .
n. doiron - leyraud , c. proust , d. leboeuf , j. levallois , j .- b .
bonnemaison , r. liang , d. a. bonn , w. n. hardy , and l. taillefer , nature * 447 * , 565 ( 2007 ) .
t. hanaguri , c. lupien , y. kohsaka , d .- h .
lee , m. azuma , m. takano , h. takagi , and j. c. davis , nature * 430 * , 1001 ( 2004 ) .
y. kohsaka , t. hanaguri , m. azuma , m. takano , j. c. davis , and h. takagi , nature physics * 8 * , 534 ( 2012 ) . m. j. lawler _
et al_. , nature * 466 * , 347 ( 2010 ) .
k. fujita _ et al_. , proc .
natl .
acad .
sci .
usa , * 110 * , e3026 ( 2014 ) .
g. ghiringhelli _
et al_. , science * 337 * , 821 ( 2012 ) .
j. chang _
et al_. , nature physics * 8 * , 871 ( 2012 ) .
e. blackburn _
et al_. , phys .
rev .
lett .
* 110 * , 137004 ( 2013 ) .
r. comin _
et al_. , science * 343 * , 390 ( 2014 ) .
e.h . da
silva neto _
et al_. , science * 343 * , 393 ( 2014 ) .
w. tabis _ et al_. , nat .
commun . * 5 * , 5875 ( 2014 ) .
m. hcker
_ et al_. , phys .
rev .
b * 83 * , 104506 ( 2011 ) .
r. comin _ et al_. , arxiv:1402.5415 .
j. c. s. davis and d .- h .
lee , proc .
natl .
acad .
sci .
usa , * 110 * , 17623 ( 2013 ) .
e. berg , e. fradkin , s. a. kivelson , and j. m. tranquada , new j. phys .
* 11 * , 115004 ( 2009 ) .
y. wang and a.v .
chubukov , phys .
rev .
b * 90 * , 035149 ( 2014 ) .
m.a .
metlitski and s. sachdev , new j. phys .
* 12 * , 105007 ( 2010 ) ; s. sachdev and r. la placa , phys .
rev .
lett . * 111 * , 027202 ( 2013 ) . c. husemann and w. metzner , phys .
rev .
b * 86 * , 085113 ( 2012 ) ; t. holder and w. metzner , phys . rev .
b * 85 * , 165130 ( 2012 ) .
s. bulut , w.a .
atkinson and a.p .
kampf , phys .
rev .
b * 88 * , 155132 ( 2013 ) .
a. bianconi , n. l. saini , a. lanzara , m. missori , and t. rossetti , h. oyanagi , h. yamaguchi , k. oka , and t. ito , phys .
rev .
lett . * 76 * , 3412 ( 1996 ) .
s. onari and h. kontani , phys .
rev .
lett .
* 109 * , 137001 ( 2012 ) .
y. ohno , m. tsuchiizu , s. onari , and h. kontani , j. phys .
soc .
jpn * 82 * , 013707 ( 2013 ) .
m. tsuchiizu , y. ohno , s. onari , and h. kontani , phys .
rev .
lett . * 111 * , 057003 ( 2013 ) .
m. tsuchiizu , y. yamakawa , y. ohno , s. onari , and h. kontani , arxiv:1405.2028 .
s. onari , y. yamakawa , and h. kontani , phys .
rev .
lett . * 112 * , 187001 ( 2014 ) .
r.m .
fernandes _
et al . _ ,
phys .
rev .
lett .
* 105 * , 157003 ( 2010 ) .
f. krger , s. kumar , j. zaanen , j. van den brink , phys .
rev .
b * 79 * , 054504 ( 2009 ) ; w. lv , j. wu , and p. phillips , phys . rev .
b * 80 * , 224506 ( 2009 ) ; c .- c .
lee , w .- g .
yin , and w. ku , phys .
rev .
lett . * 103 * , 267001 ( 2009 ) .
h. kontani and y. yamakawa , phys .
rev .
lett . * 113 * , 047001 ( 2014 ) .
p. hansmann , n. parragh , a. toschi , g. sangiovanni , and k. held , new j. phys .
* 16 * , 033009 ( 2014 ) . in the first principle study @xcite , interaction parameters are @xmath204 in ev . in the rpa
, however , we have to put @xmath205 ev to avoid the sdw order , since the self - energy correction , which describe the thermal and quantum fluctuations that destroy the magnetic order , is absent in the rpa .
in fact , @xmath97 similar to fig .
[ fig : fig1 ] ( c ) is obtained for @xmath206 ev by using the fluctuation - exchange ( flex ) approximation thanks to the self - energy correction .
the result is not sensitive for @xmath207 ev since the relation @xmath208 is assured in the flex in 2d ( mermin - wagner theorem ) ; h. kontani and m. ohno , phys .
rev .
b * 74 * , 014406 ( 2006 )
. m. s. hybertsen , m. schlter , and n. e. christensen , phys .
rev .
b * 39 * , 9028 ( 1989 ) .
k. morita _
et al_. , j. phys .
soc .
jpn * 72 * , 3164 ( 2003 ) . in the lda study ,
the @xmath1-orbital dos , @xmath209 , is about 30% of the total dos , @xmath210 .
although @xmath211 in the present @xmath0-@xmath1 model at @xmath62 , we can suppress the ratio @xmath212 by increasing @xmath213 . even in this case , the cdw due to the vc is realized by increasing @xmath29 just by @xmath214% . supplemental material .
n. harrison and s. e. sebastian , phys .
rev .
lett . * 106 * , 226402 ( 2011 ) .
h. yao , d .- h .
lee , and s. kivelson , phys .
rev .
b * 84 * , 012507 ( 2011 ) . y.k .
kim _
et al_. , science * 345 * , 187 ( 2014 ) . c. stock ,
w. j. l. buyers , r. liang , d. peets , z. tun , d. bonn , w. n. hardy , and r. j. birgeneau , phys .
rev .
b * 69 * , 014502 ( 2014 ) . s. y. savrasov and o. k. andersen , phys
. rev .
lett . * 77 * , 4430 ( 1996 ) .
d. reznik , advances in condensed matter physics , article i d 523549 ( 2010 ) .
s. johnston _
et al_. , phys .
rev .
b * 82 * , 064513 ( 2010 ) .
s. kasahara _
et al_. , nature * 486 * , 382 ( 2012 ) . |
we consider drawings of the complete graph @xmath0 in the plane in which vertices are drawn as points in the plane and edges as simple planar curves that contain no vertices other than their endpoints . as usual
, we require that all intersections are proper crossings ( no tangencies ) and that two edges share only a finite number of points . the number @xmath11 of crossings in a drawing @xmath2 is the sum of the number of intersection points of all unordered pairs of interiors of edges .
the _ crossing number _
@xmath12 is the minimum @xmath11 over all drawings @xmath2 of @xmath13 .
a drawing is _ crossing optimal ( or minimal ) _ if it realizes @xmath14 .
a long - standing conjecture is that the crossing number @xmath15 of the complete graph @xmath0 is equal to @xmath16 a very fine history of this and related problems is given by beineke and wilson @xcite .
they attribute the conjecture to anthony hill .
as it is first published by harary and hill in @xcite , we propose the notation @xmath5 used above to denote the conjectured value of @xmath15 and attribute the conjecture to harary - hill . according to @xcite , hill proposed a construction with vertices on two concentric cycles and conjectured @xmath5 to be the number of crossings for those drawings .
blaek and koman @xcite proved the following variation of hill s construction .
half of the vertices are drawn evenly spaced ( as a cycle ) at the rim of the top lid of a cylinder ( tin can ) and the remaining vertices are drawn evenly spaced at the rim of the bottom lid .
the edges are drawn either as straight - lines within the cylinder lids or ( in the case they connect two vertices from different lids ) as shortest geodesic lines on the cylinder .
this model is identical to the model with two concentric circles and it gave rise to a name for a whole class of drawings , see below .
so far , the conjecture has been verified for @xmath17 only .
pan and richter @xcite have shown that @xmath18 is 100 using a computer proof ; well - known counting arguments show @xmath19 .
mcquillan and richter @xcite present a computer - free proof that @xmath20 .
recently , an important line of research has been started by brego et al .
@xcite , who restricted the allowed drawing styles and proved that for these drawings the conjecture is true . in @xcite
, they consider 2-page book drawings of @xmath0 in which all vertices are supposed to lie on a line @xmath21 and each edge needs to be drawn on one of the two half - planes defined by @xmath21 . in @xcite , the authors extended their techniques to _ monotone _ drawings of @xmath0 , in which all the vertices have different @xmath22-coordinates and the edges are @xmath22-monotone curves . in @xcite , brego et al .
generalize their result to cylindrical drawings and @xmath22-bounded drawings .
cylindrical drawing _ has two concentric circles on which all the vertices must be placed and all edges must not intersect these circles . an _
@xmath22-bounded drawing _ requires that all edges are contained within a strip bounded by the vertical lines defined by their endpoints .
every monotone drawing is @xmath22-bounded and every 2-page book drawing is a cylindrical drawing , so their result indeed generalizes previous results .
their result , which also prompted this work , was based on the first general combinatorial condition on a drawing @xmath2 of @xmath0 that guarantees that @xmath2 has at least @xmath5 crossings . for this , they introduced the notion of _ shellability _ of a drawing of @xmath0 . in a related work , balko et al . @xcite give a combinatorial characterization of several classes of @xmath22-monotone drawings of complete graphs and show that also the odd crossing number ( a different variant of counting crossings ) of @xmath22-monotone drawings as well as shellable drawings of @xmath0 is at least @xmath5 .
the purpose of this work is to define a more general version of shellability that we call bishellability , and which is implied by shellability .
the main benefit of our approach is the simplification of the principal concept .
this allows a significantly simpler and more intuitive proof for the fact that bishellable drawings satisfy the harary - hill conjecture .
moreover , bishellability reflects better the required properties .
we are convinced that this is a further step to gain more insight into the structure of crossing minimal drawings , with the ultimate goal to prove the harary - hill conjecture .
( for the following definition , we recall that , for a drawing @xmath2 of @xmath0 , a _ face _ of @xmath2 is a component of @xmath23 $ ] .
this is the same notion as for embeddings : if we convert each crossing point of @xmath2 into a vertex , then the faces of @xmath2 are the faces of the planarly embedded graph . )
finally , if @xmath24 is a subset of vertices in the drawing , @xmath25 denotes the drawing obtained when vertices of @xmath24 and all edges incident to them are deleted from @xmath2 . for a positive integer @xmath3 , a planar drawing @xmath2 of @xmath0
is if there is a sequence @xmath26 of distinct vertices of @xmath2 so that , relative to a reference face @xmath27 , for all integers @xmath28 with @xmath29 , the vertices @xmath30 and @xmath31 are both incident with the face of @xmath32 containing @xmath27 .
their main theorem in @xcite is the following .
[ th : old ] let @xmath2 be a drawing of @xmath0 . if there is an integer @xmath33 such that @xmath2 is @xmath3-shellable
, then @xmath34 .
one of the disadvantages of the notion of shellability is that @xmath3-shellable does not imply @xmath35-shellable .
this is because the sequence @xmath26 of vertices needs to be circular to get from the reference face @xmath27 and back to @xmath27 again , and a long circular sequence does not imply a shorter circular sequence .
we introduce a more general variant of shellability that we call _
bishellability_. for a non - negative integer @xmath3 , a drawing @xmath2 of @xmath0 is _
@xmath3-bishellable _ if there exist sequences @xmath36 and @xmath37 , each sequence consisting of distinct vertices of @xmath0 , so that , with respect to a reference face @xmath27 : 1 . for each @xmath38 ,
the vertex @xmath39 is incident with the face of @xmath40 that contains @xmath27 ; 2 . for each @xmath38 ,
the vertex @xmath41 is incident with the face of @xmath42 that contains @xmath27 ; and 3 . for each @xmath43 , the set @xmath44 .
we remark that if @xmath36 and @xmath45 show that @xmath2 is @xmath3-bishellable , then the same sequences without @xmath46 and @xmath47 show that @xmath2 is @xmath35-bishellable .
moreover , the vertices @xmath48 and @xmath49 must lie on the boundary of the common face @xmath27 also , if @xmath2 is @xmath3-shellable , with witnessing sequence @xmath26 , then @xmath2 is @xmath50-bishellable with witnessing sequences @xmath51 and @xmath52 defined by @xmath53 and @xmath54 . here is the version of theorem [ th : old ] that holds for bishellable drawings ; its proof is in the next section .
[ th : main ] if @xmath2 is an @xmath55-bishellable drawing of @xmath0 , then @xmath56 .
there are two main remarks to be made here .
first is the pleasant feature that the theorem requires only one value for the amount of bishellability of the drawing .
second , even though the same principal ideas are used in the two proofs , the proof of theorem [ th : main ] involves a simpler induction than the proof of theorem [ th : old ] in @xcite .
this is due to the monotonicity of the new definition of bishellability . besides the proof of theorem [ th : main ]
, we show that bishellability leads to an extended class of drawings for which the harary - hill conjecture is true .
figure [ fig : n11bi ] shows an example of a drawing @xmath2 of @xmath8 that is @xmath57-bishellable ( so theorem [ th : main ] shows @xmath58 ) , but , for every @xmath59 , is not @xmath3-shellable ( so theorem [ th : old ] does not apply ) . to simplify the discussion
, we define a drawing @xmath2 of @xmath0 to be _
shellable _ if it is @xmath3-shellable for some @xmath60 , and _
bishellable _ if it is @xmath61-bishellable .
that is , shellable and bishellable drawings have at least @xmath5 crossings .
furthermore , we call a drawing of @xmath0 _ harary - hill optimal _ if it has @xmath5 crossings .
we use this notation to keep in mind that drawings with @xmath5 crossings are only conjectured to be optimal .
in this section , we recall the notion of good drawings and @xmath62-edges , and how to express the crossing number of a drawing @xmath2 in terms of a weighted sum of the numbers of @xmath62-edges , for @xmath63 .
then we prove that , under the assumption that a drawing @xmath2 is bishellable , for each relevant @xmath62 , the relevant weighted sum of @xmath62-edges is large enough to prove that there are at least @xmath5 crossings in @xmath2 .
we remark that we consider only _ good _ drawings ( also known as simple topological drawings ) : no edge can cross itself ; no two edges with a common incident vertex can cross ; and no two edges can cross more than once .
it is well - known that any drawing of a graph @xmath13 with fewest crossings is necessarily good .
the relation between the number of crossings in a rectilinear ( or pseudo - linear ) drawing of @xmath64 and the number of its @xmath65-edges was first described by lovsz et al .
@xcite and , independently , by brego and fernndez @xcite .
brego et al .
@xcite generalized the notion of @xmath62-edges to arbitrary good drawings , as follows .
fix a drawing @xmath66 of @xmath64 and a face @xmath67 of @xmath66 . for an edge @xmath68 of @xmath2
, we arbitrarily choose one of its orientations : here we use @xmath69 to @xmath70 .
for each other vertex @xmath71 , the three vertices @xmath72 induce a 3-cycle @xmath73 in @xmath2 . because @xmath2 is a good drawing
, @xmath73 is a simple closed curve in the sphere ( it is slightly simpler technically to consider embeddings in the sphere rather than the plane ) .
as we traverse @xmath73 from @xmath69 to @xmath70 to @xmath71 and back to @xmath69 , @xmath73 has natural right and left sides relative to the directed edge @xmath68 .
assign to @xmath71 the side @xmath74 ( right ) if @xmath27 is on the left side of @xmath73 ; otherwise , assign @xmath75 ( left ) to @xmath71 .
thus , for each @xmath76 , @xmath71 is assigned either @xmath74 or @xmath75 .
then @xmath68 is a _ @xmath62-edge _ if @xmath62 is the smaller of the number of @xmath74 s and @xmath75 s ( for @xmath68 ) .
note that being a @xmath65-edge is independent of the orientation of the edge @xmath77 as reversing @xmath77 simply exchanges all the labels @xmath78 and @xmath79 .
we make one small observation that helps the later discussion : every edge @xmath80 that has a segment incident with @xmath27 is a 0-edge . to see this ,
suppose some segment @xmath81 of @xmath80 is incident with @xmath27 and we orient @xmath80 so that @xmath27 is to the left of @xmath81 .
then , for every @xmath71 not incident with @xmath80 , the 3-cycle determined by @xmath80 and @xmath71 will have @xmath27 on the left side , showing @xmath71 is an @xmath74 .
that is , all vertices are @xmath74 with respect to @xmath80 , so @xmath80 is a 0-edge , as claimed
. one main ingredient of the proof is the relation between the number of @xmath62-edges and the number of crossings in a drawing .
this relation was shown for rectilinear drawings in @xcite and extended to good drawings in @xcite , where it is shown that the number of crossings of a good drawing can be expressed as a weighted sum of the number of @xmath62-edges of the drawing .
we include this discussion in the appendix , for the sake of completeness . specifically , if we denote by @xmath82 the number of @xmath62-edges of @xmath2 and consider @xmath83 and @xmath84 in @xcite it is shown that @xmath85 therefore , any lower bound on @xmath86 translates immediately into a lower bound for @xmath11 and , specificaly , showing that @xmath87 implies @xmath88 .
[ lm : inequality ] if a drawing @xmath2 of @xmath0 is @xmath62-bishellable and @xmath89 , then @xmath90 since @xmath2 is @xmath91-bishellable , for each @xmath92 , the drawing @xmath2 is @xmath62-bishellable .
therefore , lemma [ lm : inequality ] implies that @xmath93 as required . plugging these lower bounds into equation
[ eq : twopage ] the desired lower bound of @xmath5 on the number of crossings in @xmath2 follows ( for details see the proof of theorem 3 in @xcite ) .
we essentially follow the ideas of the proofs in @xcite , but use a simpler and more direct approach provided by the new concept of bishellability .
we proceed by induction on @xmath62 .
the base case of @xmath94 is trivial , as the face @xmath27 is incident with at least three edges and each of these is a 0-edge .
thus , @xmath95 as required . for the induction step , let @xmath96 and @xmath97 be sequences witnessing @xmath62-bishellability . consider the drawing @xmath98
. then @xmath99 , @xmath100 show it is @xmath101-bishellable and , since @xmath102 , the induction implies that @xmath103 rewritten , this is @xmath104 consider an edge @xmath80 in @xmath98 .
if @xmath80 is an @xmath105-edge with @xmath106 , then it is either an @xmath105-edge or an @xmath107-edge of @xmath2 , depending on whether @xmath48 joins the majority or minority part of the @xmath74 s and @xmath75 s with respect to @xmath80 in @xmath98 .
we call those that are @xmath105-edges in both @xmath98 and @xmath2 _ invariant_. note that the coefficient of a non - invariant @xmath105-edge in the sum for @xmath98 is @xmath108 , while its coefficient in the @xmath2-sum is @xmath109 .
thus , its contribution to each sum is the same . on the other hand
, an invariant @xmath105-edge contributes @xmath108 to the @xmath110-sum and @xmath111 to the @xmath2-sum .
there is additional contribution to the @xmath2-sum from the edges incident with @xmath48 .
we will discuss those shortly .
but these remarks allow us to conclude that @xmath112 we shall prove that there are at least @xmath113 invariant edges and the contribution of the edges incident with @xmath48 is at least @xmath114 . together with the induction assumption applied to @xmath115
, we conclude that @xmath116 as required .
we start with the edges incident with @xmath48 .
let @xmath117 and @xmath118 be the two edges incident with @xmath48 so that the corner of the reference face @xmath27 at @xmath48 contains both @xmath117 and @xmath118 .
consequently , the fact that @xmath119 shows that we may write the cyclic rotation of the edges incident with @xmath48 as @xmath120 . to make the discussion uniform , orient all edges incident with @xmath48 away from @xmath48 .
consideration of any 3-cycle @xmath121 shows that , if , relative to the edge @xmath122 , @xmath70 is @xmath75 , then , relative to @xmath123 , @xmath69 is @xmath74 .
we arbitrarily choose the orientation of the sphere so that all the vertices not incident with @xmath117 are @xmath74 s for @xmath117 . because deleting @xmath124 puts @xmath125 into the boundary of the ( extended ) reference face
, we see that @xmath125 has at most @xmath105 @xmath75 s and all the rest are @xmath74 s . from this
it is immediate that @xmath125 is an @xmath126-edge , as required .
in particular , letting @xmath127 be that integer such that @xmath125 is a @xmath127-edge , we know @xmath125 contributes @xmath128 to the sum @xmath129 . since @xmath130 , @xmath131 , so @xmath125 contributes at least @xmath111 to the sum .
therefore , @xmath132 contribute at least @xmath133 to the sum @xmath129 .
likewise , @xmath134 contribute at least the same amount , as desired for the contribution from the edges incident with @xmath48 .
the invariant edges are determined by the @xmath41 s .
the main point is that @xmath49 is incident with at least @xmath135 invariant edges upon deletion of @xmath48 . to see this
, we observe that if @xmath136 and @xmath137 are the edges incident with @xmath49 at the corner of @xmath27 incident with @xmath49 then the symmetry @xmath138 in the definition of bishellable drawings implies that the cyclic rotation of edges at @xmath49 can be taken as @xmath139 .
we may choose the labelling of the @xmath140 and @xmath141 so that @xmath142 is not one of @xmath143 .
we know that , for @xmath144 , @xmath140 has only the non-@xmath49 ends of @xmath145 as possible @xmath75 s ( say ) , with the remaining vertices being @xmath74 s .
in particular , @xmath48 is an @xmath74 for all of @xmath143 .
it follows that , for some @xmath130 , @xmath140 is a @xmath127-edge in both @xmath2 and @xmath98 .
that is , @xmath49 is incident with at least @xmath135 invariant edges .
here is the other significant simplification due to bishellability .
the identical argument applies to the vertex @xmath41 in the @xmath146-bishellable drawing @xmath147 , as witnessed by the sequences @xmath148 and @xmath149 .
the vertex @xmath41 ( in the role corresponding to @xmath49 in the preceding paragraph ) is incident with at least @xmath150 edges invariant ( relative to @xmath151 and @xmath152 ) .
the values ( @xmath74 or @xmath75 ) of @xmath153 are , for each of the edges incident with @xmath41 , independent of whether we consider @xmath151 or @xmath152 . thus , each edge of @xmath151 incident with @xmath41 is a @xmath127-edge of @xmath2 if and only if it is a @xmath127-edge in @xmath98 .
it follows that @xmath2 has , in total , at least @xmath154 invariant edges incident with @xmath97 , as required .
in this section we show that the concept of bishellability is really more general than the concept of shellability , in the sense that there exist bishellable drawings that are not shellable .
we start by considering conditions for drawings to be shellable or bishellable , respectively .
obviously a simple cycle in a drawing @xmath2 made up with at least @xmath155 non - crossed edges guarantees that @xmath2 is shellable and thus also bishellable .
also , a path of @xmath156 non - crossed edges with distinct vertices @xmath157 @xmath158 implies that @xmath2 bishellable .
of course these are sufficient but not necessary conditions .
figure [ fig : n06 ] shows on the left a bishellable drawing of @xmath159 where @xmath48 and @xmath49 are incident to the same cell @xmath27 , but the edge @xmath142 is crossed .
however , this drawing is not harary - hill optimal , as in a harary - hill optimal drawing for two vertices which are incident to the same cell the connecting edge can always be drawn within this cell without crossings .
two drawings of @xmath159 : left : bishellable drawing with a crossed edge @xmath160 ; right : non - bishellable drawing.,title="fig : " ] in @xcite it has already been argued that all cylindrical , @xmath22-bounded , monotone , and 2-page book drawings of @xmath0 are shellable and therefore have at least @xmath5 crossings . two generic classes of harary - hill optimal drawings
have been known , i.e. , for any @xmath161 there exist cylindrical and 2-page book drawings with precisely @xmath5 crossings .
only recently new families of harary - hill optimal drawings have been discovered . in @xcite
two new families have been constructed for odd @xmath162 .
both are based on the known harary - hill optimalcylindrical drawings . for one family
there is just an additional vertex added in the center of the inner cycle , plus incident edges .
the edges of the outer cycle remain uncrossed , and thus these drawings are also shellable and bishellable .
the second family is more interesting , as in addition several edges are redrawn in a way so that , for @xmath162 , every edge is crossed by at least one other edge .
it is to some extent counterintuitive that such drawings can be harary - hill optimal ; see @xcite for more details and figure [ fig : n11non ] for such a drawing of @xmath8 .
note that in these drawings not only are there no non - crossed edges , but also no two vertices share a common cell .
thus , these drawings can neither be shellable nor bishellable .
so far , this constitutes the only known family of drawings of @xmath0 which is harary - hill optimal but not bishellable .
the smallest complete graph where a non - bishellable drawing exists is @xmath159 , depicted in figure [ fig : n06 ] , right .
however , this example is not harary - hill optimal . the smallest harary - hill optimal drawing which is not bishellable is the one of @xmath163 shown in figure [ fig : n09 ] .
it is worthwhile to mention that out of the 421 harary - hill optimal drawings of @xmath163 this is the only non - bishellable drawing , up to weak isomorphism ( two drawings of @xmath0 are weakly isomorphic if they have the same ( or reversed ) rotation system ) .
the smallest example of a harary - hill optimal drawing that is not shellable but bishellable is the drawing of @xmath8 shown in figure [ fig : n11bi ] .
the two sequences @xmath164 and @xmath165 proving bishellability are indicated in the drawing .
the example is symmetric and almost a 2-page book drawing , or a monotone drawing .
there are just two edges of the spine that are crossed and two edges that are non - monotone .
recent progress on the rectilinear crossing number of @xmath0 has depended on making more refined estimates than that provided by lemma [ lm : inequality ] .
these refinements can not occur in the context of the ( topological ) crossing number considered in this paper , as the bound in lemma [ lm : inequality ] can not be improved for a drawing having @xmath5 crossings .
also , it is now known that almost half of the more than 400,000 ( up to weakly isomorphism ) harary - hill optimal drawings of @xmath8 are not shellable , and this almost surely applies in roughly equal numbers to bishellable .
it seems likely that , as @xmath161 grows , the proportion of bishellable harary - hill optimal drawings vanishes .
however , theorem [ th : main ] ( indeed theorem [ th : old ] is enough ) implies , for example , that any drawing with @xmath166 drawn in the southern hemisphere ( using an uncrossed @xmath167-cycle on the equator ) and the remaining @xmath166 drawn in the northern hemisphere ( arbitrarily ) must have at least @xmath5 crossings .
these kinds of drawings play a role in the computer - free proof @xcite that @xmath20 .
+ we close with some open problems : * can we construct a family of harary - hill optimal drawings of @xmath0 which are non - shellable , but bishellable , similar to the constructions in @xcite ?
the drawing of @xmath8 shown in figure [ fig : n11bi ] might be a good start .
* is there a concept similar to bishellability that does not require the starting vertices of the sequences to share a cell , but still implies at least @xmath5 crossings ? this would mean that each of the two sequences could have their own ( local ) reference faces .
[ [ acknowledgements ] ] acknowledgements research for this article was initiated during the _ aim workshop on exact crossing numbers _ held from april 28 to may 2 , 2014 in palo alto ( california , usa ) .
we are grateful to the american institute of mathematics ( aim ) for their support and we thank the participants of that workshop for fruitful and inspiring discussions . furthermore , o.a . and b.v .
were partially supported by the esf eurocores programme eurogiga crp compose , austrian science fund ( fwf ) : i648-n18 ; s.f .- m . was supported by the nsf grant dms-1400653 ; p.m. was partially supported by the esf eurocores programme eurogiga crp gradr , 10-eurogiga - op-003 ( dfg ) ; p.r .
was partially supported by the esf eurocores programme eurogiga crp compose , micinn project eui - eurc-2011 - 4306 ; and b.m . and r.b.r .
were supported by nserc .
10 bernardo m. brego , oswin aichholzer , silvia fernndez - merchant , pedro ramos , and gelasio salazar .
the 2-page crossing number of k@xmath168 .
, 49(4):747777 , 2013 .
bernardo m. brego , oswin aichholzer , silvia fernndez - merchant , pedro ramos , and gelasio salazar .
more on the crossing number of k@xmath168 : monotone drawings . , 44:411414 , 2013 .
bernardo m. brego , oswin aichholzer , silvia fernndez - merchant , pedro ramos , and gelasio salazar .
shellable drawings and the cylindrical crossing number of k@xmath168 . , 52(4):743753 , 2014 .
bernardo m. brego , oswin aichholzer , silvia fernndez - merchant , pedro ramos , and birgit vogtenhuber .
non - shellable drawings of k@xmath168 with few crossings . in _
cccg 2014 , halifax , nova scotia , canada _ , 2014 .
bernardo m. brego and silvia fernndez - merchant . a lower bound for the rectilinear crossing number . , 21(3):293300 , 2005 .
martin balko , radoslav fulek , and jan kynl .
crossing numbers and combinatorial characterization of monotone drawings of @xmath0 .
, pages 137 , 2014 .
lowell beineke and robin wilson . the early history of the brick factory problem .
, 32(2):4148 , 2010 .
j. blaek and m. koman .
a minimal problem concerning complete plane graphs .
in _ theory of graphs and its applications , proceedings of the symposium held in smolenice in june 1963 _ , volume 8 , pages 113117 , 1964 . frank harary and anthony hill . on the number of crossings in a complete graph . in _ proc .
edinburgh math .
_ , volume 13 , pages 333338 , 1963 .
l. lovsz , k. vesztergombi , u. wagner , and e. welzl .
convex quadrilaterals and k - sets .
towards a theory of geometric graphs . in j.
pach , editor , _ ams contemp .
volume 342 , pages 139148 , 2004 .
dan mcquillan and r. bruce richter . on the crossing number of _ _
k__@xmath169 without computer assistance . submitted .
shengjun pan and r. bruce richter .
the crossing number of _
_ k__@xmath170 is 100 .
, 56(2):128134 , 2007 .
if we consider good drawings of @xmath171 , it is well - known that there are only two such drawings up to spherical homeomorphisms .
however , the one with a crossing has , again up to spherical homeomorphisms , two different faces : one is incident with a 4-cycle and the other is incident with two vertices and the crossing .
it is easy to verify that : 1 .
if @xmath27 is bounded by the 4-cycle , then the four edges of the 4-cycle are all 0-edges while the crossing edges are both 1-edges ; 2 .
if @xmath27 is incident with just two vertices and the crossing , then the two crossing edges are 0-edges , the full edge incident with @xmath27 is a 0-edge , as is its opposite edge in the uncrossed 4-cycle , and the other two edges in the uncrossed 4-cycle are 1-edges ; and 3 . in the case of the planar drawing of @xmath171 , the three edges incident with @xmath27
are the 0-edges , and the other three edges are the 1-edges . the preceding paragraph can be used to relate crossings in a drawing @xmath2 of @xmath0 and @xmath62-edges .
each crossing determines a @xmath171 containing precisely two 1-edges ( relative to the face containing the reference face @xmath27 of @xmath2 ) .
each non - crossing @xmath171 contains three 1-edges .
it follows that if we count the number of ordered pairs @xmath172 so that @xmath80 is a 1-edge of this @xmath171 , the total number we get is @xmath173 , where @xmath174 is the number of planar @xmath171 s and @xmath175 is the number of non - planar @xmath171 s . on the other hand ,
if @xmath68 is a @xmath62-edge , and @xmath176 are distinct vertices both different from both @xmath69 and @xmath70 , then the @xmath171 induced by @xmath177 has @xmath68 as a 1-edge if and only if one of @xmath71 and @xmath178 is an @xmath74 ( relative to @xmath68 ) and the other is an @xmath75 .
if follows that every @xmath65-edge @xmath68 is a 1-edge in @xmath179 different @xmath171 s .
that is , the number of pairs @xmath172 where @xmath80 is a 1-edge of the @xmath171 is @xmath180 where , for @xmath181 , @xmath82 denotes the number of @xmath62-edges in @xmath2 .
the conclusion ( exactly as in @xcite ) is that @xmath182 on the other hand , @xmath183 is the total number of @xmath171 s in @xmath0 , so @xmath184 multiply equation ( [ eq : numberk4 s ] ) by @xmath57 , subtract equation ( [ eq:1-edge ] ) , and use the obvious fact that @xmath185 to conclude that @xmath186 equivalent to the proof of proposition 1 in @xcite this equation can be rewritten to @xmath187 if we consider @xmath83 and @xmath84 we can get the original formulation in @xcite : @xmath188 | the harary - hill conjecture , still open after more than 50 years , asserts that the crossing number of the complete graph @xmath0 is @xmath1 brego et al .
@xcite introduced the notion of shellability of a drawing @xmath2 of @xmath0 .
they proved that if @xmath2 is @xmath3-shellable for some @xmath4 , then @xmath2 has at least @xmath5 crossings .
this is the first combinatorial condition on a drawing that guarantees at least @xmath5 crossings . in this work ,
we generalize the concept of @xmath3-shellability to bishellability , where the former implies the latter in the sense that every @xmath3-shellable drawing is , for any @xmath6 , also .
our main result is that @xmath7-bishellability also guarantees , with a simpler proof than for , that a drawing has at least @xmath5 crossings .
we exhibit a drawing of @xmath8 that has @xmath9 crossings , is 3-bishellable , and is not @xmath3-shellable for any @xmath10 .
this shows that we have properly extended the class of drawings for which the harary - hill conjecture is proved . |
the cosmic microwave background ( cmb ) provides one of the most useful tools for understanding the physical processes in the early universe , giving valuable information about the origin and the dynamics of the primordial fluctuations that seeded the today large - scale structures .
there are two competing families of cosmological theories quite distinct in this regards .
+ in the models invoking most types of inflation @xcite the perturbations generation mechanism is linear and gaussian fluctuations are predicted . for a given type of inhomogeneity ( adiabatic or isocurvature )
the only degree of freedom in the initial conditions is the amplitude , and since the evolution is linear , solutions scale linearly with the initial amplitude .
linearity combined with the assumed statistical homogeneity of the universe guarantees that one may decompose the cosmological perturbations into eigenmodes with eigenfunctions that evolve independently . in the fourier space , for instance
, the evolution of a perturbation in each mode @xmath3 is independent . for a finite sample of the cmb field ,
the observed fluctuations are the convolution of the true " fluctuations of the infinite field with the fourier transform of a mask @xcite .
the 2-point function of the the power fluctuations depends only on the selection function @xcite .
+ the second class of theories , most topological defect models @xcite and some versions of inflation @xcite , predict non - gaussian statistics of the primordial perturbations . those models involve spontaneous symmetry breaking as the hot early universe cooled leading to a non - linear perturbations generation mechanism . for this type of perturbations the fourier modes @xmath3 are coupled and their evolution can not be anticipated on the basis of the initial conditions @xcite .
the predicted angular power spectrum has a position dependence with a variance larger than the variance of the angular power spectrum for the gaussian case at the same multipole order @xcite .
the defect theories are highly constrained by causality , the fluctuations being completely uncorrelated beyond the horizon scale at all times @xcite .
+ it is difficult to make difference between gaussian and non - gaussian fluctuations of the density field at large angular scales because of the tendency of any distribution to approach gaussian statistics when averaging over large area ( the central limit theorem ) , and also because of the cosmic and sampling variances .
+ in this paper we propose the wavelet transform to analyse the large scale coherence of the cmb anisotropy field with non - gaussian initial conditions .
the main advantage of the wavelet transform over the fourier transform is its capability to investigate features of the cmb anisotropy maps with a resolution according to their scales @xcite .
the wavelet base functions are highly localized in space and the information regarding the scale and the position are stored explicitly .
also , the wavelet analysis can be performed with a finite sample of data in contrast with the fourier analysis where the result depends on the selection function .
+ we employ 2-point function of power fluctuations in the wavelet space to compare cobe-_dmr _ 4-year temperature anisotropy maps to the predictions of an analytical cosmic string model @xcite .
the advantage of the analytical approach is that one can obtain statistical fluctuations with non - gaussian random phases for a given experimental configuration , while the explicit dependences on the normalization of the primordial power spectrum , string parameters and angular scales can be shown .
+ we use the non - gaussian likelihood function to constrain the dimensionless cosmic string parameter @xmath4 ( where @xmath5 is the newton s constant , @xmath6 is the speed of light and @xmath0 the the mass per unit length of the cosmic strings ) at @xmath2 cl in standard gaussian statistics , by comparing zero lag autocorrelation function from cobe-_dmr _ with the predictions of the cosmic string model .
according to the scaling solution for cosmic strings @xcite there is a fixed number @xmath7 of strings with the curvature radius of the order of the horizon per hubble volume at any given time , with orientations and velocities uncorrelated over distances larger than the horizon .
the temperature anisotropy induced by long strings is given by kaiser - stebbins formula @xcite : @xmath8 where : @xmath9 is the direction of observation , @xmath10 is the velocity of the string with the orientation @xmath11 , @xmath12 , @xmath5 is the newton s constant and @xmath0 is the mass per unit length of the string
. + the time from the last scattering surface until now is divided into @xmath13 hubble time steps @xmath14 with @xmath15 and @xmath16 .
the apparent angular size of a hubble volume at time @xmath14 is @xmath17 and @xmath18 for large redshifts in the matter dominated era assuming @xmath19 .
+ for an experiment measuring a square map of size @xmath20 at each hubble time @xmath14 the number of @xmath21 string segments random placed over the area size @xmath22 is given by @xcite : @xmath23 we use an analytical statistical string model @xcite to take into account the combined effects of the temperature fluctuations induced by strings present between the redshift @xmath24 and today ( @xmath25 ) .
+ we calculate the probability distribution functions of the temperature fluctuations induced by cosmic strings for different values of the string scaling solution parameter @xmath7 by fourier transforming the characteristic function of this model normalized to approach a gaussian distribution with zero mean and @xmath26 when @xmath27 @xcite .
+ we generated cmb independent realizations , each of 6144 pixels , using the spherical harmonic representation : @xmath28 where @xmath29 is the dmr gaussian window function with fwhm=@xmath30 .
+ the harmonic coefficients @xmath31 are random variables with zero mean and the variances given by @xcite : @xmath32\gamma[(9-n)/2 ] } { \gamma[l+(5-n)/2]\gamma[(3+n)/2]},\ ] ] where : @xmath33 is the quadrupole normalized amplitude and @xmath34 is the spectral index of the primordial power spectrum .
+ the coefficients @xmath31 are drawn from parent populations obtained in the cosmic string model the non - gaussian amplitude distributions of @xmath31 with random phases convoluted with the spherical harmonics @xmath35 are characterized by a positive kurtosis in the resulting temperature distribution .
+ we generated 800 @xmath36 full - sky realizations for each model defined by @xmath33 and @xmath7 .
to each realization we added a realization of the noise determined by the instrument sensitivity and the number of observations per pixel of dmr 4-year 53ghz ( a+b)/2 map , and rejected the pixels with galactic latitude @xmath37 obtaining @xmath38 pixels .
+ we generated also in the same conditions 800 @xmath36 full sky realizations for the same values of @xmath33 with @xmath31 drawn from a standard gaussian distribution with zero mean and @xmath26 .
a wavelet @xcite @xmath39 is a function whose binary dilatation and dyadic translations generate an orthonormal base so that any function @xmath40 can be approximated up to an arbitrarily small precision by : @xmath41 the base functions @xmath42 are labeled by the scale @xmath43 and the position @xmath44 @xmath45 .
the wavelet coefficients @xmath46 are real quantities containing all the information as that contained by the fourier coefficients .
the major difference is the number of indices , that is , in wavelet analysis the scale @xmath43 and the position @xmath44 in the scale are explicitly stored .
the wavelet base functions @xmath47 satisfy the scaling equation : @xmath48 let @xmath49 be the hubble volume spanned by the scaling function @xmath50_{j \in n}$ ] at the scale @xmath43 . from the scaling equation it follows that : @xmath51 this hierarchical structure ensures the multi - resolution capability of the wavelet analysis .
+ the orthogonality condition : @xmath52 yields to : @xmath53 this last equation represents our wavelet transform convention .
+ for a random gaussian field the ensemble - averaged power @xmath54 in the wavelet space at the separation angle @xmath55 is : @xmath56 where @xmath57 and @xmath58 are two directions in the sky separated by @xmath55 .
+ for a random non - gaussian field generated using the cosmic string model the intrinsic correlations exist only within the same scale and the equation ( 8) becomes : @xmath59 and @xmath60 contains the correlated ensemble - averaged signal .
+ we define the 2-point function of the power fluctuations in the wavelet space as : @xmath61 in this paper we present the results obtained using daubechies wavelet base functions of order 4 as a typical compact orthonormal base .
+ the wavelet transform applied on 2d map ( with @xmath38 pixels and @xmath62 uncorrelated scales ) acts in a pyramidal algorithm , stripping the initial map scale by scale into its correlated components .
the reader may refer to the references @xcite for complete information .
+ figure 1 presents the distributions of spots as a function of threshold ( expressed in number of standard deviations ) obtained for the averaged simulated maps in real and wavelet space for two types of simulations : @xmath31 with random gaussian phases ( panels a and b ) , and @xmath31 with random non - gaussian phases , obtained in the cosmic string model with @xmath63 ( panels c and d ) . for all simulations we take @xmath64k . in each case
the gaussian fits are also shown .
one can see that the differences between the distributions obtained in gaussian and non - gaussian case are more evident in the wavelet space than in the real space . + the maps obtained by monte - carlo simulations were wavelet transformed and then 2-point functions of the power fluctuations were constructed ( binned in @xmath65 bins , in the manner of cobe-_dmr _ ) . + figure 2 presents the 2-point function of the power fluctuations in the wavelet space @xmath66 normalized at @xmath67 .
the solid curve represents the theoretical prediction if the underlying distribution is gaussian and @xmath64k .
the error bars are based on monte - carlo simulations in the cosmic string model with @xmath63 and the same normalization .
according to the standard @xmath68 method , if a specific model defined by the parameters @xmath69 is more likely to have occurred higher is the probability to obtain values of the @xmath68 larger than the measured @xmath70 . the cumulative probability used to define the confidence regions on the parameters
is given by the integral : @xmath71 where @xmath72 is the likelihood function full specified by the @xmath68 value and the covariance matrix .
we define the @xmath68 as : @xmath73 where @xmath74 is the ensemble - averaged value of the 2-point function of the power fluctuations for monte - carlo realizations and @xmath75 is the two point function of the power fluctuations for _ dmr _ data .
+ the covariance matrix @xmath76 calculated for the the monte - carlo realizations is : @xmath77 the standard gaussian likelihood statistic @xmath78 associated to the two point function of the power fluctuation @xmath79 is given by : @xmath80 in the case of non - gaussian random fields the confidence regions of the parameters depend also on the higher order correlation functions .
it is shown that for mild non - gaussianity the likelihood function can be obtained within the multivariate edgeworth expansion ( mee ) @xcite . according to mee the total likelihood function
is : @xmath81 where @xmath82 is the non - gaussian correction which embodies the higher order moments of the data .
+ the integral in equation ( 11 ) can be written as : @xmath83 where @xmath84 is the integral of gaussian likelihood , that is the cumulative integral of the @xmath68 probability distribution with @xmath65 degrees of freedom , and @xmath85 is a non - gaussian correction that depends on the higher order moments . + for each set of monte - carlo simulations we estimated the cumulative probability @xmath86 .
+ figure 3 presents @xmath86 obtained for the models with @xmath64k and few values of the @xmath7 parameter compared with the same probability obtained for the gaussian case ( the upper curve ) .
+ one can see that for mild non - gaussian fields the confidence regions are larger than the confidence regions in the gaussian case at the same @xmath68 threshold .
+ using the gaussian likelihood given by equation ( 14 ) we found @xmath87 when the dmr 53ghz ( a+b)/2 map was compared with the monte - carlo simulations obtained for a gaussian underlying distribution .
+ figure 4 presents the confidence regions obtained for the parameters @xmath33 and @xmath7 in cosmic string models at @xmath88 that corresponds to @xmath2 cl for a purely gaussian statistics . + considering the region with @xmath89 and taking the extrema of this contour we found that dmr data favourize the cosmic string models with @xmath7 within 10 and 12 and @xmath33 within @xmath90 and @xmath91 at @xmath2 cl for the standard gaussian statistics .
we estimated the dimensionless cosmic parameter @xmath92 by comparing , in the wavelet space , the averaged zero lag autocorrelation function obtained in those models with zero lag autocorrelation function of dmr 4-year @xmath93 @xmath94 map .
we found : @xmath95 there are two terms that contribute to the error : a symmetrical error ( @xmath96 ) originating from the width of the confidence region and an asymmetric error given by the cosmic and sampling variances .
in this paper we propose the wavelet analysis for studying the large - scale cmb correlations induced by the cosmic strings .
+ using an analytical cosmic string model to generate cmb temperature fluctuations with non - gaussian random phases in the cobe-_dmr _ experimental configuration , and 2-point function of the power fluctuations in the wavelet space , we constrain the normalization of the primordial power spectrum @xmath33 and the parameter of the cosmic string scaling solution @xmath7 . + we found that the large - scale cmb anisotropy favourizes the cosmic string models with large number of strings per hubble volume , @xmath7=10 , 11 and 12 and @xmath33 from @xmath97k to @xmath98k at @xmath2 cl for the standard gaussian statistics . + by comparing zero lag autocorrelation function of cobe-_dmr _ 53ghz ( a+b)/2 map with our predictions we obtained the dimensionless cosmic string parameter @xmath99 , value that is compatible with the values obtained from the simulations of the evolution of the string network from @xmath100 to present @xcite , as well as with other existing studies of the expected large - scale cmb anisotropy @xcite , @xcite .
guth , phys .
d 23 * ( 1981 ) 347 a. linde , phys .
lett * b 108 * ( 1982 ) 389 a. liddle and d. lyth , phys . rep .
* 231 * ( 1993 ) 1 a. vilenkin , phys
* 121 * ( 1985 ) 263 .
a. vilenkin and e.p.s .
shellard _ cosmic strings and other topological defects _ ( cambridge university press , 1994 ) d. coulson , p. ferreira , p. graham , and n. turok nature * 368 * ( 1994 ) 27 l. kofman , a. linde and a.a starobinsky phys
lett . * 76 * ( 1996 ) 1011 a. albrecht , d. coulson , p. ferreira and j. magueijo phys .
lett * 76 * ( 1996 ) 1413 j. magueijo , a. albrecht , d. coulson and p. ferreira phys .
* 76 * ( 1996 ) 2617 p. ferreira and j. magueijo phys .
d 55 * ( 1997 ) 3358 r. crittenden and n. turok phys .
* 75 * ( 1995 ) 2642 m. kendall , a. stuart and j.k .
kendall s advanced theory of statistics _
( oxford university press , new york , 1987 ) h.a .
feldman , n. kaiser , j.a .
peacock ap .
j. * 426 * 23 .
a.j . stirling and j.a .
peacock mnras * 283l * ( 1996 ) 99 c.k .
chui , _ an introduction to wavelets _
( academic press , 1992 ) w.h .
press , b.p .
flannery , s.a .
teukolsky , w.t .
vetterling _ numerical recipes _ ( cambridge university press , 1992 ) d. bennet and f. bouchet , phys .
60 * ( 1988 ) 257 b. allen and e.p.s .
shellard phys .
* 64 * ( 1990 ) 119 n. kaiser and a. stebbins nature * 310 * ( 1984 ) 391 r. moessner , l. privolaropoulos , and r. branderberger , ap .
j. * 425 * ( 1994 ) 365 l. perivolaropoulos , phys lett .
* b298 * ( 1993 ) 305 j.r .
bond and g. efstathiou , mnras * 226 * ( 1987 ) 655 y. fujiwara and j. soda , prog .
* 95 * ( 1996 ) 1059 b. allen , r.r .
caldwell , e.p.s .
shellard , a. stebbins and s. veeraraghavan phys .
rev . lett .
* 77 * ( 1996 ) 3061 d. bennet , a. stebbins , and f. bouchet , ap.j.lett * 399 * ( 1992 ) l5 t. hara , p. mahonen , and s.miyoshi , ap.j * 414 * ( 1993 ) 412 | we analyse the large - scale coherence of the cmb anisotropy field with non - gaussian initial conditions using 2-point function of the power fluctuations in the wavelet space .
+ employing the multivariate edgeworth expansion ( mee ) we constrain the normalization for the cosmic string mass - per - unit - length @xmath0 , obtaining @xmath1 at @xmath2 cl in the standard gaussian statistics .
this value is consistent with the results obtained from the simulations of the evolution of the string network and by other large scale studies .
cosmology : cosmic microwave background , large scale structure - methodic : non - gaussian statistics , wavelet analysis |
zwicky in 1933 ( @xcite ) found that unseen mass had to be assumed to exist to explain the observational data for the coma cluster .
subsequently , it was discovered that this unseen mass called dark matter had to be incorporated into gravity to explain the rotation curves of galaxies ( @xcite ) .
underground experiments ( @xcite ) as well as astrophysical searches have failed to detect dark matter particles .
an alternative explanation for the missing dark matter is to modify newtonian and einstein gravity .
one such fully covariant and lorentz invariant alternative theory is scalar - tensor - vector - gravity ( stvg ) , also know as mog ( @xcite ) . mog has passed successful tests in explaining rotation velocity data of spiral and dwarf galaxies ( @xcite ) , ( @xcite ) , globular clusters ( @xcite ) and clusters of galaxies ( @xcite ) .
recently , it was claimed ( @xcite ) that no modified gravity theory can fit the abell 1689 acceleration data without including dark matter or heavy ( sterile ) neutrinos .
the cluster a1689 is important , for good lensing and gas data are available and we have data from 3kpc to 3mpc .
we will show that mond ( @xcite ) does not fit the a1689 acceleration data , nor does the dark matter model based on an nfw mass profile .
however , mog does fit the a1689 acceleration data without dark matter .
we will use the weak field approximation of the theory at astrophysical scales to describe the dynamics of galaxies and clusters of galaxies .
the point particle acceleration is given by ( @xcite ) : @xmath1 . \label{pointacceleration}\end{aligned}\ ] ] here , @xmath2 denotes newton s gravitational constant and @xmath3 denotes the mass .
the parameter @xmath4 determines the strength of attactive gravitation through @xmath5 , while @xmath6 is the mass of the vector field @xmath7 that produces a repulsive gravitational force through the gravitational charge : @xmath8 where @xmath9 is the current density matter source of the vector field @xmath7 . the mog acceleration ( [ mogpointacceleration ] ) satisfies the ( weak ) equivalence principle . for a distribution of matter ,
the acceleration is given by @xmath10 .
\end{split}\ ] ] the application of the theory to galaxies as well as clusters of galaxies determines the best fit values : @xmath11 and @xmath12 @xcite . the potential for a distribution of matter with density @xmath13 is given by : @xmath14
in order to see if mog fits the cluster a1689 data , we employ the normalized acceleration data given by nieuwenhuizen in his fig . 2 ( @xcite ) .
nieuwenhuizen has obtained the acceleration data by using a1689 strong and weak lensing data . in particular , the two - dimensional mass density @xmath15 has been determined by limousin et al .
( @xcite ) . from x - ray observations
the electron and gas mass density can be obtained .
we use 59 data points with 12 data points from ( @xcite ) at radii between 3 kpc and 271 kpc , 20 data points from ( @xcite ) , which overlap with the data of limousin et al . in the window between 40 kpc and 150 kpc , and 27 data points from ( @xcite ) , which cover the range between 125 kpc and 3 mpc .
the gas data is from chandra x - ray observations with a total exposure time of 150 ks ( @xcite ) and in our calculations we use a cored sersic electron density profile ( @xcite ) : @xmath16.\ ] ] from ( @xcite ) , we obtain @xmath17 , @xmath18 , @xmath19 and @xmath20 . for a typical z=0.3 solar metallicity
the gas density is ( @xcite ) : @xmath21 background galaxies are far away from the cluster , so strong lensing can be treated as occurring due to mass projected onto the plane through the cluster center .
the strong lensing yields data for the mass density along the line - of - sight : @xmath22
we investigate the possibility of fitting the a1689 data with mog , mond and with a dark matter model using a nfw profile .
although there are many galaxies in a galaxy cluster , the galaxy mass density of a1689 is dominated by the brightest cluster galaxy ( bcg ) in the center of the cluster .
we will use the bcg density profile proposed in ( @xcite ) : @xmath23 where @xmath24 and @xmath25 denote the mass and the radial size of the central galaxy , respectively , while @xmath26 denotes its core size . in our calculations
, we use in our bcg mass density profile @xmath27 kpc , @xmath28 kpc and @xmath29 from ( @xcite ) . because the bcg is spherically symmetric
, we can use the mog point mass radial acceleration equation ( [ mogpointacceleration ] ) .
we can express this acceleration by the effective equation for @xmath30 : @xmath31 where @xmath32 denotes the baryon mass . by using @xmath33 and integrating @xmath34 over the volume of the cluster , we obtain the mass @xmath35 .
we choose the parameter values : @xmath36 and @xmath37 or @xmath38 kpc .
this choice of @xmath4 and @xmath6 parameters coincides with the values employed for fits to dwarf galaxy rotation curves ( @xcite ) .
we have plotted the mog predicted acceleration in fig.1 .
the acceleration data is from ( @xcite ) .
the acceleration curve for newtonian gravity without dark matter , determined by @xmath39 , is depicted in fig.1 . and does not fit the normalized acceleration data . in the case of mond , we use the following interpolating function @xmath40 to calculate the acceleration : @xmath41 this leads to the mond acceleration : @xmath42^{1/2}.\ ] ] the predicted mond normalized acceleration curve shown in fig.1 , follows closely the newtonian acceleration curve without dark matter . to obtain the a1689 dark matter normalized acceleration
, we adopt the nfw profile ( @xcite ) : @xmath43 the dark matter model acceleration is obtained from @xmath44 where @xmath45 denotes the dark matter mass .
we have from lensing data for a1689 ( @xcite ) : @xmath46 and @xmath47 . after integrating the nfw profile and calculating @xmath48
, we obtain the acceleration curve depicted in fig.1 .
we have normalized the acceleration in all three cases to @xmath49 .
as can be seen , mog can fit the data points from small to large radii . for the nfw profile dark matter model
, just a part of the data can be fitted for radii @xmath50 200 kpc . to see
if the data can be fitted for smaller radii , we change the value of @xmath51 to fit the model for smaller radii , but then for larger values of the radii the model no longer fits the data .
these results are shown in fig.1 .
the mond prediction is the same as the newtonian prediction , and for mond there is no way to fit the data except to include dark matter or heavy sterile neutrinos .
the lensing and the chandra x - ray data can determine the total baryon density for the cluster a1689 . the acceleration for newtonian , mond and mog models is then derived from the baryon mass @xmath35 by integrating the total baryon density , assuming that the bcg is the dominant mass in the a1689 cluster . the normalized acceleration for these models
is then compared with acceleration data for a1689 .
the newtonian and mond predictions without dark matter fall far short of fitting the data at all values of the radial distance @xmath52 . a dark matter model based on a nfw density profile with the parameters @xmath53 and @xmath54 determined by strong lensing data fits the acceleration data for @xmath55 kpc but not for @xmath52 below this value .
the mog predicted acceleration without dark matter provides a good fit to the data at both small and large values of @xmath52 .
modified gravity theories have been formulated to avoid the need for dark matter .
experiments have failed to detect dark matter particles motivating studying the alternative gravitational theories .
the fully covariant and lorentz invariant mog theory fits galaxy dynamics data and cluster data .
it also fits the merging clusters bullet cluster and the train wreck cluster ( abell 520 ) without dark matter ( @xcite ) .
a mog application to cosmology without dark matter can explain structure growth and the cmb data ( @xcite ) .
the fitting of the cluster a1689 data adds an important success for mog as an alternative gravity theory without dark matter .
we thank martin green , niayesh afshordi and theodorus nieuwenhuizen for helpful discussions .
we thank marceau limousin for supplying data and for helpful discussions .
this research was supported in part by perimeter institute for theoretical physics .
research at perimeter institute is supported by the government of canada through the department of innovation , science and economic development canada and by the province of ontario through the ministry of research , innovation and science . | the galaxy cluster system abell 1689 has been well studied and yields good lensing and x - ray gas data .
modified gravity ( mog ) is applied to the cluster abell 1689 and the acceleration data is well fitted without assuming dark matter .
newtonian dynamics and modified newtonian dynamics ( mond ) are shown not to fit the acceleration data , while a dark matter model based on the navarro - frenk - white ( nfw ) mass profile is shown not to fit the acceleration data below @xmath0 .
[ firstpage ] |
quantum turbulence @xcite , as well as classical turbulence @xcite , is dominated by reconnection of vortical structures which is much simpler to treat in the framework of quantum fluids rather than in viscous fluids @xcite , while leading to similar features such as time asymmetry @xcite . despite the fundamental differences between the two forms of turbulence , there are reasons to believe that the understanding of quantum turbulence might shed new light on the understanding of its classical counterpart @xcite .
quantum fluids dynamics is properly described by the gross pitaevskii equation ( gpe ) @xcite @xmath0where @xmath1 is the complex wave function . through the madelung transformation @xmath2 , equation
can be viewed in classical fluid dynamical terms as @xmath3 where @xmath4 denotes density , @xmath5 velocity , @xmath6 pressure , and @xmath7 the so - called quantum stress ( @xmath8 ) .
defects in the wave function @xmath1 are interpreted as infinitesimally thin vortices of constant circulation @xmath9 , with healing length @xmath10 .
gpe conserves the mass and the energy @xmath11 the main reason for preferring the gpe approach to others for the study of quantum turbulence is that it guarantees a natural dynamics of interacting vortices @xcite while resolving fine scales up to the vortex core @xcite . on the contrary ,
methods based on the inviscid euler equations ( either their direct numerical simulation @xcite or vortex filament methods @xcite ) are unable to automatically perform vortex reconnections , being forbidden by euler dynamics .
the numerical solution of the gpe is normally carried out by employing time splitting fourier methods @xcite and by imposing vortices in the form of singular phase defects in a unitary background density , i.e. @xmath12 when @xmath13 . however , these methods assume periodic boundary conditions .
solutions which are not periodic must be mirrored in the directions lacking periodicity @xcite , thus imposing doubling of the degrees of freedom in each of those directions and a consequent increase of the computational effort .
recent developments @xcite have shown that reconnections in quantum fluids are strictly related to topological features characterizing the interacting vortex tubes such as writhe , total torsion and intrinsic twist .
these quantities depend on the fine details of the curve that describes the vortex centerline ( its third derivative with respect to the curvilinear abscissa is required for computing torsion ) and on the phase of the wavefunction @xmath1 in the neighborhood of the vortex centerline .
therefore it is paramount to resort to high resolution numerical simulations of equation , especially in the proximity of the reconnection event . with the goal of assessing the goodness of time - splitting fourier methods for singular solutions on uniform grids versus time - splitting finite differences on nonuniform grids , we first derive an analytic approximation of a two - dimensional steady state vortex that nullifies the right - hand - side of .
then we perform a systematic comparison between the two approaches by measuring the deviation of the numerical solution from the initial condition ( being steady the initial condition should remain preserved ) .
finally , we explore the possibility to evaluate the solution obtained by time - splitting fourier methods on nonuniform grids designed to guarantee higher spatial resolution in the proximity of vortex singularities .
we seek a two - dimensional , steady - state solution of equation that represents a straight vortex centered at the origin .
it is well - known that the classical two - dimensional euler vortex of circulation @xmath14 has azimuthal velocity @xmath15 where @xmath16 is the radius and @xmath17 $ ] is the azimuthal angle .
the cartesian components of the velocity are thus @xmath18 and @xmath19 .
therefore @xmath20 .
this shows that the velocity field is solenoidal ( @xmath21 ) , that the quantum mechanical phase , @xmath22 , is simply the azimuthal angle @xmath23 , and that the quantum of circulation , in our dimensionless units , is equal to @xmath24 . in steady conditions ,
the continuity equation ensures that @xmath25 , hence @xmath26 , which means that @xmath27 .
the solution @xmath28 leads to @xmath29 , which has infinite energy and must be rejected .
the other possibility is that @xmath30 .
since @xmath31 , then @xmath32 is parallel to @xmath33 and thus @xmath34 , @xmath33 and @xmath35 being the unitary vectors in two - dimensional polar coordinates . in a two - dimensional domain we set @xmath36 , where @xmath37 is a function to be determined . by imposing
that @xmath1 is the steady solution of equation , we find that @xmath38 satisfies the equation @xmath39 with boundary conditions @xmath40 , @xmath41 .
equation could be integrated numerically as it is , by artificially bounding the infinite domain . to avoid this problem , we resort to the change of variables @xmath42 , @xmath43 , which yields the equation for @xmath44 @xmath45 defined in the finite domain @xmath46 $ ] , with boundary conditions @xmath47 and @xmath48 .
we solve equation by central second order finite differences with equally spaced discretization points @xmath49 , @xmath50 . given the numerical solution @xmath51 of , the numerical approximation of the density is @xmath52 ^ 2\ ] ] where @xmath53 , @xmath54 .
this rescaling provides denser points @xmath55 in the neighborhood of the origin , where they are mostly needed ( more than 95% of the points @xmath55 are in the interval @xmath56 ) .
nevertheless , the computation of the initial solution for on a two - dimensional grid , for instance , requires @xmath57 to be interpolated
. it would be therefore useful to have an analytic approximation of @xmath58 .
however , since it is more convenient @xcite to find a pad approximation directly for @xmath59 ^ 2 $ ] rather than for @xmath58 , we rewrite equation in terms of @xmath60 as @xmath61 with boundary conditions @xmath62 , @xmath63 .
it is known @xcite that pad approximations of @xmath60 retain only even degrees at both the numerator and denominator , that is @xmath64 in order for this approximation to satisfy the boundary conditions , it must be @xmath65 given these simplifications , the diagonal pad approximation , with @xmath66 coefficients and both numerator and denominator of degree @xmath67 , is @xmath68 in literature this approximation is normally limited to @xmath69 @xcite , that is @xmath70 with @xmath71 despite its widespread usage , this approximation is _ qualitatively wrong _ ( see , e.g. , @xcite ) , because it reaches an unphysical maximum , above @xmath72 , at @xmath73 , unique positive solution of @xmath74 obtained by imposing @xmath75 . the physical solution of equation must reach the limit value @xmath63 monotonically , without overshooting . due to these limitations , we seek higher - order ( @xmath76 ) , monotonically increasing , pad expansions , namely @xmath77 in order to determine the coefficients of a certain approximation @xmath78 , we compute the analytic expressions @xmath78 , @xmath79 and @xmath80 and substitute them in equation obtaining the form @xmath81 the numerator @xmath82 is made of terms @xmath83 , which are in a number much larger than the @xmath66 degrees of freedom of the pad expansion . for this reason equation can not be satisfied exactly .
however , we can nullify the coefficients of @xmath66 terms @xmath83 .
we can choose to start from higher- or lower - order coefficients in @xmath82 .
we prefer to operate on lower - order powers of @xmath83 , i.e. @xmath84 , because we need a good approximation of @xmath78 in a neighborhood of the origin .
interestingly enough , we observe _ a posteriori _ that canceling the lower - order coefficients of @xmath83 results in very small values of the coefficients of larger powers of @xmath85 .
the step - by - step derivation of @xmath78 for @xmath86 is reported in [ sec : appendix ] , whereas tables [ t : coeff23 ] and [ t : coeff4 ] summarize all coefficients for the expansions @xmath87 .
^2 $ ] , numerical solution of obtained by second - order finite differences on 5000 equispaced points , and different pad approximations @xmath78 . ] in figure [ fig : rhopade ] we show the comparison between the numerical solution of by employing second - order central finite differences with 5000 points and different pad approximations @xmath78 for @xmath86 .
visual inspection confirms that @xmath69 is a poor representation of the solution of equation , especially for @xmath88 . and the numerical solution of with ( equally spaced grid , second - order finite differences , 5000 points ) . ] to appreciate quantitatively the error with respect to the numerical solution , in figure [ fig : errrho ] we report the relative error in a semilog plot .
interestingly , the relative error does not reach its maximum close to the origin , meaning that any pad approximation reproduces quite well the behavior of the vortex for @xmath89 . on the other hand ,
the maximum relative error is always reached for @xmath90 , i.e. in a region of interest for the numerical simulations that we will perform .
widely used schemes for the numerical simulation of the dynamics of are the so - called time - splitting methods and the finite difference time domain methods ( see @xcite for a review ) .
if we restrict the options to second - order accurate schemes in time , time splitting pseudospectral ( tssp ) methods , time splitting finite difference ( tsfd ) methods and crank nicolson finite difference ( cnfd ) method conserve the mass at the discretized level .
however , cnfd is implicit and requires the solution of a coupled nonlinear system at each time step . for this reason we resorted to time splitting methods .
we refer the reader to @xcite for higher - order time splitting methods . in ( * ?
* example 4.1 ) tssp is suggested when the solution is smooth and tsfd otherwise , although the hint comes from a one - dimensional numerical experiment . in what follows ,
we analyze two approaches : a classical time splitting fourier method and a time splitting _ nonuniform _ finite difference method . in any case ,
equation is split into two parts @xmath91 where @xmath92 .
the solution of the first equation depends on the space chosen for the discretization and will be described in the next two sections .
the second equation can be solved exactly , taking into account that @xmath93 is preserved by the equation .
therefore @xmath94 for any @xmath95 in the spatial domain .
if we denote by @xmath96 and @xmath97 the two partial numerical solutions , the approximation @xmath98 of @xmath99 , where @xmath100 , can be recovered by the so - called strang splitting @xmath101 equation can be solved exactly in time within the fourier spectral space . a part from the error at machine - precision level coming from the necessary direct and inverse fast fourier transforms ( ffts ) , the only possible considerable error might arise from an insufficient number of fourier modes .
this is usually not a big deal when approximating smooth solutions fastly decaying to zero , since spectral order of convergence takes place . for this to happen
, the unbounded domain has to be truncated to a computational bounded domain @xmath102 large enough to support the most of a periodic approximation of the solution . however , when simulating the dynamics of vortex solutions not decaying to zero , as in our case where @xmath103 , there are some issues to take into account : the low regularity of the solution at the origin , due both to @xmath104 and to @xmath105 , and lack of periodicity at the boundaries , also considering the usual extension of the computational domain and reflection of the solution ( see @xcite ) . in fact , after such a mirroring , the solution takes the same values at opposite boundaries , but its derivatives do not . in order to investigate the accuracy of fourier approximation for vortex solutions
, we consider the fourier series expansion of the function obtained by mirroring @xmath106 with respect to the axis @xmath107 and @xmath108 . the final computational domain is therefore @xmath109 , with @xmath110 , for which @xmath111 .
the choice of the parameters @xmath112 , @xmath113 and @xmath114 corresponds to a two - dimensional straight vortex as described above .
different choices provide more regular functions or functions fastly decaying to zero , for which the derivatives at the boundaries are almost periodic .
we compute a reference approximation by an expansion into a series with @xmath115 fourier modes and compare it with expansions ranging between @xmath116 and @xmath117 modes , in the @xmath118 norm . by mirroring ) for different choices of the parameters .
only the case of a regular and fast decaying to zero function ( @xmath119 , @xmath120 and @xmath121 , green , upward triangles ) shows the typical spectral rate of convergence . ] for a quite regular and periodic function , corresponding to @xmath119 , @xmath120 and @xmath121 we observe in figure [ fig : regularity ] a typical spectral rate of convergence . for any other choice of the parameters , which affects the regularity of the density ( @xmath112 , @xmath120 , @xmath121 ) , or the fast decay to zero of the function and its derivatives ( @xmath119 , @xmath113 , @xmath121 ) or the regularity of the phase ( @xmath119 , @xmath120 , @xmath114 ) we observe a strong order reduction .
the same reduction occurs for the straight vortex ( @xmath112 , @xmath113 , @xmath121 ) .
increasing the number of fourier coefficients so as to gain accuracy is often not an option .
in fact , due to the necessary mirroring , this corresponds to a huge growth of the degrees of freedom .
moreover , the use of hyperbolic sparse grids ( see @xcite , for instance ) is not possible , since the possibility of discarding coefficients and grid points is given only for highly regular solutions .
the low regularity of the solutions to be approximated and the needed duplication along axes in order to satisfy at least the periodicity of the values of the solutions suggest to explore the alternative of a finite difference discretization in space .
the main advantage of a finite difference approach is that the mirroring of the solution is not required , being the extension of the bounded domain replaced by the imposition of homogeneous neumann boundary conditions . given the low regularity of vortex solutions , we use centered second order finite differences . with the aim of increasing the spatial resolution around the vortex cores and keeping a reasonable degree of freedom , we employ a set of nonuniform grid points ( see @xcite , for instance , for locally adaptive finite element discretizations ) . the discretization of the laplace operator in one dimension with nonuniform finite differences on @xmath122 points provides the nonsymmetric matrix @xmath123 0 & \ldots & 0 & \frac{2}{h_{m-1}^2 } & -\frac{2}{h_{m-1}^2 } \end{bmatrix}\ ] ] where @xmath124 , @xmath125 , @xmath126 .
this is not exactly a second order approximation , although a discretization in which @xmath127 and a refinement with @xmath128 yields a first order term in the error decaying faster than the second order term ( see @xcite ) .
the approximation for the two - dimensional and the three - dimensional cases can be simply obtained by kronecker products with the identity matrix .
if we call @xmath129 the corresponding matrix , equation is transformed into the system of ordinary differential equations @xmath130 given the importance of the mass preservation , we investigate this issue for the numerical solution of system . a quadrature formula with positive weights for the computation of the mass writes @xmath131 it can be written as @xmath132 where @xmath133 denotes the transposed conjugate vector of @xmath134 and @xmath135 the matrix with diagonal @xmath136 .
we define @xmath137 such that @xmath138 with @xmath139 . if @xmath140 is _ symmetric _
, then the solution @xmath141 is an orthogonal matrix and @xmath142 this means that @xmath143 and therefore system preserves the mass at the discrete level if @xmath135 makes @xmath140 symmetric . from the structure of the matrix @xmath144 , it is clear that the vector of trapezoidal weights @xmath145 $ ] gives a matrix @xmath146 such that @xmath147 is symmetric .
the extension to @xmath135 in the two - dimensional and three - dimensional cases is trivial and this is enough to get @xmath140 symmetric as well , in fact @xmath148 .
we conclude that equation preserves the mass at the discrete level whenever the trapezoidal rule is used as quadrature formula and this is easily extended to any space dimension .
system could be solved , for instance , by the crank nicolson scheme @xmath149 which preserves the discrete mass being @xmath140 symmetric ( see @xcite ) .
this scheme is second order accurate in time , therefore the size of the time step @xmath150 has to be chosen such that the error is smaller than the time splitting error .
moreover , crank nicolson scheme requires the solution of a linear system of equations with matrix @xmath151 at each time step @xmath150 . although this is not a big deal in one space dimension , since the matrix is tridiagonal , in higher dimensions the discretization yields a large , sparse , complex symmetric matrix .
this implies the use of preconditioned krylov solvers for general matrices such as gmres or bicgstab or minimal residual methods for complex symmetric systems ( see @xcite ) .
iterative methods converge to the solution up to a specified tolerance which therefore influences the mass conservation and the whole accuracy of the result . given these complications ,
we prefer to consider a direct approximation of the exact solution @xmath152 nowadays there are several options for the computation of the action of the matrix exponential to a vector .
we refer to @xcite for a review of _ polynomial _ methods which do not require the solution of linear systems . in this way ,
the kinetic linear part is solved exactly in time , as in the fourier spectral method .
in section [ sec : pade ] we have derived various approximations of @xmath60 for a straight , two - dimensional vortex , whose wavefunction is @xmath153 . in order to quantitatively
compare the two methods introduced in section [ sec : splitting ] , we measure the preservation of such a steady solution by reporting the relative error @xmath154 with @xmath155 , where @xmath156 is either @xmath157 or @xmath158 , the latter evaluated at any required @xmath85 by linear interpolation of . the origin is excluded since @xmath159 is zero therein .
the time step @xmath160 is chosen such that @xmath161 is an integer , where @xmath162 is the final simulation time . in all our experiments , we selected @xmath163 , a reasonable value in quantum fluids simulations @xcite .
the maximum over the continuum set @xmath164 in the error above is approximated by the maximum over a discrete set which will be specified later .
although the preservation of the initial state may seem a trivial test , it is in fact a reliable and necessary experiment in order to validate the effectiveness of the proposed numerical methods .
thanks to the reliability of the analytic solution , this test can show the influence of both the spatial approximation and the time splitting error in the numerical discretization of the pde . in
what follows we will employ either tssp ( fourier ) or tsfd . for a computational grid with @xmath165 grid points in the physical domain of interest
, tssp requires a total of @xmath166 degrees of freedom due to mirroring , whereas tsfd requires only @xmath167 degrees of freedom thanks to homogeneous neumann boundary conditions .
+ we preliminary test the reliability of the three pad approximations @xmath87 and the numerical solution of equation obtained by central second order finite differences with @xmath168 uniformly distributed discretization points . for the solution of the gpe we employ tssp with fourier basis functions on a uniform two - dimensional computational grid .
for this reason , the numerical solution @xmath169 of equation must be interpolated .
results are reported in figure [ fig : diffic ] , where the relative error defined by is plotted versus time for different disks .
the number of fourier modes is fixed to @xmath170 , i.e. @xmath171 degrees of freedom .
we compare the solution at each time step with the initial condition on the grid nodes within the considered disk .
the worst approximation of the steady - state solution is the commonly used @xmath172 pad approximation , whereas @xmath173 turns out to be as accurate as the numerical solution .
for this reason , in the following experiments we will consider only @xmath173 .
all curves collapse on each other in the case of the largest disk , meaning that the maximum error occurs at the boundaries , mainly due to the non - periodicity of the solution .
+ we compare the performance of the finite difference approximation on a uniform versus nonuniform grid .
the uniform grid has the same step - size @xmath174 as the grid employed for the fourier approach and reported in figure [ fig : diffic ] .
the nonuniform grid is generated by taking into account different constraints . given the smallest step - size @xmath175 at the origin ( in the vortex core ) , we linearly increase the step - size according to @xmath127 in both @xmath95 and @xmath176 and in both positive and negative directions .
we choose @xmath177 so as to reach the boundaries exactly , and in order to keep the ratio @xmath178 , where @xmath179 .
the number of points of the nonuniform grid is chosen such that the mean value of @xmath180 equals the step - size of the uniform grid .
these constraints guarantee a reasonable nonuniform grid .
the comparison between the uniform and nonuniform grids is shown in figure [ fig : univsnon ] . in all cases
the initial condition is @xmath173 , thus the error for the uniform grid can be compared directly with that in figure [ fig : diffic ] for the case of pad approximation with @xmath181 ( red squares in both figures ) .
we first focus on the results with the same number of points and the same boundaries , i.e. red squares and blue circles in figure [ fig : univsnon ] .
the discrete mass variation along time integration is comparable and of order @xmath182 .
this confirms the conservation of mass also for the case of nonuniform grid , as discussed in section [ sec : tsfd ] . in the nonuniform case
the error is roughly one order of magnitude smaller than in the uniform case on small disks and for @xmath183 not too large , whereas the curve of the nonuniform case tends to jump onto the uniform one after a certain time as the radius of the disk of interest increases .
this suggests the idea that the error arises at the boundaries , where @xmath184 does not exactly fulfill neumann boundary conditions .
motivated by this , we have changed the boundaries from @xmath110 to @xmath185 and @xmath186 to check the dependency of the error on the choice of the truncated domain . in doing so ,
we have preserved the constraints on the nonuniform grids discussed above , obtaining @xmath187 degrees of freedom for @xmath186 ( upward green triangles ) and @xmath188 degrees of freedom for @xmath185 ( downward black triangles ) . with reference to figure [ fig : univsnon ] ,
the domain bounded at @xmath186 is clearly too small and the error is always very large compared to all the other cases . on the other hand , the curves for @xmath110 ( blue circles ) and @xmath185 ( downward black triangles ) behave roughly in the same way up to a certain value of @xmath183 , after which the case @xmath110 consistently show larger errors than the case @xmath185 .
this reinforces the claim that the error arises from the borders .
now we concentrate on our main goal , which is the comparison between tssp with fourier basis function on uniform grids and tsfd on a nonuniform grid that we fix to @xmath189 , @xmath110 , @xmath190 . in order to compare the error defined by for the two methods
, we always evaluate the tssp solution on the nonuniform grid points ( spectral solutions can be evaluated everywhere ) .
this set of points has the advantage of being denser in the vortex core , where higher spatial resolution is desirable .
results are reported in figure [ fig : tsfdvstssp ] , where sp stands for spectral and nfd for nonuniform finite differences . keeping in mind that the spectral fourier approach needs mirroring , i.e.
the number of modes in each direction must be doubled , we first choose a number of fourier modes @xmath170 in each direction to make it equal to the number of points of the reference case for nonuniform finite differences ( @xmath190 ) in the physical ( un - mirrored ) domain .
the overall behavior of the error for these two cases is comparable : tssp ( red squares ) performs better than tsfd ( blue circles ) for small values of @xmath183 , whereas the opposite happens for intermediate values of @xmath183 . for large @xmath183 the two curves collapse on each other . due to the fact that tssp needs mirroring , i.e. @xmath191 , in figure [ fig : tsfdvstssp ]
we explore also the cases with less fourier modes , namely @xmath192 ( upward green triangles ) and @xmath193 ( downward black triangles ) .
as observed for the case @xmath170 , in the long term all curves seem to provide similar errors , regardless of the disk radius . on the other hand , for small values of @xmath183 ,
the number of degrees of freedom plays a rle in that a larger number of fourier modes ensures smaller errors .
it is important to keep in mind that , for what seen in figure [ fig : regularity ] , the tssp fourier approach does not retain the spectral accuracy because of the singular nature of the solution at the origin and the lack of periodicity at the boundaries . as a final remark ,
we observe that the error of the fourier solution computed on its own uniform grid , reported in figure [ fig : diffic ] with red squares , is smaller than the error of the fourier solution evaluated on the nonuniform grid , reported in figure [ fig : tsfdvstssp ] with red squares .
as expected , from figure [ fig : tsfdvstssp ] we have seen that the smaller the number of fourier modes , the larger the relative error with respect to the initial condition .
we wish to check if there exists an upper limit to the maximum resolution of fourier spectral method . for doing so
, we increase the number of fourier modes and , proportionally , the number of time steps as suggested in @xcite .
results are shown in figure [ fig : maxrestssp ] .
we observe high accuracy in the core ( see smaller disk , left plot ) for small values of @xmath183 , immediately followed by saturation . in a larger disk ( right plot ) , saturation kicks in almost immediately .
the errors reported in figure [ fig : maxrestssp ] suggest that @xmath170 is a reasonable value of fourier modes for the preservation of a two - dimensional quantum vortex .
as explained in the introduction , our motivation to explore the nonuniform finite difference approach is based on the need , for the study of vortex reconnections @xcite , of high _ local _ spatial resolution and , possibly , accuracy . ) at nonuniform grids.,title="fig : " ] ) at nonuniform grids.,title="fig : " ] instead of increasing the number of fourier modes so as to reach higher _ global _ spatial resolution , one can resort to a tssp method with a reasonable number of modes ( considering that mirroring is needed ) and then evaluate the tssp solution on a nonuniform grid , with denser points where they are needed . in figure
[ fig : tsspcoll ] we compare the reasonable case @xmath192 , for which the number of modes is relatively small , but not too small , with different nonuniform grids .
we notice that the numerical integration itself is carried out only once and the fourier coefficients of the solution are stored at each time step .
the evaluation at the grid points is performed afterward , in the post - processing stage , as many times as desired . moreover , tools like the nonuniform fast fourier transform ( nfft , see @xcite ) can be employed for the fast evaluation of trigonometric polynomials at arbitrary point sets .
the constant spatial step - size of the tssp method is @xmath194 , whereas @xmath195 stands for the minimum value of the step - size , in the proximity of the origin , for the nonuniform grids . as seen before
, there is a substantial difference in the error only for @xmath196 , whereas for larger values of @xmath183 evaluating the fourier solution on a nonuniform grid does not worsen the solution .
it is important to note that vortex reconnections , usually , require a dynamics that takes a time of at least @xmath197 .
evaluating a tssp solution on a locally refined grid is , thus , a very promising approach to study quantum vortex reconnections .
after deriving a new accurate pad approximation for the density distribution of a two - dimensional steady - state vortex , we have used it as the initial condition for the gross pitaevskii equation to test the performance of the time - splitting fourier method .
although it can not retain its classical spectral accuracy in space , being as accurate as low - order finite difference on nonuniform grids , it preserves quite well the steady - state solution , especially in the neighborhood of the singularity .
the advantage of a post - processing evaluation on arbitrary points makes this approach suited for applications where _ local _ high resolution is required .
[ [ section ] ] 10 url # 1`#1`urlprefixhref # 1#2#2 # 1#1 w. f. vinen , an introduction to quantum turbulence , phil .
r. soc . a 366 ( 1877 )
( 2008 ) 29252933 .
m. s. paoletti , d. p. lathrop , quantum turbulence , ann .
cond . mat .
phys . 2 ( 2011 ) 213234 .
c. f. barenghi , l. skrbek , k. r. sreenivasan , introduction to quantum turbulence , proc .
usa 111 ( 1 ) ( 2014 ) 46474652 .
u. frisch , turbulence , cambridge university press , cambridge , england , 1995 .
s. b. pope , turbulent flows , cambridge university press , cambridge , england , 200 .
f. hussain , k. duraisamy , mechanics of viscous vortex reconnection , phys .
fluids 23 ( 2011 ) 021701(4 ) .
s. zuccher , m. caliari , a. w. baggaley , c. f. barenghi , quantum vortex reconnections , phys .
fluids 24 ( 125108 ) ( 2012 ) 121 .
l. p. pitaevskii , vortex lines in an imperfect bose gas , sov .
jetp 13 ( 1961 ) 45154 .
gross , hydrodynamics of a superfluid condensate , j. math .
phys . 4 ( 1963 ) 195207 .
s. zuccher , r. l. ricca , helicity conservation under quantum reconnection of vortex rings , phys .
e 92 ( 6 ) ( 2015 ) 061001 .
l. kondaurova , v. lvov , a. pomyalov , i. procaccia , structure of a quantum vortex tangle in @xmath198he counterflow turbulence , phys .
b 89 ( 2014 ) 014502 .
m. d. bustamante , r. m. kerr , 3d euler about a 2d symmetry plane , physica d 237 ( 2008 ) 19121920 .
r. hnninen , a. w. baggaley , vortex filament method as a tool for computational visualization of quantum turbulence , proc .
usa 111 ( 1 ) ( 2014 ) 46674674 .
j. koplik , h. levine , vortex reconnection in superfluid helium , phys .
71 ( 9 ) ( 1993 ) 13751379 .
a. j. allen , s. zuccher , m. caliari , n. proukakis , n. g. parker , c. f. barenghi , vortex reconnections in atomic condensates at finite temperature , phis .
a 90 ( 2014 ) 013601 .
m. w. scheeler , d. kleckner , d. proment , g. l. kindlmann , w. t. m. irvine , helicity conservation by flow across scales in reconnecting vortex links and knots , proc .
usa 111 ( 43 ) ( 2014 ) 1535015355 .
n. g. berloff , pad approximations of solitary wave solutions of the gross pitaevskii equation , j. phys . a : math .
37 ( 2004 ) 16171632 .
s. nazarenko , r. west , analytical solution for nonlinear schrdinger vortex reconnection , j. low temp .
. 132 ( 1 ) ( 2003 ) 110 .
d. p. meichle , c. rorai , m. e. fisher , d. p. lathrop , quantized vortex reconnection : fixed points and initial conditions , phys .
b 86 ( 1 ) ( 2012 ) 014509 .
w. bao , y. cai , mathematical theory and numerical methods for bose einstein condensation , kinet .
models 6 ( 1 ) ( 2013 ) 1135 .
m. thalhammer , m. caliari , c. neuhauser , high - order time - splitting hermite and fourier spectral methods , j. comput .
228 ( 3 ) ( 2009 ) 822832 .
v. gradinaru , strang splitting for the time - dependent schrdinger equation on sparse grids , siam j. numer .
46 ( 1 ) ( 2007 ) 103123 .
m. thalhammer , j. abhau , a numerical study of adaptive space and time discretisations for gross pitaevskii equations , j. comput .
physics 231 ( 20 ) ( 2012 ) 66656681 .
j. h. ferziger , m. peri , computational methods for fluid dynamics , 3rd edition , springer , 2002 .
t. choi , minimal residual methods for complex symmetric , skew symmetric , and skew hermitian systems , tech . rep .
anl / mcs - p3028 - 0812 , computation institute , university of chicago , chicago , illinois ( 2013 ) .
m. caliari , p. kandolf , a. ostermann , s. rainer , comparison of software for computing the action of the matrix exponential , bit 54 ( 1 ) ( 2014 ) 113128 .
w. bao , s. jin , p. a. markowich , on time - splitting spectral approximations for the schrdinger equation in the semiclassical regime , j. comput .
175 ( 2 ) ( 2002 ) 487524 .
j. keiner , s. kunis , d. potts , using nfft 3a software library for various nonequispaced fast fourier transforms , acm trans .
software 36 ( 4 ) ( 2009 ) 19:119:30 .
_ the case @xmath69 .
_ the coefficients of this expansion are already known , however it is instructive to proceed with their derivation in order to understand how it works .
we have to compute @xmath199 coefficients , @xmath200 , @xmath201 and @xmath202 , therefore we can use only @xmath199 equations .
these equations are obtained by nullifying , respectively , the coefficients of the terms @xmath203 , @xmath204 and @xmath205 in the numerator @xmath206 ( lower - order powers of @xmath83 ) . by nullifying the coefficient of @xmath203
, we get @xmath207 from which @xmath208 . by nullifying the coefficient of @xmath204 , and replacing @xmath202 with the expression above
, we get @xmath209 which gives @xmath210 .
if we now nullify the coefficient of @xmath205 and replace @xmath202 with @xmath211 and @xmath201 with @xmath212 , we get the following equation @xmath213 clearly , @xmath214 is not acceptable , nor is @xmath215 .
the only acceptable value is @xmath216 . as we mentioned before , equation
can not be satisfied exactly , however , an _ a posteriori _ evaluation reveals that the remaining coefficients of @xmath83 are smaller than @xmath217 and monotonically decreasing with @xmath150 . _ the case @xmath219 . _
since we have to compute @xmath220 coefficients we need @xmath220 equations , which are obtained by imposing that the coefficients of the terms @xmath203 , @xmath204 , @xmath205 , @xmath221 and @xmath222 must be zero . by nullifying the coefficient of @xmath203
we still get the same equation as for @xmath69 , @xmath223 , from which @xmath208 . by nullifying the coefficient of @xmath204 , and replacing @xmath202 with the expression above
, we get @xmath224 which is easy to solve for @xmath225 leading to @xmath226 now we collect terms in @xmath205 and impose its coefficient to be zero . in this equation
we replace @xmath202 and @xmath225 with the expressions derived above and get the equation @xmath227 which we solve for @xmath228 : @xmath229 then we nullify the coefficient of @xmath221 , substitute all previously found @xmath202 , @xmath225 and @xmath228 , getting the equation @xmath230 which gives @xmath231 finally , we nullify the coefficient of @xmath222 , substitute @xmath202 , @xmath225 , @xmath228 and @xmath201 , and get the equation for @xmath200 @xmath232 this equation must be solved numerically and leads to many real solutions .
however the only value that reproduces a physical behavior of @xmath233 for @xmath234 is @xmath235 .
it is possible to compute the first derivative and verify that @xmath236 for all @xmath237 . in other words
, @xmath238 is a physical , monotonically increasing , approximation of the density due to a two - dimensional quantum vortex .
again , the coefficients of @xmath83 that are not zero are , indeed , smaller than @xmath239 and monotonically decreasing with @xmath150 .
_ the case @xmath181 .
_ now we have 7 coefficients to compute , therefore we need @xmath241 equations , i.e. we need to nullify the coefficients of @xmath83 for @xmath242 . by canceling the term @xmath203 and solving for @xmath202
we get the usual expression @xmath208 . by nullifying the term @xmath204 , substituting @xmath202 and solving for @xmath225
we get @xmath243 , which is the same expression obtained for @xmath238 . by canceling the term @xmath205 , substituting @xmath202 and @xmath225
as found , and solving for @xmath244 we get @xmath245 by canceling the term @xmath221 , substituting @xmath202 , @xmath225 , @xmath244 and solving for @xmath246 we get @xmath247 by canceling the term @xmath222 , substituting @xmath202 , @xmath225 , @xmath244 and @xmath246 , and solving for @xmath228 we get @xmath248 by canceling the term @xmath249 , substituting all known @xmath250 , @xmath246 , @xmath228 , and solving for @xmath201 we get @xmath251 finally by canceling the term @xmath252 , substituting all @xmath250 , @xmath253 previously found , we get an equation for @xmath200 @xmath254 this equation has many real solutions , which can be determined numerically .
however , the only value that leads to a physically acceptable @xmath255 for @xmath89 is @xmath256 .
after computing all other coefficients and the first derivative , it is straighforward to verify that @xmath257 for all @xmath237 , i.e. @xmath258 is a physical , monotonically increasing , approximation of the density for a two - dimensional quantum vortex . as observed before , the coefficients of @xmath83 that are not zero are smaller than @xmath259 and monotonically decreasing with @xmath150 | we extensively study the numerical accuracy of the well - known time splitting fourier spectral method for the approximation of singular solutions of the gross pitaevskii equation . in particular
, we explore its capability of preserving a steady - state vortex solution , whose density profile is approximated by a very accurate diagonal pad expansion of order 8 , here explicitly derived for the first time .
although the fourier spectral method turns out to be only slightly more accurate than a time splitting finite difference scheme , the former is reliable and efficient .
moreover , at a post - processing stage , it allows an accurate evaluation of the solution outside grid points , thus becoming particularly appealing when high resolution is needed , such as in the study of quantum vortex interactions .
quantum fluids , nonuniform finite differences , time splitting , fourier spectral method |
maximally supersymmetric ( @xmath1 ) super yang - mills theory in four dimensions is a very special quantum field theory in several ways .
first of all , it has surprisingly simple and well - behaved amplitudes .
the @xmath1 supermultiplet is cpt self - conjugate and in addition the theory is both classically and quantum mechanically conformally invariant because the renormalization group @xmath0-function vanishes identically .
even more remarkably , although gravitational interactions are absent , the @xmath1 theory is intimately related to a supergravity theory via the celebrated ads / cft correspondence .
multiloop @xmath1 scattering amplitudes are very elegantly studied within the on - shell superspace formalism @xcite . the principle is to arrange the entire supermultiplet as a superfield expanded in grassmann variables .
all possible scattering combinations are realized by formation of superamplitudes , defined as generating functions of @xmath3 copies of superfields corresponding to the external legs .
single amplitudes are then projected out by unique strings of grassmann differential operators or pieced together in a graphical framework .
superamplitudes for general particle and helicity configurations may be constructed with the britto - cachazo - feng - witten ( bcfw ) on - shell recursion relations @xcite .
it is of course very interesting to examine features of the @xmath1 theory in more general settings .
super yang - mills theories with less than maximal supersymmetry do not possess all of the amazing properties mentioned , although the superspace formalism can be generalized @xcite . however , exactly marginal deformations of the @xmath1 super yang - mills theory preserving only reduced amount of supersymmetry in particular inherit conformal invariance @xcite , and have recently attracted considerable attention especially catalyzed by the ads / cft duality .
exactly marginal deformations have also been subject to intense investigations within the perturbative regime @xcite . in particular
, it has been of special interest to understand how scattering amplitudes are modified by the so - called @xmath0-deformation .
it can be shown that conformal invariance of the planar @xmath0-deformed super yang - mills theory requires reality of @xmath0 , the deformation parameter .
the theory enjoys only @xmath4 supersymmetry and a global @xmath5 flavor symmetry , and may be formulated with both symmetries manifest in the @xmath4 superspace formalism in terms of three charged chiral superfields and neutral vector superfield , and a star product operation which incorporates the charges of the fields under the flavor symmetry . using this lagrangian approach
the @xmath0-deformed feynman rules of the deformed theory may be obtained , and the behaviour of its scattering amplitudes elucidated @xcite .
it turns out that the @xmath1 and @xmath0-deformed theories are very similar . motivated by the developments in exactly marginal deformations and inspired by how powerfully on - shell scattering amplitudes
are constructed from analyticity and unitarity in the @xmath1 on - shell framework , we recast the deformed theory in terms of generating functions .
several modern techniques apply to present gauge theory scattering amplitudes very effectively . with all states in the adjoint representation of the gauge group ,
say @xmath6 , any tree - level amplitude can be color decomposed as @xmath7 where @xmath8 is the gauge coupling constant and @xmath9 are partial amplitudes soaking up the entire kinematical structure corresponding to a particular ordering of the @xmath3 external legs encoded by the trace of the gauge group generators @xmath10 .
the sum is over all permutations @xmath11 with trace preserving cyclic permutations @xmath12 modded out .
the partial amplitudes are written in terms of lorentz invariant , little group covariant holomorphic and antiholomorphic spinor products . in order to distinguish the two chiralities we use angle and square brackets , and define @xmath13 } = -{[ji ] } \equiv \epsilon_{\dot\alpha\dot\beta } \tilde\lambda_i^{\dot\alpha}\tilde\lambda_j^{\dot\beta}\end{aligned}\ ] ] for commuting spinors @xmath14 and @xmath15 related to momentum by @xmath16 .
the @xmath1 gauge multiplet has the unique property of cpt self - conjugacy which implies that all on - shell states can be assembled by a single holomorphic superfield @xmath17 .
it is expanded in grassmann variables @xmath18 where @xmath19 is a fundamental index of @xmath20 , the @xmath21-symmetry group of the theory . within this setup
the sixteen physical states transform in @xmath22-rank antisymmetric tensor representations as two gluons @xmath23 and @xmath24 , four fermion pairs @xmath25 and @xmath26 , plus six real scalars @xmath27 .
the tensor rank @xmath22 and particle helicity @xmath28 are related through @xmath29 .
the superfield is @xcite ( repeated indices are summed ) @xmath30 there also exists an antiholomorphic superfield @xmath31 linked to @xmath17 by the grassmann fourier transform , but with the exact same particle content encoded . due to this equivalence
either representation may be preferred . in this paper
we reserve the holomorphic and antiholomorphic descriptions for @xmath32 and @xmath33 amplitudes respectively .
proliferation of scattering amplitudes in @xmath1 super yang - mills theory is handled by introduction of superamplitudes , which are functions of @xmath3 copies of the superfield , one for each external leg .
the full @xmath3-point tree - level superamplitude is organized ascendingly according to grassmann degree in steps of four , @xmath34 ranging from eight @xmath35 s to @xmath36 .
it follows in particular that all mhv amplitudes may be packaged into a generating function , also referred to as the mhv superamplitude .
each term thus corresponds to a regular scattering amplitude involving gluons , fermions and scalars .
it is convention to extract the mhv sector from the full superamplitude as an overall factor .
the mhv tree - level superamplitude is in addition to the well - known overall momentum conservation proportional to an eightfold grassmann delta function conserving total supermomentum @xmath37 , and is defined by @xcite @xmath38 for calculational purposes it proves advantageous to expand the grassmann delta function present in the mhv superamplitude as a sum of monomials in the @xmath35 s , first in all possible values of the group index , @xmath39 and then using @xmath40 for grassmann variables , @xmath41 frequently we write holomorphic and antiholomorphic spinor products of supermomenta of the individual legs defined by @xmath42}\equiv \tilde\eta_i^a{[ij]}\tilde\eta_j^a\;.\end{aligned}\ ] ] consequently , the mhv generating function reaches the very clean form @xmath43 with four - momentum conservation stripped . for four external legs ,
some simple examples of mhv component amplitudes are @xmath44 analogous to the mhv superamplitude we may use the antiholomorphic superfields to build a @xmath33 generating function .
the @xmath33 superamplitude conserves total conjugate supermomentum @xmath45 and is apart from an ordinary momentum conserving delta function given by @xmath46 } } = i(-1)^n \frac{\prod_{a=1}^4\sum_{i < j}{[\tilde
q_i^a \tilde q_j^a ] } } { \prod_{r=1}^n{[r(r+1)]}}\;.\end{aligned}\ ] ] it is mapped from the @xmath47-coordinates to the untilded superspace using the grassmann fourier transform realized by the @xmath3-leg operator @xmath48 as a first example of @xmath33 component amplitudes consider the equivalent reinterpretation of the mhv amplitudes - in the tilded superspace , @xmath49 } } { { [ 12]}{[23]}{[34]}{[41]}}\ ; , \\
a_4^{{\overline{\mathrm{mhv } } } } ( 1 ^ -_g,2 ^ -_{f_{d}},3^+_{f_{abc}},4^+_{g_{abcd } } ) & = i\frac{{[\tilde q_{3a}\tilde q_{4a}]}{[\tilde q_{3b}\tilde q_{4b } ] } { [ \tilde q_{3c}\tilde q_{4c}]}{[\tilde q_{2d}\tilde q_{4d } ] } } { { [ 12]}{[23]}{[34]}{[41]}}\ ; , \\ a_4^{{\overline{\mathrm{mhv } } } } ( 1 ^ -_{f_{d}},2 ^ -_{f_{c}},3_{s_{ab}},4^+_{g_{abcd } } ) & = i\frac{{[\tilde q_{3a}\tilde q_{4a}]}{[\tilde q_{3b}\tilde q_{4b } ] } { [ \tilde q_{2c}\tilde q_{4c}]}{[\tilde q_{1d}\tilde q_{4d } ] } } { { [ 12]}{[23]}{[34]}{[41]}}\;.\end{aligned}\ ] ] the supersymmetric bcfw recursion relations @xcite generate all tree - level superamplitudes in @xmath1 super yang - mills theory in terms of nested sums of dual superconformal invariants @xcite from just @xmath32 and @xmath33 amplitudes . the essense is to deform the on - shell superspace by a supershift and recover the physical amplitude by residue calculus . in this paper
it suffices to examine in detail only the @xmath50 superamplitude .
it comes with the very compact result @xcite @xmath51 where , however , @xmath52 are complicated dual superconformal invariants ( see e.g. @xcite ) , @xmath53 depending on another intricate grassmann valued object , @xmath54 , defined by @xmath55 equipped with all necessary machinery of @xmath1 superamplitudes we are now ready to consider applications in @xmath0-deformed super yang - mills theory .
marginal deformations of conformally invariant supersymmetric gauge theories were first systematically studied by leigh and strassler @xcite , and have subsequently been analyzed extensively both perturbatively and at strong coupling in @xcite just to mention a few . before exactly marginal deformations are treated
, we shall first pay brief attention to the lagrangian formulation of the undeformed @xmath1 super yang - mills theory .
the standard @xmath4 superspace formalism applies by construction to field theories with @xmath4 supersymmetry .
but since @xmath56 supersymmetric theories may be reduced to multiple unextended supermultiplets coupled together , the @xmath4 framework proves useful in a much richer class of theories .
the @xmath1 particle content decomposes into one @xmath4 vector multiplet and three @xmath4 chiral multiplets .
the lagrangian can be written @xcite @xmath57+{\mathrm{h.c.}}\right\}\;,\end{aligned}\ ] ] where @xmath58 is the vector superfield and @xmath59 , @xmath60 and @xmath61 are the three chiral superfields . the kinetic term for the vector superfield involves as usual the superfield strength defined by @xmath62 where @xmath63 and @xmath64 are supercovariant derivatives .
this lagrangian is manifestly @xmath4 supersymmetric being constructed using the @xmath4 superspace prescription .
moreover , the three chiral superfields enjoy manifest @xmath65 flavor symmetry , but the full @xmath66 @xmath21-symmetry is obscured .
the general leigh - strassler theory is parametrized by complexification of the gauge coupling @xmath8 and modification of the superpotential by the substitution @xmath67 for complex @xmath68 and @xmath69 .
the deformations break supersymmetry to @xmath4 .
in order for the deformation to become exactly marginal and thereby inherit finiteness at the quantum level , the parameters must be highly constrained .
leigh and strassler demonstrated the existence of a three - complex dimensional surface in the coupling constant space of conformally invariant theories with @xmath4 supersymmetry .
this surface can be defined as a quite complicated level set @xmath70 .
the condition however has to be calculated perturbatively and unfortunately its form is not known beyond few loops except for very restrictive choices of the couplings .
the leigh - strassler deformed theory generated by the superpotential substitution includes as a special case a one - parameter family of theories known as @xmath0-deformations , which extend to quantum finiteness to all orders in the planar limit @xcite .
the model follows by tightening the assumptions on the parameters and setting @xmath71 , @xmath72 and @xmath73 , so that for @xmath74 real , the new superpotential becomes @xmath75 whence the undeformed theory is recovered by sending @xmath76 .
it turns out that the @xmath0-deformation can be understood in terms of a special operation between the superfields , namely the star product , which will prove invaluable when we evaluate deformations of scattering amplitudes @xcite . in order to define the star product
it is necessary to discuss the left - over symmetries of the @xmath0-deformed lagrangian .
the original @xmath21-symmetry is broken from @xmath20 to @xmath77 , but it is observed that the deformed theory in addition is invariant under a global @xmath5 flavor symmetry of the three chiral superfields .
more specifically the symmetries can be expressed as @xmath78 the vector superfield being neutral under these transformations .
the symmetry charges are @xmath79}\equiv ( 0,+1,-1,0)\ ; , \quad\;\ ; q^{[2]}\equiv ( -1,+1,0,0)\;.\end{aligned}\ ] ] suppose that @xmath80 and @xmath81 are chiral superfields with flavor symmetry charges @xmath82}_i$ ] and @xmath82}_j$ ] respectively .
then the star product operation together with the @xmath0-deformed commutator between these fields is defined by @xmath83}q_j^{[2]}-q_i^{[2]}q_j^{[1]}\right)}\phi_i\phi_j\ ; , \quad [ \phi_i,\phi_j]_\beta \equiv e^{i\pi\beta_{ij}}\phi_i\phi_j - e^{-i\pi\beta_{ij}}\phi_j\phi_i\;,\end{aligned}\ ] ] where @xmath84 is the antisymmetric matrix @xmath85 for @xmath74 real it follows that the star product is simply the usual product adjusted by an overall flavor dependent phase factor . of important properties of the star product we mention associativity , which is a consequence of additivity of the flavor charges , and noncommutativity , the latter being obvious from the definition
. prior to implementation in the @xmath1 lagrangian the star product must extend to the antichiral superfields @xmath86 , @xmath87 and @xmath88 .
they carry opposite charges to the chiral superfields under the @xmath5 symmetry so that @xmath89 is chargeless and @xmath90 .
we therefore note @xmath91}\equiv ( 0,-1,+1,0 ) = -q^{[1]}\ ; , \quad\;\ ; \bar q^{[2]}\equiv ( + 1,-1,0,0 ) = -q^{[2]}\;,\end{aligned}\ ] ] whereby both antichirality and mixed chirality star products become well defined . with this operation
the @xmath0-deformation can be induced in the @xmath1 lagrangian by substituting all ordinary products between superfields with star products , or equivalently replacing all commutators with @xmath0-deformed brackets . in particular , the superpotential can be written @xmath92 and hence the @xmath0-deformed lagrangian takes the form @xmath93_\beta+{\mathrm{h.c.}}\right\}\;.\end{aligned}\ ] ] by consideration of the component version of the lagrangian for the @xmath1 super yang - mills theory rather than its superspace representation it is elementary to derive the modified color - ordered feynman rules .
we mention that non - consecutive four - scalar interactions and yukawa vertices without particles from the vector multiplet exhaust the @xmath0-deformed vertices ( see fig .
[ betadependentvertices ] ) , and refer the reader to @xcite for detailed calculations .
the @xmath0-deformation adds overall phase factors to the original vertex rules of the depicted four - scalar and yukawa couplings .
we denote the three complex scalar field components of the chiral superfields @xmath80 by @xmath94 , while @xmath95 label their fermionic superpartners .
notice that the ordering of the legs is crucial.,title="fig : " ] the @xmath0-deformation adds overall phase factors to the original vertex rules of the depicted four - scalar and yukawa couplings .
we denote the three complex scalar field components of the chiral superfields @xmath80 by @xmath94 , while @xmath95 label their fermionic superpartners .
notice that the ordering of the legs is crucial.,title="fig : " ] the @xmath0-deformation adds overall phase factors to the original vertex rules of the depicted four - scalar and yukawa couplings .
we denote the three complex scalar field components of the chiral superfields @xmath80 by @xmath94 , while @xmath95 label their fermionic superpartners .
notice that the ordering of the legs is crucial.,title="fig : " ] scattering amplitudes in @xmath1 and @xmath0-deformed super yang - mills theories seem closely related when comparing their feynman rules . indeed , only three specific four - scalar and yukawa interactions are modified , and this deformation introduces nothing but prefactors to the vertex rules .
several comments are important in this connection .
the two theories have the same particle content and it is easy to see that many amplitudes are actually identical . for a @xmath0-deformed amplitude to agree with the corresponding @xmath1 expression , the phases of the @xmath0-dependent vertices should cancel each other , or such vertices should simply be absent in the diagram .
whole classes of amplitudes are insensitive to the phase deformation , the most obvious being all tree - level amplitudes with external gluons and gluinos exclusively . in general , amplitudes in the deformed theory are @xmath0-dependent .
the modifications however turn out to be surprisingly uncomplicated . at first
we realize that the original and deformed theories are equivalent at tree - level , up to a multiplicative prefactor .
this conclusion follows from the simplicity of the color - ordered feynman vertices .
but it can be shown that the product of phase factors from the individual vertices is independent of the internal structure of the diagram .
suppose the arbitrary color - ordered tree - level amplitude in consideration has any combination of @xmath3 external fields and label them @xmath96 .
then the @xmath0-dependence of this amplitude is entirely captured by @xmath97 $ ] .
actually this statement applies not only to tree - level amplitudes , but is guaranteed to hold to all orders in perturbation theory in the planar limit .
the claim has profound consequences for our applications of @xmath0-deformed super yang - mills theory .
a proof based on effective vertices of only external legs was provided in @xcite .
shortly we report a similar strategy for supervertices .
we end this section by deriving an expression for the phase factor of any given @xmath3-point amplitude in terms of only the @xmath5 charges of the superfields corresponding to the external legs . upon invoking the definition of the star product
we deduce the generalization to @xmath3 legs , @xmath98 } { { \operatorname{tr}}\left[\lambda_1\lambda_2\cdots\lambda_n\right ] } \nonumber \\ & = \prod_{i=1}^{n-1}\exp\bigg[i\pi\beta_r\sum_{j = i+1}^n\left ( q^{[1]}_{\lambda_i}q^{[2]}_{\lambda_j}- q^{[2]}_{\lambda_i}q^{[1]}_{\lambda_j}\right)\bigg]\;,\end{aligned}\ ] ] using the rather selfexplanatory notation with @xmath99 and @xmath100 denoting the symmetry charges of the field @xmath101 .
finally may be recast perhaps more conveniently as @xmath102}_{\lambda_i}q^{[2]}_{\lambda_j}- q^{[2]}_{\lambda_i}q^{[1]}_{\lambda_j}\right)\bigg]\;. \label{phasefromcharges}\end{aligned}\ ] ] this result is essential for the rest of the paper .
now that we have gained confidence with the basic structure and interactions of @xmath0-deformed super yang - mills theory , it is very natural to attempt to incorporate all scattering amplitudes sectorwise into generating functions instead of relying on traditional feynman calculations .
thereby established @xmath1 superspace applications such as intermediate state sums by grassmann integration become compatible with the @xmath0-deformed amplitudes .
we first identify particles in @xmath1 on - shell superspace with the components of the @xmath4 vector and chiral superfields .
recall that the sixteen physical states in the @xmath1 supermultiplet can be realized as two gluons @xmath23 and @xmath24 , four fermion pairs @xmath25 and @xmath26 , plus six real , self - dual scalars @xmath27 , all completely antisymmetric in the displayed fundamental @xmath66 indices .
the gluons of course belong to the vector multiplet , while the remaining fermions and scalars of the theory can be chosen such that @xcite @xmath103 for @xmath104 and @xmath105 .
we consider the @xmath32 sector and derive an expression for the @xmath3-point tree - level @xmath32 generating function . with that result at hand
the @xmath33 superamplitude will follow immediately from the fourier connection .
clearly , both expressions should reduce to the original @xmath1 superamplitudes in the limit @xmath106 .
the upshot of the preceding section was that the @xmath1 and @xmath0-deformed theories have identical planar sectors up to simple phase factors , which for any given amplitude to all orders in perturbation theory are determined by the configuration of its external legs and their @xmath5 symmetry charges according to .
the logical solution is therefore to take the @xmath1 superamplitude and just attach to each component the appropriate phase factor .
we remind ourselves that the @xmath1 @xmath32 superamplitude reads @xmath107 and therefore the task concentrates on translating the individual grassmann signatures to @xmath0-dependent expressions .
we circumvent this obstacle by assigning @xmath5 charges to the fermionic coordinates of the @xmath1 on - shell superspace instead of the superfields such that the on - shell superspace version of the star product imitates the star product defined in @xmath4 superspace .
this solution is closely related to the light cone superspace star product introduced in @xcite .
we apply the usual charges @xmath108}$ ] and @xmath109}$ ] , with opposite signs in the fourier transformed superspace .
more precisely @xmath110 carries charges @xmath108}_a$ ] and @xmath109}_a$ ] .
the trick is now to construct a pair of differential operators to decode the @xmath35-patterns and thereby count symmetry charges .
preferably the @xmath35-strings should be eigenstates of these operators , and the symmetry charges the corresponding eigenvalues . in order to achieve this
we introduce the operator @xmath111}_i \equiv \sum_{a=1}^4 q^{[1,2]}_a\eta_{ia}\partial_{ia}\end{aligned}\ ] ] with the two explicit components @xmath112}_i
\equiv \eta_{i2}\partial_{i2}- \eta_{i3}\partial_{i3}\ ; , \quad { \hat{{\mathcal}{q}}}^{[2]}_i \equiv \eta_{i2}\partial_{i2}- \eta_{i1}\partial_{i1}\end{aligned}\ ] ] such that @xmath113}_i$ ] and @xmath114}_i$ ] measure the symmetry charges @xmath115}$ ] and @xmath116}$ ] respectively for leg @xmath117 , and are by construction grassmann even .
for instance the action on the grassmann combination corresponding to the negative helicity fermion of the chiral superfield @xmath61 is @xmath113}_i(\eta_{i1}\eta_{i2}\eta_{i4 } ) = \eta_{i1}\eta_{i2}\eta_{i4}$ ] and @xmath114}_i(\eta_{i1}\eta_{i2}\eta_{i4 } ) = 0 $ ] in agreement with .
it is easy to see that the mechanism works in general .
the vector multiplet in particular has vanishing charges as it should . substituting @xmath113}_i$ ] and @xmath114}_i$ ] for the constant symmetry charges in yields the operator version of the phase factor @xmath118}{\hat{{\mathcal}{q}}}_q^{[2]}- { \hat{{\mathcal}{q}}}_p^{[2]}{\hat{{\mathcal}{q}}}_q^{[1]}\right)\bigg]\;,\end{aligned}\ ] ] which enables us to formally write the @xmath0-deformed @xmath32 superamplitude as @xmath119}{\hat{{\mathcal}{q}}}_q^{[2]}- { \hat{{\mathcal}{q}}}_p^{[2]}{\hat{{\mathcal}{q}}}_q^{[1]}\right)\bigg ] { { \mathcal}{a}}_n^{{\mathrm{mhv } } } \nonumber \\ & = i\prod_{r=1}^n{\langler(r+1)\rangle}^{-1 } \exp\bigg[i\pi\beta_r\sum_{p < q}\left ( { \hat{{\mathcal}{q}}}_p^{[1]}{\hat{{\mathcal}{q}}}_q^{[2]}- { \hat{{\mathcal}{q}}}_p^{[2]}{\hat{{\mathcal}{q}}}_q^{[1]}\right)\bigg ] \prod_{\ell=1}^4\sum_{i < j}{\langleij\rangle}\eta_{i\ell}\eta_{j\ell}\;.\end{aligned}\ ] ] the @xmath0-deformed generating function has by construction the desired property @xmath120 in the limit @xmath106 . in order to streamline the notation we first introduce an alternative version of the kronecker delta function defined by @xmath121 where for our purposes the set @xmath122 should only contain unique elements .
let us then turn to the evaluation of the action of the symmetry charge operators on a generic string of grassmann variables present in the @xmath32 superamplitude .
the vector multiplet sector commutes right through the differentiation , while numerous kronecker delta functions are produced when hitting the @xmath35-variables pertinent to the chiral multiplets . keeping track of all possible combinations we find @xmath123}{\hat{{\mathcal}{q}}}_q^{[2]}- { \hat{{\mathcal}{q}}}_p^{[2]}{\hat{{\mathcal}{q}}}_q^{[1]}\right )
\prod_{\ell=1}^4\eta_{i_\ell \ell}\eta_{j_\ell \ell } = \nonumber \\ & \left [ ( \delta_{p\{i_2,j_2\}}-\delta_{p\{i_3,j_3\ } } ) ( \delta_{q\{i_2,j_2\}}-\delta_{q\{i_1,j_1\ } } ) -(p\leftrightarrow q ) \right ] \prod_{\ell=1}^4\eta_{i_\ell \ell}\eta_{j_\ell \ell}\;.\end{aligned}\ ] ] it immediately follows that @xmath124 .
the phase factor therefore reduces slightly into @xmath125 \bigg)\end{aligned}\ ] ] such that the @xmath0-deformed @xmath32 superamplitude becomes @xmath126 with the shorthand notation @xmath127 the form of coincides with our expectations .
we see that the three spin factors corresponding to the chiral multiplets are now correlated through a phase matrix @xmath128 while the original fourth @xmath66 factor identified with the vector multiplet is left unchanged and can be separated out .
hence , @xmath4 supersymmetry is manifest .
the result also reflects that alternatively we could have considered a theory of the three chiral superfields with the appropriate phase dependent interactions and then have coupled the corresponding superspace structure to the bare @xmath4 @xmath32 superamplitude addressed in @xcite .
however , the approach presented here more efficiently generalizes to superamplitudes beyond the @xmath32 sector .
before we continue let us pause for a second and study the phase operator defined on on - shell superamplitudes in phase - deformed on - shell superspace .
formally , it is given in terms of its taylor series in the bilocal pure phase @xmath129 such that @xmath130\;.\end{aligned}\ ] ] the phase generator comes with canonical ordering of the external legs . however , with the @xmath131 different configurations introduced by color decomposition in mind , it is necessary to define the operator for an arbitrary permutation @xmath132 of @xmath133 , i.e. @xmath134}{\hat{{\mathcal}{q}}}_{\sigma(q)}^{[2]}- { \hat{{\mathcal}{q}}}_{\sigma(p)}^{[2]}{\hat{{\mathcal}{q}}}_{\sigma(q)}^{[1]}\right)\bigg]\;.\end{aligned}\ ] ] with this formulation the phase generator can be brought to depend only on the positions of the particles in the cyclic chain and is thus universal .
amplitudes should have cyclic symmetry .
we prove the crucial property of cyclicity of phase operator itself .
in general , bilocal operators such as the phase generator map cyclic functions to non - cyclic functions .
however , depending on the space upon which the operators act , the non - cyclic remainder may be brought to vanish .
the most prominent example is the level - one yangian generators .
it is straightforward to see that @xmath135}{\hat{{\mathcal}{q}}}_q^{[2]}- { \hat{{\mathcal}{q}}}_p^{[2]}{\hat{{\mathcal}{q}}}_q^{[1]}\right)- \sum_{2\leq p < q\leq 1}\left ( { \hat{{\mathcal}{q}}}_p^{[1]}{\hat{{\mathcal}{q}}}_q^{[2]}- { \hat{{\mathcal}{q}}}_p^{[2]}{\hat{{\mathcal}{q}}}_q^{[1]}\right ) = 2\sum_{p=1}^n\bigg ( { \hat{{\mathcal}{q}}}_1^{[1]}{\hat{{\mathcal}{q}}}_p^{[2]}- { \hat{{\mathcal}{q}}}_1^{[2]}{\hat{{\mathcal}{q}}}_p^{[1]}\bigg)\;,\end{aligned}\ ] ] where the two indicated configurations @xmath136 and @xmath137 differ by a cyclic transformation .
the on - shell amplitudes are neutral with respect to the flavor symmetry , which means that @xmath138}{{\mathcal}{a}}_n^{\beta_r } = \sum_{p=1}^n{\hat{{\mathcal}{q}}}_p^{[2]}{{\mathcal}{a}}_n^{\beta_r } = 0\;.\end{aligned}\ ] ] therefore the restriction of the phase factor difference to this space vanishes .
moreover , the remainder term at arbitrary order in the deformation parameter is bound to annihilate the amplitudes .
we thus conclude that @xmath139 and @xmath140 let us next consider the phase - deformed color reflection identity .
it is easy to realize that reversal of the order of the external legs inverts the phase , @xmath141 remembering that inversion of undeformed amplitudes introduces a factor of @xmath142 we thus realize that the phase - deformed analogue has to be phase dressed .
alternatively , we can compensate for the transformation of the phase via the deformation parameter , @xmath143 we calculate a number of phase factors using to obtain explicitly some of the components of the @xmath0-deformed mhv generating function .
our results agree with the phase structure obtained via usual star product defined for @xmath4 superfields .
the relevant phase matrix indices for the amplitudes written below are @xmath144 , @xmath145 , @xmath146 , @xmath147 , @xmath148 and @xmath149 respectively .
@xmath150 the @xmath0-deformed @xmath32 superamplitude allows an almost trivial continuation to the @xmath33 superspace .
it just amounts figuring out an expression for the phase factors .
but because the signs of all @xmath5 charges just get flipped , and thus cancel in , the phase factors of the two sectors are identical in form , @xmath151 completely analogous to , implementation of the phase factor matrix @xmath152 in the original @xmath1 @xmath33 superamplitude therefore yields the @xmath0-deformed @xmath33 generating function @xmath153}^{-1 } \sum_{\{i\}<\{j\ } } { { \mathcal}{p}}^{\beta_r,\,{\overline{\mathrm{mhv}}}}_{i_1j_1;i_2j_2;i_3j_3 } \prod_{\ell=1}^4 { [ \tilde q_{i_\ell}^\ell \tilde q_{j_\ell}^\ell]}\;,\end{aligned}\ ] ] or , alternatively using the grassmann fourier transform , in holomorphic superspace , @xmath154}^{-1 } \sum_{\{i\}<\{j\ } } { { \mathcal}{p}}^{\beta_r,\,{\overline{\mathrm{mhv}}}}_{i_1j_1;i_2j_2;i_3j_3 } \prod_{\ell=1}^4 { \mathcal}e_{n;\ell}(i_\ell , j_\ell ) { [ i_\ell j_\ell]}\;,\end{aligned}\ ] ] for @xmath155 defined by @xmath156
any pattern of grassmann superspace variables may be mapped to a definite phase factor , allowing easy extension to non - mhv amplitudes .
consider in full generality the @xmath1 @xmath157 tree - level superamplitude , denoted @xmath158 as usual . in this formal development
the precise expression for this superamplitude is not important . upon application of the phase factor operator to @xmath158
the @xmath0-deformed @xmath157 superamplitude can be reached in the form @xmath159}{\hat{{\mathcal}{q}}}_q^{[2]}- { \hat{{\mathcal}{q}}}_p^{[2]}{\hat{{\mathcal}{q}}}_q^{[1]}\bigg)\bigg ] { { \mathcal}{a}}_n^{{\mathrm{n}^k\mathrm{mhv}}}\;.\end{aligned}\ ] ] the @xmath157 superamplitude has grassmann degree @xmath160 and each component thus carries a grassmann string with @xmath161 distinct indices .
the general phase matrix therefore has @xmath162 labels , but the form is completely similar to .
the sets associated with the kronecker delta functions will just have @xmath163 unique elements each . to be more specific
we will establish the mhv vertex expansion of cachazo , svrcek and witten ( csw ) .
our proof is the generating function analogue of @xcite , now formulated in terms of bilocal operators .
the important point is that neutrality of all vertices implies neutrality of any amplitude .
let us quickly refresh our memory of the csw rules for constructing the @xmath157 generating tree .
the procedure is to draw all tree graphs with @xmath164 vertices , distribute @xmath3 color - ordered legs , to each of the vertices associate a @xmath32 superamplitude and finally connect them by a scalar feynman propagator and for consistency equate the grassmann coordinates on both ends of the internal lines between them .
it is now rather elementary to extract all contributions within a particular topology using grassmann integration over the @xmath165 internal lines .
the mhv superrules therefore translate into @xmath166 { { \mathcal}{a}}_{(1)}^{{\mathrm{mhv}}}{{\mathcal}{a}}_{(2)}^{{\mathrm{mhv}}}\cdots { { \mathcal}{a}}_{(i)}^{{\mathrm{mhv}}}{{\mathcal}{a}}_{(k+1)}^{{\mathrm{mhv}}}\end{aligned}\ ] ] where the discrete sum over all graphs incorporates inequivalent topologies .
although suppressed here it is important to realize that @xmath167 is an off - shell momentum .
however , in order to have a well - defined product of on - shell trees , the momenta that enters the @xmath164 superamplitudes must be null - projections constructed from the corresponding off - shell momenta using an arbitrary null reference vector .
it suffices to consider the @xmath50 case to exhaust the general factorization pattern . by cyclicity of the @xmath1 superamplitudes we can without loss of generality
arrange their internal legs alternating first and last .
thus the lines of the two supertrees can be labelled @xmath168 and @xmath169 respectively . the sum of the associated phases
is @xmath170}{\hat{{\mathcal}{q}}}_q^{[2]}- { \hat{{\mathcal}{q}}}_p^{[2]}{\hat{{\mathcal}{q}}}_q^{[1]}\bigg)+ \sum_{p = k}^{n-1}\sum_{q = p+1}^n \bigg({\hat{{\mathcal}{q}}}_p^{[1]}{\hat{{\mathcal}{q}}}_q^{[2]}- { \hat{{\mathcal}{q}}}_p^{[2]}{\hat{{\mathcal}{q}}}_q^{[1]}\bigg)\;.\end{aligned}\ ] ] we relate this expression to the phase of the full tree , e.g. legs @xmath171 , and observe that @xmath172}{\hat{{\mathcal}{q}}}_q^{[2]}- { \hat{{\mathcal}{q}}}_p^{[2]}{\hat{{\mathcal}{q}}}_q^{[1]}\bigg)\;.\end{aligned}\ ] ] by flavor charge conservation at each @xmath32 supertree the displayed double sum annihilates the expanded tree , and it can thus be discarded , @xmath173}{\hat{{\mathcal}{q}}}_q^{[2]}- { \hat{{\mathcal}{q}}}_p^{[2]}{\hat{{\mathcal}{q}}}_q^{[1]}\bigg ) { { \mathcal}{a}}_{(1)}^{{\mathrm{mhv}}}{{\mathcal}{a}}_{(2)}^{{\mathrm{mhv } } } = 0\;.\end{aligned}\ ] ] the generalization to @xmath174 is straightforward by repetition of the argument .
it follows that the phase factor respects the vertex expansion @xmath175 we therefore have the @xmath0-deformed mhv vertex expansion @xmath176 { { \mathcal}{a}}_{(1)}^{\beta_r,\,{\mathrm{mhv}}}{{\mathcal}{a}}_{(2)}^{\beta_r,\,{\mathrm{mhv}}}\cdots { { \mathcal}{a}}_{(k)}^{\beta_r,\,{\mathrm{mhv}}}{{\mathcal}{a}}_{(k+1)}^{\beta_r,\,{\mathrm{mhv}}}\;.\end{aligned}\ ] ] interestingly , this result allows us to circumvent any non - mhv amplitude deformation calculation by multiplying together simpler mhv phase factors .
what is more , a complete set of four distinct grassmann variables is also chargeless , which potentially reduces the extent of nestedness even further .
moreover , as a passing remark we note that the derivations presented here also ensure validity of the phase - dressed super bcfw on - shell recursion relations , @xmath177 to expose the general pattern and keep tediousness to a minimal extent , we resort to the simplest case beyond @xmath32 level , namely the @xmath50 sector , calculate the phase factor explicitly and construct the @xmath0-deformed generating function .
the @xmath1 @xmath50 superamplitude depends on the grassmann object @xmath52 and is given by the compact expression @xmath178 combination of the two grassmann sums that appear in @xmath52 using @xmath179 yields @xmath180 with @xmath181 denoting the heaviside step function with the convention @xmath182 . in order to maintain a hygienic labeling scheme we split each of the displayed summation indices into four , indicated by a subscript following the flavor and @xmath21-symmetry index .
the @xmath50 phase factor can now either be derived by again acting with the phase operator , or preferably inferred from the result for the @xmath32 superamplitude .
we immediately obtain @xmath183 \bigg)\;.\end{aligned}\ ] ] before the phase factor is plugged back into the amplitude a slightly compressed notation is prepared .
we take @xmath52 and strip off the grassmann delta function to get @xmath184 such that @xmath185 .
furthermore we introduce the chiral spinor @xmath186 our expression for the @xmath0-deformed @xmath50 generating tree is thus @xmath187 where the sums expand in the obvious ways @xmath188 this formula completes our analysis at tree level .
in the following we will look at the general structure of supersymmetric sums in multi - loop unitarity cuts , which break up phase deformed loop amplitudes into products of tree amplitudes .
such cuts are optimal to work with instead of using lower - loop amplitudes in the construction , because the fully developed tree - level formalism is recycled .
we first specialize to plain @xmath1 theory .
schematically we are interested in performing the intermediate state sum , @xmath189 for each cut leg . within the superspace setup
this summation is rendered very elegantly using grassmann integration over the @xmath35-variables associated with the internal lines .
let us be a bit more specific and assume that the tree - level amplitudes are represented by the superamplitudes @xmath190 which are connected by @xmath191 on - shell propagators .
all possible internal and external particle configurations are then encoded in the supercut @xmath192 { { \mathcal}{a}}_{(1)}^{\mathrm{tree}}{{\mathcal}{a}}_{(2)}^{\mathrm{tree}}\cdots { { \mathcal}{a}}_{(m)}^{\mathrm{tree}}\;.\end{aligned}\ ] ] without loss of generality the @xmath193 superamplitudes can be assumed to be either of @xmath32 or @xmath33 type .
this is of course trivially justified if all tree - level amplitudes in the supercut have at most five legs . in more complicated situations where this is not the case
, the @xmath32 vertex expansion applies to reduce non - mhv parts to products of @xmath32 superamplitudes with additional propagators .
hence , we only have to consider supercuts of the form @xmath192 { { \mathcal}{a}}_{(1)}^{\mathrm{mhv}}\cdots{{\mathcal}{a}}_{(m')}^{\mathrm{mhv}}\hat{{{\mathcal}{a}}}_{(m'+1)}^{{\overline{\mathrm{mhv}}}}\cdots \hat{{{\mathcal}{a}}}_{(n)}^{{\overline{\mathrm{mhv}}}}\;,\end{aligned}\ ] ] with @xmath194 and @xmath195 @xmath32 and @xmath33 supertrees respectively . here ,
@xmath33 superamplitudes have been fourier transformed from tilded superspace to the @xmath35-coordinates .
let us now switch on the deformation .
all ingredients are at hand .
we exploit that we can attach a subphase to each supertree @xmath122 and derive the deformed supercut @xmath196 \bigg(\prod_i{\hat{{\mathcal}{p}}}_{(i)}^{\beta_r}\bigg ) { { \mathcal}{a}}_{(1)}^{\mathrm{mhv}}\cdots{{\mathcal}{a}}_{(m')}^{\mathrm{mhv}}{{\mathcal}{a}}_{(m'+1)}^{{\overline{\mathrm{mhv}}}}\cdots { { \mathcal}{a}}_{(n)}^{{\overline{\mathrm{mhv}}}}\nonumber \\ & = \int\bigg[\prod_{i=1}^k d^4\eta_i\bigg ] { { \mathcal}{a}}_{(1)}^{\beta_r,\;{\mathrm{mhv}}}\cdots{{\mathcal}{a}}_{(m')}^{\beta_r,\;{\mathrm{mhv } } } { { \mathcal}{a}}_{(m'+1)}^{\beta_r,\;{\overline{\mathrm{mhv}}}}\cdots { { \mathcal}{a}}_{(n)}^{\beta_r,\;{\overline{\mathrm{mhv}}}}\;.\end{aligned}\ ] ] of course , for planar graphs the phases combine and reproduce the tree - level result determined by the external legs . but
the deformed supercut provides rich information about non - planar diagrams which will differ substantially from the large-@xmath197 limit .
beautiful algebraic and graphical methods for evaluating such supersums were reported in @xcite .
these techniques rely merely on the grassmann structure of the amplitudes and are hence directly compatible with our results .
the reader is encouraged to also consult @xcite .
the remarkable properties of amplitudes in conformal deformations of @xmath1 super yang - mills theory with minimal or no supersymmetry suggest that neither the ordinary or dual representations of the superconformal algebras are really natural frameworks for discussing their symmetries . in the following we therefore propose a novel phase representation of the @xmath198 algebra .
we draw attention to @xcite for thorough treatments of superconformal and yangian symmetry .
let @xmath199 be any standard @xmath1 superconformal symmetry generator , i.e. take @xmath200 we can very intuitively apply a similarity transformation to obtain a representation that manifestly annihilates the phase - deformed superamplitudes . indeed , we can formally remove the phase , apply the ordinary symmetry generator and then reinsert the deformation . on the space of amplitudes
the phase generator has a perfectly well - defined and very simple inverse given by @xmath201}{\hat{{\mathcal}{q}}}_q^{[2]}- { \hat{{\mathcal}{q}}}_p^{[2]}{\hat{{\mathcal}{q}}}_q^{[1]}\right)\bigg]\;.\end{aligned}\ ] ] our phase representation is therefore @xmath202 the transformed operators trivially satisfy the correct commutator and anti - commutator relations of the @xmath1 superconformal algebra by construction , irrespective of @xmath203 $ ] being nonzero .
moreover , if @xmath204 is a symmetry of @xmath205 , that is @xmath206 , then @xmath207 in words , the deformed generating trees are manifestly annihilated by all @xmath198 generators in this phase representation .
what is more , we can also apply the transformation to level one generators and hence lift the @xmath198 symmetry algebra to a yangian realized on the deformed amplitudes .
our discussion thus suggests that all intrinsic properties of planar @xmath1 super yang - mills theory are preserved by the phase deformation .
we emphasize that our formalism extends almost effortlessly to the @xmath2-deformation , of which the @xmath0-deformation is actually a more frequently studied special case .
it is also conformally invariant at the quantum level in the planar approximation and can be defined analogously @xcite .
indeed , the @xmath2-deformation is generated by promoting ordinary products in the @xmath1 lagrangian to star products adjusted with phases which now break the @xmath20 @xmath21-symmetry to its cartan subgroup , which is a @xmath208 flavor symmetry of the resulting theory .
customarily the star product between superfields @xmath209 and @xmath210 is @xmath211 for some basis @xmath212 .
more practically to us though , the phases can equivalently be parametrized by four - component charges @xmath213}_a$ ] and @xmath214}_a$ ] subject only to tracelessness conditions assuming that all @xmath4 multiplets are charged under the flavor symmetry .
the number of independent parameters is thus three since the charges enter the star product antisymmetrically . on the other hand , if the vector multiplet is neutral , @xmath213}_4 = { \mathcal}u^{[2]}_4 = 0 $ ]
, we recover the one - parameter @xmath0-deformation .
conservation of charge under each symmetry again implies absense of phase contributions from internal structure to all orders in planar perturbation theory .
it is now trivial to write down the appropriate phase generator @xmath215}\hat{{\mathcal}u}_q^{[2]}- \hat{{\mathcal}u}_p^{[2]}\hat{{\mathcal}u}_q^{[1]}\bigg)\bigg]\end{aligned}\ ] ] with charge counting operators defined in on - shell superspace in the usual way .
the extracted parameter @xmath74 is a reminiscence of the @xmath0-deformation and should just be considered a fixed common constant of proportionality .
we can now easily derive all tree - level amplitudes , super vertex expansions , multi - loop unitarity cuts and so forth .
in other words , everything we have said about the @xmath0-deformation is compatible with the @xmath2-deformation .
in this paper we have investigated the perturbative regime of @xmath0-deformed super yang - mills theory using on - shell methods .
we have explicitly written all mhv and nmhv tree - level scattering amplitudes in terms of new generating functions and proven generalization to arbitrary particle and helicity configurations via the mhv vertex expansion .
our results have been obtained by implementation of a phase matrix in the @xmath1 superamplitudes , derived from their grassmann structure using a sector independent operator .
several component amplitudes were given as examples .
all generating trees are manifestly @xmath4 supersymmetric and reduce to the usual maximally supersymmetric expressions when the deformation is removed . however , we transformed the @xmath1 superconformal generators in on - shell superspace and uncovered a phase dependent representation that annihilates the deformed amplitudes . in this implementation ,
all symmetries exhibited by the @xmath1 amplitudes survive the deformation .
we finally set the stage for automated computation of intermediate state sums in connection with multi - loop unitarity cuts of non - planar amplitudes in both the @xmath0- and @xmath2-deformation .
applications of generating functions in this direction seem especially promising for providing further novel insight . 99 z. bern , j. j. m. carrasco , h. ita , h. johansson and r. roiban , phys
. rev .
d * 80 * , 065029 ( 2009 ) [ arxiv:0903.5348 [ hep - th ] ] .
m. bianchi , h. elvang and d. z. freedman , jhep * 0809 * , 063 ( 2008 ) [ arxiv:0805.0757 [ hep - th ] ] .
h. elvang , d. z. freedman and m. kiermaier , jhep * 0904 * , 009 ( 2009 ) [ arxiv:0808.1720 [ hep - th ] ] .
n. arkani - hamed , f. cachazo and j. kaplan , jhep * 1009 * , 016 ( 2010 ) [ arxiv:0808.1446 [ hep - th ] ] . h. elvang , y. t. huang , c. peng , y. t. huang and c. peng , jhep * 1109 * , 031 ( 2011 ) [ arxiv:1102.4843 [ hep - th ] ] .
m. sogaard , phys .
d * 84 * , 065011 ( 2011 ) [ arxiv:1106.3785 [ hep - th ] ] .
a. brandhuber , p. heslop and g. travaglini , phys .
d * 78 * , 125005 ( 2008 ) [ arxiv:0807.4097 [ hep - th ] ] .
j. m. drummond , j. henn , g. p. korchemsky and e. sokatchev , arxiv:0808.0491 [ hep - th ] .
j. m. drummond , j. henn , g. p. korchemsky and e. sokatchev , nucl .
b * 828 * , 317 ( 2010 ) [ arxiv:0807.1095 [ hep - th ] ] .
j. m. drummond and j. m. henn , jhep * 0904 * , 018 ( 2009 ) [ arxiv:0808.2475 [ hep - th ] ] .
j. m. drummond , j. m. henn and j. plefka , jhep * 0905 * , 046 ( 2009 ) [ arxiv:0902.2987 [ hep - th ] ] .
g. georgiou , e. w. n. glover and v. v. khoze , jhep * 0407 * , 048 ( 2004 ) [ arxiv : hep - th/0407027 ] .
y. t. huang , phys .
b * 631 * , 177 ( 2005 ) [ arxiv : hep - th/0507117 ] .
h. feng and y. t. huang , jhep * 0904 * , 047 ( 2009 ) [ arxiv : hep - th/0611164 ] .
h. elvang , d. z. freedman and m. kiermaier , jhep * 0906 * , 068 ( 2009 ) [ arxiv:0811.3624 [ hep - th ] ] .
s. lal and s. raju , phys .
d * 81 * , 105002 ( 2010 ) [ arxiv:0910.0930 [ hep - th ] ] .
r. britto , f. cachazo and b. feng , nucl .
b * 715 * , 499 ( 2005 ) [ arxiv : hep - th/0412308 ] .
r. britto , f. cachazo , b. feng and e. witten , phys .
lett . * 94 * , 181602 ( 2005 ) [ arxiv : hep - th/0501052 ] .
r. g. leigh and m. j. strassler , nucl .
b * 447 * , 95 ( 1995 ) [ arxiv : hep - th/9503121 ] .
v. v. khoze , jhep * 0602 * , 040 ( 2006 ) [ arxiv : hep - th/0512194 ] . t. mansson and k. zoubos , jhep * 1010 * , 043 ( 2010 ) [ arxiv:0811.3755 [ hep - th ] ] .
a. mauri , s. penati , a. santambrogio and d. zanon , jhep * 0511 * , 024 ( 2005 ) [ arxiv : hep - th/0507282 ] . o. lunin and j. m. maldacena , jhep * 0505 * , 033 ( 2005 ) [ arxiv : hep - th/0502086 ] .
s. ananth , s. kovacs , h. shimada , jhep * 0701 * , 046 ( 2007 ) .
[ hep - th/0609149 ] .
s. ananth , s. kovacs and h. shimada , nucl .
b * 783 * , 227 ( 2007 ) [ hep - th/0702020 [ hep - th ] ] .
f. elmetti , a. mauri , s. penati and a. santambrogio , jhep * 0701 * , 026 ( 2007 ) [ arxiv : hep - th/0606125 ]
. f. elmetti , a. mauri , s. penati , a. santambrogio and d. zanon , jhep * 0710 * , 102 ( 2007 ) [ arxiv:0705.1483 [ hep - th ] ] .
d. i. kazakov and l. v. bork , jhep * 0708 * , 071 ( 2007 ) [ arxiv:0706.4245 [ hep - th ] ] . j. n. ihry , jhep * 0804 * , 051 ( 2008 ) [ arxiv:0802.3644 [ hep - th ] ] . l. v. bork , d. i. kazakov , g. s. vartanov and a. v. zhiboedov , jhep * 0804 * , 003 ( 2008 ) [ arxiv:0712.4132 [ hep - th ] ] .
y. oz , s. theisen and s. yankielowicz , phys . lett .
b * 662 * , 297 ( 2008 ) [ arxiv:0712.3491 [ hep - th ] ] .
f. fiamberti , a. santambrogio and c. sieg , arxiv:1006.3475 [ hep - th ] .
k. madhu and s. govindarajan , jhep * 0705 * , 038 ( 2007 ) [ arxiv : hep - th/0703020 ]
. m. kulaxizi , arxiv : hep - th/0612160 .
d. bundzik , jhep * 0704 * , 035 ( 2007 ) [ arxiv : hep - th/0608215 ] .
t. bargheer , n. beisert and f. loebbert , j. phys . a a * 44 * , 454012 ( 2011 ) [ arxiv:1104.0700 [ hep - th ] ] .
t. bargheer , n. beisert , w. galleas , f. loebbert and t. mcloughlin , jhep * 0911 * , 056 ( 2009 ) [ arxiv:0905.3738 [ hep - th ] ] .
q. jin and r. roiban , arxiv:1201.5012 [ hep - th ] .
s. j. gates , m. t. grisaru , m. rocek and w. siegel , front .
phys . * 58 * , 1 ( 1983 ) [ arxiv : hep - th/0108200 ] . | we present tree - level scattering amplitudes in @xmath0-deformed super yang - mills theory in terms of new generating functions , derived by construction of a phase operator and application thereof to the @xmath1 superamplitudes .
the technique is explicitly illustrated for the mhv and nmhv sectors .
along these lines we propose a phase representation of the @xmath1 superconformal algebra realized on deformed amplitudes in the planar limit .
validity of the mhv vertex expansion is proven and a connection to non - planar multi - loop unitarity cuts is established .
our derivations are also compatible with the related @xmath2-deformation . |
10 challet d and zhang y - c 1997 _ physica a _ * 246 * 407 marsili m et al 2000 _ physica a _ * 280 * 522 de martino a and marsili m 2001 _ j. phys . a ( math . gen . ) _ * 34 * 2525 challet d et al 2000 _ phys . rev . lett .
_ * 84 * 1824 heimel j a f and coolen a c c 2001 _ phys .
e _ * 63 * 56121 cugliandolo l f and kurchan j 1995 _ phil . mag .
b _ * 71 * 501 coolen a c c and heimel j a f 2001 in preparation bouchaud j - p et al 1998 in _ spin glasses and random fields _
( young a p , ed . )
world scientific ( singapore , 1998 ) marsili m _ private communication _
while investigating the possible relation between our mo line and replica - symmetry breaking , it became apparent that in the very final step of the at - line calculation in @xcite a small error has occurred .
it was found that the replica - symmetric solution becomes unstable when @xmath130 where @xmath131 and @xmath132 , with @xmath133 .
the brackets @xmath134 denote a gaussian average over @xmath58 having zero mean and variance @xmath135 , @xmath136 being the overlap between two different replicas ( off - diagonal overlap matrix element ) .
we have absorbed a spurious factor @xmath137 in @xmath138 .
if we now define @xmath139 , the at line can be written as @xmath140 . by laplace
s method we find @xmath141 , @xmath142 being the minimum of @xmath143 in @xmath144 $ ] . for @xmath145 , @xmath142 lies inside this interval and @xmath146 , while for @xmath147 @xmath142 is on the border and @xmath148 .
this gives second derivatives that are @xmath149 and @xmath150 , respectively .
the at - line is therefore given by @xmath151 .
recognizing the non - vanishing term on the l.h.s .
as the fraction @xmath152 of fickle agents , we find @xmath153 ^ 2={\overline{\phi}}\ ] ] similar to the result of @xcite where in place of @xmath152 a @xmath154 was reported .
written in terms of @xmath90 , the at line is identical to the mo line ( [ mo ] ) .
we learned that the at - line can also be derived from the dynamical stability of eqs .
( [ ddd])@xcite . | we study the dynamics of the ` batch ' minority game with market - impact correction using generating functional techniques to carry out the quenched disorder average . we find that the assumption of weak long - term memory , which one usually makes in order to calculate ergodic stationary states , breaks down when the persistent autocorrelation becomes larger than @xmath0 .
we show that this condition , remarkably , coincides with the at - line found in an earlier static calculation .
this result suggests a new scenario for ergodicity breaking in disordered systems . the minority game ( mg ) models a market of speculators interacting through a simple supply - and - demand mechanism @xcite .
one of the key behavioural assumptions of the original model is that agents act as so - called price - takers , meaning that at every stage of the game each of them only perceives the aggregate action of all agents , i.e. the total bid .
recently , in @xcite , a generalization has been introduced in which agents are able to estimate their own contribution to the total bid and use this additional information to adjust their learning dynamics and optimize their performance .
the statics of this model has been tackled by spin - glass techniques in @xcite along the lines of @xcite .
the system was found to approximately minimize a disordered hamiltonian @xmath1 whose minima could be calculated with the replica method .
it was shown that replica - symmetry breaking ( rsb ) can occur , implying the existence of multiple stationary states . in this letter
we adapt the dynamical method used in @xcite to analyze the ` batch ' version of this model . assuming time - translation invariance , finite integrated response and weak long - term memory @xcite we obtain exact results for the stationary state which are in excellent agreement with computer experiments and with earlier static approaches .
moreover , we derive a condition for the continuous onset of memory , where the assumption of weak long - term memory is found to fail while time - translation invariance still holds .
this appears to be different from the usual aging scenario in non - ergodic disordered systems .
remarkably , the memory - onset condition coincides with the at line found in statics .
we begin by recalling the definition of the model .
we consider @xmath2 agents labeled by roman indices . at each iteration
round @xmath3 all agents receive the same information pattern @xmath4 drawn at random with uniform probability from @xmath5 .
each agent has at his disposal @xmath6 different strategies ( labeled by @xmath7 ) to convert the acquired information into a trading decision .
strategies are denoted by @xmath8-dimensional vectors : @xmath9 , where @xmath10 is the trading action ( e.g. @xmath11 for ` buy ' , @xmath12 for ` sell ' ) prescribed to agent @xmath13 by his @xmath14-th strategy given receipt of information @xmath15 . by assumption ,
each component @xmath10 is selected randomly and independently from @xmath16 with uniform probabilities before the start of the game , for all @xmath13 , @xmath14 and @xmath15 .
this introduces quenched disorder into the model .
each strategy of every agent is given an initial valuation @xmath17 , which is updated at the end of every round . at the start of round @xmath3 ,
given @xmath4 , every agent selects the strategy with the highest valuation , @xmath18 , and subsequently makes a bid according to the trading decision set by the selected strategy : @xmath19 .
the total bid at round @xmath3 is defined as @xmath20 .
finally , for all @xmath13 and @xmath14 all payoffs are updated according to a reinforcement learning dynamics of the form @xmath21\ ] ] and agents move to the next round .
the first term in square brackets embodies the minority rule , in that the valuation of a strategy is increased every time it predicts the correct minority action , independently of it having been actually used .
the term proportional to @xmath22 adjusts the total bid for the possibility that agent @xmath13 is not using strategy @xmath23 . for @xmath24 one
returns to the original mg , while for @xmath25 the total bid is completely adjusted .
we focus on the case @xmath26 . introducing the variables @xmath27/2
$ ] , as well as the @xmath28-dimensional vectors @xmath29 , @xmath30 and @xmath31 , and defining @xmath32 $ ] one has @xmath33\ ] ] following @xcite we study in this paper a ` batch ' version of the model , which is obtained by averaging ( [ ddd ] ) over information patterns : @xmath34 where @xmath35 is a re - scaled time , @xmath36 and @xmath37 .
the external field @xmath38 has been added for later use .
in contrast to the more usual ` on - line ' model ( [ ddd ] ) , where the @xmath39 s are updated after every iteration step , in the ` batch ' case the updates are made on the basis of the average effect of all possible choices of @xmath15 .
this modified dynamics yields results for the stationary state which are quantitatively very similar to those of the original model @xcite .
the theoretical advantage of the ` batch ' formulation is that it circumvents the difficulty of constructing a proper continuous time limit .
the numerical advantage is that one can simulate larger systems for a longer time .
following @xcite , one derives the effective non - linear single - agent equation @xmath40 where @xmath41 $ ] and @xmath42 is a gaussian noise with zero mean and temporal correlations given by @xmath43 the matrices @xmath44 and @xmath45 appearing here are the noise - averaged single - agent correlation and response functions for the process ( [ dyna ] ) , with elements @xmath46 respectively , while @xmath47 is the identity matrix and @xmath48 denotes the matrix with all entries equal to one .
the link between the markovian multi - agent system ( [ ddd ] ) and the non - markovian single - agent process ( [ dyna ] ) is established by the fact that , for @xmath49 , @xmath50 and @xmath51 become identical to the disorder- and agent - averaged correlation and response functions of ( [ ddd ] ) : @xmath52_{{\rm dis } } \quad{\rm and}\quad g_{tt'}=\frac{1}{n}\sum_{i=1}^n{\left}[\frac{\partial s_i(t)}{\partial\theta_i(t')}{\right}]_{{\rm dis}}\ ] ] eqs ( [ dyna]-[gc ] ) describe the dynamics of the system exactly in the @xmath49 limit .
we now move to the stationary states of ( [ dyna ] ) upon making the following assumptions : [ cols= " < , < " , ] for the re - scaled quantity @xmath53 one finds @xmath54 where @xmath55 $ ] and @xmath56 , while @xmath57 is a static field .
the variance of the zero - average gaussian random variable @xmath58 can be calculated from ( [ cortemp ] ) , yielding @xmath59 with the persistent correlation @xmath60 .
the effective agent is ` frozen ' if @xmath61 , so that @xmath62 and he is always employing the same strategy . setting @xmath63 ,
this is easily seen to be the case if @xmath64 with @xmath65 $ ] , provided @xmath66 .
he is instead fickle when @xmath67 or @xmath68 , and in this case @xmath69 .
a self - consistent equation for @xmath70 can now be derived by separating the contribution of the frozen agents from that of the fickle ones . upon defining @xmath71 one
finds @xmath72\ ] ] where @xmath73 is the step function , @xmath74 is the fraction of fickle agents , and @xmath75 is the fraction of frozen agents . for @xmath76
one obtains @xmath77 equations ( [ var]-[chi ] ) form a closed set from which one can solve for @xmath78 , @xmath70 and @xmath79 for any @xmath80 and @xmath22 .
results for @xmath70 are shown in figure 1 .
for negative @xmath22 , one observes an excellent agreement between theory and experiment for all values of @xmath80 , implying that none of our assumptions is ever violated . when @xmath24 , we recover the results of @xcite , which match the simulations perfectly for @xmath80 larger than the critical value @xmath81 . at this point
the integrated response @xmath79 diverges ( fir is violated ) and a transition to a highly non - ergodic regime takes place , where the stationary state depends on the initial conditions @xmath82 .
starting with @xmath83 leads to a high volatility state , while starting with @xmath84 leads to relatively low volatility .
the latter regime can be solved using the assumption that @xmath79 remains very large for all @xmath85 .
in fact , if @xmath86 then @xmath87 so that @xmath88 , which is equivalent to @xmath89 . solving this for @xmath90 and inserting the resulting value in ( [ c ] ) , we obtain the top left branch of the @xmath24 curve in fig .
1 , which is again in excellent agreement with numerical results . for positive @xmath22
, one sees that when @xmath91 our theoretical predictions deviate from the experimental observations , whereas the agreement is perfect for @xmath92 .
finding no violation of fir , we have to conclude that either tti or wltm is violated .
however , we have found no evidence of aging .
therefore we expect the deviations to be related to the breakdown of wltm only . to find the onset of memory , we split @xmath51 in its tti part and its non - tti part : @xmath93 during the initial stages of the game , small perturbations can cause some agents , which would otherwise have remained fickle , to freeze and vice versa , thus creating a persistent part @xmath94 in the response function . as the agents freeze
, their state ( and consequently their contribution to @xmath45 ) becomes independent of @xmath35 , so that we expect @xmath95 .
after an initial equilibration period , for all frozen agents the difference between the strategy valuations have become very large , so they are virtually insensitive to perturbations . hence we must assume that @xmath96 .
the fickle agents , however , remain sensitive to small perturbations .
the effects will wear out over time ( finite response ) and are given by @xmath97 .
assuming @xmath94 is small , we expand @xmath98 in powers of @xmath94 up to first order : @xmath99 defining @xmath100 and @xmath101 , one then finds asymptotically @xmath102\!(t',t'')s(t'')\end{aligned}\ ] ] using the rectified linear function @xmath103 for @xmath104 and @xmath105 otherwise , we see that if @xmath106 then @xmath107\!(t',t'')s(t '' ) { \right}]{\right } ) , \\ \widetilde{\gamma}=\sqrt{\alpha}{\left}(\frac{1}{1+\widetilde{\chi}}-\eta{\right } ) \nonumber\end{aligned}\ ] ] as before , we have @xmath108 , whereas @xmath109\!(t',t '' ) \widetilde{g}(t'',t)\ ] ] up to first order in @xmath94 one finds @xmath110+\or(\widehat{g}^2)$ ] .
although @xmath111 is always a solution of this equation , a bifurcation occurs when @xmath112=1 $ ] , which is equivalent to @xmath113 ^ 2 $ ] , and can be written in terms of @xmath90 as @xmath114={\overline{\phi(\lambda)}}\ ] ] we call this line in the @xmath115 plane the memory - onset ( mo ) line , see figure 2 .
it coincides remarkably with the at - line ( see appendix ) , and implies that the bifurcation occurs at @xmath116 for @xmath117 . above this value
, wltm can be broken , and indeed one sees from figure 1 that numerical results deviate from our theoretical predictions for @xmath118 . to give further evidence of memory , we have analyzed the time evolution of two identical copies @xmath119 and @xmath120 of the system , starting from slightly different initial conditions .
we plotted in figure 3 the distance @xmath121 of the stationary states , given by @xmath122 , where @xmath123 is the long - time average of @xmath124 ( @xmath125 ) , versus the persistent autocorrelation of copy @xmath119 , @xmath126 .
as @xmath126 approaches @xmath127 , the two copies end up in different stationary states , proving that they remember initial conditions . at the same time , if a perturbation is applied much later during the run the copies end up in the same stationary state , indicating that indeed @xmath128 as @xmath129 . summarizing , we have shown that in this model the usual connection between broken ergodicity and broken tti ( aging ) , as seen for instance in mean - field spin glasses @xcite , does not occur .
in contrast , we derive from the dynamics a condition for breakdown of wltm and continuous onset of memory within the tti regime , which is found to be equivalent to the at - line found in the static approach .
this remarkable deviation from the well - known rsb / aging picture is possibly due to the fact that the microscopic dynamics of our model does not satisfy detailed balance .
we gratefully acknowledge support from and useful discussions with a c c coolen and m marsili , and with s franz .
we also thank sissa and king s college london for reciprocal hospitality .
this work originated at the international seminar on statistical mechanics of information processing in cooperative systems ( dresden , march 2001 ) . |
following the @xmath0-dmr detection of anisotropy in the cosmic microwave background on large angular scales ( smoot et al . 1992 ; wright et al . 1992 ; bennett et al .
1996 ; grski et al . 1996 ) , there have been many measurements of cmb anisotropy on angular scales down to @xmath9 ( ganga et al .
1994 ; gutirrez et al . 1997 ; piccirillo et al . 1997 ; netterfield et al .
1997 ; gundersen et al . 1995 ; tucker et al .
1997 ; platt et al .
1997 ; masi et al . 1996
; lim et al . 1996 ; cheng et al . 1997 ; griffin et al . 1997 ; scott et al . 1996 ; leitch et al . 1997 ; church et al . 1997 , see page 1997 for a review ) .
in addition to the dmr experiment , cmb anisotropy maps have been produced from four other experiments ( ganga et al .
1994 ; white & bunn 1995 ; scott et al . 1996
; tegmark et al .
cmb anisotropy observations are beginning to test cosmogonical models and provide interesting constraints on cosmological parameters in these models ( e.g. , bunn & sugiyama 1995 ; grski et al . 1995 , 1998 ; stompor , grski , & banday 1995 ; ganga , ratra , & sugiyama 1996 ; ganga et al .
1997a , b , 1998 ; hancock et al .
1997 ; bond & jaffe 1997 ; stompor 1997 ; lineweaver & barbosa 1997 ; ratra et al .
however , definitive results , including significant constraints on cosmological parameters , must await new cmb anisotropy data , covering a large fraction of the sky ( to minimize sample variance ) and acquired at a variety of frequencies ( to allow for control over possible non - cmb anisotropy foreground contamination ) .
of the currently available data sets , the dmr maps best meet these criteria .
plans are afoot to launch two second - generation cmb anisotropy satellite missions the microwave anisotropy probe ( map ) is expected to have data in less than four years , and the planck surveyor is expected to have data in less than a decade .
these missions are expected to measure the cmb anisotropy on angular scales larger than a fraction of a degree , at a variety of frequencies .
map will probe the anisotropy at frequencies of 22 , 30 , 40 , 60 , and 90 ghz at expected angular resolutions of 0.93 , 0.68 , 0.47 , 0.35 , and 0.21 degrees ( fwhm beamwidths ) , respectively . at 90 ghz
the map sensitivity is expected to be @xmath10k per @xmath11 pixel .
planck surveyor is expected to measure the cmb anisotropy at frequencies ranging from 30 to 900 ghz with angular resolution ranging from 0.5 to 0.075 degrees .
it is now possible to make accurate predictions for the cmb anisotropy , as a function of cosmological parameters , in many different cosmological models . in conjunction with such theoretically computed model cmb anisotropy spectra
, the data from these experiments should lead to an accurate determination of cosmological parameters , such as the mass density parameter @xmath12 , the baryonic - mass density parameter @xmath13 , the cosmological constant @xmath14 , the hubble parameter @xmath15 , the matter power spectrum , and the optical depth to the last scattering surface ( e.g. , spergel 1994 ; jungman et al .
1996 ; bond , efstathiou , & tegmark 1997 ; zaldarriaga , spergel , & seljak 1997 ; dodelson , kinney , & kolb 1997 ) . at the least
, the data from these space missions should in principle be able to rule out most cosmological models currently under consideration .
the present situation in cosmology where one is able to make definite theoretical predictions which can be judged by near - future observations is quite unusual .
prior to the @xmath0-dmr measurement of the large - scale cmb anisotropy , gott et al .
( 1990 ) had generated simulated all - sky cmb anisotropy maps predicted in the spatially - flat @xmath16 fiducial cdm model with gaussian , adiabatic , scale - invariant matter fluctuations .
the dmr data is indeed consistent with such a cmb anisotropy spectrum ( grski et al .
1996 , 1998 ) .
in this paper we generate simulated all - sky cmb anisotropy maps appropriate for map in several representative cosmological models normalized to the dmr data .
our predictions can be directly compared to the map data when it is available .
hinshaw , bennett , & kogut ( 1995 ) have generated all - sky cmb anisotropy maps expected in an @xmath2 cdm model , at angular resolutions of @xmath17 and @xmath18 , to study the effects of sky coverage , angular resolution , instrumental noise , and unresolved features in the data .
novikov & jrgensen ( 1996 ) have simulated @xmath19 sky anisotropy maps of @xmath2 cdm models with @xmath20 and 0.03 at @xmath17@xmath18 resolutions .
they have examined various statistical properties of the temperature fluctuation field , such as clustering of peaks and percolation .
we focus here on generating mock cmb anisotropy maps appropriate for the specific experimental configuration of map .
we also use these cmb anisotropy maps to examine the ability of the expected map data to discriminate between different cosmological models . in what follows
we assume that the map data will be limited by statistical noise rather than foregrounds or systematic errors .
in section 2 we present the cosmological models we study , and explain how the simulated cmb anisotropy maps are generated . in section 3 we compute the cmb anisotropy correlation function for each of the simulated maps , and show that it is sensitive to the acoustic peaks in the cmb anisotropy angular power spectrum . in section 4
we study the genus statistic of the generated maps , and explore its ability to discriminate between cosmological models and to detect non - gaussian fluctuations in the cmb anisotropy field .
the effect of instrumental noise is considered .
conclusions follow in section 5 .
to generate simulated cmb anisotropy maps for map we have chosen four representative gaussian , adiabatic cosmogonical models , with distinct cmb anisotropy angular power spectra : ( 1 ) the fiducial cdm model ( fcdm ) with @xmath2 , the hubble parameter @xmath15 divided by 100 km / sec / mpc @xmath21 , and with no reionization .
( 2 ) a spatially - flat , low - density cdm model ( @xmath22cdm ) with a cosmological constant @xmath22 , @xmath1 , and @xmath23 . ( 3 ) a spatially - open low - density cdm model ( ocdm ) with no @xmath22 , @xmath1 , and @xmath24 . ( 4 ) a reionized fiducial cdm model ( rcdm ) with @xmath16 , @xmath21 , and fully reionized at redshift @xmath25 ( i.e. , with electron ionization fraction @xmath26 at @xmath27 and so optical depth @xmath28 ) . for all models ,
the baryonic - mass density parameter is @xmath29 .
further details are given in ratra et al .
( 1997a ) and sugiyama ( 1995 ) .
the fcdm and @xmath22cdm models considered here assume a scale - invariant primordial energy - density power spectrum ( harrison 1970 ; peebles & yu 1970 ; zeldovich 1972 ) , as is predicted in the simplest spatially - flat inflation models ( guth 1981 , also see kazanas 1980 ; sato 1981 ) .
the fcdm model is inconsistent with galaxy clustering data ( e.g. , maddox et al .
1990 ; saunders et al .
1991 ; vogeley et al . 1992 ; park et al .
1994 ) , but still serves as a useful fiducial for cosmological model comparison .
@xmath22cdm models ( peebles 1984 ) became popular when it was discovered that the large - scale galaxy distribution was of the form expected in cdm models with @xmath30 ( e.g. , park 1990 , 1991 ; vogeley et al .
1992 ; da costa et al .
1994 ; peacock & dodds 1994 ) , since these models could be accommodated in the then - dominant spatially - flat inflation scenario .
however , an open inflating universe can also be created as a single bubble during the decay of an initial metastable inflation state ( gott 1982 , 1986 ) .
the ocdm model considered here assumes the primordial energy - density power spectrum of the simplest open - bubble inflation models ( ratra & peebles 1994 , 1995 ; bucher , goldhaber , & turok 1995 ; yamamoto , sasaki , & tanaka 1995 ) .
the low - density models , ocdm and @xmath22cdm , are consistent with most current large - scale structure and cmb anisotropy observations ( e.g. , gott 1997 ; turner 1997 ; cole et al .
1997 ; gardner et al .
1997 ; croft et al .
1997 ; coles et al . 1997 ; jenkins et al .
1997 ) .
a cmb temperature fluctuation field on the surface of a sphere ( i.e. , the sky ) can be decomposed into a sum of spherical harmonics , @xmath31 , where @xmath32 are the angular coordinates on the sky .
if the initial matter fluctuation is a gaussian random field , as predicted in simple inflation models , the expansion coefficients @xmath33 are independent gaussian random variables with zero means and variances @xmath34 , where @xmath35 is the cmb anisotropy angular power spectrum ( see below for more details ) .
the cmb anisotropy angular power spectra of the four models considered here are shown in figure 1 ( ratra et al . 1997a ; sugiyama 1995 ) .
following grski et al .
( 1998 ) and stompor ( 1997 ) , the angular power spectra are normalized to the arithmetic mean of the @xmath362-@xmath37 @xmath0-dmr 53 and 90 ghz data from the two extreme data sets : ( 1 ) galactic - frame maps accounting for the high - latitude galactic emission correction and including the @xmath38 moment in the analysis ; and ( 2 ) ecliptic - frame maps ignoring the high - latitude galactic emission correction and excluding the @xmath38 moment from the analysis . at large angular scales
the cmb anisotropy angular power spectra in the four models are of approximately the same amplitude . at smaller angular scales
acoustic oscillations due to the pressure of the photon - baryon fluid and matter velocity perturbations at photon decoupling cause the angular power spectra to rise . at still smaller angular scales photon diffusion and the finite thickness of the last - scattering surface smears out the fluctuations .
these effects give rise to the characteristic acoustic peaks in the angular power spectra on intermediate angular scales . in an open model the geometrical effect that the circumference of a circle is larger than @xmath39 times its radius moves these peaks to smaller angular scales relative to the spatially - flat case ( kamionkowski , spergel , & sugiyama 1994 ) . in the spatially - flat @xmath16 fcdm and low - density @xmath22cdm models ,
the first acoustic peak is at @xmath40 and @xmath41 , respectively .
the amplitudes of the peaks in these models are different . in the open case
the first peak is at @xmath42 , while in the reionized rcdm model the peaks are smeared out .
consequently , the presence , location , and amplitude of the acoustic peaks are a valuable probe of the geometry and content of the universe . in real observations and analyses ,
the temperature fluctuations on the sky are convolved with the beam pattern of the experiment , and are often further smoothed to increase the signal - to - noise ratio .
if the beam pattern is @xmath43 and the smoothing filter is @xmath44 , the observed cmb anisotropy temperature field is @xmath45 where @xmath46 and @xmath47 are the coefficients of the beam pattern and the smoothing filter expanded in legendre polynomials , and we consider only harmonics higher than the dipole . assuming the idealized case of an all - sky data map , the inverse relation for the harmonic amplitude is @xmath48 for a gaussian beam ( or a smoothing filter ) with a width @xmath49 to a very good approximation , @xmath50 we use equation ( 1 ) to simulate the cmb anisotropy maps for map .
we limit the summation to @xmath51 . for each cosmogony
the real and imaginary components of @xmath33 are randomly sampled gaussian variates with zero means , and variances equal to @xmath52 for the @xmath53 harmonics and equal to @xmath35 for the @xmath54 harmonics .
we adopt a gaussian beam with fwhm @xmath55 ( or width @xmath56 in equation [ 3 ] ) .
the temperature fluctuations are then computed at @xmath57 points in ( @xmath58 ) space on a grid with @xmath59 spacing .
the data from map is expected to include instrumental noise of @xmath60 @xmath61k per @xmath11 pixel .
if the noise is white , its power spectrum is constant @xmath62 .
consider a pure noise map convolved with a top hat filter @xmath63 with area of 0.09 square degrees .
then the variance of the noise temperature fluctuation , which is set to @xmath64k@xmath65 per 0.09 square degrees , is given by equation ( 5 ) below ( at @xmath66 with @xmath35 replaced by @xmath67 ) .
the top hat filter is @xmath68 for @xmath69 , and @xmath70 otherwise .
the coefficients of its expansion into legendre polynomials are , for @xmath71 , @xmath72^{2n}.\ ] ] we obtain from equation ( 5 ) @xmath73 for map .
one can study the effects of noise on statistics using this noise power spectrum . in practice , of course
, the map results will include an estimate of the noise at each pixel derived from the many observations taken as the satellite s horns sweep out complicated paths on the sky .
the satellite horns are paired , so subtracting their signals gives the noise map ; this allows an extremely accurate characterization of the noise as a function of position on the sky .
such a detailed noise map will allow a much more careful accounting of instrumental noise .
( for example , the noise at the ecliptic poles is expected to be about half of that at the ecliptic plane ) .
the maps resulting from the @xmath22cdm and ocdm models with simulated ( constant over the sky ) instrumental noise , projected according to the mollweide projection method ( snyder 1993 ) , are shown in figure 2 ( plate 1 ) .
these maps are smoothed with a @xmath74 fwhm gaussian , which makes the total smoothing scale of the signal @xmath5 .
the mollweide equal area projection shows the whole sky in a single map , but structures are distorted . in the topology study
we use conformal stereographic projection maps where shapes of structures are preserved locally . stereographic maps of the cmb fluctuations simulated for the ocdm and @xmath22cdm models and convolved with a @xmath75 fwhm gaussian ( which makes the total smoothing scale of the signal @xmath76 ) are shown in figure 3 ( plate 2 ) . at this smoothing length
most structures in the map generated by map will correspond to actual cmb anisotropy fluctuations ( see figure 8) .
it is to be noted that the ocdm and @xmath22cdm maps are visually very similar this is a consequence of the expected instrumental noise and smoothing ( see figures 1@xmath77 and 1@xmath78 ) .
we emphasize , however , that this does not mean that map can not statistically distinguish between these models .
differences between cosmological models are often displayed in terms of the angular power spectrum , as in figure 1 .
most models have relatively small differences at large scales ( @xmath79 ) , but they often have relatively large differences at small scales ( @xmath80 ) .
the structure of the acoustic peaks at smaller angular scales depends on the cosmogony and the cosmological parameters . hence , to test cosmogonies and measure these parameters , it is essential to have an accurate measurement of the location and amplitude of the acoustic peaks .
note that the acoustic peaks cause an oscillatory wave in the power spectrum ( see figure 1 ) .
furthermore , the amplitudes of individual harmonics in the power spectrum are destined to fluctuate with respect to the mean . to measure precisely the amplitude and location of the acoustic peaks in the power spectrum
one must make a fit over many harmonics . an important integral of the power spectrum is the two - point auto - correlation function , the real - space fourier counterpart of the power spectrum .
since the hot and cold spots due to the acoustic peaks in the power spectrum are compact in real space ( their size is fixed by the acoustic hubble radius at decoupling see figures 2 and 3 ) , it is useful to explore them in real space .
( note that in fourier space the peaks are well - approximated by a wavy sinc - function . ) while the correlation function and the power spectrum are simply related , it is useful to analyze data both in real space and in fourier space ( e.g. , hinshaw et al .
1996 ; grski et al . 1996 ; wright et al .
1996 ) .
the correlation function of cmb temperature fluctuations , corresponding to equation ( 1 ) , is @xmath81 here @xmath82 are legendre polynomials and we have ignored the monopole and dipole harmonics . the power spectra shown in figure 1@xmath83 can be approximated by an exponential minus a sinc - function .
for example , the acoustic peaks of the fcdm power spectrum can be approximately described by the function @xmath84 .
consequently , these acoustic peaks should appear as a negative top hat in the correlation function .
figure 4 shows the correlation functions of the four models considered .
thick curves are the correlation functions transformed from the power spectrum curves in figure 1 .
thin solid curves , tracing the thick ones at large angular separations , account for the convolution with a @xmath85 gaussian beam in figure 4@xmath83 and with additional @xmath86 gaussian smoothing in figure 4@xmath77 . the acoustic peaks in the power spectrum indeed appear as a valley in this plot .
the position , depth , and width of the acoustic valley depend on the model and the cosmological parameters .
we have generated seven sets of @xmath87 s for the fcdm model , and have computed their correlation functions .
these are also plotted , using dots , in figure 4 . at separations
@xmath88 cosmic variance causes a large scatter amongst the seven correlation functions .
hence it is very difficult to discriminate between models using large angular scale data alone .
however , direct measurement from these realizations shows that the fcdm acoustic valley in the function @xmath89 is consistently located at @xmath90 , comparing excellently with the theoretically predicted fcdm acoustic valley location at @xmath91 for a @xmath85 beam ( and at @xmath92 in the unsmoothed map ) .
the ocdm model acoustic valley lies at a smaller separation of @xmath93 ( @xmath94 in the unsmoothed map ) clearly distinguishable from the acoustic valleys of the fcdm and @xmath22cdm models ( some 13-@xmath37 difference for the fixed cosmological parameter values used here ) .
on the other hand , the spatially - flat fcdm and @xmath22cdm models have acoustic peaks at nearly the same angular scales .
therefore , they have to be differentiated by the amplitude of the power spectrum or the correlation function .
it can be noted from figure 4 that the correlation functions of the fcdm and @xmath22cdm models are statistically significantly different at separations @xmath95 . in the case of an ideal sample with whole - sky coverage and with no noise , the map data would distinguish the correlation function of the fcdm universe from that of the @xmath22cdm case at the 2.6-@xmath37 level at angular separations near @xmath96 .
note that this ignores the relative uncertainty between the dmr normalizations of the fcdm and @xmath22cdm models .
the numerical value of this relative uncertainty is not known , however , the absolute model - dependent uncertainty in the dmr normalization is @xmath97@xmath98 ( in @xmath99 , the quadrupole amplitude of the cmb anisotropy , grski et al .
also note that , for this comparison between fcdm and @xmath22cdm , cosmological parameters such as @xmath100 and @xmath13 are held fixed . to estimate the effects of instrumental noise on the correlation function analysis of the map data we use the noise power spectrum , @xmath101 , derived above .
this noise power is added to the cmb anisotropy signal , and the map is then further smoothed with a @xmath74 fwhm gaussian .
that is , we generate a set of @xmath102 up to @xmath103 where @xmath104 is a gaussian random variable with zero mean and @xmath105 , and @xmath46 and @xmath47 represent the @xmath106 beam and @xmath74 smoothing filter , respectively .
the corresponding correlation function is then computed . from seven realizations of the noise - added correlation functions for fcdm and ocdm
the positions of the acoustic valleys are measured to be at @xmath107 and @xmath108 , respectively .
the correlation function acoustic valleys of the fcdm and ocdm models are now 9-@xmath37 apart .
it is also noted ( from figure 4 ) that the statistical uncertainty in the amplitude of correlation function at separations @xmath109 is hardly affected by the instrumental noise expected for map .
this means that cosmic variance , rather than instrumental noise , dominates the statistical uncertainty in the correlation function at these scales .
( this is also true for the genus statistic see table 1 . ) therefore , the above argument on discrimination between the spatially - flat models on the basis of the amplitude of the correlation function remains essentially unchanged .
the power spectrum and correlation function measured from the observed cmb anisotropy maps provides important information on the nature of density fluctuations in the early universe , thus constraining structure formation mechanisms .
another important statistical characteristic of a cmb anisotropy map is whether or not its temperature fluctuation field is gaussian random - phase .
one can test for gaussianity of the temperature distribution by constructing histograms of the pixel values . to test the gaussianity and phase properties of the maps the genus topology test is useful . in this section
we study the topology of the cmb anisotropy isotemperature contour surfaces .
this provides information on the global structure of the density fluctuation field .
the genus has been used as a measure of topology of the matter / galaxy distribution . for three - dimensional topology
see gott , melott , & dickinson ( 1986 ) , gott , weinberg , & melott ( 1987 ) , vogeley et al .
( 1994 ) , and cole et al .
( 1997 ) ; for two - dimensional topology see melott et al .
( 1989 ) , park et al .
( 1992 ) , and colley ( 1997 ) .
the genus has also been used as a measure of the topology of cmb anisotropy fields ( coles 1988 ; park & gott 1988 ; gott et al .
1990 ; colley , gott , & park 1996 ; kogut et al . 1996 ; schmalzing & grski 1997 ; winitzki & kosowsky 1997 ) . for the two - dimensional cmb anisotropy temperature field ,
the genus is the number of hot spots minus the number of cold spots , i.e. , at a temperature threshold level @xmath110 the genus is @xmath111 where @xmath112 is the signed curvature of the isotemperature contours @xmath113 .
the genus curve as a function of the temperature threshold level has a characteristic shape for a gaussian random - phase field . changing the shape of the cmb anisotropy power spectrum only changes the amplitude of the genus curve .
hence the genus curve provides a good measure of both non - random phases and the shape of the power spectrum .
the above real - space definition of the genus allows it to be directly used for studying the statistics of the acoustic peak hot and cold spots .
previous measurements of the cmb anisotropy genus ( see references above ) were consistent with the gaussian random - phase hypothesis .
the theoretical genus per steradian of a two dimensional gaussian field with correlation function @xmath114 is ( gott et al . 1990 ) @xmath115 where @xmath116 , and the threshold temperature fluctuation is @xmath117 . for a cmb anisotropy temperature map convolved with beam @xmath118 and smoothing filter @xmath63 , @xmath119 and @xmath120
are given by @xmath121 where we have used the relation @xmath122 ( see equation [ 5 ] ) .
the genus per steradian is then @xmath123 we note again that the genus curve shape is fixed by the gaussian random - phase nature of the field , and its amplitude depends only on the shape of the cmb anisotropy power spectrum and not on its amplitude . for scale - free energy - density perturbations with matter power spectrum @xmath124 in a spatially - flat universe ,
the cmb anisotropy angular power spectrum is ( peebles 1982 ; bond & efstathiou 1987 ) @xmath125 we can estimate the slope of the cmb anisotropy power spectrum near the smoothing scale from the amplitude of the genus curve by using equation ( 10 ) and this formula .
when the map data becomes available one can measure the genus of the observed cmb anisotropy and so constrain cosmological models and structure formation mechanisms .
here we use our simulated cmb anisotropy maps to study the sensitivity of the genus statistic to models in the presence of instrumental noise and when the sky coverage is incomplete .
we use the stereographic projection to map the cmb anisotropy sky simulations onto a plane .
the stereographic projection is conformal and preserves local geometry . for a hemisphere
the projection is defined by @xmath126 where @xmath127 is the radius and @xmath128 is the position angle in the projected plane , @xmath77 is galactic latitude and @xmath129 is galactic longitude ( and should not be confused with the legendre polynomial index ) .
figure 3 ( plate 2 ) shows such projections for the ocdm and @xmath22cdm simulations . when computing the genus we exclude regions with @xmath130 to excise the galactic plane region where the galactic signal is likely to be important .
we use the method of gott et al .
( 1990 ) to compute the topology of the proper planar projection of the simulated cmb anisotropy sky .
in figure 5 points with error bars show the two - dimensional genus topology , at @xmath131 fwhm smoothing , derived from the simulated map data ( accounting for noise ) .
the error bars are formal errors in the mean derived from the variance among the genus measurements in each of the eight octants of the sphere .
the curves show the function @xmath132 , the form expected for a random - phase gaussian field , where @xmath133 has been adjusted to fit the points optimally via the least - squares method . for goodness of fit
, the points along the curves should be compared with student s @xmath134-variates , rather than gaussian variates , since the error bars have been estimated from the data themselves ( rather than independently ) . following gott et al .
( 1990 ) and colley ( 1997 ) , we use the variable @xmath135 to estimate the goodness of fit .
the sum runs over the 21 @xmath110 values at which the genus @xmath136 is computed ; @xmath137 is the mean genus among the eight octants of the sphere and @xmath138 is the variance .
we find for the ocdm case @xmath139 , and for fcdm @xmath140 . since @xmath141 does not follow a @xmath142 distribution , due to the non - gaussian distributions of the @xmath143 s , we generate @xmath144 monte carlo runs of @xmath141 variables with 20 degrees of freedom , and 8 samplings within each degree of freedom , exactly as we expect in equation ( [ tilchi ] ) ( since we have fit one parameter , the amplitude ) . in each run
, we generate 8 independent ( @xmath145 ) gaussian variates within 20 independent groups . from the eight samples in each group
@xmath146 we estimate the mean ( @xmath147 ) and the standard deviation in the mean @xmath148 ; @xmath149 is thus a ( @xmath145 ) @xmath134-variate . in the @xmath144 monte carlo simulations ,
we compute @xmath141 as in equation ( [ tilchi ] ) , by summing the squares of the twenty @xmath134-variates .
we find that the @xmath141 values of 23.6 and 26.3 fall at the 43% and 54% levels of the cumulative distribution of the @xmath144 simulated @xmath141 variables , demonstrating an excellent agreement with the theoretical curve .
in addition to noise , several non - cmb anisotropy foregrounds could contaminate the map data .
perhaps the foreground of most consequence for the genus analysis is the radio point - source background ( mainly agns ) . to assess its significance , we have added point sources to the @xmath5 fwhm smoothed ocdm map .
these point sources ( also smoothed ) are taken to have a number distribution in flux , @xmath150 , given by @xmath151 .
the normalization @xmath152 is chosen to reflect the real sky and has an average of one source at 3 @xmath61k or brighter per @xmath153 pixel ( holdaway , owen , & rupen 1994 ) .
comparing to the ocdm map with noise alone , we find that the point sources introduce an rms scatter in @xmath154 of 23 ( relative to the point - source - free map ) , significantly smaller than the typical uncertainty in @xmath154 , which is 42 in the map without point sources . we have also compared the genus values ( between the maps with and without point sources ) using the @xmath142 statistic .
we find @xmath155 out of an expected 21 ( 21 @xmath110 values ) , confirming that the effect of the point sources on the genus is very small . since larger smoothing lengths would diminish the signal from point sources even further ( as with the noise ) , we conclude that the effects of the point - source foreground on the genus are small .
galactic foregrounds , including warm dust and free - free emission , have been selected against in choosing the frequency coverage of map . at the optimum frequencies ,
these foregrounds are expected to be nearly an order of magnitude smaller than the cmb anisotropy .
furthermore , in the analysis of the dmr data ( bennett et al . 1996 ; grski et al .
1996 , and references therein ) the @xmath0-dirbe 140 @xmath61 m sky map ( reach et al .
1995 ) was used to correct for dust and free - free emission contamination , and the 408 mhz all - sky radio survey ( haslam et al .
1982 ) was used to correct for synchrotron emission contamination .
a similar approach , strengthened by new all - sky survey data , should ( hopefully ) suffice for dealing with these contaminants in the map data .
another possible foreground is the sunyaev - zeldovich effect in clusters of galaxies ( sunyaev & zeldovich 1972 ) . here hot electrons compton scatter the cmb photons to uniformly higher frequency , introducing an apparent decrease in brightness temperature as great as @xmath156 mk on the few arc - minute scale ( for very massive clusters with a velocity dispersion of @xmath157 km / s , markevitch 1992 ) . at smoothing lengths of @xmath5 and higher , the effect is reduced to @xmath158k , significantly below the peaks in the cmb anisotropy ( see figure 2 ) .
also , x - ray surveys of hot gas in the intracluster medium can be used to locate the massive clusters causing the shift , and to help estimate the expected magnitude of the shift ( e.g. , klein 1991 ) .
( in the map data , the large sunyaev - zeldovich signal will come from the coma cluster ; this region can be excised from the analysis .
this signal can also be estimated by projecting the thermal fluctuations out of the two highest frequency map channels .
a combination of these cuts should be able to remove most of the sunyaev - zeldovich contamination from the map data . )
non - gaussian features in the cmb anisotropy will affect the genus curve in many different ways .
here we consider two cases that can be quantified .
non - gaussianity can shift the observed genus curve to the left ( towards negative thresholds ) or right , near the mean threshold level .
it can also alter the amplitudes of the genus curve at positive and negative levels differently , causing @xmath159 .
the direction and degree of shift and asymmetry of the genus curve depend on the number , size and height / depth of hot and cold spots .
we have computed limits on these non - gaussianity measures from our simulated maps . in the fcdm model
we find that the map data will detect a shift of the genus curve near @xmath3 ( the five @xmath110 values in figure 5 centered on @xmath160 ) for @xmath161 for fwhm smoothing of @xmath162 ; the limits are the 2-@xmath37 error bars derived from the variances between fits for @xmath163 from the eight octants .
the asymmetry of the genus amplitude near @xmath164 ( the five @xmath110 values centered on 1 and @xmath165 ) can be detected for @xmath166 at @xmath167 fwhm smoothing ( also 2-@xmath37 ) .
the limits for @xmath168 and @xmath169 are very similar for ocdm and @xmath22cdm . from figure 5
, it is clear that the amplitude of the genus curves for fcdm and ocdm differ significantly .
we first consider the genus amplitude for the case when instrumental noise is ignored . in the four cosmogonies considered , and in the scale - invariant power - law model , we expect from equation ( 10 ) a theoretical genus per steradian of @xmath170 ( fcdm ) , @xmath171 ( @xmath22cdm ) , @xmath172 ( ocdm ) , @xmath173 ( rcdm ) , and @xmath174 ( @xmath175 scale - invariant power - law model ) , when the total smoothing scale ( by the beam and the filter ) is @xmath176 fwhm .
the uncertainty limits here account for cosmic variance , computed from actual realizations of @xmath87 s .
note that these limits are computed from the galactic - plane - excluded maps . also , since the genus amplitude does not depend on the amplitude of the cmb anisotropy angular power spectrum , these results are insensitive to the uncertainty in the dmr normalization of the models .
we have also computed the genus directly from simulated maps that include receiver noise , for a single realization of each model .
( since all the model computations use the same random number seed , the cosmic variance effect is minimal . ) in figure 6 we have plotted these genus amplitudes at @xmath177 for the fcdm , ocdm and @xmath22cdm models . for comparison , we have also plotted the genus for the same cosmogonical simulations but now ignoring noise . in the inset , we have zoomed in on the values at @xmath178 and @xmath76 total fwhm smoothing , with 1-@xmath37 error bars derived from comparison of the fits to the genus values within each of the eight octants ( these error bars agree very well with the @xmath179 level , as expected , press 1992 ) .
we have listed in table 1 the genus values , with statistical uncertainty and expected cosmic variance , from figure 6 .
note that the @xmath5 column values are subject to finite pixelization of both the model map results and the projected maps ( melott et al .
tests have shown that amplitude ratios between models remain very nearly constant when moving to higher resolution in the projected maps .
hence , it is the finite pixelization of our simulations which ultimately limits the resolution for the genus computation .
the three cosmogonies of figure 6 and table 1 have significantly different genus amplitudes , even when the expected instrumental noise is accounted for .
they can thus be discriminated between on the basis of the genus amplitude .
for instance , at the @xmath5 smoothing scale each model is distinguished from the other two at greater than 6-@xmath37 confidence ( table 1 and we have added all error bars in quadrature ) .
( we again emphasize that the other cosmological parameters , such as @xmath100 and @xmath13 , are not allowed to vary in these models . ) these results may be directly compared to the genus measurements from map , as an independent check of the results inferred from the power spectrum and correlation function analyses . in order to understand more completely the effects of receiver noise on the genus
, we can look in detail at what is happening , structure by structure , when noise is added . at a high temperature cut ,
for instance , noise can add a hot spot by adding a small amount of positive temperature to a region near the cut ; conversely , noise can remove a small hot spot by cooling a high - temperature region to below the cut .
we have illustrated schematically in figure 7 the possible effects of noise on the genus . in figure 7@xmath77 and
7@xmath180 the noise has neither added nor subtracted hot spots , leaving the genus unaltered . in figure 7@xmath83 , however , the noise has removed a hot spot and reduced the genus by one ; in 7@xmath78 the noise has split a hot spot into two and added one to the genus . in figure
7@xmath181 the noise has added an artificial hot spot and one to the genus while in 7@xmath150 the noise has merged two hot spots into one and reduced the genus by one .
these splittings and mergings can occur in higher multiples .
we now examine our simulated maps for these effects . in figure 8@xmath83
we have plotted the @xmath177 threshold mask for a subsection of the noiseless ocdm @xmath76 fwhm smoothed map . in figure 8@xmath77
we have plotted the same mask , but for a map with noise added .
several of the structures in figure 8@xmath83 become subdivided in figure 8@xmath77 noise has split hot spots .
also , several new hot spots appear , while at least one cold spot ( in the upper right ) develops within a hot spot .
to understand these changes quantitatively , we have counted the number of times each case in figure 7 arises for the ocdm simulations of figure 8 .
results from this analysis are given in table 2 .
table 2 shows that about 80% ( 660 out of [ 78 + 660 + 64 + 15 + 2 + 2 ] ) of the structures counted in the noise - free map correspond to exactly one structure in the map with noise .
78 real hot spots occur with no corresponding hot spots in the map with noise , while 83 ( = 64 + 15 + 2 + 2 ) overlap two or more hot spots in the map with noise .
very similar correspondence exists in comparing the noise - added map hot spots to the noise - free map hot spots , except in the `` 0 '' column .
about a quarter ( 249 out of [ 249 + 683 + 13 + 2 + 1 ] ) of the hot spots in the map with noise had no corresponding hot spots in the noise - free map .
this is to be expected because the noise adds artificial small - scale bumps on top of the signal .
these bumps can push a pixel that was just below the threshold ( here @xmath177 ) to a value over the threshold , so that extra hot spots appear ( similarly , extra cold spots appear at @xmath182 ) . at @xmath76 fwhm smoothing about two - thirds ( 683 out of 1011 ) of the structures in the simulated map maps are due to real signal .
therefore most @xmath76 scale structures in the map maps will correspond to real fluctuations in the microwave background .
in this paper we have generated mock cmb anisotropy maps of representative dmr - normalized cdm cosmogonies appropriate for the map satellite experiment . we have studied the sensitivity of the simulated map data to cosmology , sky coverage , and instrumental noise .
an accurate knowledge of instrumental noise is essential if the statistical tests considered here are to live up to expectations .
we have focused on the correlation function and the genus statistic as tests of cosmogonical models . if the underlying theory has gaussian random phases , then these two statistics are just integrals of the power spectrum . however , since they are computed from the data in a very different manner than the power spectrum , they will be useful checks on power spectrum computations even in gaussian random phase models .
the genus statistic will also be a powerful test of the random phase hypothesis .
the correlation function statistic provides an accurate measurement of the acoustic hubble radius at decoupling from the simulated map data .
the @xmath183 ocdm model considered here can be distinguished from the spatially - flat fcdm and @xmath22cdm models with more than 99% confidence from the location of the acoustic valley in the correlation function , even when the expected amount of instrumental noise is present .
the fcdm and @xmath22cdm models can also be distinguished at separations @xmath184 by the amplitude of the correlation function .
note however that for these models all other cosmological parameters ( such as @xmath13 and @xmath100 ) are fixed .
the genus of the cmb fluctuations can also be accurately measured from the map data .
the genus curves of gaussian cosmological models allow a 2-@xmath37 statistical horizontal shift of the zero - crossing point by only @xmath185 when the total effective smoothing is @xmath167 fwhm .
the asymmetry of the genus curve at the positive and negative threshold levels is only @xmath186 in the gaussian cosmogonies considered .
deviations of the observed map data that exceed these values will be evidence for non - gaussian behavior .
the amplitude of the genus curve , which is a measure of the shape of the power spectrum at the smoothing scale , can also be a powerful discriminator between cosmological models .
even with the expected amount of instrumental noise and partial sky coverage , the map data should discriminate between the cosmogonies considered here at confidence levels exceeding 99% just on the basis of the genus amplitude analysis . of possible outcomes from the map experiment , a more interesting one would be if the results were in close agreement with the low-@xmath12 open - bubble inflation model .
not only would this tell us the value of @xmath12 , but it would affirm inflation and support the open - bubble inflation model in particular , thus providing evidence that there are universes other than our own ( other bubble universes ) , all arising out of an initial metastable inflation state .
cbp acknowledges the support of the basic science research institute program bsri97 - 5408 , and the kosef program 96 - 1400 - 04 - 01 - 3 .
jrg and wnc are supported by nsf grant ast-9529120 and by nasa grant nag5 - 2759 .
wnc is grateful to the fannie and john hertz foundation for its gracious support .
this paper represents part of wnc s thesis at princeton university .
br acknowledges support from nsf grant eps-9550487 with matching support from the state of kansas and from a k@xmath187star first award .
dns acknowledges the map / midex project for support .
lccccccc & + map@xmath198 @xmath199 map@xmath200 & 0 & 1 & 2 & 3 & 4 & 5 & 6 + noise - free @xmath199 noise & 78 & 660 & 64 & 15 & 2 & 2 & 0 + noise @xmath199 noise - free & 249 & 683 & 63 & 13 & 2 & 1 & 0 + change of genus & -1 & 0 & 1 & 2 & 3 & 4 & 5 + banday , a. j. , grski , k. m. , bennett , c. l. , hinshaw , g. , kogut , a. , lineweaver , c. , smoot , g. f. , & tenorio , l. 1997 , , 475 , 393 bennett , c. l. , et al .
1996 , , 464 , l1 bond , j. r. , & efstathiou , g. 1987 , , 226 , 655 bond , j. r. , efstathiou , g. , & tegmark , m. 1997 , , in press bond , j. r. , & jaffe , a. 1997 , in microwave background anisotropies , ed .
f. r. bouchet , r. gispert , b. guiderdoni , & j. tran thanh van ( gif - sur - yvette : editions frontieres ) , 197 bucher , m. , goldhaber , a. s. , & turok , n. 1995 , phys .
d , 52 , 3314 bunn , e. f. , & sugiyama , n. 1995 , , 446 , 49 cheng , e. s. , et al .
1997 , , 488 , l59 church , s. e. , ganga , k. m. , ade , p. a. r. , holzapfel , w. l. , mauskopf , p. d. , wilbanks , t. m. , & lange , a. e. 1997 , , 484 , 523 cole , s. , weinberg , d. h. , frenk , c. s. , & ratra , b. 1997 , , 289 , 37 coles , p. 1988 , , 234 , 509 coles , p. , pearson , r. c. , borgani , s. , plionis , m. , & moscardini , l. 1997 , , in press colley , w. n. 1997 , , 489 , 471 colley , w. n. , gott , j. r. , & park , c. 1996 , , 281 , l82 croft , r. a. c. , weinberg , d. h. , katz , n. , & hernquist , l. 1997 , , submitted da costa , l. n. , vogeley , m. s. , geller , m. j. , huchra , j. p. , & park , c. 1994 , , 437 , l1 dodelson , s. , kinney , w. h. , & kolb , e. w. 1997 , phys .
d , 56 , 3207 ganga , k. , page , l. , cheng , e. , & meyer , s. 1994 , , 432 , l15 ganga , k. , ratra , b. , & sugiyama , n. 1996 , , 461 , l61 ganga , k. , ratra , b. , gundersen , j. o. , & sugiyama , n. 1997a , , 484 , 7 ganga , k. , ratra , b. , church , s. e. , sugiyama , n. , ade , p. a. r. , holzapfel , w. l. , mauskopf , p. d. , & lange , a. e. 1997b , , 484 , 517 ganga , k. , ratra , b. , lim , m. a. , sugiyama , n. , & tanaka , s. t. 1998 , , 114 , in press gardner , j. p. , katz , n. , weinberg , d. h. , & hernquist , l. 1997 , , 486 , 42 grski , k. m. , ratra , b. , sugiyama , n. , & banday , a. j. 1995 , , 444 , l65 grski , k. m. , banday a. j. , bennett , c. l. , hinshaw , g. kogut , a. , smoot , g. f. , & wright , e. l. 1996 , , 464 , l11 grski , k. m. , ratra , b. , stompor , r. , sugiyama , n. , & banday , a. j. 1998 , apjs , 114 , 1 gott , j. r. 1982 , nature , 295 , 304 gott , j. r. 1986 , in inner space / outer space : the interface between cosmology and particle physics , ed .
e. w. kolb , et al .
( chicago : university of chicago press ) , 362 gott , j. r. 1997 , in critical dialogues in cosmology , ed .
n. turok ( singapore : world scientific ) , 519 gott , j. r. , melott , a. l. , & dickinson , m. 1986 , , 306 , 341 gott , j. r. , park , c. , juszkiewicz , r. , bies , w. e. , bennett , d. p. , bouchet , f. r. , & stebbins , a. 1990 , , 352 , 1 gott , j. r. , weinberg , d. h. , & melott , a. l. 1987 , , 319 , 1 griffin , g. s. , nguyn , h. t. , peterson , j. b. , & tucker , g. s. 1997 , in preparation gundersen , j. o. , et al .
1995 , , 443 , l57 guth , a. 1981 , phys .
d , 23 , 347 gutirrez , c. m. , hancock , s. , davies , r. d. , rebolo , r. , watson , r. a. , hoyland , r. j. , lasenby , a. n. , & jones , a. w. 1997 , , 480 , l83 hancock , s. , rocha , g. , lasenby , a. n. , & gutirrez , c. m. 1997 , , in press harrison , e. r. 1970 , phys . rev .
d , 1 , 2726 haslam , c. g. t. , salter , c. j. , stoffel , h. , & wilson , w. e. 1982 , a&as , 47 , 1 hinshaw , g. , banday , a. j. , bennett , c. l. , grski , k. m. , kogut , a. , lineweaver , c. h. , smoot , g. f. , & wright , e. l. 1996 , , 464 , l25 hinshaw , g. , bennett , c. l. , & kogut , a. 1995 , , 441 , l1 holdaway , m. a. , owen , f. n. , & rupen , m. p. 1994 , mma memo 123 jenkins , a. , et al .
1997 , , submitted jungman , g. , kamionkowski , m. , kosowsky , a. , & spergel , d. n. 1996 , phys .
d , 54 , 1332 kamionkowski , m. , spergel , d. n. , & sugiyama , n. 1994 , , 426 , l57 kazanas , d. 1980 , , 241 , l59 klein , u. , rephaeli , y. , schlickeiser , r. , & wielebinski , r. 1991 , , 244 , 43 kogut , a. , banday , a. j. , bennett , c. l. , grski , k. m. , hinshaw , g. , smoot , g. f. , & wright , e. l. 1996 , , 464 , l29 leitch , e. m. , myers , s. t. , readhead , a. c. s. , & pearson , t. j. 1997 , in preparation lim , m. a. , et al .
1996 , , 469 , l69 lineweaver , c. h. , & barbosa , d. 1997 , , submitted maddox , s. j. , efstathiou , g. , sutherland , w. j. , & loveday , j. 1990 , , 242 , 43p markevitch , m. , blumenthal , g. r. , forman , w. , jones , c. , sunyaev , r. a. 1992 , , 395 , 326 masi , s. , de bernardis , p. , de petris , m. , gervasi , m. , boscaleri , a. , aquilini , e. , martinis , l. , & scaramuzzi , f. 1996 , , 463 , l47 melott , a. l. , cohen , a. p. , hamilton , a. j. s. , gott , j. r. , & weinberg , d. h. 1989 , , 345 , 618 netterfield , c. b. , devlin , m. j. , jarosik , n. , page , l. , & wollack , e. j. 1997 , apj , 474 , 47 novikov , d. i. , & jrgensen , h. e. 1996 , , 471 , 521 page , l. a. 1997 , astro - ph/9703054 park , c. 1990 , , 242 , 59p park , c. 1991 , , 382 , l59 park , c. , & gott , j. r. 1988 , baas , 20 , 987 park , c. , gott , j. r. , melott , a. l. , & karachentsev , i. d. 1992 , , 387 , 1 park , c. , vogeley , m. s. , geller , m. j. , & huchra , j. p. 1994
, , 431 , 569 peacock , j. a. , & dodds , s. j. 1994 , , 267 , 1020 peebles , p. j. e. 1982 , , 263 , l1 peebles , p. j. e. 1984 , , 284 , 439 peebles , p. j. e. , & yu , j. t. 1970 , , 162 , 815 piccirillo , l. , limon , m. , nichols , j. , schaefer , r. k. , femena , b. , gutirrez , c. m. , rebolo , r. , & watson , r. a. 1997 , , submitted platt , s. r. , kovac , j. , dragovan , m. , peterson , j. b. , & ruhl , j. e. 1997 , , 475 , l1 press , w. h. , rybicki , g. b. , & hewitt , j. n. 1992 , , 385 , 404 ratra , b. , & peebles , p. j. e. 1994 , , 432 , l5 ratra , b. , & peebles , p. j. e. 1995 , phys .
d , 52 , 1837 ratra , b. , sugiyama , n. , banday , a. j. , & grski , k. m. 1997a , , 481 , 22 ratra , b. , ganga , k. , sugiyama , n. , tucker , g. s. , griffin , g. s. , nguyn , h. t. , & peterson , j. b. 1997b , ksu preprint ksu-97/3 reach , w. t. , franz , b. a. , kelsall , t. , & weiland , j. l. 1995 , in unveiling the cosmic infrared background , ed .
e. dwek ( new york : aip ) , 37 sato , k. 1981 , phys .
b , 99 , 66 saunders , w. , et al . 1991 , nature , 349 , 32 schmalzing , j. , & grski , k. m. 1997 , mnras , submitted scott , p. f. , et al .
1996 , , 461 , l1 smoot , g. f. , et al .
1992 , , 396 , l1 snyder , j. p. 1993
, flattening the earth : two thousand years of map projections ( chicago : university of chicago press ) , 113 spergel , d. n. 1994 , baas , 26 , 1427 stompor , r. 1997 , in microwave background anisotropies , ed . f. r. bouchet , r. gispert , b. guiderdoni , & j. tran thanh van ( gif - sur - yvette : editions frontieres ) , 91 stompor , r. , grski , k. m. , & banday , a. j. 1995 , , 277 , 1225 sugiyama , n. 1995 , , 100 , 281 sunyaev , r. a. , & zeldovich , ya . b. 1972 , comments ap .
space phys . , 4 , 173 tegmark , m. , de oliveira - costa , a. , devlin , m. j. , netterfield , c. b. , page , l. , & wollack , e. j. 1997 , , 474 , l77 tucker , g. s. , gush , h. p. , halpern , m. , shinkoda , i. , & towlson , w. 1997 , , 475 , l73 turner , m. s. 1997 , in critical dialogues in cosmology , ed .
n. turok ( singapore : world scientific ) , 555 vogeley , m. s. , park , c. , geller , m. j. , & huchra , j. p. 1992
, , 391 , l5 vogeley , m. s. , park , c. , geller , m. j. , huchra , j. p. , & gott , j. p. 1994 , 420 , 525 white , m. , & bunn , e. f. 1995 , , 443 , l53 winitzki , s. , & kosowsky , a. 1997 , new astron .
, submitted wright , e. l. , bennett , c. l. , grski , k. m. , hinshaw , g. , & smoot , g. f. 1996 , , 464 , l21 wright , e. l. , et al .
1992 , , 396 , l13 yamamoto , k. , sasaki , m. , & tanaka , t. 1995 , , 455 , 412 zaldarriaga , m. , spergel , d. n. , & seljak , u. 1997 , apj , 488 , 1 zeldovich , ya .
b. 1972 , , 160 , 1p | we have simulated cosmic microwave background ( cmb ) anisotropy maps for several @xmath0-dmr - normalized cold dark matter ( cdm ) cosmogonies , to make predictions for the map experiment , an upcoming whole - sky cmb anisotropy space mission .
we have studied the sensitivity of the simulated map data to cosmology , sky coverage , and instrumental noise . given an accurate knowledge of instrumental noise
, map data will ably discriminate amongst the cosmogonies considered , and superbly determine the topology of the initial fluctuations . a correlation function analysis of the simulated map data results in a very accurate measurement of the acoustic hubble radius at decoupling . a low - density open cdm model with @xmath1
can be distinguished from the @xmath2 fiducial cdm model or a spatially - flat cdm model with a cosmological constant and @xmath1 with more than 99% confidence from the location of the acoustic valley " in the correlation function .
a genus analysis of the simulated map data indicates that in cosmogonies with gaussian random - phase initial conditions , the horizontal shift of the zero - crossing point of the genus curve near the mean temperature threshold level @xmath3 should not exceed @xmath4 ( 0.04 ) when the total effective fwhm smoothing is @xmath5 @xmath6 .
the asymmetry of the genus curve at the positive and negative threshold levels should not exceed @xmath7 ( 4% ) at @xmath5 @xmath8 fwhm smoothing .
deviations of the observed map data in excess of these small values will be evidence for non - gaussian behavior .
the amplitude of the genus curve is a measure of the shape of the power spectrum at the smoothing scale .
even with the expected amount of instrumental noise and partial sky coverage ( due to the galaxy ) , the map data should allow discrimination amongst the cosmogonies considered at more than 99% confidence solely from a genus amplitude analysis . |
recall , a triple @xmath0 consisting of a smooth manifold @xmath1 , a smooth riemannian metric @xmath2 on @xmath1 and a smooth vector field @xmath3 is said to be a _ ricci soliton with expansion constant _
@xmath4 , if it satisfies @xmath5 where @xmath6 denotes the lie derivative of @xmath2 with respect to @xmath7 .
for such a triple the ricci flow equation @xmath8 with initial condition @xmath9 has , in the @xmath10-interval on which @xmath11 , the solution @xmath12 where @xmath13 is the ( possibly locally defined ) time @xmath10 flow of @xmath7 . when @xmath14 , the soliton is said to be steady , when @xmath15 , the soliton is said to be shrinking , and when @xmath16 , the soliton is said to be expanding .
let @xmath17 denote the levi - civita connection of @xmath2 .
if @xmath18 for some smooth real - valued function @xmath19 on @xmath1 , then we say @xmath20 is a _
gradient ricci soliton
_ and @xmath19 is its _ soliton potential_. it has been known for some time that closed two - dimensional ricci solitons must have constant curvature see ( * ? ? ?
* theorem 10.1 ) and @xcite also @xcite and @xcite .
it was also well - known that any two - dimensional gradient ricci soliton must admit a non - trivial killing vector field
see , for instance , the editor s footnote of @xcite .
however , the possible geometries in the open case were not completely determined . the most complete analysis in this direction was in ( * ? ? ? * appendix a and b ) where all rotationally symmetric two - dimensional self - similar solutions to the logarithmic fast diffusion equation were explicitly computed . as the logarithmic fast diffusion equation is the evolution of the conformal factor with respect to a fixed flat background of a metric evolving under the ricci flow ,
the analysis of @xcite gives many examples of two - dimensional ricci solitons .
more in the spirit of the present note is @xcite , where the killing vector field is used to show that all positively curved two - dimensional expanding ricci solitons are the rotationally symmetric examples of @xcite .
a more complete overview of known results may be found in ( * ? ? ?
* chap . 1 ) .
motivated by an observation of @xcite connecting two - dimensional steady gradient ricci solitons with minimal surface metrics , we reduce the classification of two - dimensional gradient ricci solitons to the analysis of a simple autonomous first order ode . the analysis of this ode gives : [ listofsolitonsthm ] the following twelve one - parameter families of rotationally symmetric metrics are gradient ricci soliton metrics on either the disks @xmath21 or @xmath22 or the annuli @xmath23 or @xmath24 : 1 .
steady solitons : 1 .
@xmath25 the cigar solitons of hamilton @xcite i.e. complete positively curved steady solitons
the parameter @xmath26 corresponds to scaling of the metric ; 2 .
@xmath27 incomplete steady solitons with unbounded negative curvature called the exploding solitons in @xcite the parameter @xmath26 corresponds to scaling of the metric ; 3 .
@xmath28 incomplete negatively curved steady solitons asymptotic to a cylinder of radius @xmath29 at @xmath30 and to @xmath31 at @xmath32 .
shrinking solitons : 1 .
@xmath33 two families of incomplete positively curved shrinking solitons which extend smoothly to @xmath34 so that @xmath35 is totally geodesic and has length @xmath36 the parameter @xmath37 satisfies @xmath38 for @xmath39 and @xmath40 for @xmath41 ; 2 .
@xmath42 incomplete negatively curved shrinking solitons asymptotic to @xmath43 ; 3 . expanding solitons : 1 .
@xmath44 complete positively curved expanding solitons asymptotic to the end of a flat cone of angle @xmath45
see @xcite ; 2 .
@xmath46 complete negatively curved expanding solitons asymptotic to to the end of a flat cone of angle @xmath47 ; 3 .
@xmath48 complete negatively curved expanding solitons asymptotic to a hyperbolic cusp at @xmath30 and to the end of a flat cone of angle @xmath49 at @xmath32 ; 4 .
@xmath50 incomplete negatively curved expanding solitons asymptotic to to the end of a flat cone of angle @xmath51 at @xmath30 and extending smoothly to @xmath52 so that @xmath35 is totally geodesic and of length @xmath36 ; 5 .
@xmath53 incomplete negatively curved expanding solitons asymptotic to @xmath43 ; 6 .
@xmath54 incomplete negatively curved expanding solitons asymptotic to a hyperbolic cusp at @xmath30 and to @xmath43 at @xmath32
@xmath55 incomplete negatively curved expanding solitons asymptotic to @xmath43 at @xmath30 and extending smoothly to @xmath52 so @xmath35 is totally geodesic and of length @xmath36 ; the metrics @xmath56 and @xmath57 may be `` doubled '' to give @xmath58 metrics on a doubled surface .
the doubled metrics are not themselves ricci solitons .
all these metrics seem to have been written down in @xcite , however their geometric properties are not fully discussed there .
a further consequence of the analysis of the ode is that the above list contains all possible models of non - constant curvature two - dimensional gradient ricci solitons .
[ mainclassificationthm ] let @xmath59 be a smooth gradient ricci soliton with @xmath1 a connected orientable two - manifold .
if @xmath60 is non - empty and @xmath61 is cyclic , then there is a @xmath62 , a metric @xmath63 on @xmath64 and an isometric immersion @xmath65 so that : 1 .
@xmath66 ; 2 .
@xmath67 is maximal in the sense that if @xmath68 is a two - dimensional connected gradient ricci soliton and @xmath69 is a smooth isometric embedding , then @xmath70 and @xmath71 ; 3 .
@xmath63 is rotationally symmetric , that is in polar coordinates @xmath72 @xmath73 for some positive functions @xmath74 ; 4 .
there are @xmath75 so that if @xmath76 , then @xmath77 may be isometrically embedded in one of the models given in theorem [ listofsolitonsthm ] .
if @xmath78 is a metric on @xmath79 , then for any @xmath80 the metric @xmath81 on @xmath79 is locally isometric to @xmath2 . in particular
, @xmath2 is a gradient ricci soliton metric if and only if @xmath82 is . likewise scaling a ricci soliton metric
gives a new ricci soliton metric with scaled expansion constant .
neither the orientability condition on @xmath1 nor the condition on @xmath61 can be removed because several locally isometric but not globally isometric subsets of a fixed model soliton may be glued together to yield counter - examples .
theorems [ listofsolitonsthm ] and [ mainclassificationthm ] resolve two questions regarding the structure of non - compact two - dimensional solitons ( see @xcite ) . specifically
, a complete two - dimensional gradient ricci soliton has bounded curvature .
there exist complete two - dimensional expanding ricci solitons with negative curvature that are topologically disks and are not hyperbolic space namely the metrics @xmath83 and the universal covers of the metrics @xmath84 .
the metrics @xmath83 are parabolic and appear in @xcite though this does not seem to be widely known .
the universal covers of the metrics @xmath84 are non - parabolic .
finally , in section [ varsec ] we give a simple variational characterization of two - dimensional gradient ricci solitons and observe a connection with @xcite .
the authors would like to thank b. kotschwar for his comments and for pointing out a serious error in an early draft .
we also thank p. topping for bringing @xcite to our attention .
we will henceforth assume @xmath1 to be a connected two - manifold .
then for a gradient ricci soliton equation becomes @xmath85 which is equivalent to @xmath86 here @xmath87 denotes the gauss curvature , @xmath88 the laplacian , and @xmath89 the trace - free hessian of @xmath2 . following , ( * ? ? ?
* appendix a ) , covariant differentiation of @xmath90 together with straightforward manipulations gives @xmath91 at a point @xmath92 where @xmath93 we thus have @xmath94 which yields @xmath95 from this we see that @xmath96 we have shown : [ charsol ] let @xmath97 be a riemannian two - manifold with non - zero gauss curvature . then @xmath2 is a gradient ricci soliton metric with expansion constant @xmath4 if and only if @xmath98 moreover , if @xmath59 is a gradient ricci soliton , then we may take @xmath99 .
[ killing ] recall ( c.f .
@xcite ) that if @xmath100 is a gradient ricci soliton and @xmath1 is an oriented surface , then @xmath101 is a killing vector field for @xmath2 . here , as usual , @xmath102 denotes the integrable almost complex structure induced by @xmath2 and the orientation , i.e. @xmath102 rotates @xmath103 by @xmath104 in positive direction .
proposition [ charsol ] and remark [ killing ] reduce the classification of gradient ricci solitons in dimension two to the study of a simple ode . indeed , assume @xmath97 is a gradient ricci soliton and @xmath2 does not have constant curvature .
we first observe that @xmath1 has non - vanishing curvature and has non - constant curvature almost everywhere .
[ densityprop ] if @xmath20 is a connected two - dimensional gradient ricci soliton and @xmath105 , then @xmath93 on all of @xmath1 and @xmath106 consists of isolated points .
set @xmath107 and decompose @xmath108 into @xmath109 , @xmath110 and @xmath111 . as @xmath1 has non - constant curvature , @xmath112 is non - empty .
we claim it is dense in @xmath1 .
this follows from standard unique continuation results for uniformly elliptic equations @xcite and the fact that @xmath87 satisfies @xmath113 on @xmath1 see ( * ? ? ?
* lemma 1.11 ) .
hence , @xmath114 . indeed , by proposition [ charsol ] ,
@xmath115 is covariant constant on @xmath112 and so is bounded on all of @xmath1 .
fix @xmath116 and let @xmath72 be normal coordinates about @xmath117 . in these coordinates , @xmath118 where @xmath119 are smooth functions depending on @xmath120 and @xmath2 defined near @xmath117 .
we next show that @xmath121 . indeed , by (
* lemma 1.11 ) , the unique continuation result @xcite and the expansion of the laplacian in normal coordinates near any @xmath122 we have @xmath123 for some @xmath124 and @xmath125 .
however , if this is true then there would exist a point in @xmath126 and so @xmath121 . finally , we conclude that @xmath127 consists of isolated points .
indeed , @xmath128 is a conformal killing vector field on @xmath1 and so is ( locally ) the real part of a non - trivial holomorphic vector field . in particular
, it can only vanish at isolated points . by proposition [ charsol ] ,
in a neighborhood of every point @xmath129 there exists a non - trivial killing vector field for @xmath2 .
it follows
see ( * ? ? ?
* lemma 1.18 ) that there exists a local coordinate system @xmath72 defined on some neighborhood @xmath130 of @xmath117 , such that @xmath131 for some smooth positive real - valued function @xmath132 defined in some neighborhood of @xmath133 .
on @xmath130 we thus have @xmath134 since @xmath135 we can assume ( after possibly replacing @xmath117 with a nearby point ) , that @xmath136 is non - vanishing in some simply connected neighborhood @xmath137 of @xmath117 .
setting @xmath138 it follows that @xmath139 is a local coordinate system so that @xmath140 for some positive function @xmath141 defined in some neighborhood of @xmath142 . for @xmath143 in the domain of definition of @xmath141 , this metric is smooth and has gaussian curvature @xmath144 for later reference we remark that the geodesic curvature of the curves @xmath145 is @xmath146 which follows easily from the first variation formula . as @xmath147 is a killing vector field
, we may take @xmath148 for some constant @xmath149 which is non - vanishing since @xmath150 .
on the other hand , from proposition [ charsol ] we obtain @xmath151 combining and gives the autonomous first - order ode : @xmath152 conversely we have : [ smoothext ] if @xmath153 is defined on @xmath154 with @xmath155 and solves for some @xmath156 , then every solution @xmath132 to @xmath157 defined on @xmath158 gives rise to a smooth rotationally symmetric gradient ricci soliton @xmath159 on @xmath79 .
furthermore , there exists a solution @xmath132 to so that the associated metric extends smoothly to @xmath160 if and only if @xmath161 . in this case
@xmath162 and @xmath163 .
the first statement is an immediate consequence of the calculations above . to prove the second statement we first recall the standard result
see ( * ? ? ?
* lemma a.2 ) that a metric of the form @xmath164 for some smooth non - vanishing function @xmath132 on @xmath158 extends smoothly to @xmath160 if and only if @xmath165 suppose the metric associated to a solution of extends smoothly to the topological disk , then it follows with the continuity of the function @xmath141 and ( [ intode ] , [ smoothextcond ] ) that @xmath166 .
conversely , suppose the function @xmath141 satisfies @xmath166 so that @xmath141 smoothly extends to a solution of on the interval @xmath167 for some @xmath168 $ ] .
choosing the initial condition @xmath169 we obtain a solution to defined in some neighborhood of @xmath30 which satisfies @xmath170 .
thus , there exists an interval @xmath171 about @xmath172 on which @xmath132 is an odd function of @xmath173 . we obtain from and @xmath174 which , assuming @xmath166 , converges to @xmath175 as @xmath176 .
because of the rotational symmetry of the metric @xmath2 , we have @xmath177 . more generally , we have the following result relating the qualitative behavior of solutions of to the geometry of the metrics .
[ odetogeomprop ] if @xmath153 is defined on @xmath178 with @xmath179 and solves for some @xmath156 , then the following relationships hold between the behavior of @xmath141 and the geometry of the metric @xmath2 given by : 1 .
[ odetogeom1]@xmath2 has positive curvature if and only if @xmath141 is strictly increasing and has negative curvature if and only if @xmath141 is strictly decreasing ; 2 .
[ odetogeom2]@xmath2 has bounded curvature if and only if @xmath141 is strictly bounded away from @xmath30 ; 3 .
[ odetogeom3 ] @xmath2 is complete if and only if 1 .
[ odetogeom3a]@xmath180 , @xmath181 and @xmath182 ; or 2 .
[ odetogeom3b]@xmath183 , @xmath181 and @xmath141 is strictly bounded away from @xmath30 ; or 3 .
[ odetogeom3d]@xmath184 , @xmath185 and @xmath141 is strictly bounded away from @xmath30 .
[ odetogeom4 ] if @xmath186 and @xmath187 , then @xmath2 is asymptotic to the end of a flat cone of angle @xmath188 ; item is an immediate consequence of .
the converse direction of item follows from . on the other hand , if @xmath141 is not strictly bounded away from the zero , then , as is an autonomous first order ode , @xmath186 and @xmath189 so @xmath190 .
item is an immediate consequence of the form of the metric and lemma [ smoothext ] .
likewise , items and items will follow from the form of the metric and lemma [ smoothext ] provided we show that if @xmath189 , then @xmath191 .
however , from it is clear that if @xmath189 , then in all cases there is a constant @xmath192 , depending on @xmath149 and @xmath4 , and a @xmath193 , depending on @xmath141 , so for @xmath194 , @xmath195 standard ode comparison then implies that for @xmath194 , we have @xmath196 for some @xmath197 which proves the claim .
if @xmath198 , then @xmath199 and so @xmath200 .
hence , by @xmath132 is definte on @xmath201 and has the asymptotics @xmath202
so the associated metric is asymptotic to the end of the claimed cone .
we now analyze the solutions of in order to prove theorem [ listofsolitonsthm ] .
the constant solution @xmath203 separates the space of solutions of into positive and negative solutions . for our purposes it sufficies to study positive solutions .
let @xmath204 be the @xmath205-dimensional space of triples @xmath206 where @xmath207 @xmath208 possibly infinite and @xmath141 is a smooth positive real - valued function defined on some interval @xmath209 which solves @xmath210 there are three natural symmetries of @xmath204 .
first , we let @xmath211 act by @xmath212 the metrics @xmath2 associated to @xmath141 and @xmath82 associated to @xmath213 are locally isometric , but not globally isometric .
second , we let @xmath211 act by @xmath214 where @xmath215 denotes the map @xmath216 .
this corresponds to scaling the metric @xmath2 associated to @xmath141 by @xmath217 . by lemma [ smoothext ] , if @xmath218 and @xmath181 , then this is also true after acting by @xmath219 .
third , is the non - geometric symmetry where @xmath220 acts by translation in @xmath10 , i.e. @xmath221 for @xmath222 .
this action leaves @xmath223 unchanged .
there are six orbits of the actions corresponding to @xmath224 and within each orbit there are one - parameter families of geometrically distinct solitons . due to the fact that certain solutions to blow - up , some orbits contain more than one such family . in this case
@xmath14 so we are considering metrics of steady solitons
. moreover , simplifies to @xmath226 whose solutions are given explicitly by @xmath227 for some constant @xmath228 .
qualitatively , this @xmath229-parameter family of solutions splits into three types , depicted in figure [ figlmz ] .
[ figlmz ] we may assume @xmath231 , otherwise @xmath232 .
it follows that @xmath233 is strictly increasing and hence the solutions blow up in finite time . using the smooth extension condition of lemma [ smoothext ] we obtain the solutions @xmath234 whose associated complete metrics @xmath235
are hamilton s steady _ cigar solitons _ @xcite which are defined on the topological disk and given in polar coordinates by @xmath236 by inspection these solutions are complete and have bounded positive curvature and are asymptotic to a cylinder of radius @xmath26 . moreover , @xmath237 .
the solutions exist for all ( positive ) times , are bounded , and approach @xmath30 as @xmath239 .
it follows from lemma [ odetogeomprop ] that the associated metrics have unbounded negative gauss curvature and are incomplete . using the smooth extension condition of lemma [ smoothext ] we obtain the solutions @xmath240 whose associated family of metrics called _ exploding solitons _ in @xcite
@xmath43 are given in polar coordinates by @xmath241 clearly , @xmath242 .
furthermore , at @xmath243 the metrics have the asymptotics @xmath244 the positive part of the solution exists on the time interval @xmath247 , and approaches @xmath32 as @xmath248 and @xmath30 as @xmath239 . by lemma [ odetogeomprop ] ,
the associated metrics @xmath249 have unbounded negative gauss curvature and are incomplete . writing @xmath250 , and applying the @xmath219 action with @xmath251 , gives @xmath252 in cylindrical coordinates @xmath253 this gives metrics @xmath254 are given for @xmath255 as @xmath256 while for @xmath257 @xmath258 thus , the metrics are all asymptotically cylindrical with radius @xmath259 as @xmath260 . at @xmath261 ,
the metrics are asymptotically @xmath262 and so to leading order behave like @xmath31 as desired .
besides the separatrix @xmath203 , there is a positive separatrix @xmath264 and the orbits split into six families , depicted in figure [ figlmp ] .
[ figlmp ] in this case the separatrix @xmath264 is unstable and so solutions starting above it are strictly increasing .
if @xmath266 , then there is a solution @xmath267 which satisfies @xmath268 .
an ode comparison argument implies that the @xmath267 blow - up at some finite time @xmath269 given by @xmath270 by applying an appropriate @xmath219 action , we may take @xmath271 and so @xmath272 and @xmath273 .
the solutions have the following asymptotics at @xmath271 : @xmath274 clearly , @xmath267 is integrable and hence lemma [ smoothext ] and lemma [ odetogeomprop ] imply that there are positively curved , incomplete metrics , @xmath275 , on the disk @xmath276 .
one verifies that each @xmath275 extends smoothly to @xmath34 and , as @xmath271 the length of @xmath277 is @xmath36 .
moreover , it follows directly from that @xmath277 is totally geodesic .
finally , @xmath275 converges to the flat metric on the disk @xmath278 as @xmath279 , while @xmath275 converges to the metric @xmath280 as @xmath281 .
hence , @xmath282 as @xmath279 while @xmath283 as @xmath281 and @xmath284 is a strictly decreasing function of @xmath208 .
we take the parameter @xmath26 to be this distance . again
the separatrix @xmath264 is unstable and so solutions starting below it are decreasing and tend to @xmath30 as @xmath286 .
if @xmath287 , then there is a solution @xmath288 which satisfies @xmath289 and so , by lemma [ smoothext ] and lemma [ odetogeomprop ] , there exist negatively curved , incomplete metrics @xmath290 with unbounded curvature on @xmath22 .
furthermore , @xmath288 has the asymptotic form as @xmath286 @xmath291 hence , @xmath292 is a asymptotic to @xmath293 as @xmath294 and so taking @xmath295 , @xmath296 is asymptotic to @xmath43 as claimed . in this case
the separatrix @xmath264 is stable and so solutions starting below it are increasing and @xmath298 as @xmath286 .
if @xmath287 , then there is a solution @xmath299 which satisfies @xmath300 and so , by lemma [ smoothext ] and lemma [ odetogeomprop ] , there exists a family of positively curved , complete metrics @xmath301 on @xmath22 which are asymptotic to a cone of angle @xmath302 . in this case
the separatrix @xmath264 is stable and so solutions starting above it are decreasing and @xmath298 as @xmath286 .
an ode comparison argument implies that solutions blow - up at some finite initial time @xmath303 .
the geometry of the associated metrics is quite different depending on whether @xmath304 , @xmath305 or @xmath306 . if @xmath304 and @xmath266 , then there is a solution @xmath307 which satisfies @xmath308 and so , by lemma [ smoothext ] and lemma [ odetogeomprop ] , there exists a family of negatively curved , complete metrics @xmath309 on @xmath22 which are asymptotic to a cone of angle @xmath310 . if @xmath305 , then any solution @xmath311 has the following asymptotics near @xmath305 @xmath312 and so @xmath313 is not locally integrable . by the @xmath219 symmetry
we may take @xmath314 .
hence , by lemma [ odetogeomprop ] , there is a family of negatively curved , complete metrics @xmath315 on a cylinder which are asymptotic at one end to a cone with angle @xmath316 . by , @xmath317 as @xmath176 , while the length of the curve @xmath145 goes to @xmath30 as @xmath318 and hence @xmath315 is asymptotic at the other end to a hyperbolic cusp .
if @xmath306 , then any solution @xmath319 has the following asymptotics near @xmath306 @xmath320 and so @xmath321 is locally integrable . hence using lemma [ odetogeomprop ] , it follows that there is a family of negatively curved , incomplete metrics on @xmath322 which are asymptotic to a cone with angle @xmath323 at the puncture . as with the family @xmath324 , the metrics smoothly extend to @xmath52 and @xmath325 is totally geodesic . by the @xmath219 symmetry
we may take the length of @xmath326 to be @xmath36 which gives the @xmath327 . in this case
there is no positive separatrix and there are four qualitatively different solutions as pictured in figure [ figlmn ] .
[ figlmn ] in this case the separatrix @xmath203 is unstable and so solutions starting above it are strictly increasing , moreover there exists a solution @xmath329 which satisfies @xmath330 .
qualitatively this solution is the same as the solution @xmath267 which satisfies @xmath268 .
indeed , by the @xmath219 symmetry , we may take the blow - up time to be @xmath271 and so @xmath272 and @xmath273 .
hence , the metrics @xmath331 then have similar properties as the metrics produced in section . however , as @xmath332 the metrics @xmath331 converge to the round metric on the hemi - sphere and so @xmath333 . in this case
the separatrix @xmath203 is stable and so solutions starting above it are decreasing and @xmath334 as @xmath286 .
an ode comparison argument implies that solutions blow - up at some finite initial time @xmath303 .
the geometry of the associated metrics is quite different depending on whether @xmath304 , @xmath305 or @xmath306 . if @xmath304 , then there is a solution @xmath335 which satisfies @xmath336 and so , by lemma [ smoothext ] and lemma [ odetogeomprop ] , there exists a family of negatively curved , incomplete metrics @xmath337 with unbounded curvature on @xmath22 which are asymptotic to @xmath43 . if @xmath305 , then any solution @xmath338 has the following asymptotics near @xmath305 @xmath339 and so @xmath340 is not locally integrable . by the @xmath219 symmetry
we may take @xmath314 .
hence , by lemma [ odetogeomprop ] , there is a family of negatively curved , incomplete metrics with unbounded curvature on a cylinder .
these metrics are asymptotic at one end to a hyperbolic cusp and asymptotic at the other end to a member of the family @xmath43 and give the family @xmath341 as claimed . if @xmath306 , then any solution @xmath342 has the following asymptotics near @xmath306 @xmath343 and so @xmath344 is locally integrable .
hence , by lemma [ odetogeomprop ] , there is a family of negatively curved , incomplete metrics with unbounded curvature on @xmath345 . as with @xmath346
, these metrics smoothly extend to @xmath52 and @xmath347 are totally geodesic . by the @xmath219-action @xmath348 may be taken to have length @xmath36 . at the other end , these metrics are asymptotic to a member of the @xmath43 family at the puncture and give the family @xmath57 as claimed .
this completes the analysis needed to prove theorem [ listofsolitonsthm ] .
we now prove theorem [ mainclassificationthm ] .
first a simple lemma : [ dumblem ] suppose @xmath74 is given by lemma [ smoothext ] .
if @xmath349 is a non - constant curvature metric on @xmath79 , then @xmath350 , the group of orientation - preserving isometries , is isomorphic to @xmath351 .
clearly , for each @xmath352 , the maps @xmath353 are orientation preserving isometries .
hence , @xmath351 may be identified with a subgroup of @xmath350 . on the other hand , as solutions to are either strictly monotone or constant curvature , implies that if @xmath354 is an isometry , then @xmath355 and so @xmath356 for some @xmath357 .
as @xmath358 is non - empty , proposition [ densityprop ] implies that @xmath93 and @xmath106 consists of isolated points @xmath359 in particular , @xmath358 is connected . by the analysis of section [ sec2 ] and theorem [ listofsolitonsthm ] , for each point @xmath360 there is a simply - connected neighborhood @xmath361 about @xmath117 and an isometric embedding @xmath362 where @xmath363 is a scaling of a member of one of the twelve families of by theorem [ listofsolitonsthm ] .
clearly , for @xmath364 with @xmath365 , we have @xmath366 and there is an element @xmath367 so that @xmath368 on @xmath369 . as @xmath358 is connected , we conclude that @xmath370 for @xmath371 and @xmath372 a fixed member of one of the families of theorem [ listofsolitonsthm ] . as @xmath360 , we may delete a point if needed and take @xmath373 for @xmath62 and @xmath164 .
given points @xmath374 , a continuous path @xmath375\to m_*$ ] connecting @xmath117 and @xmath376 , and a fixed isometric embedding @xmath377 there is a unique isometric embedding @xmath378 determined by @xmath379 and @xmath208 in the obvious way this is a sort of developing map into @xmath372 .
for a fixed point @xmath380 , there is a group homomorphism @xmath381 given by the holonomy of the above construction which is the obstruction to extending a choice of @xmath382 to an isometric immersion @xmath383 . as @xmath1 is orientable as are ( by inspection ) all the models of theorem [ listofsolitonsthm ] we have that @xmath384 .
furthermore , as @xmath61 is cyclic , there is a generator @xmath385 .
we consider @xmath386 . by lemma , [ dumblem ]
there is a @xmath387 so @xmath356 where @xmath388 .
setting @xmath389 concludes the proof .
let @xmath1 denote a smooth two - manifold . on the space of riemannian metrics @xmath390 of non - vanishing gauss curvature on @xmath1 consider the functional defined by @xmath391=\int_m k_g\log |k_g|\mu_g\ ] ] where @xmath392 is the area - form of @xmath2 .
this functional has been applied to the study of ricci flow on surfaces by hamilton @xcite and chow @xcite in particular hamilton observed that it is monotonically increasing along the ricci flow on spheres with positive gauss curvature . in @xcite , we observed that if @xmath2 is a positively curved metric which is critical for @xmath393 with respect to compactly supported conformal variations , then @xmath394 is the metric of a minimal surface in euclidean @xmath205-space and that every negatively curved minimal surface metric arises in this way . as such , it is interesting to consider critical points of @xmath393 under other variations . in this direction , we show that gradient ricci soliton metrics are the only critical points of @xmath393 with respect to area - preserving variations .
we say that @xmath395 is an @xmath393-critical metric if @xmath393 is stationary at @xmath2 with respect to compactly supported area - preserving variations .
let @xmath2 be a smooth riemannian metric and @xmath396 a smooth symmetric @xmath229-form on @xmath1 .
writing @xmath397 , we have @xmath398 where @xmath399 is the trace of @xmath396 with respect to @xmath2 .
hence , the variations that preserve area are precisely the deformations by trace - free symmetric @xmath229-forms . moreover , we have ( cf .
@xcite ) @xmath400 where @xmath401 denotes the trace - free part of @xmath396 .
a short computation yields @xmath402= \frac{1}{2}\int_{m}h k_g \log|k_g|+\left({{\ensuremath{\operatorname{div } } } } _ g\left({{\ensuremath{\operatorname{div } } } } _ g\ !
\mathring{h}\right)-\frac{1}{2 } \delta_g h - hk_g \right ) ( \log |k_{g}|+1){\mu_g}.\ ] ] if @xmath396 is compactly supported then integrating by parts gives @xmath403= -\frac{1}{4}\int_{m } h \left(\delta_g \log |k_g| + 2 k_g \right )
-2\langle h , \mathring{\nabla}^2 \log |k_g|\rangle_g { \mu_g},\ ] ] where @xmath404 denotes the natural bilinear pairing on elements @xmath405 obtained via @xmath2 . for @xmath396 trace - free simplifies to @xmath402=\frac{1}{2}\int_m\langle h , \mathring{\nabla}^2 \log |k_g|\rangle_g \,{\mu_g}.\ ] ] hence , @xmath2 is an @xmath393-critical metric if and only if the gauss curvature @xmath87 of @xmath2 satisfies @xmath406 the functional @xmath393 is diffeomorphism invariant , i.e. for @xmath407 , a diffeomorphism , @xmath408=\mathcal{e}[g]\ ] ] for every @xmath409 . using noether s principle
we obtain a conservation law for @xmath393 : let @xmath3 be a compactly supported vector field and let @xmath410 denote its time @xmath10 flow
. then we have @xmath411 as a consequence of the diffeomorphism invariance we obtain at an @xmath393-critical metric @xmath412&=-\frac{1}{4}\int_m{{\ensuremath{\operatorname{tr } } } } _ g(l_xg)\left(\delta_g \log |k_g| + 2 k_g \right)\mu_g,\\ & = -\frac{1}{2}\int_m\left(\delta_g \log |k_g| + 2 k_g\right){{\ensuremath{\operatorname{div } } } } _ g\!x\,\mu_g \end{aligned}\ ] ] which vanishes for every compactly supported vector field @xmath7 on @xmath1 if and only if @xmath413 is constant . using proposition [ charsol ] we have thus shown : let @xmath97 be a riemannian two - manifold with non - zero gauss curvature .
then @xmath97 is a gradient ricci soliton if and only if @xmath2 is a critical point of the functional @xmath393 with respect to compactly supported area - preserving variations .
b. chow , s .- c .
chu , d. glickenstein , c. guenther , j. isenberg , t. ivey , d. knopf , p. lu , f. luo , and l. ni , _ the ricci flow : techniques and applications .
part i _ , mathematical surveys and monographs , vol .
135 , american mathematical society , providence , ri , 2007 , geometric aspects .
mr 2302600 j. l. vzquez , j. r. esteban , and a. rodrguez , _ the fast diffusion equation with logarithmic nonlinearity and the evolution of conformal metrics in the plane _ , adv .
differential equations * 1 * ( 1996 ) , no . 1 , 2150 . | in this note , we complete the classification of the geometry of non - compact two - dimensional gradient ricci solitons . as a consequence ,
we obtain two corollaries : first , a complete two - dimensional gradient ricci soliton has bounded curvature .
second , we give examples of complete two - dimensional expanding ricci solitons with negative curvature that are topologically disks and are not hyperbolic space . |
spin injection , transport , and detection are key factors in the field of magnetoelectronics . especially , magnetization reversal using spin - polarized current is of great interest @xcite due to its potential technological applications such as mram @xcite , spin transistor @xcite or spin battery @xcite . to understand and optimize spin - transport behavior in such devices ,
it is important to know a current distribution in it . particularly for mram applications
, we have to know spin - current magnitude and distribution to optimize the current density necessary for spin - injection induced magnetization reversal .
variety of formalisms calculating magnetoelectronic transport in one dimension ( 1d ) even for non - collinear magnetization has been proposed @xcite . however , up to now
, the spatial ( 3d ) calculation of the spin - polarized current have been missing . to obtain spatial distribution of spin - polarized current ( and spin accumulation ) in a given structure , we express such a structure as a 3d circuit of spin - dependent - resistor - elements ( sdre ) , wherein the propagation is regarded as 1d problem @xcite . resistor circuit network has been already used in @xcite to simulate current lines in metal / insulator multilayers . however , in this case , they did not use spin - polarized current .
ichimura et al .
@xcite have calculated 2d distribution of spin polarized current for co / al lateral spin valve structure .
they did not use the electric circuit , but directly solved poison equation by means of finite element method .
this article is organized as follows : sections [ s : bases ] and [ s : int ] present a matrix approach to calculate 1d diffusive propagation of spin - polarized current and potential in the single sdre .
section [ s : multilayer ] shows how this formalism can be applied to a simple multilayer structure .
this approach is just compact matrix rewriting the 1d valet - fert ( vf ) formalism @xcite .
section [ s : net ] explains how to solve general electric circuit and section [ s : grid ] tells how to divide nanostructure into circuit of sdres . finally ,
in section [ s : cuco ] we apply our calculations to ( cu / co)@xmath0 , and ( cu / co)@xmath1 pillar structures , where cross - sectional areas of the first and last layers are assumed either the same as the pillar , or infinitely large .
we show how the presence of such infinitely large layers influence currents and spin accumulation profiles .
finally section [ s : cr1 dm ] shows how to modify vf formalism to describe influence of the infinitely large layers .
in the diffusive transport regime , equations deriving the spatial distribution of electrochemical potential @xmath2 and spin - polarized current density @xmath3 inside ferro or non - magnetic metals are @xcite @xmath4 \label{eq : diffj } \vek{j}_{\up/\dw}=-\frac{\sigma_{\up/\dw}}{e}\nabla \mu_{\up/\dw},\end{gathered}\]]where @xmath5 is spin - flip diffusion length , @xmath6 conductivities for up and down channels , respectively and @xmath7 the electron charge assumed to be @xmath8 in this article .
our model is based on the circuit of sdre , consisting of layers and interfaces ( fig .
[ f : sdre ] ) .
we first express response of a single layer .
when eqs .
( [ eq : diffmu])([eq : diffj ] ) are solved in 1d , the profiles of @xmath9 and @xmath10 ( hereafter denoted as @xmath11 ) inside a given material along @xmath12-axis is given by @xcite @xmath13
\\ \pm d(x ) \frac{\sigma}{\sigma_{\up/\dw } } \exp[l/\lambda],\end{gathered}\]]where @xmath14 is charge current density and positive current direction is towards positive @xmath12-direction .
energies @xmath15 and @xmath16 are amplitudes of exponential dumping of @xmath11 and the energy @xmath17 is the asymptotic electrochemical potential equivalent to weighted average of electrochemical potentials @xmath18 .
the energies @xmath19 , @xmath20 and @xmath21 will be determined later from boundary conditions .
the relation between @xmath22 , @xmath23 and @xmath19 , @xmath20 , @xmath21 used in eqs .
( [ eq : diffj ] ) and ( [ eq : defmu ] ) can be expressed in a compact matrix form @xmath24 \vek{h}=\mat{d}\cdot\vek{f } , \label{eq : defd}\end{gathered}\]]where @xmath25 is up / down spin - polarized current with cross - section area @xmath26 of a given layer , @xmath27 is a charge current flowing through layer . the scaling factor @xmath28 which has no physical meaning ,
is introduced to adjust units of newly defined @xmath29 and @xmath30 vectors . for numerical calculations , value of @xmath28 should adjust order of @xmath22 and @xmath31 . when expressing any measurable quantities ( e.g. @xmath11 , @xmath32 , resitances ) , they obviously do not depend on @xmath28 . the left side of eq .
( [ eq : defd ] ) ( @xmath29-vector ) contains variables , which are conserved across interfaces or electric nodes . on the other hand , the right side of eq .
( [ eq : defd ] ) ( @xmath30-vector ) contains variables , which are used to calculate propagation of the @xmath32 and @xmath11 through layers [ eq .
( [ eq : defmu ] ) ] .
hence , the dynamic @xmath33-matrix relates coefficients at the boundary to coefficients for propagation .
this approach is well - known in the optics of anisotropic media @xcite .
propagation of @xmath30-vector through layer with the thickness ( `` length '' ) @xmath34 is expressed by a propagation @xmath35-matrix [ eq.([eq : defmu ] ) ] @xmath36 \displaystyle \vek{f}_{x+l}=[\mat{p}]^{-1}\cdot\vek{f}_x .
\label{eq : defp}\end{gathered}\]]hence , the relation between @xmath37 $ ] at both ends of the layer is expressed as @xmath38 in the previous section , we have expressed response of one layer in homogeneous material . here
, we describe the interfacial properties including ( i ) interface resistance @xcite @xmath39 and ( ii ) interfacial spin - flip scattering , described by a shunting resistance @xmath40 , `` shortcutting '' up and down channels at the interface .
the response of interface can be expressed by @xmath41 matrix @xmath42 where @xmath43 denotes for @xmath29-vector just below / above interface and @xmath44 writes @xmath45notice that alternative interfacial spin - flip scattering by @xmath46-parameter was introduced in @xcite describing spin relaxation at the interfaces by a thin interfacial layer of spin - flip - length @xmath47 where @xmath48 is interfacial layer thickness . from comparison of @xmath49-matrices from eqs.([eq : defk ] ) and ( [ eq : kint ] ) , @xmath50 where @xmath48 and @xmath51 are interfacial layer thickness and conductivity , respectively .
when interfacial spin - flip scattering does not exist , then @xmath52 .
although the formalism developed here is mainly to calculate @xmath11 , @xmath53 inside an electrical circuit , we shall first show how presented @xmath54 matrix algebra can be used to calculate electrical response of a multilayer structure [ fig .
[ f : sdre](c ) ] .
the description for a single layer by matrix @xmath49 in eq . ( [ eq : defk ] ) gives relationship between @xmath11 and @xmath32 at both ends of the layer .
hereafter , instead of continuity of spin - polarized current density @xmath53 over interface , we consider continuity of spin - polarized current @xmath32 , to take into account variable cross - sectional area @xmath26 of layers .
the reasons and validity are discussed later in section [ s : cr1 dm ] .
obviously , when all layers has the same @xmath26 , our formalism provides the same results as original vf formalism @xcite .
the boundary conditions at interfaces are continuity of @xmath11 and @xmath32 , i.e. continuity of @xmath29-vector .
consequently , the response of whole multilayer structure can be written as @xcite @xmath55where @xmath56^{-1 } \ , \mat{k}^{(0)}_\int\ , \mat{k}^{(1)}\ , \mat{k}^{(1)}_\int \ldots \\
\mat{k}_\int^{(m-1 ) } \mat{k}^{(m)}\ , \mat{k}^{(m)}_\int\ , \mat{d}^{(m+1 ) } \end{gathered}\ ] ] where upper index in parenthesis denotes for interface or layer number , @xmath57 the number of layers . because materials @xmath58 and @xmath59 are semi - infinite , @xmath11 inside them must not exponentially increase .
hence , some exponential terms in eq .
( [ eq : defmu ] ) must vanish , namely @xmath60 , @xmath61 .
hence , the vectors @xmath62 and @xmath63 are limited to form @xmath64 now , eq.([eq : defm1 ] ) can be solved .
substituting eq .
( [ eq : endf ] ) to eq .
( [ eq : defm1 ] ) , we found all unknowns in @xmath63 @xmath65where @xmath66 are elements of @xmath67-matrix .
the value of @xmath68 can be arbitrary value , as it only adds a constant to the profiles of @xmath11 . other simplification of eq .
( [ eq : ac ] ) follow from @xmath69 and @xmath70 , in consequence of ( i ) for @xmath71 the terms @xmath11 , @xmath19 must be constant and equal to each other and ( ii ) for @xmath71 , @xmath72 .
then , solution of eq .
( [ eq : ac ] ) can be written as @xmath73now , @xmath74 and @xmath75 are known and therefore the vector @xmath63 can be reconstructed from eq .
( [ eq : endf ] ) .
consequently , the profiles of @xmath11 , @xmath19 , @xmath32 etc . in the entire structure can be determined by recursive applying step - by - step matrix multiplications in eq .
( [ eq : defm ] ) finally , resistivity of the multilayer structure ( between first and last interface ) is @xmath76 .
as demonstrated in the previous section , the @xmath67-matrix [ defined by eq .
( [ eq : defm1 ] ) ] can characterize the entire multilayer structure where the same charge current @xmath77 flows across all layers . in this section ,
we extend the previous formalism to an electrical circuit ( network ) of sdres , mutually connected at nodes . in general , the sdre is composed of any sequence of materials ( layers ) and/or interfaces , as depicted on fig .
[ f : sdre ] .
there are three types of sdre , depending whether the length of sdre is finite or infinite : * close - end sdre [ fig .
[ f : sdre](a ) ] has finite length and hence both ends of this sdre are attached to nodes . because the boundary condition at nodes are described by @xmath29-vectors , whole sdre is described by @xmath78}$]-matrix relating @xmath29-vectors at both ends of sdre : @xmath79(0)}=\mat{k}^{[b]}\vek{h}^{[b](m_b)},\ ] ] where , for later purpose , @xmath78}$ ] can be rewritten into four @xmath80 submatrices @xmath81(0)}= \begin{bmatrix } \mat{k}_{\mu\mu } & \mat{k}_{\mu j } \\
\mat{k}_{j\mu } & \mat{k}_{jj } \end{bmatrix}^{[b ] } \begin{bmatrix } \vek{\mu}\\ u\vek{j } \end{bmatrix}^{[b](m_b)},\ ] ] where @xmath82 denotes a sdre number in the circuit and @xmath83 is the number of layers in @xmath82-th sdre , @xmath84(m_b)}\equiv [ \mu^{[b](m_b)}_{\up},\mu^{[b](m_b)}_{\dw}]^\mathrm{t}$ ] and @xmath85(m_b)}\equiv[j^{[b](m_b)}_{\up},j_\dw^{[b](m_b)}]^\mathrm{t}$ ] , @xmath86 denoting vector transposition .
hereafter , indices in square bracket denotes for sdre number , whereas indices in ordinary parentheses denotes for number of layer inside sdre .
positive current direction is from layer ( 1 ) to layer @xmath87 .
analogous to eq .
( [ eq : defm ] ) , @xmath78}$ ] consists of layer and interface contributions @xmath88}= \mat{k}^{[b](1)}\ , \mat{k}^{[b](1)}_\int \ldots \mat{k}_\int^{[b](m_b-1 ) } \mat{k}^{[b](m_b)}.\ ] ] when sdre contains only one layer ( i.e. it consists of single metal ) , then [ eq . ( [ eq : defk ] ) ] @xmath78}=\mat{k}^{[b](1)}\equiv \mat{d}^{[b]}\mat{p}^{[b]}[\mat{d}^{[b]}]^{-1}$ ] . * open - end sdre [ fig .
[ f : sdre](b ) ] has one end `` finite '' and connected to the node .
the other end is infinitely long and at its end the charge current @xmath89(0)}$ ] , flowing into the @xmath82-th open - end sdre , is applied .
because boundary conditions on node are described by @xmath29-vector and boundary conditions of the propagation into infinity by @xmath30-vector , the @xmath82-th open - end sdre is described by @xmath90}$]-matrix as @xmath91(0)}=\mat{z}^{[b]}\vek{h}^{[b](m_b)},\ ] ] where we have used convention that direction of sdre and positive current direction goes from infinite end of sdre toward its finite end [ fig .
[ f : sdre](b ) ] .
analogous to eqs .
( [ eq : defm ] ) and ( [ eq : defk0 ] ) @xmath92}=[\mat{d}^{[b](0)}]^{-1 } \mat{k}_\int^{[b](0 ) } \mat{k}^{[b](1)}\ , \mat{k}^{[b](1)}_\int \ldots \\ \mat{k}_\int^{[b](m_b-1 ) } \mat{k}^{[b](m_b)}. \end{gathered}\ ] ] when the sdre contains no layer ( i.e. it contains only single material continuous to infinity ) , then @xmath90}=[\mat{d}^{[b]}]^{-1}$ ] .
* multilayer structure [ fig .
[ f : sdre](c ) ] described in sect .
[ s : multilayer ] may be understood as a special type of sdre , having both ends open - ended . in general ,
the electric circuit is assumed to have @xmath93 nodes , @xmath94 close - end sdre and @xmath95 open - end sdre .
the boundary conditions valid for each node are determined by generalized kirchoff s laws : @xmath96}=\mathrm{const}_{n,\up/\dw}\equiv \mu_{n,\updw } , \\
\label{eq : kirj } \sum_{b_n=1}^{b_n } j_{n,\up/\dw}^{[b_n]}=0,\end{gathered}\]]i.e .
the values of @xmath11 has to be identical for all ends of sdre connected to each node and sum of polarized currents @xmath32 entering each node has to be zero .
subscript @xmath97 denotes for node number , @xmath98 and @xmath99 denotes the sdres connected to the @xmath97-th node .
hereafter , we use two @xmath100-notation ( i ) @xmath101 relating @xmath11 at the @xmath97-th node and ( ii ) @xmath102(0)}_\updw$ ] , @xmath102(m_b)}_\updw$ ] denoting @xmath11 at start , end of the @xmath82-th sdre , respectively . the relation between these two @xmath100-notations is given by connections between nodes and sdres , i.e. by the topology of the electric circuit .
following the previous discussion , the problem is how to treat large number of linear equations ( [ eq : closeend])([eq : openend])([eq : kirmu])([eq : kirj ] ) giving relations between @xmath11 and @xmath32 on the nodes and between ends of sdres .
we do it by solving one large matrix expression , @xmath103to which we apply all the previously mentioned rules .
the @xmath104-vector contains @xmath101 for all nodes and also @xmath105(m_b)}$ ] , @xmath106 , at the _ end _
( i.e. at the @xmath83-side ) of each sdre .
hence , @xmath104-vector has @xmath107 elements .
the role of the @xmath108-matrix consists of three parts [ eq.([eq : qex ] ) ] : * to relate @xmath11 , @xmath32 at the `` finite '' end of the open - end sdres and between boundary conditions of propagation towards infinity .
this is provided by @xmath109 linear equation , and hence this part occupies @xmath109 rows in @xmath108-matrix . * to relate @xmath11 between start and end of the close - end sdres ( @xmath110 linear equations ) .
* to realize current conservation at each node ( @xmath111 linear equations ) .
these contributions are studied in detail in following . in total
, @xmath108-matrix has @xmath107 rows , the same as length of @xmath29-vector .
so @xmath108 is a square matrix .
an example of a circuit of sdres is depicted in figure [ f : netsch ] .
the circuit consist of @xmath112 nodes , connected by sdre .
the circuit consists of @xmath113 close - end sdre and @xmath114 open - end sdres .
then , the resulting equation @xmath115 looks like : @xmath116}_\mu & \mat{0 } & \mat{0 } & \tilde{\mat{z}}^{[1]}_j & \mat{0 } & \mat{0 } & \mat{0 } & \mat{0 } \\ \mat{0 } & \mat{0 } & \tilde{\mat{z}}'^{[5]}_\mu & \mat{0 } & \mat{0 } & \mat{0 } & \mat{0 } & \tilde{\mat{z}}'^{[5]}_j \\
\hline -\mat{1 } & \mat{k}^{[2]}_{\mu,\mu } & \mat{0 } & \mat{0 } & \mat{k}_{\mu j}^{[2 ] } & \mat{0 } & \mat{0 } & \mat{0 } \\
\mat{0 } & -\mat{1 } & \mat{k}^{[3]}_{\mu,\mu } & \mat{0 } & \mat{0 } & \mat{k}_{\mu j}^{[3 ] } & \mat{0 } & \mat{0 } \\
-\mat{1 } & \mat{0 } & \mat{k}^{[4]}_{\mu,\mu } & \mat{0 } & \mat{0 } & \mat{0 } & \mat{k}_{\mu j}^{[4 ] } & \mat{0 } \\ \hline \mat{0 } & \mat{k}^{[2]}_{j\mu } & \mat{k}^{[4]}_{j\mu } & -\mat{1 } & \mat{k}^{[2]}_{jj } & \mat{0 } & \mat{k}^{[4]}_{jj } & \mat{0 } \\
\mat{0 } & \mat{0 } & \mat{k}^{[3]}_{j\mu } & \mat{0 } & -\mat{1 } & \mat{k}^{[3]}_{jj } & \mat{0 } & \mat{0 } \\
\mat{0 } & \mat{0 } & \mat{0 } & \mat{0 } & \mat{0 } & -\mat{1 } & -\mat{1 } & -\mat{1 } \end{array } \right ] % h - matrix \begin{bmatrix } \vek{\mu}_1
\\ \vek{\mu}_2 \\ \vek{\mu}_3 \\ \hline u\vek{j}^{[1](m_1 ) } \\
u\vek{j}^{[2](m_2 ) } \\
u\vek{j}^{[3](m_3 ) } \\
u\vek{j}^{[4](m_4 ) } \\
u\vek{j}^{[5](m_5 ) } \end{bmatrix } = % f - matrix \begin{bmatrix } \tilde{\vek{f}}^{[1](0 ) } \\
\tilde{\vek{f}}'^{[5](0 ) } \\
\hline \vek{0}\\ \vek{0}\\ \vek{0 } \\
\hline \vek{0 } \\\vek{0 } \\ \vek{0 } \end{bmatrix}\ ] ] where @xmath117^\mathrm{t}$ ] and @xmath118 , @xmath119 denote @xmath80 unitary , zero matrix , respectively .
each row which is shown in @xmath108-matrix in eq .
( [ eq : qex ] ) represents two rows ( for up and down channel ) , so hereafter we call it `` double - row '' .
( ad 1 ) the role of the first part in the @xmath108-matrix relates both ends of open - end sdre , described by eq .
( [ eq : openend ] ) , @xmath120(0)}=\mat{z}^{[b]}\vek{h}^{[b](m_b)}$ ] . however , we only know 2 variables out of 4 in vector @xmath120(0)}= [ \a^{[b](0 ) } , c^{[b](0 ) } , d^{[b](0 ) } , uj_\ch^{[b](0 ) } ] ^\mathrm{t}$ ] . we know
charge current @xmath89(0)}$ ] entering @xmath82-th sdre and that @xmath121(0)}=0 $ ] , because @xmath11 must not exponentially increase towards infinity [ eq .
( [ eq : endf ] ) ] . hence ,
taking from @xmath90}$]-matrix [ eq.([eq : openend ] ) ] only rows corresponding with known value in @xmath120(0)}$ ] , we get @xmath122(0)}\!\!\equiv\!\ !
\begin{bmatrix } 0 \\ uj_\ch^{[b](0 ) } \end{bmatrix } \!\!=\!\ !
\begin{bmatrix } z_{21}\ ! & \!z_{22}\ ! & z_{23}\ ! & \!z_{24 } \\
z_{41}\ ! & \!z_{42}\ ! & z_{43}\ ! & \!z_{44 } \end{bmatrix}^{[b ] } \ ! \begin{bmatrix } \mu_\up \\
\mu_\dw \\ uj_\up \\
uj_\dw \end{bmatrix}^{[b](m_b ) } \\ \equiv \begin{bmatrix } \tilde{\mat{z}}^{[b]}_{\mu } & \tilde{\mat{z}}^{[b]}_{j } \end{bmatrix } \begin{bmatrix } \vek{\mu}^{[b](m_b ) } \\ u\vek{j}^{[b](m_b ) } \end{bmatrix}.\end{gathered}\]]this equation is substituted into first part of in @xmath108-matrix [ eq . ( [ eq : qex ] ) ] in the form : @xmath123}_\mu \vek{\mu}^{[b](m_b ) } + \tilde{\mat{z}}^{[b]}_j u \vek{j}^{[b](m_b ) } = \tilde{\vek{f}}^{[b](0)}$ ] . in the case of the last open - end sdre connected to the structure ( in our example sdre number @xmath124 ) , the situation is slightly different . in the vector @xmath125(0)}= [ \a^{[b'](0 ) } , c^{[b'](0 ) } , d^{[b'](0 ) } , uj_\ch^{[b'](0 ) } ] ^\mathrm{t}$ ] ( a ) we have to set a value of @xmath126(0)}$ ] , which sets an absolute value of all @xmath2 and @xmath19 inside circuit to an arbitrary value and ( b ) getting a charge current @xmath127(0)}$ ] would be redundant as a sum of all charge currents entering circuit has to be zero .
hence , the last double - row in the first part of the @xmath108-matrix ( second double - row in eq .
( [ eq : qex ] ) ) looks like @xmath128(0 ) } \!\!\equiv\!\ !
\begin{bmatrix } \a^{[b'](0)}\\ 0 \end{bmatrix } \!=\ !
\begin{bmatrix } z_{11}\ ! & \!z_{12}\ ! & \!z_{13}\ !
& \!z_{14 }
\\ z_{21}\ ! & \!z_{22}\ ! & \!z_{23}\ ! & \!z_{24 } \end{bmatrix}^{[b ' ] } \ !
\begin{bmatrix } \mu_\up \\ \mu_\dw \\ uj_\up \\
uj_\dw \end{bmatrix}^{[b'](m_b ) } \\ \equiv \begin{bmatrix }
\tilde{\mat{z}}'^{[b']}_{\mu } & \tilde{\mat{z}}'^{[b']}_{j } \end{bmatrix } \begin{bmatrix } \vek{\mu}^{[b'](m_b ) } \\ u\vek{j}^{[b'](m_b ) } \end{bmatrix}\end{gathered}\]]and is substituted to @xmath108-matrix analogous as eq .
( [ eq : deftildez ] ) .
( ad 2 ) the second part in the @xmath108-matrix gives relation between @xmath129(0)}$ ] at the start of the @xmath82-th close - end sdre and @xmath129(m_b)}$ ] , @xmath130(m_b)}$ ] at the end of @xmath82-th sdre .
this relation is given by the first double row taken from eq .
( [ eq : closeend1 ] ) .
it is substituted into @xmath108 in the form @xmath131(0)}+\mat{k}^{[b]}_{\mu\mu}\vek{\mu}^{[b](m_b ) } + u\mat{k}^{[b]}_{\mu j}\vek{j}^{[b](m_b)}$ ] .
( ad 3 ) the last part in @xmath108-matrix describes the current conservation at each node , described by kirchoff s law eq .
( [ eq : kirj ] ) .
as specified , @xmath104-vector contains only values of current at the end of each sdre , @xmath132(m_b)}$ ] .
hence , if the @xmath82-th sdre starts at the @xmath97-th node ( as no . 2 and 4 sdres at the node 1 ) , then the current is expressed by the second double - row in eq .
( [ eq : closeend1 ] ) , from @xmath133(m_b)}$ ] , @xmath132(m_b)}$ ] as @xmath134(0 ) } = \mat{k}^{[b]}_{j\mu } \vek{\mu}^{[b](m_b ) } + u\mat{k}_{jj}^{[b]}\vek{j}^{[b](m_b)}$ ] . for example , in case of node @xmath135 , the current conservation is @xmath136(m_1)}+j_\updw^{[2](0)}+j_\updw^{[4](0)}=0 $ ] , which is substituted to @xmath108-matrix to the 6-th double - row as @xmath137(m_1 ) } + \mat{k}^{[2]}_{j\mu } \vek{\mu}^{[2](m_2 ) } + u\mat{k}_{jj}^{[2]}\vek{j}^{[2](m_2 ) } + \mat{k}^{[4]}_{j\mu } \vek{\mu}^{[4](m_4 ) } + u\mat{k}_{jj}\vek{j}^{[4](m_4)}=\mat{0}$ ] .
although the construction of the @xmath108-matrix as presented here may be tedious , it is rather direct to establish its construction numerically .
when the equation @xmath115 is solved , values of @xmath2 , @xmath32 for each sdre are directly written in @xmath104-vector ; their profiles can be determined by step - by step applying multiplication in eqs.([eq : closeend])([eq : defk0 ] ) and ( [ eq : openend])([eq : defz0 ] ) . in the previous section we have derived the formalism to calculate @xmath32 and @xmath11 in arbitrary electric circuit .
in this section , we explain , how to describe the whole nano - structure as a circuit of sdres , as sketched in fig . [
f : grid ] .
each part of the nanostructure is divided into 3d rectangular grid , the circuit nodes positions being @xmath138 , @xmath139 , @xmath140 .
then , a given sdre , for example in the @xmath12-direction , has length @xmath141 and cross - sectional area @xmath142 . due to this treatment
, the grid does not need to be equally spaced , but just rectangular . at the interface between two different materials , e.g. a and b ( see fig . [
f : grid ] ) , sdre is described by @xmath143 matrix consisting of three contributions @xmath144 [ eq .
( [ eq : defk0 ] ) ] .
the @xmath145 is contribution from grid point ( node ) inside a material to the interface with ( sdre is now along the z - direction ) @xmath146 and @xmath147 .
@xmath44 is the interface resistivity matrix given by eq.([eq : kint ] ) .
@xmath148 is contribution from the interface to the node in b material having @xmath149 and @xmath147 .
the electrical circuit network as described above describes correctly charge current .
however , inside 3d ( 2d ) sdre circuit the volume of metal in all resistors is three ( two ) times larger than in reality .
in such a case , the spin - polarized current would have larger dumping , as it would diffuse into larger volume .
to correct this , it is necessary to calculate 1d propagation of @xmath2 in each sdre by slightly modified eq .
( [ eq : diffmu ] ) @xmath150 where @xmath151 is a dimension of the sdre circuit .
in other words , when a given nanostructure is described by 3d ( 2d ) electrical circuit , the @xmath5 should be increased by a factor @xmath152 ( @xmath153 ) .
this @xmath5-normalization should not be applied for open - end sdres , as their contribution is correctly described by 1d propagation .
the advantage of this @xmath5-normalization is shown in figure [ f:2d3d ] , where @xmath154 through ( cu / co)@xmath0 structure ( described and studied in detail in the next section ) is compared between 1d model ( full line ) , and our 3d model without @xmath5-normalization ( @xmath155 ) and with @xmath5-normalization ( @xmath156 ) .
we can see that with ( without ) @xmath5-normalization , the agreement between 1d and 3d model is about 1% ( 30% ) .
the same is valid for spin - accumulation @xmath157 , where calculation without @xmath5-normalization leads to 30% smaller @xmath158 . in this section ,
we describe how to incorporate surface scattering to the presented formalism .
surface scattering can be described by a shunting resistance @xmath159 shortcutting up and down channels for a nodes situated just close to the wire surface .
@xmath159 has value @xmath160 , @xmath40 being surface scattering resistivity and @xmath161 being surface area corresponding to the @xmath97-th node .
when surface scattering is not presented , then obviously @xmath162 .
surface scattering can be described by modification of kirchoff law [ eq . ( [ eq : kirj ] ) ] @xmath163}\mp \frac{\mu_{n,\up}-\mu_{n,\dw}}{r_s}=0.\ ] ] to incorporate this modification into @xmath108-matrix for the @xmath97-th node , we add @xmath164 matrix @xmath165\ ] ] on the @xmath97-th double - column and @xmath97-th double - row in the last part of @xmath108-matrix which describes current conservation in each node . in our example given by eq .
( [ eq : qex ] ) and figure [ f : netsch ] , to add @xmath159 to @xmath166 node , we place @xmath164 to the 7-th double - row and 2-nd double - column into @xmath108-matrix .
in this section we use the above developed formalism on ( cu / co)@xmath0 and ( cu / co)@xmath1 pillar structures .
we will show how spin - polarized current density @xmath167 and electrochemical potential @xmath11 vary differently between when whole structure is a pillar or only a part of the structure is a pillar attached to an infinitely large continuous layer . in the literature ,
pillars terminated with infinitely large continuous layers are commonly used @xcite .
we have studied structures consisting of 2 and 3 co layers , called ( cu / co)@xmath0 , and ( cu / co)@xmath1 with dimensions in nm cu1/co1(40)/cu2(6)/co2(2)/cu3 and cu1/co1(40)/cu2(6)/co2(2)/cu3(6)/co3(20)/cu4 , respectively .
the square - shaped pillar 100 nm in size begins with the co1/cu2 interface .
the considered structure types are defined in fig .
[ f : types ] as ( a ) cross - sectional area s constant , ( b ) column , ( c ) constriction and ( d ) constriction with infinite co1 layer . in the following discussion , `` constriction '' corresponds to the case ( c ) .
the `` infinite '' homogeneous layers were approximated as a square pillar of 800 nm in size .
the magnetization of co1 layer is always fixed as `` up '' ( @xmath168 ) , whereas magnetic orientations of other co layers are varied .
note that in our diffusive transport calculations , the magnetization orientation with respect to the structure ( e.g. in - plane or out - of - plane ) do not play any role , but only mutual magnetization orientation ( parallel or antiparallel ) does play an important role .
the charge current passing through structure is assumed to be @xmath169ma , equivalent to the averaged charge current density in the pillar @xmath170a / m@xmath0 .
the electrical properties of materials are of room temperature @xcite : electric conductivity @xmath171@xmath172m@xmath173 , @xmath174@xmath172m@xmath173 , spin - flip - lengths @xmath175 nm , @xmath176 nm and co spin bulk assymetry @xmath177 .
we assume no interface resistance and no interface and surface scattering .
the sdre grid size is 10 nm .
figure [ f : crsp ] shows the profile of @xmath154 along center axis of the structures : ( a ) ( cu / co)@xmath0 ( b ) ( cu / co)@xmath1 with parallel co1 and co3 layers and ( c ) ( cu / co)@xmath1 with antiparallel co1 and co3 layers for s constant , column and constriction type structures . in all cases , @xmath154 for parallel co1 and co1 layer is larger than for antiparallel configuration .
furthermore , @xmath154 is enhanced at the position of free co2 layer for the column and constriction types compared to the s constant type structure .
for example , in the case of the constriction type structure with ( cu / co)@xmath0 @xmath178 configuration , @xmath154 is enhanced by a factor of 1.75 and in the case of ( cu / co)@xmath1 @xmath179 by 1.5 .
when only cu buffer layer is infinite in size ( column - type structure ) , @xmath154 at the position of co2 layer is enhanced too , but not so strongly .
figure [ f : crsp](c ) shows that for ( cu / co)@xmath1 with antiparallel co1 and co3 layers , @xmath154 is significantly reduced for all types of structures .
the origin of the @xmath154 enhancement is following : the pillar is attached to infinitely large cu layer which provides large volume for spin - current to be scattered and so acting as a strong spin - flip - scatterer .
hence , infinitely large cu layer works as a small shortcutting resistance between up and down channels .
consequently , shortcutting of up and down channels leads to an increase in @xmath154 .
the increase of @xmath154 is related with increase of spin - polarization efficiency @xmath180 , as charge current flowing though pillar @xmath77 is fixed in all our calculations .
consequently , increase of @xmath181 leads to decrease of critical switching current @xmath182 which is necessary to reverse magnetization direction of free layer .
as the constriction type structure is in the most common use for co / cu pillar structures , the @xmath154 enhancement ( i.e. @xmath182 reduction ) has been already widely used @xcite without being noticed .
similar effect can be realized by inserting a layer with small characteristic spin - flip resistance @xmath183 , such as pt , ag , au , ru above the last co layer or bellow the first co layer .
such cover layers has been used @xcite since the first pioneering experiment @xcite , but their contributions to the @xmath154 enhancement have been observed recently @xcite .
figure [ f : crch ] shows profiles of charge current @xmath14 along the center axis of the ( cu / co)@xmath1 structure for s constant , column and constriction type structures . obviously
, for s constant structure , the @xmath184 is constant . in the case of infinite cu termination
, @xmath154 decreases approximately exponentially over the characteristic length of 50 nm . the same decay of @xmath154 is presented in fig .
[ f : crsp ] for infinitely large cu layers . in table
[ t : tab ] , we summarize averadged values of @xmath154 in the position of co2 layer in all the types of studied structures
. these @xmath154 values may differ from those presented in fig .
[ f : crsp ] due to lateral inhomogenity of @xmath154 inside pillar , discussed in next section .
the largest averaged @xmath154 is obtained for ( cu / co)@xmath1 @xmath179 constriction structuture ( @xmath185a / m@xmath0 ) and ( cu / co)@xmath0 @xmath178 constriction structure ( @xmath186a / m@xmath0 ) , providing @xmath154 enhancement by a factor 2 with respect to ( co / cu)@xmath0 s - constant structure ( 14.85@xmath187a / m@xmath0 ) .
this tendency is already explained in above paragraph .
figure [ f : pl2 ] presents a map of spin - polarized current density @xmath154 inside ( cu / co)@xmath0 constriction type structure for ( a)(b ) parallel and ( c ) antiparallel magnetization configuration .
the case ( a ) is a structure without infinitely large co1 layer whereas ( b)(c ) with it .
a map of @xmath184 is not presented here as it looks similar to @xmath154 with @xmath178 .
the @xmath154 inside co2 layers is more homogenous and flows well perpendicular to the interfaces although in the surrounding cu layers the @xmath154 can have rather large inclination and inhomogenity .
the @xmath154 tends to be more homogeneous when passing co layer , causing that the @xmath154 in the adjacent cu layers can have rather large in - plane components .
this is remarkable in the case ( c ) .
very similar tendency is found for @xmath184 .
the above mentioned characteristics of @xmath154 are consequence of larger spin - flip resistance @xmath183 of co with respect to cu , @xmath188=14.3f@xmath189m@xmath0 , @xmath190=7.3f@xmath189m@xmath0 .
it means that spin - flip is more likely to occur inside cu than inside co. in other words , co is `` harder '' material than cu for @xmath154 to penetrate into it .
the above mentioned characteristics of @xmath184 are simply consequence of @xmath191 .
figure [ f : prftra ] shows cross - sectional ( in @xmath12-direction ) profiles of ( a ) @xmath154 and ( b ) @xmath184 through pillar in the middle of the co2 layer for ( cu / co)@xmath0 structure .
both @xmath154 and @xmath184 are inhomogenous having minima at the structure center .
this is due to inhomogenous current injection into the pillar from infinitely large layers .
there is no @xmath154 and @xmath184 inhomogenity for s constant and nearly no inhomogenity in case of column type structure as the @xmath154 and @xmath184 are homogenized by a co1 layer . for constriction type structure
the inhomogenity is 12% for @xmath178 and 5% for @xmath192 .
inhomogenity is defined as @xmath193 , where maximal , minimal value is taken from cross - sectional current distribution in fig . [
f : prftra ] .
the inhomogenity is reduced for @xmath192 due to different conductivities of up and down channel conductivities , leading to other current homogenization in cu2 spacer layer .
if co1 layer is infinitly large , the inhomogenity is increased to 10% and 29% for @xmath178 and @xmath192 , respectively .
the @xmath184 inhomogenity is 8% and 20% for constriction type structure with and without infinitely large co1 layer , respectively .
here remains a question whether such a @xmath154 inhomogenities make current magnetization reversal easier or not .
adventage of @xmath154 inhomogeneity may be , that it localy enhanced @xmath154 near the pillar edge , wheas decreasing @xmath154 at the pillar center .
as shown in table [ t : tab ] , the mean value of @xmath154 is about the same for ( cu / co)@xmath0 constriction structure with and without infinitely large co1 layer .
however , in the second case , @xmath154 is much more inhomogenous .
this is not well presented in fig .
[ f : prftra](a ) , we do not see the largest @xmath154 flowing in the vicinity of the corners of the square pillars .
disadventage of @xmath154 inhomogenity may be different magnitude of torgue exerted on magnetic spins of free co2 layer
. this may be particulary important in the case of high speed switching associated with magnetization precession .
figure [ f : mu ] presents profiles of @xmath9 , @xmath10 and @xmath19 ( hereafter @xmath100-profiles ) along the center axis of the ( a ) ( cu / co)@xmath0 @xmath178 , @xmath192 ( b ) ( cu / co)@xmath1 @xmath179 , @xmath194 and ( c ) ( cu / co)@xmath1 @xmath195 , @xmath196 .
s - constant structure is presented for both magnetization directions of co2 layers , although column and constriction types are presented only for @xmath168 magnetization of co2 layer .
in contrast to @xmath154 , inhomogenity of spin accumulation @xmath158 at the position of free co2 layer is very small , mostly bellow 1% .
table [ t : tab ] presents a mean value of @xmath158 for all types of the studied structures .
figures [ f : mu ] ( a)(b ) and ( c ) show that @xmath100-profiles depend slightly on magnetization of free co2 layer because @xmath197 .
figure [ f : mu](a ) exhibits suppression of @xmath158 in the vicinity of infinitely large cu layer .
the reason is exactly the same as discussed in sec .
[ s : cr - through ] : the infinitely large cu layer works as a strong spin - scatterer , causing a small spin - flip resistance ( large scattering ) between up and down channels .
obviously , such a shortcut reduces @xmath158 .
this is contradictory to @xcite , where it is argued that presence of spin - scatterer increases spin acululation @xmath158 inside pillar .
it should be emphesised that presence of spin - scatterers increases @xmath154 ( and magnetoresistance ) in the pillar , but _ reduce _
@xmath158 .
table [ t : tab ] shows that the largest @xmath158 is obtained for column type structure , by 20% larger than for s - constant type structure .
the reason is as follow : when cu1 is not infinitely large ( s constant structure ) , @xmath158 changes its sign approximately in the middle of co1 layer [ fig .
[ f : mu](a ) s constant ] .
when cu1 is infinitely large , it acts as strong spin - scatterer and shortcuts up and down channels .
hence , @xmath158 at cu1/co1 interface is nearly zero and hence larger @xmath158 is obtained at the co1/cu2 interface and inside co2 layer [ fig .
[ f : mu](a ) column ] . but to realize this , it is necessary that up and down channels above free co2 layer should not be shortcut meaning that the cover layer has not to be infinite .
figure [ f : mu](b ) shows that @xmath158 is reduced when co1 and co3 layers have parallel magnetization configuration . figure [ f : mu](c ) shows an increase in @xmath158 when co1 and co3 layers are antiparallel , enhancing @xmath158 by factor of 2 with respect to ( cu / co)@xmath0 s constant structure . in this case
the types of structure are not so important as in ( cu / co)@xmath0 case , because @xmath158 at the position of co2 free layer is `` screened '' by spin - scattering inside co1 and co3 layers .
hence , there is effectively no spin - accumulation @xmath158 in the case of a commonly used constriction type structure with two fm layers .
it may explain why this contribution to magnetization reversal ( predicted in @xcite ) was not observed in co / cu structure @xcite . to obtain non - zero @xmath158
, it is necessary to use either 3 fm layer system with antiparallel configuration of first and last fm layers , or to ensure that structure above free layer will not contain spin - scatterers .
it can be reached when pillar structure above free layer ( i ) does not contain any strong spin - scatterer layers ( as au , ag , pt , ru ) and ( ii ) the cover layer does not contain large volume of metal .
it means that pillar current drain should be realized by a long pillar or by a thin cover layer .
finally , we discuss influence of type structure on magnetoresistance ratio ( mr ) , presented in the last column in table [ t : tab ] .
the value of mr is determined as mr@xmath198 , where @xmath168 , @xmath199 denotes magnetization of free co2 layer , @xmath200 , where @xmath201 and @xmath202 are determined on cu side of the first and last cu / co interface , respectively .
the calculated value of mr in the case of ( cu / co)@xmath0 s - constant structure is 0.48% .
however , in the case of ( cu / co)@xmath0 constriction type , it reachs 1.01% ( enhancement by a factor of 2 ) and in the case of the constriction with infinite co1 layer even 1.43% ( enhancement by a factor of 3 ) .
the last value may be misleading as this increase is mainly due to resistivity reduction of co1 infinite layer .
table [ t : tab ] shows mr is strongly affected by the type structure , increasing with an increase of @xmath154 rather than @xmath158 .
this tendency may be observed also for ( co / cu)@xmath1 layer , showing increase ( decrease ) in mr when co1 and co3 layers have parallel ( antiparallel ) magnetization configuration .
figure [ f : jr](a ) shows dependance of @xmath154 on @xmath203 , where @xmath204 , where @xmath168 , @xmath199 means up , down magnetization of co2 layer , respectively .
it shows that @xmath154 is proportional to @xmath205 .
this figure is analogous to the experimental dependence of critical switching current @xmath182 as a function of @xmath205 @xcite , where they found that @xmath206 is proportional to @xmath205 .
these graphs are analogous from following reasons : to switch magnetic layer we need to overcome the critical spin - polarized current density @xmath207 where @xmath182 is critical switching charge current and @xmath181 is spin - polarized current injection efficiency . because @xmath208 and @xmath209 are constant , so @xmath210 . on the other hand , from our calculations
, we can express @xmath181 as @xmath211 , where @xmath212 is a fixed charge current flowing to the structure ; therefore @xmath213 .
however , in our model , @xmath154 is slightly larger for co1 and co2 parallel configuration ( fig .
[ f : crsp ] ) , although in @xcite , @xmath206 is larger for antiparallel configuration , in agreement with the spin - transfer model @xcite .
figure [ f : jr](a ) also provides an important consequence : when changing electric properties of surroundings of fm(fixed)/spacer / fm(free ) layers , an enhancement of @xmath154 in the position of free layer leads to enhancement of mr .
figure [ f : jr](b ) shows dependence of @xmath158 on @xmath205 .
we can see that with increasing @xmath205 , @xmath158 is reduced .
two different slopes correspond to two different `` sources '' @xmath158 acting as with different `` hardness '' .
the source hardness is determined by a presence or absence of scatterers bellow fixed co1 layer .
hard source is for constriction and column structure type , i.e. when cu1 is infinitely large .
weak one is for s constant structure type , i.e. when for cu1 is not infinitely large .
the explanation of this behavior has been already provided in previous section [ s : mu ] : when spin - flip scatterer is presented bellow fixed co1 layer , it vanishes @xmath158 on cu1/co1 interface and hence provides harder source of @xmath158 .
figure [ f : jr ] also shows that when changing surroundings of fm(fixed)/spacer / fm(free ) layers , an increase in @xmath158 at the position of free layer is related to a decrease of mr .
remark , that for constriction type structure with co1 layer infinitely large [ case(d ) in fig .
[ f : types ] ] , the `` source '' of @xmath158 becomes softer than for constriction type [ case ( c ) in fig .
[ f : types ] ] .
as we have shown , the constriction type structures modify the profile of both @xmath154 and @xmath214 with respect to s constant types , which provides equivalent results to 1d vf formalism @xcite .
however , to describe @xmath154 and @xmath11 inside structure by 3d model may be tedious procedure .
that is why here we discuss briefly , how to modified vf model to take into account infinitely large layers .
we propose , that each layer can have its own cross - sectional area @xmath215 .
then , instead of conserving spin polarized current density @xmath53 in the 1d formalism , we propose to conserve spin polarized current @xmath32 .
actually , we have used this boundary condition already in section [ s : multilayer ] .
this boundary condition is justified when ( i ) thickness of the layer @xmath216 is thick enough compared to the change of pillar diameter between neighboring layers , @xmath217 , @xmath218 , where @xmath219 , so that current has enough space to spread to different cross - sectional area @xmath215 ( ii ) pillar diameter @xmath220 is smaller than spin - flip length @xmath5 . it can be shown that if @xmath221 , then @xmath53 , @xmath11 , mr etc .
so , it is not important , how large value @xmath215 is used to describe infinitely large layers .
figure [ f:1 dm ] shows @xmath154 and @xmath11 calculated by 3d models in the case of constriction type structure . these data are compared with 1d models with and without expanding the terminating cu layers , called `` 1d constriction '' and `` s constant '' , respectively .
3d constriction is described by 1d constriction ( s constant ) model with precision of 15% ( 50% ) for @xmath154 and 40% ( 80% ) for @xmath158 [ fig .
[ f:1dm](a)(b ) ] .
the agreement between @xmath158 for constriction type structure is quite poor , because @xmath158 at this configuration is very small . in the case of @xmath158 larger than the above case
, the agreement is about 10% .
this 1d model has been used to calculate @xmath154 , @xmath158 and mr in the position of free co2 layer for all types of studied structures .
the results are presented in table [ t : tab ] in square brackets ; we can see rather good agreement in all cases .
an exception is the case of constriction type structure with infinitely large co1 layer , as in this case the condition ( ii ) is not fulfilled .
we have developed formalism which allows to calculate spin - polarized current @xmath154 , and electrochemical potential @xmath11 inside arbitrary electric circuit , consisting of ferro- or non - magnetic metallic sdre elements as well as interface and surface resistivities .
the formalism is limited to the parallel / antiparallel magnetic orientation in diffusive regime . to calculate spatial distribution of @xmath11 , @xmath53 inside nanostructure
, we divide the structure into an 1d , 2d or 3d electric circuit network of sdre elements which is successively solved . when division is carried out as described in section [ s : grid ] , the renormalization of spin - flip length @xmath5 has to be performed .
this formalism is applied to ( cu / co)@xmath0 and ( cu / co)@xmath1 , pillar structures , where pillar cross - sectional area of starting / terminating layers were assumed to be either infinitely large ( column type , constriction type ) or they have the same cross - sectional area as pillar ( s const types ) . inside the pillar surrounded by infinitely
large layers , the @xmath154 , @xmath184 can be inhomogeneous .
maximal inhomogeneity is found to be 29% and 20% in the case of infinitely large co1 layer . on the other hand , profile of @xmath158
is found more homogeneous , with found maximum inhomogeneity of 2% .
such @xmath154 inhomogeneities may locally enhance the value of @xmath154 , but they may disturb the magnetization reversal associated with spin precession . due to @xmath191 , @xmath222 , both @xmath184 and @xmath154
flow rather perpendicular to the co layers and furthermore the presence of co layer makes @xmath154 , @xmath184 more homogeneous .
when such a spin - scatterers ( but it is also valid for different spin - scatterers , as layers of au , ag , pt , ru , etc . )
are introduced bellow `` fixed '' co1 layer , they make @xmath158 source `` harder '' . when they are placed above free co2 layer , they shortcut up and down channels , reducing @xmath158 nearly to zero , and hence enhancing @xmath154 . consequently , to get maximum @xmath154 in the case of ( cu / co)@xmath0 at the position of `` free '' co2 layer , it is important to introduce spin - scatterers both bellow fixed co1 layer and above free co2 layer .
to get maximum @xmath158 , it is important to introduce spin - scatterer bellow fixed co1 layer , but reduce spin - scattering above co2 free layer , latter one can be realized by reducing volume of material above co2 layer , for example current drain can be thin long nanowire or cover layer with very thin thickness . as most of experimentally studied ( cu / co)@xmath0 structures have infinitely large layers at both ends , @xmath158 inside them is nearly zero . for ( cu / co)@xmath1 system , maximum @xmath154 ( @xmath158 )
is for parallel ( antiparallel ) magnetization of first , last co layer . when applying above described optimizations , @xmath154 and @xmath158 can be further enhanced .
furthermore , in agreement with experimental results @xcite we found that @xmath154 is linearly proportional to @xmath223 ( and hence mr is increased when increasing @xmath154 at the position of free co2 layer ) .
furthermore , dependence of @xmath224 is also linear , but @xmath158 is reduced when increasing @xmath205 ( when increasing mr ) . | we present a formalism determining spin - polarized current and electrochemical potential inside arbitrary electric circuit within diffusive regime for parallel / antiparallel magnetic states . when arbitrary nano - structure is expressed by 3-dimensional ( 3d ) electric circuit , we can determine 3d spin - polarized current and electrochemical potential distributions inside it .
we apply this technique to ( cu / co ) pillar structures , where pillar is terminated either by infinitely large cu layer , or by cu wire with identical cross - sectional area as pillar itself .
we found that infinitely large cu layers work as a strong spin - scatterers , increasing magnitude of spin - polarized current inside the pillar twice and reducing spin accumulation nearly to zero .
as most experimentally studied pillar structures are terminated by such a infinitely large layers , we propose modification of standard valet - fert formalism to simply include influence of such infinitely large layers . |
when does a graph @xmath0 have an orientation such that every out - degree is at most @xmath4 ? an obvious necessary condition is that @xmath5 for every subgraph @xmath6 . indeed , suppose @xmath0 has such an orientation and @xmath6 .
inside @xmath7 the sum of out - degrees equals the number of edges .
moreover , each vertex contributes at most @xmath4 to this sum , hence the condition .
hakimi proved that this condition is in fact sufficient .
[ thm : hakimi ] let @xmath0 be a graph and @xmath8 an integer
. then @xmath0 has an orientation such that every vertex has out - degree at most @xmath4 if and only if all subgraphs @xmath6 satisfy @xmath5 .
in fact , hakimi proved a slightly stronger result that determines all possible out - degree sequences produced by orientations of a given graph @xmath0 .
his proof uses induction on the number of edges , but the weaker statement given above is a straightforward consequence of hall s marriage theorem .
what about hypergraphs ?
suppose an @xmath1-uniform hypergraph @xmath0 ( i.e. a family of @xmath1-sets ) is given an _ orientation _ , by which we mean that for each edge @xmath9 one ordering of the vertices of @xmath9 is chosen .
this ordering is called the _ orientation of @xmath9_. note that each edge has exactly @xmath10 possible orientations . if @xmath11 then this coincides with the usual definition of graph orientation .
we will often denote an orientation of @xmath0 by @xmath12 and the corresponding orientation of an edge @xmath9 by @xmath13 .
given an orientation @xmath12 , a vertex @xmath14 and @xmath15 = \left\{1 , 2 , \dots , r\right\}$ ] , the _ @xmath16-degree _ of @xmath14 , written @xmath17 , is the number of edges @xmath9 such that @xmath14 is in the @xmath16-th position of @xmath13 . note that if @xmath11 then @xmath18 is the out - degree and @xmath19 is the in - degree of @xmath14 .
when does an @xmath1-uniform hypergraph @xmath0 have an orientation such that @xmath20 for every vertex @xmath14 ? again , an obvious necessary condition is that @xmath5 for every subgraph @xmath6 ( where as usual @xmath7 is a subgraph of @xmath0 if @xmath21 and @xmath22 ) .
indeed , inside @xmath7 the sum of @xmath18 over all vertices @xmath14 of @xmath7 is equal to the number of edges of @xmath7 , and is at most @xmath4 times the number of vertices .
caro and hansberg showed that this condition is sufficient .
[ thm : hakforhyp ] let @xmath0 be an @xmath1-uniform hypergraph and @xmath8 an integer
. then @xmath0 has an orientation such that @xmath20 for all vertices @xmath14 if and only if all subgraphs @xmath6 satisfy @xmath5 .
they proved it by constructing a suitable maximal flow on @xmath7 , and a simple proof via hall s marriage theorem is also possible .
now , in contrast to the situation for graphs , for oriented hypergraphs there is a sensible notion of degree for sets of _ multiple _ vertices .
for example , given an orientation @xmath12 and a pair of vertices @xmath23 , we can define @xmath24 to be the number of edges @xmath9 such that @xmath25 and @xmath14 ( in some order ) are in the first two positions of @xmath13 .
so if the oriented edges are @xmath26 ( where the vertex set is @xmath27 $ ] ) then @xmath28 . more generally , for a @xmath2-set of vertices @xmath29 and a @xmath2-set @xmath30 $ ] , the _ @xmath31-degree _ of @xmath32 , denoted by @xmath33 , is the number of edges @xmath9 such that the elements of @xmath13 in positions labeled by @xmath31 are @xmath34 ( in some order ) .
more formally , @xmath33 is the number of edges @xmath9 such that if we write @xmath35 then @xmath36 is exactly the set @xmath32 .
[ we mention in passing that there is a variant of this notion where the mutual order of @xmath25 and @xmath14 is important .
however , this alternative definition turns out to be less interesting for the types of questions examined by caro and hansberg and by us . in this paper
we mainly consider the former notion ( of ` unordered ' degrees ) , but a brief analysis of the latter is given in section [ sec : ordereddegrees ] . ] in @xcite caro and hansberg asked if a result similar to their theorem [ thm : hakforhyp ] can be found in the setting of degrees for multiple vertices .
[ qn : genhak ] fix integers @xmath8 and @xmath37 .
which @xmath1-uniform hypergraphs @xmath0 have an orientation such that @xmath38}(a ) \le k$ ] for all @xmath2-sets of vertices @xmath39 ?
again there is an obvious necessary condition : if @xmath0 has such an orientation then for each collection of @xmath2-sets @xmath40 at most @xmath41 edges @xmath9 satisfy @xmath42 ( here we write @xmath43 for the set of all @xmath2-subsets of @xmath44 ) .
indeed , given @xmath45 , every @xmath9 with @xmath42 contributes @xmath46 to the sum @xmath47}(a)$ ] , and this sum is at most @xmath41 . our first aim is to answer the question of caro and hansberg by showing that this condition is sufficient .
note that case @xmath48 is theorem [ thm : hakforhyp ] .
thmhakforpsets [ thm : hakforpsets ] let @xmath0 be an @xmath1-uniform hypergraph and @xmath49 integers
. then @xmath0 has an orientation such that @xmath38}(a ) \le k$ ] for all @xmath50 if and only if for each @xmath40 at most @xmath41 edges @xmath9 satisfy @xmath42 .
caro and hansberg s interest in theorem [ thm : hakforhyp ] was to answer the following question . when does an @xmath1-uniform hypergraph @xmath0 have an orientation such that for every vertex @xmath14 there is _ some _
@xmath15 $ ] such that @xmath51 ?
it is once again easy to obtain a necessary condition : such an orientation would partition the vertices into sets @xmath52 where each @xmath14 is put into some @xmath53 with @xmath51 .
theorem [ thm : hakforhyp ] applied to the induced subgraphs @xmath54 $ ] gives the necessary condition @xmath5 for all @xmath55 $ ] .
they proved that this condition is sufficient .
[ thm : alldeg ] let @xmath0 be an @xmath1-uniform hypergraph and @xmath8 an integer
. then @xmath7 has an orientation such that for each vertex @xmath14 some @xmath56 satisfies @xmath51 if and only if @xmath57 can be partitioned into @xmath1 sets @xmath52 such that for each @xmath58 and each @xmath59 at most @xmath41 edges @xmath9 satisfy @xmath60 . for degrees of multiple vertices , the first case @xmath61 is already non - obvious .
when can an @xmath1-uniform hypergraph @xmath0 be given an orientation such that for any @xmath2-set of vertices @xmath32 there is _ some _
@xmath2-set @xmath30 $ ] such that @xmath62 ?
such an orientation would partition @xmath63 into @xmath64 sets @xmath65{^{(p)}},$ ] where each @xmath50 is put into @xmath66 with @xmath62 . for any @xmath31 and any edge @xmath9 the @xmath2-set of vertices that are in positions labeled by @xmath31 in @xmath13 must not belong to @xmath66 .
so there are no edges whose @xmath2-sets would all belong to a single @xmath66 , giving a necessary condition .
caro and hansberg asked if , similarly to the case @xmath48 , this condition is sufficient .
[ qn : orient - part ] let @xmath37 be integers and @xmath0 an @xmath1-uniform hypergraph .
suppose @xmath63 is partitioned into sets @xmath67 , where @xmath68 , in such a way that for any @xmath69 there are no edges @xmath9 such that @xmath70 .
must @xmath0 have an orientation such that for any @xmath2-set of vertices @xmath32 there is some @xmath2-set @xmath30 $ ] such that @xmath62 ?
our main aim of the paper is to show that the answer to question [ qn : orient - part ] is positive for @xmath1 much larger than @xmath2 , but negative in general .
it will turn out in this paper that the following notion is crucial to understanding question [ qn : orient - part ] .
we wish to determine if a function @xmath71^{(p ) } } \to { [ n]^{(p)}}$ ] must ` fix an intersection ' in the sense explained by the next definition .
let @xmath72 be integers and @xmath71^{(p ) } } \to { [ n]^{(p)}}$ ] a function .
we say that @xmath73 _ fixes an intersection _ if there are distinct @xmath74^{(p)}}$ ] such that @xmath75 .
moreover , if every nonconstant function @xmath71^{(p ) } } \to { [ n]^{(p)}}$ ] fixes an intersection then we say that @xmath76^{(p)}}$ ] has the _ fixed intersection property_. it is not difficult to see that if @xmath77 then @xmath78^{(p)}}$ ] does not have the fixed intersection property .
this can be demonstrated by choosing @xmath79 $ ] , @xmath80 { \setminus}[p ] = \{p+1 , \dotsc , 2p\}$ ] and defining @xmath81^{(p ) } } \to { [ 2p]^{(p)}}$ ] by @xmath82 this @xmath73 is nonconstant and does not fix an intersection .
we conjecture that @xmath83 ( for @xmath77 ) is actually the only exceptional case .
[ conj : main ]
let @xmath72 be integers with @xmath84
. then @xmath76^{(p)}}$ ] has the fixed intersection property . in section [ sec : hypergraphs ]
we show that , if conjecture [ conj : main ] is true for given @xmath2 and @xmath85 , then question [ qn : orient - part ] can be answered positively ( for the same choice of @xmath2 with @xmath86 ) . in section [ sec : ramsey ] we show conjecture [ conj : main ] is true when @xmath1 is much larger than @xmath2 .
thmconjpart [ thm : conjpart ] for every integer @xmath87 there is some @xmath88 such that if @xmath89 then @xmath76^{(p)}}$ ] has the fixed intersection property .
the answer to question [ qn : orient - part ] is positive if @xmath1 is sufficiently large , given @xmath2 .
what if @xmath90{^{(p)}}$ ] does not have the fixed intersection property ?
the simplest such case is @xmath91 .
the main part of this paper is devoted to showing how the trivial failure of the fixed intersection property of @xmath92{^{(2)}}$ ] can be ` lifted ' to a failure of question [ qn : orient - part ] for the case @xmath93 .
our work relies on a new version of the ` amalgamation ' technique in sparse ( structural ) ramsey theory .
thmnogen [ thm : nogen ] there is a @xmath94-uniform hypergraph @xmath7 satisfying : a. [ it : nilhak ] for every orientation of @xmath7 there is some pair of vertices @xmath23 such that @xmath95 for all @xmath96{^{(2)}}$ ] ; b. [ it : sixpart ] there is a partition of @xmath97 into six sets @xmath98 such that @xmath99 is not contained in a single @xmath100 for any edge @xmath9 . in particular ,
the answer to question [ qn : orient - part ] is negative for @xmath93 .
one such hypergraph is constructed in section [ sec : sparseramsey ] . in section [ sec : ordereddegrees ] we give a brief analysis of the ` ordered ' notion of degrees . finally , in section
[ sec : openproblems ] we suggest some open problems .
in this section we prove theorem [ thm : hakforpsets ] and give a positive answer to question [ qn : orient - part ] for the choices of @xmath1 and @xmath2 such that @xmath101{^{(p)}}$ ] has the fixed intersection property .
we start with an easy proof of theorem [ thm : hakforpsets ] .
construct a bipartite graph @xmath7 with vertex classes @xmath102 and @xmath103 .
join @xmath104 and @xmath105 by an edge if @xmath106 .
in other words , we join each edge of @xmath0 to its @xmath2-subsets .
given an orientation @xmath12 on @xmath0 , let @xmath107 be the set of edges in @xmath7 of the form @xmath108 where @xmath32 is the set of vertices in the first @xmath2 positions of @xmath13 , that is , @xmath109 where @xmath110 .
note that @xmath107 covers every element of @xmath44 exactly once and , conversely , every subset of @xmath111 that covers every element of @xmath44 exactly once is given by some choice of @xmath12 . moreover ,
each @xmath105 is covered by @xmath107 exactly @xmath38}(a)$ ] times .
therefore by hall s marriage theorem , @xmath0 has an orientation such that all @xmath112 satisfy @xmath38}(a ) \le k$ ] if and only if @xmath113 for all @xmath114 , where @xmath115 is the neighbourhood of @xmath116 in @xmath7 .
this can be restated as @xmath117 for all @xmath114 and @xmath118 with the property that @xmath119 .
having chosen @xmath45 , the largest @xmath116 such that @xmath119 is @xmath120 .
hence , this statement is equivalent to @xmath121 for all @xmath122 .
now we switch our attention to question [ qn : orient - part ] .
we are trying to give an @xmath1-uniform hypergraph @xmath0 an orientation such that for each @xmath2-set of vertices _ some _ @xmath31-degree is zero .
in particular , if @xmath9 is an edge and @xmath106 is a @xmath2-set , we require @xmath123 for some @xmath124 .
so we need a tool that would give @xmath9 an orientation when for each @xmath2-set @xmath106 there is a @xmath2-set @xmath125 $ ] such that the vertices @xmath32 can not take exactly the positions labeled by @xmath124 .
of course , no such orientation exists if all sets @xmath124 are equal ( some @xmath2 vertices must be in positions labeled by that set ) . however , if the sets @xmath124 are not all equal then it turns out that such an orientation exists provided @xmath101{^{(p)}}$ ] has the fixed intersection property .
[ lem : orient - edge ] let @xmath37 be integers such that @xmath101{^{(p)}}$ ] has the fixed intersection property and let @xmath9 be an @xmath1-set . for each @xmath2-set @xmath106 choose a @xmath2-set @xmath125 $ ] .
if the chosen @xmath2-sets @xmath124 are not all equal then the elements of @xmath9 can be ordered in such a way that the positions labeled by @xmath124 are not taken exactly by the elements of @xmath32 for any @xmath2-set @xmath106 . without loss of generality
assume that @xmath126 $ ] .
then we have a nonconstant function @xmath127^{(p ) } } \to { [ r]^{(p)}}$ ] that sends @xmath128^{(p)}}$ ] to @xmath124 .
let @xmath129 be a permutation of @xmath101 $ ] , chosen uniformly at random .
for each @xmath2-set @xmath130 $ ] , @xmath131 with probability @xmath132
. therefore @xmath133 \right ] & \le \sum_{a \in { [ r]^{(p ) } } } { \mathbb{p}}\left [ \sigma(a ) = i_a \right ] \\ & = 1 \end{aligned}\ ] ] with equality if and only if the events @xmath131 are pairwise disjoint . since @xmath127^{(p)}}$ ] has the fixed intersection property , @xmath134 for some distinct @xmath2-sets @xmath135 $ ] .
hence there is a permutation @xmath136 of @xmath101 $ ] such that @xmath137 and @xmath138 , so the events @xmath139 and @xmath140 are not disjoint , and so the above inequality is strict .
therefore there is a choice of @xmath129 such that @xmath141 for any @xmath32 .
the required ordering of @xmath9 is @xmath142 for this choice of @xmath129
. this immediately gives a positive answer to question [ qn : orient - part ] for the case when @xmath127^{(p)}}$ ] has the fixed intersection property .
[ prop : partthenorient ] let @xmath37 be integers such that @xmath127^{(p)}}$ ] has the fixed intersection property and let @xmath0 be an @xmath1-uniform hypergraph
. suppose @xmath63 is partitioned into sets @xmath67 , where @xmath68 , in such a way that for any @xmath69 there are no edges @xmath9 such that @xmath70
. then @xmath0 has an orientation such that for any @xmath2-set of vertices @xmath32 there is some @xmath2-set @xmath30 $ ] such that @xmath62 .
relabel the sets @xmath67 with @xmath2-subsets of @xmath101 $ ] ( instead of integers @xmath143 ) .
given an edge @xmath9 of @xmath0 , associate to each @xmath2-set @xmath106 the @xmath2-set @xmath125 $ ] where @xmath144 . by assumption
not all of these @xmath2-sets are equal , so lemma [ lem : orient - edge ] can be applied to get an orientation @xmath13 that does not contribute to @xmath145 for any @xmath32 .
do this for every edge to get an orientation of @xmath0 . for any @xmath2-set of vertices
@xmath32 , @xmath124 does not depend on the choice of @xmath9 , and @xmath123 .
can the requirement for @xmath127^{(p)}}$ ] to have the fixed intersection property be removed ? as we shall see in section [ sec : sparseramsey ] , proposition would not be true for which is in fact the smallest case when @xmath127^{(p)}}$ ] does not have the fixed intersection property .
therefore some condition on @xmath1 and @xmath2 must be imposed .
it might be interesting to know if some weaker condition would be enough .
on the other hand , this result can be strengthened to match the form of theorem [ thm : alldeg ] .
we fix an integer @xmath8 and ask if @xmath0 can be given an orientation such that for every @xmath2-set of vertices some @xmath31-degree is at most @xmath4 . with theorem [ thm : hakforpsets ] , it is once again it is easy to find a sufficient condition
. let @xmath8 , @xmath37 be integers and @xmath0 an @xmath1-uniform hypergraph .
suppose that @xmath0 has an orientation such that for each @xmath2-set of vertices @xmath32 there is some @xmath2-set @xmath30 $ ] such that @xmath146 .
then @xmath63 can be partitioned into @xmath147 sets @xmath67 such that for each @xmath58 and each @xmath148 at most @xmath41 edges @xmath149 satisfy @xmath42 .
partition @xmath63 into @xmath150 disjoint sets @xmath151^{(p)}},$ ] by putting each into @xmath66 such that @xmath146 . for
each @xmath31 apply theorem [ thm : hakforpsets ] to the induced subgraph @xmath152 $ ] .
( to make the definition of induced subgraphs precise , given @xmath153 , @xmath154 $ ] is the @xmath1-uniform hypergraph with vertex set @xmath155 and edge set .
informally , we keep only those edges whose @xmath2-subsets are in @xmath116 . )
the argument used for the case @xmath156 together with theorem [ thm : hakforpsets ] is enough to prove sufficiency when @xmath127^{(p)}}$ ] has the fixed intersection property .
[ prop : partthenorient - genk ] let @xmath49 be integers such that @xmath127^{(p)}}$ ] has the fixed intersection property and let @xmath0 be an @xmath1-uniform hypergraph .
suppose @xmath63 is partitioned into sets @xmath67 , where @xmath68 , in such a way that for each @xmath58 and each @xmath148 at most @xmath41 edges @xmath149 satisfy @xmath42 .
then @xmath0 has an orientation such that for each @xmath2-set of vertices @xmath32 there is some @xmath2-set @xmath30 $ ] such that @xmath146 .
relabel the sets @xmath67 with @xmath2-subsets of @xmath101 $ ] ( instead of integers @xmath143 ) . by theorem [ thm : hakforpsets ] we can give orienations to the edges of the induced subgraphs @xmath152 $ ] , @xmath157^{(p)}}$ ] , in such a way that , only counting these edges , @xmath146 whenever @xmath158 .
now , same as in the proof of proposition [ prop : partthenorient ] , lemma [ lem : orient - edge ] allows us to give orientation to the remaining edges of @xmath0 without increasing @xmath33 for any @xmath158 .
here we prove theorem [ thm : conjpart ] which says that @xmath76^{(p)}}$ ] has the fixed intersection property for pairs @xmath159 with @xmath85 much larger than @xmath2 .
we start by extending the definitions given in the introduction to slightly greater generality .
let @xmath87 be an integer and @xmath160 sets .
we say that a function @xmath161 _ fixes an intersection _ if there are distinct @xmath162 such that @xmath75 .
moreover , we say that @xmath163 has the _ fixed intersection property _ if every nonconstant function @xmath164 fixes an intersection .
here is a technical lemma .
[ lem : technical ] let @xmath87 be an integer and @xmath165
. if @xmath166 is a nonconstant function that is constant on @xmath167 for some @xmath168 with @xmath169 then @xmath73 fixes an intersection .
suppose for contradiction @xmath73 does not fix an intersection .
note that @xmath170 as @xmath73 is nonconstant .
write @xmath171 and @xmath172 .
suppose @xmath173 where @xmath174 .
then either @xmath175 or @xmath176 for all @xmath177 . in the latter case ,
pick @xmath177 such that @xmath178 .
this is possible because @xmath179 .
then @xmath75 , so @xmath73 fixes an intersection , contradicting our assumption .
therefore for any @xmath180 as defined above and any @xmath177 we have @xmath175 .
this means that @xmath73 is constant on @xmath181 where @xmath182 .
now write @xmath183 and repeat the argument to conclude that @xmath73 is constant on @xmath184 where @xmath185 .
repeat it again to show that @xmath73 is constant on @xmath186 for every @xmath85 where @xmath187 are nested sets that cover @xmath107 .
but this implies @xmath73 is constant on @xmath163 . contradiction .
we use this lemma to get the result for infinite domains .
[ lem : innat ] for any positive integer @xmath2 , @xmath188 has the fixed intersection property .
use induction on @xmath2 .
clearly , theorem holds for @xmath48 so suppose @xmath77 .
let @xmath189 be a nonconstant function . without losing generality
assume that @xmath190 ) = [ p]$ ] .
now @xmath191 $ ] for @xmath192 defines a finite colouring of @xmath188 where colours are subsets of @xmath193 $ ] . by ramsey s theorem
there is an infinite set @xmath194 such that @xmath195 is monochromatic on @xmath167 .
say , @xmath196 $ ] for all @xmath197 . if @xmath198 pick any @xmath199){^{(p)}}$ ] . then @xmath200| = 0 = |f(x ) \cap f([p])|$ ] , so @xmath73 fixes an intersection .
if @xmath201 $ ] then @xmath202 $ ] for all @xmath197 but this can not happen by lemma [ lem : technical ] . now suppose @xmath203 $ ] and write @xmath204 . for fixed @xmath205
define @xmath206 by @xmath207 for all @xmath208 . as @xmath73 is not constant on @xmath2-sets of @xmath209 by lemma [ lem : technical ] ,
there is a choice of @xmath210 for which @xmath211 is not constant .
then the induction hypothesis implies that @xmath212 for some distinct @xmath213 and so @xmath214 .
a compactness argument extracts the result for finite domains .
[ cor : infrang ] let @xmath87 be an integer . then there is some @xmath215 such that every nonconstant function @xmath216{^{(p ) } } \to { \mathbb{n}}{^{(p)}}$ ] fixes an intersection .
suppose not .
then for any integer @xmath217 there is a nonconstant function @xmath218{^{(p ) } } \to { \mathbb{n}}{^{(p)}}$ ] that does not fix an intersection . by reordering @xmath219 , if necessary , we can assume that @xmath220{^{(p ) } } ) \subset \left[p \binom{s}{p } \right]{^{(p)}}$ ] for each @xmath217 and each @xmath221 .
we define @xmath189 as follows . for any @xmath222 , there are only finitely many choices for @xmath223 on @xmath224{^{(p)}}$ ] so , in particular , infinitely many @xmath223 agree on @xmath193 $ ] . define @xmath190)$ ] to be @xmath225)$ ] for any such @xmath85 . among these @xmath223 ,
infinitely many agree on @xmath226{^{(p)}}$ ] .
define @xmath73 on @xmath226{^{(p)}}$ ] to be the same as any such @xmath223 . among these @xmath223 ,
infinitely many agree on @xmath227{^{(p)}}$ ] and we define @xmath73 on @xmath227{^{(p)}}$ ] to agree with these @xmath223 .
continue , at each step restricting to an infinite subsequence of the @xmath223 .
we get @xmath73 that , for every @xmath228 , agrees with some @xmath229 on @xmath230{^{(p)}}$ ] . in particular , @xmath73 is nonconstant since @xmath231 is nonconstant on @xmath232{^{(p)}}$ ] by lemma [ lem : technical ] . by lemma [ lem : innat ] , @xmath75 for some distinct @xmath233 . but @xmath234{^{(p)}}$ ] for some @xmath228 so @xmath235 , hence @xmath229 fixes an intersection . contradiction .
we get theorem [ thm : conjpart ] as an immediate corollary .
in light of conjecture [ conj : main ] and proposition [ prop : partthenorient ] , we seek @xmath77 such that question [ qn : orient - part ] has a negative answer when @xmath236 .
it turns out that @xmath237 works .
we recall the exact statement that we will prove .
the choice @xmath237 allows us to consider the elements of @xmath238 as edges and non - edges of a graph on vertices @xmath239 . with this idea in mind
we will deduce theorem [ thm : nogen ] from a statement about graphs rather than hypergraphs .
[ defn : special ] take six colours and partition them into three pairs . call two colours _ opposite _ if they are in the same pair .
suppose @xmath32 is a @xmath94-clique of a graph whose edges are coloured with these six colours .
we say that @xmath32 is _ special _ if there is a pair of opposite colours @xmath240 such that two independent edges of @xmath32 have colour @xmath241 and the remaining four edges have colour @xmath242 .
[ lem : main ] there is a graph @xmath0 satisfying : a. [ it : nomono ] edges of @xmath0 can be coloured with six colours without forming a monochromatic @xmath94-clique ; b. [ it : monoorspecial ] whenever the edges of @xmath0 are coloured with six colours there is a monochromatic @xmath94-clique or a special @xmath94-clique .
let @xmath0 be a graph as given by lemma [ lem : main ] .
form a @xmath94-uniform hypergraph @xmath7 on the vertices of @xmath0 by taking the @xmath94-cliques of @xmath0 as its edges ( so @xmath243 and @xmath244 ) .
property ( [ it : nomono ] ) of lemma [ lem : main ] for @xmath0 directly implies property ( [ it : sixpart ] ) of theorem [ thm : nogen ] for @xmath7 .
suppose for contradiction @xmath7 does not have property ( [ it : nilhak ] ) of theorem [ thm : nogen ] , that is @xmath7 has an orientation @xmath245 such that for any pair of vertices @xmath23 there is some @xmath246^{(2)}}$ ] such that @xmath247 .
in particular , there is an edge colouring @xmath248^{(2)}}$ ] such that @xmath249 for all @xmath149 . as is evident from the proof of lemma [ lem : orient - edge ] ( or by a simple check ) , the induced colouring @xmath250^{(2)}}$ ] on any @xmath94-clique @xmath251 fixes an intersection .
partition the set of colours @xmath252^{(2)}}$ ] into three pairs , each consisting of two disjoint @xmath253-sets .
let these be the pairs of opposite colours .
property ( [ it : monoorspecial ] ) of lemma [ lem : main ] for @xmath0 implies that some @xmath94-clique @xmath251 is monochromatic or special under @xmath195 . if we identify the vertices of @xmath32 with @xmath254 then on @xmath32 @xmath195 induces a mapping @xmath255^{(2 ) } } \to { [ 4]^{(2)}}$ ] .
if @xmath32 is monochromatic then this mapping is @xmath256^{(2)}}\ ] ] under some permutation of @xmath254 .
if @xmath32 is special then under some permutation the mapping is @xmath257 in both cases @xmath258 does not fix an intersection , giving a contradiction .
this finishes the proof .
the proof of lemma [ lem : main ] is based on a new amalgamation method in sparse ramsey theory .
amalgamation ( also known in literature as _ partite construction _ ) was introduced by neetil and rdl in @xcite .
there have been many more applications of this technique .
see , for example , @xcite or @xcite for general introduction . here is a brief outline of the method .
suppose we wish to find a graph @xmath0 that is sparse ( e.g. , does not contain a copy of some fixed graph ) but has the property that every colouring of its _ vertices _ with @xmath4 colours contains a monochromatic copy of some fixed graph @xmath7 .
we start with a sparse @xmath259-partite graph @xmath260 with the property that if its vertices are coloured so that every vertex class is monochromatic then there must be a monochromatic copy of @xmath7 .
then a larger @xmath259-partite graph @xmath0 is formed by taking many copies of @xmath260 and glueing them together in a specific way ( this process usually consists of several steps ) .
the goal is to ensure that @xmath0 is as sparse as @xmath260 and that any vertex colouring of @xmath0 contains a copy of @xmath260 with monochromatic vertex classes ( and therefore contains a monochromatic copy of @xmath7 ) .
a variant of this technique is useful in finding sparse @xmath0 with the property that every colouring of its _ edges _ with @xmath4 colours contains a monochromatic copy of @xmath7 .
the main difference is that we start with a sparse @xmath259-partite graph with the property that if its edges are coloured so that every pair of vertex classes spans a monochromatic bipartite graph then there must be a monochromatic copy of @xmath7 .
our result is of slightly different type .
the sparsity condition ( [ it : nomono ] ) in lemma [ lem : main ] is standard but the ramsey property ( [ it : monoorspecial ] ) is not .
the difference is that instead of requiring a monochromatic structure that is smaller than the forbidden structure we require a structure of the same size but not necessarily monochromatic . to adjust for this , instead of working with @xmath259-partite graphs we use graphs that are ` almost ' @xmath259-partite .
more specifically , vertices are partitioned into @xmath259 sets that span sparse rather than empty subgraphs .
let @xmath0 and @xmath7 be graphs and let @xmath4 be a positive integer .
we say that @xmath0 is _ @xmath4-edge - ramsey for _ @xmath7 if every colouring of the edges of @xmath0 with @xmath4 colours contains a monochromatic copy ( not necessarily induced ) of @xmath7 .
similarly , @xmath0 is _ @xmath4-vertex - ramsey for _ @xmath7 if this holds for every colouring of the vertices of @xmath0 with @xmath4 colours .
note that @xmath0 containing @xmath7 as a subgraph is equivalent both to @xmath0 being @xmath46-edge - ramsey for @xmath7 and to @xmath0 being @xmath46-vertex - ramsey for @xmath7 .
any @xmath259-partite graph ( where @xmath259 is a fixed positive integer ) is considered as having its vertex classes listed in a fixed order .
more formally , any such @xmath0 is a pair @xmath261 where @xmath262 is the list of vertex classes and @xmath263 is the set of edges of @xmath0 .
we write @xmath264 to mean @xmath262 when there is no danger of confusion .
the order of vertex classes is respected when standard operations on graphs are carried out .
for example , * if @xmath265 are a family of @xmath259-partite graphs labeled by @xmath266 then their union is @xmath267 provided @xmath268 when @xmath269 .
so the @xmath16-th vertex class of the union of graphs is the union of @xmath16-th vertex classes ; * if @xmath0 and @xmath7 are @xmath259-partite graphs then @xmath0 contains @xmath7 as a subgraph if there is an injective graph homomorphism @xmath270 that maps each vertex class of @xmath7 into the corresponding vertex class of @xmath0 ; + contains @xmath7 . ] * if @xmath0 and @xmath7 are @xmath259-partite graphs then @xmath0 is a copy of @xmath7 if there is a graph isomorphism @xmath270 that maps each vertex class of @xmath7 onto the corresponding vertex class of @xmath0 .
let @xmath4 be a fixed positive integer .
given a @xmath259-partite graph @xmath271 for any @xmath272 we can construct a @xmath259-partite graph @xmath273 , called the _
@xmath58-th amalgam of _ @xmath0 .
the main property of @xmath273 is that whenever its vertices are coloured with @xmath4 colours there must be a copy of @xmath0 with all vertices in @xmath100 of the same colour .
we construct @xmath273 as follows .
let @xmath274 be any set with @xmath275 .
for any @xmath276 let @xmath277 be a copy of @xmath0 whose @xmath58-th vertex class is exactly the set @xmath32 . choose these copies in such a way that they would intersect only at the @xmath58-th vertex class ( so @xmath278 for any @xmath279 ) . now define @xmath280 to be the union of all these copies of @xmath0 .
-th amalgam . ]
the pigeonhole principle implies that whenever the vertices of @xmath273 are coloured with @xmath4 colours , some set @xmath276 is monochromatic . then @xmath277 is a copy of @xmath0 in @xmath273 with the @xmath58-th vertex class monochromatic .
furthermore , it is not difficult to see that if for some @xmath222 there are no @xmath222-cliques in @xmath0 then there are none in @xmath273 either . we will need a slightly stronger result . [
prop : amal - noter ] let @xmath281 be an edge colouring without a monochromatic copy of @xmath282 . then there is an edge colouring @xmath283 without a monochromatic copy of @xmath282 . in other words , if for fixed @xmath4 and @xmath222 @xmath0 is not @xmath4-edge - ramsey for @xmath282 then neither is @xmath273 .
every edge @xmath9 of @xmath273 is an edge of @xmath277 for a unique @xmath276 .
let @xmath284 where on the right hand side @xmath9 is considered to be an edge of @xmath285 .
suppose for contradiction @xmath286 produces a monochromatic copy of @xmath282 on vertices @xmath287 . at most one of them
is in @xmath274 as it is an independent set so we can assume @xmath288 .
now @xmath289 for a unique @xmath32 and we must have @xmath290 for some @xmath291 since otherwise the vertices @xmath287 would span a monochromatic copy of @xmath282 in @xmath292 under the colouring induced by @xmath195 , contradicting our assumption on @xmath195 .
but then @xmath293 is not an edge of @xmath273 which is absurd .
[ lem : v - vs - e ] let @xmath294 and @xmath4 be positive integers and suppose @xmath7 is a graph that is not @xmath295-edge - ramsey for @xmath282 .
then there is a graph @xmath0 such that a. @xmath0 is @xmath4-vertex - ramsey for @xmath7 , and b. @xmath0 is not @xmath295-edge - ramsey for @xmath282 . define @xmath296 and take a @xmath259-partite graph @xmath297 that is a union of vertex disjoint copies of @xmath7 such that for any choice of @xmath298 vertex classes of @xmath297 some copy of @xmath7 intersects all of them .
note that if the vertices of @xmath297 are coloured with @xmath4 colours and each vertex class is monochromatic then there must be a monochromatic copy of @xmath7 . since @xmath7 is not @xmath295-edge - ramsey for @xmath282 , neither is @xmath297 .
take @xmath299 . by proposition [ prop : amal - noter ] , @xmath0 is not @xmath295-edge - ramsey for @xmath282 .
however , suppose the vertices of @xmath0 are coloured with @xmath4 colours . from previous observations we know that @xmath0 contains a copy @xmath300 of @xmath301 with the @xmath259-th vertex class monochromatic .
moreover , @xmath300 is @xmath4-vertex - coloured so contains a copy @xmath302 of @xmath303 with the @xmath304 vertex class monochromatic ( so in fact , both the @xmath259-th and the @xmath304 vertex classes of @xmath302 are monochromatic ) .
repeat this argument to deduce that @xmath0 contains a copy of @xmath297 with all vertex classes monochromatic .
our construction of @xmath297 now ensures that there is a monochromatic copy of @xmath7 .
a special case is particularly useful .
[ cor : v ] let @xmath222 and @xmath4 be positive integers and suppose @xmath7 is a graph that contains no copy of @xmath282 .
then there is a graph @xmath0 such that a. @xmath0 is @xmath4-vertex - ramsey for h , and b. @xmath0 does not contain @xmath282 as a subgraph .
this is lemma [ lem : v - vs - e ] with @xmath305 .
a slightly more involved application of the method produces the following result .
it is folklore , but for completeness we outline its proof as described in @xcite .
[ lem : e - vs - ks ] let @xmath222 and @xmath4 be positive integers and suppose @xmath7 is a graph that does not contain a copy of @xmath282 .
then there is a graph @xmath0 such that a. @xmath0 is @xmath4-edge - ramsey for @xmath7 , and b. @xmath0 does not contain @xmath282 as a subgraph .
again , we will start with a @xmath259-partite graph @xmath297 but this time choose @xmath306 , that is @xmath259 is the smallest integer such that whenever the edges of @xmath307 are coloured with @xmath4 colours we are guaranteed to get a monochromatic copy of @xmath308 .
construct @xmath297 by taking vertex disjoint copies of @xmath7 in such a way that whenever @xmath298 vertex classes of @xmath297 are taken there is a copy of @xmath7 that intersects them all .
note that if the edges of @xmath297 are coloured with @xmath4 colours and if every pair of vertex classes spans a monochromatic bipartite graph then there is a monochromatic copy of @xmath7 .
we will apply a slighty modified process of amalgamation to @xmath297 . given a @xmath259-partite graph @xmath309 and integers @xmath310 ,
define a @xmath259-partite graph @xmath311 as follows .
let @xmath312 $ ] be the bipartite graph induced by @xmath313 on vertex classes @xmath53 and @xmath100 .
let @xmath314 be a bipartite graph on vertex classes @xmath44 and @xmath315 such that whenever the edges of @xmath314 are coloured with @xmath4 colours , one of its induced subgraphs isomorphic to @xmath316 ( with @xmath317 and @xmath318 ) is monochromatic .
it is not obvious that @xmath314 exists but we defer the explanation to the end of the proof . for every induced subgraph @xmath319 that is isomorphic to @xmath316 ( with @xmath317 and @xmath318 ) , take a copy @xmath277 of @xmath313 making sure that @xmath320 = a$ ] and that the copies are disjoint everywhere else . finally , define @xmath321 .
then @xmath311 does not contain @xmath282 as a subgraph ( taking induced copies of @xmath316 in the previous step was crucial ) and whenever the edges of @xmath311 are coloured with @xmath4 colours we are guaranteed to get a copy of @xmath313 with the edges joining vertex classes @xmath53 and @xmath100 having the same colour . list all pairs @xmath322 , @xmath310 , in some order @xmath323 where @xmath324 .
define @xmath325 if the edges of @xmath0 are coloured with @xmath4 colours then @xmath0 is bound to have a copy of @xmath297 every pair of vertex classes spanning a monochromatic bipartite subgraph , and hence a monochromatic copy of @xmath7 .
moreover , @xmath0 does not contain @xmath282 as a subgraph .
it remains to prove the existence of @xmath314 .
for this we use the well known hales jewett theorem ( see @xcite ) . let the vertex classes be @xmath326 and @xmath327 where @xmath85 is a large integer . join @xmath328 and @xmath329 by an edge if @xmath330 are edges in @xmath316 .
so if @xmath331 is the set of edges of @xmath316 then @xmath314 has edges @xmath332 . by the hales
jewett theorem , for sufficiently large @xmath85 every colouring of @xmath332 with @xmath4 colours has a monochromatic combinatorial line .
it is not difficult to see that a combinatorial line in @xmath332 corresponds to induced copy of @xmath316 .
we proceed by proving a proposition which we will use to extend the technique to non - partite graphs . before stating it we introduce a simple construction .
given graphs @xmath0 and @xmath7 , define the _ pair graph _ of @xmath0 and @xmath7 to be the graph obtained by taking vertex disjoint copies of @xmath0 and @xmath7 and joining each vertex in the copy of @xmath0 to each vertex in the copy of @xmath7 by an edge .
denote this graph by @xmath333 and call the edges joining the copy of @xmath0 to the copy of @xmath7 _
intermediate_. [ prop : technical ] let @xmath4 and @xmath222 be positive integers and suppose @xmath297 and @xmath334 are graphs that do not contain @xmath282 .
then there are graphs @xmath0 and @xmath7 such that a. @xmath0 and @xmath7 do not contain @xmath282 , and b. whenever the intermediate edges of @xmath333 are coloured with @xmath4 colours , there is a copy of @xmath335 with @xmath336 , @xmath337 and all intermediate edges monochromatic . by corollary [ cor :
v ] , there is a graph @xmath7 that is @xmath4-vertex - ramsey for @xmath334 but contains no @xmath222-cliques .
there is also a graph @xmath0 that is @xmath338-vertex - ramsey for @xmath297 but contains no @xmath222-cliques .
let @xmath195 be any colouring of the intermediate edges of @xmath333 with colours @xmath339 $ ] .
it induces a vertex colouring @xmath340^{v(h)}$ ] given by @xmath341 . by our choice of @xmath0 ,
some copy @xmath336 is monochromatic under @xmath342 . in other words ,
the colour of an intermediate edge of @xmath343 only depends on its endpoint in @xmath7 and not on the one in @xmath297 .
now define a vertex colouring @xmath344 $ ] by letting each vertex of @xmath7 have the colour of any edge joining it to @xmath297 .
some copy @xmath337 is monochromatic under @xmath345 which means exactly that all intermediate edges of @xmath335 have the same colour .
let @xmath260 be a fixed graph and suppose that for each vertex @xmath346 a graph @xmath347 is chosen , and write @xmath348 for the collection of chosen graphs .
define the _
@xmath349-blowup of @xmath260 _ , denoted by @xmath350 , as follows .
* each vertex @xmath14 of @xmath260 is replaced by a copy of @xmath347 and these copies are pairwise vertex disjoint .
* for every edge @xmath351 of @xmath260 , each vertex of @xmath352 is joined to each vertex of @xmath347 by an edge .
call such edges _ @xmath351-intermediate_.
* for every nonedge @xmath351 of @xmath260 there are no edges between @xmath352 and @xmath347 .
note that the previously introduced pair graph is a special case of a blowup : @xmath333 is the same as @xmath353 .
lemma [ lem : e - vs - ks ] implies the existence of a graph @xmath331 that is @xmath354-edge - ramsey for @xmath355 but contains no copies of @xmath356 .
set @xmath357 , that is , a copy of @xmath331 is chosen for each vertex @xmath346 .
enumerate the edges of @xmath260 as @xmath358 and consider them one by one in this order .
suppose @xmath359 is considered and we have a list of graphs @xmath360 . by proposition [ prop : technical ] , there are graphs @xmath361 and @xmath362 such that a. @xmath361 and @xmath362 contain no copies of @xmath356 , and b. whenever the intermediate edges of @xmath363 are coloured with six colours , we can find copies @xmath364 and @xmath365 with all intermediate edges of @xmath366 of the same colour .
replace @xmath367 by @xmath361 and @xmath368 by @xmath362 to obtain a new list @xmath369 .
do this for all edges in turn and use the final list @xmath370 to form the blowup @xmath371 .
take any colouring @xmath195 of the edges of @xmath0 with six colours . by the choice of @xmath0
, it contains a copy of @xmath372 with the @xmath373-intermediate edges monochromatic .
furthermore , the edges of this copy are also coloured with six colours and so it contains a copy of @xmath374 with the @xmath375-intermediate edges monochromatic ( so in fact , both the @xmath373-intermediate and the @xmath375-intermediate edges are monochromatic ) .
repeat this argument to conclude that there is a copy of @xmath376 with the @xmath9-intermediate edges monochromatic for each edge @xmath377 .
moreover , the edges of the copy of @xmath331 which corresponds to a vertex @xmath346 are coloured with six colours so it contains a monochromatic @xmath355 . having this in mind
, @xmath195 induces a colouring @xmath378 of both the edges and the vertices of @xmath260 ( so @xmath378 is a total colouring ) with six colours where @xmath379 notice that if @xmath378 contains 1 .
[ item : e - k4 ] a @xmath356 with all edges of the same colour , or 2 .
[ item : ve ] a vertex and an incident edge of the same colour then @xmath195 contains a monochromatic @xmath356 . moreover , if @xmath378 contains 1 .
[ item : eopep ] an edge @xmath9 both of whose endpoints have the colour opposite to the colour of @xmath9 then @xmath195 contains a special @xmath356 .
conversely , suppose that @xmath378 is a total colouring of @xmath260 with six colours .
it induces an edge colouring @xmath195 of @xmath0 with six colours where @xmath380 a monochromatic copy of @xmath356 appears in @xmath195 if and only if @xmath378 contains ( [ item : e - k4 ] ) or ( [ item : ve ] ) .
putting this together with earlier considerations reduces the initial problem to finding a graph @xmath260 that admits a total colouring with six colours with no ( [ item : e - k4 ] ) and ( [ item : ve ] ) but whose every total colouring with six colours produces ( [ item : e - k4 ] ) , ( [ item : ve ] ) or ( [ item : eopep ] ) . start with a graph @xmath7 that is @xmath94-edge - ramsey for @xmath356 but not @xmath381-edge - ramsey for @xmath356 .
lemma [ lem : v - vs - e ] enables us to choose @xmath260 that is @xmath382-vertex - ramsey for @xmath7 but not @xmath381-edge - ramsey for @xmath356 .
if we use five colours to colour the edges of @xmath260 avoiding a monochromatic @xmath356 and use another colour for all the vertices , we get a total colouring without ( [ item : e - k4 ] ) and ( [ item : ve ] ) . on the other hand ,
given a total colouring of @xmath260 with six colours , first consider the colours in each pair of opposite colours as being the same , thereby reducing the number of colours to three . by the choice of @xmath260
, there is a copy of @xmath7 with monochromatic vertex set .
this means that in the original total colouring the vertices of this copy of @xmath7 are coloured using only one pair of opposite colours , say , red and blue .
if some edge of this copy of @xmath7 is red or blue then we get ( [ item : ve ] ) or ( [ item : eopep ] ) .
otherwise , only four colours are used for these edges so we have ( [ item : e - k4 ] ) .
this concludes the proof of lemma [ lem : main ] .
in this section we define and briefly examine the notion of ` ordered ' degrees for sets of multiple vertices .
let @xmath12 be an orientation of an @xmath1-uniform hypergraph @xmath0 .
given a pair of vertices @xmath23 , we can define @xmath383 to be the number of edges @xmath9 such that @xmath25 is in the first position of @xmath13 and @xmath14 is in the second .
for example , if @xmath384 then @xmath385 . more generally , for an _ ordered _
@xmath2-tuple of distinct vertices @xmath386 and an _ ordered _
@xmath2-tuple @xmath387 $ ] , the _ ordered @xmath31-degree _ of @xmath32 , denoted by @xmath388 , is the number of edges @xmath9 such that the elements of @xmath13 in positions labeled by @xmath31 are the vertices @xmath34 in _ this order_. more formally , @xmath388 is the number of edges @xmath9 such that if we write @xmath389 then @xmath390 . in this section by a _
@xmath2-tuple _ we will always mean an ordered @xmath2-set without repeated elements .
for any set @xmath107 we denote by @xmath391 the family of all @xmath2-tuples with elements from @xmath107 .
for example , @xmath392{^{\underline{2 } } } = \left\ { ( 1,2 ) , ( 1,3 ) , ( 2,1 ) , ( 2,3 ) , ( 3,1 ) , ( 3,2 ) \right\}$ ] . following the spirit of theorems [ thm : hakimi ] , [ thm : hakforhyp ] and question [ qn : genhak ] , we can ask when an @xmath1-uniform hypergraph @xmath0 can be given an orientation such that @xmath393 for all @xmath2-tuples of vertices @xmath32 , where @xmath8 is a fixed integer .
it is easy to find a necessary condition : for any collection of @xmath2-sets @xmath40 there can be no more than @xmath394 edges @xmath9 such that @xmath395 .
indeed , every such edge contributes to the sum @xmath396 by @xmath46 and this sum does not exceed @xmath394 .
the proof of sufficiency is almost identical to theorem [ thm : hakforpsets ] .
[ thm : hakforptuples ] fix integers @xmath49 and let @xmath0 be an @xmath1-uniform hypergraph .
suppose that for any @xmath40 there are at most @xmath394 edges @xmath9 such that @xmath42 .
then @xmath0 has an orientation such that @xmath393 for every @xmath2-tuple @xmath39 . construct a bipartite graph @xmath7 with vertex classes @xmath102 and @xmath397 .
join @xmath104 and @xmath105 by an edge if @xmath9 contains all elements of @xmath32 .
suppose @xmath398 and let @xmath399 be the neighbourhood of @xmath107 in @xmath7 .
all @xmath400 permutations of any element of @xmath401 are in @xmath401 so by assumption @xmath402 . by hall
s marriage theorem there is a @xmath46-to-@xmath4 matching from @xmath44 to @xmath315 .
orient each edge @xmath403 in such a way that the initial @xmath2-positions of @xmath13 are the @xmath2-tuple to which @xmath9 is joined in this matching .
this produces an orientation of @xmath0 with the required property .
a version of question [ qn : orient - part ] can be asked for ordered degrees .
when does an @xmath1-uniform hypergraph @xmath0 have an orientation such that for each @xmath2-tuple of vertices @xmath32 there is a @xmath2-tuple @xmath30 $ ] such that @xmath404 ? if @xmath48 then this is covered by theorem [ thm : alldeg ] so let us assume @xmath77 .
in contrast to the notion of ` unordered ' degrees , it turns out that every @xmath0 has such an orientation .
in fact , this can be achieved by a simple explicit construction .
let @xmath405 be integers and @xmath0 an @xmath1-uniform hypergraph .
then @xmath0 has an orientation such that for each @xmath2-tuple of vertices @xmath32 there is a @xmath2-tuple @xmath30 $ ] such that @xmath404
. without loss of generality assume that @xmath406 $ ] . for each edge
@xmath9 , order its vertices in the increasing order .
let @xmath407 be a @xmath2-tuple of vertices .
if @xmath408 then @xmath409 . otherwise , @xmath410 .
question [ qn : orient - part ] has a negative answer if @xmath411 which is the smallest pair for which there is a nonconstant @xmath412{^{(p ) } } \to [ r]{^{(p)}}$ ] that does not fix an intersection , and our proof went by constructing a graph that is not @xmath354-edge - ramsey for @xmath356 but whose every edge - colouring with colours @xmath92{^{(2)}}$ ] induces a colouring on a @xmath356 that does not fix an intersection
. it might be interesting to know if question [ qn : orient - part ] has a negative answer exactly when the conclusion of conjecture [ conj : main ] is false , and perhaps this could be done by adapting the same construction for hypergraphs .
let @xmath37 be integers such that @xmath127^{(p)}}$ ] does not have the fixed intersection property .
must there be an @xmath1-uniform hypergraph @xmath0 satisfying : a. @xmath63 can be partitioned into @xmath68 sets @xmath67 in such a way that for any @xmath69 there are no edges @xmath9 such that @xmath70 ; b. whenever @xmath0 is given an orientation there is some @xmath2-set of vertices @xmath32 such that @xmath413 for all @xmath2-sets @xmath30 $ ] ? using brute - force computer search we were able to check conjecture [ conj : main ] for cases @xmath414 .
our main progress in the general case is achieved by showing that pairs @xmath415 with @xmath1 much larger than @xmath2 have the stronger property that every nonconstant @xmath416{^{(p ) } } \to { \mathbb{n}}{^{(p)}}$ ] fixes an intersection ( corollary [ cor : infrang ] ) . [
it : infrang ] for what choices of @xmath1 and @xmath2 does every nonconstant function @xmath417^{(p ) } } \to { \mathbb{n}}{^{(p)}}$ ] fix an intersection ?
the interesting choices of @xmath1 and @xmath2 are when @xmath1 is slightly larger than @xmath418 .
computer search reveals that this question is satisfied by @xmath419 but not by @xmath420 .
therefore conjecture [ conj : main ] can not be proved solely by answering this question . however
, if @xmath421{^{(3 ) } } \to [ m]{^{(3)}}$ ] does not fix an intersection then @xmath422 must be large .
in fact , the smallest such @xmath422 is @xmath423 .
curiously , every such @xmath73 ( even if @xmath424 ) is injective
. therefore it might be that the additional injectivity condition , would not change the answer .
[ it : infranginj ] for what choices of @xmath1 and @xmath2 does every injection @xmath417^{(p ) } } \to { \mathbb{n}}{^{(p)}}$ ] fix an intersection ? in conjecture
[ cor : infrang ] we deal with functions @xmath127^{(p ) } } \to { [ r]^{(p)}}$ ] . in this setting the analogue of question [ it : infranginj ] is as follows .
[ it : finrangbij ] for what choices of @xmath1 and @xmath2 does every bijection @xmath417^{(p ) } } \to { [ r]^{(p)}}$ ] fix an intersection ?
there is a simple counting argument that shows question [ it : finrangbij ] is satisfied by @xmath425 .
suppose @xmath416{^{(p ) } } \to [ r]{^{(p)}}$ ] is a bijection and assume @xmath190 ) = [ p]$ ] for convenience .
there are exactly @xmath426 elements of @xmath101{^{(p)}}$ ] that do not intersect @xmath193 $ ] and @xmath427 elements that intersect @xmath193 $ ] .
so if there is no @xmath428{^{(p)}}$ ] such that @xmath429| = |x \cap [ p]| = 0 $ ] then @xmath430 which is equivalent to @xmath431 .
but then @xmath432 . if @xmath433 then the left hand side is at most @xmath434 which is not greater than @xmath253 provided @xmath195 is large enough .
we would like to thank imre leader for his invaluable comments regarding the presentation of this article . | let @xmath0 be an @xmath1-uniform hypergraph .
when is it possible to orient the edges of @xmath0 in such a way that every @xmath2-set of vertices has some @xmath2-degree equal to @xmath3 ? ( the @xmath2-degrees generalise for sets of vertices what in - degree and out - degree are for single vertices in directed graphs . ) caro and hansberg asked if the obvious hall - type necessary condition is also sufficient .
our main aim is to show that this is true for @xmath1 large ( for given @xmath2 ) , but false in general .
our counterexample is based on a new technique in sparse ramsey theory that may be of independent interest . |
the discovery of ultramassive black holes ( umbh ) , with masses above @xmath5 by mcconnell et al ( 2011 , 2012 ) has challenged our understanding of black hole growth .
their objects lie at the centres of brightest cluster galaxies , which have the highest stellar masses known .
the mystery has now deepened by the recent claim of a umbh in the lenticular galaxy ngc1277 ( van den bosch et al 2012 ) .
the estimated black hole mass of @xmath0 corresponds to 14 per cent of the total stellar mass of that galaxy , much larger than the 0.1 per cent found in a normal massive galaxy .
moreover , van den bosch et al ( 2012 ) imply that there may be many more such objects , based on a table including 5 other galaxies with similar properties .
all of the known examples lie within about 100 mpc from us .
presumably they were among the most luminous quasars when they were growing but are certainly not quasars now .
ngc1277 lies in the core of the perseus cluster where we have amassed a very deep x - ray exposure with chandra ( fabian et al 2006 ; 2011 ) .
it appears as a weak compact source with properties similar to other early - type galaxies in the cores of clusters , showing a power - law continuum in an extended thermal minicorona ( sun et al 2007 ; santra et al 2007 ) .
we present here a re - analysis of the x - ray data .
the minicorona of 10 million k gas lies within the bondi radius so should be accreting .
if the accretion flow were radiatively efficient with a radiative efficiency of 10 per cent then the luminosity should exceed @xmath2 , i.e. it should be a quasar .
it clearly is not , since the power - law component has an x - ray luminosity of only @xmath1 .
we discuss possible ways in which accretion may be reduced .
we then briefly look at ngc1270 which is another object from the table of umbh candidates of van den bosch ( 2012 ) and lies also in the perseus core and has roughly similar properties .
we note the surprisingly large mean black hole mass density implied by the newly discovered umbh candidates .
we then consider the possibility in which extreme active galactic nucleus feedback ( silk & rees 1998 ; fabian 2012 ) during the growth phase of a umbh pushes all surrounding gas away , except in the directions along the filaments fuelling the black hole accretion .
star formation would then be inhibited close to the umbh and when growth ceases around redshift 3 , all umbh are surrounded by whatever compact stellar bulge has survived from the earliest growth phases . those umbh which now lie at the centre of a cluster or group of galaxies will have since been able to accrete a large stellar envelope due to cannibalism , cooling flows etc and now appear as a bcg .
those orbiting in the core of a cluster , like ngc1277 , will have been unable to accrete much of a stellar envelope and so have a large fraction of their mass in the central black hole .
we concentrate on the umbh interpretation of ngc1277 , but briefly note and discuss the possible influence of a bottom - heavy stellar mass function , which may reduce the black hole mass required by the observations .
the core of the perseus cluster has been observed several times with chandra acis - s , centred on the brightest cluster galaxy ngc1275 fabian et al ( 2006 , 2012 ) .
4345 background - subtracted counts are detected in a 3 arcsec radius aperture on ngc1277 ( fig . 1 ) , 3.75 arcmin to the north of ngc1275 , with a total effective exposure time of 698,800 s. the source flux in the 0.57 kev band is @xmath6 . since the source is about 3 arcmin from the aimpoint of the telescope ,
the point spread function is slightly worse than on axis .
fitting the profile of the source by a gaussian gives a width , @xmath7 of about 1 arcsec , which is slightly larger than the nominal width for this off - axis distance . using a simulation with the chandra tool marx
, we can compare it to the observed profile of the source ( fig . 2 ) .
the chandra data can be fitted by the simulated point source of gaussian width ( standard deviation ) @xmath8 arcsec plus an extended part of intrinsic gaussian width @xmath9 arcsec .
the ratio of the flux in the two components is @xmath10 . adopting a cluster redshift of 0.018
the angular scale is 360 pc per arcsec , so the extended component has a half - light radius of approximately 0.36 kpc . the spectrum ( fig .
3 ) is characteristic of minicoronae ( vikhlinin et al 2001 ; sun et al 2007 ) .
it is well - modelled by thermal gas plus a power - law continuum ( @xmath11 , the thermal model alone gives @xmath12 and the power - law alone gives @xmath13 ) .
the peak just below 1 kev is due to fe - l emission lines from the hot gas .
we obtain a temperature for the gas of @xmath14 .
the best - fitting metal abundance is 0.6 but is poorly constrained provided that it is at least 0.05 ( the thermal continuum is degenerate with the power - law ) .
galactic absorption of @xmath15 is included , which is typical for the core of the perseus cluster ( fabian et al 2006 ) .
the power - law component has a photon index @xmath16 if the metallicity is fixed at @xmath17 ( anders & grevesse 1982 ) , and the absorption - corrected 0.57 kev luminosity is @xmath18 .
the normalization of the power - law component correlates with both @xmath19 and metal abundance @xmath20 .
provided that @xmath21 ( i.e. at least comparable to the icm ) then the uncertainty of the normalization is less than 30 per cent .
these results are similar to those obtained previously ( sun et al 2007 ) .
the ratio of the flux of the power - law to the thermal power - law component is about 2 , in fair agreement with the power - law arising from an unresolved central point source and the thermal emission being extended , as expected .
a good fit is also obtained with two thermal components , at 0.96 and 4.6kev .
modelling the thermal component as a sphere of uniform density and radius @xmath22 gives a density of @xmath23 .
the half - light radius of sphere is 0.36 kpc if @xmath24 .
the bondi radius @xmath25 for a black hole of mass @xmath0 is @xmath26 for gas at @xmath27 , which is slightly larger than the radius at which most of the thermal gas is found . assuming that the gas is flowing inward at its sound speed , at a radius of about 0.6 kpc
, we obtain an approximate accretion rate , @xmath28 which gives an accretion luminosity of @xmath29 where we use 0.25 for the factor @xmath30 , which is appropriate for bondi accretion of an adiabatic gas .
the radiative efficiency of accretion is @xmath31 , the gas density @xmath32 , the sound speed @xmath33 . @xmath34 and @xmath35 are the mass of the black hole and the proton , respectively .
there is then a mismatch between the predicted accretion luminosity and the observed luminosity of the power - law source , by a factor of nearly @xmath3
. some of this could be in bolometric corrections , but a factor of more than @xmath36 will remain . we see no obvious disturbance in surrounding intracluster gas
so there is no major mechanical energy loss through winds or jets .
the overall accretion efficiency must be @xmath37 or less .
inefficient accretion flows have been inferred before for sgr a * at the centre of our galaxy ( rees 1983 ; narayan & yi 1994 ) .
the flow can become convectively unstable at low accretion rates leading to a convectively dominated accretion flow ( narayan et al 2000 ; quatert & gruzinov 2000 ) . mismatches with supermassive black holes have been observed before , from m87 ( di matteo et al 2003 ) and ngc3115 ( wong et al 2011 ) where @xmath38 .
the most massive black holes provide therefore stringent tests of accretion theory .
we note that there is a very deep dust lane ( fig .
4 ) in the centre of ngc1277 , with a radius of @xmath39 pc .
this presumably indicates that matter is orbiting the black hole there , with a velocity of at least @xmath40 , if the black hole mass is @xmath41 any cold accretion flow may be stalled at the location of the dust lane : thermal gas at a temperature of 1 kev will however be in a thick atmosphere .
ngc1277 is listed as a radio source with flux 2.7 mjy at 1.4 ghz by miller & owen ( 2001 ; from work by sijbring 1993 ) .
the presence of umbh in galaxies with stellar masses as low as that of ngc1277 needs to be confirmed .
this can be done by spectroscopy of the stellar bulge at higher spatial resolution .
if the dust ring in ngc1277 is accompanied by gas , as is likely , then the high expected circular velocity ( @xmath42 ) of that ring would provide convincing confirmation .
we shall proceed assuming that there _ is _ a umbh in ngc1277 .
van den bosch et al ( 2012 ) list a further 5 galaxies which are candidates for hosting ultramassive black holes , including 2 more in the perseus cluster ( ngc1270 and ugc2689 ) .
ngc1270 is also in the core of the cluster and lies in some of the deep chandra images .
it is also characterised by a minicorona of temperature 1.07 kev plus a power - law component of 0.57kev luminosity @xmath43 ( sun et al 2007 ; pearce et al 2012 submitted ) .
ngc1270 is nearly 10 arcmin away from the chandra aimpoint so the psf is too broad for an extent analysis .
we note that ngc4889 , a bcg in the coma cluster with a black hole of mass of @xmath44 ( mcconnell et al 2011 ) also has a minicorona , one of the first discovered ( vikhlinin et al 2001 ) .
if we add the 6 candidate ultramassive black holes ( umbh ) in the table of van den bosch ( 2012 ) to the 2 found by mcconnell et al ( 2012 ) , then we have 8 within about 105 mpc .
this represents a mass density of @xmath45 assuming a mean mass of @xmath5 .
if they are all @xmath44 the mass density would be a factor of two larger and if we account for the fact that they are all northern objects , which is the minimum selection effect operating , then the mass density increases by another factor of two .
their mass density could then easily be @xmath46 which is a significant fraction of the mean mass density of black holes , estimated by e.g. marconi et al ( 2004 ) at @xmath47 or hopkins et al ( 2006 ) at @xmath48 .
the above 8 objects translate to a present umbh space density of at least @xmath49 . with three umbh in the perseus cluster
they appear to be strongly clustered .
the objects of mcconnell et al ( 2012 ) are bcgs , 3 of the van den bosch ( 2012 ) sample are in the perseus cluster and one other is in abell 347 .
hlavacek - larrondo et al ( 2012 ) have indirect arguments for umbh in a distant population of bcgs .
there is thus some evidence that the environment of a cluster , or whatever turns into a cluster , is conducive to the growth of umbh .
the galaxies that host them are of two types , the first are bcgs , with a high stellar mass , the second are undistinguished early - type galaxies with relatively low stellar mass , such as ngc1277 .
if there is a umbh in ngc1277 it is growing little at the present epoch , since its current mass doubling time is many hubble times .
its bolometric eddington limit is @xmath50 which rivals the most luminous objects known and requires an accretion rate of 100s @xmath51 .
even if it took billions of years to grow , we can expect that it was a powerful quasar or blazar in the past .
hopkins et al ( 2006 ) , ghisellini et al ( 2010 ) and volonteri et al ( 2011 ) have all considered the growth of black holes of mass exceeding @xmath52 , based on observations of quasars and blazars .
the present number density they estimate for black holes above that mass is less than we are inferring for the umbh , which are ten times more massive .
this may indicate that the umbh , if real , are now too numerous to have been regular quasars or blazars , but must have lost their accretion energy through mechanical means , for example powerful winds , or been highly obscured .
a possible scenario then emerges where the enormous power of the accreting umbh stops any significant star formation in its neighbourhood .
such extreme agn feedback could have been too fierce for new star formation to occur and only a compact stellar region or bulge survives from the earliest stages of growth .
of course the black hole has to be fuelled , which requires relatively low angular momentum gas to be fed into the centre , more or less continuously .
dubois et al ( 2012 ) show that this can occur at high redshifts .
this would require that feedback occurs in a bipolar fashion , as expected if jets or winds from an accretion disc are involved , in order that the fuelling along filaments can continue undisturbed .
the large mass of the black hole means that the orientation of the object will be relatively fixed . in this model , where the major umbh growth phase is likely to happen in deep potential wells before a redshift of 3 , we are left with the most massive black holes lying in compact bulges . at redshift 3 , most umbh would then resemble ngc1277 .
( van den bosch et al 2012 have already noted the resemblance between ngc1277 and typical red , passive galaxies at earlier times . )
what the host galaxy eventually looks like then depends on how that bulge accretes more gas and stars , and how further feedback shapes the gas and thus star formation .
such a scenario is consistent with recent observational evidence that the growth of massive galaxies from redshifts of 2.5 to 1 occurs in an inside - out mode ( van dokkum et al 2010 , szomoru et al 2011 and references therein ) .
if the umbh host lies in a cluster then it only later accretes or accumulates significant gas for star formation if it lies at the centre of the potential well and so has a low velocity with respect to its surroundings .
cannabalism , cooling flows and mergers can supply that object with an extensive halo of stars and gas .
the object thus becomes a bcg .
if the umbh bulge orbits in the core of a cluster at @xmath53 , then it accumulates few stars and little gas and could end up resembling ngc1277 .
its outer dark matter halo will have dissolved into the general cluster halo .
note that there is a requirement in the model , which is not easy to fulfill , that fuelling can proceed while powerful feedback is taking place and we have appealed to a special geometry to allow this to occur .
this special requirement may be why few galaxies host a umbh .
only when the geometry of the fuelling filaments is appropriate can a umbh form , otherwise the central black hole is 10 to 100 times less massive .
the energy released in growing a black hole of mass @xmath54 is @xmath55 which will heat @xmath56 of gas , comparable to the gas in the present - day perseus cluster , to @xmath57 .
this could expel much of the gas in any early subcluster if the energy is widely distributed and little is radiated .
widespread heating to a few kev per particle by the growth of black holes has been invoked to explain the the x - ray properties of groups and clusters ( e.g. the x - ray luminosity temperature relation , wu , fabian & nulsen 2000 ) , but 10kev is too much .
some inefficiency in the heating is required , such as could occur due to infrared radiation in an obscured scenario , or if the accretion process is intrinsically inefficient ( @xmath58 ) .
the data on ngc1277 presented by van den bosch et al ( 2012 ) indicate that it is an unusual galaxy with a high central mass - to - light ratio .
future observations will confirm whether this is due to a umbh .
an alternative possibility is that the initial mass function ( imf ) of the stars formed in the core of the galaxy is very bottom heavy , i.e. it has much of its mass in low mass stars which are not directly detected .
recent work by van dokkum & conroy ( 2010 ) , conroy & van dokkum ( 2012 ) and cappellari et al ( 2012 ) have shown that the imf in early type galaxies is increasingly skewed to low mass stars as the velocity dispersion of the stars increases .
this could be due to the higher pressure of extended gaseous atmospheres in deep potential wells ( krumholtz 2012 ) .
the high velocity dispersions found in the umbh host galaxies already imply a bottom - heavy imf in the candidate umbh hosts .
it is possible that the observations can be explained by a combination of low - mass stars and a less massive black hole than a umbh .
van den bosch et al ( 2012 ) do however show that the mass - to - light ratio of the stellar component needs to be exceptional ( e.g. @xmath59 ) for the black hole mass to be significantly reduced .
we thank the referee for helpful comments .
we thank the chandra x - ray observatory team for taking the observations .
the image of ngc1277 used in fig .
4 is from the mikulski archive for space telescopes ( mast ) .
acf thanks the royal society for support .
anders e. , grevesse n. , 1989 , gecoa , 53 , 197 cappellari m. et al 2012 , nature , 484 , 485 conroy c. , van dokkum p.g . , 2012 ,
apj , 760 , 71 di matteo t. , allen s.w . , fabian a.c . , wilson a.s . , young a.j . , 2003 ,
apj , 582 , l33 dubois y. , pichon c. , haehnelt m. , kimm t. , slyz a. , devriendt j. , pogosyan d. , 2012 , mnras , 423 , 3616 fabian a.c . , sanders j.s . , taylor g.b . ,
allen s.w .
, crawford c.s . , johnstone r.m .
, iwasawa k/. 2006 .
mnras , 366 , 417 fabian a.c . , et al . , 2011 , mnras , 418 , 2154 fabian a.c . , 2012 , araa , 50 , 455 ghisellini g. , et al 2010 , mnras , 405 , 387 hopkins p.f . , herquist l. , cox t.j . ,
di matteo t. , robertson b. , springel v. , 2006 , apjs , 163 , 1 krumholz m.r . , apj , 743 , 110 mcconnell n.j . , ma c - p . , gebhardt k. , wright s.a . , murphy j.d . , lauer t.r . ,
graham j.r . ,
richstone d.o .
, 2011 , nature , 480 , 215 mcconnell n.j . ,
, murphy j.d . , gebhardt k. , lauer t.r . , graham j.r . ,
wright s.a . ,
richstone d.o . , 2012 ,
apj , 756 , 179 miller n.a . ,
owen f.n . , 2001 , apjs , 134 , 355 narayan r. , yi i. , 1995 , 452 , 710 narayan r. , igumenshchev i.v .
, abramowicz m.a . , 2000 , apj , 539 , 798 quataert e. , gruzinov a. , 2000 , apj , 539 , 809 rees m.j . , 1982 , in ` the galactic centre ' , ed . g. reigler and r.d .
blandford ( aip ) , p.166 santra s. , sanders j.s .
, fabian a.c . , 2007 , mnras , 382 , 895 sijbring d. , 1993 , phd thesis , leiden silk j. , rees , m.j .
, 1998 , a&a , 331 , l1 sun m. , jones c. , forman w. , vihkhlinin a. , donahue m. , voit m. , 2007 , apj , 657 , 197 szomoru d. , franx m. , bouwens r. j. , van dokkum p. g. , labb i. , illingworth g. d. , trenti m. , 2011 , apj , 735 , l22 van den bosch r.c.e . , gebhardt k. , gltekin k. , van de ven g. , van der wel a. , walsh j.l , 2012 , nature , 491 , 729 van dokkum p.g . , conroy c. , 2010 , nature , 468 , 440 van dokkum p.g .
et al 2010 , apj , 709 , 1018 vikhlinin a. , markevitch m. , forman w. , jones c. , 2001 , apj , 555 , l87 volonteri m. , haardt f. , ghisellini g. , della ceca r. , 2011 , mnras , 416 , 216 wong k - w . , irwin j.a .
kukita m. , million e. , mathews w.g .
, bregman j.n . , 2011 , apj , 736 , l23 wu kks , fabian ac , nulsen pej 2000 , mnras , 318 , 889 | we study the x - ray emission from ngc1277 , a galaxy in the core of the perseus cluster , for which van den bosch et al .
have recently claimed the presence of an ultramassive black hole ( umbh ) of mass @xmath0 , unless the imf of the stars in the stellar bulge is extremely bottom heavy .
the x - rays originate in a power - law component of luminosity @xmath1 embedded in a 1 kev thermal minicorona which has a half - light radius of about 0.36 kpc , typical of many early - type galaxies in rich clusters of galaxies . if bondi accretion operated onto the umbh from the minicorona with a radiative efficiency of 10 per cent , then the object would appear as a quasar with luminosity @xmath2 , a factor of almost @xmath3 times higher than observed
. the accretion flow must be highly radiatively inefficient , similar to past results on m87 and ngc3115 .
the umbh in ngc1277 is definitely not undergoing any significant growth at the present epoch .
we note that there are 3 umbh candidates in the perseus cluster and that the inferred present mean mass density in umbh could be @xmath4 , which is 20 to 30 per cent of the estimated mean mass density of all black holes .
we speculate on the implied growth of umbh and their hosts , and discuss the possibiity that extreme agn feedback could make all umbh host galaxies have low stellar masses at redshifts around 3 . only those which end up at the centres of groups and clusters later accrete large stellar envelopes and become brightest cluster galaxies . ngc1277 and
the other perseus core umbh , ngc1270 , have not however been able to gather more stars or gas owing to their rapid orbital motion in the cluster core .
. x - rays : galaxies galaxies : clusters intergalactic medium galaxies : individual ( ngc1277 ) |
many of the predictions of the bogoliubov theory of superfluidity@xcite have been confirmed experimentally , in particular in ultracold atomic gases.@xcite nevertheless a clear understanding of the infrared behavior of interacting bosons at zero temperature has remained a challenging theoretical issue for a long time .
early attempts to go beyond the bogoliubov theory have revealed a singular perturbation theory plagued by infrared divergences due to the presence of the bose - einstein condensate and the goldstone mode.@xcite in the 1970s , nepomnyashchii and nepomnyashchii proved that the anomalous self - energy vanishes at zero frequency and momentum in dimension @xmath0.@xcite this exact result shows that the bogoliubov approximation , where the linear spectrum and the superfluidity rely on a finite value of the anomalous self - energy , breaks down at low energy . as realized latter
on,@xcite the singular perturbation theory is a direct consequence of the coupling between transverse and longitudinal fluctuations and reflects the divergence of the longitudinal susceptibility a general phenomenon in systems with a continuous broken symmetry.@xcite in this paper , we review the infrared behavior of interacting bosons .
a more detailed discussion together with a comparison to the classical o(@xmath1 ) model can be found in ref .
@xcite . in sec .
[ sec_bosons_pt ] , we briefly review the bogoliubov theory and the appearance of infrared divergences in perturbation theory .
we introduce the ginzburg momentum scale @xmath2 signaling the breakdown of the bogoliubov approximation . in sec .
[ sec_bosons_nprg ] , we discuss the non - perturbative renormalization group ( nprg ) and show how it enables to obtain the exact infrared behavior of the normal and anomalous single - particle propagators without encountering infrared divergences.@xcite
we consider interacting bosons at zero temperature with the ( euclidean ) action @xmath3 , \label{action}\ ] ] where @xmath4 is a bosonic ( complex ) field , @xmath5 , and @xmath6 .
@xmath7 $ ] is an imaginary time , @xmath8 the inverse temperature , and @xmath9 denotes the chemical potential .
the interaction is assumed to be local in space and the model is regularized by a momentum cutoff @xmath10 .
we consider a space dimension @xmath11 . introducing the two - component field @xmath12 ( with @xmath13 and @xmath14 a matsubara frequency ) ,
the one - particle ( connected ) propagator becomes a @xmath15 matrix whose inverse in fourier space is given by @xmath16 where @xmath17 and @xmath18 are the normal and anomalous self - energies , respectively , and @xmath19 .
if we choose the order parameter @xmath20 to be real ( with @xmath21 the condensate density ) , then the anomalous self - energy @xmath22 is real .
the bogoliubov approximation is a gaussian fluctuation theory about the saddle point solution @xmath23 .
it is equivalent to a zero - loop calculation of the self - energies,@xcite @xmath24 this yields the ( connected ) propagators @xmath25 where @xmath26^{1/2}$ ] is the bogoliubov quasi - particle excitation energy .
when @xmath27 is larger than the healing momentum @xmath28 , the spectrum @xmath29 is particle - like , whereas it becomes sound - like for @xmath30 with a velocity @xmath31 . in the weak - coupling limit , @xmath32 ( @xmath33 is the mean boson density ) and @xmath34
can equivalently be defined as @xmath35 . in the bogoliubov approximation ,
the occurrence of a linear spectrum at low energy ( which implies superfluidity according to landau s criterion ) is due to @xmath36 being nonzero .
the lowest - order ( one - loop ) correction @xmath37 to the bogoliubov result @xmath38 is divergent for @xmath0 . retaining only the divergent part , we obtain ^(1)(p ) ^(1)(p ) - 2 a_d+1 ( ^2+)^(d-3)/2 [ sigma4a ] if @xmath39 and ^(1)(p ) ^(1)(p ) - a_4 ( ) [ sigma4b ] if @xmath40 , where a_d = lcc - & & d<4 , + & & d=4 . . we can estimate the characteristic ( ginzburg ) momentum scale @xmath2 below which the bogoliubov approximation breaks down from the condition @xmath41 or @xmath42 for @xmath43 and @xmath44 , @xmath45 this result can be rewritten as @xmath46 where g = gm |n^1 - 2/d ~()^2 is the dimensionless coupling constant obtained by comparing the mean interaction energy per particle @xmath47 to the typical kinetic energy @xmath48 where @xmath49 is the mean distance between particles.@xcite a superfluid is weakly correlated if @xmath50 , i.e. @xmath51 ( the characteristic momentum scale @xmath52 does however not play any role in the weak - coupling limit ) . in this case , the bogoliubov theory applies to a large part of the spectrum where the dispersion is linear ( i.e. @xmath53 ) and breaks down only at very small momenta @xmath54 .
when the dimensionless coupling @xmath55 becomes of order unity , the three characteristic momentum scales @xmath56 become of the same order .
the momentum range @xmath57 $ ] where the linear spectrum can be described by the bogoliubov theory is then suppressed .
we expect the strong - coupling regime @xmath58 to be governed by a single characteristic momentum scale , namely @xmath52.@xcite the exact values of @xmath59 and @xmath36 can be obtained using the u(1 ) symmetry of the action , i.e. the invariance under the field transformation @xmath60 and @xmath61 . on the one hand ,
the self - energies satisfy the hugenholtz - pines theorem,@xcite ( p=0)-(p=0)=. [ sigma6a ] on the other hand , the anomalous self - energy vanishes , ( p=0 ) = 0 .
[ sigma6b ] the last result was first proven by nepomnyashchii and nepomnyashchii.@xcite it shows that the bogoliubov theory , where the linear spectrum and the superfluidity rely on a finite value of the anomalous self - energy , breaks down at low energy in agreement with the conclusions drawn from perturbation theory ( sec . [ subsec_bosons_ir ] ) .
the nprg enables to circumvent the difficulties of perturbation theory and derive the correlation functions in the low - energy limit.@xcite the strategy of the nprg is to build a family of theories indexed by a momentum scale @xmath62 such that fluctuations are smoothly taken into account as @xmath62 is lowered from the microscopic scale @xmath10 down to 0.@xcite this is achieved by adding to the action ( [ action ] ) an infrared regulator term s_k[^ * , ] = _ p ^*(p ) r_k(p ) ( p ) .
[ irreg ] the main quantity of interest is the so - called average effective action _
k[^ * , ] = - z_k[j^*,j ] + _ p [ j^*(p)(p)+c.c . ] - s_k[^ * , ] , defined as a modified legendre transform of @xmath63 $ ] which includes the subtraction of @xmath64 $ ] .
@xmath65 denotes a complex external source that couples linearly to the boson field @xmath66 and @xmath67 is the superfluid order parameter .
the cutoff function @xmath68 is chosen such that at the microscopic scale @xmath10 it suppresses all fluctuations , so that the mean - field approximation @xmath69=s[\phi^*,\phi]$ ] becomes exact .
the effective action of the original model ( [ action ] ) is given by @xmath70 provided that @xmath71 vanishes .
for a generic value of @xmath62 , the cutoff function @xmath72 suppresses fluctuations with momentum @xmath73 and frequency @xmath74 but leaves those with @xmath75 unaffected ( @xmath76 is the velocity of the goldstone mode ) .
the dependence of the average effective action on @xmath62 is given by wetterich s equation@xcite _ k[^ * , ] = trr_k(^(2)_k[^ * , ] + r_k)^-1 , [ rgeq ] where @xmath77 and @xmath78 .
@xmath79 $ ] denotes the second - order functional derivative of @xmath80 $ ] . in fourier space ,
the trace involves a sum over momenta and frequencies as well as the internal index of the @xmath81 field .
because of the regulator term @xmath82 , the vertices @xmath83 are smooth functions of momenta and frequencies and can be expanded in powers of @xmath84 and @xmath85 .
thus if we are interested only in the low - energy properties , we can use a derivative expansion of the average effective action.@xcite in the following we consider the ansatz@xcite _ k[^ * , ] = dx , [ effaction ] where @xmath86 .
@xmath87 denotes the condensate density in the equilibrium state .
we have introduced a second - order time derivative term .
although not present in the initial average effective action @xmath88 , we shall see that this term plays a crucial role when @xmath0.@xcite in a broken u(1 ) symmetry state with real order parameter @xmath89 , the normal and anomalous self - energies are given by _
k , n(p ) & = & + v_a , k^2 + ( 1- z_c , k)i - ( 1-z_a , k)_+ _ k n_0,k , + _ k , an(p ) & = & _ kn_0,k .
these expressions imply the existence of a sound mode with velocity @xmath90 at the initial stage of the flow , @xmath91 , @xmath92 , @xmath93 and @xmath94 , which reproduces the results of the bogoliubov approximation .
a crucial property of the rg flow is that _ k~k^3-d [ lambk1 ] vanishes with @xmath62 when @xmath0 .
( [ lambk1 ] ) follows from the numerical solution of the rg equations , but can also be anticipated from the expected singular behavior of the longitudinal propagator.@xcite the parameters @xmath95 , @xmath96 and @xmath97 can be related to thermodynamic quantities using ward identities,@xcite @xmath98 where @xmath99 is the mean boson density and @xmath100 the superfluid density . here
we consider the effective potential @xmath101 as a function of the two independent variables @xmath102 and @xmath9 .
the first of equations ( [ ward1 ] ) states that in a galilean invariant superfluid at zero temperature , the superfluid density is given by the full density of the fluid.@xcite equations ( [ ward1 ] ) also imply that the goldstone mode velocity @xmath103 coincides with the macroscopic sound velocity,@xcite i.e. @xmath104 since thermodynamic quantities , including the condensate `` compressibility '' @xmath105 should remain finite in the limit @xmath106 , we deduce from ( [ ward1 ] ) that @xmath107 vanishes in the infrared limit , and @xmath108 both @xmath109 and the macroscopic sound velocity @xmath103 being finite at @xmath110 , @xmath97 ( which vanishes in the bogoliubov approximation ) takes a non - zero value when @xmath106 . ,
@xmath96 and @xmath97 vs @xmath111 where @xmath112 for @xmath113 , @xmath114 and @xmath115 [ @xmath116 .
the inset shows @xmath2 vs @xmath117 obtained from the criterion @xmath118 [ the green solid line is a fit to @xmath119 .
( right panel ) condensate density @xmath87 , superfluid density @xmath100 and goldstone mode velocity @xmath103 vs @xmath111.,title="fig:",width=226 ] , @xmath96 and @xmath97 vs @xmath111 where @xmath112 for @xmath113 , @xmath114 and @xmath115 [ @xmath116 .
the inset shows @xmath2 vs @xmath117 obtained from the criterion @xmath118 [ the green solid line is a fit to @xmath119 .
( right panel ) condensate density @xmath87 , superfluid density @xmath100 and goldstone mode velocity @xmath103 vs @xmath111.,title="fig:",width=226 ] the suppression of @xmath96 , together with a finite value of @xmath120 shows that the average effective action ( [ effaction ] ) exhibits a `` relativistic '' invariance in the infrared limit and therefore becomes equivalent to that of the classical o(2 ) model in dimensions @xmath121 .
in the ordered phase , the coupling constant of this model vanishes as @xmath122,@xcite which agrees with ( [ lambk1 ] ) . for @xmath106 , the existence of a linear spectrum
is due to the relativistic form of the average effective action ( rather than a non - zero value of @xmath123 as in the bogoliubov approximation ) . to obtain the @xmath110 limit of the propagators ( at fixed @xmath124 )
, one should in principle stop the flow when @xmath125.@xcite since thermodynamic quantities are not expected to flow in the infrared limit , they can be approximated by their @xmath110 values . making use of the ward identities ( [ ward1 ] ) , we deduce the exact infrared behavior of the normal and anomalous propagators ( at @xmath110),@xcite g_n(p ) & = & - - - g_(p ) , + g_an(p ) & = & - g_(p ) , [ de3 ] where @xmath126 is the propagator of the longitudinal fluctuations .
the constant @xmath127 follows from the replacement @xmath128 .
the leading terms in ( [ de3 ] ) agree with the results of gavoret and nozires.@xcite the contribution of the diverging longitudinal correlation function was first identified by nepomnyashchii and nepomnyashchii.@xcite the conclusions of the preceding section can be obtained more rigorously from the rg equation ( [ rgeq ] ) satisfied by the average effective action .
the flow of @xmath129 , @xmath96 and @xmath97 is shown in fig .
[ fig_boson_flow ] for a two - dimensional system in the weak - coupling limit .
we clearly see that the bogoliubov approximation breaks down at a characteristic momentum scale @xmath130 . in the goldstone regime @xmath131
, we find that both @xmath129 and @xmath96 vanish linearly with @xmath62 in agreement with the conclusions of sec .
[ subsec_bosons_de ] . furthermore , @xmath97 takes a finite value in the limit @xmath106 in agreement with the limiting value ( [ velir ] ) of the goldstone mode velocity .
figure [ fig_boson_flow ] also shows the behavior of the condensate density @xmath87 , the superfluid density @xmath132 and the velocity @xmath103 . since @xmath133 , the mean boson density
@xmath134 is nearly equal to the condensate density @xmath87 .
apart from a slight variation at the beginning of the flow , @xmath87 , @xmath132 and @xmath103 do not change with @xmath62 .
in particular , they are not sensitive to the ginzburg scale @xmath2 .
this result is quite remarkable for the goldstone mode velocity @xmath103 , whose expression ( [ cdef ] ) involves the parameters @xmath129 , @xmath96 and @xmath97 , which all strongly vary when @xmath135 .
these findings are a nice illustration of the fact that the divergence of the longitudinal susceptibility does not affect local gauge invariant quantities.@xcite
interacting bosons at zero temperature are characterized by two momentum scales : the healing ( or hydrodynamic ) scale @xmath34 and the ginzburg scale @xmath2 .
@xmath2 sets the scale at which the bogoliubov approximation breaks down . for momenta @xmath30 ,
it is possible to use a hydrodynamic description in terms of density and phase variables .
this description allows one to compute the correlation functions without encountering infrared divergences.@xcite in this paper , we have reviewed another approach , based on the nprg , which enables to describe the system at all energy scales and yields the exact infrared behavior of the single - particle propagator .
a nice feature of the nprg is that it can be used to study models of strongly - correlated bosons such as the bose - hubbard model.@xcite note however that in the strong - coupling limit ( i.e. @xmath136 in three dimensions , with @xmath137 the @xmath138-wave scattering length ) , the bound states of the interaction potential can not be ignored . as a result
, stability is not guaranteed and the system is at best in a quasistable state : see e. braaten , h .- w .
hammer , and t. mehen , phys .
88 * , 040401 ( 2002 ) .
the strong - coupling limit of interacting bosons can nevertheless be reached in the bose - hubbard model.@xcite one could consider a more general ansatz where @xmath139 , @xmath140 and @xmath141 are functions of the condensate density @xmath102 , and the full effective potential @xmath142 is retained rather than expanded to quadratic order about @xmath87 . in the superfluid phase , @xmath142 is convex but otherwise the conclusions regarding the infrared behavior are unchanged . | we review the infrared behavior of interacting bosons at zero temperature . after a brief discussion of the bogoliubov approximation and the breakdown of perturbation theory due to infrared divergences ,
we show how the non - perturbative renormalization group enables to obtain the exact infrared behavior of the correlation functions . |
in the standard qcd perturbation theory ( pt ) we have : * the renormalization group ( rg ) equation @xmath0 for the effective coupling @xmath1 with @xmath2 ; * the one - loop solution generates landau pole singularity : @xmath3 ; * the two - loop solution generates square - root singularity : @xmath4 ; * pt series is a series in powers of effective coupling : @xmath5 in the analytic perturbation theory ( apt ) we have : * different effective couplings in minkowskian ( radyushkin @xcite , and krasnikov and pivovarov @xcite ) and euclidean ( shirkov and solovtsov @xcite ) regions ; * apt is based on the rg and causality that guaranties standard perturbative uv asymptotics and spectral properties ; * in euclidean domain , @xmath6 , @xmath7 , apt generates the following set of images for the effective coupling and its @xmath8-th powers , @xmath9 ; * in minkowskian domain , @xmath10 , @xmath11 , apt generates another set of images for the effective coupling and its @xmath8-th powers , @xmath12 ; * pt power series @xmath13 transforms into non - power series @xmath14 in apt , where @xmath15 are _ numbers _ in -scheme . by the analytization in apt for an observable @xmath16 we mean the `` kllen lehman '' representation @xmath17_\text{an } = \int_0^{\infty}\ ! \frac{\rho_f(\sigma ) } { \sigma+q^2-i\epsilon}\ , d\sigma \end{aligned}\ ] ] with the spectral density @xmath18/\pi$ ] .
then in the one - loop approximation ( note pole remover @xmath19 in ( [ eq : a_1 ] ) ) @xmath20 whereas analytic images of the higher powers ( @xmath21 ) are : @xmath22
in the standard qcd pt we have not only power series @xmath23 , but also : * the factorization procedure in qcd gives rise to the appearance of logarithmic factors of the type : @xmath24 ; * the rg evolution generates evolution factors of the type : @xmath25\,b(\mu^2)$ ] , which reduce in the one - loop approximation to @xmath26 with @xmath27 being a fractional number ; * the rg in the two - loop approximation for the coupling @xmath28^\nu\ln\left(a_s(l)\right)$ ] .
that means we need to think how to obtain analytization of new functions : @xmath29 , @xmath30 , @xmath31 .
let us first do it for the one - loop apt .
here we have a very nice recursive relation ( [ eq : a_1 ] ) .
we will use it to construct analytic images of fractional powers of qcd effective coupling in the euclidean ( fapt ) and minkowskian ( mfapt ) domains . consider the laplace transform @xmath32 which is well defined for all @xmath33 .
then @xmath34\ , e^{-l t } dt\,.\end{aligned}\ ] ] moreover , we can define for all @xmath35 @xmath36\ , e^{-l t } dt\,.\end{aligned}\ ] ] the only things one needs to know are @xmath37 and @xmath38 . eqs . (
[ eq : a_1 ] ) and ( [ eq : u_1 ] ) produce the answer : @xmath39 and @xmath40 $ ] .
this allows us to obtain explicit expressions for @xmath41 ( @xmath42 ) and @xmath43 ( @xmath44 ) using eq .
( [ eq : a.nu.laplace ] ) : @xmath45 } { \pi(\nu -1 ) \left(\pi^2+l^2\right)^{(\nu-1)/2}}\,.~\end{aligned}\ ] ] here @xmath46 is reduced lerch transcendental function .
it is an analytic function in @xmath47 .
interesting to note that @xmath48 appears to be an entire function in @xmath47 , whereas @xmath43 is determined completely in terms of elementary functions .
these expressions can be analytically continued to negative values of @xmath49 , though in derivation we assume @xmath33 .
let us discuss the main properties of @xmath41 ( @xmath42 ) and @xmath43 ( @xmath44 ) : ( left panel ) and @xmath43 ( right panel ) for fractional @xmath50 $ ] .
[ fig : u23_a23],title="fig:",scaledwidth=45.0% ] ( left panel ) and @xmath43 ( right panel ) for fractional @xmath50 $ ] .
[ fig : u23_a23],title="fig:",scaledwidth=45.0% ] these couplings have the following properties : + @xmath51 ; + @xmath52 for @xmath53 ; @xmath54 , @xmath55 , @xmath56 ; + @xmath57 for @xmath58 ; + @xmath59 for @xmath60 ; + @xmath61_\text{an } = \left[a^\nu\ln^k(a)\right]_\text{an}$ ] ; + @xmath62 for @xmath63 .
the convergence of this expansion is very fast .
( left panel ) and @xmath43 ( right panel ) for integer @xmath64 . in order to show all curves on the same panel we scale different curves by factors @xmath65 .
[ fig : ua_2345],title="fig:",scaledwidth=45.0% ] ( left panel ) and @xmath43 ( right panel ) for integer @xmath64 . in order to show all curves on the same panel we scale different curves by factors @xmath65 .
[ fig : ua_2345],title="fig:",scaledwidth=45.0% ] we display graphics of @xmath41 and @xmath43 in fig . [
fig : u23_a23 ] : one can see here a kind of distorting mirror on both panels .
next , in fig .
[ fig : ua_2345 ] we show graphics for @xmath64 . here
we can trace the partial values @xmath66 graphics for @xmath41 as functions of @xmath47 at fixed values of @xmath49 can be found in our last papers @xcite .
we compare the basic ingredients of ( m)fapt in table [ tab : pt.apt.fapt ] with their counterparts in conventional pt and apt .
.comparison of pt , apt , fapt ( @xmath67 ) , and mfapt ( @xmath68 ) . in the row , named ` inverse powers ' , we put @xmath69 that symbolically encodes just the item ( 2 ) of @xmath43 properties , see the list on the previous page .
[ tab : pt.apt.fapt ] [ cols="^,^,^,^,^",options="header " , ]
the two - loop equation for the normalized coupling @xmath70 is @xmath71 \quad \text{with}~c_1\equiv\frac{b_1}{b_0 ^ 2}\,.\end{aligned}\ ] ] ( dotted line ) and @xmath72 ( solid line ) in fapt .
right panel : comparison of relative errors @xmath73 ( dashed line ) in standard pt and @xmath72 ( solid line ) in fapt .
[ fig : fapt23_pt3],title="fig:",scaledwidth=45.0% ] ( dotted line ) and @xmath72 ( solid line ) in fapt .
right panel : comparison of relative errors @xmath73 ( dashed line ) in standard pt and @xmath72 ( solid line ) in fapt .
[ fig : fapt23_pt3],title="fig:",scaledwidth=45.0% ] rg solution of this equation assumes the following form : @xmath74 = l = \frac{1}{a_{(1)}(l)}\,.\end{aligned}\ ] ] we can expand @xmath75 in terms of @xmath76 with inclusion of terms @xmath77 : @xmath78 analytic version of this expansion is @xmath79 in fig . [
fig : fapt23_pt3 ] we demonstrate nice convergence of this expansion using relative errors of the 2- and 3-term approximations : @xmath80 @xmath81 we see that relative accuracy of the 3-term approximation in fapt ( see the left panel of fig . [ fig : fapt23_pt3 ] ) is better than 2% for @xmath82 . in the same time , the right panel of fig .
[ fig : fapt23_pt3 ] demonstrates that relative accuracy of the same 3-term approximation in standard pt even at @xmath83 is much higher about 10% , whereas in fapt it is smaller than 1% ! , computed analytically via eq .
( [ eq : a2_fapt3 ] ) ; dashed line represents the result of a numerical integration , while stars correspond to the available numerical results of magradze in @xcite .
right panel : the solid line represents @xmath87 , computed analytically via eq .
( [ eq : a2_fapt3 ] ) , while the dashed line stands for @xmath88 .
[ fig : fapt - pt.a2],title="fig:",scaledwidth=45.0% ] , computed analytically via eq .
( [ eq : a2_fapt3 ] ) ; dashed line represents the result of a numerical integration , while stars correspond to the available numerical results of magradze in @xcite .
right panel : the solid line represents @xmath87 , computed analytically via eq .
( [ eq : a2_fapt3 ] ) , while the dashed line stands for @xmath88 .
[ fig : fapt - pt.a2],title="fig:",scaledwidth=45.0% ]
scaled hard - scattering amplitude truncated at the next - to - leading order ( nlo ) and evaluated at renormalization scale @xmath92 reads @xcite @xmath93 with shorthand notation ( @xmath94 ) @xmath95 + b_0\,t_{h}^{(1,\beta)}\left(x , y;\lambda_{r}\right ) + t_{h}^{(\text{fg})}(x , y ) \right]\,.\end{aligned}\ ] ] the leading twist-2 pion distribution amplitude ( da ) @xcite at normalization scale @xmath96 is given by @xcite @xmath97\,.\end{aligned}\ ] ] all nonperturbative information is encapsulated in gegenbauer coefficients @xmath98 . to obtain
factorized part of pion form factor ( ff ) one needs to convolute the pion da with the hard - scattering amplitude : @xmath99 in order to obtain the analytic expression for the pion ff at nlo in @xcite the so - called `` naive analytization '' has been suggested .
it uses analytic image only for coupling itself , @xmath100 , but not for its powers .
in contrast and in full accord with the apt ideology the receipt of `` maximal analytization '' has been proposed recently in @xcite .
the corresponding expressions for the analytized hard amplitudes read as follows : @xmath101_\text{nai - an } & = & { \cal a}_{1}^{(2)}(\lambda_{r } q^2)\,t_{h}^{(0)}(x , y ) + \frac{\left({\cal a}_{1}^{(2)}(\lambda_{r } q^2)\right)^2}{4\pi}\ , t_{h}^{(1)}\left(x , y;\lambda_{r},\frac{\mu_{f}^2}{q^2}\right)\,;\\ \left[q^2 t_{h}\left(x , y , q^2\right ) \right]_\text{max - an } & = & { \cal a}_{1}^{(2)}(\lambda_{r } q^2)\,t_{h}^{(0)}(x , y ) + \frac{\acal_{2}^{(2)}(\lambda_{r } q^2)}{4\pi}\ , t_{h}^{(1)}\left(x , y;\lambda_{r},\frac{\mu_{f}^2}{q^2}\right)\,.\end{aligned}\ ] ] gev@xmath102 , dashed lines to @xmath103 , dotted lines to the blm prescription , whereas dash - dotted lines to the @xmath104-scheme .
right panel : predictions for the scaled pion form factor calculated with the bms bunch of the pion das .
the dashed lines inside the strip indicate the corresponding area of predictions obtained with the asymptotic pion da .
the experimental data are taken from ( diamonds ) and @xcite , @xcite ( triangles ) .
[ fig : pionff_nai_max_strip],title="fig:",scaledwidth=32.5% ] gev@xmath102 , dashed lines to @xmath103 , dotted lines to the blm prescription , whereas dash - dotted lines to the @xmath104-scheme .
right panel : predictions for the scaled pion form factor calculated with the bms bunch of the pion das .
the dashed lines inside the strip indicate the corresponding area of predictions obtained with the asymptotic pion da .
the experimental data are taken from ( diamonds ) and @xcite , @xcite ( triangles ) .
[ fig : pionff_nai_max_strip],title="fig:",scaledwidth=32.5% ] gev@xmath102 , dashed lines to @xmath103 , dotted lines to the blm prescription , whereas dash - dotted lines to the @xmath104-scheme .
right panel : predictions for the scaled pion form factor calculated with the bms bunch of the pion das .
the dashed lines inside the strip indicate the corresponding area of predictions obtained with the asymptotic pion da .
the experimental data are taken from ( diamonds ) and @xcite , @xcite ( triangles ) .
[ fig : pionff_nai_max_strip],title="fig:",scaledwidth=32.5% ] in fig .
[ fig : pionff_nai_max_strip ] we show the predictions for the factorized pion ff in the `` naive '' and in the `` maximal analytization '' approaches .
we see that in the `` maximal analytization '' approach the obtained results are practically insensitive to the renormalization scheme and scale - setting choice ( already at the nlo level ) .
we show also the graphics for the whole pion ff , obtained in apt with the `` maximally analytic '' procedure using the ward identity to match the non - factorized and factorized parts of the pion ff , see the right panel of fig .
[ fig : pionff_nai_max_strip ] .
the green strip in this figure contains both nonperturbative uncertainties from nonlocal qcd sum rules @xcite and renormalization scheme and scale ambiguities at the level of the nlo accuracy .
it is interesting to note here that fapt approach , used in @xcite for analytization of the @xmath105-terms in the hard amplitude ( [ eq : t_hard_nlo ] ) , diminishes also the dependence on the factorization scale setting in the interval @xmath106 gev@xmath102 .
we conclude with the following resume : + the implementation of the analyticity concept ( the dispersion relations ) from the level of the coupling and its powers to the level of qcd amplitudes as a whole generates extension of the apt to ( m)fapt ; + we formulate the rules how to apply ( m)fapt at the two- and three - loop levels ; + we show that convergence of the perturbative expansion is significantly improved when using non - power ( m)fapt expansion ; + as an additional advantage we obtain the minimal sensitivity to both the renormalization and factorization scale setting , revealed on the example of the pion electromagnetic form factor .
* acknowledgments : * this investigation was supported in part by the deutsche forschungsgemeinschaft ( projects dfg 436 rus 113/881/0 ) , the heisenberg
landau programme , grant 2007 , the russian foundation for fundamental research , grants no .
05 - 01 - 00992 and 07 - 02 - 91557 , and the brfbr
jinr cooperation programme , contract no . f06d-002 .
10 a. v. radyushkin , jinr rapid commun . *
78 * , 96 ( 1996 ) .
n. v. krasnikov and a. a. pivovarov , phys .
b116 * , 168 ( 1982 ) .
d. v. shirkov and i. l. solovtsov , jinr rapid commun .
* 2[76 ] * , 5 ( 1996 ) ; phys .
* 79 * , 1209 ( 1997 ) .
a. i. karanikas , n. g. stefanis , phys .
b 504 ( 2001 ) 225 ; ibid .
b 636 ( 2006 ) 330 . a. p. bakulev , s. v. mikhailov , and n. g. stefanis , phys . rev . * d72 * , 074014 ( 2005 ) ; phys
. rev . * d75 * , 056005 ( 2007 ) .
b. a. magradze , reprint rmi-2003 - 55 , 2003 [ hep - ph/0305020 ] .
r. d. field , r. gupta , s. otto , and l. chang , nucl .
b186 * , 429 ( 1981 ) .
f. m. dittes and a. v. radyushkin , sov .
. phys . * 34 * , 293 ( 1981 ) .
e. braaten and s .- m .
tse , phys .
rev . * d35 * , 2255 ( 1987 ) .
b. meli , b. nii , and k. passek , phys . rev . *
d60 * , 074004 ( 1999 ) .
a. v. radyushkin , ubna preprint p2 - 10717 , 1977 [ hep - ph/0410276 ] .
a. v. efremov and a. v. radyushkin , phys . lett . *
b94 * , 245 ( 1980 ) .
n. g. stefanis , w. schroers , and h .- c .
kim , phys .
b449 * , 299 ( 1999 ) .
n. g. stefanis , w. schroers , and h .- c .
kim , eur .
j. * c18 * , 137 ( 2000 ) .
a. p. bakulev , k. passek - kumeriki , w. schroers , and n. g. stefanis , phys . rev . * d70 * , 033014 ( 2004 ) .
j. volmer _
et al . _ ,
lett . * 86 * , 1713 ( 2001 ) . c. n. brown _ et al .
_ , phys . rev . *
d8 * , 92 ( 1973 ) . c. j. bebek _
et al . _ ,
phys . rev .
* d13 * , 25 ( 1976 ) .
a. p. bakulev , s. v. mikhailov , and n. g. stefanis , phys . lett .
* b508 * , 279 ( 2001 ) ; in _ proceedings of the 36th rencontres de moriond on qcd and hadronic interactions , 1724 mar 2001 , les arcs , france _ , edited by j. t. t. van ( world scientific , singapore , 2002 ) , pp .
133136 ; phys .
rev . * d67 * , 074012 ( 2003 ) ; phys .
b578 * , 91 ( 2004 ) .
a. p. bakulev and a. v. pimikov , acta phys .
polon . * b37 * , 3627 ( 2006 ) ; pepan lett .
* 4 * , 637 ( 2007 ) ; int .
a22 * , 654 ( 2007 ) .
a. p. bakulev , in _ new trends in high - energy physics , proceedings of the conference , yalta ( crimea ) , 1623 sept . , 2006 _ , edited by p. n. bogolyubov , l. l. jenkovszky , v. v. magas , and z. i. vakhnenko ( bitp nasu ( kiev ) , jinr ( dubna ) , kiev , 2006 ) , pp . 203212
. a. p. bakulev , a. i. karanikas , and n. g. stefanis , phys . rev . * d72 * , 074015 ( 2005 ) . | we describe the development of analytic perturbation theory ( apt ) in qcd , called fractional apt ( fapt ) , which has been suggested to apply the renormalization group evolution and qcd factorization technique in the framework of apt . *
fractional apt in qcd + in the euclidean and minkowski regions * + alexander p. bakulev + _ bogoliubov lab . of theoretical physics , + joint institute for nuclear research , 141980 , dubna , russia _ |
a schematic of the experimental setup is shown in fig . [
fig : supp1 ] .
a microwave generator creates a tone that is routed into a @xmath25he cryostat and attenuated .
the tone impinges on the sum port of a 180 degree hybrid coupler ( krytar 4040124 ) , which divides the power into two coherent components .
the signals then travel through the two tibs , and are summed together with a 90 degree hybrid ( fairview microwave sh7219 ) .
finally , the tone propagates through the microwave receiver and is readout in a spectrum analyzer or vector network analyzer . to control the transmission of the tibs , three control currents are used .
first , an off - chip magnetic coil ( orange circle ) creates a uniform magnetic flux @xmath26 across the entire chip ( @xmath27 moe ) .
the second and third control currents run in a pair of on - chip bias lines ( blue inductors ) that create gradiometric fluxes for the two tibs . to operate a tib as a double - balanced modulator
, we fix the current in the off - chip magnetic coil and dynamically modulate the current @xmath28 in its on - chip bias line .
this changes the imbalance of the tib , and thus its transmission @xmath29 .
judicious choice of this control current @xmath28 can substantially improve modulator performance . to calibrate this
, we first use a network analyzer to measure the transmission through each tib as a function of an applied dc current in its on - chip bias line .
the other tib is unbiased during this measurement ( its inductor bridge is balanced and incident signals will be reflected ) allowing for independent calibration of both tibs .
[ fig : supp2]a shows the squared amplitude ( top row ) and phase ( bottom row ) of signals transmitted through each tib , as a function of the probe frequency and a gradiometric flux @xmath30 that is proportional to the control current . in the squared amplitude plots
, one observes low transmission when @xmath31 is small , and maximal transmitted power when @xmath32 .
here @xmath33 is the reduced flux quantum , with @xmath34 planck s constant over 2@xmath21 and @xmath35 the charge of an electron .
the uniform flux is fixed at @xmath36 . because half the signal is dissipated in the 50 ohm termination on the 90 degree hybrid , and another half is reflected by the un - probed arm of the network ,
the maximum possible transmission in this configuration is -6 db . from the plots of phase in the bottom row
, one can see that inverting the polarity of the bias current causes a @xmath21 phase shift in the transmission .
, with probe frequency fixed at 6 ghz . when bias current is swept sinusoidally , transmission changes in a non - sinusoidal manner .
( c ) applying a bias current @xmath37 produces approximately sinusoidal modulation of transmission , as seen in ( d ) . for tib # 1 ,
the map @xmath38 is not invertible , which causes some distortion ( higher harmonics of @xmath5 ) in the transmitted spectrum .
( e ) power spectral density of transmission when current is modulated sinusoidally ( filled circles ) , and when the current is chosen to produce sinusoidal transmission ( unfilled diamonds ) .
spectra are normalized to the total power in each sideband .
pulse shaping reduces the weight of unwanted higher harmonics by approximately 20 db .
data for tib#1 ( tib#2 ) are shown in blue ( orange ) . ]
linecuts of the real part of the transmission at 6 ghz are shown in fig .
[ fig : supp2]b .
these are maps @xmath38 between the applied control current and the transmission through the bridges .
if these maps were linear , a sinusoidal control current would induce a sinusoidal modulation of the transmission through the tib .
transmission , however , is a nonlinear function of the control current.@xcite to ensure sinusoidal modulation by the tibs at the desired angular frequency @xmath5 , we use an arbitrary waveform generator to apply the control current @xmath39 ( fig .
[ fig : supp2]c ) .
here @xmath40 denotes the inverse of the map @xmath38 .
the time - dependence of the transmission , @xmath41 , is then ensured by function composition to be the desired value of @xmath42 ( fig .
[ fig : supp2]d ) .
when @xmath38 is not a one - to - one mapping ( fig .
[ fig : supp2]b , blue curve ) this procedure only produces an approximation of sinusoidal modulation @xcite ( fig .
[ fig : supp2]d , top panel ) .
the overall effect of this procedure can be seen in fig .
[ fig : supp2]e , which plots the power spectral density of @xmath41 for both tibs , when the bias currents are sinusoidal ( circles ) and when the bias currents are chosen to produce sinusoidal transmission ( diamonds ) .
the latter method reduces the spectral weight of higher harmonics by about 20 db .
harmonics in the spectrum of @xmath41 are imprinted directly onto the spectrum of signals transmitted through the ssbm . fig .
[ fig : supp3 ] shows the nearest 10 harmonics around the carrier as the iq phase difference is swept .
the lo port of the device is driven by a microwave tone at @xmath11 ghz and the i and q ports are driven by angular frequencies of @xmath14 mhz ( fig .
[ fig : supp3]a ) , @xmath43 mhz ( fig .
[ fig : supp3]b ) , and @xmath44 mhz ( fig .
[ fig : supp3]c ) . 3 mhz ( a ) , 20 mhz ( b ) , and 60 mhz ( c ) , with @xmath45 ghz . a gray dashed line at -3 db shows the expected upper - bound on transmitted power .
] as @xmath5 is increased and signals are converted farther from the carrier , both the carrier and the spurious harmonics increase in magnitude . this is a consequence of the finite bandwidth of our arbitrary waveform generator ( tektronix awg5014c ) , which samples at 1.2 gs / s and has a bandwidth of 180 mhz .
( for input powers much less than the 1 db compression point , the size of spurious harmonics does not depend on the power of the input signal . ) as the if modulation rate @xmath5 grows , the number of points available to represent the control current pulse @xmath46 decreases , and the bandwidth needed to capture fine corrections to @xmath46 increases .
this diminishes the ability of the pulse - shaping routine to ensure approximately sinusoidal transmission .
for instance , as @xmath47 changes from 3 to 60 mhz , the image rejection goes from 30 db to 20 db , and the sideband contrast goes from 25 db to 13 db .
the previous sections shows frequency conversion of pure microwave tones with an ssbm . in the envisaged application ,
however , the ssbm must convert wavepackets with more complex spectral composition .
a realistic example of such a signal is a microwave tone that is phase modulated in time , at a rate set by the bandwidth of the qubit readout cavity .
such a signal simulates the dataflow from a qubit - cavity system that is being continuously monitored , say , for the detection of quantum jumps : as the qubit state stochastically jumps between ground and excited , the phase of the transmitted microwave tone concomitantly acquires a phase shift of 0 or @xmath21 radians .
the phase modulation of the transmitted signal creates a wavepacket with a richer frequency content than the orginal tone , and extraction of the information in the phase modulation requires single - sideband modulation of the entire wavepacket . to simulate this process
, we use the ssbm to frequency - convert a microwave tone that is amplitude modulated with a square wave at frequency @xmath48 .
( for this demonstration , we use amplitude modulation in place of phase modulation to allow detection with a spectrum analyzer . )
we measure the modulation of the signal s amplitude in time ( fig . [
fig : supp4]a - c ) , and the spectrum of the converted signal ( fig .
[ fig : supp4]d - f ) .
amplitude modulations of @xmath49 100 khz , 1 mhz , and 5 mhz are shown in the top , middle , and bottom rows of the figure .
for each modulation , the four traces show frequency conversion by 0 , 3 , 20 , or 60 mhz , as evidenced by the shifted spectra in the figure s right column .
ghz tone after frequency conversion by the ssbm . transmitted power versus time ( a - c ) and power spectral density ( d - f )
are shown for modulation rates of 100 khz ( top row ) , 1 mhz ( middle row ) , and 5 mhz ( bottom row ) . at each of the three modulation rates
, we use the ssbm to convert the frequency of an input tone by 0 , 3 , 20 , and 60 mhz .
these are depicted by the four traces in each panel . for visual clarity , the amplitude of the unconverted signals ( blue traces ) in ( a - c )
are scaled by a factor of 1/2 . ] from fig .
[ fig : supp4 ] , one can see that the instantaneous bandwidth of the ssbm exceeds 5 mhz . at higher frequencies , for example 5 mhz ,
distortion of the square - wave begins to become apparent .
we believe the instantaneous bandwidth of the device is currently limited by the relaxation of eddy currents in the chip ground plane .
this relaxation time can be substantially reduced in future devices .
to assess the power - handling of the ssbm we sweep the on - chip power during operation at @xmath50 mhz , @xmath45 ghz , and @xmath51 degrees ( up - conversion ) , and measure the power transmitted at various harmonics of the modulation rate .
[ fig : linearity ] shows the results .
we measure the power in the nearest 18 harmonics @xmath52 to the carrier . for visual clarity ,
only the four nearest odd harmonics and the carrier are plotted , though the other harmonics exhibit similar behavior .
ghz , @xmath53 = 3 mhz , and @xmath51 degrees .
1 db compression occurs at -85 dbm . ]
when the incident power is below -95 dbm , the output power in the carrier and all the sidebands scales linearly . as the input power increases
, 1 db compression in the first blue sideband occurs at -85 dbm .
this value is consistent with previous results , which found the 1 db compression point of a single tib to be -88 dbm @xcite .
( as the tibs in our ssbm need only process one - half the power incident on the device s input , the ssbm s compression point is 3 db below that of a single tib . ) in a dispersive readout scheme the power of a readout tone depends on the dressed cavity bandwidth , but most tones fall between -130 and -120 dbm @xcite .
the ssbm s 1 db compression point at -85 dbm thus provides ample power handling for this application .
as the ssbm is created with two instances of a tib , design improvements for the single - sideband modulator may be partitioned into two categories : changes of the tib , and changes of the composite device ( the two tibs , hybrid couplers , and internal microwave connections ) .
we address these in turn .
as mentioned in the main text , immediate improvements on the tib may be realized by eliminating the chip - mode around 5 ghz ( apparent in fig .
[ fig : supp2]a ) , and by improving the impedance match .
the chip - mode is an lc resonance between the inductance of the balun coils and a parasitic capacitance to ground , which can be tuned out of the operation band by manipulating the width of the balun microstrips ( their length is constrained to be a quarter wavelength at the operation frequency ) . in the current generation of devices ,
impedance matching a tib to 50 ohms is inhibited by the finite tunability of the inductors in the bridge and the impedance of those inductors ( about 20 ohms at 6 ghz ) with respect to 50 ohms .
we overcome this problem by tuning out the residual inductance with capacitors in series with the four nodes of the bridge .
although this is an expedient solution , the resulting match is effective only within several hundred mhz of the frequency for which the capacitors have the appropriate impedance . a more sophisticated approach , such as tapered lines or matching transformers , could easily expand this bandwidth to several ghz , as the bode - fano criterion @xcite is relatively lenient for low @xmath54 circuits .
more quantitatively , an imbalanced tib may be modeled as an inductor in series with a 50 ohm load .
the thevenin equivalent inductance of a tib with inductors parametrized as @xmath1 and @xmath2 ( as in fig .
2b of the main text ) is @xmath55 , which for the tibs in this device amounts to 800 ph when the bridges are maximally imbalanced .
the bode - fano criterion thus allows for -20 db reflections over an 8 ghz bandwidth .
improved impedance matching may therefore drastically reduce return loss over the 4 to 8 ghz band .
better impedance matching can also be achieved directly by improving the tunability ( dynamic range ) of the squid critical currents .
while reducing the number of squids in each array will accomplish this by reducing geometric inductance and enhancing participation ratios , power handling will suffer . another way to improve tunability
is to more uniformly distribute the gradiometric flux across the length of the array , which reduces variation in the critical currents of the squids that make up the array . in the current layout , squids near the inside corners of the bias line couple more strongly to it , and are threaded by greater amounts of flux .
laying out the figure - eight of the bridge so that the bias line can bisect it without making turns will alleviate this problem , increase the degree to which the bridge can be imbalanced , and reduce impedance mismatch .
finally , the instantaneous bandwidth of the tib may also be improved by elimination of long timescale eddy currents in the chip ground plane .
this can be done by altering the resistance ( modifying the @xmath56 time ) of a normal metal layer placed in the chip ground plane to break supercurrent loops .
in addition to improving the constituent components , a next - generation ssbm can also benefit from complete on - chip integration .
first , on - chip superconducting hybrids can be extremely low - loss .
second , path length differences in the two arms of the network can be easily constrained to micron length scales , making them smaller than the operation wavelengths by more than four orders of magnitude .
( to contrast , in our current device this path length difference is 3 mm ) .
suppressing path length difference far below the operation wavelength allows for simultaneous minimization of the power in the image sideband and maximization of the power in the desired sideband .
a third considerable benefit of moving to a completely monolithic circuit is the recovery of an additional 3 db of modulation gain . in the current implementation ,
an off - chip power splitter divides the input power and delivers it to the two on - chip tibs . during operation
, this distributed network suffers from return loss , as each tib is mismatched from 50 ohms during portions of the modulation period ( a portion of this reflected power is dissipated in the power divider at the input , and another portion exits the device s input as return loss ) .
it also results in an additional 3 db of insertion loss ; when signals in the two arms of the interferometer are summed in the 90 degree hybrid , the image ( undesired ) sideband is dissipated in the terminated port of the hybrid . to remedy this ,
we propose using a 90-degree hybrid coupler as the power - divider at the device s input , and combining the signals in the interferometer with a one - to - one transformer . although this layout does not recover the 3 db of return and insertion loss at the device s input , it eliminates the 3 db of insertion loss from the image sideband at the output .
an ssbm designed in this way could operate with a modulation gain of -3 db . | we introduce and experimentally characterize a superconducting single - sideband modulator compatible with cryogenic microwave circuits , and propose its use for frequency domain multiplexing of superconducting qubit readout .
the monolithic single - quadrature modulators that comprise the device are formed with purely reactive elements ( capacitors and josephson junction inductors ) and require no microwave - frequency control tones .
microwave signals in the 4 to 8 ghz band , with power up to -85 dbm , are converted up or down in frequency by as much as 120 mhz .
spurious harmonics in the device can be suppressed by up to 25 db for select probe and modulation frequencies .
recent advances have allowed many groups to demonstrate superb control over one or several superconducting qubits .
this is seen , for example , in the creation and transmission of fock and cat state superpositions@xcite , or the ability to continuously tune both the strength@xcite and axis@xcite of a quantum non - demolition@xcite measurement .
efforts to combine multiple superconducting qubits into a ( more coherent ) logical qubit are progressing in parallel.@xcite as algorithms on multiple logical qubits become experimentally realizable , the challenge of scaling readout for many - qubit systems grows in relevance .
multiplexing , a signal processing technique in which a transmission medium is shared among multiple signals , is a natural way to combat such a scaling challenge . in this letter
, we introduce a single - sideband modulator ( ssbm ) for cryogenic microwave applications and propose its use for frequency domain multiplexing of superconducting qubit readout .
we describe its design , theory of operation , and experimental performance . in a circuit quantum electrodynamics architecture
, qubits are readout by measuring the transmission of a cavity that has a resonance frequency that is modified ( dressed ) by a qubit.@xcite specifically , the cavity resonance is dressed by an amount that depends on the qubit state .
to - date , multi - qubit experiments have scaled - up this readout protocol directly @xcite or with frequency domain multiplexing ( fdm ) hardwired into the measurement architecture.@xcite in direct `` brute - force '' scaling , each qubit channel has a dedicated input and output line .
conversely , qubits in an fdm architecture can share a single input and output line , and their readout cavities are designed in advance to ensure their dressed resonances are separated by multiple cavity linewidths ( spectrally resolvable ) . as the number of readout channels in a multi - qubit measurement rises , direct scaling becomes increasingly untenable .
recognition of this fact has spurred new proposals for scalable architectures @xcite , as well as a slew of devices for in - fridge signal processing , from non - reciprocal elements @xcite to mixers @xcite and switches.@xcite the `` hard - wired '' fdm approach achieves scalability at the cost of flexibility .
the spectral allotment for each channel is determined prior to fabrication and typically@xcite remains fixed through the measurement process .
this constraint inhibits efficient use of the available bandwidth , as densely packing the spectra allotted for each channel makes nearby channels sensitive to any deviation from the designed dressed cavity frequencies , linewidths , or dispersive shifts .
it also precludes the use of identical qubit - cavities , which may be desirable in quantum simulation applications .
finally , the hard - wired fdm architecture fundamentally limits the degree to which a many - qubit experiment can be reconfigured without fabrication of new devices .
there is thus a need for _ flexible _ multiplexed architectures that _ efficiently _ utilize the available measurement bandwidth . to this end
, we introduce a microwave component for cryogenic analog signal processing , which could enable flexible fdm in a many - qubit measurement .
this component is a single - sideband modulator ( ssbm ) , a two - port device that converts incident signals up or down in frequency . unlike commercial ssbms which utilize resistive diodes , in our device
the nonlinear elements are purely reactive . for such systems ,
the manley - rowe relations permit frequency - conversion without dissipation.@xcite moreover , and in contrast to other cryogenic microwave frequency converters@xcite , the nonlinear elements are driven by rf signals with frequencies no greater than several hundred mhz , obviating the need for high - bandwidth ghz control lines .
finally , the device can be realized as a monolithic integrated circuit in a nbalo@xmath0nb tri - layer process @xcite , allowing for high - yield wafer - scale production .
a proposed use for the device is shown in fig .
[ fig : fig0 ] . by positioning a ssbm downstream of the qubit / cavity system in each measurement channel
, a single microwave line can be used to readout all the channels simultaneously in an fdm scheme .
measurement bandwidth can be dynamically allocated among channels by converting each transmitted tone to its assigned band .
this ability to assign spectrum in - situ facilitates efficient use of the available measurement bandwidth , as there is no risk of spectral overlap from uncertainty in fabrication parameters . furthermore , it allows for the use of fdm even if the dressed cavity frequencies are spectrally irresolvable .
although other multiplexing proposals for time and code - domain schemes@xcite share these benefits , the flexible fdm architecture depicted in fig .
[ fig : fig0 ] distinguishes itself by requiring no timing coordination between channels .
the ssbm is created from a hartley - type@xcite in - phase quadrature ( iq ) modulator with i and q ports driven 90 degrees out of phase .
this iq modulator is itself built from a pair of double - balanced modulators ( fig .
[ fig : fig1]a ) realized with inductive bridge circuits ( fig .
[ fig : fig1]b ) , which we call tunable inductor bridges ( tibs).@xcite two pairs of tunable inductors form the bridge , with inductances @xmath1 and @xmath2 .
signals couple differentially to the input port ( top and bottom bridge nodes ) and output port ( left and right bridge nodes ) .
transmission between these ports is controlled by the imbalance in the bridge @xmath3 .
and @xmath2 arranged opposite one - another .
input and output ports are defined by the east - west and north - south nodes of the bridge .
transmission through the bridge scales with the inductor imbalance @xmath4 .
( c ) a tunable inductor bridge realized with arrays of squids .
bridge imbalance is modulated with a static background magnetic flux applied with an off - chip magnetic coil and a time - dependent gradiometric flux applied with an on - chip bias line .
( d ) schematic of an ssbm composed of a power divider , two double - balanced modulators with i ports driven in quadrature , and a 90 degree hybrid coupler .
( e ) phasor representation of signals traveling through an idealized ssbm .
modulated signals are recombined using a hybrid coupler and the output signal is converted up in frequency by @xmath5 . ] manipulation of that imbalance is accomplished through the use of four flux - tunable inductors arranged in a figure - eight geometry , which tune in concert to balance ( or imbalance ) the bridge .
their coordinated tuning is controlled with an off - chip magnetic coil and an on - chip bias line that bisects the figure eight ( fig .
[ fig : fig1]c).@xcite series arrays of superconducting quantum interference devices ( squids ) are used to realize the flux - tunable inductors .
arrays are employed in place of individual squids to dilute the josephson nonlinearity.@xcite to operate the tib as a double - balanced modulator , we use the flux - control bias line to sinusoidally modulate the transmission through the bridge at angular frequency @xmath5.@xcite borrowing language from the field of nonlinear - microwave elements , we denote the two microwave ports of the tib as the local oscillator ( l ) and radio frequency ( r ) ports of the modulator , and the flux - control line as the intermediate frequency ( i ) port .
however , as a tib has no galvanic connection between its i and l or i and r ports , we refer to it as a modulator rather than a mixer .
note also that in contrast to a common way of operating mixers by driving their nonlinear elements with a strong signal in their l ports , the i port of the tib actuates the modulation process and signals at the l and r ports can be arbitrarily small .
we create an iq modulator by dividing the power of an input tone into two double - balanced modulators and then summing their outputs with a 90 degree hybrid coupler ( fig .
[ fig : fig1]d ) . in the device
reported in this letter , both double - balanced modulators are integrated on a single chip and connected to a commercial power splitter and 90 degree hybrid
. future versions of the ssbm , however , could also integrate the passive components on - chip.@xcite fig .
[ fig : fig1]e provides a phasor representation of a microwave tone as it propagates through the ssbm .
this tone is first split into two arms with equal phase shift , each connected to the l port of a double - balanced modulator .
control currents are applied to the i ports of both modulators , modulating the signal in each arm into two tones spectrally shifted from the original by @xmath6 ( blue / red sidebands , respectively ) .
the phase of the control current in the lower arm , @xmath7 , is advanced @xmath8 radians with respect to that in the upper .
recombination at the 90 degree hybrid causes the blue ( red ) sidebands in each arm to constructively ( destructively ) interfere at the device s output .
the input tone is thus converted up in frequency by @xmath5 .
generally , the amplitude of the red and blue sidebands scales as @xmath9 , allowing for selection of the red sideband when @xmath10 radians .
we denote the i port in the top arm as the i port of the iq modulator , and the i port of the double - balanced modulator in the bottom arm as the q port .
d , as a function of iq phase difference , @xmath7 , with l port driven at @xmath11 ghz and transmission modulated at @xmath12 3 mhz .
the first upper and lower sidebands are shown in solid blue and solid red , respectively .
they exceed the input frequency ( green ) by up to 30 db .
power is also detected at higher harmonics of the modulation frequency .
for the plotted l and i frequencies , the largest harmonic is the fifth ( dashed blue and red , respectively ) . a gray dashed line at -3 db is a guide to the eye .
at the operating points @xmath13 , the power in the first sideband exceeds all other harmonics by more than 20 db . ]
this angular dependence of sideband power is depicted in fig .
[ fig : fig2 ] , which shows the transmitted power at various frequencies as a function of the iq phase difference @xmath7 . for this measurement ,
the ssbm is driven by a microwave tone at @xmath11 ghz through its l port while we modulate the transmission across the bridges at @xmath14 mhz .
as this modulation is not purely comprised of a single spectral component at angular frequency @xmath5 , higher harmonics of @xmath5 at frequencies @xmath15 , with @xmath16 an integer greater than 1 , are also observed .
we monitored the nearest 16 higher harmonics ( @xmath17 ) , and found the @xmath18 harmonic to be the largest at this particular choice of @xmath19 and @xmath5 .
the plot shows the input frequency ( green ) , along with the first upper ( blue ) and lower ( red ) sidebands , and the largest of the higher harmonics ( @xmath18 , dashed red and blue ) .
more harmonics and phase sweeps at different iq modulation frequencies are shown in fig .
s3 of the supplementary material .
as @xmath7 is varied between @xmath20 and @xmath21 we observe that all measured harmonics oscillate with @xmath7 at a rate commensurate with their order .
two clear operation points are visible at @xmath13 radians , where the first blue ( red ) sideband is suppressed in favor of the first red ( blue ) sideband .
the contrast between them exceeds 30 db at these points , and the contrast between the desired sideband and the nearest harmonic ( the 5th , in this case ) , which we call the sideband contrast , is approximately 23 db .
we set the scale of the y - axis in units of db relative to an unmodulated input tone , which we call modulation gain .
this is done by statically biasing both tibs to their state of maximal imbalance .
we normalize the sideband power in this way to reveal inefficiencies in the modulation process and separate them from the intrinsic return and insertion loss of the ssbm s constituent components . in fig .
[ fig : fig2 ] , the modulation gain in the desired sideband never exceeds -3 db ( dashed gray line ) because the tib in each arm of the interferometer is biased to modulate transmission sinusoidally .
as the device is realized as a distributed network , that means each arm is reflecting half its incident power on average power which is ultimately dissipated in the power splitter , or directed out the device s input port ( return loss).@xcite to characterize the device over a broader range of operation frequencies , sideband contrast for a variety of l and i frequencies is plotted in fig .
[ fig : fig3]a . in the proposed application of the ssbm for fdm of qubit readout ,
this is an important measure of device performance : when the measurement spectrum is densely filled , spurious sidebands are a source of cross - talk among channels . .
the different traces depict different l frequencies @xmath22 ( color ) , for selection of the blue ( solid lines ) and red ( dashed lines ) sidebands . ] in addition to sideband contrast , other important specifications for a ssbm are the modulation gain , l , r , and i frequency limits , instantaneous bandwidth , linearity , and return and insertion losses .
modulation gain is plotted in fig .
[ fig : fig3]b as a function of @xmath5 , for tones at several different input frequencies .
two major trends are visible : first , modulation gain decreases as @xmath5 increases .
this effect is a consequence of the bandwidth requirement of the pulse - shaping scheme used to sinusoidally modulate transmission through the tibs , which begins to exceed our control hardware s bandwidth as @xmath5 increases.@xcite the second trend is a variation in the modulation gain achieved at small modulation frequencies as the probe frequency @xmath19 changes .
this is caused by a change in the phase of a tib s transmission as its magnitude is modulated .
this effect is evident in fig .
s2 of the supplementary material , and may be partially alleviated in a future device by removal of a chip - mode at 5 ghz .
figs .
[ fig : fig3]a and b also show the ranges of l , r , and i frequencies that can be processed by the ssbm .
the device converts frequencies from / into the 4 to 8 ghz band , shifting them by as much as 120 mhz to the red or blue . for an fdm application in superconducting qubit readout , the quantity with which to compare
this is the coupling strength of a readout cavity s strongly coupled port @xmath23 .
allowing 3 @xmath23 of spectrum per readout channel , and taking a representative value of several mhz for @xmath24@xcite , such an i range allows for 10 to 40 ( or more@xcite ) independent readout channels per input line .
conversion of amplitude - modulated tones with the ssbm , and measurements of its linearity , are shown in supplementary figs .
s4 and s5 .
the instantaneous bandwidth of the device exceeds 5 mhz , and its 1 db compression point is -85 dbm , providing sufficient bandwidth and power - handling for dispersive qubit readout .
we now discuss return and insertion loss of our ssbm . although the insertion loss of the tibs is less than 0.5 db@xcite , our ssbm dissipates 3 db of the input power ( the entire image sideband ) in the fourth port of the 90 degree hybrid at the device s output .
as noted previously , an additional 3 db ( or more ) of power is dissipated / returned by our distributed network when the tibs are operated as modulators this is precisely the modulation gain plotted in fig .
[ fig : fig3]b .
finally , the tibs themselves are not perfectly transmitting over the entire range of operating frequencies . to quantify this ,
unnormalized transmission in the 4 - 8 ghz band is plotted in fig .
s2 of the supplementary material , with independent measurements on both tibs .
although this transmission approaches unity at some frequencies , it is degraded elsewhere by the presence of a chip mode around 5 ghz and impedance mismatches outside the 5 - 7 ghz band .
a future integrated design for the ssbm could improve on many of these specifications : removal of the chip mode and matching the circuit impedance over the entire 4 - 8 ghz band are well - posed problems in microwave engineering .
an on - chip implementation will improve sideband contrast by suppressing path length differences in the network below an operation wavelength and shrink device footprint by two orders of magnitude .
an additional 3 db of conversion gain can also be recovered by using the quadrature hybrid as the power divider on the input and employing a one - to - one transformer to combine the outputs of the two tibs this eliminates the power dissipated in the image sideband .
a next - generation device could thus operate over a several ghz frequency range with negligible return and insertion loss and modulation gain approaching -3 db .
further design considerations for a future device are provided in the supplementary material .
a single sideband modulator constructed from the repeated instancing of broadband modulation elements like tibs is thus an appealing , general purpose way to engineer frequency conversion .
the l and r bandwidths allow for conversion of tones in the 4 - 8 ghz band , and the device s 1 db compression point at -85 dbm far exceeds the power typically@xcite used for dispersive readout . in addition , low return , insertion , and modulation losses are achievable in future design iterations .
finally , the ssbm s nonlinear elements require no ghz frequency control lines , and are actuated solely with radio frequency signals ( several hundred mhz or less ) .
it is therefore suitable for construction of the scalable and flexible multiplexed architectures needed in future many - qubit experiments with superconducting circuits .
_ acknowledgment _ this work is supported by the aro under contract w911nf-14 - 1 - 0079 and the national science foundation under grant number 1125844 .
the authors thank bradley a. moores and andrew p. higginbotham for helpful discussions .
39ifxundefined [ 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 * * , ( ) \doibase http://dx.doi.org/10.1063/1.4764940 [ * * , ( ) , http://dx.doi.org/10.1063/1.4764940 ] @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevapplied.4.034002 [ * * , ( ) ] @noop * * , ( ) @noop ( ) @noop ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) link:\doibase 10.1103/physrevapplied.6.024009 [ * * , ( ) ] @noop @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ , vol .
( , ) @noop @noop * * , ( ) @noop link:\doibase 10.1103/physreva.93.033817 [ * * , ( ) ] @noop * * , ( ) @noop @noop * * , ( ) @noop * * , ( ) * supplementary material for + `` single - sideband modulator for frequency domain multiplexing of superconducting qubit readout '' * |
we have explored two avenues for the _ ab initio _ calculations of nuclear - reaction properties starting from the nuclear structure of the ncsm @xcite : i ) the lorentz integral transform ( lit ) approach @xcite , and ii ) the resonating - group method ( rgm ) @xcite . in sect .
[ lit ] , we highlight a recent application of the ncsm / lit approach to the _ ab initio _
calculation @xcite of the @xmath0he total photo - absorption cross section . in sect .
[ rgm ] , we outline the ncsm / rgm approach and present its application to neutron-@xmath0he scattering .
we performed an _ ab initio _
calculation @xcite of the @xmath0he total photo absorption cross section in unretarded dipole approximation , using the high quality nn potential at the fourth order ( n@xmath1lo ) in the chiral effective field theory ( @xmath2eft ) expansion of ref .
@xcite , and the nnn interaction at the highest order presently available ( n@xmath3lo ) @xcite .
the two low - energy constants that parameterize the short - range nnn interaction were selected as discussed in ref .
the microscopic treatment of the continuum problem is achieved by means of the lit method , applied within the ncsm approach .
accurate convergence of the ncsm expansions is reached using three - body effective interactions for both the nn and nnn terms of the potential .
the results , presented in fig
. [ he4cs ] , show a peak around the excitation energy of @xmath4 mev , with peak - height mildly sensitive to the nnn force .
the experimental situation in the near - threshold region is controversial @xcite : we find an overall good agreement with the photo - disintegration data from bremsstrahlung photons of nilsson et al . and the indirect measurements of nakayama et al .
for reactions involving two nuclear fragmets , the rgm many - body wave function is expanded in terms of binary - cluster channel wave functions @xmath5
= \sum_{\nu}\int d\vec{r}\,\varphi_{\nu}(\vec{r}\,)\,\hat{\mathcal a}\,\phi_{\nu\vec{r}}^{(a - a , a)}\,,\ ] ] with @xmath6 here , @xmath7 is the antisymmetrizer accounting for the exchanges of nucleons between the two clusters ( which are already antisymmetric with respect to exchanges of internal nucleons ) . the relative - motion wave functions @xmath8 depend on the relative distance between the center of masses of the two clusters in channel @xmath9 .
they can be determined by solving the many - body schrdinger equation in the hilbert space spanned by the basis functions ( [ rgm_basis ] ) : @xmath10\varphi_{\nu}(\vec{r}\,)\,,\ ] ] where @xmath11 denotes the full hamiltonian of the system , and the two integration kernels are defined as @xmath12 in the ncsm / rgm approach we use _ ab initio _ ncsm wave functions for each of the two clusters , and ncsm effective interactions derived from realistic forces .
the most challenging task is to evaluate the hamiltonian and norm kernels : so far we have developed formalism for cluster states with a single - nucleon projectile and we have applied it to the scattering of low - energy neutrons on @xmath0he . in fig .
[ figphase ] , we present the converged @xmath13 phase shift for the @xmath14he system obtained using the bare @xmath15 @xcite potential , and two - body effective interactions derived from the @xmath2eft nn potential of ref . @xcite .
we find an overall good agreement with experiment .
this work was performed under the auspices of the u.s .
department of energy by lawrence livermore national laboratory under contract de - ac52 - 07na27344 .
support from u.s .
doe / sc / np ( work proposal number scw0498 ) and the department of energy under grant de - fc02 - 07er41457 is acknowledged . 9 navrtil , p. , et al .
. lett . * 84 * , 5728 ( 2000 ) ; phys . rev .
* c62 * , 054311 ( 2000 ) ; phys .
lett . * 88 * , 152502 ( 2002 ) ; phys . rev . *
c68 * , 034305 ( 2003 ) efros , v. d. , et al .
: phys . lett . *
b338 * , 130 ( 1994 ) tang , y. c. , et al : phys . rep .
* 47 * , 167 ( 1978 ) ; langanke , k. , friedrich , h. : advances in nuclear physics .
newyork : plenum ( 1987 ) quaglioni , s. , navrtil , p. : phys . lett . *
b652 * , 370 ( 2007 ) entem , d. r. , machleidt , r. : phys . rev . *
c68 * , 041001(r ) ( 2003 ) van kolck , u. : phys . rev .
* c49 * , 2932 ( 1994 ) ; epelbaum , e. , et al .
: phys . rev .
* c66 * , 064001 ( 2002 ) navrtil , p. , et al .
lett . * 99 * , 042501 ( 2007 ) shima , t. , et al . : phys . rev . *
c72 * , 044004 ( 2005 ) ; nillson , b. , et al .
: phys . rev .
* c75 * , 014007 ( 2007 ) ; nakayama , s. , et al .
c76 , 021305(r ) ( 2007 ) bogner , s. k. , kuo , t. t. s. , schwenk , a. : phys . rep .
* 386 * , 1 ( 2003 ) | we report on recent microscopic calculations of reaction properties based upon the nuclear structure of the _ ab initio _ no - core shell model ( ncsm ) . |
planetary systems are assembled from dust and gas in the disks surrounding pre - main sequence stars .
the nature of the formed planets are intimately linked to the structure , composition and evolution of the parent circumstellar disk .
molecular emission lines serve as probes of disk characteristics , such as density , temperature and ionization fraction , that are not accessible by other observations .
for example , n@xmath0h@xmath1 along with the deuterated ions like h@xmath0d@xmath1 and dco@xmath1 are believed to trace the ionization fraction near the midplane of the disks @xcite .
molecular distributions in disks are also important to characterize because of their connection with the composition of forming planetesimals .
this especially true of organic molecules , of which h@xmath0co is an important representative .
while most molecules are expected to be reprocessed in larger planetary bodies , the disk - chemical composition may survive quite intact in icy planetesimals , including comets @xcite .
such planetesimals may have seeded the earth with water and organics , connecting disk chemistry with the origins of life . predicting the organic composition of these planetesimals depend on our understanding of the distribution of the organic composition of grains in the disks .
the overall disk chemical structure is set by a combination of photochemistry at the disk surface and sequential freeze - out of molecules in the disk interior ( e.g. * ? ? ?
the distributions of n@xmath0h@xmath1 and h@xmath0co and their relationship to co present important test cases of the chemical models .
the co molecule is one of the last to freeze out and predicted to deplete quickly from the gas - phase at t@xmath5 k for typical disk mid - plane densities .
abundant n@xmath0h@xmath1 is expected only where co is depleted because n@xmath0h@xmath1 forms from protonation of n@xmath0 , which remains in the gas - phase at temperatures a few degrees lower than co @xcite , and is destroyed mainly by reactions with co @xcite . in dense cores in star forming regions , the abundance of co is observed to show a strong anti - correlation with n@xmath0h@xmath1 ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in disk models ,
this effect is manifested as a jump in the n@xmath0h@xmath1 column density at the co `` snow line '' , one order of magnitude in a recent calculation @xcite .
the snow line is here defined as the disk radius where the midplane dust temperature is cold enough for volatiles to condense into ice grains .
the h@xmath0co chemistry is more complicated than n@xmath0h@xmath1 because h@xmath0co can form through several different pathways , both in the gas - phase and on grain - surfaces .
grain - surface formation of h@xmath0co should depend directly on co freeze - out ; constrained by theory @xcite and experiments ( e.g. * ? ? ?
* ; * ? ? ?
* ) , h@xmath0co ( and ch@xmath4oh ) form readily from co ice hydrogenation .
if this is the dominant formation pathway of h@xmath0co in disks , and if ices are partially desorbed non - thermally ( e.g. * ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , then h@xmath0co gas should coincide with n@xmath0h@xmath1 exterior to the co snow line . to date , emission from millimeter
wavelength h@xmath0co and n@xmath0h@xmath1 lines has been detected toward 8 and 6 protoplanetary disks , respectively @xcite .
most detections are toward t tauri stars with massive disks , and the detection fraction toward more luminous herbig ae stars is low @xcite . based on these observations , h@xmath0co and n@xmath0h@xmath1 are mainly abundant in disks with large reservoirs of cold dust and gas , where co freeze - out is expected to occur .
a direct connection between co freeze - out and n@xmath0h@xmath1 and h@xmath0co in disks has yet to be observationally established , however .
in this paper we present submillimeter array ( sma ) observations of h@xmath0co and n@xmath0h@xmath1 toward the disks around hd 163296 and tw hya , and we use these new observations along with h@xmath0co and n@xmath0h@xmath1 observations from discs ( disk imaging survey of chemistry with sma ) to constrain the h@xmath0co and n@xmath0h@xmath1 distributions .
the new data and their calibration are described in [ sec : obs ] . in [ sec : res ] , we present the h@xmath0co and n@xmath0h@xmath1 images and spectra toward hd 163296 and tw hya , models of the h@xmath0co distribution toward hd 163296 , h@xmath0co excitation temperature calculations , and examine the relationship between h@xmath0co and n@xmath0h@xmath1 emission across the sample of disks . in [ sec : disc ] , we discuss the implications of these results , summarize the mounting evidence for co - ice regulated chemistry and make predictions for future observations of h@xmath0co and n@xmath0h@xmath1 emission from disks with better sensitivity and resolution .
the observations of hd 163296 ( @xmath6 , @xmath7 ; j2000.0 ) were made between 2008 and 2012 , and of tw hydrae ( @xmath8 , @xmath9 ; j2000.0 ) between 2008 and 2012 , using the eight - antenna submillimeter array ( sma ) located atop mauna kea , hawaii . table [ tab : obs ] provides a summary of the observational parameters and results .
for the 2007 and 2008 observations , the sma receivers operated in a double - sideband mode with an intermediate frequency ( if ) band of 46 ghz from the local oscillator frequency , sent over fiber optic transmission lines to 24 overlapping `` chunks '' of the digital correlator .
the 2012 observations were made after an upgrade that enabled a second if band of 68 ghz , effectively doubling the bandwidth .
the sma observations of hd 163296 were carried out in the compact - north ( com - n ) , compact ( com ) and subcompact ( sub ) array configurations .
the 2007 observations included the dco@xmath1 32 line at 216.1126 ghz and the h@xmath0co @xmath10 line at 225.698 ghz .
the 2012 observations included the n@xmath0h@xmath1 32 at 279.512 ghz and the h@xmath0co @xmath11 line at 281.527 ghz .
the observing loops used j1733130 as the main gain calibrator and observed j1744312 every other cycle to check the phase calibration .
flux calibration was done using observations of titan and uranus .
the derived fluxes of j1733 - 130 were 1.19 jy ( 2007 mar 20 ) , 1.25 jy ( 2012 jun 10 ) , 1.40 jy ( 2012 aug 12 and 14 ) .
the bandpass response was calibrated using observations of 3c279 , uranus and j1924292 .
the sma observations of tw hya were carried out in the compact ( com ) and subcompact ( sub ) array configurations .
the 2008 observations included the h@xmath0co @xmath12 line at 351.769 ghz .
the 2012 jan 13 sub observation included the h@xmath0co @xmath11 and n@xmath0h@xmath1 32 lines , like hd 163296 , but unfortunately the chunk containing n@xmath0h@xmath1 was corrupted and unusable .
the 2012 jun 04 com observation using a similar setting successfully included the n@xmath0h@xmath1 line .
the observing loops used j1037295 as the gain calibrator .
flux calibration was done using observations of titan and callisto .
the derived fluxes of j1037295 were 0.73 jy ( 2008 feb 23 ) , 0.73 jy ( 2012 jan 13 ) and 0.82 jy ( 2012 jun 4 ) .
the bandpass response was calibrated using observations of 3c279 and 3c273 .
routine calibration tasks were performed using the mir software package , cqi / mircook.html ] and imaging and deconvolution were accomplished in the miriad software package .
in this section we present detections of h@xmath0co and n@xmath0h@xmath1 emission lines toward hd 163296 and tw hya , display their respective distributions ( [ det ] ) , and compare the higher quality observations of h@xmath0co in hd 163296 with models ( [ model ] ) .
we then combine the new data with previously reported h@xmath0co and n@xmath0h@xmath1 detections in disks to examine trends with respect to h@xmath0co excitation temperature ( [ temp ] ) and with each other ( [ corr ] ) .
figure [ fig : obs_mom ] shows images of the spectrally integrated emission toward tw hya and hd 163296 at the rest frequencies of two h@xmath0co lines and the n@xmath0h@xmath1 @xmath13 line .
h@xmath0co @xmath11 and @xmath12 and n@xmath0h@xmath1 are detected toward tw hya , and h@xmath0co @xmath10 and h@xmath0co @xmath11 and n@xmath0h@xmath1 are detected toward hd 163296 .
the emission toward tw hya appears to be centrally peaked at the size scale of the beams ( fwhm @xmath14 ) . toward hd 163296 , however , neither the h@xmath0co lines nor the n@xmath0h@xmath1 line emission peaks at the location of the continuum peak that marks the stellar position , but instead show significant offsets .
this is most readily apparent for the h@xmath0co @xmath10 line that was observed with a slightly smaller and more advantageously rotated beam , where the emission appears ring - like .
this is the second reported observation of a ring - like h@xmath0co distribution after dm tau @xcite . interpreting the h@xmath0co emission toward dm tau is however complicated by a large central cavity in dust emission @xcite .
figure [ fig : obs_spec ] shows the spatially integrated spectra .
the line shapes and central velocities agree with what has been previously observed for other molecular lines toward these disks .
the line fluxes are listed in table [ tab : obs ] , and the values are comparable to detections of these lines toward other large protoplanetary disks @xcite .
figure [ fig : obs_mom ] demonstrates that h@xmath0co emission toward hd 163296 is spatially resolved and thus contains information on the radial distribution of h@xmath0co in the disk .
the relative excitation of the two h@xmath0co transitions should probe primarily the vertical distribution and thus provide complementary constraints .
we explore the h@xmath0co distribution based on a previously developed accretion disk model with a well - defined temperature and density structure , constrained by the hd 163296 broadband spectral energy distribution , spatially resolved millimeter dust continuum , and multiple co and co isotopologue line observations @xcite .
we adopt the same methods as @xcite for constraining the h@xmath0co distribution , here fitting models that assume a radial power - law ( [ power ] ) and a simple ring with inner boundary at the co `` snow line '' ( [ ring ] ) . for a first - order analysis of the distribution of h@xmath0co , we model the radial variation in the column density as a power law n@xmath15 between an inner radius r@xmath16 and outer radius r@xmath17 , where n@xmath18 is the column density at 100 au in @xmath19 , @xmath20 is the distance from the star in au , and @xmath21 is the power - law index . for the vertical distribution , we assume that h@xmath0co is present with a constant abundance in a layer with boundaries toward the midplane and toward the surface of the disk ( similar to @xcite ) .
this assumption is motivated by chemical models ( e.g. * ? ? ?
* ) that predict a three - layered structure where most molecules are photodissociated in the surface layer , frozen out in the midplane , and have an abundance that peaks at intermediate disk heights .
the surface ( @xmath22 ) and midplane ( @xmath23 ) boundaries are presented in terms of @xmath24 , where @xmath25 is the hydrogen column density measured from the disk surface .
this simple model approximates the vertical location where h@xmath0co is most abundant .
the excitation of multiple transitions can constrain both @xmath22 and @xmath23 , but in this case of very modest signal - to - noise , we fix @xmath22 to 0.79 , the surface boundary found for co by @xcite , and we fit @xmath23 for the midplane boundary and the power - law parameters ( n@xmath18 , @xmath21 , r@xmath16 and r@xmath17 ) . using the structure model , we compute a grid of synthetic h@xmath0co visibility datasets over a range of r@xmath17 , r@xmath16 , @xmath21 , @xmath23 and n@xmath18 values and compare with the observations .
the best - fit model is obtained by minimizing @xmath26 , the weighted difference between the real and imaginary part of the complex visibility measured in the ( @xmath27)-plane sampled by the sma observations of both h@xmath0co transitions .
we use the two - dimensional monte carlo model ratran @xcite to calculate the radiative transfer and molecular excitation .
the collisional rates are taken from the leiden atomic and molecular database @xcite .
table [ tab : best - fit ] lists the best - fit parameters of the model .
the power - law index of 2 implies an h@xmath0co column density that strongly increases with radius .
figure [ fig : model ] shows the best - fit radial distribution of the h@xmath0co column density .
figures [ fig:32_chmap ] and [ fig:43_chmap ] present comparisons between the observed channel maps and the best - fit model .
the model reproduces the main features of the observations remarkably well , in particular the flux ratio between the inner and outer channels , and the lack of emission at the location of the continuum peak .
figure [ fig : chi2 ] shows the @xmath26 surfaces for the r@xmath16 and r@xmath17 versus the power law index @xmath21 , which enables us to quantify the uncertainties associated with the inner and outer region sizes and the power - law index .
we find that @xmath21 is constrained between 0.53.0 ( within 1@xmath28 ) while r@xmath16 is constrained to be @xmath29200 au .
the @xmath26 value does not change significantly for inner radii @xmath2990 au , as expected for the 2@xmath30 beam size of the observations .
the outer radius is better constrained , since the emission is very sensitive to the value of r@xmath17 with a positive power - law index .
figure [ fig : mod_spec ] shows the h@xmath0co 32 line spectrum compared with the spectra derived from models with different radial column densities power - law indices .
the spectra suggest a lack of high velocity line wings associated with emission originating in the inner regions of the disk , consistent with the results of the @xmath26 analysis .
the positive power - law index found for the h@xmath0co radial distribution implies that h@xmath0co is present mainly in the outer disk .
this is expected if h@xmath0co forms _ in situ _ from co ice hydrogenation and is therefore present mainly beyond the co snow - line , previously determined to be at 160 au by @xcite
. guided by this astrochemical _ ansatz _ , we have tried a second `` ring '' model where the h@xmath0co gas is only present where co has frozen out .
the vertical surface boundary is then defined by the co freeze - out temperature of 19 k @xcite , while the midplane boundary can be constrained by the excitation of multiple h@xmath0co transitions , as in the power - law model . within this layer
, the abundance of h@xmath0co is assumed to follow that of h nuclei with a constant fractional abundance , which is also a parameter fit to the data .
the best - fit abundance is @xmath31 and the midplane boundary @xmath23 is consistent with what we find in the power - law model ( table [ tab : best - fit ] ) .
figure [ fig : model ] shows that the vertical surface boundary at 19 k effectively results in a ring - like radial structure , where the inner edge of the ring is at co snow line .
figure [ fig : model ] also shows that the best - fit power - law and ring models result in similar h@xmath0co column densities beyond the co snow line .
the profile of the ring model is considerably flatter than the power - law model and even drops outside of 300 au exponentially to the edge of co emission .
figures [ fig:32_chmap][fig:43_chmap ] show that the power - law model and the ring model channel maps display some subtle differences . but both provide good fits to the data within the noise of the sma observations .
the same model also provides a good match to the n@xmath0h@xmath1 flux , but the combination of low signal - to - noise and observations of just one transition preclude any independent modeling of the n@xmath0h@xmath1 distribution .
hd 163296 is very favorable for h@xmath0co and n@xmath0h@xmath1 imaging since its relatively high luminosity and massive disk puts the co snow line at a large angular distance compared to other disks , enabling us to resolve the h@xmath0co and n@xmath0h@xmath1 emission . without such spatial information , however , we can still obtain a constraint on where h@xmath0co
emission originates in disks based on the average h@xmath0co excitation temperatures . as a gross approximation
, h@xmath0co that coexists with co ice should be cold ,
i.e. present at an excitation temperature comparable to the co freeze - out temperature . to test the viability of using excitation temperatures to constrain the h@xmath0co distribution we extracted spectra from three of the hd 163296 simulations presented in 3.2 , selecting models with the best outer radius , no inner hole and power - law indices of @xmath32 , @xmath33 , and @xmath34
. these distributions approximately correspond to a h@xmath0co abundance that follows the h@xmath0 column , that keeps constant with radius and that increases steeply , forming a ring .
assuming lte , that both h@xmath0co transitions trace the same underlying populations , and a single rotational excitation temperature , @xmath35 , we calculate @xmath36 where @xmath37 and @xmath38 are the upper energy levels for the low and high h@xmath0co transitions used in the calculation ( h@xmath0co @xmath39 and @xmath11 for most disks ) , @xmath40 and @xmath41 the corresponding line frequencies and temperature independent transition strengths and dipole moments , @xmath42 the integrated line intensity , which is calculated from the integrated fluxes based on @xmath43 , where @xmath44 is the flux in jy , @xmath45 the intensity in k , @xmath46 the line wavelength in millimeters , and @xmath47 and @xmath48 the emission diameters in @xmath30 .
because of both vertical and radial temperature gradients in the disk , the size dependence of the emission regions from two transitions is complicated but the emission area is not expected to be very different . for simplicity
we assume the extent of the emission is the same for both transitions on account of the model dependent effects of the temperature gradients .
all of the line parameters were gathered from splatalogue ( a transition - resolved compilation of several spectroscopic databases ) , with the data originating from cdms @xcite .
lte is a reasonable approximation if h@xmath0co is mainly present at high densities .
the critical densities for the observed h@xmath0co transitions vary between 10@xmath49 and 5@xmath5010@xmath51 @xmath52 @xcite , dependent on assumed kinetic temperatures . at radii @xmath29300 au ,
all gas colder than 25 k is at densities higher than 10@xmath53 @xmath52 @xcite , justifying this assumption . using the simulated spectra we derive an excitation temperature of 27 k for the model with @xmath54 and excitation temperatures close to or below 20 k for the other two models .
excitation temperatures thus provide some constraints on the h@xmath0co distribution , but a temperature close to the expected co freeze - out temperature only implies an increasing abundance with radius ; it can not be used to assess how steeply the abundance increases . by combining the new h@xmath0co detections toward tw hya and hd 163296 with h@xmath0co data from discs @xcite
, we have a sample of 10 disks with two h@xmath0co line detections or one h@xmath0co line detection and one upper limit ( table [ tab : star ] ) .
these data are sufficient to calculate excitation temperatures , albeit with substantial uncertainties .
figure [ fig : temp ] shows the calculated excitation temperatures for 9 of the 10 disks ; the chemically peculiar herbig ae star hd 142527 is not included because its excitation temperature of 250 k is probably not due to thermal excitation .
the h@xmath0co excitation temperature , listed in table [ tab : star ] , is consistent with , or lower than , the co freeze - out temperature of @xmath320 k for all of these disks .
the average h@xmath0co temperature in the sample ( excluding hd 142527 ) is @xmath55 k. we note that the relatively high excitation temperature toward hd 163296 is most likely due to the fact that some h@xmath0co 32 emission is resolved out by the sma observations , as the compact - north antenna configuration has few short baselines .
we examine the disk sources to test if n@xmath0h@xmath1 and h@xmath0co emission are correlated across the sample , as would be expected if 1 ) the two molecules form under the same physio - chemical conditions , i.e. only in the regions where co has frozen out , 2 ) the line emission trace the total n@xmath0h@xmath1 and h@xmath0co column well , and 3 ) midplane ionization levels and co hydrogenation efficiencies do not differ ` too much ' across the sample ( i.e. the size of the co freeze - out region is the most important regulator of n@xmath0h@xmath1 and h@xmath0co column across the sample ) .
it should be noted that there are other scenarios that could produce a correlation as well , and that the correlation analysis below should only be considered as a constraint on the h@xmath0co distribution in combination with the results in the previous sections . in particular a constant h@xmath0co / n@xmath0h@xmath1 across a disk sample
would be expected if the relative fractions of chemically characteristic disk regions is always similar in disks .
to conclusively test this requires a larger sample than currently available , but the fact that we did not find that h@xmath0co emission correlates with any other molecular emission than n@xmath0h@xmath1 already challenges this scenario .
where possible , we base the comparison on the h@xmath0co @xmath39 line that has @xmath56 k , similar to n@xmath0h@xmath1 @xmath13 ( @xmath57 k ) , to minimize variations in fluxes caused by the different detailed temperature structures in different disks . excluding hd 142527 ,
this line has been observed toward 6/9 of the sample . for the remaining 3/9 disks , we calculate the expected h@xmath0co @xmath39 line flux based on the h@xmath0co excitation temperatures and fluxes of other h@xmath0co lines toward each source .
we then normalize the flux of each h@xmath0co @xmath39 line to the ( taurus ) distance of 140 pc , and we further normalize to a disk mass of 0.01 m@xmath58 to account for the fact that more nearby and more massive disks tend to have overall stronger line emission .
figure [ fig : corr ] shows that there is a strong correlation between the normalized h@xmath0co and n@xmath0h@xmath1 fluxes in the disk sample ; the rank correlation is statistically significant at the 95% level .
as expected , figure [ fig : corr2 ] shows that this implies a nearly constant n@xmath0h@xmath1 3 - 2 / h@xmath0co 3@xmath59 flux ratio across the sample .
our modeling strategy in this study and in @xcite has been to first constrain the overall structure of molecular emission in disks using a parametric model with a minimum of free parameters , i.e. to determine whether the radial column density profile of a species decreases , increases or is flat as a function of disk radius .
as demonstrated in @xcite , the slope of the radial column density profile already can place significant constraints on the formation pathway of a molecule . here
we find that h@xmath0co toward hd 163296 belongs to the family of molecules that display an increasing column density with radius .
this first - order constraint on the h@xmath0co distribution motivated us to consider h@xmath0co formation pathways that would result in an increase of h@xmath0co with disk radius .
h@xmath0co formation through co - ice hydrogenation results in a simple prediction that h@xmath0co should be present in a ring , with the inner edge at the co snow line .
we therefore set up a second model , based on this prediction , to test if the observations are consistent with this hypothesis for h@xmath0co formation .
we propose that this combination of backward and forward modeling both provides a fair view of the constraints obtained by fitting the data , and challenges our basic understanding of disk chemistry .
h@xmath0co can form through multiple chemical pathways .
we have shown that the h@xmath0co distribution towards hd 163296 and the sample statistics are consistent with formation through _ in situ _ co ice hydrogenation . here
we consider the effects of additional pathways for h@xmath0co formation , in particular ( 1 ) _ in situ _ gas phase formation , and ( 2 ) formation in the pre- and proto - stellar phases , followed by incorporation into the disk .
h@xmath0co can form in the gas - phase through ion - neutral reactions involving , e.g. ch@xmath60 or through neutral - neutral reactions between ch@xmath4 and o ( e.g. * ? ? ?
the neutral - neutral formation pathway is expected to result in a radially flat column density structure for a typical t tauri disk @xcite , with most emission originating at temperatures of 20 - 40 k @xcite .
it is not clear from existing disk models whether the structure will look substantially different if the neutral - ion reactions dominate . in a protostellar chemistry model @xcite
, ch@xmath60 disappears when co depletes ( e. bergin , private communication ) .
this suggests that h@xmath0co forming through this pathway should be anti - correlated with co - freeze - out .
neither of these gas - phase pathways thus predicts excess h@xmath0co column densities in the outer disk , or at low ( t@xmath2920 k ) temperatures .
the observed low excitation temperature could on its own be explained by efficient turbulent mixing of h@xmath0co formed in the gas - phase and then cooled down in the midplane regions , similarly to what has been proposed to explain cold co gas in disks @xcite .
turbulent mixing would not , however , explain the observed deficiency of h@xmath0co towards the inner disk in hd 163296 .
gas - phase formation of h@xmath0co then seems an unlikely dominant source of h@xmath0co in disks in light of the new observations , but the case is unlikely to be conclusively settled until the exclusive grain - surface product ch@xmath4oh is observed to display a similar distribution .
h@xmath0co in disks could be a product of protostellar or molecular cloud chemistry , as h@xmath0co is commonly observed in pre- and protostellar sources , and this molecular content may be preserved , at least in part , through the process of disk formation .
@xcite has modeled this scenario , starting with a h@xmath0co ice abundance of 10@xmath61 inherited from the cold cloud .
the model also includes h@xmath0co formation through gas - phase processes in the disk and results are presented with and without photodesorption . in the model without photodesorption , h@xmath0co follows co in the inner disk and has an additional outer disk component with an abundance that decreases with radius beyond the co snow line . when photodesorption is included , the h@xmath0co column is flat across the disk , corresponding to a power - law index of 0 .
neither predicted abundance pattern is consistent with the new observations .
in addition , h@xmath0co is common towards protostellar sources of a range of luminosities ( e.g. * ? ? ? * ; * ? ? ? * ) , while it is pre - dominantly detected towards disks around the low luminosity t tauri stars . in short , h@xmath0co formation through _ in situ _ co - ice hydrogenation is not only consistent with the observations , but it is the _ only _ pathway proposed ( so far ) that naturally explains the observations . to conclusively demonstrate a co - ice hydrogenation origin would , however , require the detection of co - spatial emission of ch@xmath4oh ; ch@xmath4oh has no known efficient gas - phase formation pathway and is predicted to form together with h@xmath0co whenever co ice is hydrogenated @xcite .
the `` snow line '' is typically used to denote the midplane disk radius at which the temperature is low enough for water to condense out on dust grains .
outside of the snow line , grain accretion will be faster because of larger and stickier grains , which may substantially speed up the formation of planetesimals and eventually planets ( e.g. * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
. in the solar system , the dividing line between rocky planets and gas giants coincide with the h@xmath0o snow line @xcite .
co is another abundant volatile in disks and its snow line may boost planet formation in the outer disk by providing extra solid masses @xcite and inducing planet traps @xcite , and could affect the elemental make - up of the forming gas - giants @xcite . because of its high volatility ( @xmath62 k ) , the co snow line is expected at disk radii of 10s100s of au . this should make it a far more accessible target than the h@xmath0o snow line for millimeter interferometry studies aimed at examining the general effects of snow lines on disk structures .
localizing the co snow line directly from millimeter co data is challenging , however .
disks have both a radial temperature gradient away from the central star , and a vertical one set by radiative heating at the disk surface ( e.g. * ? ? ?
this results in a co condensation front that is located at different radii at different disk heights ( fig .
[ fig : cartoon ] ) , and also that some co is present at all radii in the upper disk layers .
this fact , together with a complex radiative transfer ( most co lines are expected to be optically thick in the disk center and optically thin in the outer parts of the disk ) , means that the location of a co snow line in a disk can not be inferred from simply inspecting a co disk image .
the best constraint that exists to date on a co snow line radius is toward the herbig ae star hd 163296 based on the analysis of multi - transition , multi - isotope , spatially resolved co line data on a self - consistent physical disk model @xcite .
the temperature structure of the model has been constrained by optically thick multiple co lines and detailed analysis of the optically thinner @xmath63co emission reveals a significant column density reduction at around 165 au that can not be explained by the overall disk column distribution as traced by the dust .
this is interpreted as the result of co freeze - out and the location as the co snow line .
uncertainties in the temperature structure will mainly affect the determination of the co freeze - out temperature at the location of the co snow line , rather than the location itself , which is constrained between 135 and 175 au in the disk of hd 163296 . while fruitful , this is a time consuming approach that will be difficult to apply to larger samples of disks and will always include some degree of model dependency .
another approach to constrain the co snow line location is to identify trace species that are only present where co has begun to freeze out .
such molecules should display a ring - like structure with the inner edge corresponding to the midplane co snow line .
as we have described , n@xmath0h@xmath1 and h@xmath0co are good candidate probes of co snow lines . in principle , these species can be used as powerful chemical imaging tools to constrain co snow line locations in large samples of disks , rather than relying on complex analysis of the co isotopologue observations .
it seems clear that both n@xmath0h@xmath1 and h@xmath0co are outer - disk species from the chemical perspective . to connect their formation to the onset of co freeze - out
conclusively requires a combination of higher sensitivity , and higher resolution imaging .
this kind of imaging can be done with the newly available capabilities of the alma telescope in chile , which is nearing completion of construction . to demonstrate the astrochemical predictions generated by our analysis of the sma data , and the ease at which they can be tested with alma
, we present a set of simulations of hd 163296 alma observations using the antenna configuration 5 ( corresponding to 0.3 resolution ) in figure [ fig : sim ] .
the predicted h@xmath0co and n@xmath0h@xmath1 rings are readily detected at this resolution , and the power - law and ring models are clearly distinguished .
for both models , the diameter of the emission maximum can be estimated directly from the images within a fraction of the beam size , enabling us to determine if the h@xmath0co and n@xmath0h@xmath1 emission truly trace the co snow line .
in addition , alma should have sufficient sensitivity to detect ch@xmath4oh if it is present with a similar abundance and distribution as h@xmath0co , as expected if both of these species are formed through co hydrogenation and then non - thermally desorbed .
we have presented three observational results that support the idea that co freeze - out regulates the h@xmath0co and n@xmath0h@xmath1 chemistry in disks : 1 .
h@xmath0co and n@xmath0h@xmath1 emission towards hd 163296 appears offset from the continuum peak at a size scale consistent with the co snow line at 160 au .
these observations are matched well by a ring model where h@xmath0co is present only in the disk regions with co freeze - out .
the h@xmath0co excitation temperature in a sample of 9 disks is typically below @xmath64 k , consistent with the bulk of h@xmath0co emission originating in disk regions where co is expected to freeze - out .
the h@xmath0co and n@xmath0h@xmath1 emission are correlated across the disk sample , consistent with the hypothesis of coexistence beyond the co snow line .
these results suggest that both n@xmath0h@xmath1 and h@xmath0co should be present in rings , with the inner edge at the co snow line .
this may be used as a probe of co snow line locations across samples of disk and is also important for predicting the organic content of comets forming at different disk radii . in general ,
the radial and vertical distributions of molecules constitute strong probes of the basic chemistry used into astrochemical models , while molecular abundances and column densities are probably best used to test our understanding of the structure and history of individual objects .
the sma is a joint project between the smithsonian astrophysical observatory and the academia sinica institute of astronomy and astrophysics and is funded by the smithsonian institution and the academia sinica .
we thank edwin bergin and paola dalessio for their helpful suggestions , and a referee for constructive comments on the paper .
support for k. i. o. is provided by nasa through a hubble fellowship grant awarded by the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, for nasa , under contract nas 5 - 26555 .
we also acknowledge nasa origins of solar systems grant no .
nnx11ak63 .
l c parameter & hd 163296 + + @xmath65 ( @xmath19 ) & 1.7@xmath5010@xmath66 + @xmath67 & 2 + @xmath68 ( au ) & 410 + @xmath69 ( au ) & @xmath29200 + _ surface boundary @xmath22 _ & _ 10@xmath70=0.79 _ + midplane boundary @xmath23 & 10@xmath71=31.6 + @xmath26 & 720101 + + fractional abundance & 5.5@xmath72 + _ @xmath73 ( au ) _ & _ 500 _ + _ surface boundary _ & _ t@xmath2919 k _ + midplane boundary @xmath23 & 10@xmath71=31.6 + @xmath26 & 720101 + tw hya@xmath74 & 11 01 51.91 & @xmath7534 42 17.0 & k7 & 56 & 0.005 & @xmath11 , @xmath12 & @xmath31.3@xmath76 & 11(6 ) + hd 163296@xmath77 & 17 56 21.29 & @xmath7521 57 21.9 & a1 & 122 & 0.089 & @xmath10 , @xmath11 & @xmath30.46@xmath76 & 29(10 ) + dm tau@xmath78 & 04 33 48.73 & @xmath7918 10 10.0 & m1 & 140 & 0.02 & @xmath39 , @xmath11 & & 11(4 ) + aa tau@xmath78 & 04:34:55.42 & @xmath7924:28:53.2 & k7 & 140 & 0.01&@xmath39 , ( @xmath11)@xmath80 & & @xmath2918 + lkca 15@xmath78 & 04 39 17.78 & @xmath7922
21 03.5 & k5 & 140 & 0.05 & @xmath39 , @xmath11 & & 16(3 ) + gm aur@xmath78 & 04 55 10.98 & @xmath7930 21 59.4 & k3 & 140 & 0.03 & @xmath39 , @xmath11 & & 12(5 ) + i m lup @xmath81 & 15 56 09.23 & @xmath7537 56 05.9 & m0 & 150 & 0.08 & @xmath39 , @xmath11 & & 22(6 ) + as 209@xmath82 & 16 49 15.29 & @xmath7514 22 08.6 & k5 & 125 & 0.028 & @xmath39 , ( @xmath11)@xmath80 & & @xmath2930 + v4046 sgr@xmath83&18 14 10.47 & @xmath7532 47 34.5 & k5 & 73 & 0.01 & @xmath10 , @xmath11 & @xmath30.92@xmath76 & 12(3 ) + hd 142527@xmath84&15 56 41.89 & @xmath7542 19 23.3 & f6 & 145 & 0.1 & @xmath39 , @xmath11 & & 250(120 ) + + star and disk data from : @xmath74@xcite , @xmath77@xcite , @xmath78@xcite , @xmath85@xcite , @xmath86@xcite , @xmath82@xcite , @xmath83@xcite , @xmath84@xcite . + @xmath87from this work and @xcite .
upper limits in parentheses .
@xmath76calculated from other h@xmath0co transition strengths and the determined excitation temperature .
@xmath80the @xmath88 lines were reported as upper limits in @xcite , but after re - examination of the emission maps , we conclude that these lines are better described as tentative detections ( 2 - 3@xmath28 ) . | we present submillimeter array observations of h@xmath0co and n@xmath0h@xmath1 emission in the disks around the t tauri star tw hya and the herbig ae star hd 163296 at 2@xmath26@xmath2 resolution and discuss the distribution of these species with respect to co freeze - out .
the h@xmath0co and n@xmath0h@xmath1 emission toward hd 163296 does not peak at the continuum emission center that marks the stellar position but is instead significantly offset . using a previously developed model for the physical structure of this disk ,
we show that the h@xmath0co observations are reproduced if h@xmath0co is present predominantly in the cold outer disk regions . a model where h@xmath0co is present only beyond the co snow line ( estimated at a radius of 160 au ) matches the observations well . we also show that the average h@xmath0co excitation temperature , calculated from two transitions of h@xmath0co observed in these two disks and a larger sample of disks around t tauri stars in the discs ( the disk imaging survey of chemistry with sma ) program , is consistent with the co freeze - out temperature of @xmath320 k. in addition ,
we show that n@xmath0h@xmath1 and h@xmath0co line fluxes in disks are strongly correlated , indicative of co - formation of these species across the sample . taken together ,
these results imply that h@xmath0co and n@xmath0h@xmath1 are generally present in disks only at low temperatures where co depletes onto grains , consistent with fast destruction of n@xmath0h@xmath1 by gas - phase co , and _ in situ _ formation of h@xmath0co through hydrogenation of co ice . in this scenario h@xmath0co ,
ch@xmath4oh and n@xmath0h@xmath1 emission in disks should appear as rings with the inner edge at the co midplane snow line .
this prediction can be tested directly using observations from alma with higher resolution and better sensitivity . |
the large - scale integral - sign warps of the outer regions of spiral galaxies has been a long - standing theoretical puzzle .
the strongest warps are observed in the hi disks when these extend well beyond the optical disks ( briggs 1990 ) .
the warps are antisymmetric and consequently can be well - fitted by the kinematic tilted - ring model of rogstad , lockhart , and wright ( 1974 ) . for a set of 12 galaxies with high - quality data on the extended hi disks , briggs ( 1990 ) found empirically a set of ` rules of behavior ' of warps : 1 .
the warps develop with increasing radius @xmath8 for @xmath8 between @xmath9 and @xmath10 ( the holmberg radius , @xmath11 ) , where the subscript denotes the surface brightness in @xmath12 in mag / arcsec@xmath13 ; 2 .
the line - of - nodes of the warp is straight for @xmath14 ; and 3 . for @xmath15 ,
the line - of - nodes forms an open leading spiral .
an illustrative case is shown in figure 1 which gives a surface plot of the warp of the disk of the spiral galaxy m 83 obtained from the data of briggs ( 1990 ) .
the rotation of m 83 is clockwise so that the warp is a _ leading _ spiral wave .
figure 2 also shows the data of briggs ( 1990 ) for m 83 , in panel ( * a * ) the radial ( @xmath8 ) dependence of the tilt angle @xmath16 , in ( * b * ) the @xmath8-dependence of the angle of the line - of - nodes @xmath17 , and in ( * c * ) the dependence of @xmath16 on @xmath17 , which we refer to as a briggs plot .
a different behavior is observed for the galaxy ngc 3718 which shows an approximately straight line - of - nodes out to @xmath18 and a large tilt angle ( @xmath19 ) at this radius ( briggs 1990 ) .
a partial listing of the theoretical works on warps of disk galaxies include kahn and woltjer ( 1959 ) , hunter and toomre ( 1969 ; hereafter ht ) , toomre ( 1983 ) , sparke and casertano ( 1988 ) , binney 1992 , hunter ( 1994 ) , merritt and sellwood ( 1994 ) , and nelson and tremaine ( 1995 ) . since the pioneering analysis by hunter and toomre ( 1969 ) and the earlier related work by lynden - bell ( 1965 ) ,
the theoretical studies have focused mainly on determining the eigenmodes and eigenfrequencies of warped self - gravitating disks .
the studies used the ht dynamical equation for the vertical displacement of the disk surface @xmath20 , where @xmath21 is the azimuthal angle .
in contrast with these earlier works , the present study investigates different initial value problems of a warped disk in an oblate ( or prolate ) halo potential , and it develops a different representation for the disk warping .
instead of describing the disk in terms of @xmath20 , we develop a representation in terms of @xmath0 independently tilted , self - gravitating , concentric rings which is suggested by the kinematic ring model of rogstad et al .
may and james ( 1984 ) mention a model of this kind , but they give no details . we show that the @xmath1ring representation is equivalent to the ht equation for @xmath20 for @xmath22 .
however , new insight is given by the @xmath1ring model : a new constant of the motion for a warped disk is found .
the notion of the phase - locking of the line - of - nodes of nearby rings due to self - gravity is demonstrated .
we propose that this explains briggs s rule no .
furthermore , treatment of initial value problems for the age of a galaxy can be done with sufficient accuracy with relatively small values of @xmath0 ( @xmath23 ) . in 2
we first give basic parameters of the disk and the halo . in
2.1 we treat the case of one tilted ring and in 2.2 comment on the interpretation of observations of a ring .
subsection 2.3 considers the case of two tilted rings .
subsection 2.4 treats the case of two tilted , counter - rotating rings .
subsection 2.5 treats the case of three corotating rings .
section 3 treats the general case of @xmath0 tilted corotating rings . in this section
we consider different possible excitations or origins of warps - that due to a passing satellite ( 3.1 ) , a sinking satellite ( 3.2 ) , a tilted halo potential ( 3.3 ) , and an initially tilted outer disk plane ( 3.4 ) .
section 4 discusses the continuum limit where @xmath24 .
section 5 gives conclusions of this work .
the equilibrium , unwarped galaxy is assumed to be axisymmetric and to consist of a thin disk of stars and gas and a slightly oblate or prolate halo of dark matter .
we use an inertial cylindrical @xmath25 and cartesian @xmath26 coordinate systems with the disk and halo equatorial planes in the @xmath27 plane .
the total gravitational potential is written as @xmath28 where @xmath29 is the potential due to the inner part of the disk ( as discussed below ) , @xmath30 is that for the outer disk , and @xmath31 is that for the halo .
the particle orbits in the equilibrium disk are approximately circular with angular rotation rate @xmath32 , where @xmath33 the vertical epicyclic frequency ( squared ) , @xmath34 gives a measure of the restoring force in the @xmath35direc - tion .
the surface mass density of the inner ( optical ) disk is assumed to be @xmath36 with @xmath37 and @xmath38 constants and @xmath39 the total disk mass . for @xmath40 , @xmath41 @xmath42 ( binney and tremaine 1987 ; hereafter bt , p. 409 ) . note that @xmath43 .
the outer , neutral hydrogen disk can be described approximately by a fermi function for the surface density @xmath44 } \eqno(5)\ ] ] where @xmath45 is the radius inside of which half the neutral hydrogen mass is located , @xmath46 , and @xmath47 .
this dependence provides a good fit to the data of broeils and rhee ( 1966 ; their figure 6a ) .
the total neutral hydrogen mass is @xmath48 .
the halo potential is taken to be @xmath49~ , \eqno(6a)\ ] ] where @xmath50 const .
is the circular velocity at large distances , @xmath51 const .
is the core radius of the halo , and @xmath52 is the axial ratio of the equipotential surfaces with @xmath53 their ellipticity .
an oblate ( prolate ) halo corresponds to @xmath54 ( @xmath55 ) .
we have @xmath56 the halo is usually assumed to be oblate with @xmath57 ( nelson and tremaine 1995 ) .
this gives @xmath58 which is of interest in the following .
for the radii of the warps observed in galaxies ( @xmath59 ) , @xmath60 . for @xmath61 , the halo mass within a radius @xmath8 is @xmath62.$ ] perturbations of the galaxy
are assumed to consist of small angle tilting ( @xmath63 ) of the outer disk with azimuthal mode number @xmath64 ( or @xmath65 ) .
the halo and the inner disk are assumed to be unaffected by the perturbation .
we can describe the outer disk by a number @xmath0 of tilted plane circular rings .
this description is _ general _ for small tilting angles where the linearized equations are applicable .
we do not need to assume that the disk is ` razor thin ' but rather that the disk half - thickness @xmath66 is small , @xmath67 .
the vertical displacements of the disk @xmath68 are in general much larger than @xmath66 .
the tilting of the ring is completely described by the tilt angle @xmath69 , which is the angle between the upward normal to the ring and the @xmath35axis , and the azimuthal angle @xmath70 , which is the angle between the line - of - nodes ( where the ring intersects the @xmath27 plane ) and the @xmath71axis .
the geometry is shown in figure 3 .
the perturbation is assumed to be small in the respect that @xmath72 .
the position vector of a point on the ring is @xmath73~ , ~~0\leq \phi \leq 2\pi~ , \eqno(7)\ ] ] ( cartesian components ) , where @xmath74 is the height of the ring above the @xmath27 plane .
we have @xmath75~,\ ] ] @xmath76 where @xmath77 \eqno(8b)\ ] ] is the angle of the tilt about the @xmath71axis , and @xmath78 \eqno(8c)\ ] ] is the angle about the @xmath79axis .
we have @xmath80 and @xmath81 . the angular momentum of the tilted ring is @xmath82 @xmath83 where @xmath84
, @xmath85 is the ring mass , and @xmath86 is its angular rotation rate which is unaffected by the perturbation to first order in @xmath16 .
we have @xmath87 @xmath88 here , @xmath89 @xmath90 $ ] is the unit tangent vector to the ring ( to first order in @xmath16 ) , @xmath91 is the distance along the ring circumference measured from , say , the ascending node where @xmath92 , and @xmath93~. \eqno(10b)\ ] ] from equation ( 9 ) we have @xmath94 @xmath95 the torque on the ring is @xmath96 or @xmath97 @xmath98 the ring is assumed to exert no torque on itself .
consequently , the force @xmath99 ( per unit circumference ) which contributes to @xmath100 in equation ( 12a ) is due to the inner disk and the halo .
thus we have @xmath101 where @xmath102 is the vertical epicyclic frequency excluding the contribution due to the outer disk [ see equation ( 3 ) ] . also , @xmath103 where @xmath104 which excludes the outer disk contribution to @xmath105 [ see equation ( 2 ) ] .
thus we find @xmath106 @xmath107 the equations of motion @xmath108 are @xmath109 @xmath110 after multiplying by two and letting @xmath111 denote the moment of inertia of the ring .
the terms of the left - hand side @xmath112 are due to the coriolis force .
it is useful to introduce @xmath113~. \eqno(16)\ ] ] this representation is well - known from treatments of spin precession in quantum mechanics .
we can then combine equations ( 15a ) and ( 15b ) to obtain @xmath114 where @xmath115 note that we can have @xmath116 for a prolate halo .
solutions of equation ( 17 ) can be taken as @xmath117 with @xmath118 a complex constant . from equation
( 17 ) we have the dispersion relation @xmath119 which gives @xmath120 in all cases considered here , @xmath121 . for @xmath122 , the plus sign corresponds to the _ fast mode _
, @xmath123 , with prograde precession [ @xmath70 increasing ] , and the minus sign to the _ slow mode _ , @xmath124 , with retrograde precession [ @xmath70 decreasing ] ( ht ) . for @xmath125 and @xmath122 , the slow mode has @xmath126 .
these two modes are the analogues of the _ normal modes _ of vibration of a non - rotating mechanical system . that there are two modes rather than one for each ring is due to the doppler splitting from the rotation ( @xmath86 ) . a general solution of equation ( 17 ) is @xmath127 where @xmath124 and @xmath128 are the two roots given in equation ( 19 ) , and @xmath129 and @xmath130 are complex constants
the four real quantities determine the initial values @xmath131 , @xmath132 , @xmath133 , and @xmath134 . if only the fast or slow mode is excited ( @xmath129 or @xmath135 ) , then @xmath16 = const .
and @xmath136 const.@xmath137 with @xmath138 . a simple way to excite only the slow or fast mode is to take @xmath139 , @xmath140 , @xmath141 , and @xmath142 .
figure 4 illustrates the two limiting cases of @xmath143 and @xmath144 .
we can recast equations ( 15 ) in a form equivalent to that given by ht by multiplying equation ( 15a ) by @xmath145 and equation ( 15b ) by @xmath146 and subtracting the two .
noting that @xmath147 and @xmath148 , we obtain @xmath149 note that @xmath20 is non - zero only for @xmath8 equal to the ring radius . from equation ( 2 )
we have @xmath150 for a toroidal ring of minor radius @xmath151 , this difference is of the order of @xmath152 . in the absence of a halo , the right hand side of the equation for @xmath153 is @xmath154 which is the same as the ht expression @xmath155}\over { |{\bf r - r'}|^3}}~.\ ] ] the @xmath156 term comes from the integration of over @xmath157 while the @xmath158 term is from the integration over @xmath159 .
multiplying equation ( 15a ) by @xmath160 and ( 15b ) by @xmath161 and adding the two gives @xmath162 , where the energy of the tilted ring is @xmath163 where a dot denotes a time derivative .
this expression agrees with the result of ht ( their appendix b ) . in the absence of a halo or with an oblate halo ( @xmath164 ) , we have @xmath165 , so that both terms in @xmath166 are non - negative which implies stability of the ring tilting .
in the case of a prolate halo with @xmath116 instability is possible . in order to understand the stability of the ring tilting ,
it is useful to consider the influence of a small newtonian drag or friction on the ring motion , a force @xmath167 per unit circumference , with @xmath168 but @xmath169 .
this drag could result from dynamical friction of the ring with the dark halo matter ( nelson and tremaine 1995 ) .
the full calculation of @xmath170 due to dynamical friction on a precessing ring is complicated ( nelson and tremaine 1995 ) , but an estimate based on treating the ring as two point masses @xmath171 at the points @xmath172 ( weinberg 1985 ) gives @xmath173 @xmath174 where @xmath175 is the coulomb logarithm , @xmath176 is the circular velocity , and @xmath177 accounts for the density profile of the halo , with @xmath178 the halo core radius .
later , in 3 , where we represent the outer disk as @xmath0 tilted interacting rings , the pertinent estimate for @xmath179 for a single ring is this formula with @xmath85 the total mass of the outer disk rather than the mass of a single ring ( weinberg 1985 ) . the drag torque due to @xmath180 is @xmath181 @xmath182 and @xmath183 .
solution of equation ( 17 ) including @xmath184 gives @xmath185 thus , both fast and slow precession modes with @xmath186 are stable in the presence of friction , whereas the slow precession mode with @xmath187 is unstable @xmath188 $ ] .
for the slow precession mode , the kinetic energy term @xmath189 in @xmath166 is smaller than the magnitude of the potential energy term @xmath190 for @xmath191 . in the presence of the friction force , @xmath192 .
thus , for the slow precession mode , @xmath193 is negative if @xmath194 , and the tilt angle grows .
the instability is therefore a _ negative energy dissipative _ instability .
an analogous instability was predicted and observed for the precession of laboratory collisionless relativistic electron rings ( furth 1965 ; beal et al .
the instability also occurs in the presence of a non - newtonian ( nonlinear ) friction force on the ring .
a second quadratic integral of the motion of equations ( 15 ) , analogous to angular momemtum ( see following ) , can be obtained by multiplying equation ( 15a ) by @xmath195 and equation ( 15b ) by @xmath196 and adding .
this gives @xmath197 , where @xmath198 @xmath199 an analogous constant of the motion exists for cases of two or more rings ( see 2.3 , 3 , and 4 ) . this constant of the motion is a new result of the present work . for the case of instability , with @xmath200 complex , @xmath201 . with
the above - mentioned friction force included , we have @xmath202 , which is compatible with equation ( 23 ) which shows instabililty for @xmath116 and gives @xmath203 and @xmath204 . the second integral ( 24 )
can be understood by noting that equations ( 15 ) follow from the lagrangian @xmath205 @xmath206 a canonical transformation to the variables @xmath16 and @xmath17 gives @xmath207 thus the canonical momentum @xmath208
@xmath209 @xmath210 = const .
@xmath211 because @xmath212 . on the other hand , @xmath213 depends on time .
the hamiltonian @xmath214 is @xmath215 and @xmath216 = const . in that @xmath217
has no explicit time dependence .
the @xmath218dependent terms of @xmath219 can evidently be viewed as an effective potential for the @xmath16 motion . in general ( for @xmath220 ) , @xmath16 nutates between a minimum and a maximum value .
the angular frequency of the nutation is @xmath221 so that the _ nutation period _ is @xmath222 , where @xmath223 is the period of the orbit .
this can also be seen by noting that with @xmath224 given in general by equation ( 20 ) , we have @xmath225|$ ] and @xmath226 . for @xmath227 , the period of the slow mode @xmath228 , while the fast mode period @xmath229 is slightly longer than the nutation period @xmath230 .
the nutation of the ring is clearly evident in figure 4 . for @xmath231 ( figure 4a )
, @xmath224 nutates many ( @xmath232 ) times in the period @xmath233 of motion of @xmath234 about the origin . on the other hand , for @xmath235 ( figure 4b )
, @xmath224 has an elliptical path with the azimuth of say the maximum of @xmath236 precessing slowly in the clockwise direction with a period @xmath237 .
a counter - rotating ring with @xmath238 behaves in the same way as a co - rotating ring . for @xmath186 ,
the ring slow mode precession is retrograde relative to the ring particle motion but of course prograde relative to our coordinate system .
warps of spiral galaxies are deduced from measurements of @xmath239 cm line neutral hydrogen emission . here , we consider the spectral signature of a single ring in either the slow precession mode or the fast precession mode .
the observer is considered to be in the @xmath240 plane at an angle @xmath241 to the @xmath35axis .
the velocity of the ring matter is given by equation ( 10 ) .
thus , the velocity of the hi in the direction of the observer is @xmath242 @xmath243 the spectrum of the ring is @xmath244 @xmath245 .
the ring motion @xmath246 enters through the term involving @xmath247 .
for a ring in the slow precession mode with @xmath248 , the ring motion has only a small affect on @xmath249 .
the situation is very different for a ring in the fast precession mode where @xmath250 if @xmath165 . in this case
the sign of the term in @xmath249 proportional to @xmath16 is reversed which corresponds to a change of the line - of - nodes angle @xmath17 by @xmath251 .
observations are interpreted assuming negligible ring motion , @xmath252 , and this is wrong if the fast precession mode(s ) is excited
. the fast precession modes of a system with many rings may disappear from view over a long enough time due to phase - mixing ( see 3.1 and 3.2 ) .
consider now the case of two tilted plane circular rings of mass @xmath253 and angular rotation rate @xmath254 at radii @xmath255 with @xmath256 .
we have @xmath257 @xmath258 where @xmath259 and @xmath260 are the tilt angles and @xmath261 the moments of inertia for the two rings . to obtain the torque on , say , ring 1 , we exclude as before the force due to this ring on itself .
thus the horizontal force on ring @xmath262 is @xmath263 where the @xmath264 is the gravitational potential due to ring 2 , and the @xmath262-subscript on the parenthesis indicates evaluation at @xmath265 . note that @xmath266 @xmath267 where @xmath268 @xmath269 there is also the vertical force on ring @xmath262 , @xmath270 where @xmath271 \over \left[r_1 ^ 2+r_2 ^ 2 - 2r_1r_2{\rm cos}(\phi-\phi')\right]^{3\over2 } } \eqno(31b)\ ] ] is the vertical force on ring 1 due to ring 2 .
this integral can be simplified by noting that @xmath272 @xmath273 the sin@xmath274 term in equation ( 31c ) does not contribute to the integral ( 31b ) .
thus we have @xmath275 @xmath276 @xmath277 combining terms in equations ( 12b ) and ( 12c ) gives @xmath278 @xmath279 where @xmath280 measures the strength of the gravitational interaction between the two rings , and @xmath281 is defined in equation ( 18 ) . the two terms involving the integral @xmath282 cancel .
note that @xmath283 is symmetric between the two rings and that it is non - negative because @xmath284 [ see equation ( 45e ) and figure 5 ] .
the torque on ring 2 is found in the same way to be @xmath285 @xmath286 the equations of motion are @xmath287 for the limit where the only torques are those due to the two rings , @xmath288 and @xmath289 , we find @xmath290 which is a necessary result .
we can follow the steps leading to equation ( 22 ) to obtain @xmath291 for the energy @xmath292 of the two tilted rings , where @xmath293 @xmath294 where @xmath295 and @xmath296 are the moments of inertia of the two rings .
this energy is non - negative and the rings are stable if both @xmath297 and @xmath298 are non - negative .
also , in this case the rings are stable in the presence of dissipative forces such as the friction force of subsection 2.1 .
thus a necessary condition for instability with or without dissipation is that @xmath297 and/or @xmath298 be negative .
a second quadratic integral of the motion can be obtained by following the steps leading to equation ( 24 ) .
we find @xmath299 , where @xmath300 for the case of instability , @xmath301 , @xmath302 , we have @xmath303 .
thus , a further necessary condition for instability in the absence of dissipation ( in addition to @xmath297 and/or @xmath298 being negative ) is that @xmath304 and @xmath305 have different signs ; that is , @xmath306 for @xmath307 . for the slow precession modes with @xmath308
, this corresponds to counter - rotating rings , @xmath309 .
the lagrangian for the two ring system is @xmath310 @xmath311~. \eqno(40)\ ] ] the corresponding hamiltonian @xmath312 is equal in value to @xmath292 of equation ( 38 ) .
the second constant of the motion , @xmath313 , where @xmath314 . note that @xmath315 and that @xmath316 so that @xmath317 .
as in the case of a single ring , it is useful to introduce @xmath318~ , \eqno(41a)\ ] ] @xmath319~. \eqno(41b)\ ] ] the equations of motion ( 36 ) can then be written as @xmath320 @xmath321 after dividing through by @xmath322 . for @xmath323 and @xmath324
, we get the dispersion relation @xmath325 \left[\left(\omega-\omega_2\right)^2-k_2 ^ 2\right]={c_{12}^2 \over i_1i_2}~ , \eqno(43)\ ] ] where @xmath326 with @xmath256 .
equation ( 43 ) can readily be solved for the four roots .
two of these roots are high - frequency , _ fast precession modes _ with @xmath327 , assuming @xmath328 and @xmath329 .
the other two roots are the _ slow precession modes _ with much smaller frequencies , typically , @xmath330 . in order to get simple analytic results ,
we first consider the low - frequency modes @xmath330 , assuming @xmath331 and @xmath332 .
we can then neglect the second time derivatives in equations ( 42 ) and the @xmath333 terms in equation ( 43 ) , and this leads to a quadratic dispersion equation for @xmath200 .
the roots of this equation are @xmath334\ ] ] @xmath335 ^ 2 + 4\omega_{g12}^2\ell_{12}\right\}^{1\over2}~. \eqno(44)\ ] ] here , @xmath336 for @xmath256 , are the slow precession frequencies of the two rings in the absence of gravitational interaction between them ( @xmath337 ) ; @xmath338 where @xmath339 is a frequency which measures the strength of the gravitational interaction between the two rings , and where @xmath340 are the angular momenta of the two rings .
we have @xmath341 @xmath342 where @xmath343^{3\over2}}~~ , \eqno(45e)\ ] ] with @xmath344 for @xmath345 , @xmath346 . for @xmath347 ,
@xmath348 $ ] .
we assume that @xmath349 , where @xmath66 is the disk thickness .
the dependence of @xmath350 on @xmath351 is shown in figure 5 .
a simple limit of equation ( 44 ) is that where @xmath352 .
the two roots are then @xmath353 and @xmath354 .
the zero frequency mode corresponds to both rings tilted by the same angle , @xmath355 , which is the rigid tilt mode ( ht ) .
the other mode has @xmath356 ; that is , the rings are tilted in opposite directions for @xmath357 .
we first consider co - rotating rings where the four @xmath200 roots of equation ( 43 ) are real in the absence of dissipation .
note that the low frequency roots of equation ( 44 ) are real for @xmath358 .
a general solution for the motion of the two rings is then @xmath359 where @xmath360 are the four mentioned frequencies , and @xmath361 are @xmath362 complex constants .
however , only @xmath362 real quantities are needed to specify general initial conditions because the @xmath363 are related to the @xmath364 by equations ( 42 ) . if only a _
single _ mode of the system is excited , then @xmath365 and @xmath366 .
equations ( 42 ) then imply that @xmath367 is real .
because @xmath368 , we conclude that @xmath369 $ ] is real so that @xmath370 is either @xmath371 or @xmath372 .
that is , the rings are either tilted in the same direction and precess together or they are tilted in opposite directions and also precess together . from equations ( 42 ) , @xmath373 @xmath374/c_{12}$ ] .
figure 6a shows the @xmath339 dependence of the frequencies of the two slow - precession modes obtained from equations ( 43 ) and ( 44 ) for the case where both rings are co - rotating and the halo is _
oblate_. the ` slower ' of the two modes has @xmath375 , while the other has @xmath376 as shown in figure 6b .
the larger gravitational torque in the second case accounts for the faster retrograde precession . as @xmath339 increases
, the frequency of the slower mode approaches @xmath377 and @xmath378 for @xmath379 are defined in equation ( 45a ) .
this result can be derived from equation ( 43 ) or ( 44 ) . for the case of @xmath0 rings and sufficiently large @xmath380
, we find in general that the lowest frequency approaches the angular momentum weighted average of the single ring precession frequencies @xmath381 [ equation ( 45a ) ] .
in contrast with the dependence of @xmath124 , @xmath128 decreases monotonically as @xmath339 increases . from the approximate equation ( 44 ) for the slow precession modes
, we can distinguish _ weak _ and _ strong_-_coupling _ limits of two co - rotating rings .
the coupling strength is measured by the dimensionless parameter @xmath382 @xmath383 which is symmetrical in the ring indices ( unlike @xmath339 ) . for _ weak coupling _
, @xmath384 , the two rings are affected little by their gravitational interaction , and the two slow mode frequencies are @xmath385 with @xmath386 and @xmath387 with @xmath388 . in the opposite limit of _ strong coupling _ , @xmath389 , the gravitational interaction of the two rings is important , and @xmath124 is given approximately by equation ( 47 ) with @xmath390 , and @xmath391 with @xmath392 . for nearby rings , @xmath393 , of equal mass @xmath394 ( with @xmath395 the surface mass density of the disk at @xmath8 ) , we have @xmath396 $ ] , @xmath397 , and thus @xmath398 evidently , there is strong ( weak ) coupling for @xmath399 ( @xmath400 ) , where @xmath401 corresponds to @xmath402 so that @xmath403^{1\over2}~. \eqno(50a)\ ] ] assuming a flat rotation curve ( @xmath404 const . ) and @xmath405 const .
, we obtain @xmath406 @xmath407 rings closer together than @xmath401 will be strongly coupled .
figure 7 shows sample orbits of @xmath323 and @xmath408 for two co - rotating rings in an oblate halo for initial conditions @xmath409 and @xmath410 . the top panel shows a case of weak coupling ( small @xmath411 ) where the line - of - nodes @xmath4 of the two rings regress essentially independently with the result that @xmath370 increases linearly with time .
the bottom panel is for a case of strong coupling ( large @xmath411 ) where the line - of - nodes are _ phase - locked _ in the sense that the time average of @xmath370 is zero .
figure 8 shows the azimuthal angle difference of the line - of - nodes for the two cases of figure 7 . in the case of weak coupling , a low frequency initial
perturbation excites both low frequency modes which have similar frequencies ( @xmath412 ) , and this gives the unlocked behavior . for strong coupling
, the initial perturbation excites mainly the @xmath124 mode because @xmath391 , and this gives the locked behavior . for the parameters of figures 4 and 5 , note that @xmath413 .
figure 9a shows the @xmath339 dependence of the frequencies of the two slow - precession modes obtained from equation ( 43 ) for the case where both rings are co - rotating and the halo is _
prolate_. ( we have adopted the convention of ordering the mode frequencies by their magnitudes for @xmath414 . ) figure 9b shows that the @xmath128 mode has @xmath375 . as in the case of an oblate halo , for large @xmath339 or @xmath411 ,
@xmath128 approaches the limit given by equation ( 47 ) and @xmath378 . as in the case of a single ring in a prolate halo potential
[ see equation ( 23 ) ] , small friction torques , with friction coefficients @xmath415 and @xmath416 , on two co - rotating rings in a prolate halo lead to a negative energy dissipative instability .
we find that the slow modes with prograde precession , @xmath417 , are unstable in the presence of dissipation .
that is , the @xmath124 mode in figure 9 is unstable ( stable ) for @xmath418 @xmath419 @xmath420 , whereas the @xmath128 mode is unstable for all @xmath339 .
the two high frequency modes are prograde but are damped by friction for all @xmath339 .
as in the case of an oblate halo ( figure 8) , we observe phase - locking for sufficiently large values of @xmath411 .
frictional torques between two co - rotating rings in a prolate halo also give rise to a negative energy dissipative instability .
such torques could arise from the viscous interaction between adjacent gaseous rings .
these frictional torques do not change the total angular momentum of the two rings .
therefore , these frictional torques , if newtonian ( linear ) in nature , are _ necessarily _ of the form that they contribute terms @xmath421 @xmath422 on the right hand sides of equations ( 42a ) and ( 42b ) , respectively , where @xmath423 is the friction coefficient ( with units of angular frequency ) . for a gas with kinematic shear viscosity @xmath424 ( with units of @xmath425 ) the terms ( 51 ) arise from the momentum flux density or stress @xmath426 due to the different vertical velocities of the two rings as a function of @xmath21 .
this gives @xmath427 , where @xmath428 .
the relevant viscosity is that due to turbulence in the gas @xmath429 owing to the smallness of the microscopic viscosity .
an estimate of @xmath430 can be made following the proposal of shakura ( 1973 ) and shakura and sunyaev ( 1973 ) that @xmath431 , where @xmath432 is the sound speed , @xmath66 is the half - thickness of the disk , and @xmath433 is a dimensionless constant thought to be in the range @xmath434 to @xmath262 . in sections 3 and 4
we discuss further the frictional torques between adjacent rings . here , we consider the case of two tilted rings which are rotating in opposite directions .
this situation is pertinent to observed counter - rotating galaxies ( see , for example , jore , broeils , and haynes 1996 ) . from equation ( 44 ) , instability is possible ( in the absence of dissipation ) only for @xmath435 in agreement with our discussion of equation ( 39 ) .
thus , consider ring @xmath436 to be counter - rotating ( @xmath437 and @xmath438 ) , while ring @xmath262 is co - rotating .
the region of instability is bounded by the two curves @xmath439 @xmath440 on which the square root in equation ( 44 ) is zero .
the curves are shown in figure 10 . for parameters in the region between the two curves
there is instability , which is seen to occur only for @xmath297 and/or @xmath298 negative . for fixed values of @xmath441 , @xmath442 , and @xmath443
, instability occurs in general ( @xmath444 ) for @xmath339 larger than a critical value but smaller than a second larger critical value .
the maximum growth rate is for @xmath445 ( so that @xmath446 ) , and it is @xmath447 .
this is a _ dynamical instability _ in that it does not depend on dissipation as in the case of a single ring with @xmath116 . for conditions of maximum growth , @xmath448 , so that @xmath449 and @xmath450 .
we can thus view the co - rotating ring as ` torquing up ' the more slowly precessing counter - rotating ring by the gravitational interaction or vica - versa .
figure 11a shows the @xmath339 dependence of the frequencies of the two slow - precession modes obtained from equation ( 43 ) for the case where ring @xmath451 is counter - rotating ( @xmath452 ) and ring 1 is co - rotating ( @xmath453 ) and the halo is _
oblate_. figure 11b shows that the mode @xmath124 with prograde precession involves mainly motion of ring 2 , whereas the @xmath128 mode with retrograde precession involves mainly ring 1 . for the conditions studied ( @xmath454 ) ,
we find no phase - locking .
extension of the results of subsection 2.2 to the case of three tilted rings is straightforward .
figure 12a shows the dependence of the three slow precession frequencies on the strength of the gravitational interaction measured by @xmath339 with @xmath455 const . for increasing @xmath339 , the slowest mode ( @xmath124 ) _ slowly _
approaches the limiting frequency given by the generalization of equation ( 47 ) to three rings .
the other two frequencies decrease monotonically with increasing @xmath339 .
figure 12b shows the geometrical nature of the three slow precession modes .
figure 13 shows sample orbits @xmath456 for three rings for cases of weak and strong coupling .
figure 14 shows different cases of phase - locking for a three ring system .
the generalization of equation ( 40 ) gives the lagrangian for an @xmath0 ring system as @xmath457 @xmath458 .
\eqno(53a)\ ] ] the hamiltonian , @xmath459 is @xmath460\ ] ] @xmath461~. \eqno(53b)\ ] ] because the @xmath462 are non - negative , all of the terms in @xmath219 are non - negative if @xmath329 which is the case for oblate halo .
thus the @xmath1ring system with @xmath463 is stable in the presence of dissipative forces .
the total canonical angular momentum is @xmath464 in the absence of dissipation @xmath465 and @xmath466 are constants of the motion . as in the case of one or two rings ,
it is useful to let @xmath467~. \eqno(54)\ ] ] the equations of motion can then be written as @xmath468 @xmath469 with @xmath6 , where the @xmath470 generalize @xmath283 of equation ( 34 ) . as discussed in 2.1 , a newtonian drag force on the ring motion due to dynamical friction with the halo matter
can be taken into account by including the term @xmath471 on the right hand side of equation ( 55 ) .
the influence of the relative friction discussed in 2.3 can be accounted for by including the terms @xmath472 on the right hand side of equation ( 55 ) . here , @xmath473 and @xmath474_{j+{1 \over 2}}$ ] with @xmath475 and @xmath476 .
with the dissipation terms ( 56 ) and ( 57 ) included one can readily show that @xmath477 ( see also 4 ) .
we have not found an analogous result for @xmath478 in the presence of dissipation . we have developed and tested codes to solve equations ( 55 ) ( including the terms ( 56 ) and ( 57 ) ) for @xmath479 and @xmath480 rings . for the results presented here ,
the rings are taken to be uniformly spaced in @xmath8 with ring one at @xmath481 and ring @xmath0 at @xmath482 so that @xmath483 , with @xmath484 in units of kpc .
the ring masses @xmath485 and moments of inertia @xmath486 are calculated with @xmath487 the sum of the surface density of the inner exponential disk and that of the neutral hydrogen given in 2 .
the @xmath488 are calculated using equations ( 4 ) , ( 6 ) , and ( 18 ) .
the coupling coefficients @xmath489 are evaluated using equations ( 34 ) and ( 30b ) and stored .
equations ( 55 ) are solved as four first order equations for @xmath490 and @xmath491 . at the same time , an additional equation , @xmath492 is also solved to give @xmath493 , which is the line - of - nodes angle relative to the @xmath71axis .
comparisons of the temporal responses of @xmath494 and @xmath480 ring sytems for different initial conditions show generally good agreement of the warps @xmath7 for the time intervals studied .
this indicates that @xmath495 rings gives a valid representation of a continuous disk .
however , in contrast with a continuous disk ( ht ) , an @xmath1ring disk has @xmath496 discrete modes with separate frequencies - @xmath0 ` low - frequency ' modes and @xmath0 ` high - frequency ' modes .
figure 15 shows the power spectrum of @xmath497 for a @xmath480 ring disk - halo system which is initially given a random perturbation with all radial wavelengths excited and no damping .
the frequencies @xmath200 are normalized by the angular velocity of the inner ring @xmath498 .
note that @xmath499 gyr for our reference values .
the frequency differences between adjacent modes is @xmath500 . for treatment of initial value problems of disk warping for time intervals @xmath501 , frequency differences @xmath502 less than @xmath503 are irrelevant .
it is sufficient to have @xmath504 which corresponds to @xmath505 @xmath506 rings .
this suggests that a larger number of rings may be needed for our cases with @xmath507 gyr ( figures 18 and 19 ) .
this will be investigated in a future work .
the lowest frequency mode of the power spectrum , @xmath124 in figure 15 ( see also 2.3 and 2.5 ) , remains discrete ( that is , isolated ) as @xmath24 because our disk model has a sharp outer edge so that @xmath508 is integrable ( ht ; sparke and casertano 1988 ) .
power spectra obtained for cases with @xmath509 show that @xmath124 increases with @xmath510 much more slowly than the higher frequency modes which is consistent with the behavior observed in figures 6 and 12 . for the conditions of figure 15 ,
the magnitude @xmath124 is considerably smaller than that predicted for the limit of strong self - gravity between the rings where @xmath511 [ equation(47 ) ] .
this limiting frequency is the same as the frequency @xmath512 of sparke and casertano ( 1988 ) . in any case
, the frequency @xmath124 is so small ( the corresponding period is @xmath513 gyr ) that it is irrelevant to the initial value problems considered below . in the following four subsections we consider different possible origins of warps in an otherwise flat galaxy .
subsection 3.1 discusses warp excitation by a passing compact satellite ; 3.2 treats warp excitation by a compact sinking satellite ; 3.3 treats the case of a tilted halo potential .
subsection 3.4 considers the tilt evolution for the case where the initial plane of the disk material is tilted .
consider the excitation of a warp in a galaxy due to the passage of a satellite of mass @xmath514 much less than the galaxy s mass .
the satellite s orbit @xmath515 can easily be calculated exactly for @xmath516 , neglecting the back reaction of the perturbed galaxy on the satellite .
however , we assume that even at closest approach the satellite is far from the center of the galaxy and therefore calculate the orbit in the halo potential , equation ( 6a ) , with the ellipticity of the halo neglected for simplicity ( @xmath61 ) . at the closest approach at @xmath517 ,
the satellite s speed is @xmath518 , and the angle between the satellite and the @xmath519 plane is @xmath520 .
also , at closest approach , the satellite is taken to be at @xmath521 and @xmath522 .
we consider both prograde and retrograde satellite passages . in the presence of a satellite of finite mass @xmath514 ,
the description of the galaxy disk in terms of tilted circular rings breaks down .
the centers of the rings are shifted from the origin , and the rings become non - circular .
for this reason we consider the response of the galaxy to a ` symmetrized ' perturbation of a satellite obtained by replacing the actual satellite by two satellites , each of equal mass @xmath523 , with one satellite in the orbit @xmath515 described above and the other in the orbit @xmath524 . with this prescription
we then calculate the torque of the satellites on each ring as a function of time , @xmath525\ ] ] @xmath526 where @xmath527 $ ] , with @xmath528 given by equation ( 8a ) , is the position vector to a point on the @xmath529th ring , and @xmath530 runs over the different rings . here ,
@xmath531^{1\over2}$ ] is the gravitational potential of the satellite and @xmath532 is its characteristic radius . to obtain the warp response of the galaxy
we evaluate equations ( 59 ) numerically for @xmath530 and include the complex torques @xmath533 on the right hand sides of equations ( 55 ) .
figure 16a shows the briggs plot of the warp response @xmath534 of the galaxy resulting from the retrograde passage of a satellite of mass @xmath535 .
note that the dependence of @xmath534 at the times shown is of the form of a _ leading _ spiral wave and that it is qualitatively similar to the leading spiral waves observed in many warped galaxies ( briggs 1990 ) .
note also that the form of the spiral wave is preserved for a long time .
however , the amplitude of the warp @xmath16 is significantly smaller than the warps observed in many cases , and this agrees with the conclusions of ht . nevertheless , it is of interest to understand the behavior .
figure 16b shows the warp response for the retrograde case with and without gravitational interactions ( @xmath470 ) between the rings .
clearly these interactions give a strong phase - locking of the inner rings of the disk ( roughly , @xmath536 to @xmath537 or @xmath538 to @xmath539 kpc ) which have the same @xmath540 as a function of time .
the phase - locking of two and three rings was disscussed in 2.2 and 2.4 .
figure 17a shows the briggs plot for the warp resulting from the prograde passage of the satellite .
for early times ( @xmath541 gyr ) the form of @xmath534 is a _ trailing _ spiral wave which is distinctly different from that for the retrograde passage at a similar time .
the amplitude of the warp is noticably larger than for the retrograde case but still smaller than many observed warps . for later times ( @xmath542 gyr ) the trailing wave `
unwraps , ' evolving into a leading spiral wave of form similar to that for the retrograde passage for radial distances out to about @xmath543 kpc . for larger @xmath8 , the @xmath534 curve has a ` spur ' where @xmath16 increase with @xmath8 while @xmath17 remains roughly constant ( see 3.2 ) .
this unwrapping appears to be due to the phase - mixing of the fast precession modes which propagate over the radial extent of the disk in times less than @xmath544 gyr .
the different behavior of the prograde and retrograde cases for early times ( @xmath545 gyr ) can be understood by considering the response of a single ring with or without gravitational interactions between rings .
figure 17b shows the orbits @xmath546 $ ] of ring no .
25 ( radius @xmath547 kpc ) including the gravitational interactions for prograde and retrograde encounters .
consider the torque on the @xmath5 ring @xmath548 which enters on the right hand side of equation ( 55 ) .
the tilting of the ring during the encounter is relatively small and can be neglected for this discussion .
consequently , @xmath549 is an even function of @xmath550 if the pericenter is taken as @xmath551 .
further , @xmath549 is the same for prograde and retrograde encounters . for the considered geometry of the satellite orbit , @xmath552 , and
the magnitude of @xmath549 decreases monotonically with increasing @xmath553 . on the other hand , the @xmath71torque @xmath554 ,
although smaller in magnitude , is an odd function of time , and it has opposite signs for prograde and retrograde encounters .
very roughly @xmath555/dt$ ] for the prograde case whereas @xmath556/dt$ ] for the retrograde case . as a function of time
, @xmath557 traces out a closed _ clockwise _ path in the complex plane for the prograde case , whereas this path is _ counterclockwise _ for the retrograde case . for the prograde case ,
the path of @xmath557 is in the same clockwise sense as the path of the orbit @xmath558 for the _ fast _ precession mode ( see 2.1 ) , whereas for the retrograde case the path of @xmath557 is in the same direction of motion as @xmath7 for the _ slow _ precession mode . for this reason a prograde encounter couples strongly to the fast precession mode , whereas a retrograde encounter couples mainly to the slow precession mode . here
, we discuss the behavior of warps excited by a slowly sinking compact minor satellite .
the satellite is assumed to be minor in the respect that its mass is much smaller than that of the galaxy plus halo matter within say @xmath559 kpc .
the sinking of a more massive satellite is likely to cause substantial thickening of the disk ( walker , mihos , and hernquist 1996 ) which is not observed .
initially , the satellite is assumed to be in an approximately circular bound orbit in the halo gravitational potential at a large radius @xmath560 in the @xmath240 plane at an ange @xmath520 above the @xmath561 plane .
the satellite slowly sinks owing to dynamical friction with the halo matter [ equation ( 6a ) with @xmath61 ] which in the simplest description ( bt , p. 428 ) causes the specific angular momentum of the satellite @xmath563 to decrease as @xmath564 where @xmath565 is the coulomb logarithm .
the time for the satellite to sink from @xmath560 to the galaxy s center is @xmath566 @xmath567 the disk response is found by integrating equations ( 55 ) including the numerically evaluated torques ( 59 ) for @xmath515 at each time step for a ` symmetrized ' sinking analogous to the approach discussed in 3.1 .
figure 18 shows the briggs plots of the warp response @xmath534 of the galaxy at different times after the sinking ( @xmath568 ) of a satellite of mass @xmath569 in retrograde and prograde orbits .
a number of points are observed : \(1 ) the warp amplitudes are larger than for the passing satellite ( of mass @xmath535 ) discussed in 3.1 , but the amplitudes are still smaller than some observed warps ( briggs 1990 ) .
\(2 ) coiling in the @xmath534 curves at early times ( @xmath570 gyr ) tends to disappear at later times ( @xmath571 gyr ) .
we believe but have not proven that this is due to the phase - mixing of the fast precession modes which propagate over the radial extent of the disk in times @xmath572 gyr ( see also 3.1 ) .
\(3 ) for both the retrograde and prograde cases @xmath534 exhibits a leading spiral wave for @xmath573 gyr qualitatively of the form observed for m 83 ( see figure 2c ) . however , the tilt amplitude @xmath16 is significantly smaller than that for m 83 . the dissipative torques due to dynamical friction [ equation ( 56 ) ] and that due to relative friction [ equation ( 57 ) ] have different consequences for the warp evolution . for a dynamical friction coefficient @xmath574 ,
the warp amplitudes @xmath16 for the cases of figures 18a and 18b are reduced by factors @xmath575 while the line - of - nodes angles are roughly the same . on the other hand , for relative friction coefficients @xmath576 ,
short radial wavelengths are damped out without appreciably affecting the overall shape of the @xmath534 curves .
dekel and shlosman ( 1983 ) and toomre ( 1983 ) proposed that observed warps of galaxies may result from the fact that the dark matter halo is oblate and is rigidly tilted with respect to the inner disk .
to consider this possibility we generalize equation ( 55 ) to the case of @xmath0 tilted rings in a halo potential which is rigidly tilted by an angle , say , @xmath577 ( with @xmath578 ) with respect to the @xmath35axis of the inner disk .
we find @xmath579 @xmath580 where @xmath581 and @xmath582 .
the damping terms are still given by equations ( 56 ) and ( 57 ) .
for a special disk tilt , @xmath583 , the right - hand - side of equation ( 61 ) is zero so that @xmath584 is time - independent .
this @xmath584 corresponds to the laplacian surface of the disk ( bt , p. 413 ) .
it is given by @xmath585 @xmath586 where @xmath587 is the kronecker delta and the prime on the summation means that the diagonal terms are omitted .
the damping terms ( 56 ) and ( 57 ) are of course zero for the laplacian tilt
. if the self - gravity of the rings ( @xmath470 ) is negligible , @xmath588\theta_h$ ] . equation ( 62 ) can always be inverted to give @xmath584 if @xmath589 .
this is because the @xmath1ring system is stable for @xmath463 ( see discussion following equation ( 53b ) ) which implies @xmath590 . a general solution of equation ( 61 )
can evidently be written as @xmath591 , where @xmath592 obeys equation ( 55 ) which has no reference to the halo tilt . over a sufficiently long period of time ,
damping given by equation ( 56 ) and/or ( 57 ) will give @xmath593 as discussed in the paragraph after next .
figure 19a shows the laplacian tilt @xmath584 for a representative case with and without the self - gravity of the rings .
figure 19b shows the briggs plots @xmath534 at times @xmath594 and @xmath362 gyr for a disk started from a laplacian tilt and from a @xmath595 deviation from a laplacian tilt .
the briggs plot for the laplacian tilt is in contrast with the leading spiral wave observed , for example , in m 83 ( figures 1 and 2 ) ( briggs 1990 ) .
however , the laplacian tilt may be relevant to cases such as ngc 3718 which shows a relatively straight line - of - nodes .
there is no evident reason for a disk to be initially ` set up ' in a laplacian tilt .
figure 19b shows that deviations from the laplacian tilt evolve to give a line - of - nodes which is not straight .
as mentioned , dissipative torques due to dynamical friction [ equation ( 56 ) ] and/or relative friction [ equation(57 ) ] act to damp out the deviation ( @xmath592 ) from the laplacian tilt over a period of time .
for example , for @xmath596 in equation ( 56 ) and no relative friction , the maximum line - of - nodes deviation ( @xmath597 ) at @xmath362 gyr is reduced by a factor @xmath598 from the case with no dissipation shown in figure 19b . on the other hand for @xmath599 and no dynamical friction , the maximum line - of - nodes deviation
is reduced by a factor @xmath575 .
here we consider the possibility that at some intial time an outer gaseous disk is formed which is tilted with respect to the plane of the inner disk .
this situation could arise by the capture , tidal - disruption , radial - spreading , and cooling of a low mass gas cloud by a disk galaxy . for a spherical cloud of mass @xmath600 and radius
@xmath601 in an approximately circular orbit of radius @xmath8 , tidal breakup occurs roughly for @xmath602 @xmath603 @xmath604 , where @xmath176 is the circular velocity in the flat rotation region of the galaxy [ equation ( 6b ) ] .
treatment of the cloud disruption is beyond the scope of the present work .
instead , we study the initial value problem where at @xmath551 there is a smooth transition from an untilted inner disk to a tilted outer disk with a straight line - of - nodes .
specifically , we take @xmath605\bigg\}~ , \eqno(63a)\ ] ] @xmath606 where @xmath607 is the initial tilt angle of the outer disk .
similar behavior is found for other smooth variations of @xmath7 .
the smooth variation of @xmath7 avoids the excitation of short radial wavelength modes .
figure 20 shows the nature of the warp evolution resulting from the initial conditions ( 63 ) .
figure 21 shows a surface plot of the disk at @xmath608 gyr .
a leading spiral wave appears in the briggs plot ( panel * c * ) , and its amplitude grows with time .
progressively , the interior region of the disk ( say , @xmath609 kpc ) flattens to the plane of the inner disk while the exterior region warp amplitude @xmath16 grows .
the dependences shown in figure 20 for @xmath610 gyr are _ qualitatively _ similar to those observed for m 83 ( see figure 2 ) .
the outer part of the disk shows an inverse power law dependence of the line - of - nodes @xmath17 on @xmath8 ( panel * b * of figure 20 ) which is qualitatively similar to that for m 83 where @xmath611 .
the dissipative torques due to dynamical friction [ equation ( 56 ) ] and that due to relative friction [ equation ( 57 ) ] affect the results of figure 20 in different ways . for a dynamical friction coefficient @xmath574 , and no relative friction , the maximum warp amplitude at @xmath594 gyr is reduced by about @xmath612 . on the other hand , for no dynamical friction , but a relative friction coefficient @xmath613 , the warp at @xmath594 gyr is essentially unchanged .
in the continuum limit , the hunter and toomre ( 1969 ) equation for the vertical displacement of the disk @xmath20 can be obtained from equation ( 55 ) by recalling that @xmath614 @xmath615 $ ] .
we omit for the moment the damping terms ( 56 ) and ( 57 ) .
multiplying equation ( 55 ) by @xmath616 and taking the imaginary part gives @xmath617 @xmath618 where the left hand side of this equation follows from the same steps as for equation ( 21 ) , where @xmath619 , and where @xmath620 with @xmath621 taken as a constant for simplicity .
the continuum limit is @xmath22 and @xmath622 .
combining equations ( 64)-(66 ) and dropping the @xmath529 subscripts gives @xmath635 @xmath636^{3\over 2}}~ , \eqno(67)\ ] ] where @xmath637 .
note that @xmath638 includes the vertical restoring force of the inner disk and the halo .
thus , equation ( 67 ) is the same as hunter and toomre s ( 1969 ) equation for @xmath153 , in that they did not separate out the inner disk and did not include a halo potential .
the newtonian drag term ( 56 ) due to dynamical friction gives a contribution @xmath639 on the right hand side of equation ( 67 ) .
the damping term ( 57 ) due to relative friction between the rings gives a contribution @xmath640 \eqno(69)\ ] ] on the right hand side of equation ( 67 ) .
multiplying equation ( 67 ) by @xmath641 , integrating over the outer disk , and following the steps of ht ( their appendix b ) gives , in the absence of dissipation , the constant of the motion @xmath642\ ] ] @xmath643 @xmath644 ^ 2~ , \eqno(70)\ ] ] where @xmath645 is given by equation ( 30b ) . substituting @xmath646 $ ] from equation ( 8a ) ,
one readily finds that @xmath647 is identical to @xmath648 of equation ( 53b ) .
note that @xmath647 is non - negative and therefore the disk is stable if @xmath649 . including the dissipation terms ( 68 ) and ( 69 )
gives @xmath650 @xmath651 ^ 2~. \eqno(71)\ ] ] the relative friction is of course zero if @xmath652 is a function only of time which corresponds to a rigid tilt .
the second constant of the motion @xmath653 , found in 2.1 and 3 [ equation ( 53c ) ] , can of course be expressed in terms of @xmath20 in the continuum limit .
instead , we obtain this new constant directly from the ht equation for @xmath153 by multiplying equation ( 67 ) by @xmath654 and integrating over the outer disk .
this gives @xmath655 = { \rm const.}\ ] ] @xmath656 and @xmath657 in the continuum limit , in the absence of dissipation .
with dissipation included , we have not found a simple expression for @xmath658 .
this paper develops a representation for the antisymmetric small amplitude warp dynamics of a self - gravitating disk in terms of @xmath0 tilted concentric rings .
that is , we consider only azimuthal mode number of the warp @xmath64 ( or @xmath65 ) .
this representation is suggested by the kinematic ring model of rogstad et al .
( 1974 ) which is used in interpreting hi disk warps ( briggs 1990 ) .
the rings are considered to be in a fixed oblate ( or prolate ) halo potential and in the potential of an inner , untilted disk .
different initial value problems are studied using our @xmath1ring model .
we first consider in detail the tilting dynamics of one ring in the potential of the halo and inner disk .
the equations of motion are shown to have a particularly simple form when written in terms of the complex tilt angle @xmath659 , where @xmath16 is the actual tilt angle , and @xmath17 is the angle of the line - of - nodes relative to the inertial @xmath71axis . the single ring has a slow and a fast precession mode which are analogous to the normal modes of vibration of a non - rotating mechancial system .
for the slow ( fast ) mode @xmath224 rotates in the clockwise ( counter clockwise ) direction in the complex plane .
the dynamical equation for @xmath234 of course has an energy constant of the motion , @xmath166 .
additionally , we show that there is a second constant of the motion . by writing out the lagrangian for one ring , @xmath660
, this constant of the motion is recognized as the canonical angular momentum for the @xmath17 coordinate , @xmath661 const .
we examine the influence of the drag on the motion of one ring through the halo matter due to dynamical friction .
the full calculation of this drag ( nelson and tremaine 1995 ) is beyond the scope of this work .
instead , we include a linear drag torque in the equation of motion for @xmath234 which we estimate following the approach of weinberg ( 1985 ) .
inclusion of this term shows that dissipation can destabilize the ring tilting in a prolate halo potential .
the instability is of the negative energy type ; that is , it occurs only for @xmath662 .
we next consider the particularly interesting case of two tilted , gravitationally interacting rings of different radii in the potential of the halo and inner disk .
we derive dynamical equations for @xmath663 and @xmath664 .
this case shows the phenomenon of phase - locking of the line - of - nodes of the two rings . the gravitational interaction between the two rings
is appropriately measured by the coupling strength @xmath411 of equation ( 48 ) . as @xmath411 increases the ring motion changes from that of independent precession for @xmath665 to phase - locked precession for @xmath666 where the lines - of - nodes coincide approximately , @xmath667 .
this phase - locking persists for the case of @xmath668 rings .
we examine the influence of dissipative forces on the tilting motion of two rings .
in addition to dynamical friction , which is included as linear drag torques on each ring , we include the ` relative ' friction which can result from the differential vertical motions of two gaseous rings due to turbulent viscosity .
the corresponding drag torques on each ring have a definite form owing to the requirement that the total angular momentum of the two rings be conserved in the absence of external torques .
the turbulent viscosity is assumed to be described by the @xmath433 model of shakura ( 1973 ) and shakura and sunyaev ( 1973 ) .
we find that the relative friction , as well as the dynamical friction , can destabilize the motion of two rings in a prolate halo potential .
we study the case of two tilted rings which rotate in opposite directions .
this situation , which is pertinent to observed counter - rotating galaxies ( see , for example , jore , broeils , and haynes 1996 ) , is found to be unstable in a prolate halo potential for certain conditions in the absence of dissipation .
that is , a dynamical instability may occur . for the general case of @xmath1tilted , gravitationally interacting rings of radii @xmath669 in the potential of the halo and inner disk
we obtain equatins of motion for @xmath670 for @xmath530 .
the lagrangian has the form @xmath671 , and this gives , in the absence of dissipative torques , both an energy constant of the motion and a total canonical angular momentum constant of the motion , @xmath672 const .
we comment on the numerical solutions for @xmath7 for different values of @xmath0 , and we give a sample power spectrum for @xmath495 which shows the distribution of the @xmath496 mode frequencies .
we argue that for treatment of intial value problems of time duration @xmath673 , mode frequency differences @xmath502 smaller than @xmath503 are irrelevant .
this leads to the estimate @xmath674 for the number of rings needed to treat warps in galactic disks . in the continuum limit @xmath22 ,
the ring model is shown to give the hunter and toomre ( 1969 ) dynamical equation for the vertical displacement of the disk @xmath20 .
dissipative torques due to dynamical friction and/or relative friction ( for gaseous rings ) are shown to cause the energy of the warp @xmath647 to decrease with time .
we have numerically solved the dynamical equations for @xmath7 for @xmath495 for four different types of initial conditions which may give rise to observed warps in galaxies : ( 1 ) warp excitation by a passing satellite with relatively large impact parameter ; ( 2 ) excitation by a slowly sinking compact minor satellite ; ( 3 ) warp evolution in a tilted halo potential ( dekel and shlosman 1983 ; and toomre 1983 ) ; and ( 4 ) warp evolution resulting from an initially tilted disk plane due to the tidal breakup of a gas cloud .
the nature of the disk response at different times is most clearly shown in the polar plots of @xmath3 versus @xmath4 which we refer to as briggs plots .
this is one of the forms used by briggs ( 1990 ) to describe the warp geometry as deduced by fitting a kinematic ring model ( rogstad et al .
1974 ) to hi observations .
for case ( 1 ) we find that the polar plots @xmath675 have the shape of a leading spiral wave qualitatively of the form observed , but the warp amplitude @xmath3 is smaller than observed in many galaxies .
the small warp amplitudes found here are in agreement with the conclusions of ht .
the leading spiral in the @xmath675 curves in the outer part of the disk ( where the self - gravity between the rings is small ) results from the dominance of the slow precession mode which gives very roughly @xmath676 , where @xmath381 is the single ring slow precession frequency .
the fast precession modes disappear in a few gyrs due to phase - mixing . for case
( 2 ) of a sinking satellite , we find that the @xmath675 curves have a large noise component and are unlike the observed curves in the absence of relative friction between the rings .
this is due to the violent excitation from the satellite passing through the disk .
however , with relative friction included , the @xmath675 curves are smooth leading spiral waves qualitatively similar in shape to the observed curves .
the warp amplitude , although significantly larger than for case ( 1 ) , is still smaller than that for many observed warps .
we find that the inner part of the disk where @xmath3 is small has an approximately straight line - of - nodes ( @xmath4 independent of @xmath529 ) for both cases ( 1 ) and ( 2 ) as observed by briggs ( 1990 ) ( his ` rule of behavior ' no .
we show that this is due to the above - mentioned phase - locking of the lines - of - nodes of the inner rings of the disk due to self gravity between the rings . for case
( 3 ) of an oblate halo potential rigidly tilted with respect to the inner disk of the galaxy , there is a unique time independent laplacian warp @xmath677 ( bt , p. 413 ) which has a straight line - of - nodes ( @xmath678 const . ) .
however , it is unlikely that a galactic hi disk is ` set up ' with the laplacian tilt .
deviations from @xmath584 evolve to give a line - of - nodes which is not straight . over the age of the galaxy
, the deviations from @xmath584 may damp out due to dynamical friction and/or relative friction between the rings .
this case is possibly relevant to the warp of ngc 3718 which shows an approximately straight line - of - nodes out to @xmath679 ( briggs 1990 ) .
case ( 4 ) of an initially tilted disk plane with a straight line - of - nodes gives : ( 1 ) polar plots @xmath675 for @xmath680 gyr with the leading spiral shape qualitatively of the form observed ; ( 2 ) straight line - of - nodes ( @xmath681 const . ) in the inner part of the disk as observed as a result of phase - locking due to self - gravity ; and ( 3 ) a warp amplitude linearly dependent on the initial tilt of the disk plane .
further , the line - of - nodes angle @xmath682 in the outer part of the disk is found to have an inverse power law dependence on @xmath8 qualitatively of the form of shown for m 83 ( see figure 2b ) .
beal , j.w . ,
brettschneider , christofilos , n.c . ,
hester , r.e . , lamb , w.a . ,
sherwood , w.a . ,
spoerlein , r.l . ,
weiss , p.b .
, & wright , r.e .
1969 , in _ plasma physics and controlled nuclear fusion research _
( international atomic energy agency , vienna ) , vol . | the present work represents an attempt to understand the ` rules of behavior ' of observed warps in the hi disks of spiral galaxies found by briggs ( 1990 ) .
in contrast with most earlier theoretical work , the present study investigates different initial value problems of a warped disk in an oblate ( or prolate ) halo potential , and it represents the disk warp in terms of @xmath0 independently tilted , self - gravitating , concentric rings .
this representation gives new insight into the disk warping .
a new constant of the motion of @xmath0 tilted rings is identified ( in addition to the energy ) .
the phenomenon of phase - locking of the lines - of - nodes of nearby rings due to self - gravity is demonstrated .
we consider the influence of dynamical friction due to ring motion through the halo matter as well as friction between gaseous rings with different vertical motions due to turbulent viscosity .
we first consider the dynamics of one , two , and three tilted rings of different radii in a halo potential .
we go on to develop dynamical equations for @xmath1rings which are most simply expressed in terms of the complex tilt angles @xmath2 , where @xmath3 is the actual tilt angle and @xmath4 the line - of - nodes angle for the @xmath5 ring ( @xmath6 ) .
we numerically solve the equations for @xmath7 for four different types of initial conditions : ( 1 ) warp excitation by a passing satellite , ( 2 ) excitation by a sinking compact minor satellite , ( 3 ) warp evolution due to a tilted halo potential , and ( 4 ) warp evolution resulting from an initially tilted disk plane . |
josephson - junction charge qubits are known to be candidates for scalable solid - state quantum computing circuits @xcite , @xcite , @xcite . here
we consider a superconducting charge qubit which consists of two josephson junctions embedded in a loop with very small inductance @xmath0 , typically in the ph range .
this insures effective decoupling from the environment .
however , in the practical implementation of qubit circuitry it is important to have the loop inductance as much as possible consistent with a proper operation of a qubit .
a relative large loop inductance facilitates a qubit control biasing schemes and the formation , control and readout of two - qubit quantum gates .
these considerations stimulated some investigations of the role the loop inductance plays in the dynamic properties of charge qubits @xcite , @xcite , @xcite , where for the small loop inductance the corrections to the energy levels due to finite inductance of the loop have been found .
the corrections have been obtained by perturbation expansion of the energy over small parameter @xmath1 , where @xmath2 is the josephson junction inductance . for complex josephson circuit
the construction of quantum hamiltonian which accounts for finite inductances of superconducting loops can be made with the aid of the graph theory @xcite .
this approach has been developed in @xcite for systematic derivation of the hamiltonian of superconducting circuits and has been applied for the calculations of the effects of the finite loop inductance both for flux @xcite and charge @xcite qubits . in principle , the account for a finite loop inductance ( even if it is small ) requires for the magnetic energy to be included in quantum mechanical hamiltonian of a qubit from the very beginning .
it allows one to obtain the effects of the interaction between two - level qubit and its own lc circuit .
in addition it allows a correct definition of the current operator in terms of its matrix elements in a two level basis . in this paper
we investigate the effect of finite loop inductance and gate capacitance for a asymmetric charge qubit , which consists of two josephson junctions embedded in a superconducting loop .
the construction of exact lagrangian and hamiltonian for the charge qubit is given in section ii and section iii , respectively .
the approximation for exact hamiltonian for small @xmath0 is made in section iv .
it is shown that hamiltonian is decomposed in three parts : qubit part , lc - oscillator part and the qubit - lc oscillator interaction part . in this approximation
the energy levels of the charge qubit are explicitly dependent on the gate capacitance and critical current asymmetry and , in addition , are shifted due to vacuum fluctuations of lc oscillator .
the current operator both in charge and in eigenstate basis is obtained in section v. it is shown that the asymmetry of critical currents of the josephson junctions results in additional terms in the operator of critical current .
the corrections to the qubit energies due to its interaction with lc circuit and their dependence on critical current asymmetry and on gate capacitance are calculated in section vi .
we consider here a charge qubit in the arrangement , which has been first proposed in @xcite ( see fig.[fig1 ] ) .
the qubit consists of two josephson junctions in a loop with very small inductance @xmath0 , typically in the ph range .
this insures effective decoupling from the environment . as a general case
we assume that two junctions have different critical currents @xmath3 , @xmath4 and capacitance @xmath5 , @xmath6 .
the josephson energy @xmath7 , where @xmath8 is the flux quantum , is assumed to be much less than the coulomb energy @xmath9 , so that the charge at the gate is well defined .
the lagrangian of this qubit is the difference between the charge energy in the junction capacitors and the josephson plus magnetic energy : @xmath10 where @xmath11 is the electric energy of jj s and gate capacities @xmath12 @xmath13 , where @xmath14 is the voltage drop across the inductance . in virtue of josephson relations
@xmath15 the voltage drop across the inductance is : @xmath16\ ] ] where @xmath17 is the external flux . in terms of the phases
@xmath18 , @xmath19 , @xmath20 lagrangian ( [ lag1 ] ) takes the form : @xmath21 where @xmath22 .
next we make the known ( likharev and averin ) redefinition of the josephson phases : @xmath23 ; @xmath24 .
lagrangian ( [ lag2 ] ) takes the form : @xmath25
conjugate variables are defined in a standard way : @xmath26 @xmath27 from these equations we express phases in terms of conjugate variables : @xmath28\ ] ] @xmath29\ ] ] where @xmath30 , @xmath31 , @xmath32 , @xmath33 , @xmath34 .
now we construct hamiltonian : @xmath35 eliminating time derivatives of the phases from ( [ ham1 ] ) with the aid of ( [ deltadot ] ) , ( [ phidot ] ) , we obtain the final expression for hamiltonian of the asymmetric charge qubit : @xmath36 where @xmath37 , @xmath38 , @xmath39 , @xmath40 .
the first two equations of motion @xmath41 are given by eqs .
( [ phidot ] ) and ( [ deltadot ] ) .
two other equations are as follows : @xmath42 @xmath43 below we consider hamiltonian ( [ ham2 ] ) as quantum mechanical with commutator relations imposed on its variables @xmath44 = i;\quad \left [ { \delta , n_\delta } \right ] = i\ ] ] where @xmath45 , @xmath46 .
obviously , hamiltonian ( [ ham2 ] ) is 2d nonlinear oscillator .
we assume l is small , so that its frequency @xmath47 . therefore we can consider @xmath48 as fast variable with fast oscillations near the point @xmath49 , the minimum of potential @xmath50 ( see ( [ ham2 ] ) ) : @xmath51 we single out of this potential the fast variable @xmath48 , which describes the interaction of the qubit with its own @xmath52 circuit .
the point of minimum @xmath49 of @xmath53 ( [ pot ] ) with respect to @xmath48 is defined from @xmath54 : @xmath55 in what follows we consider @xmath56 as slow variable and expand @xmath57 near the point of minimum to the third order in @xmath48 ( @xmath58 ) . in the vicinity of @xmath49
the potential @xmath57 can be written as : @xmath59 where @xmath60 is the operator conjugate to @xmath61 . with the aid of ( [ phic ] )
we write @xmath62 to the first order in @xmath63 : @xmath64 therefore , we decompose hamiltonian ( [ ham2 ] ) into oscillator , qubit and interaction parts : @xmath65 , where @xmath66 @xmath67 @xmath68 in the above equations we disregard the constant term which is proportional to @xmath69 .
the first term in ( [ hint ] ) describes the interaction of the phase variables of the qubit , @xmath48 and @xmath56 via the gate , the other terms are responsible for the interaction of the qubit with its own lc circuit .
first we quantize ( [ hosc ] ) according to ( @xmath70 = - i$ ] ) : @xmath71 where @xmath72=1 $ ] . in addition
, we use the two level approximation in the charge basis : @xmath73 ; @xmath74 ; @xmath75 with pauli operators @xmath76 in this approximation @xmath77=@xmath78 , since these operators couple charge states which differs by two cooper pairs
. therefore , @xmath79=@xmath80=@xmath81 .
now we write down hamiltonian ( [ hosc],[hqb ] , [ hint ] ) within two level subspace in terms of pauli operators @xmath82 and oscillator operators @xmath83 .
we obtain the following result : @xmath84 where @xmath85 @xmath86\left ( { a^ + - a } \right)\\\nonumber + \frac{{e_j \beta } } { { 64}}\eta ^2 \left ( { \left ( { 1 + \frac{{\xi ^2 } } { 8 } } \right ) - \left ( { 1 - \frac{{\xi ^2 } } { 8 } } \right)\cos \varphi _ x } \right)\left ( { a^ + + a } \right)^2 \\\nonumber + \frac{{e_j \beta } } { { 384\sqrt 2 } } \eta ^3 \left ( { 1 - \frac{{\xi ^2 } } { 4 } } \right ) \sin \varphi _ x \left ( { a^ + + a } \right)^3\end{aligned}\ ] ] @xmath87 @xmath88 where @xmath89 here we neglect the interaction of the qubit with its own lc circuit . it is justified if @xmath63 is sufficiently small so that the energy levels of the qubit oscillator are located much higher than the ground level of the qubit . the approximation we make here is to average hamiltonian ( [ hamtl ] ) over the vacuum state , @xmath90 , of the qubit oscillator .
the result is as follows : @xmath91 where @xmath92\cos \varphi _ x\ ] ] @xmath93 @xmath94 @xmath95\ ] ] hamiltonian ( [ h ] ) has the corrections on the order of @xmath96 which are due to the vacuum fluctuations of the lc oscillator . since in a charge
qubit @xmath97 these corrections in principle might be not very small .
hamiltonian ( [ h ] ) can be made diagonal in the eigenbasis with the aid of the matrix @xcite : @xmath98 where @xmath99 , @xmath100 , @xmath101 , @xmath102 ; @xmath103 , @xmath104 .
the qubit hamiltonian in eigenstate basis , therefore , reads : @xmath105 where @xmath106 is given by ( [ w ] ) and @xmath107 ^ 2 } \ ] ] as is seen from ( [ de ] ) the energies of the charge qubit account for its full asymmetry and are explicitly dependent on the inductance and on the gate capacitance .
from the first principles the average current in the loop is equal to the first derivative of the eigenenergy relative to the external flux : @xmath108 this expression can be rewritten in terms of exact hamiltonian of a system : @xmath109 from ( [ currb ] ) we would make ansatz that the current operator is as follows : @xmath110 however ( [ currc ] ) is not a consequence of ( [ currb ] ) .
therefore , the ansatz ( [ currc ] ) must be proved in every case , since the current operator in the form of eq .
( [ currc ] ) has to be consistent with its definition in terms of variables of hamiltonian @xmath111 .
the prove for our case is given below .
the current operator across every junction is a sum of a supercurrent and a current through the capacitor : @xmath112 we are interested in the current through the inductance coil , @xmath113 in ( [ curr1 ] ) ( see fig.[fig1 ] )
. direct calculation of @xmath113 with the aid of ( [ phidot],[deltadot],[nphidot],[ndeltadot ] ) yield the result : @xmath114 which is independent of parameters of a particular junction in the loop . from the other hand the expression ( [ currop1 ] ) can be obtained from our hamiltonian ( [ ham2 ] ) with the aid of ( [ currc ] ) .
therefore , the equation ( [ currc ] ) gives us the true expression for the current operator .
it is important to note that the proper expression for the current operator ( [ currop1 ] ) can not be obtained without magnetic energy term in the original lagrangian ( [ lag1 ] ) .
it follows from ( [ ham2 ] ) and ( [ currop1 ] ) that @xmath115 \ne 0 $ ] .
therefore , an eigenstate of @xmath111 can not possess a definite current value . for the charge
qubit the current operator can be obtained from ( [ h ] ) with the aid of its definition ( [ currc ] ) : @xmath116\ ] ] the transformation of ( [ currqb ] ) in the eigenstate basis yeild the result : @xmath117 where @xmath118 @xmath119 @xmath120}}{{\cos ^2 \frac{{\varphi _ x } } { 2 } + \frac{{\xi ^2 } } { 4}\sin ^2 \frac{{\varphi _ x } } { 2}}}\ ] ] @xmath121}}{{\cos ^2 \frac{{\varphi _ x } } { 2 } + \frac{{\xi ^2 } } { 4}\sin ^2 \frac{{\varphi _ x } } { 2}}}\ ] ] therefore , the current operator is not diagonal neither in the charge basis no in eigenstate basis .
if we neglect the inductance and the current asymmetry ( @xmath122 , @xmath123 ) , we obtain : @xmath124 , @xmath125 @xmath126 where @xmath127 the existence of nondiagonal elements of the current operator in eigenstate basis is important if we consider the inductive coupling of several qubits .
here we enlarge the hilbert space to add to two qubit states two photon states of lc resonator , @xmath128 .
the transformed hamiltonian ( [ hamtl ] ) , which accounts for transitions between ground and excited state of lc resonator will read : @xmath129 where @xmath130\ ] ] @xmath131\ ] ] @xmath132 @xmath133 @xmath134 @xmath135 @xmath136 @xmath137 @xmath138 the operators @xmath139 are defined on the eigenstates @xmath140 , @xmath141 : @xmath142 in addition , we restrict photon subspace to photon numbers n=0,1 .
the basis set for our photon+ qubit system is : @xmath143 within this basis the wave function for hamiltonian ( [ hamqblc ] ) is decomposed as : @xmath144 the schrodinger equation @xmath145 takes the form : @xmath146\left| { 0 \uparrow } \right\rangle \\
+ \left [ { b\left ( { w + \frac{1}{2}\delta e - e } \right ) + c\left ( { - ra + irb + sx + isy } \right ) + d\left ( { q_2 - q_1 + rc - sz } \right ) } \right]\left| { 0 \downarrow } \right\rangle
\\ + \left [ { a\left ( { q_1 + q_2 + rc + sz } \right ) + b\left ( { ra + irb + sx - isy } \right ) } \right . \\
{ + c\left ( { w + p - \frac{1}{2}\delta e + t - e } \right ) - d\frac{{tc}}{{\varepsilon ^2 } } \left ( { a + ib } \right ) } \right]\left| { 1 \uparrow } \right\rangle
\\ + \left [ { a\left ( { ra - irb + sx + isy } \right ) + b\left ( { q_1 + q_2 - rc - sz } \right ) } \right . \\
{ - c\frac{{tc}}{{\varepsilon ^2 } } \left ( { a - ib } \right ) + d\left ( { w + p + \frac{1}{2}\delta e - t - e } \right ) } \right]\left| { 1 \downarrow } \right\rangle = 0 \\ \end{array}\ ] ] the energy levels are defined by equating of the determinant of this equation to zero . in order to simplify the problem we assume the inductance of the qubit is very small . in this limit
we may put @xmath147 , @xmath148 , @xmath149 , @xmath150 , @xmath151 .
the hamiltonian ( [ hamqblc ] ) is reduced to : @xmath152 it is important that the inductance can not be eliminated at all in the two photon approximation , since the quantity @xmath63 is in the denominators of @xmath153 and @xmath154 .
the schredinger equation for hamiltonian ( [ hamred ] ) is as follows : @xmath155\left| { 0 \uparrow } \right\rangle
\\ + \left [ { b\left ( { \frac{1}{2}\delta e - e } \right ) + cr\left ( { - a + ib } \right ) + d\left ( { - q_1 + rc } \right ) } \right]\left| { 0 \downarrow } \right\rangle
\\ + \left [ { \left [ { a\left ( { q_1 + rc } \right ) + br\left ( { a + ib } \right ) + c\left ( { e_0 - \frac{1}{2}\delta e - e } \right ) } \right . } \right]\left| { 1 \uparrow } \right\rangle
\\ + \left [ { ar\left ( { a - ib } \right ) + b\left ( { q_1 - rc } \right ) + d\left ( { e_0 + \frac{1}{2}\delta e - e } \right ) } \right]\left| { 1 \downarrow } \right\rangle = 0 \\ \end{array}\ ] ] the energy levels are defined by equating of the determinant of the following matrix to zero : @xmath156 below we calculate the energies for the following set of the qubit parameters : @xmath157 , @xmath158 , @xmath159 ; @xmath160 .
the calculations have been performed for two cases . for noninteracting qubit we used the expression ( [ de ] ) , where the lc circuit only renormalizes the energy due to the vacuum fluctuations of lc ocsillator ( factor @xmath161 in ( [ de ] ) ) . for the qubit which interacts with its own lc circuit we solved the equation ( [ matrix ] ) .
the plots of the qubit energy levels are shown on fig.[fig2 ] . as is seen from these plots the finite value of @xmath63 ,
though it is rather small , modifies the gap between ground and first excited energy levels .
the calculations show that at the point @xmath162 , where the gap is minimum , the gap for noninteracting qubit is @xmath163 .
the interaction of the qubit with its lc circuit reduces the gap to @xmath164 .
the dependance of minimum gap ( at the point @xmath159 , @xmath162 ) on @xmath63 is shown on fig.[fig3 ] . as is seen from fig.[fig3 ] the noninteracting qubit is slightly modified by the inductance .
a small reduction of the gap with the increase of @xmath63 is due to the factor @xmath161 in ( [ de ] ) .
however , the reduction of the gap with the increase of @xmath63 for interacting qubit is much more significant ( black triangles on fig.[fig3 ] ) .
it is important to note that this effect is more pronounced for relative large @xmath165 s . for @xmath166
the interaction with lc circuit does not alter the energies ( compared to those for noninteracting case ) . | the hamiltonian of a charge qubit , which consists of two josephson junctions is found within well known quantum mechanical procedure .
the inductance of the qubit is included from the very beginning .
it allows a selfconsistent derivation of the current operator in a two state basis .
it is shown that the current operator has nonzero nondiagonal matrix elements both in the charge and the eigenstate basis .
it is also shown that the interaction of the qubit with its own lc resonator has a noticeable influence on the qubit energies .
the influence of the junctions asymmetry and the gate capacitance on the matrix elements of the current operator and on the qubit energies are calculated .
the results obtained in the paper are important for the circuits where two or more charge qubits are coupled with the aid of inductive coil . |
for @xmath2 and @xmath3 , we define the dilation operator on @xmath4 by @xmath5 and the translation operator by @xmath6 . given a @xmath7 real , expansive matrix @xmath8 and a lattice of the form @xmath9 for @xmath10
, we consider wavelet systems of the form @xmath11 where the fourier transform of @xmath1 has compact support .
our aim is , for any given real , expansive dilation matrix @xmath8 , to construct wavelet frames with good regularity properties and with a dual frame generator of the form @xmath12 for some explicitly given coefficients @xmath13 and @xmath14 .
this will generalize and extend the one - dimensional results on constructions of dual wavelet frames in @xcite to higher dimensions .
the extension is non - trivial since it is unclear how to determine the translation lattice @xmath15 and how to control the support of the generators in the fourier domain .
this will be done by considering suitable norms in @xmath16 and non - overlapping packing of ellipsoids in lattice arrangements .
the construction of redundant wavelet representations in higher dimensions is usually based on extension principles @xcite . by making use of extension principles one is restricted to considering expansive dilations @xmath8 with integer coefficients .
our constructions work for any real , expansive dilation .
moreover , in the extension principle the number of generators often increases with the smoothness of the generators .
we will construct pairs of dual wavelet frames generated by one smooth function with good time localization .
it is a well - known fact that a wavelet frame need not have dual frames with wavelet structure . in @xcite frame wavelets with compact support and explicit analytic form
are constructed for real dilation matrices .
however , no dual frames are presented for these wavelet frames .
this can potentially be a problem because it might be difficult or even impossible to find a dual frame with wavelet structure .
since we exhibit _ pairs _ of dual wavelet frames , this issue is avoided .
the principal importance of having a dual generator of the form ( [ eq:26 ] ) is that it will inherit properties from @xmath1 preserved by dilation and linearity , _
e.g._vanishing moments , good time localization and regularity properties . for a more complete account of such matters we refer to @xcite . in the rest of this introduction
we review basic definitions .
a frame for a separable hilbert space @xmath17 is a countable collection of vectors @xmath18 for which there are constants @xmath19 such that @xmath20 if the upper bound holds in the above inequality , then @xmath21 is said to be a bessel sequence with bessel constant @xmath22 . for a bessel sequence @xmath21
we define the frame operator by @xmath23 this operator is bounded , invertible , and positive .
a frame @xmath21 is said to be _ tight _ if we can choose @xmath24 ; this is equivalent to @xmath25 where @xmath26 is the identity operator .
two bessel sequences @xmath21 and @xmath27 are said to be _ dual _ frames if @xmath28 it can be shown that two such bessel sequences are indeed frames . given a frame @xmath21 ,
at least one dual always exists ; it is called the canonical dual and is given by @xmath29 . only a frame , which is not a basis , has several duals . for @xmath30
the fourier transform is defined by @xmath31 with the usual extension to @xmath4 .
sets in @xmath16 are , in general , considered equal if they are equal up to sets of measure zero .
the boundary of a set @xmath32 is denoted by @xmath33 , the interior by @xmath34 , and the closure by @xmath35 .
let @xmath36 . a _
multiplicative tiling set _
@xmath32 for @xmath37 is a subset of positive measure such that @xmath38 in this case we say that @xmath39 is an almost everywhere partition of @xmath16 , or that it tiles @xmath16 .
a multiplicative tiling set @xmath32 is _ bounded _ if @xmath32 is a bounded set and @xmath40 . by @xmath41-dilative periodicity of a function
@xmath42 we understand @xmath43 for a.e .
@xmath44 , and by a @xmath41-dilative partition of unity we understand @xmath45 ; note that the functions in the `` partition of unity '' are not assumed to be non - negative , but can take any real or complex value .
a ( full - rank ) lattice @xmath15 in @xmath16 is a point set of the form @xmath9 for some @xmath10 .
the determinant of @xmath15 is @xmath46 ; note that the generating matrix @xmath47 is not unique , and that @xmath48 is independent of the particular choice of @xmath47 .
fix the dimension @xmath49 .
we let @xmath50 be expansive , _ i.e.,_all eigenvalues of @xmath8 have absolute value greater than one , and denote the transpose matrix by @xmath51 . for any such dilation @xmath8
, we want to construct a pair of functions that generate dual wavelet frames for some translation lattice .
our construction is based on the following result which is a consequence of the characterizing equations for dual wavelet frames by chui , czaja , maggioni , and weiss ( * ? ? ?
* theorem 4 ) .
[ thm : dual - charac - rn ] let @xmath2 be expansive , let @xmath15 be a lattice in @xmath16 , and let @xmath52 , @xmath53 .
suppose that the two wavelet systems @xmath54 and @xmath55 form bessel families . then @xmath56 and @xmath57 will be dual frames if the following conditions hold @xmath58 by @xmath59 for @xmath60 , condition ( [ eq : nondiagonalterm - rn ] ) becomes @xmath61 we use the notation as in @xcite , thus @xmath62 and @xmath63 . since @xmath64 for any @xmath65 , equation ( [ eq:1 ] ) yields @xmath66 for @xmath67 .
by @xmath68 , we can rewrite as @xmath69 using that @xmath70 for all @xmath71 .
gathering the two equations displayed above yields @xmath72 for all @xmath73 .
the conclusion follows now from ( * ? ? ?
* theorem 4 ) . the following result
, lemma [ thm : bessel2-rn ] , gives a sufficient condition for a wavelet system to form a bessel sequence ; it is an extension of ( * ? ? ?
* theorem 11.2.3 ) from @xmath74 to @xmath4 .
[ thm : bessel2-rn ] let @xmath2 be expansive , @xmath15 a lattice in @xmath16 , and @xmath75 .
suppose that , for some set @xmath76 satisfying @xmath77 , @xmath78 then the wavelet system @xmath79 is a bessel sequence with bound @xmath22 .
further , if also @xmath80 holds , then @xmath81 is a frame for @xmath4 with frame bounds @xmath82 and @xmath22 .
the statement follows directly by applying theorem 3.1 in @xcite on generalized shift invariant systems to wavelet systems . in the general result for generalized shift invariant systems ( *
* theorem 3.1 ) , the supremum / infimum is taken over @xmath16 , but because of the @xmath41-dilative periodicity of the series in ( [ eq:14 ] ) and ( [ eq : lemma - suff - lower ] ) for wavelet systems , it suffices to take the supremum / infimum over a set @xmath76 that has the property that @xmath83 up to sets of measure zero .
theorem [ thm : dual - charac - rn ] and lemma [ thm : bessel2-rn ] are all we need to prove the following result on pairs of dual wavelet frames .
[ thm : constr - dual - wavelet - rn - general ] let @xmath2 be expansive and @xmath0 .
suppose that @xmath84 is a bounded , real - valued function with @xmath85 for some @xmath86 and some bounded multiplicative tiling set @xmath32 for @xmath87 , and that @xmath88 let @xmath89 for @xmath90 and let @xmath91 and @xmath92 .
take a lattice @xmath15 in @xmath16 such that @xmath93 and define the function @xmath94 by @xmath95 if @xmath96 and @xmath97 for @xmath98 , then the functions @xmath1 and @xmath94 generate dual frames @xmath99 and @xmath100 for @xmath4 . on the fourier side , the definition in ( [ eq : dual - generator1-rn ] )
becomes @xmath101 since @xmath84 by assumption is compactly supported in a `` ringlike '' structure bounded away from the origin , this will also be the case for @xmath102 .
this property implies that @xmath1 and @xmath94 will generate wavelet bessel sequences .
the details are as follows .
the support of @xmath84 and @xmath102 is @xmath103 note that @xmath104 .
the sets @xmath39 tiles @xmath16 , whereby we see that @xmath105 since @xmath106 , condition ( [ eq:13 ] ) implies that @xmath107 for @xmath108 and @xmath109 .
therefore , using , we find that @xmath110 an application of lemma [ thm : bessel2-rn ] with @xmath111 shows that @xmath1 generates a bessel sequence .
similar calculations using will show that @xmath94 generates a bessel sequence ; in this case the sum over @xmath112 will be finite , but it will in general have more than one nonzero term . to conclude that @xmath1 and @xmath94 generate dual wavelet frames
we will show that conditions and in theorem [ thm : dual - charac - rn ] hold . by @xmath41-dilation periodicity of the sum in condition , it is sufficient to verify this condition on @xmath113 .
for @xmath114 we have by , @xmath115 \\ & \phantom{= } \;+ \hat \psi(b \xi ) \left [ b_{-1 } \hat \psi ( \xi ) + b_0 \hat \psi ( b \xi ) + \dots + b_{{d}-1}\hat \psi(b^{{d}}\xi ) \right ] + \cdots \\ & \phantom{= } \;+ \hat \psi(b^{{d}}\xi ) \left [ b_{-{d}}\hat \psi(\xi ) + \dots + b_{-1 } \hat \psi(b^{{d}-1}\xi ) + b_0 \hat
\psi(b^{{d}}\xi ) \right ] , \intertext{and further , by an expansion of these terms , } & = \sum _ { j , l = 0}^{{d } } b_{l - j } \hat\psi(b^j \xi ) \hat\psi(b^l
\xi)\\ & = b_0 \sum_{j=0}^{{d } } \hat\psi(b^j \xi)^2 + \sum _ { \substack{j , l = 0 \\ j > l}}^{{d } } ( b_{j - l}+b_{l - j } ) \hat\psi(b^j \xi ) \hat\psi(b^l \xi ) .\end{aligned}\ ] ] using that @xmath96 and @xmath116 for @xmath117 and @xmath118 , we arrive at @xmath119 exhibiting that @xmath1 and @xmath94 satisfy condition . by we see that condition implies that the functions @xmath120 and @xmath121 will have disjoint support for @xmath122 , hence ( [ eq : nondiagonalterm - rn ] ) is satisfied .
the use of the parameters @xmath123 in the definition of the dual generator together with the condition @xmath124 was first seen in the work of christensen and kim @xcite on pairs of dual gabor frames .
we can restate theorem [ thm : constr - dual - wavelet - rn - general ] for wavelet systems with standard translation lattice @xmath125 and dilation @xmath126 , where @xmath127 is so that @xmath128 .
the result follows directly by an application of the relations @xmath129 for @xmath60 and @xmath130 for @xmath131 , and the fact that @xmath132 is unitary as an operator on @xmath4 .
suppose @xmath1 , @xmath133 , @xmath8 and @xmath15 are as in theorem [ thm : constr - dual - wavelet - rn - general ] .
let @xmath134 be such that @xmath135 , and let @xmath136 .
then the functions @xmath137 and @xmath138 , where @xmath94 is defined in ( [ eq : dual - generator1-rn ] ) , generate dual frames @xmath139 and @xmath140 for @xmath4 .
the following example [ ex : adhoc - construc - rn ] is an application of theorem [ thm : constr - dual - wavelet - rn - general ] in @xmath141 for the quincunx matrix .
in particular , we construct a partition of unity of the form ( [ eq : dyadic - part2-rn ] ) for the quincunx matrix . [ ex : adhoc - construc - rn ] the quincunx matrix is defined as @xmath142 and its action on @xmath143 corresponds to a counter clockwise rotation of 45 degrees and a dilation by @xmath144 .
( -.13,-.13)(1.2,1.2)(0,0)(-.13,-0.13)(1.2,1.2 ) ( 1,0)(1,1)(0.5,0.5 ) ( 0,1)(1,1)(0.5,0.5 ) ( 0,1)(0,0.5)(0.5,0.5 ) ( 0.5,0)(0,0.5)(0.5,0.5 ) ( 0.5,0)(1,0)(0.5,0.5 ) ( 0.5 , 0.75)@xmath145 ( 0.75 , 0.5)@xmath146 ( 0.625 , 0.175)@xmath147 ( 0.175 , 0.625)@xmath148 ( 0.35 , 0.35)@xmath149 ( 1.17,-2.4pt)@xmath150 ( -2.4pt,1.17)@xmath151 define the tent shaped , piecewise linear function @xmath152 by @xmath153 where the sets @xmath154 are the triangular domains sketched in figure [ fig : quincunx - support - sets ] .
note that the value at `` the top of the tent '' is @xmath155 .
define @xmath156 as a mirroring of @xmath152 in the @xmath150 axis and the @xmath151 axis : @xmath157 since the transpose @xmath41 of the quincunx matrix also corresponds to a rotation of 45 degrees ( but clockwise ) and a dilation by @xmath144 , we see that @xmath158 .
we are now ready to apply theorem [ thm : constr - dual - wavelet - rn - general ] with @xmath159}}^2 \setminus b^{-1}({\ensuremath{\left[{-1},{1}\right]}}^2 ) = { \ensuremath{\left[{-1},{1}\right]}}^2 \setminus i_1 $ ] and @xmath160 ; the set @xmath32 is the union of the domians @xmath146 and @xmath145 and their mirrored versions .
we choose @xmath161 and @xmath162 , hence @xmath163 and @xmath164 .
therefore , @xmath165}}^2,\ ] ] that shows that we can take @xmath166 or @xmath167 , since @xmath168}}^2 + { \gamma } ) \cap { \ensuremath{\left[{-1},{1}\right]}}^2 = \emptyset$ ] whenever @xmath169 .
defining the dual generator according to yields @xmath170 using that @xmath171 , and we remark that @xmath102 is a piecewise linear function since this is the case for @xmath172 .
the conclusion from theorem [ thm : constr - dual - wavelet - rn - general ] is that @xmath1 and @xmath94 generate dual frames @xmath173 and @xmath174 for @xmath141 .
the frame bounds can be found using lemma [ thm : bessel2-rn ] since the series ( [ eq:14 ] ) and ( [ eq : lemma - suff - lower ] ) are finite sums on @xmath32 ; for @xmath175 one finds @xmath176 and @xmath177 . when the result on constructing pairs of dual wavelet frames is written in the generality of theorem [ thm : constr - dual - wavelet - rn - general ] , it is not always clear how to choose the set @xmath32 and the lattice @xmath15 .
in example [ ex : adhoc - construc - rn ] we showed how this can be done for the quincunx dilation matrix and constructed a pair of dual frame wavelets . in section [ sec : spec - case - rn ] and theorem [ thm : constr - dual - wavelet - rn - special ] we specify how to choose @xmath32 and @xmath15 for general dilations .
the issue of exhibiting functions @xmath1 satisfying the condition ( [ eq : dyadic - part2-rn ] ) is addressed in section [ sec : examples - rn ] . in one dimension , however , it is straightforward to make good choices of @xmath32 and @xmath15 as is seen by the following corollary of theorem [ thm : constr - dual - wavelet - rn - general ] .
the corollary unifies the construction procedures in theorem 2 and proposition 1 from @xcite in a general procedure .
[ thm : constr - dual - wavelet - gen - phi ] let @xmath178 , @xmath179 , and @xmath180 .
suppose that @xmath181 is a bounded , real - valued function with @xmath182 for some @xmath183 , and that @xmath184 let @xmath89 for @xmath90 , let @xmath185 , and define the function @xmath94 by @xmath186 let @xmath187}}$ ] .
if @xmath188 and @xmath189 for @xmath98 , then @xmath1 and @xmath94 generate dual frames @xmath190 and @xmath191 for @xmath74 .
in theorem [ thm : constr - dual - wavelet - rn - general ] for @xmath192 and @xmath193 we take @xmath194 as the multiplicative tiling set for @xmath195 .
the assumption on the support of @xmath84 becomes @xmath196 moreover , since @xmath197 } } , \quad \bigcup_{j =- m}^{2{d } } a^{-j}(e ) \subset { \ensuremath{\left[{-a^{c+m}},{a^{c+m}}\right]}},\ ] ] and @xmath198}}+{\gamma } ) \cap { \ensuremath{\left[{-a^{c+m}},{a^{c+m}}\right ] } } = \emptyset \quad \text{for $ { \ensuremath{\left\lvert{\gamma}\right\rvert } } \ge a^c + a^{c+m } = a^c ( 1+a^m)$,}\ ] ] the choice @xmath199 for @xmath200 satisfies equation ( [ eq:13 ] ) .
this corresponds to @xmath201 for @xmath202 .
the assumptions in corollary [ thm : constr - dual - wavelet - gen - phi ] imply that @xmath203 ; we note that in case @xmath204 , the corollary reduces to ( * ? ? ?
* theorem 2 ) .
we aim for a more automated construction procedure than what we have from theorem [ thm : constr - dual - wavelet - rn - general ] , in particular , we therefore need to deal with good ways of choosing @xmath32 and @xmath15 .
the basic idea in this automation process will be to choose @xmath32 as a dilation of the difference between @xmath205 and @xmath206 , where @xmath205 is the unit ball in a norm in which the matrix @xmath207 is expanding `` in all directions '' ; we will make this statement precise in section [ sec : dilation - matrix ] .
this idea is instrumental in the proof of theorem [ thm : constr - dual - wavelet - rn - special ] .
we need the following well - known equivalent conditions for a ( non - singular ) matrix being expansive .
[ thm : expansive - matrix - equiv ] for @xmath208 the following assertions are equivalent : @xmath41 is expansive , _
i.e.,_all eigenvalues @xmath209 of @xmath41 satisfy @xmath210 .
[ item : pairsrn1 ] for any norm @xmath211 on @xmath16 there are constants @xmath212 and @xmath213 such that @xmath214 for any @xmath215 .
[ enu : almost - expanding ] there is a hermitian norm @xmath216 on @xmath16 and a constant @xmath217 such that @xmath218 for any @xmath44 .
[ enu : really - expanding ] @xmath219 for some ellipsoid @xmath220 , @xmath127 , and @xmath221 .
[ enu : expanding - ellipsoid ] by proposition [ thm : expansive - matrix - equiv ] we have that for a given expansive matrix @xmath41 , there exists a scalar product with the induced norm @xmath216 so that @xmath222 holds for some @xmath221 .
we say that @xmath216 is a norm associated with the expansive matrix @xmath41 .
note that such a norm is not unique ; we will follow the construction as in the proof of ( * ? ? ?
* lemma 2.2 ) , so let @xmath223 and @xmath224 be as in in proposition [ thm : expansive - matrix - equiv ] for the standard euclidean norm with @xmath225 for @xmath226 , where @xmath209 are the eigenvalues of @xmath41 . for @xmath227 satisfying @xmath228
we introduce the symmetric , positive definite matrix @xmath229 : @xmath230 the scalar product associated with @xmath41 is then defined by @xmath231 .
it might not be effortless to estimate @xmath223 and @xmath224 for some given @xmath41 , but it is obvious that we just need to pick @xmath227 such that @xmath232 becomes positive semi - definite for some @xmath221 since this corresponds to @xmath233 , that is , @xmath234 for all @xmath44 .
we let @xmath205 denote the unit ball in the hermitian norm @xmath235 associated with @xmath41 , _
i.e.,_@xmath236 and we let @xmath237 denote the annulus @xmath238 the ringlike structure of @xmath237 is guaranteed by the fact that @xmath41 is expanding in all directions in the @xmath216 norm , _
i.e.,_@xmath239 which is in proposition [ thm : expansive - matrix - equiv ] .
we note that by an orthogonal substitution @xmath205 takes the form @xmath240 , where @xmath241 are the positive eigenvalues of @xmath242 and @xmath243 with the @xmath244th column of @xmath245 comprising of the @xmath244th eigenvector of @xmath242 .
the annulus @xmath237 is a bounded multiplicative tiling set for @xmath37 .
this is a consequence of the following result .
[ thm : dilated - annuli ] let @xmath246 be an expansive matrix .
for @xmath247 there is a unique @xmath248 so that @xmath249 ; that is , @xmath250 from equation ( [ eq:9 ] ) we know that @xmath251 is a nested sequence of subsets of @xmath16 , thus @xmath252 are disjoint sets . since @xmath253 and @xmath254 for @xmath108 and @xmath217 , we also have @xmath255 taking the limit @xmath256 we get ( [ eq:8 ] ) .
[ ex : hermitian - norm ] let the following dilation matrix be given @xmath257 here we are interested in the transpose matrix @xmath207 with eigenvalues @xmath258 , hence @xmath41 is an expansive matrix with @xmath259 .
the dilation matrix @xmath41 is not expanding in the standard norm @xmath260 in @xmath16 , _
i.e.,_@xmath261 , as shown by figure [ fig : ex3-euclidean - norm ] .
, @xmath262 , @xmath263 , and @xmath264 marked by solid , long dashed , dashed , and dotted lines , respectively . note that @xmath265 is non - empty , and even @xmath266 is non - empty . ] in order to have @xmath41 expanding the unit ball we need to use the hermitian norm from in proposition [ thm : expansive - matrix - equiv ] associated with @xmath41 . in
we take @xmath267 so that the real , symmetric , positive definite matrix @xmath242 is @xmath268 and let @xmath269 .
the choice @xmath267 suffices since it makes @xmath232 semi - positive definite for @xmath270 and thus @xmath271 holds for @xmath272 .
figure [ fig : ex3-adaptive - norm ] and [ fig : ex3-adaptive - norm - zoom ] illustrate that @xmath41 indeed expands the hermitian norm unit ball @xmath205 in all directions . in the hermitian norm @xmath216 associated with @xmath41 and its dilations @xmath273 .
only the boundaries are marked . ]
we also remark that the hermitian norm with @xmath274 will not make the dilation matrix @xmath41 expanding in @xmath16 ; in this case we have a situation similar to figure [ fig : ex3-euclidean - norm ] . .
boundaries of the sets @xmath205 , @xmath275 , @xmath276 , and @xmath277 marked by solid , long dashed , dashed , and dotted lines , respectively . ]
let us consider the setup in theorem [ thm : constr - dual - wavelet - rn - general ] with the set @xmath278 for some @xmath279 , where the norm @xmath280 is associated with @xmath41 .
let @xmath281 be the smallest eigenvalue of @xmath242 such that @xmath282 is the largest semi - principal axis of the ellipsoid @xmath205 , _ i.e.,_@xmath283 .
then we can take any lattice @xmath284 , where @xmath47 is a non - singular matrix satisfying @xmath285 as our translation lattice in theorem [ thm : constr - dual - wavelet - rn - general ] . to see this , recall that we are looking for a lattice @xmath286 such that
, for @xmath109 , @xmath287 for our choice of @xmath32 we find that @xmath288 and @xmath289 . since @xmath290 and similar for @xmath291 , we have the situation in ( [ eq:19 ] ) whenever @xmath292 . here
we have used that for the @xmath293-norm @xmath294 .
for @xmath295 we have @xmath296 therefore , by @xmath297 for @xmath298 , we have @xmath299 now , by assuming that @xmath47 satisfies ( [ eq:16 ] ) , we have @xmath300 for @xmath301 , hence the claim follows
. a lattice choice based on ( [ eq:16 ] ) can be rather crude , and produces consequently a wavelet system with unnecessarily many translates . from equation ( [ eq:16 ] ) it is obvious that any lattice @xmath9 with @xmath302 sufficiently small will work as translation lattice for our pair of generators @xmath1 and @xmath94 .
hence , the challenging part is to find a sparse translation lattice whereby we understand a lattice @xmath15 with large determinant @xmath303 . in the dual lattice system
this corresponds to a dense lattice @xmath286 with small volume @xmath304 of the fundamental parallelotope @xmath305 since @xmath306 . in theorem
[ thm : constr - dual - wavelet - rn - special ] in the next section we make a better choice of the translation lattice compared to what we have from ( [ eq:16 ] ) .
using a crude lattice approach as above , we can easily transform the translation lattice to the integer lattice if we allow multiple generators .
we pick a matrix @xmath47 that satisfies condition ( [ eq:16 ] ) and whose inverse is integer valued , _
i.e. , _ @xmath307 .
the conclusion from theorem [ thm : constr - dual - wavelet - rn - general ] is that @xmath308 and @xmath309 are dual frames .
the order of the quotient group @xmath310 is @xmath311 , so let @xmath312 denote a complete set of representatives of the quotient group , and define @xmath313 since @xmath314 and likewise for the dual frame , the statement follows . we list some standing assumptions and conventions for this section . [
[ general - setup . ] ] * general setup .
+ + + + + + + + + + + + + + + + we assume @xmath2 is expansive .
let @xmath315 be a hermitian norm as in in proposition [ thm : expansive - matrix - equiv ] associated with @xmath316 , let @xmath205 denote the unit ball in the @xmath216-norm , and let @xmath229 be the symmetric , positive definite matrix such that @xmath317 .
let @xmath318 , where @xmath319 are the eigenvalues of @xmath242 , and let @xmath320 be such that the spectral decomposition of @xmath242 is @xmath321 .
the following result is a special case of theorem [ thm : constr - dual - wavelet - rn - general ] , where we , in particular , specify how to choose the translation lattice @xmath15 .
since we in theorem [ thm : constr - dual - wavelet - rn - special ] define @xmath15 , it allows for a more automated construction procedure .
[ thm : constr - dual - wavelet - rn - special ] let @xmath322 be as in the general setup .
let @xmath178 and @xmath0 .
suppose that @xmath84 is a bounded , real - valued function with @xmath323 for some @xmath324 , and that holds .
take @xmath325 .
then the function @xmath1 and the function @xmath94 defined by @xmath326 \quad \text{for } x \in { { \ensuremath{\mathbb{r}}}^n},\end{aligned}\ ] ] generate dual frames @xmath327 and @xmath328 for @xmath4 note that @xmath329 and @xmath330 .
the annulus @xmath237 is a bounded multiplicative tiling set for the dilations @xmath331 by lemma [ thm : dilated - annuli ] , hence this is also the case for @xmath332 for @xmath333 .
the support of @xmath172 is @xmath334 .
therefore we can apply theorem [ thm : constr - dual - wavelet - rn - general ] with @xmath335 , @xmath336 and @xmath337 for @xmath338 so that @xmath163 and @xmath339 .
the only thing left to justify is the choice of the translation lattice @xmath15 .
we need to show that condition ( [ eq:13 ] ) with @xmath163 and @xmath339 in theorem [ thm : constr - dual - wavelet - rn - general ] is satisfied by @xmath340 . by the orthogonal substitution @xmath341 the quadratic form @xmath342 of equation ( [ eq:25 ] )
reduces to @xmath343 where @xmath344 , hence in the @xmath345 coordinates @xmath205 is given by @xmath346 which is an ellipsoid with semi axes @xmath347 .
therefore , in the @xmath348 coordinates , @xmath349 or , in the @xmath350 coordinates , @xmath351 by applying @xmath352 to this relation it becomes @xmath353 whereby we see that condition ( [ eq:13 ] ) is satisfied with @xmath163 and @xmath340 .
the dual lattice of @xmath286 is @xmath354 .
it follows from theorem [ thm : constr - dual - wavelet - rn - general ] that @xmath1 and @xmath94 generate dual frames for this choice of the translation lattice .
the frame bounds for the pair of dual frames @xmath355 and @xmath328 in theorem [ thm : constr - dual - wavelet - rn - special ] can be given explicitly as @xmath356 respectively .
the frame bounds do not depend on the specific structure of @xmath15 , but only on the determinant of @xmath15 ; in particular , the condition number @xmath357 is independent of @xmath15 . to verify these frame bounds
, we note that equation ( [ eq:20 ] ) together with the fact @xmath358 , @xmath359 imply that @xmath360 therefore , by equations ( [ eq:18 ] ) and ( [ eq:28 ] ) with @xmath335 , @xmath163 and @xmath339 , we have @xmath361 and @xmath362 for @xmath363 .
the stated frame bounds follow from lemma [ thm : bessel2-rn ] .
let @xmath8 and @xmath242 be as in example [ ex : hermitian - norm ] .
the eigenvalues of @xmath242 are @xmath364 and @xmath365 .
let the normalized ( in the standard norm ) eigenvectors of @xmath242 be columns of @xmath366 and @xmath367 , hence @xmath321 . by the orthogonal transformation @xmath341
the hermitian norm unit ball @xmath205 becomes @xmath368 which is an ellipse with semimajor axis @xmath369 and semiminor axis @xmath370 .
since @xmath371 , we have @xmath372 by the orthogonal substitution back to @xmath350 coordinates , we get @xmath373 suppose that @xmath84 is a bounded , real - valued function with @xmath374 for @xmath375 that satisfies the @xmath41-dilative partition . since @xmath376 we need to take @xmath377 and @xmath378 , see figure [ fig : ex3-gamma - star - c1 ] and [ fig : ex3-gamma - c1 ] . for @xmath376
is shown by dots , and the boundary of the set @xmath379 by a solid line .
boundaries of the set @xmath379 translated to several different @xmath380 are shown with dashed lines . recall that @xmath381 , hence @xmath382 for @xmath380 . ] for @xmath376 . ]
let the setup up and assumptions be as in theorem [ thm : constr - dual - wavelet - rn - special ] , except for the lattice @xmath15 which we want to choose differently .
as in section [ sec : choos - transl - latt ] the dual lattice @xmath286 needs to satisfy ( [ eq:19 ] ) for @xmath383 .
we want to choose @xmath286 as dense as possible since this will make the translation lattice @xmath15 as sparse as possible and the wavelet system with as few translates as possible . since @xmath384
, we are looking for lattices @xmath286 that packs the ellipsoids @xmath385 , @xmath112 , in a non - overlapping , optimal way . by the coordinate transformation @xmath386
, the ellipsoid @xmath379 turns into the standard unit ball @xmath387 in @xmath16 .
this calculations are as follows .
@xmath388 and we arrive at a standard sphere packing problem with lattice arrangement of non - overlapping unit @xmath389-balls .
the proportion of the euclidean space @xmath16 filled by the balls is called the density of the arrangement , and it is this density we want as high as possible .
taking @xmath15 as in theorem [ thm : constr - dual - wavelet - rn - special ] corresponds to a square packing of the unit @xmath389-balls @xmath390 by the lattice @xmath391 , _
the density of this packing is @xmath393 , where @xmath394 is the volume of the @xmath389-ball : @xmath395 and @xmath396 .
this is not the densest packing of balls in @xmath16 since there exists a lattice with density bigger than @xmath397 for each @xmath398 @xcite ; a slight improvement of this lower bound was obtained in @xcite for @xmath399 .
moreover , the densest lattice packing of hyperspheres is known up to dimension @xmath400 , see @xcite ; it is precisely this dense lattice we want to use in place of @xmath391 ( at least whenever @xmath401 ) . in @xmath143 lagrange
proved that the hexagonal packing , where each ball touches @xmath402 other balls in a hexagonal lattice , has the highest density @xmath403 . hence using @xmath404 with @xmath405 instead of @xmath406
improves the packing by a factor of @xmath407 it is easily seen that this factor equals the relation between the area of the fundamental parallelogram of the two lattices @xmath408 . in figure
[ fig : ex3-gamma - star - c1 ] we see that each ellipse only touches @xmath409 other ellipses corresponding to the square packing @xmath410 ; in the optimal packing each ellipse touch @xmath402 others . in @xmath411 gauss proved that the highest density is @xmath412 obtained by the hexagonal close and face - centered cubic packing ; here each ball touches @xmath413 other balls .
with theorem [ thm : constr - dual - wavelet - rn - special ] at hand the only issue left is to specify how to construct functions satisfying the partition of unity ( [ eq : dyadic - part2-rn ] ) for any given expansive matrix .
in the two examples of this section we outline possible ways of achieving this . as usual
we fix the dimension @xmath49 and the expansive matrix @xmath414 . in the examples in this section
we construct functions satisfying the assumptions in theorem [ thm : constr - dual - wavelet - rn - special ] , that is , a real - valued function @xmath415 with @xmath416 for some @xmath324 and @xmath178 so that the @xmath41-dilative partition @xmath417 holds . in the construction
we will use that the radial coordinate of the surface of the ellipsoid @xmath418 , @xmath60 , can be parametrized by the @xmath419 angular coordinates @xmath420 .
the radial coordinate expression will be of the form @xmath421 for some positive , trigonometric function @xmath422 , where @xmath422 is bounded away from zero and infinity with the specific form of @xmath422 depending on the dimension @xmath389 and the length and orientation of the ellipsoid axes .
we illustrate this with the following example in @xmath423 .
we want to find the radial coordinate @xmath424 of the ellipsoid @xmath425 as a function the angular coordinates @xmath426 and @xmath427 .
we express @xmath428 in the hyperspherical coordinates @xmath429 \times [ 0,\pi ] \times [ 0 , 2\pi)$ ] as follows : @xmath430 then we substitute @xmath431 , @xmath432 , in the expression above and factor out @xmath433 to obtain @xmath434 , where @xmath435 the conclusion is that @xmath436 .
[ exa : d - one ] for @xmath437 in theorem [ thm : constr - dual - wavelet - rn - special ] we want @xmath438 for any given @xmath439 . the choice @xmath437 will fix the `` size '' of the support of @xmath152 so that @xmath440 for some @xmath333 .
now let @xmath441 and @xmath442 denote the radial coordinates of the surface of the ellipsoids @xmath443 and @xmath444 parametrized by @xmath419 angular coordinates @xmath445 , respectively .
let @xmath446 be a continuous function on the annulus @xmath447 satisfying @xmath448 and @xmath449 . using the parametrizations @xmath450 of the surfaces of the two ellipsoids and fixing the @xmath419 angular coordinates we realize that we only have to find a continuous function @xmath451 } } \to { { \ensuremath{\mathbb{r}}}}$ ] of one variable ( the radial coordinate ) satisfying @xmath452 and @xmath453 .
for example the general function @xmath454 of @xmath455 variables can be any of the functions below : define @xmath461 by : @xmath462 this way @xmath152 becomes a @xmath41-dilative partition of unity with @xmath463 , so we can apply theorem [ thm : constr - dual - wavelet - rn - special ] with @xmath464 and @xmath160 .
we can simplify the expressions for the radial coordinates @xmath450 of the surface of the ellipsoids @xmath465 and @xmath444 from the previous example by a suitable coordinate change .
the idea is to transform the ellipsoid @xmath466 to the standard unit ball @xmath387 by a first coordinate change @xmath467 .
this will transform the outer ellipsoid @xmath379 to another ellipsoid .
a second and orthogonal coordinate transform @xmath468 will make the semiaxes of this new ellipsoid parallel to the coordinate axes , leaving the standard unit ball @xmath387 unchanged .
here @xmath469 comes from the spectral decomposition of @xmath470 , _
i.e.,_@xmath471 . in the @xmath472
coordinates @xmath473 is a constant and @xmath474 with @xmath446 of the form ( [ eq:27 ] ) for @xmath475 and likewise for @xmath476 .
[ exa : all - d ] for sufficiently small @xmath479 define @xmath480 by @xmath481 this makes @xmath482 a subset of the annulus @xmath483 ; it is exactly the subset , where points less than @xmath484 in distance from the boundary have been removed , or in other words @xmath485 for this to hold , we of course need to take @xmath479 sufficiently small , _ e.g._such that @xmath486 holds for some @xmath487 .
let @xmath488 satisfy @xmath489 , @xmath490 , and @xmath491 , and define @xmath492 . by convoluting the characteristic function on @xmath482 with @xmath493
we obtain a smooth function living on the annulus @xmath494 .
so let @xmath495 be defined by @xmath496 and note that @xmath497 since @xmath498 .
normalizing the function @xmath499 in a proper way will give us the function @xmath152 we are looking for .
we will normalize @xmath499 by the function @xmath500 : @xmath501 for a fixed @xmath502 this sum has either @xmath455 or @xmath503 nonzero terms , and @xmath500 is therefore bounded away from @xmath504 and @xmath505 : @xmath506 hence we can define a function @xmath507 by @xmath508 the function @xmath152 will be an almost everywhere @xmath41-dilative partition of unity as is seen by using the @xmath41-dilative periodicity of @xmath500 : @xmath509 since @xmath499 is supported on the annulus @xmath510 , we can simplify the definition in ( [ eq : definition - g - rn ] ) to get rid of the infinite sum in the denominator ; this gives us the following expression @xmath511 we can obtain a more explicit expression for @xmath499 by the following approach .
let @xmath441 and @xmath442 denote the radial coordinates of the surface of the ellipsoids @xmath512 and @xmath444 parametrized by @xmath419 angular coordinates @xmath445 , respectively .
finally , let @xmath513 be defined by @xmath514 where @xmath420 can be found from @xmath350 , and @xmath515 b. han , _ compactly supported tight wavelet frames and orthonormal wavelets of exponential decay with a general dilation matrix _
, j. comput .
appl . math .
* 155 * ( 2003 ) , 4367 .
b. han , q. mo , _ multiwavelet frames from refinable function vectors _ , adv .
comput . math .
* 18 * ( 2003 ) , 211245 . | given a real , expansive dilation matrix we prove that any bandlimited function @xmath0 , for which the dilations of its fourier transform form a partition of unity , generates a wavelet frame for certain translation lattices
. moreover , there exists a dual wavelet frame generated by a finite linear combination of dilations of @xmath1 with explicitly given coefficients .
the result allows a simple construction procedure for pairs of dual wavelet frames whose generators have compact support in the fourier domain and desired time localization .
the construction relies on a technical condition on @xmath1 , and we exhibit a general class of function satisfying this condition . |
the authors owe a debt of gratitude to t. dppner for bringing up the question of the origin of the resonances predicted in average - atom calculations . the work of j.n . and k.t.c .
was performed under the auspices of the u.s.department of energy by lawrence livermore national laboratory under contract de - ac52 - 07na27344 . | recent calculations [ nilsen et al .
arxiv:1212.5972 ] predict that contributions to the scattered photon spectrum from @xmath0 and @xmath1 bound states in chromium ( @xmath2 ) at metallic density and @xmath3 ev resonate below the respective bound - state thresholds .
these resonances are shown to be closely related to continuum lowering , where @xmath4 bound states in the free atom dissolve into a resonant @xmath5 partial wave in the continuum .
the resulting @xmath6-state resonance dominates contributions to the bound - free dynamic structure function , leading to the predicted resonances in the scattered x - ray spectrum .
similar resonant features are shown to occur in all elements in the periodic table between ca and mn ( @xmath7 ) .
52.25.os : emission , absorption and scattering of radiation , 52.38.-r : laser - plasma interaction , 52.70.-m : plasma diagnostic techniques , 56.65.rr : particle in cell method . as discussed in the accompanying article nilsen et al .
@xcite , resonances in bound - free contributions to thomson scattering of x - rays for a cr plasma can complicate the determination of plasma density and temperature from measurements of the scattered x - ray spectrum .
inasmuch as bound - state features in dense plasmas of lighter elements show up as smooth broad features , it is of interest to explore the origin of the resonances that show up in cr . to this end
, we first note that the free cr atom has the configuration [ ar ] @xmath8 . in a dense plasma ,
the bound electrons outside the ar - like ionic core dissolve into the continuum .
in particular , the @xmath4 electrons in cr form a resonant state near 12 ev above threshold for the conditions comsidered here . it is the contribution of this @xmath6-state resonance that leads to resonances in the thomson scattering cross section . in the following paragraphs , we study these resonances in greater detail for cr and other elements between ca and mn ( @xmath9 )
. .properties of elements @xmath10 at @xmath11 ev .
@xmath12 is the atomic weight , @xmath13 ( g / cc ) is plasma density , @xmath14 and @xmath15 are average - atom eigenvalues ( ev ) of @xmath1 and @xmath0 bound states , @xmath16 and @xmath17 are the number of bound and continuum electrons per ion .
@xmath18 is the average ion charge .
@xmath19 , @xmath20 and @xmath21 are the number of continuum electrons in @xmath22 , @xmath23 and @xmath6 states , respectively .
[ tab1 ] [ cols="^,>,>,>,>,>,>",options="header " , ] in table [ tab1 ] , we show results from average - atom calculations of plasmas of elements in row four of the periodic table . in each case
, we assume that the plasma density is the metallic density and that the plasma temperature is @xmath24 ev . in the first four rows of the table
, we list the atomic symbol , atomic number @xmath25 , atomic weight @xmath12 and density @xmath13 ( g / cc ) . in the next two lines ,
we show average - atom eigenvalues @xmath26 and @xmath27 of @xmath0 and @xmath1 states ( ev ) .
the theoretical inner shell ionization thresholds -@xmath28 are 20 - 30% smaller than measured values @xcite .
rows seven and eight list @xmath16 and @xmath17 , the average number of bound and continuum electrons inside the wigner - seitz ( ws ) sphere .
the average - atom model predicts that there are approximately 18 bound electrons and @xmath29 continuum electrons for each element .
the ninth row gives @xmath18 , the average ionic charge or , equivalently , the number of free electrons per ion . for elements listed in the table , @xmath18 , which is near 2 ,
is significantly smaller than @xmath17 owing to the fact that the continuum charge is concentrated inside the ws sphere . in the final three rows of the table , we list @xmath19 , @xmath20 and @xmath21 , contributions to @xmath17 from @xmath22 , @xmath23 and @xmath6 partial waves , respectively .
it should be noted that these three partial - wave contributions account almost completely for @xmath17 .
moreover , for the elements in question , @xmath6-states provide the dominant contribution to the sum . in summary
, the continuum is concentrated inside the ws sphere and is dominated by @xmath6-state partial waves .
-state phase shift @xmath30 and twice the continuum @xmath6-state distribution function @xmath31 are shown as functions of electron energy @xmath32 ( ev ) for 5 ev plasmas at metallic density of elements in row four of the periodic table.[fig1 ] ] in fig .
[ fig1 ] , we plot the @xmath6-state phase shift @xmath30 and the continuum @xmath6-wave distribution inside the ws sphere @xmath31 for elements in table [ tab1 ] .
the distribution function @xmath33 for a continuum state with angular momentum @xmath34 is defined by @xmath35 } } \int_0^{r_{\text{\tiny ws}}}\hspace{-1em } p_{p l}(r)^2 dr , \ ] ] where @xmath36 is the radial wave function for a continuum state with momentum @xmath23 and angular momentum @xmath34 . when integrated over energy , @xmath37 gives the total number of continuum electrons for the @xmath38 partial waves listed in the final three rows of table [ tab1 ] . in all cases , the continuum @xmath6-waves
exhibit resonant behavior near 12 ev , with strong peaks in the distribution functions @xmath33 and corresponding changes in the phase shifts by almost a factor of @xmath39 .
the contribution to the thomson scattering dynamic structure function from bound states @xmath40 is the sum over contributions from individual subshells with quantum numbers @xmath41 : @xmath42 where @xmath43 is the momentum of the continuum electron , @xmath44 is the fractional occupation number of subshell @xmath41 , and @xmath45 and @xmath46 are the momentum and energy transfer from the incident photon @xmath47 to the scattered photon @xmath48 , respectively
. atomic units in which @xmath49 are used here . in eq .
( [ snl ] ) , @xmath50 is an average - atom wave function that approaches a plane wave @xmath51 asymptotically and @xmath52 is the wave function for a bound state with quantum numbers @xmath53 .
the bound - state wave function is expressed in terms of the radial wave function @xmath54 as @xmath55 moreover , the scattering wave function , which consists of a plane wave plus an _ incoming _ spherical wave is expanded as @xmath56 where the radial function @xmath57 is normalized to a phase - shifted sine wave asymptotically : @xmath58 with the above definitions in mind , eq .
( [ snl ] ) can be expressed as @xmath59 where @xmath60 and @xmath61 as shown earlier , the dominant contribution to the continuum wave function @xmath62 is from @xmath6 waves ( @xmath63 ) and occurs near continuum energy @xmath64 ev above threshold .
it follows that the dominant contribution to @xmath65 is from the two terms with @xmath66 and @xmath67 , while the dominant contribution to @xmath68 is from the single term @xmath69 .
energy conservation implies that the resonant contribution to the scattered photon spectrum from bound state @xmath70 occurs at scattered x - ray energy @xmath71 , with @xmath72 ev . for cr at metallic density and @xmath24 ev
is plotted against @xmath73 , where @xmath74 is the scattered photon energy .
the incident photon energy is @xmath75 ev and the scattering angle is 40@xmath76 .
resonance peaks and thresholds associated with transitions from @xmath0 and @xmath1 bound states are shown by the vertical construction lines .
right panel : contributions from bound - free @xmath77 , free - free @xmath78 , and the total structure function @xmath79 , which includes elastic scattering , are plotted against @xmath73 .
[ fig2 ] ] as a specific example , we consider scattering of a 4750 ev photon at 40@xmath76 from cr at metallic density and @xmath24 ev . in the left panel of fig . [ fig2 ] , we show the bound - free contribution @xmath40 to the dynamic structure function .
the vertical construction lines show the @xmath80 and @xmath81 resonance peaks and thresholds .
the resonance peaks are shifted 12 ev below the thresholds of 29.3 and 56.2 ev for the @xmath0 and @xmath1 electrons , respectively , as expected from the position of the peak in @xmath31 shown in the cr panel of fig .
[ fig1 ] .
( short black dashes ) , bound - free contributions including @xmath0 and @xmath1 resonances @xmath40 ( solid blue lines ) and free contributions @xmath82 ( long red dashes ) for scattering from elements listed in table [ tab1 ] are plotted as functions of @xmath73 , where @xmath74 is the scattered x - ray energy and @xmath75 ev is the incident x - ray energy .
the scattering angle for the illustrated cases is 130@xmath76 .
[ fig3 ] ] in fig .
[ fig3 ] , we present results of calculations of the inelastic bound - free contribution to the dynamic structure function @xmath40 , the inelastic free - free contribution @xmath82 and the total dynamic structure function including elastic scattering @xmath83 for elements listed in table [ tab1 ] .
the incident x - ray energy in these examples is @xmath84 ev and the scattering angle is 130@xmath76 .
the relative amplitude of the resonant peaks is seen to increase systematically throughout the sequence as expected from the @xmath6-electron distributions shown in fig .
[ fig1 ] .
by contrast , the relative amplitude of the free - free contribution remains relatively constant from element to element as a consequence of the fact that @xmath18 , the number of free electrons per ion , is relatively constant throughout the sequence . in conclusion , the resonances predicted in the bound - free contribution to the thomson scattering x - ray spectrum arise from a @xmath6-state resonance in the continuum associated with continuum lowering in a dense plasma .
for the examples considered herein , elements between ca and mn at metallic density and @xmath24 ev , the @xmath6-state resonances occur at continuum - electron energy @xmath72 ev . the amplitude of the @xmath6-state resonances increase systematically from element to element .
the @xmath6-state resonance dominate the continuum and the bound - free contribution to the dynamic - structure function @xmath40 .
the resonance associated with transitions from subshell @xmath70 occurs at scattered x - ray energy @xmath85 , where the resonance energy for the cases considered here is @xmath86 ev . |
gamma ray bursts ( grb ) are sudden and powerful gamma - ray flashes , occurring at a rate of @xmath1 1 per day in the universe .
the duration of grbs at mev energies ranges from @xmath2 sec to about @xmath3 sec , with long bursts being characterized by a duration @xmath4 seconds . in the past years
, it has been established that some long grbs are associated to supernovae ( sne ) originating from the death of massive stars .
the grbs have been associated with powerful supernovae ib / c having energies in excess of the majority of such sne and for this reason they have been called hypernovae(iwamoto et al . 1998 ; see also paczy@xmath5ski 1998 ) . in particular
, most of the evidence points towards type ic sne ( see woosley & bloom 2006 , hjorth & bloom 2011 ) .
the collapsarmodel proposed to explain long grbs takes into account this phenomenological aspect and proposes a wolf - rayet progenitor which undergoes core collapse , thus producing a rapidly rotating black hole surrounded by an accretion disk which injects energy into the system and thus acts as a central engine(woosley 1993 , macfayden & woosley 1999 , zhang , woosley & macfayden 2003 ) .
however , the collapsar can originate also in massive stars in binary systems , as suggested by several authors ( e.g. baron , 1992 ; kobulnicky & fryer , 2007 ; yoon et al .
detailed galaxy evolution models are able to predict the temporal behaviour of sn rates in galaxies of all morphological types .
therefore , a comparison between theoretical sn ib / c rates and observed grb rates seems appropriate at the present time .
bissaldi et al .
( 2007 ) already attempted such a comparison but the data relative to grbs were much less than at the present time and limited to lower redshifts . at the present time , grbs have been observed up to @xmath6 ( salvaterra et al .
2009 ; tanvir et al .
2009 ) . in this paper
we aim at studying the behaviour of the sn ib / c rate in galaxies and as a function of redshift and compare it with the most recently derived cosmic grb rate . in computing the type
ib / c sn rate in galaxies we will adopt both single wolf - rayet and massive binaries as grb progenitors and we will consider a dependence of the sn ib / c progenitors on the initial stellar metallicity , not considered in any previous similar work ( e.g. bissaldi et al .
our intention is to make predictions for the rates of type ib / c supernovae at various redshift in spiral and irregular galaxies . at the moment
, no observations can constraint such predictions , but in the future new powerful observational devices as the james webb telescope or elt will provide extensive and constraining observational constraints on those rates .
the comparisons with the present predictions will then allow to confirm or reject the present predictions and will bring new clues on the nature of the type ib / c supernovae progenitors and on the star formation histories in spiral and irregular galaxies . from the comparison between observed grb cosmic rate and predicted cosmic sn ib / c rate , we aim first to check whether the present ratio of grb to that of sn ib / c can be well reproduced by our models , second to see how this ratio may change with redshift in the frame of our model . in section 2
we will describe the chemical evolution model adopted to compute the evolution of galaxies of different morphological type , as well as the computation of the sn ib / c rate . in section 3
the computed sn ib / c rates for irregular and spiral galaxies of different masses are presented and compared to the observed rates . in section 4
we assemble different cosmic histories of star formation , including the one computed by means of our galaxy models , to compute the cosmic type i b / c sn rate , then we compare this cosmic sn rate with the cosmic grb rate . finally , in section 5 a discussion and some conclusions are presented .
in order to compute the type ib / c sn rate in galaxies we need to know the galaxy star formation history .
galaxies of different morphological type ( ellipticals , spirals , irregulars ) are characterized by different star formation histories ( see matteucci 2001 ) . in particular , ellipticals should have suffered an intense and short star formation episode , whereas spirals and irregulars should have had milder star formation rates ( sfr ) and are still forming stars now .
irregular galaxies must have suffered the mildest sfr since they contain more gas than galaxies of other types . here , we focuse on irregular and spiral galaxies for the following reasons : i ) because the sne ib / c are observed only in star forming galaxies and ii ) because the observations on the hosts of grbs have revealed that long grbs are associated with faint , blue and often irregular galaxies ( conselice et al .
2005 ; fruchter et al . 2006 ; tanvir & levan 2007 ; wainwright et al . 2007 ; li 2008 ) and tend to occur in galaxies with low metallicities ( fynbo & al .
2003 , 2006a ; prochaska et al . 2004
; soderberg et al .
2004 ; gorosabel et al . 2005 ; berger et al . 2006 ; savaglio 2006 ; stanek et al . 2006
; wolf & podsiadlowsky 2007 ; modjaz et al .
2008 ; savaglio et al .
2009 ) .
however , we do not exclude that this could be a selection effect ( see e.g. mannucci et al .
2011 ; campisi et al .
in particular , in the paper of mannucci et al .
( 2011 ) , the authors suggest that probably the region high-@xmath7/high mass is populated by the dark grbs .
this idea is partially confirmed by the observations of some dark grb host galaxies provided by kruhler et al .
( 2011 ) , where the mass of the dark grb hosts seem to be higher than the mass of the normal grb hosts . to do that we use a detailed self - consistent chemical evolution model reproducing the majority of the properties of irregular and spiral galaxies .
the irregular galaxies play an important role in the study of chemical evolution and star formation owing to their simpler structure and lack of evolution compared to spiral galaxies .
we also adopt a relatively simple model for spirals without taking into account gradients along the disk .
we assume that both irregular and spiral galaxies assemble all of their mass by means of a continuous infall of pristine gas .
this is certainly true for spiral disks such as that of the milky way ( see e.g. chiappini et al . 1997 ; 2001 ,
boissier & prantzos 1999 ) .
the basic equations are : @xmath8 where @xmath9 is the gas mass in the form of an element i normalized to the present time total luminous infall mass .
the quantity @xmath10 represents the abundance by mass of an element i and by definition the summation over all the elements present in the gas mixture is equal to unity .
the quantity @xmath11 is the total fractional mass of gas .
the quantity , @xmath12 represents the rate at which the element @xmath13 is restored into the ism by the dying stars .
finally , @xmath14 and @xmath15 represent the infall and wind rate , respectively .
the sfr ( @xmath16 ) , namely the amount of interstellar gas , expressed in solar masses , turning into stars per unit time , is assumed to be continuous and defined as a schmidt ( 1959 ) law : @xmath17 where the quantity @xmath18 is the star formation efficiency ( sfe ) , namely the inverse of the typical time - scale for star formation , and is expressed in @xmath19 .
the sfe for spirals is assumed to be higher than in irregulars .
we explored one initial mass functions ( imf ) , in particular the salpeter ( 1955 ) one ( x=1.35 in the mass range 0.1 - 100 @xmath20 ) .
a main assumption of the model for irregulars is the existence of galactic winds triggered by sn explosions . in particular
, it is assumed that a fraction ( @xmath21 ) of the initial blast wave energy of sne is transformed into thermal gas energy and the wind starts when the thermal energy of gas equates the binding energy of gas the wind can be normal .
namely each element is lost at the same rate , or differential , in the sense that some elements ( metals for example ) are preferentially lost ( see bradamante et al .
the galactic wind is likely to occur in these systems because of their relatively low potential well .
moreover , galactic outflows are observed in irregular galaxies ( see e.g. martin , 1999 ; martin et al .
2002 ) . to do
that we follow the method described in bradamante et al .
( 1998 ) and yin et al.(2010 ) ; in particular , the wind rate is assumed to be proportional to the sfr through a free parameter @xmath22 larger than zero : @xmath23 where @xmath13 represents a specific chemical element . to have a preferential loss of metals , as indicated by dynamical models ( e.g. mc low & ferrara , 1999 ) , we use a differential wind in which @xmath24 and @xmath25 for the other elements . in the case of spirals , the galactic wind is less likely to occur , owing the deep potential well in which the spiral disks lie .
in fact , in spiral disks it is more likely to have galactic fountains rather than galactic winds ( spitoni et al .
2009 ) . finally , the assumed rate of infall is the same for irregulars and spirals and follows the law : @xmath26 where @xmath27 is a suitable constant , derived by integrating eq .
( 3 ) over the galactic lifetime , @xmath28 is the abundance by mass of the element @xmath13 in the infalling gas , assumed to be primordial and @xmath29 is the infall timescale .
this timescale is expressed in gyrs and defined as the characteristic time at which half of the total mass of the system has assembled .
the values of @xmath29 are derived as the best ones in order to reproduce the majority of the observational constraints .
this timescale is different for galaxies of different morphological type , being quite short in spheroids and increasing for spirals and irregulars . in table 1
we show the adopted model parameters for a typical spiral galaxy ( milky way - like ) and for a typical irregular galaxy .
the infall mass @xmath30 is the mass that eventually would be accreted if no galactic wind would occur .
model irr is described by an infall mass of @xmath31 which is one order of magnitude lower than the infall mass of model sp .
the infall timescale for the spiral galaxy is assumed to be 6 gyr , although in the milky way disk the timescale for disk formation should have been shorter in the internal than in the external regions ( inside - out formation , matteucci & franois 1989 ; chiappini et al .
1997 ) ; here we want only to show some averaged properties and 6 gyr is an average timescale between the internal timescales ( @xmath32 gyr ) and the external ones ( @xmath33 gyr ) .
the timescale for the formation of irregulars is assumed to be smaller ( 4 gyr ) .
always in table 1 we report the wind parameter for irregular galaxies ( @xmath34 , eq . 3 ) and
the sfe for irregulars and spirals . by following the work of calura et al .
( 2009 ) we assumed that the sfe increases with galactic mass .
.parameter sets used for describing our models:@xmath30 is the final total assembled mass if nothing is lost , @xmath29 is the infall time scale , @xmath34 is the wind parameter ( see eq .
3 ) and @xmath35 is the star formation efficiency . [ cols="^,^,^",options="header " , ] in figure [ figsfrmie ] , the time evolution of the star formation rates ( expressed in @xmath36 ) is plotted for the three different types of galaxies ( elliptical , spiral and irregular ) .
the sfr increases until the energy injected into the ism by stellar winds and sn(ia , ib , and ii ) explosions triggers a galactic wind . at that time , the thermal energy is equivalent to the binding energy of gas , and the gas is lost at a rate proportional to the sfr ( eq.3 ) with a consequent drop of the sfr .
moreover , we can see that the present models well fit the present time averaged sfr both in spirals and in irregulars . as functions of time and redshift ; the redshift of galaxy formation is @xmath37 in a @xmath38cdm cosmology .
the infall masses for each type of galaxies are : @xmath39 ( elliptical , dashed line ) @xmath40 ( spiral , dotted line ) and @xmath41 ( irregular , solid line ) . in the figure are shown also some average values for the present time sfr in spirals ( square , chomiuk & povich , 2011 ) and in irregulars ( circle , harris & zaritsky , 2009).,scaledwidth=50.0% ] figure [ figmzrel ] shows the predicted and observed mass - metallicity relation at the present time relative to small mass galaxies . in particular , we show both the best fit of maiolino et al .
( 2008 ) of the data provided by kewley & ellison ( 2008 ) concerning star forming galaxies and the data and best fit of dwarf irregulars , as inferred by lee et al .
( 2006 ) . as one can see , our models lie close to the best fits . in our galaxy models
the mass - metallicity relation arises naturally by adopting a smaller sfe in smaller galaxies .
the above comparisons show that our chemical models very well reproduce many observed properties of the present day spirals and irregulars .
since , these present day properties results from the whole previous evolution , these good correspondence give some confidence that these models can be used to explore the much less well known early phases of these galaxies , which will hopefully be observed in greater details in a near future thanks to more powerful observational devices as the james webb telescope and the elt .
and sfe=0.02 ( open green circle).,scaledwidth=50.0% ] in figure [ evozmwr ] we show the evolution of the metallicity @xmath42 for our galaxies and the corresponding evolution of the minimum wr mass , according to the relations of table 2 and 3 .
as one can see , owing to the milder increase of metallicity in the irregular galaxy , the @xmath43 varies more gradually as a function of time in this galaxy than in the spiral one . on the other hand ,
the variation of @xmath43 in an elliptical is very fast in the very early phases . in figure
[ snr ] , the sn ib / c , ic rates are shown as a function of time for model irr and model sp .
different rates arising under different assumptions concerning the snib / c and snic progenitors and their dependence on the initial stellar metallicity ( see sect .
the percentage of sne from binary systems represents @xmath1 30% of the total predicted rate .
the points in figure [ snr ] are the observed sn ib / c rates for an irregular ( circle ) and a spiral ( square ) galaxy provided by li et al .
( 2011 ) .
it is worth noting that the samples of sn - host galaxies usually include ( due to an observational bias ) small numbers of irregulars and there is no way to get out of this as long as the sn surveys will not include a large number of such irregular galaxies .
following li et al.(2011 ) , one can conservatively assume that the true value of the sn rate in irregulars is between li et al .
( 2011 ) and mannucci et al ( 2005 ) estimates . in figure [ snr ]
this uncertainty is included in the size of the error bars .
as one can see , the predicted present time rate of sne ib / care consistent , within the error bars , with the observed ones .
an interesting quantity , often shown in literature , is the ratio between the sn ib / c and the sn ii rate , @xmath44 ( i.e. prantzos & boissier 2003 ; boissier & prantzos 2009 ; prieto et al .
2008 ; smartt et al .
2009 and smith et al .
2011 ) . in order to have a more consistent comparison with the data we have plotted the @xmath44 and @xmath45 ratios as functions of metallicity ( see figure [ i / iivsz]).it is worth noting that when using @xmath44 and @xmath45 vs metallicity , we can not specify the galaxy type because the relation between mass and metallicity and the imf are the same for all galaxy types . as one
can see from figure [ i / iivsz ] , our predictions are in good agreement with the data both in the case of sne ib / c ( solid line ) and sne ic ( dashed line ) and with previous calculations ( boissier & prantzos , 2009 ) . in a @xmath38cdm cosmology .
the various model predictions for each rate depend on the different assumptions concerning the snib / c and snic progenitors and their dependence on the initial stellar metallicity : @xmath46 ( solid line for model irr and dotted - double dashed line for model sp ) , @xmath47 ( dashed line for snibc and dotted line for snic in model sp ; dashed- dotted line for snibc and double dotted - dashed line for snic in model irr ) .
the points are the observed sn ib / c rates , obtained by multiplying the observed rate per unit mass ( li et al .
2011 ) by the present time stellar mass of the galaxy in mod .
irr ( circle ) and mod .
sp ( square ) . in the figure
is shown also the local grb rate provided by guetta et al.(2005 ) , triangle , and salvaterra et al .
( 2011 ) , diamond.,scaledwidth=51.0% ]
figure [ snr ] shows the existence of a good match between the expectedsnib / c localrate , computed for spirals and irregulars , and the observedsn rates derived by li et al .
this agreement assesses quantitatively the reliability of the prescriptions that have been used in sections 2 and 3 to derive the snib / c rate as a function of time and provides more weight to the results that will be presented in section 5 .
when the metallicity effect on the minimum wr mass is taken into acount , there is a difference only for irregulars , mainly at early times , due to the slower growth of metallicity in these systems .
it is interesting to compute the ratio of type ib / c sne to grbs ( in the local universe ) and to do that we should compare the theoretical snib / c rates , shown in figure [ snr ] , with the local rate of grbs at the present time ( triangle ) .
the canonicalvalue for the latter quantity ranges between @xmath1 0.5 grb @xmath49 ( schmidt 2001 ) to @xmath1 1 event @xmath50 ( guetta et al .
2005 ) . in order to compare the snib / c and the grb rate in the right units we use the local grb rate of 1.1 event @xmath50 ( guetta et al . 2005 ) taking into account the local density of b luminosity , @xmath51 ( e.g. madau , della valle & panagia 1998 ) and the b luminosity of the milky way , @xmath52 .
this approach gives @xmath53 .
this observedrate has to be re - scaled by using the beaming factor , @xmath54 .
the beaming factor accounts for the fact that a grb does not light up the full celestial sphere but rather a fraction .
there are several estimates of this parameter , @xmath55 ( guetta & della valle , 2007 ) for local and low luminosity grbs , corresponding to @xmath56 , and @xmath57 ( guetta et al .
2005 , yonetoku et al . 2005 ; van putten & regimbau 2003 , frail et al .
2001 ) , for high - luminosity grbs , corresponding to beaming angles of @xmath58 .
if we conservatively assume , for localgrbs , @xmath59 , we derive a localratio grb / snib / c of @xmath60 and @xmath61 in spirals and irregulars , respectively .
this is an expected result , since not all sne ib / c will end up as long grbs . in figure [ snr ]
we have shown the grb rate provided by guetta et al .
( 2005 ) in unit of @xmath62 and re - scaled using the beaming factor @xmath63 .
the cosmic sn and grb rates are defined in an unitary comoving volume of the universe .
this definition is necessary to study the rates at high redshift where the morphology of the observed galaxies is not known .
the cosmic rates refer , in fact , to a mixture of galaxies which can be different at every redshift . both the cosmic type ib / c and grb rates depend upon the sfr in galaxies but also on the galaxy and grb luminosity functions .
if these functions do not evolve with redshift then both the sn and grb rates will trace the csfr . on the contrary
, the observed behaviour can be due to the evolution of the luminosity functions of galaxies ( e.g. number density evolution ) .
the csfr has now been measured up to very high redshift ( z@xmath18 ) , especially thanks to galaxies hosting grbs ( see kistler et al .
2009 ) . in figure
[ csfr ] , we show a revised version of the csfr predicted by calura & matteucci ( 2003 ) and obtained by taking into account the evolution of galaxies of different morphological type ( ellipticals , spirals and irregulars ) , as described in the previous sections . in particular , the csfr has been computed by assuming a pure luminosity evolution of galaxies , in other words , the main parameters of the schechter ( 1976 ) galaxy luminosity function have been kept constant with redshift . to compute the csfr we have adopted the following relation : @xmath65,\ ] ] where @xmath66 identifies a particular galaxy type ( elliptical , spiral , irregular ) and @xmath67 represents the history of star formation in each galaxy , as shown in figure [ figsfrmie ] .
the quantity @xmath68 is the galaxy number density , expressed in units of @xmath69 for each morphological galaxy type and it has been assumed to be constant and equal to the present time one , as derived by marzke et al .
this csfr is therefore obtained by assuming that all galaxies started forming stars at the same time and that there is no number density evolution ; these are very simple assumptions but they can be useful to disentangle the effect of the sfr from that of the luminosity function and to compare these results with predictions from hierarchical galaxy formation models .
this predicted csfr shows a high peak of star formation at very high redshift due to the contribution of the ellipticals which have formed their stars very early .
this predicted high redshift csfr is probably too high and unrealistic since the number density of ellipticals at high redshift could have been overestimated .
in other words , the hypothesis of no number density evolution could be incorrect .
on the other hand , the predicted csfr for redshift @xmath70 seems underestimated relative to the data .
this does not mean that our galactic sf histories are wrong but again it could be due to neglecting the number density evolution of galaxies .
we have adopted also csfrs computed in the framework of the hierarchical clustering galaxy formation scenario , as well as the fit to the observed csfr . in fact , in figure [ csfr ] is also shown the csfr obtained by cole et al .
( 2001 ) best fitting the data collected by hopkins ( 2004 ) from 1995 onwards .
this same parametric form has been used later on by many authors such as hopkins & beacom ( 2006 ) and blanc & greggio ( 2008 ) , since it fits also more recent data up to redshift z=6 .
it is worth noting that also all the other theoretical csfrs shown in figure [ csfr ] are underestimating the csfr at intermediate and low redshifts . to derive observationally
the csfr one should adopt some of the well known tracers of sf , in particular @xmath71 , @xmath72 , uv continuum . in these wavebands the effect of dust can not be neglected and therefore the dust correction is necessary to obtain the correct csfr .
the differences between corrected and uncorrected data are generally large , as shown by strolger et al .
in particular , uncorrected data tend to show a strong decline of the csfr for @xmath73 , whereas the corrected data show an almost constant csfr for @xmath74 .
another important effect in the derivation of the csfr is related to the uncertainty in the faint end of the luminosity function of galaxies . by means of these different csfrs
we have then computed the cosmic sn ib / c rate shown in figures [ csnr_grb ] where it is reported also the observed cosmic grb rate .
the adopted progenitors for sne ib / c are assumed to be single wr stars with constant minimum mass of @xmath75 plus binary systems , as described in section 2.1 .
as one can see , the theoretical error in the csnr increases towards high and very high redshift and it is roughly a factor of ten at z=6 .
clearly at these high redshifts ( @xmath76 ) the uncertainties are still too large too draw any conclusion . in figure [ csnr_z ]
we show the predicted cosmic type ib / c sn rate obtained by adopting the cole et al .
( 2001 ) csfr and @xmath43 depending on z , all the other assumptions being the same .
here we have considered the cosmic evolution of z , and in particular we assumed the z vs. time evolution typical of an elliptical galaxy of @xmath77 as shown in figure [ evozmwr ] ; this is because by weighting the z vs. time of each galaxy on their number density , the z vs.time relation of the ellipticals dominates at all redshifts .
in fact , spheroids are very likely to be the responsible for the production of the bulk of metals in the universe ( see calura & matteucci 2004 ) . as one can see in figure [ csnr_z ] , the effect of metallicity on the type ib / c sn progenitors is stronger at early times and produces a lower cosmic type ib / c sn rate .
this effect would be similar if applied to all the csfrs of figure [ csfr ] coming back to figure [ csnr_grb ] , a visual inspection of this figure confirms that grbs originating from the explosion of massive stars are only a tiny fraction of sne ib / c class . in particular , the comparison of the snib / c rates with the matsubayashi et al .
( 2005 ) semi - empirical track , suggests the ratio grb / snib / c to be @xmath58 in the local universe and to increase up to @xmath78 all over the redshift range z= 1 @xmath79 8 .
interestingly enough , the grb / sneib / c ratio at z @xmath1 0 nicely reproduces the observedratio between the local grb rate and the snib / c rates obtained by other authors : @xmath80 ( guetta et al .
2005 ) and @xmath81 ( guetta & della valle 2007 ) , respectively . after taking these figures at their face values ,
we conclude that local and low luminositygrbs ( @xmath82 ) barely need the correction for beaming and therefore we can infer that they emit almost isotropically .
this result is in good agreement with observations ( admittedly on scanty statistic ) .
for example , for grb 060218 soderberg et al .
( 2006 ) find @xmath83 , for grb 031203 , @xmath84 ( malesani , private communication ) and @xmath85 ( corresponding to @xmath86 ) for low luminositygrbs population ( guetta & della valle , 2007 ) . the increasing ratio grb / snib / c @xmath87 for @xmath88 , in the case of the matsubayashi et al .
( 2005 ) grb rate , may suggest either a different behavior for cosmological grbs due to the existence of a different grb population ( e.g. bromberg et al .
2011 and references therein ) characterized by a larger beaming factor , likely of the order of @xmath89 ( corresponding to a jet opening angle of @xmath1 20deg-6deg ) or the matsubayashi et al .
semi - empirical track is still affected by the obvious bias which favors the discovery at high-@xmath7 of only highly beamed grbs .
similar conclusions ( short of a constant ) can be obtained by comparing the snib / c rate track with the grb track by wanderman & piran ( 2010 ) and with that by salvaterra et al .
in particular , the salvaterra et al .
( 2011 ) rate is derived from a redshift complete sample of bright _ swift _ grbs under the assumption that grbs did not experience luminosity evolution with redshift .
it is worth noting that on the basis of the current studies , it is not possible to distinguish between a pure density and a pure luminosity evolution . in general
, there is no agreement among different authors on this issue .
butler et al .
( 2010 ) suggest that pure density evolution models produce the observed number of grbs at high - z , but in other works the luminosity evolution is used to explain the grb rate ( i.e. see salvaterra et al . 2009b , petrosian et al .
2009 ) increasing faster than some csfr , such as that of hopkins & beacom ( 2006 ) .
this is because the hopkins & beacom ( 2006 ) csfr is decreasing at high z. however , this behaviour of the csfr needs to be confirmed by more data and at the moment we can not exclude the csfr to be flat at high z. : solid line refers to a constant @xmath43 whereas dotted line refers to the case of @xmath47.,scaledwidth=50.0% ]
in this paper we have computed the type ib / c sn rates expected at the present time in irregular and spiral galaxies of different masses with the aim of predicting the variation with redshift of the snib / c rate based on successful models for the chemical evolution of irregulars and spirals .
we considered both single wr stars and massive stars in binary systems as sn ib / c progenitors .
we used stellar evolution results indicating that the minimum mass of wr stars is a function of the stellar metallicity thus suggesting a higher rate of sne ib / c in more metal rich galaxies .
then we considered various csfrs as functions of cosmic time , both theoretically and observationally derived , and computed the cosmic type ib / c sn rates expected from the assumptions on type ib / c sn progenitors .
these cosmic type ib / c sn rates were then compared to the observationally derived cosmic grb rate .
our main conclusions can be summarized as follows : * by taking into account wr progenitors depending on the metallicity and a fraction of massive close binary systems equal to 15% of all massive stars as sn ib / c progenitors , it is possible to reproduce the present observed type ib / c sn rate both in dwarf metal poor irregular and in spiral galaxies .
it is worth noting that the galactic evolution models adopted here are well reproducing the main chemical properties of these galaxies . *
if a dependence on stellar metallicity is assumed for the wr stars , differences arise in the type ib / c sn rates only at early evolutionary times in galaxies .
negligible differences are produced on the predicted local rates .
* we have compared the local observed type ib / c rates in spirals and irregulars with the local grb rate and derived a local ratio grb / sne ibc of @xmath90 .
as expected , only a fraction of these sne gives rise to grbs .
* we took various csfr histories and computed the cosmic type ib / c sn rates .
also in this case we considered both a constant minimum wr mass and a mass varying with metallicity .
the effect of the dependence of @xmath43 on the metallicity is to predict lower cosmic type ib / c sn rates at very high redshift .
we have then compared the cosmic type ib / c sn rates with the cosmic grb rate derived from _ swift _
data and found that the ratio grb / sne ibc @xmath91 .
this confirms previous results that only a small fraction of all sne ib / c gives rise to grbs , but our factor is smaller than what found in bissaldi et al .
the reason for this resides in the fact that we have adopted the recent grb cosmic rate derived from swift data , whereas in bissaldi et al .
( 2007 ) the cosmic grb rate was derived on the basis of semi - empirical estimates ( matsubayashi et al . 2005 ) . on the other hand , if we compare our cosmic sn rates with the matsubayashi et al .
( 2005 ) rate we confirm the results of bissaldi et al .
( 2007 ) , indicating a ratio grb / snib / c rates of @xmath92 . *
studies of grbs and their hosts have revealed to be extremely important to trace galaxy evolution at very high redshifts , although the interpretation of cosmic diagrams is difficult since it involves assumptions on the luminosity function of both galaxies and grbs .
it is interesting to note that salvaterra & chincarini ( 2007 ) pointed out that by adopting the csfr derived by cole et al .
( 2001 ) and assuming a grb luminosity function independent of redshift , one largely underestimates the number of high redshift grbs detected by swift .
this fact could be interpreted in two ways : either the characteristic luminosity of grbs increases with redshift or the csfr at very high redshift is higher than in cole et al .
we have shown that a high csfr can be achieved by means of monolithic like models of ellipticals producing stars at a very high rate and at very high redshift .
however , no firm conclusions can be drawn on the csfr at very high redshift because of the large uncertainties due to dust corrections .
v.g . thanks anah granada , cyril georgy and luca vincoletto for many useful discussions . | supernovae ib / c are likely to be associated to long grbs , therefore it is important to compare the sn rate in galaxies with the grb rate . to do that we computed type ib / c sn rates in galaxies of different morphological type ( ellipticals , spirals and irregulars ) by assuming different histories of star formation and different supernova ib / c progenitors .
we included some recent suggestions about the dependence of the minimum mass of single wolf - rayet ( wr ) stars upon the stellar metallicity and therefore upon galactic chemical evolution .
we adopted several cosmic star formation rates ( i.e. relative to a comoving unitary volume of the universe ) as functions of cosmic time , either observationally or theoretically derived , including the one computed with our galaxy models .
then we computed the cosmic type ib / c sn rates .
our results show that the predicted type ib / c sn rates in spirals and irregulars can well reproduce the present time observed rates if both single wr stars and massive binary systems are taken into account as type ib / c sn progenitors .
the metallicity effects on the minimum mass for single wr stars are evident mainly in the early phases of galaxy evolution and do not influence substantially the predicted local type ib / c rates .
we derived the following conclusions : i ) the ratio cosmic grb - type ib / c rate varies in the range @xmath0 in the whole redshift range , thus suggesting that only a small fraction of all the type ib / c sne gives rise to grbs .
ii ) the metallicity dependence of type ib / c sn progenitors produces lower cosmic sn ib / c rates at early times , for any chosen cosmic star formation rate .
iii ) different theoretical cosmic star formation rates , computed under different scenarios of galaxy formation , produce sn ib / c cosmic rates which differ mainly at very high redshift .
however , it is difficult to draw firm conclusions on the high redshift trend because of the large uncertainties in the data .
iv ) grbs can be important tracers of star formation at high redshift if their luminosity function does not vary with redshift and they can help in discriminating among different galaxy formation models .
[ firstpage ] supernovae gamma - ray bursts galaxy evolution . |
the quantum - mechanical description of ground state properties of interacting many - body systems is a long - standing challenge in theoretical physics . in general the problem
is two - fold : the first task is to specify the interaction between the constituents in terms of a two - body potential .
the second is to solve the quantum - mechanical many - body problem with that interaction . for many - body systems composed of helium atoms the first task ,
the construction of an appropriate interatomic potential , is solved rather precisely .
the hfdhe2 potential of aziz _ et al . _
@xcite is able to reproduce experimental transport coefficients in a wide temperature range . a simpler and less precise description of the interaction is given by the lennard - jones potential @xcite @xmath1\ ] ] with @xmath2 and @xmath3 .
since the emphasis of the investigations presented in this paper is more on novel concepts , which remarkably simplify the treatment of the many - body problem , we will restrict ourselves to this simple and widely used parameterization . a common property of many realistic microscopic interactions like van der waals type potentials between neutral atoms or the nucleon - nucleon interaction in atomic nuclei
is a strong short - range repulsion , which induces strong short - range correlations in the quantum many - body state .
the repulsive core prevents the particles to approach each other closer than the radius of this core such that the two - body density matrix is strongly depleted for particle distances smaller than the core radius .
this effect can not be described in a mean - field like picture or in terms of a superposition of a finite number of symmetrized product states . in this work
we treat these correlations with the unitary correlation operator method ( ucom ) , that has already been applied successfully in the framework of the nuclear many - body problem @xcite . in the following section we introduce the general formalism of the unitary correlation operator method and
derive analytic expressions for correlated wave functions and the correlated hamiltonian . in the third section
we determine the optimal correlation operator for liquid @xmath0he from the two - body system and discuss the structure of the effective interaction generated by the unitary transformation . in section iv we investigate the homogeneous @xmath0he liquid with that correlator and introduce a density - dependent two - body correlator that describes very well the influence of three- and more - body correlations .
finally , we successfully apply the density - dependent correlator to describe the ground state of small @xmath0he droplets . considering the great numerical simplification
only a few one - dimensional integrals have to be calculated compared to monte - carlo - type calculations @xcite , the agreement of the results is excellent .
the goal of the unitary correlation operator method ( ucom ) is to describe short - range interaction - induced correlations by a unitary transformation of an uncorrelated many - body state with @xmath4 particles .
the unitary operator @xmath5 associated with this transformation can be written as an exponential of a hermitian generator @xmath6 @xmath7 .\ ] ] the generator is chosen to be an irreducible two - body operator @xmath8 which describes genuine two - body correlations only . in principle
we could add a three - body part to account for genuine three - body correlations .
for the moment we omit this step to develop the basic formalism on a simple level . the particular structure of the unitary transformation which should reflect the properties of the short - range correlations
is described by the two - body generator @xmath9 .
the repulsive core of the interaction leads to a suppression of the probability density within the range @xmath10 of the core .
uncorrelated many - body states that are symmetrized products of single - particle states lead , however , to a two - body density that does not show any depletion for relative distances @xmath11 between the pair of particles .
therefore the unitary transformation should shift those particles that are in the forbidden region of the repulsion to larger distances .
this is done by means of a distance - dependent radial shift in the relative coordinate of two - particles .
the hermitian generator of this shift in two - body space reads @xmath12 , \ ] ] where @xmath13 is the operator of the relative momentum , @xmath14 is the relative coordinate , and @xmath15 is the associated unit vector . the function @xmath16 describes the size of the shift as function of the distance of the particles . first we want to apply the unitary correlation operator @xmath5 to an uncorrelated many - body state @xmath17 out of a low - momentum model space .
the correlated many - body state @xmath18 is defined by @xmath19 here and in the following all correlated quantities are marked by a tilde . in order to demonstrate the generic effect of the correlation operator we discuss first a two - body system with an uncorrelated two - body state @xmath20 .
using the definition of the correlation operator @xmath5 and the generator @xmath9 we can evaluate the correlated two - body wave function in coordinate representation explicitly @xcite @xmath21 the respective transformation with the hermitian adjoint correlation operator @xmath22 leads to @xmath23 the correlation operator acts in the relative coordinate @xmath24 of the two - body system only , the center of mass coordinate @xmath25 is invariant . according to the construction of the the generator the unitary transformation corresponds to a radial shift in the relative two - body coordinate @xmath26 the transformed relative distances are given by the correlation functions @xmath27 and @xmath28 , respectively . due the unitarity of the correlation operator @xmath29 the correlation functions @xmath27 and @xmath28 are inverse to each other @xmath30 = r .\ ] ] the conservation of the norm of the correlated two - body wave function is guaranteed by the metric factor @xmath31 in and @xmath32 where @xmath33 denotes the derivative of @xmath34 .
it corresponds to the square root of the jacobi determinant associated with the transformation of the normalization integral .
the correlation functions @xmath34 that describe the coordinate transformation are connected to the function @xmath16 , which enters in the generator , by the integral equation @xmath35 for small shift distances the approximate relation @xmath36 holds . since all correlated quantities can be expressed directly in terms of the correlation functions @xmath27 and @xmath28 we will use one of these as basic quantity to describe the detailed structure of the short - range correlations .
the shift function @xmath16 is only used for formal considerations and never specified explicitly .
complementary to the definition of a correlated state we can use the correlation operator to define correlated operators . for an arbitrary observable @xmath37
the correlated operator @xmath38 is given by the similarity transformation @xmath39 it is evident that the formulations in terms of correlated states and correlated operators are equivalent when expectation values or matrix elements are calculated @xmath40 thus we can choose the formulation that is technically or intuitively better suited for the specific application under consideration . in most cases the formulation in terms of correlated operators is easier to apply in the many - body system .
moreover it provides a systematic approximation scheme that will be discussed in section [ sec : ucom_clusterexp ] . as for the correlated wave function
we first discuss specific correlated operators for a two - body system .
the generalization to many - body systems will be done in the following sections .
the simplest operator of interest is the local two - body potential @xmath41 that depends on the relative distance @xmath42 only . to evaluate the correlated potential we use the results on the correlated two - body wave function to determine a general two - body matrix element of the form @xmath43 & \quad= \int\!\!{\ensuremath{\mathrm{d}}}^3r\,{\ensuremath{\mathrm{d}}}^3\!x\;{\ensuremath { \big < { \psi } \big| \,{{\ensuremath{{\ensuremath{{{{\ensuremath{\bm{c}}}}}^{\dag}}}}}}\ , \big| { { \ensuremath{\vec{x}}},{\ensuremath{\vec{r } } } } \big > } } \,v(r)\ , { \ensuremath { \big < { { \ensuremath{\vec{x}}},{\ensuremath{\vec{r } } } } \big| \,{{\ensuremath{{\ensuremath{\bm{c}}}}}}\ , \big| { \psi ' } \big > } } \\[-2pt ] & \quad= \int\!\!{\ensuremath{\mathrm{d}}}^3r\,{\ensuremath{\mathrm{d}}}^3\!x\;{\ensuremath { \big < { \psi } \big| { { \ensuremath{\vec{x}}},{\ensuremath{\vec{r } } } } \big > } } \;v[{\ensuremath{r_+}}(r)]\ ; { \ensuremath { \big < { { \ensuremath{\vec{x}}},{\ensuremath{\vec{r } } } } \big| { \psi ' } \big > } } \\[2pt ] & \quad= { \ensuremath { \big < { \psi } \big| \,{v[{\ensuremath{r_+}}({\ensuremath{{\ensuremath{\bm{r}}}}})]}\ , \big| { \psi ' } \big > } } . \end{split}\ ] ] in going from the second to the third line we use equations and and substitute the integration variable @xmath24 by @xmath44 .
since the two - body states are arbitrary this relation is valid on the operator level @xmath45 .\ ] ] thus the correlated two - body potential @xmath46 is given by the original potential with transformed radial coordinate @xmath47 .
the correlated operator of the kinetic energy in the two - body system is slightly more complicated due to its momentum dependence .
first we decompose the operator of the two - body kinetic energy @xmath48 in a center of mass contribution @xmath49 and an relative part @xmath50 @xmath51 the operator of the center of mass momentum @xmath52 commutes with the correlation operator , i.e. , the operator of the center of mass kinetic energy is invariant under the unitary transformation .
the correlation operator acts on the relative part @xmath50 of the kinetic energy only .
a similar calculation as for the local potential leads to the following structure of the correlated relative kinetic energy @xcite @xmath53 besides @xmath50 an additional momentum dependent term appears in the correlated relative kinetic energy , which can be formulated in terms of a tensorial effective mass correction .
it is conveniently split into a correction for the radial part @xmath54 and a different correction for the angular component @xmath55 @xmath56 where @xmath57 is the operator of the relative two - body angular momentum .
the distance dependent effective mass corrections @xmath58 and @xmath59 depend on the correlation function @xmath28 only @xmath60 in addition to these momentum dependent terms a local contribution @xmath61 appears in @xmath62 .
\end{split}\ ] ] this additional two - body potential is also determined by the correlation function @xmath28 and its derivatives . in the preceding section we evaluated correlated operators explicitly in the two - body system
. however , correlated operators in the many - body system contain additional terms . the unitary transformation of an operator , e.g. , an one - body operator like the kinetic energy or a two - body operator like the potential , generates a correlated operator , which contains irreducible contributions to all particle numbers @xmath63 } + { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{o}}}}}}}}^{[2 ] } + { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{o}}}}}}}}^{[3 ] } + \dots .
\end{split}\ ] ] we use the notation @xmath64}$ ] for the irreducible @xmath65-body part of the correlated operator @xmath38 .
this decomposition according to irreducible particle number is called cluster expansion @xcite .
the cluster expansion gains physical meaning by the so called cluster decomposition principle @xcite : two localized subsystems , which are separated beyond the range of interactions , are independent of each other .
therefore the state of the total system decouples into a direct product of the states of the two subsystems .
this implies an analogous decomposition property of the correlation operator and of correlated operators . because of the cluster decomposition principle the cluster expansion is a natural starting point for approximations of the correlated operators . for a selected particle there will be a certain number of other particles within the range of the correlation depending on the density .
the number of particles in this cluster gives the maximum order of the cluster expansion that contributes .
we can truncate the cluster expansion at low orders if the range of the correlations is sufficiently small compared to the average distance of the particles . the simplest nontrivial approximation results from the truncation of the cluster expansion beyond two - body order .
the correlated operator in two - body approximation reads @xmath66 } + { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{o}}}}}}}}^{[2 ] } .\ ] ] this approximation requires that the system is sufficiently dilute such that contributions of the three - body order of the cluster expansion are small . or in other words , the probability for three and more particles to be in the range of the repulsive core simultaneously has to be small .
the technical advantage of the two - body approximation is that closed analytic expressions for the correlated operators can be deduced by just considering the two - body system .
the correlated hamiltonian in two - body approximation is of the form @xmath67 } + { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{t}}}}}}}}^{[2 ] } + { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{v}}}}}}}}^{[2 ] } \\ & = \sum_i^n { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{t}}}}}}}}^{[1]}_i + \sum_{i < j}^n { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{t}}}}}}}}^{[2]}_{ij } + \sum_{i < j}^n { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{v}}}}}}}}^{[2]}_{ij } . \end{split}\ ] ] the one - body part of the correlated kinetic energy is simply the uncorrelated kinetic energy operator .
the two - body part consists of the effective mass corrections and the additional local potential as determined in the previous section @xmath68 } & = { \ensuremath{{\ensuremath{\bm{t}}}}}\\ { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{t}}}}}}}}^{[2 ] } & = { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{t}}}}}}}}_{\omega } + { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{t}}}}}}}}_r + { \ensuremath{\tilde{u}}}({\ensuremath{{\ensuremath{\bm{r } } } } } ) .
\end{split}\ ] ] the correlated potential has only a two - body contribution that is given by @xmath69 } = { \ensuremath{\tilde{v}}}({\ensuremath{{\ensuremath{\bm{r } } } } } ) .\ ] ] in order to get a rough measure for the validity of the two - body approximation we define a smallness parameter @xmath70 as a product of the density @xmath71 of the system and a typical volume in which the correlations between a pair of particles change their relative wave function .
the correlation volume @xmath72 is defined by the norm of the defect wave function , i.e. , the difference between the uncorrelated uniform wave function @xmath73 and its correlated companion @xmath74 ^ 2 = \int\!\!{\ensuremath{\mathrm{d}}}^3r\ ; [ { \ensuremath{\mathcal{r}_+}}(r)-1]^2 .\ ] ] the smallness parameter @xmath75 is a measure for the probability to find a third particle within the volume where the correlations between two particles change their relative wave function significantly .
the two - body approximation is valid only if this probability is small such that three - body correlations are negligible .
we have shown for different physical systems @xcite that the relative contribution of the three - body order to the energy exceeds 10% if the smallness parameter reaches a value of typically @xmath76 .
if the two - body approximation is not sufficient one can include higher orders of the cluster expansion successively . however , already the calculation of a three - body correlated wave function starting from the many - body correlation operator in analogy to section [ sec : ucom_wavefct ] is not practicable
. therefore we reverse the procedure and start from a general many - body coordinate transformation in analogy to the correlation function @xmath34 and determine correlated many - body wave functions and operators . in a second step we connect the explicit structure of the coordinate transformation with the many - body correlation operator , at least in an approximate way . consider a @xmath4-body system with a collective coordinate vector @xmath77 .
the short - range correlations are described by a @xmath4-body coordinate transformation @xmath78 the transformation function @xmath79 corresponds to the correlation function @xmath27 in the two - body case .
similar to @xmath28 we define an inverse transformation function @xmath80 by @xmath81 = x .\ ] ] the correlated @xmath4-body wave function @xmath82 is defined via the coordinate transformation @xmath83 with a metric factor @xmath84 that ensures norm conservation .
the metric factor is given by the square root of the jacobi determinant of the transformation @xmath85 with the matrix elements @xmath86_{ij } = { \ensuremath{\frac{\partial{x'_i}}{\partial{x_j } } } } = { \ensuremath{\frac{\partial{[{\ensuremath{\mathcal{x}_-}}(x)]_i}}{\partial{x_j } } } } .\ ] ] formally we can construct a unitary many - body correlation operator @xmath5 that generates this particular coordinate transformation @xmath87 the check of the unitarity relation @xmath88 with this definition is straightforward . using this formulation
we can evaluate many - body correlated operators in the same way as shown in section [ sec : ucom_operator ] for the two - body approximation .
we will only present selected results that are needed for the following investigations .
the full formalism is discussed in @xcite . a general local potential @xmath89 , which depends on the coordinates @xmath77 of all @xmath4 particles , is transformed as in the two - body case @xmath90 , \ ] ] i.e. , the correlated potential is given by the original potential with transformed coordinate dependence .
the expression for the @xmath4-body correlated kinetic energy operator @xmath91 is more involved @xmath92_{jk } \ ; { \ensuremath{{\ensuremath{\bm{p}}}}}_k + { \ensuremath{\tilde{u}}}({\ensuremath{{\ensuremath{\bm{x } } } } } ) , \ ] ] where @xmath93 denotes the collective momentum vectors of all @xmath4 particles . as in the two - body case
the correlated kinetic energy contains an effective mass tensor and an additional local potential .
the effective mass tensor is given by the square of the jacobi matrix @xmath94_{jk } = \frac{1}{2 m } \sum_{i=1}^{3n } [ j_-(x)]_{ji}\ , [ j_-(x)]_{ik } .\ ] ] the additional local potential @xmath95 given by @xmath96_{x ' = { \ensuremath{\mathcal{x}_+}}(x ) } .\ ] ] with these expressions it is in principle possible to calculate expectation values of the correlated hamiltonian up to arbitrary order in the cluster expansion .
in practice all extensions beyond two - body approximation are very costly , thus only the three - body order will be considered
. we will use the general formulation of the many - body coordinate transformation to estimate the contribution of the three - body order of the cluster expansion . since we can not derive the many - body coordinate transformation directly from the correlation operator
we will construct the transformation by generalization of the known one in the two - body system .
it is useful to reformulate the two - body coordinate transformation discussed in section [ sec : ucom_wavefct ] in terms of the general transformation function @xmath97 , where the upper index indicates the number of particles involved .
the transformation of the relative coordinate with the correlation function @xmath34 is equivalent to the following transformation of the single - particle coordinates @xmath98 with a shift vector @xmath99\ , { \ensuremath{\hat{r}}}.\ ] ] this notation reveals the intuitive picture behind the description of short - range correlations by means of a coordinate transformation . as a consequence of the short - range repulsion the particles are displaced along their connecting axis by the shift vector @xmath100 . the two - body transformation can be readily generalized to describe three - body correlations .
the corresponding three - body coordinate transformation @xmath101 is given by @xmath102 thus the coordinate of the first particle is transformed with respect to the second and the third particle simultaneously .
as a necessary prerequisite this transformation obeys the cluster decomposition principle . if one of the three particles is separated beyond the range of the correlations , i.e. , to distances where @xmath100 vanishes , then the transformation reduces to the two - body transformation for the remaining pair .
we will use this three - body transformation to calculate the local three - body contributions of the correlated hamiltonian , i.e. , the correlated potential and the local part of the correlated kinetic energy .
the irreducible three - body part @xmath103}$ ] of the correlated potential is given by @xmath104 } & = { \ensuremath{{\ensuremath{{{{\ensuremath{\bm{c}}}}}^{\dag}}}}}({\ensuremath{{\ensuremath{\bm{v}}}}}_{12 } + { \ensuremath{{\ensuremath{\bm{v}}}}}_{13 } + { \ensuremath{{\ensuremath{\bm{v}}}}}_{23})\ , { \ensuremath{{\ensuremath{\bm{c}}}}}\\ & - ( { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{v}}}}}}}}^{[2]}_{12 } + { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{v}}}}}}}}^{[2]}_{13 } + { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{v}}}}}}}}^{[2]}_{23 } ) , \end{split}\ ] ] where @xmath105 is the uncorrelated two - body potential and @xmath106}_{ij}$ ] is the two - body part of the correlated potential .
the first term describes the fully correlated two - body potential in a three - body system , which is given by a three - body coordinate transformation according to .
the second term is the sum of the two - body correlated potentials .
an analogous expression results for the local three - body part of the correlated kinetic energy @xmath107}$ ] , which involves and .
the only input needed to evaluate correlated quantities explicitly is the correlation function @xmath34 .
since @xmath34 reflects the properties of the correlations induced by the two - body interaction it can be extracted from the interacting two - body system .
one method to construct an optimal correlation function is based on the exact solution of the two - body schrdinger equation .
the exact relative wave function @xmath109 contains all information on interaction induced two - body correlations . for the case of energy @xmath108 and relative angular momentum @xmath110 the solid line in figure [ fig : opt_e0solution ] shows the radial wave function ( solid line ) for the lennard - jones potential ( thin dotted line ) .
the short - range part is dominated by a huge correlation hole , i.e. , a region where the strong repulsive core of the potential enforces a vanishing wave function .
we can extract a correlation function that describes these short - range correlations by requiring that the exact solution is reproduced by a correlated ansatz wave function @xmath111 for all @xmath112 inside some maximum radius @xmath113 .
thus the optimal correlation operator maps the short - range part of the uncorrelated wave function @xmath114 onto the exact solution @xmath115
. in coordinate representation this condition leads with and to @xmath116 which can be written as an implicit integral equation for the optimal correlation function @xmath28 . ]
@xmath117 ^ 3 = 3 \int_0^r\!\ !
{ \ensuremath{\mathrm{d}}}\xi\;\xi^2 \bigg [ \frac{\phi_{0}(\xi)}{\phi_{{\ensuremath{\mathrm{ex}}}}[{\ensuremath{r_+}}(\xi ) ] } \bigg]^2 .\ ] ] the uncorrelated wave function @xmath118 should be chosen in accordance with the uncorrelated many - body trial state .
for example , if a condensate of @xmath4 uncorrelated bosons is described by a @xmath4-fold product of a gaussian - shaped single - particle wave functions the relative wave function for each pair of bosons is again of gaussian shape .
the choice of the uncorrelated wave function @xmath118 determines , which structures of the exact solution are generated by the correlation operator and which have to be described by the ansatz state . in
order to isolate short - range correlations the uncorrelated state should be able to describe the long - range behavior of the exact solution . according to this requirement we construct a hybrid wave function with a long - range behavior given by the exact zero - energy solution outside the range @xmath119 of the interaction : @xmath120 where @xmath121 is the s - wave scattering length of the interaction . for two @xmath122he atoms interacting via the lennard - jones potential
we get @xmath123 .
the short - range part of the ansatz state is described by a gaussian . at some radius @xmath124
both functions are matched with continuous derivative .
the final form of the uncorrelated wave function reads @xmath125 & ; r
\le \chi \\[3 mm ] 1 - a / r & ; r > \chi \end{cases } .\ ] ] the matching radius @xmath124 is chosen such that the correlation function @xmath126 determined from vanishes at some finite radius @xmath113 .
this allows a natural separation of long- and short - range correlations in the two - body system . for the @xmath108 scattering problem ( solid line ) in comparison with the hybrid ansatz @xmath118 for @xmath127 ( dashed line ) .
in addition the radial dependence of the lennard - jones potential @xmath128 is shown ( thin dotted line ) . ] resulting from the integral equation ( solid line ) and the fitted parameterization ( thin solid line ) .
for comparison the parameterization with parameters resulting from energy minimization ( thin dotted line ) is depicted . ] in figure [ fig : opt_e0solution ] the hybrid ansatz with matching radius @xmath129 is shown in comparison with the exact @xmath108 solution .
both wave functions agree nicely at large distances . at small distances
the exact solution shows the pronounced depletion within the core and an enhancement in the attractive region of the potential .
these short - range correlations are not contained in the uncorrelated wave function ( dashed curve ) .
instead the unitary correlation operator @xmath5 , which maps the uncorrelated wave function onto the exact eigenfunction ( solid curve ) , takes over the task . as discussed in section [ sec : ucom_operator ] this has the great advantage to allow the definition of correlated interactions that act between simple uncorrelated many - body states like product states .
the solution of the integral equation for the correlation function @xmath28 using these two wave functions is shown by the dashed line in figure [ fig : opt_correlator ] .
the quantity @xmath126 corresponds to the net distance by which two particles with an initial separation @xmath112 are shifted away from each other .
the shape of the curve shows that particles with distances smaller than the core radius ( @xmath130 ) are shifted out of the core .
if the initial distance is already larger than the core radius , then the shift decreases rapidly and finally vanishes at @xmath131 .
the small shift for larger radii ( @xmath132 ) originates from the minimal deviations of exact wave function and hybrid ansatz in this range
. however these are long range correlations , which should not be described by the unitary correlation operator .
therefore we will set the function @xmath126 to zero for @xmath133 in order to define the short - range correlation function .
this intrinsic separation between long- and short - range correlations works only if the potential has a sufficiently strong attractive region in addition to the repulsive core . for this class of interactions
the exact two - body wave function is depleted inside the core and enhanced in the attractive region . since the unitary correlation operator conserves the norm of the wave function it can generate this kind of structure very easily : the amplitude shifted out of the core is placed in the attractive region . for purely repulsive potentials
the situation is different .
the exact wave function of a scattering state is depleted inside the core and the excess probability is spread over a large volume .
therefore the unitary mapping leads to a correlator of large range . in order to stay within the two - body approximation one has to restrict the range of the correlations and compensate by improving the uncorrelated trial states . an alternative method to determine the correlation function @xmath28 emerges from the variational principle .
if we think of the energy expectation value to be calculated with correlated states and the uncorrelated hamiltonian then the correlation function introduces additional degrees of freedom into the trial state . by choosing a suitable parameterization for @xmath28 with few variational parameters we can minimize the correlated energy expectation value and obtain an optimal correlation function .
according to the structure of the correlation function determined by the mapping procedure described in the previous section we use the following parameterization @xmath134\ ] ] with three free parameters .
the parameter @xmath135 determines the range of the correlation function , @xmath136 is related the maximum shift distance and thus the radius of the core , and @xmath137 influences the slope of @xmath28 at small radii , i.e. , the `` hardness '' of the core .
we used this parameterization successfully for the description of the structure of atomic nuclei and nuclear matter @xcite . in order to describe a homogeneous @xmath122he liquid we choose a constant wave function as two - body trial state @xmath138 . in that case only the local potentials @xmath139 and @xmath61 contribute to the expectation value of the correlated hamiltonian @xmath140 . \end{split}\ ] ] the expectation values of all momentum - dependent contributions @xmath141 in vanish . by inserting the parameterization of the correlation function into and we can express the integrand as function of the three variational parameters @xmath142 and @xmath137 .
the integration as well as the minimization can easily be done numerically .
we obtain a unique set of parameters for the optimal correlation function @xmath143 the solid line in figure [ fig : opt_correlator ] shows the resulting correlation function .
it is in very good agreement with the correlation function determined by the mapping procedure ( dashed line ) .
this demonstrates that the parameterization used in the variation is well suited for the correlation function . for numerical simplicity we will use the parameterized correlation function with parameters according to for the following investigations .
( thin dotted line ) , correlated potential @xmath144 ( dashed line ) and sum of correlated potential and local part of the correlated kinetic energy @xmath145 ( solid line ) . ]
( solid line ) and radial effective mass correction @xmath146 ( dashed line ) . ] with the explicit form of the correlation function we can study the structure of the correlated hamiltonian in detail .
as shown in section [ sec : ucom_operator ] the unitary transformation generates a momentum - dependent effective interaction .
the local part of this interaction consists of the correlated interaction potential @xmath46 and a local contribution @xmath147 from the correlated kinetic energy operator .
figure [ fig : opt_corrint_local ] compares the radial dependencies of the uncorrelated lennard - jones potential , the correlated lennard - jones potential and the sum of all local contributions . the unitary transformation of the potential shifts the repulsive core to smaller radii .
this leads to a substantial reduction of the strength of the core , only a moderate repulsion at very short ranges remains . if we add the local part of the correlated kinetic energy the repulsive part vanishes completely ; the local component of the correlated hamiltonian is purely attractive .
actually , it can be shown analytically that mapping of an uncorrelated wave function , which is constant within the range of the repulsive core , onto the exact @xmath108 solution results in a local part @xmath148 of the correlated hamiltonian that vanishes identically inside the core @xcite .
the small contributions of the correlated local potentials within the core range that show up in figure [ fig : opt_corrint_local ] result from the use of the restricted parameterization for the correlation function and the slight variation of the trial state .
the momentum - dependent parts of the correlated hamiltonian , which are formulated in terms of distance - dependent effective mass corrections , are shown in figure [ fig : opt_corrint_effmass ] .
the radial effective mass correction ( dashed line ) generates a repulsion if the uncorrelated relative wave function has non - vanishing radial derivatives .
the angular effective mass correction ( solid line ) generates an attraction if the uncorrelated wave function has non - vanishing relative angular momentum . in general the contribution of these components
is small compared to the total energy expectation value as the uncorrelated states have small gradients .
formally we can combine all the two - body terms of the correlated hamiltonian to a momentum - dependent effective interaction @xmath149 } = { \ensuremath{\tilde{v}}}({\ensuremath{{\ensuremath{\bm{r } } } } } ) + { \ensuremath{\tilde{u}}}({\ensuremath{{\ensuremath{\bm{r } } } } } ) + { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{t}}}}}}}}_{\omega } + { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{t}}}}}}}}_r\ ] ] with a purely attractive local part given by and and a momentum dependence given by . with this effective interaction the correlated hamiltonian in two - body approximation
has the standard form @xmath150}_{ij } .\ ] ] the next orders of the cluster expansion contribute additional three - body , four - body , and higher order effective interactions .
the tamed effective two - body interaction @xmath151}$ ] has a different operator structure and completely different radial dependencies than the uncorrelated potential @xmath41 .
however , both interactions are phase - shift equivalent , i.e. , in a two - body scattering process both interactions generate identical phase - shifts for any collision energy , by construction .
this generic property of the effective interaction results from the finite range of the unitary transformation , which does not influence the asymptotic behavior of the scattering solutions .
it is completely independent of the shape of the correlation function at short ranges . in that way
the unitary correlator provides a recipe to generate an infinite manifold of phase - shift equivalent interactions .
this aspect is of special interest in cases where the interaction is determined only on the basis of two - body scattering data like the nucleon - nucleon interaction .
as a first application of the ucom formalism we want to investigate the equation of state , i.e. , the energy per particle as function of the density , for a homogeneous @xmath122he liquid at temperature @xmath152k .
in this case all effects that the interaction has on the many - body state are described by the unitary correlation operator .
the uncorrelated many - body state @xmath153 is a direct product of identical constant one - body states @xmath154 @xmath155 for the calculation of expectation values we assume a finite volume @xmath156 containing @xmath4 particles . accordingly
the constant one - body states read @xmath157 the limit @xmath158 at constant density @xmath159 is performed in the final step of the calculation .
first we calculate the energy per particle with the correlated hamiltonian in two - body approximation @xmath160 } + { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{u}}}}}}}}^{[2 ] } + { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{t}}}}}}}}_{\omega}^{[2 ] } + { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{t}}}}}}}}_r^{[2 ] } .\ ] ] the energy per particle in two - body approximation is defined by the expectation value of this hamiltonian calculated with the uncorrelated many - body state @xmath161 since the uncorrelated state is a product of constant one - body states only the local components @xmath162 and @xmath163 contribute ; the expectation values of all momentum - dependent terms vanish .
a simple calculation for the expectation value of the local components leads to the following expression for the energy per particle in two - body approximation @xmath164 } + c_u^{[2]})\;\rho\ ] ] with constant coefficients given by @xmath165 } = \frac{1}{2 } \int\!\!{\ensuremath{\mathrm{d}}}^3r\ ; { \ensuremath{\tilde{v}}}(r ) , \qquad c_u^{[2 ] } = \frac{1}{2 } \int\!\!{\ensuremath{\mathrm{d}}}^3r\ ; { \ensuremath{\tilde{u}}}(r ) .\ ] ] the integrals over the correlated potential @xmath46 given by and the local part of the correlated kinetic energy @xmath147 according to are evaluated numerically @xmath166 } = -856.5 \ , \text{k\aa}^3 , \qquad c_u^{[2 ] } = 401.5 \ , \text{k\aa}^3 .\ ] ] obviously , for homogeneous bose fluids at zero temperature the two - body approximation is not able to describe saturation , i.e. , a minimum of the energy at some finite density .
the energy in two - body approximation is always proportional to density and due to the negative sum of the coefficients drops with increasing density .
he liquid as function of density calculated in two - body approximation @xmath167 ( dashed line ) and in three - body approximation @xmath168 ( dotted line ) .
the vertical line marks the density at which the smallness parameter @xmath75 reaches the value @xmath169 .
the circles show the results of an exact gfmc calculation @xcite ; statistical errors are smaller than the size of the symbols . ]
the dashed line in figure [ fig : liq_energy_2b3b ] shows this behavior in comparison with results of a green s function monte carlo ( gfmc ) calculation @xcite , which are exact up to the statistical errors of the monte carlo sampling .
despite the fact that the two - body approximation fails to generate a minimum , the predicted orders of magnitude for the energy are correct .
this is remarkable , considering that the uncorrelated lennard - jones potential would give an infinite positive energy expectation value for all densities .
the special unitary transformation that we apply tames the strong core of the potential and leads to a bound many - body system .
we note that the two - body approximation does generate saturation in other systems like homogeneous fermi liquids . in these cases
the mass corrections contribute , because of the fermi motion , and generate a positive contribution to the energy that grows with a higher power in density than the attractive local terms .
a straightforward step to go beyond the two - body approximation is the inclusion of the three - body terms of the cluster expansion .
in addition to the terms in the correlated hamiltonian in three - body approximation contains three - body contributions from the correlated potential and the correlated kinetic energy @xmath170 } + { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{u}}}}}}}}^{[3 ] } + { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{t}}}}}}}}^{[3 ] } .\ ] ] again only the local three - body terms @xmath171}$ ] and @xmath172}$ ] contribute to the energy expectation value with the uncorrelated states .
the resulting energy per particle in three - body approximation reads @xmath173 } + c_u^{[2]})\;\rho + ( c_v^{[3 ] } + c_u^{[3]})\;\rho^2 .\ ] ] the coefficients are defined by the six - dimensional integrals @xmath174 } = \frac{1}{6}\int\!\!{\ensuremath{\mathrm{d}}}^3x_1{\ensuremath{\mathrm{d}}}^3x_2\;{\ensuremath{\tilde{v}}}^{[3]}({\ensuremath{\vec{x}}}_1,{\ensuremath{\vec{x}}}_2,{\ensuremath{\vec{x}}}_3=0)\ ] ] and the corresponding expression for @xmath175}$ ] .
the integration is done with a standard monte carlo algorithm ( `` vegas '' from @xcite ) and yields @xmath176 } & = ( -2512 \pm 50)\ , \text{k\aa}^6 , \\
c_u^{[3 ] } & = ( 29739 \pm 1100)\ , \text{k\aa}^6 . \end{split}\ ] ] the local three - body terms of the correlated hamiltonian generate a positive contribution to the energy per particle which is proportional to @xmath177 .
the dominant term is the local three - body contribution of the correlated kinetic energy .
together with the negative density - proportional contribution of the local two - body potentials the energy in three - body approximation provides a minimum at finite density .
the density dependence of the energy per particle in three - body approximation @xmath168 is shown by the dotted curve in figure [ fig : liq_energy_2b3b ] .
obviously the three - body contribution is not a small correction to the two - body approximation for densities larger than @xmath178 , which is only one fourth of the saturation density .
this demonstrates that many - body correlations have a strong influence in these systems and that the two - body approximation alone is not sufficient to describe ground state properties .
a similar conclusion can be drawn from the smallness parameter @xmath179 .
it reaches the phenomenological limit @xmath76 for the validity of the two - body approximation at about @xmath180 of the expected saturation density
. however , the naive inclusion of the three - body terms of the cluster expansion of the correlated hamiltonian does not improve the result beyond a density of about @xmath178 .
although we obtain saturation in three - body approximation , both , energy and density at the minimum of @xmath168 are substantially smaller than the results of the gfmc calculation .
this discrepancy has several reasons : firstly , the optimal correlation function was determined in the two - body system , i.e. , in two - body approximation . for a consistent treatment on the level of the three - body approximation the optimal correlation function should be determined by minimizing the energy in three - body approximation .
secondly , we expect that genuine three - body correlations play a very strong role @xcite .
they are described by an irreducible three - body part @xmath181 in the generator that was not included in the present treatment .
finally , in view of the large three - body contributions there is no good reason to neglect terms beyond three - body order .
each of the points mentioned can in principle be explicitly included in the calculation .
however , none of these corrections are expected to yield reasonably converged results in the particular case of the @xmath122he liquid .
the density range where the description should be applied is far above the range where the two - body approximation is applicable , i.e. , where the smallness parameter @xmath75 is below the typical value of @xmath169 .
the successive inclusion of higher orders of the cluster expansion will extend this range but one would have to include very high orders which are beyond the numerical possibilities to reach the expected saturation density .
the only reasonable way to extend the treatment is a partial summation over all cluster orders .
this is done in the framework of jastrow correlations with the so called hypernetted chain ( hnc ) summation schemes @xcite .
however , this type of partial summation is not feasible for the unitary correlation operator because of its more complicated operator structure .
we aim at an effective description of the higher orders of the cluster expansion that still allows a compact analytic formulation of correlated observables and does not require extensive numerical efforts .
one general way to simulate the effect of higher cluster orders is to introduce density - dependent correlation functions on the level of the two - body approximation . at low densities
the correlator shifts a pair of particles in an optimal way from the repulsive into the attractive region .
when the density grows other particles will be nearby and obstruct the correlation that was optimal for a free particle pair . to simulate this effect
the range of the correlation function @xmath28 should be more and more reduced for increasing density .
this will reduce the binding as the particles will not be able any more to fully exploit the attractive region of the potential . in practice
this is accomplished by scaling the parameters @xmath136 and @xmath135 , which have the dimension of a length , in the parameterization of the correlation function @xmath28 with some factor @xmath182 that depends on density . at very low densities
we assume that the higher - order contributions are negligible such that @xmath183 . with growing density
the effect of the higher cluster orders grows such that the scaling factor @xmath184 has to drop .
the most simple ansatz for the density dependence of @xmath184 is @xmath185 with one free parameter @xmath186 .
the density - dependent correlation function reads @xmath187^{\!\eta } \exp\!\!\big[\!\!-\exp\!\big(\frac{r}{\beta\xi(\rho)}\big)\!\big].\ ] ] using this correlation function all components of the correlated hamiltonian like correlated potential , local part of the correlated kinetic energy , and effective mass corrections become explicitly density - dependent .
the energy per particle is given by @xmath188,\ ] ] where the subscript @xmath71 indicates that the density - dependent correlation function is used .
due to the explicit density - dependence of the correlated local potentials the equation of state is not proportional to @xmath71 any more . from a taylor expansion of around @xmath189
we obtain terms to all powers of the density .
thus all orders of the cluster expansion are represented by the density - dependent correlator in an effective way .
the phenomenological parameter @xmath186 in the scaling factor is in general adjusted such that agrees with one experimental data point , or for some density with the result of a realistic calculation .
we will use the result of the gfmc calculation @xcite for the density @xmath190 and energy @xmath191 at the saturation point and require @xmath192 this ensures that the equation of state calculated with the density - dependent correlator runs through that point .
however , this does not imply that the minimum of coincides with the gfmc minimum . using the values @xcite @xmath193 and @xmath194
we obtain @xmath195 we have chosen to fix the density dependence , i.e. , @xmath186 , with a calculation instead of an experimental value , because that calculation uses the same potential and it is known that this simple lennard - jones potential does not reproduce exactly the experimental two - body data .
he liquid calculated in two - body approximation with a two - body optimized correlation function @xmath167 ( dashed line ) and with density - dependent correlation function @xmath196 ( solid line ) .
the vertical line marks the density at which the smallness parameter @xmath75 reaches the value @xmath169 .
the gray circles show the results of an exact gfmc calculation @xcite , the minimum is indicated by the black circle . ]
figure [ fig : liq_energy_2bdd ] shows the energy per particle @xmath196 calculated with the density - dependent correlation function .
the phenomenological description of the higher order contributions by a density - dependent correlation function indeed generates a minimum of the energy with @xmath197 the position of the minimum ( indicated by the cross ) as well as the shape of the equation of state around the minimum agree very nicely the the gfmc result . as noted above this
is not an immediate consequence of the adjustment of @xmath186 .
the agreement shows that the phenomenological density - dependence is well suited to describe the major aspects of many - body correlations on a numerically very simple level .
actually , all calculations have been performed within a simple _ mathematica _ notebook . once the density - dependence is fixed the correlation function can be used to evaluate any correlated observable of interest .
the analytic expressions for many correlated observables in two - body approximation are rather simple , but the evaluation of higher cluster orders becomes complicated if not intractable .
thus the inclusion of the phenomenological treatment of higher order effects by means of the density - dependent correlator is a very useful tool .
interesting quantities are , e.g. , the radial two - body distribution function @xmath198 , the static structure factor @xmath199 or the one - body momentum distribution @xmath200 .
the corresponding observables for fermionic liquids were discussed elsewhere @xcite . in the following section
we will use the density - dependent correlation function to investigate the structure of small droplets of @xmath122he .
this will demonstrate the universality of the concept .
in a next step we want to use the unitary correlation operator method to investigate the ground state properties of small @xmath122he droplets .
the many - body problem is treated in a simple variational framework .
the many - body trial state @xmath201 is like for the homogeneous liquid assumed to be a direct product of identical one - body states .
the one - body states are described by a gaussian wave function @xmath202\ ] ] with a common width parameter @xmath121 as the only variational degree of freedom .
the gaussian trial state is applicable for small droplets with @xmath203 ; larger droplets show a saturation of the central density , which can not be modeled with this ansatz @xcite .
an advantage of the gaussian ansatz is that the two - body wave function @xmath204 can be separated analytically into a center of mass component @xmath205 and a relative wave function @xmath206 @xmath207 , \\
{ \ensuremath { \big < { { \ensuremath{\vec{r } } } } \big| { \phi } \big > } } & = ( 2\pi a)^{-3/4 } \exp[-{\ensuremath{\vec{r}}}^{\,2}\!/(4 a ) ] .
\end{split}\ ] ] this simple trial state is able to describe mean - field like effects of the interaction , i.e. , the spatial localization of the atoms .
however , any correlation effect beyond the mean - field level has to be described by the unitary correlation operator .
first we use the two - body approximation with the correlation function obtained from the two - body scattering solution to investigate the ground state properties of small droplets .
the correlated intrinsic hamiltonian @xmath208 in two - body approximation has the structure @xmath209 } + { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{u}}}}}}}}^{[2 ] } + { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{t}}}}}}}}_{\omega}^{[2 ] } + { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{t}}}}}}}}_r^{[2 ] } , \ ] ] where capital letters indicate summation over all particles , like @xmath210 } = \sum_{i < j}^n { \ensuremath{\tilde{{\ensuremath{{\ensuremath{\bm{v}}}}}}}}^{[2]}_{ij}$ ]
. the uncorrelated intrinsic kinetic energy @xmath211 , i.e. , the total kinetic energy @xmath212 reduced by the center of mass contribution @xmath213 , can be expressed by the relative two - body kinetic energy @xmath50 @xmath214 where @xmath215 ^ 2 .\ ] ] due to the simple structure of the many - body trial state the expectation value of the correlated hamiltonian in the @xmath4-body system can be expressed by that of the two - body system @xmath216 moreover we can use that the intrinsic hamiltonian acts on the relative part @xmath206 of the two - body wave function only .
this leads to a simple form of the correlated energy expectation value of the @xmath4-body system @xmath217 we used that the expectation value of the angular effective mass correction @xmath218 vanishes due to the spherical symmetry of the relative wave function ( relative s - wave state ) .
the two - body expectation value of the relative kinetic energy is given by @xmath219 for the correlated two - body potential and the local part of the correlated kinetic energy we obtain @xmath220\;\big|{\ensuremath { \big < { r } \big| { \phi } \big > } } \big|^2 .\ ] ] finally , the expectation value of the radial effective mass correction reads @xmath221 for a fixed correlation function @xmath28 the radial dependencies of the correlated potential @xmath46 , the local part of the correlated kinetic energy @xmath147 , and the radial effective mass correction @xmath146 are known analytically ( see section [ sec : ucom_operator ] ) .
the minimization of the energy @xmath222 as function of the width parameter @xmath121 as well as the calculation of the integrals in equations and is done numerically .
another interesting observable is the mean - square radius of the droplet , defined by the expectation value of the operator @xmath223 where @xmath224 is the center of mass coordinate of the many - body system .
the formulation in terms of the two - body operator @xmath225 reveals a similar structure like for the intrinsic kinetic energy .
the many - body product state under consideration allows a direct calculation of the uncorrelated expectation value using the relative two - body wave function @xmath226^{1/2 } \\ & = \bigg[\frac{n-1}{2n } { \ensuremath { \big < { \phi } \big| \,{{\ensuremath{{\ensuremath{\bm{r}}}}}^2}\ , \big| { \phi } \big > } } \bigg]^{1/2 } = \bigg[\frac{n-1}{2n}\ , 3 a\bigg]^{1/2 } .
\end{split}\ ] ] the correlated ms - radius @xmath227 in two - body approximation has a similar structure like the kinetic energy @xmath228 } .\ ] ] in addition to the uncorrelated mean - square radius @xmath229 the unitary transformation generates a two - body contribution @xmath230 } = \frac{1}{2n } \sum_{i < j}^n \big[{\ensuremath{r_+}}^2({\ensuremath{{\ensuremath{\bm{r}}}}}_{ij } ) - { \ensuremath{{\ensuremath{\bm{r}}}}}_{ij}^2\big ] .\ ] ] the expectation value of the correlated ms - radius can again be expressed by the relative two - body wave function alone @xmath231^{1/2}.\ ] ] in order to study the quality of the two - body approximation we first look at very small droplets with @xmath232 . for these droplet sizes an early variational monte carlo ( vmc ) calculation on the basis of the lennard - jones potential exists @xcite .
the authors use a jastrow - type parameterization of two - body correlations with a long - range form adjusted to the behavior of the exact two - body solution .
the calculation of the many - body expectation value with these trial states involves a monte carlo integration routine .
the resulting energy expectation values show a statistical error of approximately 10% .
our results of the minimization of @xmath222 with respect to @xmath121 using the system independent correlation function given in equations and for @xmath232 are summarized in figure [ fig : drop_2b ] .
the upper panel shows the correlated energy per particle , the middle panel the correlated rms - radius of the droplet and the lower panel the smallness parameter @xmath233 versus the number of particles @xmath4 .
( lower panel ) for small @xmath122he droplets in two - body approximation with two - body optimized correlation function .
the gray circles show results of a variational monte carlo calculation @xcite , statistical errors for the energy are indicated by the gray bars .
the lines should guide the eye . ] as for the homogeneous liquid the expectation values of the uncorrelated lennard - jones potential would diverge .
but the unitary transformation of the hamiltonian using the correlation function determined in the two - body system removes the strong short - range repulsion of the potential completely . with this
tamed correlated potential and a simple product ansatz for the many - body wave function the energies of very loosely bound droplets actually agree very nicely for @xmath234 with the vmc result . for @xmath235
the energy expectation value is below the vmc value and the difference grows with increasing particle number or density .
since expectation values are evaluated in two - body approximation , i.e. , the full cluster expansion is truncated above two - body order , a basic property of the ritz variational principle is lost : the variational minimum of the energy is not necessarily bounded from below by the exact energy eigenvalue .
the observed overbinding indicates that the higher orders of the cluster expansion have to give a sizeable positive contribution to raise the expectation value above the exact eigenvalue . in direct connection with the overbinding in two - body approximation the correlated rms - radii of energy minimized states are systematically smaller than the vmc result .
accordingly the densities are too high .
in contrast to the homogeneous liquid the two - body approximation does not produce a pathological collapse towards high densities for small droplets . this is due to the positive contributions of the kinetic energy and the effective mass correction that are absent in the homogeneous case . as a measure for the validity of the two - body approximation we used the smallness parameter @xmath236 with correlation volume @xmath72 defined by .
in order to specify the smallness parameter for inhomogeneous systems we have to define a measure for the average density . one possible definition is the value of the density for a step - like density profile with same ( uncorrelated ) rms - radius as the original distribution . for identical gaussian single - particle wave functions with width parameter @xmath121 this equivalent density
is given by @xmath237^{3/2 } = \frac{3 n}{4 \pi } \bigg[\frac{2}{5 } \ ; \frac{n}{n-1}\ ; \frac{1}{a } \bigg]^{3/2}.\ ] ] the smallness parameter @xmath238 defined with this density is shown in the lower panel of figure [ fig : drop_2b ] as function of particle number .
like for the homogeneous liquid the energy in two - body approximation starts to deviate from the exact result if the smallness parameter @xmath233 exceeds the value @xmath169 ( indicated by the dashed line ) .
this again confirms that the two - body approximation is valid as long as the smallness parameter fulfills the condition @xmath239 . to account for the effect of many - body correlations in a simple but efficient way we employ the concept of density - dependent correlation functions introduced in section [ sec : liq_eos_densdep ] for the homogeneous liquid . to apply the density - dependent correlator in an inhomogeneous system we have to specify in which way the density - dependence
should be evaluated .
the most simple approach is to insert the equivalent density of the droplet , as defined in , into the density - dependent correlation function .
that means we neglect that the effects from stronger many - body correlations in the center of the droplet do not exactly cancel the weaker ones at the surface . a more elaborate ansatz to account for
the inhomogeneity would be to implement the density - dependence in local density approximation , i.e. , the correlation function which acts on the relative coordinate of a particle pair depends on the local density at their center of mass position .
this is not done here . [
cols="^,^,^,^,^,^,^,^,^,^,^,^,^ " , ] the variational calculation of the ground state properties of small @xmath122he droplets including a correlation function that depends on the equivalent density is straightforward .
the parameterization of the correlation function @xmath240 is used with the parameters determined in the two - body system .
the parameter of the density - dependent scaling function is taken from the investigations of the homogeneous liquid .
the energy expectation value in two - body approximation with density - dependent correlations is of the same form as for the density - independent correlation functions discussed in the preceding section .
( lower panel ) for small @xmath122he droplets in two - body approximation with the density - dependent correlation function fixed in the homogeneous liquid .
the gray circles show results of a variational monte carlo calculation @xcite .
the lines should guide the eye . ]
the results of the energy minimization with respect to the width parameter @xmath121 are summarized in figure [ fig : drop_2bdd ] .
the energy expectation values are in full agreement with the vmc calculation @xcite for all particle numbers .
the overbinding observed with a density - independent correlation function is completely compensated by the density - dependence .
the correlated rms - radii of droplets with @xmath241 are consistently larger by 0.2 than the vmc result .
unfortunately the authors @xcite give no estimate for the statistical or systematic errors for their procedure to obtain the rms - radii .
for the smallest droplets our result for the rms - radii are below the vmc results .
this may be caused by the restriction to gaussian trial states , which are not able to describe the long - range exponential tail present in these extremely weakly bound systems . at the same time the insufficient trial state causes a slight underbinding of the smallest droplets . for completeness figure
[ fig : drop_2bdd ] also shows the smallness parameter @xmath238 .
since both , the correlation volume @xmath72 and the density of the droplets are reduced by the density - dependent correlator the product stays small . in any case
@xmath233 has lost its meaning as a measure for three - body correlations as the density - dependence of the correlator already includes many - body correlations .
table [ tab : drop ] summarizes the results of the energy minimization with two - body optimized and density - dependent correlator for droplet sizes up to @xmath242 .
in addition to the correlated energy and the rms - radius the individual terms of the correlated hamiltonian are shown . the attractive correlated potential and the repulsive local part of the correlated kinetic energy are the major contributions and show a large cancellation .
the intrinsic kinetic energy and the effective mass correction give rather small contributions .
we presented a simple and elegant approach to describe strong short - range correlations in a many - body system that result from the strong repulsive core of the two - body interaction .
these correlations induce high - momentum components into the relative wave function that can not be treated within a low - momentum model space .
the correlation operator @xmath5 generates the short - range correlations by a unitary transformation in the relative coordinate of each pair of particles .
it shifts the particles out of the repulsive core of the two - body interaction .
the correlation operator is then used to define correlated observables , most notably the correlated hamiltonian .
the unitary transformation yields a tamed effective two - body interaction , which consists of a purely attractive local part and a weak momentum dependence .
in addition three - body and higher order terms are generated .
for dilute systems , i.e. , if the range of the correlations is small compared to the average distance between the particles , it is sufficient to include the terms up to two - body order . in high - density systems the higher order terms can be included explicitly or simulated by a density - dependent correlation function . in order to study the applicability of the method in many - body systems , where short - range correlations play a dominant role
, we investigated the ground state properties of homogeneous @xmath122he liquids and small droplets .
the correlation function is fixed by energy minimization in the two - body problem with a constant uncorrelated wave function .
alternatively the correlation function can be determined by mapping an uncorrelated wave function onto the exact energy @xmath108 two - body wave function .
we use this correlation function to calculate the energy of the homogeneous @xmath122he liquid as function of density .
the uncorrelated many - body state is a direct product of constant one - body states , thus the influence of the interaction on the state has to be described by the correlation operator alone . already in two - body approximation
the strong short - range repulsion of the lennard - jones potential is tamed completely and we obtain a bound system .
however , the two - body approximation does not generate saturation , i.e. , the energy per particle decreases proportional to density . to simulate the effect of many - body correlations
we employ a density - dependent correlation function .
the one parameter introduced by the density dependent scaling of the correlation function is fixed by adjusting the correlated energy for one value of the density to the result of a gfmc calculation .
the resulting equation of state shows a minimum and is in very good agreement with the predictions of the gfmc calculation over the whole density range . in a second step
we use the density - independent as well as the density - dependent correlation function to perform an _ ab initio _ calculation of the ground state structure of small @xmath122he droplets in a simple variational framework .
it turns out that the two - body approximation with the density - independent correlation function gives a good description for small droplets with particle numbers @xmath234 . for larger droplets an overbinding compared to a quasi exact variational monte carlo calculation appears .
due to the higher density in these droplets many - body correlations become relevant .
if we use the density - dependent correlation function with one parameter adjusted in the homogeneous liquid then the energies and rms - radii are in very good agreement with the exact calculation for all particle numbers under consideration .
we conclude that the unitary correlation operator method is a powerful and universal tool to handle short - range correlations in the many - body system . for sufficiently low densities one can use the two - body approximation to construct the correlated hamiltonian as well as other correlated observables in closed form .
the resulting effective two - body interaction can be used for an ab initio description of the many - body system based on a low - momentum model space .
a gross measure for the validity of the two - body approximation is the smallness parameter @xmath75 , which should be below a phenomenological limit of @xmath169 . for higher densities one can utilize density - dependent correlation functions to describe many - body correlations effectively within the two - body approximation .
the additional parameter in the density - dependence can be fixed by one measured data point or by comparison with an exact calculation .
nevertheless , the resulting density - dependent correlator can then be used for a broad range of applications , e.g. , _ ab initio _ calculations of correlated observables like energies , densities or momentum distributions of different systems without further adjustments .
the great advantage of the unitary correlation operator method with density - dependent correlation functions is that besides its transparent physics it requires only minimal computational power compared to the huge computational resources needed for monte - carlo - type calculations .
a set of _ mathematica _ notebooks that illustrate the application of the unitary correlation operator method to @xmath0he liquids can be found at the url : ` http://theory.gsi.de/~rroth/math/ ` . | the unitary correlation operator method ( ucom ) is employed to treat short - range correlations in both , homogeneous liquid and small droplets of bosonic @xmath0he atoms .
the dominating short - range correlations in these systems are described by an unitary transformation in the two - body relative coordinate , applied either to the many - body state or to the hamiltonian and other operators .
it is shown that the two - body correlated interaction can describe the binding energy of clusters of up to 6 atoms very well , the numerical effort consisting only in calculating one two - body matrix element with gaussian single - particle states .
the increasing density of bigger droplets requires the inclusion of correlation effects beyond the two - body order , which are successfully implemented by a density - dependent two - body correlator . with
only one adjusted parameter the binding energies and radii of larger droplets and the equation of state of the homogeneous @xmath0he liquid can be described quantitatively in a physically intuitive and numerically simple way . |
observations at long wavelengths , particularly those made by the _ infrared astronomical satellite _ ( _ iras _ ) and the _ infrared space observatory _ ( _ iso _ ) , have revealed a low - redshift population of galaxies which emit most of their luminosity in the far infrared .
ultraluminous infrared galaxies ( ulirgs ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , which have far - infrared luminosities greater than @xmath10 , represent the luminous tail of this population . infrared number counts and the cosmic infrared background
provide strong evidence that the ulirg population becomes progressively more important at high redshift @xcite .
in addition , submillimeter galaxies ( smgs ) have been revealed as likely ulirg analogs at high redshift @xcite .
however , although the energy generation mechanisms of these extremely luminous objects have been a topic of great interest and study , the relative fraction of high redshift smgs powered by buried agn is still a matter of debate , as are the relative contributions of agn and starbursts to the energy output of ulirgs at low redshift .
mid - infrared diagnostics of the energetics of galaxies have been developed using ground - based @xcite and space - based @xcite observations . in general , the mid - infrared spectra of starbursts are characterized by low - excitation fine structure lines , polycyclic aromatic hydrocarbon ( pah ) features , and a weak 36 @xmath1 rest - frame continuum .
in contrast , the mid - infrared spectra of agn exhibit high - excitation emission lines , very weak or absent pahs , and a strong 36 @xmath1 rest - frame continuum . until recently
, the sensitivity of available spectrometers has limited the application of these diagnostics to the brightest galaxies at low redshift .
the highly sensitive instruments aboard the _
spitzer space telescope _
@xcite , especially the _ infrared spectrograph _
( _ irs _ ; * ? ? ?
* ) , are powerful tools for performing detailed studies of ulirgs at higher redshifts than were previously possible with _ iras _ or _
iso_. in order to identify high - redshift ulirgs for detailed study , the _ irs _ and _ mips _ instrument teams have completed a mid - infrared imaging survey of the 9 deg@xmath11 botes region of the noao deep wide - field survey ( ndwfs ; jannuzi et al . , in preparation ) .
the ndwfs is a @xmath12 imaging survey reaching 3@xmath13 point - source depths of approximately 27.7 , 26.7 , 26.0 , and 19.6 vega magnitudes , respectively .
in particular , we used the ndwfs data release 3 catalog .
the mid - infrared imaging was carried out with the _ multiband imaging photometer for spitzer _ ( _ mips _ ; * ? ? ?
* ) and reaches 5@xmath13 limits of 0.28 , 35 , and 100 mjy at 24 , 70 , and 160 @xmath1 , respectively .
most ( 8.5 deg@xmath11 ) of this field has also been surveyed using the _ infrared array camera _ ( _ irac _ ; * ? ? ?
* ) , also aboard _ spitzer _ , to 5@xmath13 sensitivities of 6.4 , 8.8 , 51 , and 50 @xmath14jy at wavelengths of 3.6 , 4.5 , 5.8 , and 8.0 @xmath1 , respectively @xcite .
the acis - i instrument @xcite aboard the _ chandra x - ray observatory _ was also used to map the entire botes field down to a limiting sensitivity of @xmath15 erg @xmath16 s@xmath17 in the energy range 0.57 kev @xcite .
the botes field also overlaps with the 20 cm first survey , which has a limiting sensitivity of approximately 1 mjy @xcite .
the ndwfs botes field was chosen for the mips imaging survey because of its low infrared background and high - quality , deep @xmath18 imaging .
these properties allow the selection of mid - infrared sources which are bright enough for _ irs _ spectroscopy but which have very faint optical counterparts , and are therefore likely at high redshift .
@xcite describe the results of _ irs_observations of 30 extreme ( @xmath19 mjy ; @xmath20 mag ) sources identified in botes .
the vast majority of the 17 objects for which the mid - infrared spectra yielded redshifts are probably obscured agn at @xmath21 .
we present here a detailed analysis of two objects which were also selected from the botesfield , but which have unique infrared and optical properties .
in particular , we present _ irs _ and optical spectroscopy , and constrain the spectral energy distributions of these objects using data from the previously described multiwavelength botes surveys , as well as new submillimeter data .
mips j142824.0 + 352619 ( hereafter source 1 ) was selected from a catalog of objects that were detected in all three mips bands ( thus the prefix `` mips '' ) . following the naming scheme of @xcite
, we use the ndwfs @xmath22-band coordinates coordinates are @xmath23 , + 35@xmath2426@xmath251914 ( j2000 ) . ] to form the name of this object .
source 1 was chosen for follow - up because it is bright at 160 @xmath1 ( @xmath26 mjy ) , has red optical near - infrared colors ( @xmath27 ) , and is unresolved in the optical .
the large 160 @xmath1 flux indicates that source 1contains large amounts of cold dust .
the red optical near - infrared colors and compact morphology suggest that it resides at @xmath9 .
the combination of the 160 @xmath1 flux and the high redshift implies that source 1 probably has a large far - infrared luminosity ( @xmath28 ) .
@xcite presented the spectral energy distribution ( sed ) of this object from the optical through the radio .
they argued that while the emission shortward of 1 @xmath1 ( observed frame ) is likely dominated by a @xmath29 lens , the near - infrared emission is dominated by the 24 @xmath1 source .
based on a cool ( @xmath3043 k ) dust temperature , adherence to the far - infrared radio correlation , and the presence of a prominent 1.6 @xmath1 stellar bump , they argued that source 1is a dusty starburst .
as discussed in [ sec : firstpah ] , the _ irs _ spectrum we present provides a rough redshift of @xmath2 .
@xcite used this estimate to identify h@xmath31 in the near - infrared spectrum and obtained @xmath32 .
in combination with their sed , this redshift confirms a high luminosity of @xmath33 .
they also noted that the instrinsic luminosity may be up to a factor of 10 lower if lensing is important .
here we present additional diagnostics to determine the energetics of source 1 .
sst24 j142827.19 + 354127.71 ( hereafter source 2 ) , was selected from a catalog of objects detected at 24 @xmath1 ( thus the prefix `` sst24 '' ) . following the convention of @xcite
, we use the 24 @xmath1 position to derive the name of this object .
source 2 has the brightest 24 @xmath1 flux ( @xmath34 mjy ) among mid - infrared sources with faint optical counterparts ( @xmath5 ) .
the faint @xmath35-band magnitude ( @xmath36 ) implies that source 2 is at high redshift , and the 24 @xmath1 flux indicates a large luminosity .
the plan of this paper is as follows : [ sec : observations ] describes the mid - infrared _ irs
_ observations of both sources , as well as optical spectroscopy and submillimeter imaging of source 2 .
the results for each object are presented in
[ sec : firstobjectlong ] and [ sec : secondobjectlong ] .
finally , our conclusions are discussed in [ sec : conclusions ] . throughout
, we use the following cosmological parameters : @xmath37 , @xmath38 , and @xmath39 km s@xmath17 mpc@xmath17 .
both objects were observed with the _ irs _ short low ( sl ) module in first order for a total of 480 seconds and with the long low ( ll ) module in first and second orders for a total of 1920 seconds each .
targets were placed on the slits by performing moderate - accuracy peak - ups on and offsetting from nearby 2mass stars . observed two dimensional spectra were processed with version s11.0 of the spitzer science center pipeline .
further processing was carried out using the basic calibrated data ( bcd ) products from the pipeline . since the sl data were taken in first order only , a background image for each position ( nod )
was constructed from data taken in the other position .
in contrast , the background images for each ll position were constructed from data taken when the object was in the other order .
extraction of one dimensional spectra corresponding to each position was accomplished with the smart analysis package @xcite .
the data from each position were then averaged to produce the final spectra , which provide low resolution coverage from @xmath30838 @xmath1 .
optical spectra of both source 1 and source 2 were obtained with the deep imaging multi - object spectrograph ( deimos ; * ? ? ?
* ) on the w. m. keck ii 10-meter telescope on the night of ut 2005 may 06 .
the observations , three 30 minute exposures , were obtained in 0.5 `` seeing through a 1.0 '' wide and 10 " long slitlet oriented at pa=130 .
a 600 line mm@xmath17 ( 7500blaze ) grating setting with the gg400 blocking filter was used .
the central wavelength was set to 7500 .
this corresponds to a 0.65per pixel mean spectral dispersion .
the deep2 data reduction pipeline was used to perform cosmic ray removal , flat - fielding , co - addition , sky - subtraction , and wavelength calibration .
the one - dimensional spectra were extracted using standard iraf routines .
relative spectrophotometric calibration was performed using observations of wolf 1346 @xcite .
the optical spectrum of source 1 is presented and discussed by @xcite .
submillimeter images of source 2 were taken on ut 2004 april 1012 with the 350 @xmath1 filter on the sharc - ii camera at the caltech submillimeter observatory .
pointing and flux calibration were performed on nearby arp 220 , and data were reduced using the crush software package @xcite .
weather conditions were not optimal , with an average opacity of @xmath40 .
the total exposure time was roughly 1 hour .
the object was not detected , though we derive a 3@xmath13 upper limit of 150 mjy .
ccccc & & ( source 1 ) + wavelength & & & & + 445 nm ( @xmath41 ) & 0.25 @xmath42 0.03 @xmath14jy & mosaic - i & 0.98 @xmath42 0.03 @xmath14jy & mosaic - i + 658 nm ( @xmath35 ) & 2.02 @xmath42 0.08 @xmath14jy & mosaic - i & 2.29 @xmath42 0.10 @xmath14jy & mosaic - i + 806 nm ( @xmath22 ) & 6.21 @xmath42 0.11 @xmath14jy & mosaic - i & 5.21 @xmath42 0.13 @xmath14jy & mosaic - i + 1.22 @xmath1 ( @xmath43 ) & 31.9 @xmath42 5.3 @xmath14jy & wirc & & + 1.63 @xmath1 ( @xmath44 ) & 38.4 @xmath42 5.2 @xmath14jy & wirc & & + 2.19 @xmath1 ( @xmath45 ) & 72.4 @xmath42 9.4 @xmath14jy & wirc & 134.8 @xmath42 10.6 @xmath14jy & onis + 3.6 @xmath1 & 250.9 @xmath42 7.2 @xmath14jy & irac & 296.3 @xmath42 8.9 @xmath14jy & irac + 4.5 @xmath1 & 280.6 @xmath42 8.7 @xmath14jy & irac & 621.3 @xmath42 18.6 @xmath14jy & irac + 5.8 @xmath1 & 198.3 @xmath42 16.6 @xmath14jy & irac & 1600.9 @xmath42 48.0 @xmath14jy & irac + 8.0 @xmath1 & 211.1 @xmath42 14.2 @xmath14jy & irac & 3938.9 @xmath42 118.2 @xmath14jy & irac + 24 @xmath1 & 0.72 @xmath42 0.07 mjy & mips & 10.55 @xmath42 0.13 mjy & mips + 70 @xmath1 & 34 @xmath42 6 mjy & mips & & + 160 @xmath1 & 430 @xmath42 90 mjy & mips & @xmath4645 mjy ( 3@xmath13 ) & mips + 350 @xmath1 & 226 @xmath42 45 mjy & sharc - ii & @xmath46150 mjy ( 3@xmath13 ) & sharc - ii + 850 @xmath1 & 21.9 @xmath42 1.3 mjy & scuba & & + 20 cm & 0.937 @xmath42 0.039 mjy & vla & @xmath460.98 mjy ( 5@xmath13 ) & vla + [ table : multiwavelengthphotometry ]
as seen in the top panel of figure [ fig : spectra ] , the mid - infrared spectrum of source 1 is rich in emission features , including the pah bands at 6.2 , 7.7 , 11.3 , and 12.7 @xmath1 .
these features indicate a redshift of @xmath47 . based on a preliminary reduction of the _ irs _ data
, @xcite identified the narrow h@xmath31 and adjacent [ ] lines in a near - infrared spectrum to determine a more accurate redshift of @xmath48 .
all plots and calculations herein adopt this more accurate redshift .
based on the optical spectrum , @xcite also considered whether or not the mass of a foreground @xmath29 object in the field is significantly magnifying the background @xmath2 source .
luminosities have not been corrected , since the magnification is unknown .
the pah features in the _ irs _ spectrum of source 1 not only provide its redshift , but are also clues to the source of its energy generation .
the strength of pah features roughly correlates with the starburst contribution to the mid - infrared flux ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* armus et al . , in preparation ) . in the spectrum of source 1 ,
the 7.7 and 11.3 @xmath1 pah features are adjacent to the broad 9.7 @xmath1 silicate absorption band , making the continuum at these wavelengths uncertain , while the 12.7 @xmath1 pah feature is blended with the [ ] emission line at 12.8 @xmath1 .
we therefore use the 6.2 @xmath1 pah rest - frame equivalent width ( @xmath49 ) as our primary diagnostic of the starburst contribution to the energy generation .
for source 1 , @xmath50 , 2550% lower than the typical value measured for nearby starbursts ( 0.5 0.7 @xmath1 ; brandl et al .
, in preparation ) , but much larger than that measured for nearby agn ( 0.005 0.02 @xmath1 ; * ? ? ?
the equivalent width diagnostic therefore indicates that the mid - infrared spectrum of source 1 is starburst - dominated .
compared to local ulirgs ( armus et al .
, in preparation ) , source 1has a large 6.2 @xmath1 pah luminosity of @xmath51 .
however , the top panel of figure [ fig : seds ] shows that this object has strong @xmath52 emission from large amounts of @xmath3043k dust .
we therefore quantified the contribution of the pah feature to the bolometric luminosity of the source .
for source 1 , @xmath53 , where the error bar includes contributions from the uncertainties in both @xmath54 and @xmath55 .
the latter was determined by @xcite , and includes uncertainties associated with fitting a modified blackbody to the sed .
it does not include an estimate of the range of alternate models that could be used for this fit . since source structure depends on wavelength , differential magnification as a function of wavelength represents an added complication in measuring @xmath56 .
the value of @xmath56 measured for source 1 is a factor of @xmath3023 smaller than that seen in other prototypical starbursts and starburst - dominated ulirgs , including ngc7714 , m82 , and ugc5101 , which have @xmath57 , @xmath58 , and @xmath59 , respectively .
it is within a factor of 1.5 smaller than the values measured for the @xmath21 starburst - dominated ulirgs recently discovered by @xcite , which have @xmath60 and @xmath61 . given the substantial uncertainties discussed above , the normalized pah luminosity of source 1 is consistent with all of these comparison objects .
recently , @xcite presented the _
spectra of two @xmath62 smgs at @xmath63 .
they found that these smgs appear to be scaled - up versions of local ulirgs , based on a comparison of their pah to far - infrared ratios pah feature to the rest - frame 222 @xmath1 ( observed - frame 850 @xmath1 at @xmath64 ) continuum flux density ] . using the best - fitting modified blackbody derived for source 1 by @xcite
, we find its pah to far - infrared ratio is @xmath65 .
this is very similar to the values reported for the @xmath63 smgs and local ulirgs .
in addition to pah features , the mid - infrared spectra of galaxies may contain additional emission features , such as [ ] at 6.9 @xmath1 , [ ] at 8.9 @xmath1 , [ ] at 10.5 @xmath1 , [ ] at 12.8 @xmath1 , [ ] at 15.6 @xmath1 , [ ] at 7.6 @xmath1 , and [ ] at 14.3 @xmath1 ( all rest wavelengths ) .
the expected positions of these features are marked in figure [ fig : spectra ] .
due in part to the low signal - to - noise and resolution of our spectra , many of these lines are not detected .
however , it is possible that a blend of [ ] and h@xmath66 s(5 ) is detected near 7 @xmath1 , and [ ] at 12.8 @xmath1 is clearly seen on top of the broader pah feature , with a flux of @xmath67 w m@xmath68 .
we measure a 3@xmath13 lower limit of [ ] / [ ] @xmath69 1.24 , a high value which is consistent with those predicted and measured for starbursts @xcite . because ob stars can not ionize
, the detection of the 14.3 @xmath1 [ ] emission line would imply the presence of an agn .
although we do not detect this line , our 3@xmath13 upper limit ( @xmath70 w m@xmath68 ) implies [ ] 14.3 @xmath1 / [ ] 12.8 @xmath71 .
it is not uncommon for agn to exhibit similar ratios @xcite , so this limit can not rule out the presence of an agn in source 1 .
the mid - infrared spectrum of source 1 displays a broad silicate absorption trough near 9.7 @xmath1 .
it is difficult to accurately estimate a silicate optical depth for a pah - dominated spectrum because of the uncertainty in determining the continuum level . using a simple procedure , we assume a power - law continuum between the measured flux densities at 5.5 and 14.5 @xmath1 ( rest - frame ) .
the 5.5 @xmath1 continuum anchor point lies sufficiently shortward of the 6.2 @xmath1 pah feature to remain free of pah emission .
in addition , source 1 shows no evidence of water ice or hydrocarbon absorption , which are sometimes seen in ulirg spectra at these wavelengths .
our long wavelength continuum anchor is well beyond both the 12.7 @xmath1 pah feature and the ( unobserved ) 12.8 @xmath1 [ ] emission line . at 14.5 @xmath1 ,
ulirg spectra are dominated by thermal emission from warm , small grains ( see [ sec : continuumslope ] ) .
this procedure yields an optical depth of @xmath72 , presented as a lower limit because of the unknown extinction at the chosen continuum points . adopting the extinction law of @xcite ,
our limit on @xmath73 implies @xmath74 mag .
this large visual extinction may explain why the optical spectrum indicates a different redshift than the mid - infrared spectrum @xcite .
the mid - infrared spectra of galaxies can be modeled as combinations of the spectra of regions ; the photo - dissociation regions ( pdrs ) around regions ; and agn - heated dust @xcite .
as we showed in [ sec : firstpah ] , source 1 is powered primarily by a starburst .
thus , we would expect its mid - infrared spectrum to be dominated by pah - rich pdrs and regions , which typically exhibit steeply rising continua at 1216 @xmath1 due to very small grains with radii less than 10 nm ( vsgs ; * ? ? ?
* ) . in figure
[ fig : absorption ] , we compare the mid - infrared spectrum of source 1 with those of the unobscured galactic reflection nebula ngc7023 ( pure pdr ; * ? ? ? * ) and the prototypical starbursts ngc7714 @xcite and m82 @xcite .
all three spectra are similar at rest - wavelengths shortward of 8 @xmath1 . like most starbursts ,
ngc7714 and m82 have strong vsg continua which set in longward of 8 @xmath1 .
however , the mid - infrared spectrum of source 1 has only a weak vsg continuum .
a large hard x - ray flux can indicate the presence of an agn .
source 1 was not detected in a 5 ks chandra survey of the botes field @xcite .
this non - detection corresponds to an upper limit of @xmath76 for neutral hydrogen column densities of @xmath77 @xmath16 , assuming a galactic column density of @xmath78 @xmath16 and a power law spectrum with photon index @xmath79 .
this limit is consistent with the values of @xmath80 measured for starburst - dominated galaxies @xcite .
however , we can not rule out the presence of a seyfert - like nucleus behind neutral hydrogen column densities of @xmath81 @xmath16 .
the shape of the sed also provides information on the power source .
@xcite presented the optical to radio sed of source 1 , and argued that it is a starburst based upon its cool dust temperature , adherence to the far - infrared radio correlation , and the prominence of the 1.6 @xmath1 stellar bump . using a large spectroscopically - classified sample , @xcite
have developed _ irac _
color criteria for identifying agn .
interestingly , by these standards , source 1 would be classified as an agn
. however , those authors point out that their criteria may result in contamination from @xmath82 ulirgs .
the _ irs _ spectrum of source 2 , shown in the bottom panel of figure [ fig : spectra ] , differs dramatically from that of source 1 .
the only obvious feature is the broad 9.7 @xmath1 silicate absorption band . assuming a power - law continuum between 7.8 and 14.5 @xmath1 ( rest - frame ) , we obtain an optical depth of @xmath83 .
this corresponds to @xmath84 mag @xcite . as with source 1 , these are lower limits due to the unknown extinction at the continuum points .
the placement of the absorption feature also indicates a redshift of @xmath2 .
we obtained a more accurate redshift of @xmath85 from the [ ] 3727 emission line in a keck - deimos optical spectrum ( see figure [ fig : deimos ] ) .
unfortunately , no optical spectroscopic diagnostics of the energetics of source 2 could be identified .
source 2 has an extremely large mid - infrared luminosity . at 5 @xmath1 ( rest - frame )
, it has @xmath86 .
out of the 17 objects with redshifts from the @xcite sample , only one ( sst24j143001.91 + 334538.4 at @xmath87 ) has a comparable 5 @xmath1 luminosity .
the remainder were either not sampled at 5 @xmath1 ( 4 sources ) or are at least a factor of @xmath302 less luminous .
the 5 @xmath1 luminosity of source 2 is also over an order of magnitude brighter than any of the six @xmath88 ulirgs observed with the _ irs _ by @xcite and
at least a factor of three brighter than even the most luminous local ulirgs .
the powerful soft x - ray and ultraviolet radiation associated with an agn may destroy pah carriers ( e.g. * ? ? ?
* ; * ? ? ?
the lack of pah features in source 2 ( @xmath89 @xmath1 ; 3@xmath13 ) suggests that its luminosity is generated primarily by an agn , since the pah equivalent width is a factor of @xmath303040 lower than found in pure starbursts ( brandl et al .
, in preparation ) .
the sed of source 2 is shown in black in the bottom panel of figure [ fig : seds ] .
the 81000 @xmath1 luminosity is poorly constrained by the available data .
however , given its extremely high mid - infrared luminosity , it is likely that source 2 has @xmath62 .
in addition , the optical and near - infrared data do not display a stellar bump , suggesting that this feature is swamped by a hot dust continuum produced by an agn . for reference
, we have overplotted the normalized spectra of f00183 - 7111 @xcite , mrk231 ( * ? ? ?
* armus et al . , in preparation ) , and the nucleus of ngc1068 @xcite .
all three objects are probably agn - dominated with energetically significant circumnuclear star formation , but the agn - heated dust in f00183 - 7111 and mrk231 is deeply obscured , and the ngc1068 spectrum is dominated by nuclear , rather than circumnuclear , emission .
source 2 shows a much smaller silicate absorption than f00183 - 7111 , and its 160 and 350 @xmath1 limits indicate that it has smaller relative amounts of cool dust than either f00183 - 7111 or mrk231 .
its sed is most similar to that of the nucleus of ngc1068 , where we have a direct line - of - sight to the hot dust heated by the agn .
source 2 was not detected in the 5 ks chandra survey , but we are unable to put useful limits on @xmath80 because @xmath55 is poorly constrained .
it was also undetected in the 20 cm first survey @xcite , providing a 5@xmath13 upper limit of 0.98 mjy .
we have presented the _ irs _
spectra of two @xmath2 ulirgs selected from the botes region of the ndwfs .
source 1 has a high 6.2 @xmath1 pah equivalent width , high [ ] / [ ] ratio , and low @xmath90 .
these properties , in combination with the prominent 1.6 @xmath1 stellar bump , cold ( @xmath3043 k ) dust temperature , and concordance with the radio - fir correlation noted by @xcite , indicate that source 1 is dominated by a starburst . with @xmath91 ( modulo its unknown lensing amplification ) ,
source 1 is extremely luminous compared to low - redshift starbursts .
however , pah - dominated spectra of comparably luminous objects are beginning to be discovered at @xmath92 @xcite .
source 1 may be analogous to these more distant luminous starbursts , but is at a redshift where it is more easily studied in detail .
its use as a template for higher redshift populations rests on our ability to understand its lensing properties @xcite , which will require high - resolution imaging to separate out the contribution from the foreground object . at a similar redshift , source 2 is among the most luminous mid - infrared sources known ( as measured at 5 @xmath1 , rest - frame ) .
it appears to be powered mainly by an agn : it has a very small 6.2 @xmath1 pah equivalent width ; exhibits a mid - infrared sed similar to those of agn - dominated ulirgs ; and lacks large amounts of cool dust , as evidenced by its moderate silicate optical depth and non - detections at 160 and 350 @xmath1 .
the first sizable high - redshift samples of dusty sources observed with the _ irs _ have focused on objects at @xmath92 @xcite .
the submillimeter galaxies also lie at this redshift .
relatively less is known about the population of dusty sources at @xmath93 .
this redshift range has been traditionally difficult to access through optical spectroscopy , and has come to be known as the `` redshift desert '' .
indeed , the optical spectra of both objects and the near - infrared spectrum of source 1 show very few emission lines .
a redshift identification for either source would have been dubious without its mid - infrared _ irs _ spectrum .
nevertheless , a large fraction of the stars in local l@xmath94 galaxies were likely formed in this redshift range ( e.g. * ? ? ?
* ; * ? ? ?
objects such as source 1 and source 2 provide examples of how the _ irs _ will allow the identification and detailed study of objects in this critical redshift range .
this work made use of images and data products provided by the noao deep wide - field survey ( * ? ? ?
* dey et al . , in preparation ; jannuzi et al .
, in preparation ) , which is supported by the national optical astronomy observatory ( noao ) .
noao is operated by aura , inc .
, under a cooperative agreement with the national science foundation .
the authors wish to recognize and acknowledge the very significant cultural role and reverence that the summit of mauna kea has always had within the indigenous hawaiian community .
we are most fortunate to have the opportunity to conduct observations from this mountain . ,
s. m. et al .
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 .
1657 - 1669 ( 2003 ) . , 16571669 , g. p. , bautz , m. w. , ford , p. g. , nousek , j. a. , & ricker , g. r. 2003 , in 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 ) . , 2844 | we present low - resolution ( @xmath0 ) mid - infrared ( 838 @xmath1 ) spectra of two @xmath2 ultraluminous infrared galaxies with @xmath3 : mips j142824.0 + 352619 and sst24 j142827.19 + 354127.71 .
the spectra were taken with the _ infrared spectrograph _
( _ irs _ ) on board the _ spitzer space telescope_. both objects were discovered in a _ spitzer_/_mips _
survey of the botes field of the noao deep wide - field survey ( ndwfs ) .
mips j142824.0 + 352619 is a bright 160 @xmath1 source with a large infrared - to - optical flux density ratio .
previous authors provide evidence for a foreground lens and estimate an amplification of @xmath410 , although this factor is presently poorly constrained . the 6.2 , 7.7 , 11.3 , and 12.8 @xmath1 pah emission bands in its _ irs_spectrum indicate a redshift of @xmath2 .
the large equivalent width of the 6.2 @xmath1 pah feature indicates that at least 50% of the mid - infrared energy is generated in a starburst , an interpretation that is supported by a large [ ] / [ ] ratio and a low upper limit on the x - ray luminosity .
sst24 j142827.19 + 354127.71 has the brightest 24 @xmath1 flux ( 10.55 mjy ) among optically faint ( @xmath5 ) galaxies in the ndwfs .
its mid - infrared spectrum lacks emission features , but the broad 9.7 @xmath1 silicate absorption band places this source at @xmath6 .
optical spectroscopy confirms a redshift of @xmath7 .
given this redshift , sst24 j142827.19 + 354127.71 has among the largest rest - frame 5 @xmath1 luminosities known .
the similarity of its sed to those of known agn - dominated ulirgs and its lack of either pah features or large amounts of cool dust indicate that the powerful mid - infrared emission is dominated by an active nucleus rather than a starburst .
our results illustrate the power of the _ irs _ in identifying massive galaxies in the well - known `` redshift desert '' between @xmath8 and in discerning their power sources . because they are bright , mips j142824.0 + 352619 ( pending future observations to constrain its lensing amplification ) and sst24 j142827.19 + 354127.71 are useful @xmath9 templates of a high luminosity starburst and agn , respectively . |
in brief , a bit commitment ( bc ) is the following task that is executed in two steps ( ( a ) and ( b ) below ) by two mistrustful parties , a sender , alice , and a receiver , bob.(a ) commit phase ( c - phase ) : alice chooses a bit ( @xmath0 or @xmath1 ) and commits it to bob .
that is , she gives bob a piece of evidence that she has a bit @xmath2 in mind and that she can not change it ( in this case , the commitment is said to be binding ) .
bob can not learn the value of the committed bit from that evidence until alice reveals further information ( in this case , the commitment is said to be concealing).(b ) opening phase ( o - phase ) : at a later time , alice opens the commitment .
that is , she tells bob the value of @xmath2 and convinces him that it is indeed the genuine bit that she chose during the c - phase .
if alice changes the value , it can be discovered by bob .
a bc is an important cryptographic primitive with many applications in more sophisticated tasks and is of great theoretical and practical interest .
current classical bc protocols are proven secure by invoking some unproven computational assumption ; that is , complexity of some kind of mathematical problems such as the hardness of factoring large integers .
after the invention of the quantum computing algorithm that makes the computational assumption totally invalid , it has been brought to many researchers attention whether there exists a bc protocol that is guaranteed secure solely by physical principles . in recent years , mayers , lo , and chau have proven that an unconditionally secure bc is impossible ( no - go theorem for a bc ) under the standard nonrelativistic assumption.@xcite however , although their discussions are quite correct , their proofs are a bit formal and nonconstructive .
it is not yet clear what prevents us from implementing the unconditionally secure bc protocol . in this paper , we give a constructive proof of the no - go theorem for a bc that would make things more transparent and convincing from the viewpoint of quantum information theory .
we clarify why quantum mechanics does not help a quantum bc protocol to achieve more than a classical one does .
first , let us consider an honest protocol .
the most important point concerning the bc protocol is that alice needs to unveil a value of @xmath2 in the o - phase consistently with information transmitted in the c - phase . from the information - theoretic point of view
, this implies that one bit of classical information should be transmitted from alice to bob at the end of the protocol .
therefore , when we set @xmath3 and @xmath4 as the amounts of information in a bit transmitted in the c - phase and o - phase , the following identity holds : @xmath5namely , the bc protocol is essentially a split transmission of one - bit information in two temporally separated steps : one in the c - phase and the other in the o - phase .
only a fraction of the bit information needs to be transmitted in each step .
noting this fact , we can formulate the quantum bc protocols reported so far@xcite as follows . in order to demand unconditional security
, alice reveals to bob quantum information as a piece of evidence of her commitment by transmitting a system , such as a photon or an electron , in the c - phase . in the o - phase , she reveals to bob classical information which consists of the value of @xmath2 and the measurement basis on the system .
finally , alice and bob test the consistency between the reported value of @xmath2 and the measurement results of the system . according to the quantum description of the protocol involving classical communication suggested by tal mor ,
let subsystem @xmath6 ( bob s system ) be the system with arbitrary dimensional state space @xmath7 that carries quantum information in the c - phase and subsystem @xmath8 ( alice s system ) be the system with arbitrary dimensional state space @xmath9 that carries classical information in the o - phase.@xcite let @xmath10 be the genuine states of the joint system @xmath11 to be prepared by alice according to her choice of @xmath2 . then , eq .
( [ eq1 ] ) is equivalent to the condition that @xmath12 and @xmath13 are orthogonal in the joint hilbert space , @xmath14 ; i.e. , @xmath15here , to avoid confusion throughout this paper , we use superscripts to denote the appropriate state space for a state or an operator .
according to the protocol , by transmitting a subsystem @xmath6 to bob , alice reveals , in general , nonorthogonal marginal states @xmath16(@xmath0 or @xmath1 ) in the c - phase .
it is proven in appendix a that from condition ( [ eq2 ] ) , we can always find two mutually orthogonal purifications @xmath17 of @xmath18 in the orthogonal subspace in which the support of the state @xmath10 lies .
any set of two orthonormal states in @xmath19 should be represented by two orthonormal states in the two - dimensional subspace @xmath20 spanned by @xmath21 in @xmath19 .
furthermore , any such subspace @xmath20 can be defined by a set of two orthonormal states in @xmath19 : @xmath22with an appropriate choice of two sets of orthonormal states @xmath23 and @xmath24 and coefficients @xmath25 and @xmath26 , where @xmath27 @xmath28 makes up a schmidt basis for @xmath9 @xmath29 .
therefore , given a set of the mutually orthogonal purifications @xmath30 in @xmath19 , we can always find the following form of schmidt decomposition@xcite , @xmath31 \left| \psi _
{ 1}^{ab}\right\rangle = -\sin \theta \left| 0^{ab}\right\rangle + \cos \theta \left| 1^{ab}\right\rangle , % \end{array}% \right .
\label{eq4}\ ] ] by choosing appropriate schmidt coefficients and absorbing any phase factors in the definition of the bases . as a result of eqs .
( [ eq5 ] ) and ( eq4 ) , the marginal states @xmath32 and @xmath33 are commutable and diagonalized simultaneously by schmidt basis as @xmath34where @xmath35 denotes a partial trace over subsystem @xmath8 , and two states @xmath36are orthogonal on @xmath7 ; i.e. , @xmath37 .
the forms of @xmath38 and @xmath39 can be freely chosen in the protocol and various complex forms have been proposed to prevent cheating of both parties , but the concrete forms are irrelevant to the subject in the following discussion .
according to the model given in sec [ sec2 ] , we will evaluate the performance of alice s and bob s cheating .
the purpose of bob s cheating is to obtain as much information as possible about @xmath2 during the c - phase from the marginal states @xmath18 . in the following ,
the amount of available information about @xmath2 for bob during the c - phase is evaluated as a measure of his cheating performance . from the protocol agreed by alice and bob ,
the states @xmath10 to be prepared by alice are known to them .
therefore , bob can calculate the schmidt bases , @xmath40 and @xmath41 , that diagonalize the marginal states @xmath42 and @xmath33 beforehand .
bob can perform an optimal measurement for distinguishing @xmath32 and @xmath33 by making use of the orthogonality between @xmath38 and @xmath43 . to describe a measure of the available information about @xmath2 for bob from @xmath18 , consider the fidelity between @xmath32 and @xmath44 .
@xcite it is given by @xmath45noting that @xmath38 and @xmath46 are orthogonal , we can calculate @xmath47 from eq .
( eq6 ) as @xmath48 in the representation in which @xmath38 and @xmath46 are diagonal .
therefore , eq . ( [ eq6 ] ) gives@xmath49 the smaller the fidelity is , the more bob can distinguish between @xmath50 and @xmath33 correctly ; therefore , he can gain more information about @xmath2 . to confirm this
, we consider the quantum error probability which gives the lower limit of error rate for distinguishing @xmath42 and @xmath33.@xcite it is given as @xmath51noting again that @xmath38 and @xmath46 are orthogonal , it follows from eq .
( [ eq6 ] ) that @xmath52 in the representation in which @xmath38 and @xmath46 are diagonal .
thus , it follows that @xmath53let us now introduce distinguishability between @xmath32 and @xmath54 as @xmath55then , the larger the distinguishability is , the more bob can distinguish between @xmath32 and @xmath33 correctly .
thus , the distinguishability gives a measure of the available information about @xmath2 for bob from @xmath18 .
it is easily seen that @xmath56 and @xmath57 satisfy @xmath58therefore , there is a trade - off relationship between @xmath56 and @xmath57 ; that is , the smaller @xmath59 is , the larger @xmath60 is and the more correctly bob can distinguish @xmath32 and @xmath33 , and vice versa .
let us turn to the information - theoretic measure of available information for bob .
mutual information between the value of genuine @xmath2 and the value of @xmath2 that is judged from the measurement of @xmath32 and @xmath54 is an appropriate measure from the viewpoint of information theory .
when alice chooses the value of commit bit @xmath2 between 0 and 1 with equiprobability , this measure depends only on @xmath32 and @xmath54 and is given by @xmath61where @xmath62 is an entropy function ( in bit ) .
the purpose of alice s cheating is to unveil her commit bit @xmath2 at her will in the o - phase while ensuring unveiled @xmath2 does not conflict with bob s measurement of his subsystem @xmath6 revealed by her during the c - phase . from the agreed protocol
, alice can calculate the purification given in eqs .
( eq4 ) and ( [ eq5 ] ) beforehand . in the following ,
her ability to change the commit bit is evaluated for two known cheating strategies as a measure of her cheating performance .
this is a strategy which was first proposed by mayers.@xcite alice honestly reveals either @xmath32 or @xmath33 in the c - phase by transmitting subsystem @xmath6 of the joint system @xmath11 prepared in the arbitrary purification associated to either @xmath32 or @xmath63 . in the o - phase , by a local unitary operation on subsystem @xmath8 in her hand , she can change the joint state into any purification @xmath64 of her chosen @xmath18 that satisfies @xmath65and @xmath66where @xmath67.@xcite then , according to her necessity , she changes the joint state into the fake states , @xmath68here , the state @xmath69 saturates the upper bound of @xmath70 in eq .
( [ eq18 ] ) and is most parallel to the state @xmath71.@xcite she tells bob the basis to be used for his measurement on subsystem @xmath6 that is found from her projection measurement of subsystem @xmath8 by an appropriate basis @xmath72 .
bob performs projection measurement on his subsystem @xmath6 according to her instruction and checks her commitment from the consistency between the value of @xmath2 that is unveiled by alice in the o - phase and his measurement results .
the fake state @xmath73 given in eq .
( [ eq19 ] ) is optimal in mayer s strategy . to confirm this
, we consider the probability @xmath74 that alice causes and bob finds an inconsistency between the unveiled value of @xmath2 and bob s measured data .
it is proven in appendix b that @xmath74 is zero when alice prepares the genuine state @xmath10 or the purification @xmath75 , and @xmath76when she prepares the fake state @xmath77 , where equality holds if and only if @xmath64 lies in the subspace @xmath20 spanned by a set of orthonormal states @xmath78 .
applying eq .
( [ eq18 ] ) to eq .
( [ eq20 ] ) yields @xmath79the state @xmath73 in eq .
( [ eq19 ] ) yields the lower bound @xmath80 for the probability @xmath74 that depends only on @xmath50 and @xmath33 .
therefore , the fake states in eq .
( [ eq19 ] ) give the least possibility of disclosing her cheating to bob and they are optimal for this strategy .
the lower limit @xmath81 is a convenient measure of alice s ability to change her commitment in the o - phase .
it should be noted that mayer s strategy is asymmetric with respect to the value of @xmath2 that alice unveils in the o - phase .
for example , consider alice reveals @xmath32 in the c - phase .
then , if she unveils @xmath0 honestly in the o - phase , bob s measured data on his subsystem @xmath6 is perfectly consistent with her disclosure and @xmath74 is zero .
conversely , if she wants to unveil @xmath82 , she can cheat bob successfully with the probability @xmath83by preparing the fake state @xmath84 . here ,
( [ eq10 ] ) and ( [ eq21 ] ) are used to derive eq .
( [ eq22 ] ) .
thus , the lower limit @xmath80 depends on @xmath2 that alice unveils in the o - phase .
it should be further noted that @xmath85 if @xmath86 .
this means that mayer s strategy can be applicable only when @xmath87 .
now let us turn to the information - theoretic measure of alice s cheating performance .
let @xmath88 be mutual information between the value of @xmath89 that alice unveils and the value of @xmath2 that bob judged from his measurement on his subsystem @xmath6 in the o - phase .
taking into account the asymmetry noted in the previous paragraph , we get the upper bound of @xmath90 as a function only of @xmath32 and @xmath33 as follows : @xmath91here , @xmath92 is considered to be a good information - theoretic measure of alice s ability to change her commit bit for this strategy .
this is a strategy which was first given by koashi and imoto in the context of quantum key distribution@xcite , but later applied to the bc protocol by hardy and kent.@xcite according to this strategy , alice reveals @xmath93 in the c - phase by transmitting the subsystem @xmath6 of the joint system @xmath11 prepared in the arbitrary purification of @xmath94 . when she unveils her commitment in the o - phase , she can change the joint state into any purification @xmath69 of @xmath94 satisfying @xmath95and @xmath96by performing a local unitary operation on her subsystem @xmath8 .
then , according to her choice of @xmath2 , she changes the joint state into the fake state , for example , so that when @xmath97 , @xmath98here , @xmath69 saturates the upper bound of @xmath99 in eq .
( [ eq25 ] ) and is the most parallel to the state @xmath75.@xcite she tells bob the basis to be used for his measurement on subsystem @xmath6 that is found from her projection measurement on her subsystem @xmath8 by an appropriate basis @xmath100 .
bob performs projection measurement on his system according to her instruction and checks her commitment from the consistency between the value of @xmath2 that is unveiled by alice in the o - phase and his measurement results .
it can also be proven from appendix b that the lower bound @xmath101 of the probability @xmath102 in this strategy depends only on @xmath32 and @xmath54 , and it is given by @xmath103the states @xmath69 in eq .
( [ eq26 ] ) yield the lower bound @xmath104 .
therefore , they are optimal for this strategy .
the lower limit @xmath105 gives a convenient measure of alice s ability to change her commitment in the o - phase .
in contrast to mayer s strategy , hardy - kent s strategy is symmetric with respect to the value of @xmath2 that alice unveils in the o - phase .
the lower limit @xmath106 is independent of her disclosure of @xmath2 .
the upper bound of the mutual information @xmath107 for hardy - kent s strategy is written in terms of @xmath106 as @xmath108which is considered to be a good information - theoretic measure of alice s ability to change commit bit @xmath2 in this strategy . to compare the cheating performances of alice and bob for both mayer s and hardy - kent s strategies , we plot the three information theoretic measures @xmath109 , @xmath110 , and @xmath111 in fig . f1 as a function of the fidelity @xmath112 chosen as a common parameter .
this figure clearly shows that there is a trade - off relationship between bob s available information in the c - phase ( @xmath109 ) and alice s ability to change commit bit @xmath2 in the o - phase ( @xmath113 ) .
it is clear that the sum is bounded ; i.e. , @xmath114(for @xmath115 ) .
this equation is a direct consequence of eq .
( [ eq15 ] ) , showing a trade - off relationship between the distinguishability @xmath116 , a measure of bob s information gain in the c - phase , and the fidelity @xmath112 , a measure of alice s ability to change commit bit in the o - phase .
therefore , there is a trade - off in the performance of alice s and bob s cheating .
figure [ f1 ] also shows that hardy - kent s strategy is superior to mayer s with respect to ability to change commit bit @xmath2 when the value of @xmath56 is large .
a secure bc protocol must not allow cheating by either parties , alice or bob . to satisfy this condition , both @xmath117 and @xmath113 should vanish simultaneously . however , fig .
[ f1 ] indicates that this requirement is never satisfied because of the trade - off relationship between @xmath118 and @xmath113 .
in addition , even if we choose a balanced condition for both parties , @xmath119 , both @xmath118 and @xmath113 are already large enough .
therefore , it is concluded that a law of quantum physics does not help to improve the security of the bc protocol .
as already proven generally by mayers , lo , and chau , this conclusion should be valid not only for the particular cheating strategies described in this paper but also for any strategies that alice and bob can choose . to understand this conclusion ,
consider the entropy of entanglement ( or entanglement in brief ) , which is known to be a unique measure of the amount of entanglement for the pure state.@xcite entanglement of the purification @xmath17 is defined as the von neumann entropy of the marginal state @xmath18 of @xmath71 or equivalently as the shannon entropy of the squares of the schmidt coefficients of @xmath75 . from eqs .
( [ eq4 ] ) , ( [ eq5 ] ) , ( [ eq13 ] ) , and ( [ eq16 ] ) , it is easily calculated as @xmath120here , it should be noted that @xmath117 is equivalent to maximum information @xmath3 available from @xmath18 that is transmitted from alice to bob in the c - phase ; that is , @xmath121applying eqs .
( [ eq1 ] ) and ( [ eq29 ] ) to ( [ eq31 ] ) , we obtain the inequality , @xmath122this inequality implies two things .
first , the performance of alice s cheating , when it is measured by @xmath113 , is bounded by the entanglement of the purification @xmath123 , which is determined only by its marginal state @xmath18 ( see eq .
( [ eq30 ] ) ) .
second , it is also bounded by the amount of information @xmath4 that is revealed in the o - phase .
these implications are reasonable for the following reason .
when alice wants to cheat bob , what she can do is restricted to local operation and measurement on the subsystem @xmath8 in her hand after she has revealed @xmath124 by transmitting the subsystem @xmath6 .
it is known to be a fundamental law of quantum information processing that entanglement can not be increased if we are allowed to perform only local operations and subselection on the subsystem of a joint system . in this restricted situation , the best she can do to cheat is use the local unitary operation that conserves the entanglement shared in the joint system and keeps the marginal states @xmath124 unchanged .
otherwise , the strategy must be by far an optimal one because a fraction of the entanglement must be lost from the joint system and dissipate into the environment during the local operation .
under such circumstances , alice can change the information content encoded only in the relative phase between coefficients of each term in the purification @xmath125 , but she can not change the information content encoded in their absolute values . it is the entanglement resource that is responsible for alice s cheating , and there is no cheating strategy that can break the bound given by entanglement @xmath123 as shown in eq .
( [ eq32 ] ) .
in addition , it is also reasonable that only partial information that is to be revealed in the o - phase can be used for alice s cheating but the partial information already revealed in the c - phase can not .
conversely , we must be aware that alice makes use of partial information that is reserved to be revealed in the o - phase as an entanglement resource for cheating .
it is worth noting that the present proof can be regarded as a concrete example of the general proof of the no - go theorem for a zero - knowledge - convincing protocol recently given by horodecki et al.horodecki our proof clearly indicates that if alice wants to convince bob that she has a definite value of a commit bit ( which is , of course , _ classical _ information ) in mind in the c - phase , the information provided by her to him in the c - phase has to carry nontrivial information about the commit bit in her mind . if the information revealed in the c - phase is independent of her commit bit , alice can always try to cheat by proposing the test which would give some result with certainty and independently of her commit bit in the o - phase .
our proof suggests an information - theoretic ground for the no - go theorem of the zero - knowledge - convincing protocol .
namely , any protocol with a test message that convinces bob that alice knows some state @xmath126 ( which is , in general , _ quantum _ information ) , the test message has to carry non - zero information about state @xmath126 to prevent alice s cheating .
the present proof implies that the conjecture of mayers about two - party secure computation , which states that the symmetric protocol might be possible whereas the asymmetric tasks , such as unidirectional secure computations , would be impossible , is correct.@xcite in the unidirectional two - party computation , which allows only one of the two parties to learn the result , both members of the party can be a cheater and security requirements for both members are incompatible .
such unidirectional protocols under the standard nonrelativistic assumption are necessarily insecure .
we believe that unidirectional quantum communication does not achieve more than classical communication alone in the two - party model .
however , it has not yet been proven that no non - trivial cryptographic tasks in the two - party model using bidirectional quantum communication are unconditionally secure . indeed
, there are some proposals on the quantum protocols for non - trivial weaker tasks in two - party bidirectional quantum communication such as quantum coin - tossing@xcite and quantum gambling.@xcite it will still be important to solve the general problem concerning what is possible and what is impossible in two - party secure computation when unproven computational assumptions are abandoned .
in conclusion , we have given constructive proof why an unconditionally secure quantum bc is impossible in the light of quantum information theory .
the bc protocol is in essence the protocol in which one - bit information is split and revealed in two temporally separated steps : the c - phase and the o - phase .
it ensures only a fraction of the bit information is revealed at a time . in the quantum bc protocol ,
increasing the information revealed in the c - phase is to bob s advantage ; conversely , increasing the information revealed in the o - phase makes things to alice s advantage .
this situation is similar to the classical protocol .
furthermore , the protocol that is unbiased to both alice and bob is not secure for both . therefore , it is impossible to design a bc protocol whose security is established solely on the law of quantum physics .
in addition , it has been clarified that , alice can make use of the entanglement resource , which is equal to the amount of information reserved to be revealed in the o - phase , to cheat . to prevent alice s cheating ,
the information revealed in the c - phase must depend on her commit bit , and it must inevitably carry non - zero information about her commitment . it can be concluded that quantum mechanics itself makes designing an unconditionally secure bc protocol impossible .
suppose that the states @xmath10 ( @xmath128 ) of joint system @xmath11 that is to be prepared by alice are mutually orthogonal on the joint space @xmath129 ; i.e. , @xmath130because @xmath131 and @xmath13 commute , they can be diagonalized simultaneously in terms of orthonormal bases @xmath132 and @xmath133 in @xmath19 as follows : @xmath134where @xmath135 and @xmath136 are mutually orthogonal ; i.e. , @xmath137 , and @xmath138 and @xmath139 are sets of real eigenvalues of @xmath10 satisfying @xmath140 and @xmath141 .
thus , @xmath12 and @xmath13 have orthogonal supports in @xmath19 .
marginal states revealed by alice to bob in the c - phase are commutable and , in general , nonorthogonal states . using this representation
, we can write them as @xmath142now , we consider mutually orthonormal states @xmath143 and
@xmath144 @xmath145 that lie in the subspace to which @xmath131 and @xmath13 belong respectively ; i.e. , @xmath146then , the marginal states for them are @xmath147because the states of subsystem @xmath8 represent the classical information transferred from alice to bob in the o - phase , different states @xmath148 and @xmath149 are orthogonal on the subspace @xmath9 . therefore , @xmath150by noting that the second terms in eq .
( [ a5 ] ) vanishes , it is concluded that by choosing @xmath151 and @xmath152 so that @xmath153it is always possible to obtain mutually orthogonal purification @xmath125 of @xmath154 in the subspace in which the support of the state @xmath10 lies .
suppose that the state of subsystem @xmath8 is measured to be @xmath155 when @xmath8 of the joint system @xmath11 prepared in the state @xmath156 is subjected to projection measurement by the orthonormal basis @xmath72 for @xmath9 .
according to general results of quantum measurement theory , the state of the subsystem @xmath6 is projected onto the pure state @xmath157from eqs .
( [ eq4 ] ) and ( [ eq5 ] ) , it is easily seen that @xmath158therefore , we find that , if and only if the basis @xmath100 is chosen so that the overlap between @xmath159 and @xmath160 and that between @xmath161 and @xmath162 are the same , i.e. , @xmath163the states @xmath164 and @xmath165 become mutually orthogonal . in the quantum bc protocol , alice and bob agree to use the measurement basis @xmath166 on subsystem @xmath6 that has a one - to - one correspondence to a state @xmath155 on @xmath8 through the joint state @xmath167 .
she reveals to bob the measurement basis @xmath168 associated with her state @xmath155 in the o - phase , and he measures his subsystem @xmath6 by this basis .
now , we consider the probability @xmath74 that alice causes an inconsistency between the value of @xmath2 that is unveiled by her in the o - phase and bob s measured data when alice prepares an honest state @xmath169 .
alice projects her subsystem @xmath8 of the joint system @xmath11 prepared in @xmath10 onto a state @xmath161 among the complete orthonormal basis @xmath170 for space @xmath9 .
she can perform such a projection on her subsystem at her own free will .
correspondingly , the state of bob s system is projected to be @xmath171here , from appendix a , the states @xmath172 and @xmath10 satisfy @xmath173then , we obtain the identity @xmath174 .
this identity implies that if bob follows alice s instruction and measures his system by the measurement basis given by her , the value of @xmath2 unveiled by alice in the o - phase is perfectly correlated with bob s measurement result , no matter what alice prepares @xmath172 or @xmath10 .
therefore , if bob is honest enough to follow alice s instruction , the probability @xmath102 that bob finds an inconsistency in his data vanishes if alice prepares @xmath175 or @xmath10 .
consequently , alice can transmit one bit of classical information to bob with certainty .
next , we consider the probability @xmath74 when alice prepares a fake state @xmath176 that lies in joint space @xmath177 . when the joint system @xmath11 prepared in the state @xmath178
is subjected to the projection measurement by using the orthonormal basis @xmath100 for @xmath9 and the result is @xmath155 , the state of the subsystem @xmath6 is projected onto the pure state @xmath179let the fidelity between @xmath180 and @xmath181 be @xmath182 .
then , the probability @xmath74 that bob finds an inconsistency in his data is given by @xmath183under the condition of eq .
( [ b3 ] ) , it follows that @xmath184@xmath185where @xmath186 is the projector onto two - dimensional subspace @xmath20 in @xmath19 that is spanned by a set of orthonormal states @xmath78 , and @xmath187 is the overlap between the state @xmath155 in @xmath9 and subspace @xmath20 . the equal sign in inequality ( [ b9 ] ) holds if and only if state @xmath188 lies within subspace @xmath20 . from eqs .
( [ b1]),([b6]),([b8 ] ) , and ( [ b9 ] ) , we obtain @xmath189applying eq . ( [ b10 ] ) to eq .
( [ b7 ] ) , we finally obtain @xmath190here , equality holds if and only if the state @xmath188 lies within subspace @xmath20 .
g. brassard , c. crpeau , r. jozsa , and d. langlois , in _ proceedings of the 34th annual ieee symposium on the foundation of computer science _
( ieee computer society press , los alamitos , california 1993 ) , p. 362 - 371 . | we give a comprehensive and constructive proof of the no - go theorem of a bit commitment given by mayers , lo , and chau from the viewpoint of quantum information theory .
it is shown that there is a trade - off relation between information acquired by bob during the commitment phase and the ability to change a commit bit by alice during the opening phase .
it is clarified that a protocol that is unbiased to both alice and bob can not be , at the same time , secure against both parties .
fundamental physical constraints that govern this no - go theorem are also discussed . |
a canonical quantum theory for the gravitational field is performed starting from an hamiltonian formulation , hence from the definition of a space - time splitting .
a criticism to this approach is that it is based on the identification of a time - like vector field ; but it is expected that , once the geometry has been quantized , only a probabilistic response can be given for a vector to be time - like or space - like . a physical way to introduce a space - time splitting of the space - time is to include in the dynamics a real fluid . by virtue of the matter - time dualism @xcite , the introduction of matter provides a solution to the time - problem in quantum gravity @xcite . here
the link between the identification of a reference and the appearance of a matter term is investigated from the point of view of lagrangian symmetries .
in particular , we are going to discuss the dynamical implications of the covariance principle when it is restricted to a synchronous frame of reference .
the hamiltonian constraints will be inferred , showing how fixing the reference leads to the materialization of a dust - like contribution .
finally , the description of a generic cosmological model will be given , together with the picture of a quantum to classical transition for the adopted evolutionary approach .
a synchronous ( gaussian ) reference frame is characterized by a metric tensor having the following fixed components @xmath0 and @xmath1 ( @xmath2 ) , as soon as a 3 + 1 splitting @xmath3 of the full space - time manifold is performed .
this feature implies @xmath4 and @xmath5 for the lapse function and for the shift vector , respectively , in the adm formalism .
it can be shown that the synchronous character is preserved by the following set of restricted coordinate transformations @xcite , @xmath6 @xmath7 and @xmath8 being three generic space functions . by restricting the invariance requirement for the lagrangian ,
one finds the following set of constraints , replacing the super - hamiltonian and the super - momentum ones , @xmath9 @xmath10 being a scalar density of weight @xmath11 , hence it can be written as @xmath12 , with @xmath13 a scalar function .
these expressions outline that , while the super - momentum constraint is not modified , in fact the invariance under 3-difeomorphisms still stands , the super - hamiltonian is non - vanishing , in order to preserve the synchronous character of the reference . the modification induced on the form of constraints
can be interpreted as the emergence of a dust fluid co - moving with the slicing , whose energy density is given by the function @xmath13 .
einstein equations underlying such a picture read @xmath14 @xmath15 being orthogonal to the space - like slicing .
one can think at this contribution as the physical realization of the synchronous reference .
however , it is clear that we are not dealing with an external matter field since the energy density @xmath13 is not always positive and @xmath16 is fixed , once initial conditions are assigned on a non - singular hypersurface .
we perform the quantization of the synchronous gravitational field by promoting the 3-geometry @xmath17 and its conjugated momentum @xmath18 to operators acting on the state function @xmath19 in a canonical way , _
@xmath20 the constraints are implemented according with the dirac prescription , hence they read as follows @xmath21 the condition ( [ 2 ] ) implies that the wave functional depends on the 3-geometries @xmath22 , not on 3-metrics .
but a relevant achievement is the relation ( [ 1 ] ) : @xmath23 being the generators of time displacements , it fixes an evolutionary character for wave functional , which can be described by the schrdinger equation @xmath24 therefore , the quantum features of the dust contribution outline its behavior as a clock - like matter .
the next task is to find out a negative portion of the super - hamiltonian spectrum , which allows to interpret the additional contribution as a physical matter field . in this respect ,
a change of variables is performed on the 3-metric and it is written as @xmath25 with @xmath26 and @xmath27 . + this way the hamiltonian density takes the form @xmath28 @xmath29 and @xmath30 being the conjugate momenta to @xmath31 and @xmath32 , respectively . here , the potential term @xmath33 comes from the 3-ricci scalar and @xmath34 refers to first and second order spatial gradients .
in this scheme , equations ( [ 1 ] ) can be written as @xmath35 @xmath36 this equation has a klein - gordon - like structure , hence there exist solutions with negative values of @xmath10 .
it is worth noting that in a synchronous reference the metric determinant always vanishes monotonically at a time @xmath37 where all the geodesics lines cross each other , i.e @xmath38 , with @xmath39 ( as a consequence of the landau - raichoudhuri theorem ) .
this property supports i)@xmath40 as an internal time , ii ) to take the limit @xmath41 , where the system ( [ newhaux ] ) admits an asymptotic solution .
in fact , in this limit , the potential term turns out to be drastically suppressed with respect to the term @xmath42 , thus the dynamics of different spatial points decouples and the quantization can be performed in a local minisuperspace . + an approximated solution
is provided in this scheme , point by point in space , by the following expression @xmath43 @xmath44 and @xmath45 satisfying the two equations respectively @xmath46 the potential results to be negligible , also on a quantum level , when the condition below holds @xmath47 a wider set of wave packets are in agreement with this condition , approaching @xmath48 .
for instance , one of them is characterized by a deviation @xmath49 , where @xmath50 is a small uncertainty around the picked value @xmath51 ( @xmath52 ) .
indeed , this approach works only if @xmath51 is greater than a fiducial one @xmath53 and this is consistent with the prescription that a quantum - classical correspondence stands for high quantum numbers . by writing @xmath54 ,
the function @xmath55 obeys the following equation for the negative part of the spectrum @xmath56
@xmath57 a solution is given for @xmath58 by @xmath59 @xmath60 being the corresponding bessel functions , while @xmath61 and @xmath62 are two integration constants .
therefore , a quantum range of positive energies exists for the dust contribution near enough to the `` singular '' point @xmath63 .
however , there remain some open points : i ) the existence of a stable ground level of negative @xmath10 .
ii ) the inclusion of spatial gradients of the dynamical variables .
iii ) the physical meaning of the limit @xmath64 .
iv ) the behavior of @xmath10 in the classical limit .
most of these issues can be solved in a generic inhomogeneous cosmological setting , where the 3-metric is given by @xmath65 with @xmath66 and @xmath67 six scalar functions and @xmath68 a @xmath69 matrix .
the dynamics of different points decouples near the singularity and the schrdinger functional equation splits to the sum of @xmath70 independent point - like contributions as follows @xcite ( we denote by the subscript @xmath71 any minisuperspace quantity ) @xmath72\psi _ x - \frac{3\hbar ^2}{8\pi } e^{-3\alpha } \partial ^2_{\varphi } \psi _ x - \nonumber\\ - \left ( \frac{1}{2k\mid j\mid ^2}e^{\alpha } v(\beta _ { \pm } ) - \frac{\lambda } { k}e^{3\alpha } \right ) \psi _
\psi _ x = \psi _ x ( t , \ ; \alpha , \ ; \beta _ { \pm } , \ ; \varphi ) \ , , \end{aligned}\ ] ] where a cosmological constant @xmath73 and a scalar field @xmath74 have been added to the dynamical description .
if an integral representation is taken for the wave function @xmath75 @xmath76 where @xmath77 is fixed by the initial conditions at @xmath78 , the dynamics is given by @xmath79 \xi _ x = \mathcal{e}_x\xi _ x \ , .
\end{aligned}\ ] ] in the expression above , it has been neglected any dependence of @xmath80 on @xmath81 because , asymptotically to the singularity ( @xmath82 ) , it has to be of higher order ( adiabatic approximation ) .
let us now consider wave packets which are flat over the width @xmath83 ( @xmath84 being the standard deviation in the momenta space ) . in the new variable @xmath85 ,
the equation ( [ eigenvp2 ] ) reads @xmath86 a solution to equation ( [ ph9 ] ) is provided by @xmath87 @xmath88 being the characteristic length associated to the universe `` energy '' , while @xmath89 denotes the planck scale length .
however , the validity of the solution above requires the condition @xmath90 .
+ hence , the quantum dynamics in a fixed space point ( _ i.e. _ over a causal portion of the universe ) is described , in the considered approximation ( @xmath91 ) , by a free wave - packet for the variables @xmath92 and @xmath74 and by a profile in @xmath93 which has a maximum in @xmath94 .
if a lattice structure for the space - time is assumed on the planckian scale @xcite , we can take @xmath95 and , noting that @xmath96 , we obtain @xmath97 furthermore , to preserve the reality of @xmath10 , we have to impose that the square root above is real , hence we get @xmath98 . finally , from the request @xmath99 , then the inequality ( [ fininequa ] ) can be written as @xmath100 @xmath101 being the planck mass .
therefore , the existence of a cut - off implies that a ground state exists for the evolutionary approach .
hence it is a natural request to assume the universe to approach this state during its evolution . in order to infer the influence of the dust contribution on the cosmological evolution , we evaluate the associated critical parameter .
being the super - hamiltonian a constant of motion and assuming the universe to be in the ground state , we find @xmath102 , with @xmath103 the present universe radius of curvature . since the actual critical density can be expressed as @xmath104)$ ] ( being @xmath105 the total universe critical parameter ) , then we have @xmath106 therefore , the dust contribution can not play the role of dark matter .
this conclusion can be modified by the introduction of matter from the thermal bath , which could have a relevant interaction with the universe in the trans - planckian region ( see the model addressed in @xcite , where ultrarelativistic matter and a perfect gas were included ) .
we are going to discuss if a scenario can be defined where the transition to a classical regime happens , before the spatial curvature contribution becomes relevant .
this way , the obtained spectrum of @xmath10 would result to be a good approximation during the whole cosmological evolution .
where in the last equation it has been used the identification @xmath112 , coming out from the first one ( which is the hamilton - jacobi equation for the considered system ) .
it is worth noting that the evolution of the wave function @xmath113 is described by a schrdinger equation .
the dynamics for @xmath113 is that of a 3-dimensional non - relativistic free particle with the time variable given by @xmath116 .
hence , if at the initial time @xmath113 is in the form of a sufficiently narrow packet , then it will remain peaked over the whole regime where our approximations hold .
therefore , within the adopted evolutionary approach , _ it can be inferred a quasi - classical description for the universe volume , together with a quantum - like schrdinger dynamics for the anisotropies and the scalar field_. this feature demonstrate that a quantum - classical transition can be defined , where the dust contribution although being relevant during the planckian regime , nevertheless it becomes negligible in the classical phase .
an open point remains the transition to a classical description for anisotropies and the scalar field . in this respect
it could be significant the scalar curvature , which here has been neglected by the approximation scheme .
in fact , such potential terms must approach a quadratic dependence on these variables , which allows to construct stable coherent states . | the implications of restricting the covariance principle within a gaussian gauge are developed both on a classical and a quantum level .
hence , we investigate the cosmological issues of the obtained schrdinger quantum gravity with respect to the asymptotically early dynamics of a generic universe . a dualism between time and the reference frame fixing is then inferred . |
the standard icosahedral projection is defined by the particular embedding of the icosahedral group @xmath0 into the hyperoctahedral group @xmath1 upon subducing the defining representation @xmath2 of @xmath3 to @xmath4 , the reduction to irreducible form is obtained as @xmath5}(g)+d^{[31 ^ 2_-]}(g)$ ] , where @xmath6 is chosen @xmath7,\ ] ] and @xmath8 is the golden ratio , @xmath9 .
the symbol @xmath10 stands for @xmath11 .
the columns of this matrix form six orthogonal basis vectors of @xmath12 .
the first three rows of the matrix @xmath6 give the six basis vectors of @xmath12 , @xmath13 , `` icosahedrally projected '' to the `` parallel space '' , @xmath14 ( 3d space of the irreducible representation @xmath15}(g)$ ] ) , and the next three rows give the six vectors `` icosahedrally projected '' to the `` orthogonal space '' ( 3d space of the irreducible representation @xmath16}(g)$ ] ) , @xmath17 .
the weyl group of the lattice @xmath18 , @xmath19 is a subgroup of the hyperoctahedral group , @xmath20 .
the icosahedrally projected points of the @xmath21lattice into @xmath22 we call @xmath23module .
it can be shown that the @xmath24 transformation of @xmath18 points such that it commutes with @xmath4 reduced to @xmath25 and acts in @xmath22 as a minimal stretching is @xmath26 . \ ] ] this matrix is written in the @xmath27 basis . when referred to a @xmath18 basis
it becomes an element of @xmath24 .
the minimal stretching in ogawa s sense , such that the quasilattice points are mapped into quasilattice points , is a factor @xmath28 , and @xmath29 transformed to @xmath25 is @xmath30 where @xmath31 is a projector to @xmath22 and @xmath32 to @xmath33 .
the basis of the 2f module can be chosen as for example @xmath34 the coefficients in each of the three directions are in @xmath35 $ ] , so it is a @xmath35$]module with inflation factor @xmath8 ( the smallest stretching factor by which module points are mapped into module points ) . the acceptance domain , i.e. the vertex window @xmath36 of the tilings 2f @xcite is the voronoi domain ( @xmath37 ) of the root
lattice @xmath18 icosahedrally projected to @xmath38 , @xmath39 . in @xmath38 ,
its outer shape is a triacontahedron with the edge length @xmath40 ( parallel to the @xmath41fold symmetry axes of an icosahedron ) , @xmath42 .
the vertices of the voronoi domain of the root
lattice @xmath18 are the three translationally nonequivalent
holes"@xcite of the @xmath18 lattice .
we denote them @xcite by @xmath43 , @xmath44 and @xmath45 . after the projection of the voronoi domain to @xmath38 , the vertices of the window @xmath46 are of type @xmath43 and @xmath45 , marked by black and white balls ( figure 1 ) respectively .
the vertices of the tiles ( prototiles ) in the class of tilings 2f are the @xmath18 root lattice points projected to @xmath22 .
the tilings are obtained by the `` cut and project '' method : cut the 3d boundary @xmath47 of @xmath37 , project its dual boundary @xmath48 to @xmath49 .
@xmath50 is coded by @xmath47 projected to @xmath38 , @xmath51 .
all codings are the projected boundaries of @xmath37 ( that we together with @xmath37 call the voronoi complex ) to @xmath38 .
dual boundaries are the 3d boundaries of delaunay cells . in general :
all tiles , faces , edges , vertices , vertex configurations from the tilings in @xmath49 are coded by corresponding dual objects in @xmath33 , i.e. by the voronoi complex projected to @xmath33 .
consequently , all properties of the tilings in @xmath49 can be determined in @xmath38 . this projection species " 2f are equivalent @xcite to de bruijn s strip projection species of danzer @xcite .
the prototiles of the class of nonsingular locally isomorphic tilings 2f are six tetrahedra with the edges parallel to the directions of 2fold symmetry axes of an icosahedron .
they have two lengths , the standard @xmath52 and @xmath53 .
the tiles are denoted @xcite by @xmath54 and @xmath55 .
they are coded in @xmath38 by corresponding dual boundaries @xmath56 and @xmath57 , respectively . in figure 1 we denote the tiles by @xmath58 . in the paper we are not going to mention explicitly @xmath59 any more , so we simplify the notation , instead of @xmath50 we write @xmath60 .
the tiles ( prototiles ) @xmath61 and @xmath62 have four kinds of faces @xmath63 , @xmath64 that are coded in @xmath38 by the corresponding ( dual ) 4boundaries of the voronoi domain of the @xmath18lattice , @xmath65 , @xmath64@xcite .
all faces @xmath63 of tetrahedra are clearly triangles , from now on denoted by @xmath66 .
@xmath67 is an equilateral triangle with edge length @xmath68 , @xmath69 is also equilateral with edge length @xmath70 .
these faces are perpendicular to the directions of 3fold symmetry axes of an icosahedron . @xmath71 and @xmath72 are the golden triangles : @xmath71 with one edge @xmath73 and two @xmath68 , @xmath72 with two edges @xmath53 and one @xmath68 .
these faces are perpendicular to the directions of 5fold symmetry axes of an icosahedron .
the face and edge content of the tiles is presented by the table @xmath74 let us mention three important characteristics of the tilings 2f .
the tilings 2f are dissectable into the tilings 4 @xcite .
the tilings 4 are particular 2d triangular tilings of a plane by two golden triangles @xmath71 and @xmath72 @xcite . to a fixed 5fold direction of the triacontahedron ( the vertex window for the 2f tilings , @xmath75 )
there is a corresponding decagonal belt of height @xmath76 , see figure 2 .
let a plane orthogonal to the chosen 5fold direction intersect the belt .
this intersection is a decagon , the vertex window for the tilings 4 .
the same plane intersects with a set of corresponding projected to @xmath33 are such that in the 3d tilings they code the edges and the 2d faces , all orthogonal to the chosen 5fold direction . ]
5boundaries ( @xmath77 , @xmath78 ) and 4boundaries ( @xmath79 , @xmath80 ) of @xmath81 .
these intersections form the coding of the edges and tiles of the tilings 4 . as a consequence , in @xmath22
there appear the planar ( 2d ) tilings 4 which tiles are faces of the 3d tilings 2f .
consider the set of all these 2d tilings 4 perpendicular to the chosen axes .
the height of the belt allows ( as a necessary condition ) for a spacing of these tilings at fibonacci distances @xmath82 and @xmath83 , @xmath84 the same holds true for all 5fold directions .
a projection class of tilings by the mosseri sadoc tiles @xcite , can be locally derived from 2f @xcite . the inflation
deflation class of mosseri
sadoc tilings@xcite has been ( not accurately ) defined by the substitution rules for non decorated tiles @xcite .
we will come again to this property .
the third property of the tilings 2f is that also the tilings @xcite can be locally derived from 2f @xcite .
an inflation procedure in a tiling requires two steps : \(i ) one stretches all the edges of the tiles by the same factor , ( ii ) one replaces each stretched tile by the tiles of the previous size .
let us call the result the inflated tiling .
the rules which tell how one replaces each of the tiles are the * inflation rules*. in order to find the inflation rules for the tilings obtained by the projection we determine for each stretched representative tile ( the tile coded by the representative dual in @xmath33 ) all its possible replacements by the original tiles .
the analysis is done in the orthogonal space . when the tiles are stretched in @xmath22 by the factor @xmath28 , the window in @xmath33 , @xmath85 and
all its coding polytopes ( the sub boundaries of the voronoi domain projected to @xmath33 ) are transformed by -@xmath86 , i.e. we act by the inflation matrix for the module in 6d space by @xmath87 . by construction
the inflation leaves the projection species ( the tilings obtained by the cut and project " method ) 2f invariant . in order to determine all possible coverings of the stretched tiles by the original tiles
, it turns out , like in a planar tilings 4 @xcite to be enough to study how the 36 vertex configurations@xcite go into vertex configurations under the inflation .
a representative vertex configuration ( up to the icosahedral symmetry ) is a set of tiles which fills the solid angle around a vertex point . the derivation of vertex configurations proceeds as follows : any tile @xmath88 in the tilings has a coding polytope @xmath89 located inside the window @xmath90 of the tiling .
@xmath90 can be decomposed into intersections @xmath91 .
any non empty intersection determines a vertex configuration @xmath92 .
one must analyze only the intersections up to the icosahedral symmetry .
for this aim we choose in @xmath33 a fundamental sector w.r.t .
the icosahedral group from the triacontahedron , see figure 3 , transform it by @xmath93 , obtain the transformation law of all 36 vertex configurations under the inflation , and get the following map @xmath94 we show that an inflation of the tilings exists . in order to define uniquely
how each tile of 2f inflates we find that one has to put arrows on each of them such that they break its symmetry .
one could introduce the arrows on all edges of the tiles such that the common edges for the tiles have a common orientation of an arrow .
moreover , for two tiles @xmath62 and @xmath95 we have to introduce two different colours , let them be blue and red ( @xmath96 , @xmath97 , @xmath98 and @xmath99 ) . the substitution matrix @xmath82 for the tiling 2f is presented as a table @xmath100 and is to be read as for example : @xmath101 , the tile @xmath102 stretched by @xmath8 , is covered by the original tiles @xmath103 , @xmath104
, @xmath105 , @xmath23 , @xmath106 , @xmath107 .
both the arrows on edges that break the symmetry of the tiles and the different colours are coded in @xmath33 .
the inflation is not a stein inflation " , the union of the tiles that cover a stretched tile needs not to have the shape of the stretched tile itself .
the tiles @xmath6 , @xmath108 and @xmath2 inflate by a stein inflation , but not the tiles @xmath102 , @xmath95 and @xmath62 .
the faces @xmath109 and @xmath110 ( golden triangles ) inflate as in the tilings 4 @xcite .
the face @xmath111 ( small equilateral triangle ) transforms into @xmath112 , @xmath113 ( big equilateral triangle ) .
the face @xmath112 ca nt be replaced by the union ( in the same plane ) of other faces .
the tiles @xmath102 , @xmath95 and @xmath62 , having each at least one face @xmath112 , do nt have a stein inflation . in figure 4
we show the stein inflation of the tiles @xmath6 , @xmath2 and @xmath108 , and in figures 5 till 8 the inflation of the red tile @xmath99 , of the tile @xmath102 and the red tile @xmath97 , of the blue tile @xmath98 and finally of the blue tile @xmath96 .
this inflation rules define the inflation
deflation species , the inflation deflation class of tilings by six , or rather eight decorated tetrahedra @xmath114 and @xmath115 @xcite . the question is whether the class of tilings defined by inflation deflation ( inflation deflation species @xcite ) is equivalent to the class of tilings 2f obtained by projection ( projection species @xcite ) . from the inflation rules for the prototiles of the class of tilings 2f we have derived @xcite the inflation rules for the prototiles of the projection species @xcite .
it turns out that the mosseri
sadoc prototiles have also to be decorated in order to define the inflation procedure uniquely @xcite
. the tilings by decorated tiles are not only locally derivable from the projection class 2f @xcite , but equivalent to the class 2f .
moreover , the derived inflation rules for the projection class are indeed the same as for the inflation species introduced by mosseri and sadoc ( up to a nonexisting decoration)@xcite .
we determine the volume inflation matrix @xmath116 for the tiling 2f : @xmath117 in the tabular representation of the matrix @xmath116 the symbol @xmath118 stands for vol@xmath118 .
let us write it as a matrix equation @xmath119 then the frequencies of the tiles , @xmath120 fulfill the equation @xmath121 so we can test the inflation rules .
the frequencies calculated from the volumes of the corresponding codings in @xmath33 are @xmath122 the vector with these entries is indeed an eigenvector of the matrix @xmath123 with the eigenvalue @xmath124 , hence fulfills equation ( 10 ) .
one eigenvalue of the matrix @xmath116 is @xmath125 .
the other eigenvalues fulfill the equation @xmath126=0\ ] ] and are @xmath127 , @xmath128 , @xmath129 , @xmath130 ; @xmath131 and + @xmath132 where @xmath133^{1/2 } \right\}^{1/3 } $ ] .
hence @xmath134 and @xmath135 .
financial support of deutsche forschungsgemeinschaft is gratefully acknowledged .
we also thank the geometry center at the university of minnesota for making geomview freely available .
p. kramer , z. papadopolos and d. zeidler , `` symmetries of icosahedral quasicrystals '' , in proc .
of the symp .
_ symmetries in science v : algebraic structures , their representations , realizations and physical applications _ , ed .
b. gruber et al . , plenum press 1991 , pp . 395427 r. mosseri and j. f. sadoc , `` two and three dimensional non periodic networks obtained from self similar tiling '' , _ the physics of quasicrystals _ , eds .
p. j. steinhardt and s. ostlund , world scientific 1987 , pp .
72034 z. papadopolos and o. ogievetsky , on inflation rules of mosseri sadoc tilings " , preprint ( to be presented at the _ 7th international conference on quasicrystals _ ) , stuttgart , germany , 20 - 24 september 1999 | the tiles of the canonical tilings 2f are six tetrahedra .
we determine their inflation rules by the projection method .
dedicated to ludwig danzer on the occasion of his 70th birthday |
the presence of competing interactions and multiple electronic instabilities often leads to emergent phenomena and new phases .
in particular , the complex relationship between superconductivity and magnetic or charge order has attracted considerable interest . in many of the high temperature iron pnictide superconductors ,
the superconductivity occurs in the vicinity of spin - density - wave ( sdw ) order @xcite .
meanwhile upon applying pressure to the heavy fermion superconductor cecu@xmath0si@xmath0 @xcite , evidence was found for superconductivity being in close proximity to multiple instabilities , with one superconducting dome in the vicinity of magnetic order and another near a possible valence instability @xcite .
the interplay between charge - density - wave ( cdw ) order and superconductivity has also been of particular interest recently @xcite , mainly due to the coexistence of these competing phases in some systems , with a similar phase diagram to systems with sdw and superconductivity @xcite , as well as the observation of cdw in the high temperature cuprate superconductors . in the cuprates ,
the role of cdw order in the formation of superconductivity remains a central and open issue @xcite , but the microscopic relationship is still unresolved .
+ structural distortions have also been found to occur in various iron based superconductors @xcite .
for instance , upon lowering the temperature of electron doped bafe@xmath0as@xmath0 , the symmetry of the crystal lattice is reduced from the four - fold symmetry of the thcr@xmath0si@xmath0 structure , to a two fold symmetric orthorhombic structure , while sdw ordering onsets at lower temperatures @xcite .
there is an increasing body of evidence that this distortion does not arise due to phonons @xcite , but is from an electronic instablity , the nematic order , which may be driven by either spin or orbital / charge fluctuations @xcite .
the suppression of both the nematicity and sdw ordering in the region of the superconducting dome , again indicates the importance of understanding the relationship between the various ordered phases and underlying interactions .
on the other hand , there also exist some non - iron based 122 pnictide superconductors @xmath3 ( @xmath4ba or sr , @xmath5transition metal and @xmath6pnictogen ) which undergo structural distortions upon cooling @xcite . a comparison with the iron based materials may provide crucial insights into the roles of various interactions and the resulting collective phenomena .
as@xmath0 at various pressures .
the different high temperature transitions are denoted by @xmath7 , @xmath8 , @xmath9 and @xmath10 .
the resistivity upon warming and cooling is shown in the vicinity of ( b ) @xmath7 at 4 kbar , ( c ) @xmath8 at 9 kbar , and ( d ) @xmath9 and @xmath10 at 17 kbar . in ( c )
the derivative @xmath11 is also shown by the solid line , which shows a clear change at @xmath8 . ] unlike the 122 iron pnictide superconductors which crystallize in the thcr@xmath0si@xmath0 structure ( @xmath12 ) , srpt@xmath0as@xmath0 forms in the cabe@xmath0ge@xmath0-type structure ( @xmath13 ) at high temperatures .
both are layered tetragonal structures but they have different stacking of the layers of transition metal and as atoms along the @xmath14 axis .
a further difference from the iron pnictides is that in the pt based materials , spin fluctuations are not generally expected to play a significant role .
below @xmath15 = 470 k , the crystal structure of srpt@xmath0as@xmath0 undergoes an orthorhombic distortion and cdw ordering @xcite , with a superconducting transition at a lower temperature of 5.2 k@xcite .
similarly in the newly discovered bapt@xmath0as@xmath0 with a resistive transition at @xmath2 = 1.67 k @xcite , a first order phase transition appears at about 275 k , at which a structural phase transition from the tetragonal cabe@xmath0ge@xmath0-type structure to an orthorhombic structure was found using powder x - ray diffraction ( xrd ) @xcite . in this article
, we demonstrate using single crystal xrd measurements that a cdw state already exists at room temperature in bapt@xmath0as@xmath0 and the structural transition at @xmath16 k corresponds to a change in the cdw superlattice .
to study the relationship between superconductivity and the cdw state , we performed resistivity measurements under pressure , which reveal a rich temperature - pressure phase diagram with multiple pressure - induced transitions , the emergence of which are correlated with sudden changes in @xmath2 .
single crystals of bapt@xmath0as@xmath0 were synthesised using a self - flux method described in ref .
resistivity measurements under pressure were carried out using a piston - cylinder - type pressure cell up to 27 kbar , and daphne 7373 was used as the pressure transmitting medium , to ensure hydrostaticity .
the applied pressure was determined by the shift in @xmath2 of a high quality pb single crystal .
measurements were performed using a physical property measurement system ( quantum design ppms-14 t ) in the temperature range of 2 k to 300 k , while measurements from 0.3 k to 5 k were performed using a @xmath17he refrigerator .
single - crystal xrd from room temperature to 100 k was carried out on a stoe ipds ii diffractometer at the argonne national laboratory , using an mo source ( @xmath18 ) .
reciprocal space reconstruction of the single crystal xrd data was performed with the software x - area .
a simulation of the diffraction peaks in the ( @xmath19 ) plane for the tetragonal cabe@xmath0ge@xmath0-type structure ( @xmath13 ) is shown in fig.[xrdfig](a ) , while the reciprocal space reconstruction of single crystal xrd data at 300 k and 100 k are shown in figs.[xrdfig ] ( b ) and ( c ) respectively .
reflections are labeled with the miller indices of the cabe@xmath0ge@xmath0-type structure ( @xmath13 ) . in the previous powder xrd measurements
@xcite , we reported a structural phase transition at 275 k from a tetragonal cabe@xmath0ge@xmath0-type structure to an orthorhombic one . based on the current single crystal xrd results , it should be noted that neither the reflections at 300 k nor 100 k could be satisfactorily indexed with a perfect cabe@xmath0ge@xmath0-type crystal structure . from comparing the experimental data at both 300 k and 100 k with the simulated reflections , we observe a number of additional peaks arising from complex cdw modulations . if we index the single crystal xrd data by introducing an orthorhombic distortion , the lattice parameters at 300 k are @xmath20 , @xmath21 and @xmath22 , while at 100 k , @xmath23 , @xmath24 and @xmath25 .
these results indicate that even at room temperature , the cdw modulations induce a small difference in @xmath26 and @xmath27 , which becomes larger below 280 k. subsequent analysis of the data at 300 k reveals peaks corresponding to 2@xmath282@xmath282 superlattice structure , where the unit cell is doubled in all directions , along with an incommensurate modulation @xmath29 .
reflections which are forbidden for the cabe@xmath0ge@xmath0-type structure such as @xmath30 and @xmath31 and the half fractional peaks such as @xmath32 arise from the 2@xmath282@xmath282 superlattice structure while the peaks at the position of @xmath33 around the superlattice peaks arise from the incommensurate modulation .
a comparison of the data from 300 k to 285 k shows no change in the structure of bapt@xmath0as@xmath0 . from 280 k to 100 k , bapt@xmath0as@xmath0 enters a new phase in which the symmetry further lowers to a 6@xmath286@xmath282 superlattice structure while the incommensurate modulation @xmath34 remains the same , as indicated by the appearance of new fractional peaks such as @xmath35 and @xmath36 .
therefore these results demonstrate that the cdw state is already present at room temperature , but there is a change in cdw structure below 280 k , along with an increased orthorhombic distortion .
+ as@xmath0 at various pressures , near the superconducting transition .
( b ) the upper critical field ( @xmath37 ) of bapt@xmath0as@xmath0 as a function of @xmath38 at various pressures , normalized by the product of @xmath2 and the slope of @xmath37 near @xmath2 . ] the temperature - pressure phase diagram of bapt@xmath0as@xmath0 was determined by measuring the resistivity under pressure . as shown in fig .
[ fig2](a ) , at @xmath39 kbar , there is a sudden increase of the resistivity upon cooling at about @xmath40 k , which is similar to the previously reported results at ambient pressure@xcite and corresponds to a change of the cdw structure . with increasing pressure
, the temperature of this transition decreases , before abruptly disappearing above @xmath41 kbar .
after @xmath7 disappears , a new transition labelled as @xmath8 is observed , where there is only a small anomaly in the resistivity and little change with increasing pressure . for pressures greater than @xmath42
kbar , @xmath8 disappears and two new transitions emerge at @xmath9 and @xmath10 , with the former corresponding to an increase of the resistivity upon cooling while the latter shows a sudden decrease .
these transitions have a different pressure dependence with the lower transition @xmath9 being rapidly suppressed with increasing pressure , while @xmath10 increases with pressure .
note that apart from @xmath8 , all the transitions display pronounced hysteresis , as shown in figs.[fig2](b)-(d ) , indicating a first - order nature .
it can also be seen that there is an increase in resistivity at @xmath7 and @xmath9 while it drops at @xmath10 . at several pressures with the residual values of @xmath43 subtracted .
the solid lines show fits to a @xmath44 dependence in the normal state . ]
figure [ fig3](a ) displays the resistivity under pressure at low temperatures , in the vicinity of the superconducting transitions . at ambient pressure
, @xmath2 is about 1.6 k , determined from where the resistivity reaches @xmath45 of the normal state value @xmath46@xmath1@xmath47 cm , similar to that previously reported@xcite . with increasing pressure
, the superconducting transition sharpens and @xmath2 undergoes a moderate enhancement at around 7 kbar which is close to @xmath48 , to around 2 k , along with a small decrease of @xmath43 . further increasing the pressure up to 13 kbar results in little change of @xmath2 .
however , upon applying pressures greater than 13 kbar , there is another jump in @xmath2 which is larger in size , reaching a near constant value above 17 kbar .
the superconducting transition at 17 kbar is significantly broader than that at both lower and higher pressures , which may be due to a pressure inhomogeneity , leading to the coexistence of both superconducting phases .
the newly emerged superconducting phase has the highest @xmath2 @xmath1 3.8 k , the lowest residual resistivity in the normal state , and @xmath2 changes little with pressure up to 26 kbar . the upper critical field @xmath49 at different pressures is shown in fig .
[ fig3](b ) .
the values of @xmath37 were measured in three regions of the temperature - pressure phase diagram ( 0 , 7 and 19 kbar ) , and were normalized by the product of @xmath2 and the derivative of @xmath37 near @xmath2 . when the normalized @xmath37 is plotted as a function of @xmath38 , all the data falls on one curve , in good agreement with the orbital limited values of @xmath37 calculated using the werthamer - helfand - hohenberg ( whh ) model shown by the dashed line @xcite .
this suggests that there is no change of the superconducting pairing state under pressure and that orbital limiting rather than pauli limiting is the dominant pair breaking mechanism .
meanwhile as displayed in fig .
[ fig4 ] , the low temperature normal state resistivity could be well fitted with fermi liquid behavior , @xmath50
. + as@xmath0 showing both the high temperature and superconducting transitions .
( b ) pressure dependence of the residual resistivity @xmath43 and @xmath51 coefficient . ]
figure [ fig5](a ) displays the temperature - pressure phase diagram obtained from resistivity measurements .
two abrupt enhancements of @xmath2 occur under pressure , both of which clearly coincide with changes in the high temperature phase transitions . at @xmath52 @xmath1 7
kbar , the cdw transition at @xmath7 disappears and a new transition @xmath8 is observed .
this change is accompanied by an enhancement of @xmath2 . upon further increasing the applied pressure ,
there is little change of @xmath8 up to @xmath53 kbar , where the transition at @xmath8 is no longer observed and two new phase transitions emerge .
the transition at @xmath9 decreases with increasing pressure before being suppressed at less than 22 kbar , while @xmath10 increases with pressure , reaching 290 k at 24 kbar .
the appearance of these transitions is accompanied by another pronounced enhancement of @xmath2 by almost a factor of two .
the critical pressure for the emergence of @xmath9 and @xmath10 is slighty lower than that corresponding to the jump in @xmath2 .
this may be due to the variation in the pressure transmitted by the daphne 7373 medium with temperature , which can cause a pressure increase of around 1 - 2 kbar between 300 k and low temperatures @xcite .
furthermore , the results from fitting the resistivity in the normal state are shown in fig .
[ fig5](b ) .
each step like increase of @xmath2 coincides with a drop in @xmath43 , with a small decrease at the disappearance of @xmath7 and a larger drop at the emergence of @xmath9 and @xmath10 .
meanwhile the resistivity coefficient @xmath51 undergoes a sharp drop at @xmath48 followed by a sudden increase at @xmath54 .
furthermore , neither @xmath43 or @xmath51 display any noticeable anomaly upon the disappearance of the transition at @xmath9 .
the above results demonstrate an intricate relationship between the superconductivity and the high temperature transitions in bapt@xmath0as@xmath0 .
the nature of the transitions at @xmath8 , @xmath9 and @xmath10 are yet to be uncovered . in comparison
, srpt@xmath0as@xmath0 undergoes a structural phase transition below @xmath55 k , from a tetragonal to orthorhombic structure @xcite , with two incommensurate cdw modulations emerging @xcite .
however , in bapt@xmath0as@xmath0 at ambient pressure the orthorhombic distortion and cdw modulation already exist above @xmath7 and this transition corresponds to a change of the periodicity of the superlattice . furthermore , there is a sudden increase in the resistivity at @xmath7 as in the case of the structural transition in srni@xmath0p@xmath0 @xcite , whereas there is a drop in the resistivity at @xmath15 in srpt@xmath0as@xmath0 .
however , the behaviour of bapt@xmath0as@xmath0 above 22 kbar more strongly resembles that of srpt@xmath0as@xmath0 , since we observe a drop in the resistivity at @xmath10 and there is a comparable value of @xmath2 .
if then the transition under pressure at @xmath10 is similar to that of srpt@xmath0as@xmath0 at ambient pressure , the differences in the cdw modulation and high temperature crystal structures may explain the different behavior of the resistivity seen at @xmath7 and @xmath10 .
this suggests the possibility that the cdw state below @xmath7 disappears at @xmath48 , above which only a weak anomaly at @xmath8 is observed .
furthermore , it is not clear whether the cdw state observed at ambient pressure and room temperature persists upon the application of pressure and it may be that the transition at @xmath8 lies beneath a cdw transition at higher temperatures . at @xmath54
a different cdw state may emerge .
however , while cdw states generally compete with superconductivity , in this instance the emergence of the transition at @xmath10 leads to a significant enhancement of @xmath2 , rather than a decrease .
the dramatic change in both @xmath43 and the @xmath51 coefficient at @xmath54 suggests this may correspond to a sudden change in the electronic structure and fermi surface topology .
such changes may lead to an enhanced density of states , which in turn may cause the significant increase of @xmath2 and @xmath51 .
the transition at @xmath9 , which appears under @xmath10 up to around 19 kbar , may then correspond to a further lowering of the crystal symmetry or a change in the modulation vector * q * of the cdw state . a change of * q * can lead to different behavior of the resistivity since this changes the reconstructed fermi surface , with the cdw gap opening across different regions .
however for these scenarios , it would be unusual that there is little change in @xmath2 upon the disappearance of @xmath9 .
as previously discussed , at ambient pressure there is a structural distortion at @xmath7 corresponding to a significantly increased orthorhombicity .
since both @xmath9 and @xmath10 have a similar first order nature , whether these transitions correspond to changes in either the crystal symmetry or structural parameters needs to be checked with xrd measurements under pressure .
all these results suggest the presence of both structural and electronic instabilities in the temperature - pressure phase diagram . in light of the recent evidence that the nematic order in the iron pnictide based superconductors is driven by electronic rather than structural instabilities @xcite , it will be of interest to determine whether any of the transitions or structural changes in bapt@xmath0as@xmath0 also have an electronic origin .
the relationship between the high temperature transitions and the superconductivity also needs to be further investigated and compared to both the cuprate and iron based materials , but the significantly lower values of @xmath2 may be a result of the lack of a spin fluctuation mediated pairing mechanism . in conclusion , using single crystal xrd measurements we have identified that a cdw state exists in bapt@xmath0as@xmath0 in ambient conditions , and that the transition at @xmath56 k corresponds to a change of the superlattice structure .
the application of pressure results in a rich phase diagram .
pressure suppresses the cdw transition at @xmath7 until it disappears at around @xmath48 = 7 kbar , which leads to an enhancement of @xmath2 . upon further increasing the pressure ,
new high temperature transitions are observed in the resistivity , which are correlated with an additional increase of @xmath2 .
since these results indicate the presence of multiple instabilities at high temperatures , bapt@xmath0as@xmath0 may provide a platform for studying the competition between different interactions and their interplay with superconductivity .
further experiments such as nmr and xrd under pressure , as well as theoretical calculations are needed to characterize the origin of these transitions and elucidate the mechanisms which lead to these exotic behaviors .
we thank c. cao for interesting discussions .
the work at zhejiang university was supported by the national natural science foundation of china ( no.11474251 and no.11374257 ) , national key r&d program of the most of china ( no .
2016yfa0300202 ) and the science challenge program of china .
the work at argonne was supported by the office of science , office of basic energy sciences , division of materials sciences , us department of energy , under contract de - ac02 - 06ch11357 ( bes - dmse ) .
10 j. paglione and r. l. greene , high - temperature superconductivity in iron - based materials .
* 6 , * 645 - 658 ( 2010 ) .
x. h. chen , p. c. dai , d. l. feng , t. xiang and f. -c .
zhang , iron - based high transition temperature superconductors .
sci . rev .
* 1 , * 371 - 395 ( 2014 ) .
f. steglich , j. aarts , c. d. bredl , w. lieke , d. meschede , w. franz , and h. schfer , superconductivity in the presence of strong pauli paramagnetism : ce@xmath57@xmath58 .
rev . lett .
* 43 , * 1892 - 1896 ( 1979 ) .
p. gegenwart , c. langhammer , c. geibel , r. helfrich , m. lang , g. sparn , f. steglich , r. horn , l. donnevert , a. link , and w. assmus , breakup of heavy fermions on the brink of `` phase a '' in cecu@xmath0si@xmath0 .
* 81 , * 1501 ( 1998 ) . h. q. yuan , f. m. grosche , m. deppe , c. geibel , g. sparn , and f. steglich , observation of two distinct superconducting phases in ce@xmath57@xmath58 .
science * 302 , * 2104 - 2107 ( 2003 ) .
a. t. holmes , d.jaccard and k. miyake , signatures of valence fluctuations in cecu@xmath0si@xmath0 under high pressure , phys .
b * 69 , * 024508 ( 2004 ) . for a review ,
see a. m. gabovich , a. i. voitenko , and m. ausloos , charge- and spin - density waves in existing superconductors : competition between cooper pairing and peierls or excitonic instabilities , phys .
* 367 * , 583 ( 2002 ) .
m. hoesch _ et al .
_ , splitting in the fermi surface of zrte@xmath59 : a surface charge density wave system .
b * 80 , * 075423 ( 2009 )
. e. morosan _
et al . _ ,
superconductivity in cu@xmath60tise@xmath0 , nat .
* 2 , * 544 ( 2006 ) .
j. m. tranquada _
_ , evidence for stripe correlations of spins and holes in copper oxide superconductors , nature * 375 , * 561 - 563 ( 1995 ) .
t. wu _ et al .
_ , magnetic - field - induced charge - stripe order in the high - temperature superconductor yba@xmath0cu@xmath59o@xmath61 , nature * 477 , * 191 - 194 ( 2011 )
. g. ghiringhelli _
et al . _ ,
long - range incommensurate charge fluctuations in ( y , nd)ba@xmath0cu@xmath59o@xmath62 , science * 337 , * 821 - 825 ( 2012 ) r. comin _ et al .
_ , symmetry of charge order in cuprates , nature materials * 14 , * 796 - 800 ( 2015 )
. y. x. wang _ et al .
_ , enhancement of superconductivity at the onset of charge - density - wave order in a metal , phys .
b * 92 , * 125108 ( 2015 ) .
h. freire _
_ , renormalization group analysis of the pair - density - wave and charge order within the fermionic hot - spot model for cuprate superconductors .
b * 92 , * 045132 ( 2015 )
. m. rotter , m. tegel , and d. johrendt , superconductivity at 38 k in the iron arsenide ( ba@xmath63k@xmath60)fe@xmath0as@xmath0 .
lett * 101 , * 107006 ( 2008 ) .
m. yi _ et al .
_ , symmetry - breaking orbital anisotropy observed for detwinned ba(fe@xmath63co@xmath60)@xmath0as@xmath0 above the spin density wave transition , proc .
usa * 108 , * 6878 - 6883 ( 2011 ) .
m. g. kim _
et al . _ ,
character of the structural and magnetic phase transitions in the parent and electron - doped bafe@xmath0as@xmath0 compounds , phys .
b * 83 , * 134522 ( 2011 ) .
et al . _ , divergent nematic susceptibility in an iron arsenide superconductor , science * 337 , * 710 ( 2012 ) .
r. m. fernandes , a. v. chubukov and j. schmalian , what drives nematic order in iron - based superconductors ?
, nature phys .
* 10 , * 97 - 104 ( 2014 ) .
f. ronning , n. kurita , e. d. bauer , b. l. scott , t. park , t. klimczuk , r. movshovich , and j. d. thompson , the first order phase transition and superconductivity in bani@xmath0as@xmath0 single crystals . j. phys . : condens .
matter * 20 , * 342203 ( 2008 ) .
et al . _ ,
coexistence of superconductivity and charge density wave in srpt@xmath0as@xmath0 .
jpn * 79 , * 123710 ( 2010 ) .
et al . _ ,
superconductivity in 122 antimonide srpt@xmath0sb@xmath0 supercond .
* 26 , * 075001 ( 2013 ) .
a. f. fang__et al . _
_ , single - crystal growth and optical conductivity of srpt@xmath0as@xmath0 superconductors , phys .
b * 85 , * 184520 ( 2012 ) .
_ , modulated crystal structures and phase transitions - the compounds srpt@xmath0a@xmath64 and eupt@xmath0as@xmath0 . z. anorg . allg . chem .
* 633 , * 2037 ( 2007 ) .
_ , two - coupled structural modulations in charge - density - wave state of srpt@xmath0as@xmath0 superconductor , chin .
b * 23 , * 086103 ( 2014 ) .
s. kawasaki _ et al .
_ , coexistence of multiple charge - density waves and superconductivity in srpt@xmath0a@xmath64 revealed by @xmath65as - nmr / nqr and @xmath66pt - nmr , phys .
b * 91 , * 060510(r ) ( 2015 ) . w. b. jiang _ et al .
_ , superconductivity and structural distortion in bapt@xmath0as@xmath0 , j. phys .
: condens .
matter * 27 , * 022202 ( 2015 ) .
r.werthamer , e. helfand , and p. c. hohenberg , temperature and purity dependence of the superconducting critical field , @xmath37 .
electron spin and spin - orbit effects , phys . rev . * 147 , * 295 ( 1965 ) .
k. murata , h. yoshino , h. o. yadav , y. honda , and n. shirakawa , pt resistor thermometry and pressure calibration in a clamped pressure cell with the medium , daphne 7373 , rev .
* 68 , * 2490 ( 1997 ) .
f. ronning , e. d. bauer , t. park , s .- h .
baek , h. sakai , and j. d. thompson , superconductivity and the effects of pressure and structure in single - crystalline srni@xmath0p@xmath0 , phys .
b * 79 , * 134507 ( 2009 ) . | the newly discovered bapt@xmath0as@xmath0 shows a structural distortion at around 275 k , followed by the emergence of superconductivity at lower temperatures . here
we identify the presence of charge density wave ( cdw ) order at room temperature and ambient pressure using single crystal x - ray diffraction , with both a superlattice and an incommensurate modulation , where there is a change of the superlattice structure below @xmath1 275 k. upon applying pressure , bapt@xmath0as@xmath0 shows a rich temperature - pressure phase diagram with multiple pressure - induced transitions at high temperatures , the emergence or disappearance of which are correlated with sudden changes in the superconducting transition temperature @xmath2 .
these findings demonstrate that bapt@xmath0as@xmath0 is a promising new system for studying competing interactions and the relationship between high - temperature electronic instabilities and superconductivity . |
the kpz equation @xcite , in its apparent simplicity , involves many issues that need clarification .
the continuum equation ( that describes the local growth of an interface profile ) is @xmath7 the real behavior of this equation is not well understood .
the absence of a complete mean field theory does not help , and the fact that we have do understand the effects of a strong coupling fixed point makes very difficult the development of a perturbative renormalization approach . a practical approach for numerical studied of the problem is to consider lattice restricted solid on solid ( rsos ) surfaces ( see for example @xcite for a review ) : one considers a discretized height field on a @xmath0 dimensional lattice , and imposes the constrain that the absolute value of the distance among two neighboring surface elements can only take the values zero and one
. old extensive numerical simulations @xcite do not clarify the situation completely . after a number of interesting field theoretical results @xcite , the introduction of new renormalization group based techniques is a potentially promising direction of development @xcite .
the main theoretical points which still deserve proper explanation are two : it is still not clear whether equation ( [ kpz_equation ] ) has a finite upper critical dimension ( @xmath8 ) or not , and the exact quantification of the related critical exponents .
regarding arguments in favor of @xmath9 we address the reader to @xcite , while arguments supporting @xmath10 can be found in @xcite .
in this light we introduce here a new numerical technique and present some precise numerical simulations that allow us to estimate critical exponents .
thanks to the new , precise technique we succeed to clarify some questions : we show for example that a conjecture of @xcite is not founded , and we give quantitative estimates about the behavior of sub - leading corrections .
the precise results on which this paper is based are due to a new technique we introduce to simulate rsos surfaces , where adjacent elements of the surface can only be at a distance of @xmath5 , @xmath11 or @xmath12 lattice spacings .
the technique is a generalization of the so called _ multi - spin coding _
technique ( well discussed by rieger @xcite in the context of the study of disordered systems ) , where , by using the fact that the @xmath13 spins can be represented by boolean logical variables , one stores @xmath14 copies of the system in one single computer word ( we will assume in the following we are using a @xmath14 bit computer ) .
the main idea is that we can simulate , basically at the cost of one single usual simulation , @xmath14 copies of the system , by rewriting the basic operations ( like summing spins for computing the effective force ) as boolean operations , and by exploiting the fact that when , for example , the computer is calculating a logical bit @xmath15 it is indeed doing that @xmath14 times at once .
the gain of such an approach over an usual single spin simulation is of order @xmath6 : one gains a factor @xmath14 for the number of systems that updates at once , and looses a factor of order @xmath1 in computational complexity ( adding the spin in the boolean form takes some more time @xcite ) .
we generalize here the ising case to the case of a _ field of differences _ , that can have three values , since a given element of the surface can be at the same height of a given neighbor , or one step behind it , or one step ahead of it .
the method is new as applied to such a system , where one does not have to compute the value of an energy , but uses the boolean operations to determine if the element can be moved without violating the geometrical constraint .
let us thinks for simplicity about @xmath16 , where we have a @xmath16 support where a height field ( the surface height ) @xmath17 is defined .
one surface element @xmath18 has @xmath19 first neighbors : in the rsos model the difference @xmath20 , where @xmath21 , can only take the three values @xmath12 , @xmath11 and @xmath22 . we store the @xmath23 values of the @xmath24 ( and , as we will discuss better ,
each @xmath24 needs two bit to be stored .
this is a redundant way to store the information , since a factor two in memory could be easily saved , but it is very convenient from the point of view of computer time : as usual one trades memory for time , and avoids a cumbersome reconstruction by storing more information ) .
if a given element is behind its neighbor the difference is @xmath12 , if it has the same height than the neighbor the difference is @xmath11 , while if it is ahead of the neighbor the difference is @xmath22 . since we have three allowed values we can represent each of these difference with two bits ( that we will call respectively @xmath25 , high bit , and @xmath26 , low bit ) : we can for example code with @xmath27 the situation where the given element is behind its neighbor , with @xmath28 the situation where they have the same height , and with @xmath29 the situation where it is ahead ( the value @xmath30 is forbidden ) .
if the element is ahead of even a single one of the @xmath16 neighbors the move is forbidden ( it would violate the rsos constrain ) .
it is clear this is easily implemented in our coding : one just needs to do a logical @xmath31 of the @xmath16 @xmath25 bits related to the site @xmath32 , and if it is @xmath5 at least one of the neighbors is behind and the move is not allowed .
let us see better why .
we start comparing our element @xmath32 to the first of its @xmath19 neighbors ( let us say the one in the positive direction @xmath22 ) : if @xmath32 is ahead the neighbor the relevant @xmath25 bit ( that we have stored in @xmath33 ) is @xmath5 , and we already know we can not move . if @xmath32 is on the contrary at the same height of the neighbor or behind it we have that , as far as this neighbor is concerned , the element @xmath32 can advance of one unit .
now we look at the second neighbor , where the same reasoning holds : if we look at the logical @xmath31 of the two relevant high bit we will find that we can not move if this quantity is one .
looking at all the @xmath16 neighbor we see that the move is forbidden if the logical @xmath31 of the @xmath16 @xmath25 bit is @xmath5 : clearly this operation , as all the other present in the core of the code , updates with a single computer cycle the @xmath14 copies of the system . after doing that ,
if the element @xmath32 can not be moved there is nothing that has be done : if it gets moved we have to update the @xmath24 related to the site @xmath32 and to the neighbors , to describe the new situation .
the codes simulating systems in a different number of spatial dimensions ( in our case from @xmath34 to @xmath35 ) are simple generalization one of the other ( when adding a dimension one has to add checks on the new neighbors and their update ) . as usual in these kind of simulations the random noise is implemented with a random choice of the site to be updated .
we also implement a fifty percent probability of really updating a surface element that according to the rsos constrain could be updated .
this is very important , since starting from random independent surfaces is not enough : because of our parallel scheme the sites of our @xmath14 copies have to be updated in the same order , and such an updating algorithm with updating probability one is attractive , and the @xmath14 configurations asymptotically at large times become equal @xcite .
the probability of not accepting an allowing change , that depends on each of the @xmath14 configurations , solve this problem .
we are aware of a parallel algorithm to update surfaces @xcite : it is very different in spirit from our algorithm ( since it parallelizes on different sites of the lattice ) .
we have not compared in detail the performances of the two algorithms , but we believe that on one side the algorithm of @xcite is more general , and not limited to rsos models , but on the other side our algorithms is more regular ( there are no exceptional loops ) , and is probably better performing for the model we study .
we have based this work on numerical simulations of lattices of volume @xmath36 .
the spatial dimensionality @xmath0 is the spatial support , where a @xmath5 dimensional surface take values .
we study the @xmath34 , @xmath37 and @xmath35 cases . at a given time @xmath38 of the dynamical evolution
the position of the surface can be expressed by the values @xmath39 ( that we reconstruct by the differences we store in our code , see the former section ) .
@xmath32 represents , in lexicografic order , a @xmath0-ple labeling the spatial sites .
we consider a dynamics which generates a restricted solid on solid ( rsos ) growth : distances of first neighboring elements of the surface can not be larger than one . at each trial step
we move elements of the surface that are not constrained not to do so because of the rsos restriction with probability @xmath40 : we have to use a probability different from one in order to keep independent the different surfaces we simulate in the same computer word .
we consider a large number of different lattice sizes . in @xmath34
we take @xmath26 going from @xmath41 to @xmath42 , in @xmath37 we consider @xmath26 going from @xmath41 to @xmath43 while in @xmath35 @xmath26 goes from @xmath41 to @xmath44 .
all our data have been averaged over @xmath14 different dynamical runs . .
number of full lattice sweeps for each run on different lattice sizes and number of dimensions .
[ tab - one ] [ cols="^,^,^,^,^,^",options="header " , ] our recent are not incompatible with the recent ones of @xcite , but our small error bars allow us to reach precise conclusions .
also the comparison with the exponents found in @xcite is fair : we have stressed the length of our runs , in order to be able to give a clean estimate of the time asymptotic behavior , so that our result is hopefully unbiased .
the numerical technique we have introduced works well , and has allowed us to run very precise numerical simulations with a limited amount of computer time ( a few months of pentium ii processor ) .
we have been able to determine critical exponents of the kpz universality class with high accuracy .
we have falsified the guess that the exponents are simple rational numbers .
it is also unambiguous from our data that the upper critical dimension is larger than @xmath3 ( as opposed to the claims of @xcite ) .
thanks to our precise measurements ( and fitting together the first @xmath2 non trivial moments of @xmath45 ) we have also been able to determine the exponent @xmath46 of the first non - leading scaling corrections .
it is interesting to notice that the estimated value of @xmath46 is always very close to @xmath5 , independent from the dimensionality of the system .
the next interesting step , following the approach of @xcite , would be to try and implement a systematic monte carlo renormalization group : the multi - surface coding technique we have discussed could be a very important ingredient of such a development .
for the numerical simulations described here we have used the _ kalix2 _
parallel computer ( built on pentium ii chips ) , funded by italian murst 1998 cofin .
we thank claudio castellano and matteo marsili for interesting discussions , and marcel den nijs and kay wiese for a relevant correspondence . c. castellano , m. marsili and l. pietronero , phys .
* 80 * ( 1998 ) 3527 , cond - mat/9802284 ; c. castellano , a. gabrielli , m. marsili , m. a. muoz and l. pietronero , phys .
e * 58 * ( 1998 ) r5209 , cond - mat/9809197 ; c. castellano , m. marsili , m. a. muoz and l. pietronero , cond - mat/9904434 .
ballesteros , l. a. fernandez , v. martin - mayor and a. munoz - sudupe , phys .
b * 387 * ( 1996 ) 125 ; h.g .
ballesteros , l. a. fernandez , v. martin - mayor , a. munoz - sudupe , g. parisi and j.j .
ruiz - lorenzo , phys .
b * 58 * ( 1998 ) 2740 . | we study the kpz equation ( in @xmath0 = @xmath1 , @xmath2 and @xmath3 spatial dimensions ) by using a rsos discretization of the surface .
we measure the critical exponents very precisely , and we show that the rational guess is not appropriate , and that @xmath4 is not the upper critical dimension .
we are also able to determine very precisely the exponent of the sub - leading scaling corrections , that turns out to be close to @xmath5 in all cases .
we introduce and use a _ multi - surface coding _
technique , that allow a gain of order @xmath6 over usual numerical simulations . |
the recent discovery of colossal magnetoresistance ( cmr)@xcite in mixed - valence compounds of the form @xmath0 , ( where @xmath1 can be ca , sr or ba ) has revived the interest in these materials with perovskite structure .
the interest is focused on the phase diagram and the magneto - transport properties .
both features have been determined experimentally@xcite . for @xmath2 and low temperatures
the system is metallic and presents ferromagnetic order . as the temperature increases the system becomes insulator and paramagnetic .
the ferro - paramagnetic transition occurs at an @xmath3 dependent critical temperature @xmath4 .
the colossal magnetoresistance effects are observed for temperatures close to @xmath5 . for @xmath6 and low temperatures
the system is a layer antiferromagnetic with ferromagnetic coupling inside planes . in these compounds
the electronically active orbitals are the @xmath7 @xmath8 orbitals .
the number of @xmath8 electrons per @xmath7 is @xmath9 , namely , four electrons per @xmath10 and three per @xmath11 . the cubic symmetry and the strong hund s rule coupling make that three electrons get trapped in the @xmath12 states and therefore these electrons become electrically inert , forming a core spin @xmath13 of magnitude @xmath14 .
the rest of the electrons go to the @xmath15 orbitals and they get strongly coupled to @xmath13 by the hund s rule coupling .
consequently , all the spins on each @xmath7 prefer to be parallel . for small values of @xmath3 ,
the perovskites show a long range jahn - teller order which selects a preferred @xmath8 orbital and therefore it is possible to assume that the electrons move only through one @xmath8 orbital .
the double exchange ( de ) mechanism developed by zener @xcite , explains the existence of ferromagnetism and metallic behavior at low temperatures . in this model ,
the electrons get mobility between the manganese ions using oxygen , which is magnetically inert , as an intermediate . as a consequence
, the tunneling takes place between two configurations in which the @xmath7 ions of different charge ( @xmath16 and @xmath17 ) interchange their valence states .
this conduction process is proportional to the electron transfer integral @xmath18 if the core spins of @xmath16 and @xmath17 are aligned .
otherwise , the transfer integral is inversely proportional to the hund s rule coupling energy , which for the @xmath7 ions is much larger than @xmath18 . in the de model
the ferromagnetism is then induced via this electron conduction process . in the limit of large hund s coupling the spin of
the electron in the active @xmath8 orbital is parallel to the local core spin @xmath19 , which is treated as a classical rotator with the normalization @xmath20 and characterized by the angles @xmath21 and @xmath22 .
then , semiclassically , the effective hopping hamiltonian is @xmath23 here @xmath24 creates an electron at site @xmath25 with spin parallel to @xmath19 , @xmath26 denotes the nearest - neighbor pairs , which are only counted once , and the hopping amplitude acquires a berry s phase and it becomes a complex number given by@xcite , @xmath27 this complex hopping appears after rotating the conduction electron spins so that the spin quantization axis at site @xmath25 is parallel to @xmath19 , and then project onto the spin parallel to @xmath28 . in this paper
we will call complex de ( cde ) model the system governed by the hamiltonian ( [ hcomplex ] ) with the hopping given by ( [ tcomplex ] ) . for topologies where electrons close paths do not occur , it is possible to choose wave functions phase factors such that @xmath29 being @xmath30 the angle between the semiclassical spins @xmath19 and @xmath28 . in this work
we will use the name de model for the system controlled by equation ( [ hcomplex ] ) with the hopping ( [ treal ] ) .
this is the hamiltonian proposed by anderson and hasegawa@xcite as a generalization of the hamiltonian describing the tunneling between two fixed sites @xmath25 and @xmath31 .
de gennes@xcite and kubo and ohata@xcite calculated , in a mean field theory approximation , the magnetic phase diagram of the de model .
their results show a ferro - paramagnetic phase transition at a critical temperature @xmath32 this transition is accompanied by a change in the temperature dependence of the resistivity . in equation ( [ tmf ] )
, @xmath33 means the expectation value of the operator @xmath34 at zero temperature . however , recently it has been pointed out@xcite that the critical temperatures predicted by the mean field theory of the de model is much bigger than the observed experimentally .
furthermore , the resistivity implied by the hopping ( [ treal ] ) is incompatible with many aspects of the experimental information .
et al_@xcite proposed that in addition to the de mechanism a strong electron - phonon interaction plays a crucial role in these perovskites .
the interplay between these two effects could reproduce the experimental critical temperature and the observed behavior of the resistivity at the magnetic transition temperature@xcite . on the other hand , yunoki _
et al_@xcite claim that the inclusion of the phase in the hopping term can lower the critical temperature down to the experimental values .
also , for large but finite hund s coupling , they obtain that at low concentrations phase separation between hole - poor antiferromagnetic and hole - rich ferromagnetic regions occurs .
the existence of phase separation is also obtained by solving the de model plus an antiferromagnetic coupling between next neighbors core spins@xcite . in this paper
we are interested in two points ; first , in comparing the magnetic phase diagram of the two de models , equations ( [ tcomplex ] ) and ( [ treal ] ) , and second , in obtaining an approximated expression for the de energy which allows us to perform monte carlo calculations in bigger unit cells and to obtain accurate values of the magnetic critical temperature . the main results of our calculations are shown in fig.[tc ] . here
we show the critical temperature versus concentration , for the complex de model and for the real de model , in the case of a unit cell of size @xmath35 .
we obtain that there is not big difference between the critical temperatures of both models .
this result indicates that the contribution to the partition function of electronic configurations in which fermions move on closed loops in real space can be neglected . in fig.[tc ] we also show monte carlo results obtained by using an approximated expression for the electron kinetic energy .
we have checked in small systems that this approximation is realistic . from our results
we obtain that the critical temperature of the de model is in the range of the experimental one@xcite .
the paper is organized as follows : in section ii we describe briefly the monte carlo algorithm we use and in section iii we present the results obtained for the cde and de hamiltonians . in section
iv we introduce an approximation to the de energy which allows us to perform monte carlo simulations in bigger systems .
section v is devoted to show the results obtained in the frame of this approximation .
finally , we present the conclusions in section vi .
for calculating the magnetic phase diagrams , we have performed classical monte carlo ( mc ) simulations on the classical core spin angles .
the simulations are done in @xmath36 cubic lattices with periodic boundary conditions .
although the localized spins are considered classical , the kinetic energy of the conduction electrons is calculated by diagonalizing the de hamiltonian because we consider it as a quantum quantity .
the standard metropolis algorithm@xcite was used in the mc simulations .
the sites to be considered for a change in the spin orientation are randomly chosen .
once a site is selected for a spin reorientation , the angle associated with an attempted change of the spin is chosen at random from within a specified range@xcite .
then the energy change , @xmath37 , associated with the attempted update , is calculated .
if the quantity @xmath38 is smaller than a random number between @xmath39 and @xmath40 , the change is allowed , otherwise , it is rejected .
typically , 5000 - 7000 mc steps per spin are used for equilibration and 3000 - 5000 steps for spin are used for calculating averages . in the simulations we calculate the average of the internal energy , @xmath41 , and the average of the absolute value of the magnetization , @xmath42 , @xmath43 @xmath44 where @xmath45 denotes statistical average .
since the mc updating procedure generates uniform rotations of the spin system , a calculation of the mc average of the direction of the magnetization is not meaningful .
also we calculate the average value of the width of the density of states , @xmath46 , and the value of the chemical potential , @xmath47 , measured with respect to the bottom of the density of states , @xmath48 where @xmath49 and @xmath50 are the minimum and the maximum energies obtained from diagonalizing the electron hamiltonian , and @xmath51 is the higher energy of the occupied states .
in order to obtain the internal energy in each of the monte carlo steps , it is necessary to diagonalize the electronic hamiltonian .
the diagonalization is very expensive in terms of cpu time , and therefore we can only study small size systems . here
we present the results obtained for a cubic unit cell of size @xmath52 and for the electron concentrations @xmath53 , @xmath54 and @xmath55 . in figure
[ de - cde ] , it is presented , for the de and the cde hamiltonians , the magnetization , @xmath42 , as a function of the temperature , @xmath56 . due to the finite size of the unit cells used in the simulations
, @xmath42 is different from zero at any temperature , and we define the critical temperature , @xmath57 , as the point where the second derivative of @xmath42 with respect @xmath56 changes sign .
this way of obtaining the critical temperatures , implies uncertainties of around @xmath58 of the value of @xmath57 . in any case
we obtain that the critical temperatures of the de and the cde hamiltonians are practically the same . in figure [ e - t ]
we plot the internal energy @xmath41 as a function of @xmath56 .
note that the difference between the internal energies obtained by using the cde and using the de hamiltonians is rather small .
that is the reason why both models give very similar magnetic critical temperatures .
these results imply that in order to calculate total kinetic energies the close paths are almost irrelevant . in figure [ w - mu - t ]
we plot , as a function of the temperature , the width of the density of states @xmath46 and the chemical potential @xmath47 , for the de case .
note that at temperatures close to @xmath57 , the bandwidth and the chemical potential have a value bigger than the obtained in the fully disordered case , @xmath59 .
this is because in the de models the ferromagnetic to paramagnetic transition is a second order phase transition , and therefore the internal energy is continuous at the critical temperature .
moreover , the bandwidth and the chemical potential present a continuous behavior near @xmath57 . in the de models
we do not observe any change in the electronic states at the ferro - paramagnetic critical temperature .
in the mc calculations of the de models , the size of the matrix to diagonalize impose a restriction on the dimension of the unit cell used in the simulation . in order to be able to perform simulations in bigger systems ,
we have developed a second order perturbation theory for obtaining an expression for the electron kinetic energy in a background of randomly oriented core spins .
we start writing the hamiltonian ( [ hcomplex ] ) in the form , @xmath60 with @xmath61 being @xmath62 the average of the absolute value of the hopping amplitude , @xmath63 and @xmath64 here @xmath65 is the total number of @xmath7 ions in the system . note that since the hamiltonian is hermitic , @xmath62 is a real quantity .
given a disordered system , characterized by a set of @xmath66 , we want to obtain the expectation value of @xmath67 .
the hamiltonian @xmath68 can be diagonalized by using the bloch states , @xmath69 where @xmath70 represents the atomic orbital at site @xmath25 , @xmath71 are the lattice vectors and @xmath72 is a wave vector in the first brillouin zone .
the energy of the state @xmath73 is @xmath74 being @xmath75 the lattice parameter . in function of these eigenvalues ,
the expectation value of @xmath68 is , @xmath76 where the sum is over the occupied states . from the value of @xmath77
we can obtain the expectation value of @xmath78 , @xmath79 which is independent of the value of @xmath80 . to obtain the expectation value of @xmath81
, we first notice that since the set @xmath82 are randomly distributed , @xmath83 , and it is necessary second order perturbation theory in order to get a correction to the expectation value of @xmath68 , @xmath84 where @xmath85 is the green function of the perfect crystal .
because the values @xmath82 are not correlated , it is easy to obtain the expression , @xmath86 with @xmath87 adding @xmath77 and @xmath88 , we obtain the following expression for the energy of the system , @xmath89 in figure [ coef ] we plot @xmath90 and @xmath91 as a function of the electron density @xmath3 . the first term in the above expression is a ferromagnetic coupling for the core spins and the second term is an antiferromagnetic coupling .
note that @xmath90 is only linear on @xmath3 at small values of the electron concentration .
in figure [ de - cde ] we plot , for @xmath53 , @xmath92 and @xmath93 , the absolute value of the magnetization , @xmath42 , as a function of the temperature , @xmath56 , for the case where the internal energy is obtained by using the approximated de energy , equation ( [ energy - approx ] ) , for comparing with the case in which it is obtained by diagonalizing the de hamiltonian . the difference between both curves
is very small , and the corresponding critical temperatures are practically the same , see figure [ tc ] .
therefore we conclude that equation ( [ energy - approx ] ) is a good approximation for the energy of the de model .
using equation ( [ energy - approx ] ) , for the de energy , we can perform mc simulations in much bigger systems , and obtain more precise values of the critical temperatures . in figure [ n=20 ]
we plot the magnetization as a function of the temperature for a sample of size @xmath94 . for this size
the critical temperature is obtained with a big precision .
we have checked that for @xmath94 the values of the critical temperatures are accurate in two digits and that it is not worthwhile to increase more the size of the system .
we have performed mc simulations for different values of the electron concentration and , in figure [ tc ] , we have plotted the magnetic phase diagram .
kubo and ohata@xcite have developed an expression for the de energy which coincides with @xmath77 . in order to know the importance of @xmath88 in the calculation of the value of the critical temperature , we have also computed the value of @xmath42 as a function of @xmath56 by using only @xmath77 for the internal energy , see figure [ n=20 ] . comparing these results with the ones obtained with the approximated de energy , @xmath95 , we estimate that the inclusion of @xmath88 in the internal energy lowers the critical temperature more than a 20@xmath96 . in fig.[tprome ] , we plot for different values of the concentration @xmath3 , the statistical average of the absolute value of the hopping amplitude , @xmath97 , as a function of the temperature .
this quantity is proportional to the electron bandwidth .
we find that @xmath98 is a continuous function of @xmath56 , and there are not signals of any change in the electronic structure near @xmath57 .
the bandwidth near @xmath57 is around 1.15 times bigger than the bandwidth in the @xmath59 limit .
et al_@xcite have obtained that in the @xmath59 limit of the cde model , only @xmath99 of the electron states are localized . in order to know the difference between the density of states at @xmath100 and at @xmath59 , we have calculated the statistical average of the electron density of states .
we find that the number of electron states with energies between @xmath101 and @xmath102 is less than @xmath103 of the total number of states .
therefore , even if all these states were localized , the percentage of localized states near @xmath57 is not larger than @xmath104 .
we conclude that in the de models it is not possible to relate the metal - insulator transition with the ferro - paramagnetic one .
the critical temperatures we obtain are around 1.5 times smaller than the obtained in mean field theory by kubo and ohata@xcite .
however , it is around 8 times smaller than the obtained by millis _
et al._@xcite using also a mean field approximation .
we think that the difference occurs because they use a lineal dependence , see fig .
[ coef ] , of the kinetic electron energy which overestimate the value of @xmath105 , and therefore the value of @xmath57 . from kubo and ohata model ( [ tmf ] ) ,
it is clear that @xmath57 scales with @xmath106 and the magnetization versus @xmath107 curve is independent of the electron concentration x. but there is no reason to think that this must be the case when the second term , @xmath88 , of our approach is included in the calculations .
however , we have found that the @xmath42 versus @xmath107 curves almost coincide ( see fig.[n=20 ] ) .
this fact is not only found in the frame of our approximation to the de energy but also when the de and cde hamiltonians are diagonalized .
the presented results are in qualitatively agreement with those obtained by dagotto _
et al_@xcite .
however our estimations of @xmath57 are @xmath108 times bigger .
the discrepancy is due to the difference in the criterion used to obtain @xmath57 . in reference @xcite the @xmath57 is defined as the temperature where the spin - spin correlation function in real space becomes zero at the maximum distance available in the unit cell .
they used a unit cell of size @xmath109 .
our criterium is based in the change of sign of the second derivative of @xmath42 with respect to @xmath56 .
we have checked that both criteria give the same @xmath57 in big unit cells , but the criterium based on the correlation function underestimate @xmath57 in small unit cell calculations .
this is clear in fig.[correlation ] where we plot the spin - spin correlation function versus distance for several temperatures and for two unit cells with @xmath110 and @xmath111 .
the electron concentration is @xmath112 to compare with the results presented in reference @xcite . in the case of @xmath110
it seems that the critical temperature is smaller than @xmath113 , because the correlation function is zero at a separation of 8@xmath75 .
however , when @xmath111 it is clear that the ferromagnetic correlation at @xmath113 is not zero at 8@xmath75 , and in fact it seems that at this temperature the system is still ferromagnetic . on the other hand ,
our criterium overestimate the @xmath57 for small unit cells and the possibility of increasing the size leads to more realistic values of the critical temperature . to compare with the experimental critical temperatures ,
it is necessary an estimation of @xmath114 .
the experimental bandwidth is around 1 - 4ev@xcite , and therefore the value of @xmath114 is around 0.08 - 0.3ev . with this result
our estimate of the critical temperature is between 150k and 500k at @xmath115 .
this value of @xmath57 is in the range of the observed experimentally .
using monte carlo techniques , we have obtained the magnetic phase diagram of the double exchange hamiltonian . comparing the results obtained from the double exchange hamiltonian with a complex hopping ( cde ) and with its absolute value ( de ) we have found that the berry phase of the hopping has a negligible effect in the magnetic critical temperature .
this implies that it is possible to choose wave functions phase factors that counteract the phase of the hopping such that it is not relevant in the mc calculations . to avoid the limitations on the size of the systems studied
we have developed a second order perturbative approach to calculate the electron kinetic energy without diagonalizing the hamiltonian . within this approach
we have calculated the critical temperature for a bigger size of the system ( @xmath94 ) and , consequently , more accurately .
the values of @xmath57 obtained are in the range of the experimental ones and in qualitative agreement with those in ref .
neither the average of the electron density of states nor the kinetic energy reveal any strange behavior near the magnetic critical temperature .
we conclude that , in order to understand the metal - insulator transition which occurs in the manganites , another effects , not included in the double exchange hamiltonian , should be taken into account . _
_ this work was supported by the cicyt of spain under contract no .
pb96 - 0085 , and by the fundacin ramn areces .
helpful conversations with f. guinea , r. ramrez , s. das sarma and j.a .
vergs are gratefully acknowledged .
et al . _ , physica ( amsterdam ) * 155b * , 362 ( 1989 ) ; k. chahara _ et al . _
lett . * 63 * , 1990 ( 1993 ) ; r. von helmolt _
lett . * 71 * , 2331 ( 1993 ) ; s. jin _
science * 264 * , 413 ( 1994 ) .
see for example , a.p .
ramirez , j. phys . : condens .
matter * 9 * , 8171 - 8199 ( 1997 ) ; j.m.d .
coey , m. viret and s. von molnar , adv . in phys .
, in press . c. zener , phys
82 * , 403 ( 1951 ) .
anderson and h. hasegawa , phys .
rev * 100 * , 675 ( 1955 ) .
degennes , phys . rev .
* 118 * , 141 ( 1960 ) .
e. mller - hartmann and e. dagotto , phys .
b * 54 * , r6819 ( 1996 ) .
k. kubo and n. ohata , j. phys .
. jpn . * 33 * , 21 ( 1972 ) .
a.j . millis , p.b .
littlewood and b.i .
shraiman , phys .
* 74 * , 5144 ( 1995 ) .
h. rder , j. zang and a.r .
bishop , phys .
lett . * 76 * , 1536 ( 1996 ) ; j. zang , a.r .
bishop and h. rder , phys .
b * 53 * , r8840 ( 1996 ) .
millis , b.i .
shraiman and r. mueller , phys .
* 77 * 175 ( 1996 ) ; a.j .
millis , r. mueller and b.i .
shraiman , phys .
b * 54 * 5405 ( 1996 ) .
s. yunoki , j. hu , a.l .
malvezzi , a. moreo , n. furukawa and e. dagotto , phys .
. lett . * 80 * , 845 ( 1998 ) .
d. arovas and f. guinea , cond - mat/9711145 . edited by k. binder ( springer - verlag , berlin , 1979 ) .
serena , n. garca and a. levanyuk , phys .
b , * 47 * , 5027 ( 1993 )
. q. li , j. zang , a.r . bishop and c.m .
soukoulis , phys .
b , * 56 * , 4541 , ( 1997 ) .
e. dagotto , s. yunoki , a.l .
malvezzi , a. moreo , j. hu , s. capponi , d. poilblanc and n. furukawa , cond - matt/9709029 . | we have used monte carlo simulation techniques to obtain the magnetic phase diagram of the double exchange hamiltonian .
we have found that the berry phase of the hopping amplitude has a negligible effect in the value of the magnetic critical temperature . to avoid finite size problems in our simulations
we have also developed an approximated expression for the double exchange energy .
this allows us to obtain the critical temperature for the ferromagnetic to paramagnetic transition more accurately . in our calculations
we do not observe any strange behavior in the kinetic energy , chemical potential or electron density of states near the magnetic critical temperature .
therefore , we conclude that other effects , not included in the double exchange hamiltonian , are needed in order to understand the metal - insulator transition which occurs in the manganites
. -1.0 cm pacs number 71.10.-w , 75.10.-b |
weinstein manifolds are symplectic counterparts of affine ( stein ) complex manifolds .
they are exact symplectic manifolds which admit symplectic handle decompositions with isotropic core disks . in particular the index of any handle
is at most half of the dimension of the manifold .
we refer the reader to @xcite for precise definitions .
middle dimensional handles are called _ critical _ and handles below middle dimension _
subcritical_. given a weinstein manifold @xmath1 , its symplectic homology @xmath2 is a symplectic invariant which was first defined by a. floer and h. hofer in @xcite .
it vanishes if @xmath1 is subcritical , see @xcite .
a general weinstein manifold @xmath1 is obtained by attaching critical handles to a subcritical manifold @xmath3 along a collection @xmath4 of legendrian spheres in the contact manifold @xmath5 which is the ideal boundary of @xmath3 . in @xcite , @xmath2 was expressed in terms of the legendrian homology algebra of @xmath6 .
symplectic homology comes with tqft - like operations which on the homology level are all expressed in terms of a ( pair - of - pants ) product and a bv - operator , see @xcite . in this paper
we express the product and the bv - operator in terms of the legendrian homology algebra of @xmath4 and some additional operations on it .
the bv - operator is determined by the algebra itself , whereas the expression of the product involves a duality operation which can be identified with the first term in the extension of legendrian homology algebra to legendrian rational symplectic field theory ( sft ) .
the paper is organized as follows . in section [
sec : alg - prelim ] we introduce algebraic constructions which we later apply to legendrian homology algebras . in section
[ sec : leg - algebra ] we express the legendrian homology algebra results from @xcite in the terminology of section [ sec : alg - prelim ] and also include a discussion on the construction of a version of rational sft in the legendrian setting ( completion of this construction is work in progress ) . in section
[ sec : bv - leg ] we present the symplectic homology product and the bv - operator in terms of legendrian homology algebra and a duality operator and in section [ sec : examples ] we use the result to compute the symplectic homology , with product and bv - operator , of cotangent bundles of spheres .
while proofs are mostly omitted , we note that the proofs of the algebraic results from section [ sec : alg - prelim ] are fairly straightforward .
the authors thank m. abouzaid , s. ganatra , e. getzler , l. ng , and p. seidel for inspiring discussions .
in this section we describe a number of constructions in a purely algebraic setting . the constructions will be applied in geometrically relevant situations in sections [ sec : leg - algebra ] and [ sec : bv - leg ] .
we associate to a differential graded algebra @xmath7 over a field @xmath8 a pair of dual @xmath7-modules and a corresponding tensor algebra .
we also discuss cyclic versions of these objects , which are vector spaces over @xmath8 .
let @xmath9 be an algebra over a field @xmath8 of characteristic @xmath10 generated by idempotents @xmath11 , where @xmath12 , where @xmath13 is the kronecker delta .
we define a graded module @xmath14 over @xmath9 , which is generated by a countable set @xmath15 decomposed as a disjoint union @xmath16 .
the degree of an element @xmath17 will be denoted by @xmath18 .
the decomposition of @xmath15 induces a decomposition @xmath19 , where @xmath20 is generated by @xmath21 .
when applied on the left , the element @xmath22 , @xmath23 , acts as the identity on @xmath20 and as the @xmath10-map on @xmath24 , @xmath25 .
similarly , when applied on the right , the element @xmath22 acts as the identity on @xmath26 and as @xmath10 on @xmath27 , @xmath25 .
we define an extension @xmath28 of the module @xmath14 as follows .
the module @xmath28 is generated by @xmath29 , where @xmath30 and @xmath31 , where @xmath32 , for @xmath33 and @xmath34 , @xmath35 .
we define @xmath36,\ ] ] where @xmath37 denotes the @xmath38-dimensional module freely generated by the set @xmath39 and where @xmath40 $ ] denotes @xmath14 with grading shifted up by @xmath41 .
the module @xmath28 inherits a decomposition @xmath42 from @xmath14 , where @xmath43 & \text{if } i\ne j,\\ kx_j\oplus v^{(jj)}[1 ] & \text{if } i = j .
\end{cases}\ ] ] here the idempotents act as before on @xmath44 $ ] as a subset of @xmath28 , and as follows on the new generators : @xmath45 and @xmath46 , @xmath47 .
we will use the following notation . if @xmath17 then @xmath48 denotes the element in @xmath40\subset{\widehat}v$ ] that corresponds to @xmath49 . furthermore ,
if @xmath50 and @xmath51 are either @xmath14 or @xmath28 and if @xmath52 is an ordered pair of elements in @xmath53 then @xmath54 is _ composable _ if @xmath55 .
let @xmath56 be the tensor algebra over @xmath9 generated by @xmath14 with its natural grading .
then @xmath7 is generated additively by composable monomials @xmath57 , @xmath58 , @xmath59 , where we say that @xmath60 is composable provided @xmath61 is composable for @xmath62 .
( here , as often in the following , we will suppress the tensor sign in the notation for tensor products ) .
the decomposition of @xmath14 induces a corresponding decomposition of @xmath7 , @xmath63 , where @xmath64 is generated by composable monomials @xmath65 with @xmath66 for some @xmath67 and @xmath68 for some @xmath69 .
we will assume that the algebra @xmath7 has a differential @xmath70 of degree @xmath71 which leaves the subspaces @xmath64 , @xmath72 invariant ( i.e. , @xmath73 , @xmath74 satisfies the leibniz rule , and @xmath75 ) .
we will usually write @xmath76 for the homology algebra @xmath77 .
define @xmath80 then @xmath78 is a graded left - right module over the differential graded algebra @xmath7 which is additively generated by composable monomials @xmath81 , where @xmath82 , @xmath83 , @xmath84 , and @xmath85 , @xmath86 .
we say that a composable monomial @xmath60 is _ cyclically composable _ if the pair @xmath87 is composable .
let @xmath88 be the submodule generated ( over @xmath8 ) by cyclically composable monomials and define @xmath89 as the quotient space of @xmath90 in which monomials which differ by graded cyclic permutation are identified . if @xmath91 then we write @xmath92 for its equivalence class in @xmath89 , i.e. if @xmath93 denotes the natural projection then @xmath94 .
the differential @xmath70 induces a differential @xmath95 as follows . if @xmath96 , @xmath97 , and @xmath98 , then @xmath99 here @xmath100 where the @xmath9-module homomorphism @xmath101 is defined by @xmath102 and @xmath103 , for @xmath104 .
[ lm : dm2 ] the map @xmath105 is a differential ( i.e. @xmath106 ) which leaves @xmath90 invariant and descends to a differential @xmath107 , still denoted @xmath108 , in such a way that the natural projection @xmath109 is a chain map .
we write @xmath110 and @xmath111 , respectively , for the homologies @xmath112 and @xmath113 .
note that @xmath110 is a left - right module over @xmath76 , while @xmath111 is just a @xmath8-vector space .
[ rem : satellites ] we can write @xmath114 : = { ' } \!m^{{\rm cyc}}\oplus { ' ' } \!m^{{\rm cyc}}\ ] ] ( i.e. @xmath115 is the submodule of @xmath89 generated by @xmath116 and @xmath117 its complement ) .
let @xmath118 be the projection and let @xmath119 .
the second summand @xmath120 can be viewed as a non - commutative version of the `` lie algebra of vector fields '' on the space @xmath14 .
indeed , we can identify the cyclic monomial @xmath121 , @xmath122 , @xmath123 , with the vector field @xmath124 .
the differential @xmath125 can be naturally viewed as a vector field on @xmath14 , and hence an element of @xmath120 , given by @xmath126 .
for any @xmath127 its differential @xmath128 is just the lie bracket @xmath129 $ ] of vector fields , or equivalently the lie derivative @xmath130 .
this interpretation corresponds to the `` satellite philosophy '' from @xcite .
another interpretation was pointed out to the authors by m. abouzaid : @xmath111 is isomorphic to the hochschild homology @xmath131 .
recall that the module @xmath14 is generated over @xmath9 by a set @xmath15 and that the module @xmath78 is generated over @xmath7 as a left - right module by @xmath134 , where @xmath135 and @xmath31 .
let @xmath136 denote the @xmath8-vector space dual to @xmath28 .
then @xmath137 is the _ direct product _ of the 1-dimensional @xmath8-vector spaces which are generated by elements of @xmath138 , where @xmath139 and @xmath140 .
furthermore , we have @xmath141 in @xmath137 , the idempotent @xmath22 , @xmath23 , acts from the left as the identity on @xmath142 and as @xmath10 on @xmath143 , @xmath25 , and acts from the right as the identity on @xmath144 and as @xmath10 on @xmath145 , @xmath25 .
fix an integer @xmath146 ( it will correspond to half the dimension of the liouville domain @xmath3 in section [ sec : leg - algebra ] ) .
we endow @xmath137 with a grading defined as follows : @xmath147 define @xmath148 we define the tensor algebra @xmath149 as a certain completion of the free associative algebra generated by @xmath150 .
its elements are infinite series in the @xmath151-variables with polynomial coefficients in the @xmath152-variables ( compare to the above definitions of @xmath78 as an infinite sum and @xmath153 as an infinite product ) . for simplicity of notation and since all completions we consider below are of the same type as that of , we will , as in that equation , suppress completions from notation and use the direct sum symbol in decompositions .
consider the extension @xmath132 of @xmath154 obtained by adding multiplicative generators @xmath155 and @xmath156 of degrees @xmath157
which satisfy the following relations : @xmath158 and @xmath159 the algebra @xmath132 can be decomposed according to tensor type : @xmath160 , where @xmath161 is the @xmath8-subspace generated by monomials with @xmath162 factors from @xmath78 and @xmath163 from @xmath164 .
we decompose further : @xmath165 , where @xmath166 is the subspace spanned by all monomials of total @xmath167-degree @xmath168 and total @xmath169-degree @xmath170 .
in particular , @xmath171 , @xmath172 , and @xmath173 .
we also write @xmath174 , @xmath175 , @xmath176 , @xmath177 , and @xmath178 . for @xmath179 ,
let @xmath180 in analogy with the definition of @xmath181 in section [ sec : mmcyc ] , we define @xmath182 .
further , define @xmath183 as the quotient of @xmath184 obtained by dividing by the following _ graded cyclic permutation _ rule : if @xmath185 and if @xmath186 then @xmath187 , and we identify @xmath186 with @xmath188 , where @xmath189 we will often think of cyclic monomials as written on a circle @xmath190 with the first letter at @xmath191 .
this allows us to speak about generators in a cyclic monomial as ordered in the ( counter ) clockwise direction staring from some fixed generator in the monomial . in this language
, the above permutations correspond to rotations of the circle . the vector space @xmath192 inherits the decomposition by tensor type : @xmath193 if the monomials @xmath1 and @xmath194 differ by a graded cyclic permutation , i.e. if @xmath195 , then , for @xmath196 , @xmath197 and @xmath198 .
we thus define @xmath199 and @xmath200 , and note that this definition does not depend on choice of representative @xmath1 . a monomial @xmath201 is _ balanced _ if @xmath202 and @xmath203
. the subspace of @xmath183 generated by balanced monomials will be denoted by @xmath133 .
it contains balanced versions @xmath204 and @xmath205 of @xmath79 and @xmath206 , respectively .
more precisely , @xmath207 consider a monomial @xmath209 and fix a factor @xmath210 in @xmath1 .
then @xmath211 for monomials @xmath212 .
define @xmath213 where the sum is taken over all factors @xmath214 in @xmath1 .
it is straightforward to check that this gives a well defined operation and by definition @xmath215 . in particular , if @xmath216 then @xmath217 satisfies @xmath218 where @xmath219 is the operator @xmath220 defined in above .
furthermore , @xmath221 we next effectively break the cyclic symmetry of @xmath133 .
excited _ monomial is a cyclic monomial @xmath222 with one of its factors from @xmath223 distinguished .
we call the distinguished factor of an excited monomial @xmath1 the _ excited _ generator of @xmath1 .
the @xmath8-vector space generated by excited monomials from @xmath133 will be denoted by @xmath224 .
note that @xmath225 inherits decompositions from @xmath226 , @xmath227 , etc .
we set @xmath228 .
define the _ excitation operator _
@xmath229 to be the linear map which maps a monomial @xmath1 the sum of all its excitations . note that @xmath225 contain excited versions @xmath230 and @xmath231 of @xmath204 and @xmath205 , respectively , which are isomorphic to their non - excited counterparts .
( in the context of @xmath225 we think of @xmath232 and @xmath205 as generated by monomials in @xmath233 and @xmath234 _ without _ excited generator . ) we will use the following notation to identify excited generators in monomials : we underline the excited generator in a monomial in @xmath226 and write e.g. @xmath235 , @xmath236 @xmath237 , and @xmath238 .
we next define a product on @xmath224 .
consider two excited cyclic monomials @xmath239 .
fix a factor @xmath240 from @xmath1 and a factor @xmath241 from @xmath242 .
write @xmath243 and @xmath244 , where @xmath245 is the sign which arises from cyclically rotating @xmath1 so that @xmath246 is its last factor and where @xmath247 is the sign which arises from cyclically rotating @xmath242 so that @xmath248 is its first factor .
define the _ contraction _ operation @xmath249 , @xmath250 where the sum is taken over all pairs @xmath54 of factors @xmath240 from @xmath1 and @xmath241 from @xmath242 , and where @xmath251 if @xmath252 and @xmath253 then we define , for @xmath254 , the _ partial contraction _ @xmath255 as follows : @xmath256 where the sum is taken over all @xmath240 from @xmath1 and where @xmath257 is the @xmath258 factor from @xmath259 in @xmath242 counting _ counter - clockwise _ from the unique factor from @xmath260 .
similarly , if @xmath261 and @xmath262 then we define , for @xmath263 the _ partial contraction _ @xmath264 : @xmath265 where the sum is taken over all @xmath241 from @xmath242 and where @xmath266 is the @xmath267-th factor from @xmath268 in @xmath1 counting _ clockwise _ from the unique factor from @xmath259 .
let @xmath239 then @xmath269 .
further if @xmath270 then @xmath271 , and if @xmath272 then @xmath273 .
define the _ bracket _ or _ commutator _
@xmath274\colon { { \mathbf u}}^{{{\mathrm{bal}}}}\otimes{{\mathbf u}}^{{{\mathrm{bal}}}}\to{{\mathbf u}}^{{{\mathrm{bal}}}}$ ] as @xmath275=x\star y -(-1)^{|x| |y|}\,y\star x.\ ] ] [ l : jacobi ] the bracket satisfies the graded jacobi identity : @xmath276+(-1)^{|z||y|}[z,[x , y]]+(-1)^{|y||x|}[y,[z , x]]=0.\ ] ] let @xmath277 we associate with @xmath261 a multi - linear operation @xmath278 defined by @xmath279 note that if @xmath280 then @xmath281 , that if exactly one of the arguments lies in @xmath204 then the image lies in @xmath204 , and that the operation vanishes on @xmath163-tuples with more than component in @xmath204 .
similarly , @xmath282 defines a multi - linear operation @xmath283 , @xmath284 if @xmath285 then @xmath286 , if exactly one of the arguments lies in @xmath287 then the image lies in @xmath287 , and the operation vanishes on @xmath162-tuples with more than one component in @xmath287 . in the special case @xmath288
, we associate with @xmath289 two more operations , @xmath290 and @xmath291 where @xmath292 similarly , @xmath293 induces an operation @xmath294 , @xmath295 we observe that @xmath296 and @xmath297 .
we define special elements @xmath298 which derive from the differential on @xmath7 and which , in the spirit of sft , see @xcite , are called _ hamiltonians_. [ dfn : hamiltonian1 ]
@xmath299 we use the hamiltonian @xmath300 to define a differential on @xmath301 .
first define the _ exterior differential _ @xmath302 as follows .
consider the map @xmath303 which acts as the algebra differential @xmath70 on generators of @xmath132 from @xmath7 , which maps all other generators to @xmath10 , and which satisfies the graded leibniz rule .
then @xmath74 preserves @xmath184 , descends to @xmath192 where it preserves @xmath226 .
it then induces a map @xmath304 as follows .
the map is already determined on @xmath305 and if @xmath306 is a monomial in @xmath225 , where @xmath307 then @xmath308 .
then @xmath73 and we observe that @xmath309 satisfies the following _ master equation _ @xmath310 since @xmath311 , @xmath312 $ ] and can be rewritten in maurer - cartan form : @xmath313=0.\ ] ] decomposing @xmath314 , is equivalent to the following sequence of identities for the @xmath315-components of @xmath316 : @xmath317 the first identity in this sequence is @xmath318 define the differential @xmath319 , @xmath320.\ ] ] ( note that is equivalent to @xmath321 . )
the differential interacts with the bracket in the following way .
[ lm : star ] for @xmath322 , @xmath323)=[d_ux , y]+(-1)^{|x|}[x , d_uy].\ ] ] in particular , the operation @xmath274 $ ] descends to the homology @xmath324 .
we observe that @xmath325 leaves @xmath204 , @xmath326 , @xmath287 , and @xmath327 invariant . for simple notation
we write @xmath328 for the restriction of @xmath325 to any one of these subspaces .
consider the isomorphisms @xmath329 and @xmath330 , @xmath331 the maps @xmath332 and @xmath333 are chain isomorphisms .
the following result is a consequence of lemma [ lm : star ] .
if @xmath334 satisfies @xmath335 and has the correct tensorial type to define one of the operations , see section [ sec : convol ] , then this operation is a chain map . if furthermore @xmath336 for some @xmath337 then the operation is chain homotopic to @xmath10 .
we introduce product operations on @xmath326 and @xmath327 .
recall the operation @xmath339 defined by the formula @xmath340 see .
. then @xmath342 .
[ prop : dual - product ] the operation @xmath343 descends to an associative product on the homology @xmath344 which satisfies the following graded commutativity relation for the grading shifted by @xmath41 : @xmath345 the homology class of the cycle @xmath346 is the unit for this product on @xmath338 .
in fact , @xmath347 is a unit for @xmath348 already on the chain level , before passing to homology .
consider the second and third identities satisfied by the hamiltonian @xmath349 : @xmath350=0,\\\label{eq : i_3 } & d{{h}}^{1}_{3}+[{{h}}^1_1,{{h}}^1_3]+ { { h}}^1_2{\mathop{\star}\limits}_1 { { h}}^1_2 + { { h}}^1_2{\mathop{\star}\limits}_2 { { h}}^1_2=0.\end{aligned}\ ] ] here is the chain map equation for @xmath351 , and shows that this product is associative on the homology level . indeed , @xmath352 the unit property can be checked directly if one observes that the sum @xmath51 of all terms in @xmath353 with excited generator from @xmath354 and which contain at least one @xmath355-factor , @xmath35 is @xmath356 hence , taking into account that @xmath357 we have for any @xmath358 @xmath359 to verify the commutativity relation , let @xmath1 be a @xmath360-cycle in @xmath361 .
applying the @xmath362-operator to the equation @xmath363=0 $ ] gives @xmath364 and thus , on the level of homology , @xmath365 and we calculate @xmath366 if there exists an element @xmath368 such that @xmath369 satisfies the master equation @xmath370=0,\ ] ] then we say that @xmath371 is _ compatible _ with @xmath372 and we call @xmath373 the _ full ( rational ) hamiltonian_. as we shall see later , when @xmath7 is a legendrian homology algebra then we can always find @xmath371 compatible with @xmath372 using a relative sft formalism associated with parallel copies of the legendrian submanifold in the spirit of @xcite .
assume now that @xmath374 , and in particular @xmath375 , which satisfies exists and is given .
then we define the operation @xmath376 , @xmath377 note that if @xmath378 then @xmath379 .
[ thm : product - properties ] @xmath380 1 .
the operation @xmath381 descends to an operation on the homology @xmath382 where it can be expressed by the following three equivalent expressions : @xmath383 equivalently , the following diagram commutes on the homology level , @xmath384^{({{h}}^1_2)^{{\widetilde}{{\downarrow\uparrow}}}}&\\ { { \mathbf m}}^{{\mathrm{bal}}}\otimes { { \mathbf m}}^{{\mathrm{bal}}}\ar[ur]^ { ( { { h}}^2_0)^{\circlearrowleft}\otimes{\mathrm{id}}\;}\ar[rr]^{\boxdot\;\simeq\ ; -({{h}}^2_0{\mathop{\star}\limits}_2 { { h}}^1_2)^{\downarrow}}\ar[dr]_{{\mathrm{id}}\otimes({{h}}^2_0)^{\circlearrowleft}\;\ ; } & & { { \mathbf m}}^{{\mathrm{bal}}}\\ & { { \mathbf m}}^{{\mathrm{bal}}}\otimes{\left(}{{\mathbf m}}^\ast{\right)}^{{\mathrm{bal}}}\ar[ur]_{({{h}}^1_2)^{{\widetilde}{{\uparrow\downarrow}}}}&}\ ] ] where @xmath385 for @xmath386 and @xmath387 .
the operation @xmath388 is associative and graded commutative on homology , i.e. if @xmath389 then @xmath390 the above result is a consequence of the master equation which in particular implies the identities @xmath391=d{{h}}^{2}_{0}+{{h}}^2_0\star { { h}}_1 ^ 1=0,\\\label{eq : h21 } & dh^{2}_{1}+[h^1_1,h^2_1]+h^2_0{\mathop{\star}\limits}_1 h^1_2 + h^2_0{\mathop{\star}\limits}_2h^1_2=0.\end{aligned}\ ] ] a straightforward calculation shows that @xmath392 . by , @xmath393 , by , @xmath394 , and therefore @xmath395 is a cycle in @xmath396 .
similarly , @xmath397 is a cycle and implies that their homology classes coincide .
this shows that the product descends to homology and implies the first equation of part 1 of the theorem .
the alternative expressions for @xmath388 in part 1 are immediate from the definitions of the operations @xmath398 , @xmath399 , and @xmath400 .
@xmath401 at -30 460 @xmath402 at 90 460 @xmath402 at 208 460 @xmath402 at 316 460 @xmath402 at 428 460 @xmath403 at 550 460 @xmath404 at -30 300 @xmath402 at 72 300 @xmath402 at 175 300 @xmath402 at 264 300 @xmath402 at 360 300 @xmath402 at 450 300 @xmath403 at 550 300 @xmath405 at 170 120 @xmath406 at 400 120 upwards and @xmath162 downwards @xmath41-valent vertices represents @xmath407 .
the dot represents the @xmath408-operator .
equations @xmath401 and @xmath404 are consequences of @xmath409 ( terms from @xmath410 are not shown in the figures).,title="fig : " ] to verify the commutativity we observe that given a cycle @xmath411 , using the homological identity @xmath412 , we calculate @xmath413 finally , the proof of associativity is pictorially illustrated on figure [ fig : calc ] .
the products @xmath388 on @xmath326 and @xmath348 on @xmath327 are related via the chain map @xmath414 in the following way .
the map induced by @xmath415 on homology is a homomorphism of rings @xmath416 .
moreover , if there is a unit @xmath417 for @xmath388 on @xmath367 , i.e. @xmath418 , for all @xmath419 , then @xmath420 is injective . indeed , if @xmath421 then , taking into account that @xmath422 , we calculate @xmath423 to verify the injectivity of @xmath420 we use one of equivalent definitions of @xmath388 from theorem [ thm : product - properties ] , part 1 : @xmath424 and hence , @xmath425 implies @xmath426 using the isomorphism @xmath427 , @xmath428 defined above we transport the product to @xmath79
. given @xmath429 , we define @xmath430 note that the homomorphism @xmath431 shifts the grading by @xmath432 , and hence the commutation equation for @xmath388 translates to the equation @xmath433 let us also point out that the degree of the product @xmath434 is equal to @xmath435 .
in this section we first reformulate the main result relating legendrian homology algebra and symplectic homology ( * ? ? ?
* corollary 5.7 ) in the terminology of section [ sec : alg - prelim ] , and second give a brief sketch of the geometry that enters the construction of the full rational hamiltonian in the framework of legendrian sft with focus on the parts of it that enter the expression of the product .
let @xmath437 be a liouville domain with contact boundary @xmath438 and let @xmath439 be a union of legendrian spheres . let @xmath15 be the set of reeb chords of @xmath4 and consider the decomposition @xmath440 , where a chord in @xmath21 starts on @xmath441 and ends on @xmath442 .
let @xmath14 be the graded module @xmath443 generated by the set @xmath15 with decomposition @xmath444 , where @xmath20 is generated by @xmath21 .
then the legendrian homology algebra @xmath445 is the tensor algebra @xmath7 associated to @xmath14 as in and the differential @xmath446 gives a differential graded algebra @xmath447 , see ( * ? ? ?
* section 4.1 ) .
let @xmath448 be the chain complex defined in .
the following result is a consequence of ( * ? ? ?
* remark 7.4 ) .
[ lm : lhho ] the chain complex @xmath89 is canonically isomorphic to the chain complex @xmath436 defined in ( * ? ? ?
* section 4.5 ) . in ( * ? ? ?
* corollary 5.7 ) it was shown that if @xmath449 ( e.g. if @xmath3 is a subcritical weinstein manifold ) and if @xmath1 is obtained from @xmath3 by attaching lagrangian @xmath450-handles along @xmath4 , then there is a quasi - isomorphism @xmath451 where @xmath452 denotes a chain complex with homology equal to @xmath2 .
the differential in the legendrian algebra @xmath453 is defined by counting elements of moduli spaces of rigid holomorphic disks in @xmath454 , with boundary on @xmath455 , and with exactly one positive and several negative punctures , see e.g. ( * ? ? ?
* section 4.1 ) .
the formalism of legendrian rational sft comes from algebraic structures associated with holomorphic disks with an arbitrary number of positive and negative punctures . as is well known , in order to construct such a theory one needs to handle boundary cusp - degenerations of holomorphic disks .
the problem of boundary cusp - degeneration can be solved by incorporating string topology operations into the sft formalism .
the corresponding theory has not yet been constructed in sufficient generality for it to be applicable in the setting of this paper .
( see , however @xcite for the case of @xmath41-dimensional legendrian knots in @xmath456 ) .
however , the algebraic formalism described in section [ sec : alg - prelim ] requires a different version of the full rational hamiltonian @xmath374 , in the spirit of @xcite .
it arises naturally when one considers moduli spaces of holomorphic disks with boundaries on multiple parallel copies of @xmath455 .
let us first consider the term @xmath375 .
this is the only term besides @xmath353 which is needed for defining the product on @xmath79 , see figure [ fig : copies ] .
we recall here that @xmath457 for any @xmath458 is determined by the differential on the legendrian algebra @xmath7 . for
each @xmath459 fix a point @xmath460 .
let @xmath461 , where @xmath462 and @xmath463 , @xmath464 , be an excited cyclic monomial .
we associate to it a moduli space @xmath465 . here
the moduli space @xmath466 consists of holomorphic disks anchored in @xmath3 , see ( * ? ? ?
* section 2.2 ) , @xmath467 where @xmath468 is a set of cyclically ordered boundary punctures on @xmath469 , with the following properties : * at the puncture @xmath470 , @xmath471 is asymptotic to the chord @xmath472 at @xmath473 ; * if @xmath474 then at the puncture @xmath475 @xmath471 is asymptotic to the chord @xmath476 at @xmath477 ; * if @xmath478 then , @xmath471 extends continuously over @xmath475 , and @xmath479 . the moduli space @xmath480 consists of holomorphic disks connected by morse flow lines . to define it , choose auxiliary morse functions @xmath481 , @xmath482 , each with exactly two critical points . for @xmath483 define the auxiliary moduli space @xmath484 of pairs @xmath485 of holomorphic disks anchored in @xmath3 , @xmath486 where @xmath487 are sets of cyclically ordered boundary punctures on @xmath469 , with the following properties . at @xmath488 , @xmath489 satisfies condition @xmath490 if @xmath491 or condition @xmath492 if @xmath493 . at @xmath494 , @xmath489 satisfies condition @xmath495 . at the puncture @xmath496 , @xmath497 , @xmath489 extends continuously and the following condition holds : * the points @xmath498 and @xmath499 lie in the same component @xmath500 and there exists a trajectory of @xmath501 which starts at @xmath498 and ends at @xmath499 , where @xmath502 is the auxiliary morse function on @xmath503 . here the gradient is taken with respect to the metric induced by the symplectic form and the almost complex structure . finally , define @xmath504 given @xmath505 , let @xmath506 if @xmath507 or @xmath508 .
define @xmath509 , @xmath35 .
then we have the dimension formula @xmath510 in particular in terms of the grading on @xmath132 , for the balanced monomial @xmath511 we have @xmath512 define @xmath513 as @xmath514 , where @xmath515 where the sum is taken over all balanced monomials @xmath516 which additively generate @xmath517 and for which @xmath518 . here
@xmath519 is the algebraic number of @xmath41-dimensional components of the moduli space @xmath520 .
as will be briefly discussed below , the above definition of @xmath521 can generalized to more than two positive punctures and as a consequence we get the following result .
[ prop : legendrian - master ] there exists an element @xmath522 , with trivial component in @xmath523 , such that @xmath524 complements @xmath372 to the full rational sft hamiltonian ( i.e. if @xmath525 then @xmath526 ) . here
the components @xmath527 , @xmath528 , of the full rational hamiltonian @xmath529 satisfies the relation @xmath530 , @xmath531 , and are defined through a construction which generalizes that used in the definition of @xmath375 above .
these higher order components of the hamiltonian are not needed for the definition of the product ( but some of them are needed for proving its properties ) and will not be defined in this paper .
we give however a brief description : @xmath532 is defined using moduli spaces @xmath533 of so - called _ generalized holomorphic disks_. generalized holomorphic disks are combined objects consisting of holomorphic disks with boundary on one copy of @xmath4 connected by gradient flow trees associated to auxiliary morse functions on @xmath4 , an example is depicted in figure [ fig : gendisk ] , see @xcite for similar considerations .
the elements of the moduli space @xmath480 defined above are examples of generalized holomorphic disks where the gradient flow tree consists of only one gradient trajectory . a more general configuration is shown in figure [ fig : gendisk ] . more precisely , if @xmath534 is a monomial in @xmath535 , where @xmath536 and where @xmath537 , @xmath83 , then we consider the moduli space @xmath533 of generalized holomorphic disks with asymptotic data according to @xmath538 , i.e. positive punctures at @xmath539 if @xmath540 , negative punctures at @xmath541 if @xmath542 , auxiliary negative punctures at the reeb chords in the monomials @xmath543 , and these punctures appearing along the boundary of the disk in the cyclic order given by @xmath538 .
the dimension of @xmath533 is given by @xmath544 in particular for @xmath545 we have @xmath546 and @xmath407 is defined as @xmath547 where the sum ranges over all @xmath548 with @xmath549 .
with boundary on @xmath455 . ] in this geometric context , the excitation operator @xmath550 has the following meaning .
given a generalized holomorphic disk with boundary in the ordered union of ( symplectizations of ) parallel copies of a legendrian manifold @xmath4 we always insist that the boundary arcs of the disk between punctures corresponding to generators from @xmath551 be mapped to different copies of @xmath4 .
we also require that when going around the boundary of the disk in the counter - clockwise direction we jump at every such puncture except one to a higher copy of @xmath4 in terms of the natural ordering corresponding to the direction of the reeb vector field .
the unique puncture where we jump down corresponds to the excited generator of the monomial corresponding to the boundary punctures . here
the moduli spaces corresponding to different assignment of the distinguished jump - down " punctures are canonically diffeomorphic , and the operator @xmath550 just associates with this universal moduli space the moduli spaces corresponding to all possible assignments of the jump - down " vertex .
[ r : gluingsign ] the geometric motivation for the `` gluing sign '' in the equations @xmath552 , @xmath35 is as follows .
the algebraic variables @xmath355 and @xmath553 , @xmath35 has two dual geometric meanings as follows : when @xmath554 is glued to @xmath555 , @xmath554 corresponds to a positive puncture at a short reeb chord between copies of @xmath4 at the minimum of the morse function @xmath556 and @xmath555 corresponds to a negative puncture at this chord , when @xmath557 is glued to @xmath355 , @xmath557 corresponds to a negative puncture at the maximum of @xmath558 and @xmath355 to a positive puncture at this maximum . in a coherent orientation scheme
there is a sign difference ( a factor @xmath559 ) between the gluings at these two types of punctures .
let @xmath3 , @xmath6 , and @xmath1 be as above .
let @xmath453 and @xmath448 . then gives a quasi - isomorphism @xmath560 .
in this section we present the operations on @xmath89 which correspond to the product and the bv - operator on @xmath452 under @xmath420 .
let @xmath561 be the full rational hamiltonian determined by @xmath6 , see .
then proposition [ prop : legendrian - master ] implies that @xmath375 satisfies .
hence , according to theorem [ thm : product - properties ] and , there exists a product @xmath562 of degree @xmath435 which is associative and commutative on homology and which is defined by conjugating one of the following homotopically equivalent formulas for the product on @xmath361 with @xmath431 : @xmath563 see figure [ fig : copies ] .
[ thm : sh - product ] if @xmath449 ( e.g. if @xmath3 is subcritical ) then , under the map on homology induced by the quasi - isomorphism @xmath564 , see , the pair of pants product on @xmath2 corresponds to the product @xmath434 on @xmath565 .
@xmath566 at 31 114 @xmath567 at 140 114 @xmath568 at 88 35 @xmath566 at 340 130 @xmath568 at 340 50 @xmath569 at 672 190 @xmath570 at 713 57 @xmath569 at 515 80 , @xmath571 are parallel copies of @xmath4 .
the leftmost disk has three punctures at mixed ( i.e. connecting different components ) reeb chords , one positive and two negative , and auxiliary negative punctures at pure ( i.e. connecting a component to itself ) chords which are not shown in the picture .
the middle disk has two mixed positive punctures and auxiliary pure negative punctures , the latter again not shown .
the rightmost picture shows how the disks are joined at a common mixed chord between @xmath568 and @xmath566 to form the tensor , with inputs and output as indicated , which represents the product.,title="fig : " ] let @xmath453 and consider the linear map @xmath572 of degree @xmath41 defined as follows on monomials @xmath573 , @xmath84 , @xmath122 : @xmath574 the map @xmath575 is a chain map and if @xmath449 then , under the map on homology induced by the quasi - isomorphism @xmath564 , see , the bv - operator on @xmath2 corresponds to the operator @xmath575 on @xmath565 .
in this section we compute the symplectic homology with product and bv - operator for cotangent bundles of spheres presented as the result of attaching a lagrangian @xmath450-handle to the legendrian unknot in the boundary of the ball .
consider the legendrian unknot in @xmath577 with coordinates @xmath578 and contact form @xmath579 .
figure [ fig : unknot ] shows the projection of the legendrian unknot @xmath4 for @xmath580 to the @xmath581-plane . to define the legendrian unknot of dimension @xmath583 consider the @xmath41-dimensional unknot @xmath4 with reeb chord over @xmath584 and symmetric with respect to the rotation by a half turn around the @xmath585 axis . include @xmath4 into the @xmath586-subspace of @xmath577 and take the @xmath587-dimensional unknot as the orbit of @xmath4 under the action by @xmath588 given by @xmath589 for @xmath590 . for simplicity , we let @xmath4 denote the legendrian unknot in @xmath577 for any @xmath450 .
consider @xmath591 and include it into a darboux chart in the boundary of the @xmath592-ball @xmath593 . then attaching a lagrangian @xmath450-handle to @xmath593 along @xmath4 gives a symplectic manifold which is symplectomorphic to the cotangent bundle @xmath594 of the @xmath450-sphere . as explained in @xcite , in order to determine
@xmath445 it suffices to study reeb chords and holomorphic disks that are contained in the darboux chart .
the unknot @xmath4 has only one reeb chord which we denote @xmath49 and @xmath595 .
the moduli space of holomorphic disks with positive puncture at @xmath49 and a marked point on the boundary evaluates with degree @xmath596 to @xmath4 . to see this
, note that for @xmath580 , holomorphic disks that are rigid up to translation correspond to polygons with boundary on the knot diagram and with convex corners , where at positive punctures the boundary orientation goes from the lower to the upper strand and vice versa at negative punctures .
there are thus two disks with evaluation maps that give a degree @xmath596 map since we use the null - cobordant spin structure on the circle to orient moduli spaces .
for @xmath597 we note that we may take the almost complex structure to be invariant with respect to the @xmath588 action and it follows that the moduli space of holomorphic disks with positive puncture at @xmath49 is diffeomorphic to @xmath598 and that , when equipped with a marked point on the boundary , it evaluates with degree @xmath599 .
the above description of moduli spaces implies in particular that the differential @xmath74 on @xmath600 is trivial , i. e. , @xmath601 .
consequently , the differential on @xmath361 satisfies @xmath602=[{{h}}_1^{1},x],\ ] ] where @xmath603 thus @xmath604 and @xmath565 ( recall the chain isomorphism @xmath605 , see ) are generated by the cycles listed in tables [ tab : shn-1odd ] and [ tab : shn-1even ] when @xmath583 is even and odd , respectively . in order to compute the product
we need to determine @xmath607 and @xmath608 .
consider first @xmath607 ( see definition [ dfn : hamiltonian1 ] ) : @xmath609 see remark [ r : signs ] for a discussion of the sign in this formula . in our description of @xmath170 in terms of generalized holomorphic disks , the disks which give rise to are degenerate .
the first two terms comes from the trivial strip of @xmath49 with an unconstrained morse puncture lying in one of the two boundary components .
the last term comes from the @xmath610-family of three - punctured constant disks lying in @xmath611 with one puncture constrained to lie in @xmath611 and remaining two punctures unconstrained .
next , consider @xmath612 . since the evaluation map from the moduli space of holomorphic disks with positive puncture at @xmath49 into @xmath4 has degree @xmath596
there is algebraically @xmath596 disk which satisfies a point constraint at @xmath613 and we conclude from the description of generalized holomorphic disks that @xmath614 see remark [ r : signs ] for a discussion of the sign in this formula .
[ r : signs ] orientations of moduli spaces are fixed by orienting capping operators at reeb chords , see @xcite for details .
the overall sign in is independent of the choice of capping operator for @xmath49 since variables associated to @xmath49 appears quadratically in @xmath615 , the overall signs in changes with the orientation data at @xmath616 , and the overall sign in changes , independently , with the orientations of capping operators at both @xmath616 and @xmath49 .
it follows that we can choose any sign combination in the two equations and in particular the one under consideration . to compute @xmath617 acting on @xmath618 we use the components of @xmath619 in which the puncture at the second leg is excited .
we will denote the sum of these by @xmath620 .
the relevant parts of the hamiltonian are thus , see and , ( recall gradings : @xmath621 , @xmath622 , @xmath623 , and @xmath624 ) @xmath625 and we get @xmath626 using for the tensor of the product @xmath627 then gives the following expression for the product , with @xmath628 : if @xmath583 is odd then @xmath629 and if @xmath583 is even then for @xmath630 @xmath631 we conclude from this calculation that the commutative product @xmath632 which corresponds to the pair of pants product on @xmath633 has unit represented by the class @xmath634 and multiplication according to the above formulas after substituting @xmath635 with @xmath41 everywhere .
y. eliashberg , m. gromov , convex symplectic manifolds , in `` several complex variables and complex geometry , part 2 '' ( santa cruz , ca , 1989 ) , 135162 , _ proc . sympos .
pure math . _ * 52 * , part 2 , amer .
soc . , providence , ri , 1991 .
a. floer , h. hofer , symplectic homology .
i. open sets in @xmath640 , _ math .
z. _ * 215 * ( 1994 ) , no .
1 , 3788 . p. seidel , a biased survey of symplectic cohomology . in _ current developments in mathematics _
( harvard , 2006 ) , intl . press , 2008 , 211253 . a. weinstein , contact surgery and symplectic handlebodies , _ hokkaido math . j. _
, * 20*(1991 ) , 241251 . | this research announcement continues the study of the symplectic homology of weinstein manifolds undertaken in @xcite where the symplectic homology , as a vector space , was expressed in terms of the legendrian homology algebra of the attaching spheres of critical handles .
here we express the product and @xmath0-operator of symplectic homology in that context . |
dynamical mass generation in gauge theories has long been studied using the gap equation ; i.e. , the dyson - schwinger equation ( dse ) for the dressed fermion propagator @xcite .
the gap equation is nonlinear in the dressed - fermion mass - function , @xmath0 ; and it was found early @xcite that a simple ( rainbow ) truncation of this equation possess a countable infinity of solutions , with each one characterized by : its magnitude and sign at @xmath1 ; and the number of zeros it possesses on @xmath2 .
the existence of this tower of solutions to the truncated equation almost immediately fostered speculation about their possible physical consequences @xcite .
the non - uniqueness of solutions to the rainbow gap equation was later rediscovered independently , with the explicit identification of three distinct solutions that bifurcate from the trivial @xmath3 solution as the coupling strength is increased @xcite ; and then again about a decade later @xcite .
more recently , solutions of the gap equation with repeated zero crossings were also found using a simple generalization of the rainbow truncation @xcite ; and in ref.@xcite their existence was firmly established using a procedure that enables one to obtain all solutions of the gap equation in the presence of a variety of external control parameters . in considering all these contributions one
is naturally led to wonder about the general conditions under which a gauge theory s gap equation will admit solutions with zeros in addition to that single positive - definite solution which is conventionally associated with dynamical chiral symmetry breaking ( dcsb ) .
herein we provide a partial answer , using quantum electrodynamics in three dimensions ( qed3 ) as an illustrative tool .
our choice of theory is motivated first by a number of properties that qed3 shares with quantum chromodynamics ( qcd ) .
for example , quenched qed3 possesses a nonzero string tension @xcite ; and this feature persists in the unquenched theory if massive fermions circulate in the photon vacuum polarization @xcite .
in addition , since qed3 is super - renormalizable , it has a well - defined chiral limit and therefore admits the possibility and study of dcsb .
dcsb in qcd underlies the success of chiral effective field theory , explains the origin of constituent - quark masses @xcite and hence the vast bulk of visible mass in the universe @xcite , and quite probably shares its origin with light - quark confinement @xcite .
there is also an applied interest and relevance because qed3 is used in condensed matter physics as an effective field theory for high - temperature superconductors @xcite and graphene @xcite .
our partial answer to the problem of anticipating the appearance of solutions to the gap equation with one or more damped oscillations is based upon the following observation @xcite : in rainbow truncation the dressed - fermion mass function satisfies a nonlinear hammerstein integral equation @xcite .
the conditions under which such equations admit oscillatory solutions were long ago elucidated in the mathematics literature @xcite .
we find that if the dressed - fermion - photon vertex produces a gap equation kernel that satisfies those conditions , then solutions with zeros are readily found numerically .
this is not the case when the vertex produces a kernel that violates the conditions .
our material is organized as follows . in sec.[theory ] we recapitulate briefly upon the theory of nonlinear hammerstein integral equations , listing the criteria which guarantee the existence of oscillatory solutions for the mass function and providing a simple example .
we analyze qed3 from the perspective of hammerstein integral equations in sec.[qed3-gap ] ; and in sec.[gauge ] we explain that if oscillatory solutions exist in landau gauge , then they are present in all covariant gauges .
we summarize in sec.[final ] .
a nonlinear , homogeneous hammerstein integral equation of the first kind has the form @xcite @xmath4 where @xmath5 is the equation s solution , @xmath6 is a nonlinear functional of @xmath5 , @xmath7 is the kernel of the equation and @xmath8 is a real number . a solution of eq.([eq : hammer ] )
is a pair @xmath9 , where @xmath10 and @xmath5 is a continuous function on @xmath11 $ ] .
suppose now that : * @xmath12 and @xmath13 are continuous , and @xmath14 for @xmath15 @xmath16\times\mathbb{r}$ ] ; * @xmath17 for @xmath18\times\mathbb{r}$ ] ; * @xmath19 for @xmath20 $ ] and @xmath21 ; * @xmath22 when @xmath23 uniformly for @xmath20 $ ] ; * @xmath24 for @xmath18\times\mathbb{r}$ ] ; and * @xmath7 is a symmetric oscillation kernel on @xmath25\times[0,1]$ ] . in such circumstances
it is known @xcite that if @xmath8 exceeds @xmath26 , where @xmath27 , @xmath28 , is the @xmath29-th eigenvalue of the linear operator associated with eq.([eq : hammer ] ) ; i.e. , the integral equation obtained through the replacement @xmath30 then the hammerstein equation possesses at least @xmath31 nontrivial solutions with zeros in @xmath32
. it was subsequently demonstrated @xcite that eq.([eq : hammer ] ) possesses infinitely many solutions , each distinguished by its number of zeros , for any prescribed value of @xmath33 thus , as with differential equations , at least one large class of integral equations possesses an enumerable infinity of solutions with damped oscillations .
a simple , physical realization of the hammerstein integral equation may arise in connection with graphene , a one - atom - thick layer of graphite , wherein quasiparticle excitations are described by the massless dirac equation in @xmath34-dimensions .
this is the euler - lagrange equation for fermions in a @xmath34-dimensional version of quantum electrodynamics , except that the speed of light is replaced by the fermi velocity ; viz .
, @xmath35 thus , the interaction strength in suspended graphene is : @xmath36 in which case the model s gap equation can produce solutions that dynamically break chiral symmetry ; and such solutions express a realignment of the ground - state within the sample .
for graphene on a substrate , however , the effective coupling is screened , owing to a dielectric constant , @xmath37 , associated with the substrate : @xmath38 and for @xmath37 sufficiently large , the symmetry breaking solutions will disappear .
the phase transition in this model is described by the following gap equation @xcite @xmath39 with @xmath40 where @xmath41 is the complete elliptic integral of the first kind and all mass - dimensioned quantities have been rescaled by an ultraviolet cutoff @xmath42 , where @xmath43 characterizes the lattice spacing in the sample .
it is straightforward to verify that the elements in eq .
satisfy the conditions @xmath44@xmath45 , @xmath46 and hence there is an @xmath47 such that for @xmath48 the gap equation possesses a countable infinity of distinct solutions , distinguished one from another by the number of zeros they exhibit . in this case , @xmath49 @xcite ; and the first three solutions are depicted in fig.[fig : graphene ] . ) , plotted as @xmath50 .
solid curve no zeros ; long - dashed curve one zero ; and short - dashed curve two zeros . with @xmath51 , in units of @xmath42 , the true magnitude of the solutions is , respectively , o@xmath52 , o@xmath53 , o@xmath54 .
[ fig : graphene ] ]
the fully dressed fermion propagator is commonly written in one of the following forms : @xmath55\,,\\ & = & -i \gamma\cdot p\ , \sigma_v(p ) + \sigma_s(p)\ , , \label{spsigmavs}\end{aligned}\ ] ] where @xmath56 is a current - fermion lagrangian mass and the self - energy is given by @xmath57 here , @xmath58 and @xmath59 are the dressed photon propagator and dressed fermion - photon vertex , respectively ; and @xmath60 is the coupling , which has mass dimension one .
since qed3 is super - renormalizable and therefore no ultraviolet divergence can arise whose regularization would introduce a new mass - scale , @xmath60 defines the natural scale of the theory .
the quenched , rainbow truncation of eq . is obtained with @xmath61\frac{1}{q^2}\\ \gamma_\nu(q , p ) & \to & \gamma_\nu\,.\end{aligned}\ ] ] we work primarily in landau gauge because it occupies a special place in gauge theories @xcite .
it is the gauge in which any sound _ ansatz _ for the fermion - photon vertex can most legitimately be described as providing a pointwise accurate approximation .
the vertex in any other gauge is then defined as the landau - khalatnikov- fradkin ( lkf ) transform @xcite of the landau gauge _
ansatz_. the sensible implementation of this procedure guarantees gauge covariance and hence obviates any question about the gauge dependence of gauge invariant quantities .
we expand on these points in sec.[gauge ] .
after a little dirac algebra and evaluation of the angular integrals in eq .
, one finds @xmath62 and a single equation for the dressed - fermion mass function ( @xmath63 $ ] ) : @xmath64 one may now exploit the fact that qed3 is a super - renormalizable theory , in which the explicit mass - scale is defined by @xmath60 and effects associated with dynamical mass generation are an order of magnitude smaller ( see fig.[fig : graphene ] ) .
ultraviolet momenta therefore have no influence on nonperturbative phenomena and hence , at no cost , one may introduce a mass - scale @xmath65 , @xmath66 but @xmath67 , @xmath68 fixed , such that the following gap equation is equivalent to eq . :
@xmath69 where @xmath70 , @xmath71 , @xmath72 .
in some of our subsequent analysis and illustrations , we will employ the approximation @xmath73 which is precise for @xmath74 and @xmath75 , and , in fact , quite accurate in general for such a weakly singular kernel @xcite .
this leads from eq . to @xmath76
\ ; , \label{sde - masslog}\end{aligned}\ ] ] an inhomogeneous hammerstein equation , which becomes homogeneous when @xmath77 . if one sets @xmath78 in the denominator of the integrand in eq .
, then one arrives at a linear equation that can be treated analytically . in the chiral limit , however , unlike its parent , that equation is invariant under rescaling : @xmath79 , with @xmath80 an arbitrary constant , and does not possess an infrared regularizing mechanism .
these qualitative differences are both remedied by the simple expedient of introducing an infrared cutoff , @xmath81 , with @xmath82 : @xmath83 \;. \label{sde - masslogk}\ ] ] if one relaxes @xmath84 mildly , to the extent that the condition is satisfied on @xmath85\times\mathbb{r}$ ] , this is an homogeneous hammerstein equation . .
the nature of the solution evolves with @xmath81 : it acquires an additional zero each time @xmath81 decreases through one of a countable infinity of thresholds .
[ fig : nodan ] ] so long as the following boundary conditions are applied : @xmath86 eq . is equivalent to the following second - order differential equation : @xmath87 where @xmath88 .
this problem has the solution @xmath89 where @xmath90 is a bessel function . for @xmath91 ,
the solution is monotonically decreasing and positive definite .
this remains true as @xmath81 is reduced until , at @xmath92 , there is a qualitative change in the solution , with the appearance of a zero .
a second zero appears for @xmath93 ; and this pattern , illustrated in fig.[fig : nodan ] , continues _ ad infinitum_. this discussion illustrates the statements of sec.[theory - h ] and may be viewed as establishing that , in the chiral limit , the gap integral equation , eq .
, possesses a countably infinite number of simultaneous solutions , which are distinguished by their magnitude and the correlated number of zeros .
sturm - liouville theory provides another perspective on the guaranteed simultaneous existence of a countable infinity of solutions to eq .. consider the linearized version of eq . in the chiral limit : @xmath94 the solutions of this equation can not be obtained analytically .
however , it can be viewed as one in a class of integral equations with the general form : @xmath95 where the kernel @xmath96 is real and symmetric under @xmath97 .
this is an homogeneous fredholm equation of the first kind . introducing a @xmath98-point quadrature rule , eq .
is replaced by a system of coupled algebraic equations : @xmath99 where @xmath100 , @xmath101 , and the set @xmath102 contains the weights associated with the chosen quadrature .
one may always choose a rule with @xmath103 , so that @xmath104 corresponds to a symmetric matrix .
then , defining @xmath105 and @xmath106 , eq . can be expressed @xmath107 this is an eigenvalue problem for a real , symmetric matrix , @xmath108 ; and different eigenvectors correspond to distinct solutions of the gap equation . .
[ fig : num ] ] we have applied the procedure just described to eq . , using a gaussian quadrature .
the first four eigenfunctions are depicted in fig.[fig : num ] .
all solutions fall as @xmath109 in the ultraviolet and all possess the sturm - liouville property ; namely , the @xmath110 eigenfunction , @xmath111 , possesses one more zero than @xmath112 .
the number of independent eigenfunctions equals the number of quadrature points and hence , in the continuum limit , there is a countable infinity of solutions , each distinguished by its magnitude and number of zeros .
armed with the knowledge accumulated above , we are ready to find all solutions of the original problem ; i.e. , eq .. in the chiral limit , besides the trivial @xmath3 solution that is admitted in perturbation theory , and the well - known positive - definite dcsb solution , we find the now expected series of solutions , each one of which may be labeled uniquely by the number of zeros it possesses and its sign at @xmath1 .
one must include that sign because the chiral - limit gap equation is even under @xmath113 , so each solution has a mirror image .
the first few solutions are illustrated in fig.[fig : allnoded ] .
all nonzero solutions exhibit at least one inflection point in the infrared and decay as @xmath114 in the ultraviolet .
( recall that the coupling @xmath60 has mass - dimension one . hence , without loss of generality , hereafter we set @xmath115 and measure all dimensioned quantities with respect to this scale . ) . )
[ fig : allnoded ] ] in fig.[fig : multinode ] we display an example of a multi - zero solution of eq .. with the numerical techniques typically employed to solve the gap equation , such solutions are difficult to find and also unstable , in the sense that they are lost when even a very small current - fermion mass ( @xmath116 ) is introduced .
such difficulties are overcome if a more sophisticated numerical algorithm is used ; e.g. , the homotopy continuation method described and employed in ref.@xcite .
indeed , ref.@xcite presents a detailed analysis of the case @xmath117 .
little therefore needs to be explained herein .
it suffices to record that , at fixed interaction strength , the number of distinct solutions to the gap equation diminishes rapidly with increasing @xmath56 .
moreover , there is a critical value , @xmath118 herein , such that eq .
supports only the positive - definite zero - free solution for @xmath119 .
it is natural to ask whether the existence of multiple solutions to the gap equation is a peculiar feature of the rainbow truncation . in partial answer
we note that refs.@xcite found this property persisted when the so - called central ball - chiu _ ansatz _ was employed ; viz . , @xmath120\ , .
\ ] ] we will therefore reconsider this case herein , bringing to bear the knowledge we ve gained from sec.[theory - h ] . .
once the mass - scale is set via @xmath115 , the magnitude of the solution is fixed by the nonlinearity of eq .. [ fig : multinode ] ] using eq . in landau gauge
, the gap equation produces @xmath62 and the following equation for the mass function : @xmath121 this is not different from eq . and hence the pattern of behavior described in secs.[sec : linearized ] , [ sec : nonlinear ] is repeated with eq
.. now consider a simple extension of eq . ;
@xmath122 where @xmath123 is a symmetric function that approaches unity as either or both of @xmath124 , @xmath125 approach infinity . in this case , again , @xmath62 in landau gauge and the mass function satisfies @xmath126 plainly , therefore , with any function @xmath123 that preserves assumption@xmath46 in sec.[theory - h ] , the chiral limit gap equation will possess a countable infinity of distinct solutions , distinguished by their number of zeros and their sign and magnitude at @xmath1 .
as will become plain in sec.[gauge ] , @xmath62 is not required to ensure this outcome .
( a further illustration of these general statements is presented in ref.@xcite . ) the class of vertex _ anstze _ described by eq .
it includes the form in eq .. however , it is known that the true dressed - fermion gauge - boson vertex is more complicated still ( especially in the presence of dcsb ) , involving as many as eleven other dirac matrix structures @xcite . we have therefore solved the qed3 gap equation using this range of more sophisticated _ anstze _ and the numerical algorithms that delivered every solution explained above , irrespective of its complexity ( see app.[sec : appendix ] ) . in each of these cases we find the positive definite dcsb solution of the chiral - limit gap equation and its mirror image but no solutions that possess even a single zero . naturally , despite the effectiveness of our numerical methods when employed with simple vertex _
, the explanation might be that they are inadequate to the task when these more complicated and realistic _
anstze _ are used . on the other hand , such _
produce gap equations that are not of the hammerstein form .
in fact , with extant vertex _ anstze _
, the gap equations are typically elements in the more general class of nonlinear urysonh integral equations @xcite .
little is known about the nature of solutions to such equations .
moreover , the full theory must be even more complicated because , in addition to the explicit dependence on @xmath127 expressed in _ anstze _
, the true dressed vertex will possess an implicit dependence on the dressed - propagator whose nature is impossible to guess .
these observations indicate that the known conditions under which the gap equation can possess multiple solutions are quite strict and , moreover , that they are usually not met when the dressed - fermion gauge - boson vertex possesses what might justifiably be called a _ realistic _
form . as a counterpoint ,
we have considered a combination of the qcd - based interactions in refs.@xcite and the complete ball - chiu _ ansatz_. in these cases the homotopy continuation algorithm reveals that there are no solutions with zeros unless the interaction strength is inflated to an unrealistically large value ; i.e. , more than five - times the strength required to explain contemporary experiment @xcite . owing to the capacities of the homotopy continuation algorithm , which are detailed elsewhere @xcite , one can safely conclude that no solution has been overlooked . applying this experience to qed3
, one can reasonably argue that a qed3 gap equation with the ball - chiu vertex possesses no solutions with a zero .
this is because the mass - scale is set by the coupling strength , which therefore does not provide a tool for inflating the interaction strength relative to other scales ; and , furthermore , fermion loops provide screening , so that unquenching acts to further suppress the interaction strength .
the fermion propagator is gauge covariant but not gauge invariant ; and in the preceding discussion we have focused on landau gauge . as we noted above ,
this is because landau gauge occupies a special place in gauge theories .
in addition to other important properties , such as being a fixed point of the renormalization group and the gauge in which any sensitivity to model - dependent differences between _ anstze _ for the fermion - photon vertex are least noticeable , it is also the sole covariant gauge in which the infrared behavior of the fermion propagator is not modified by a non - dynamical gauge - dependent exponential factor whose presence can obscure truly observable features of the theory @xcite .
moreover , as we now explain , by capitalizing on the landau - khalatnikov - fradkin ( lkf ) transformations @xcite , a mechanical realization of the gauge covariance property , one may focus on landau gauge in abelian theories without loss of generality . in configuration space
, the fermion propagator in any covariant gauge , @xmath128 , is obtained from its landau gauge ( @xmath129 ) form via the following operation : @xmath130 for @xmath131 , one may readily translate this transformation into momentum space @xcite : [ lkfts ] @xmath132 , \label{gistv}\\ \sigma_s(p;\xi ) & = & \frac{\varsigma}{\pi p}\int_0^\infty \!\ !
dk\ , k\ \sigma_s(k,0 ) \left[\frac{1}{\lambda^-}-\frac{1}{\lambda^+}\right]\ ! , \label{gist}\end{aligned}\ ] ] where @xmath133 and we have used eq .. the analogues for @xmath134 are given in ref.@xcite .
plainly , with a solution for the dressed - quark propagator in hand , obtained using any landau - gauge _ ansatz _ for the dressed - fermion gauge - boson vertex , one can straightforwardly find the result in another covariant gauge . ; and _ lower panel _ , @xmath0 .
both panels : @xmath129 , solid curve ; and @xmath135 , dashed curve , obtained via eqs .. [ fig : lkf2 ] ] ; and _ lower panel _ , @xmath0 . both panels : @xmath129 , solid curve ; and @xmath135 , dashed curve , obtained via eqs .. [ fig : lkf2 ] ]
we illustrate the result for a two - zero solution in fig.[fig : lkf2 ] .
it is true in general that the wavefunction renormalization , @xmath136 , remains zero - free in all covariant gauges and the lkf transformation does not shift the location of any zero in @xmath0 .
it is clear , therefore , that whatever is the qualitative character of the gap equation s landau gauge solution , it is the same in all covariant gauges .
namely , if there are multiple solutions for the mass function in landau gauge , each with a different number of zeros , then these solutions exist in all gauges .
notably , however , for @xmath131 their magnitude is rapidly damped ( see eq . and lower panel of fig.[fig : lkf2 ] ) , an effect that will make them difficult to locate numerically if one chooses to solve the gap equation directly in a different gauge .
of course , the gap equation solution obtained directly in a different gauge would only be meaningful if the gap equation were built with the lkf transform of the original landau - gauge vertex _
[ final ] using quenched qed3 as an example , we explored conditions under which a gauge theory s chiral - limit gap equation can exhibit multiple solutions .
we argued that so long as the landau - gauge gap equation can be rewritten as an hammerstein integral equation of the first kind , it will possess a countable infinity of simultaneous solutions .
those solutions will be distinguished , one from another , by the number of zeros they possess and their magnitude and sign at the origin .
whilst a large class of truncations of gauge theory gap equations possess the hammerstein property , a far larger class does not .
membership of the hammerstein class is not usually determined by the form of the dressed - gauge - boson propagator but , instead , it is decided by the structure of the dressed - fermion gauge - boson vertex .
we can not now say anything definite in general about the existence of multiple solutions for members of the non - hammerstein class of gap equations .
however , by analyzing examples from that class , we were led to conjecture that there is at least a large subclass that do not possess solutions with zeros .
indeed , it appears to us that the general conditions under which a gauge theory s chiral - limit gap equation can possess solutions with zeros are quite restrictive ; and it is therefore probable that when a realistic _ ansatz _ is employed for the dressed vertex , the only dynamical chiral symmetry breaking solutions that exist are those with no zeros .
although we have only considered qed3 explicitly , the mathematical framework we have described applies more broadly .
we therefore judge that our results are equally relevant to existing realistic models for the gap equation in qcd .
we thank v.p .
gusynin , v.a .
miransky , i.a .
shovkovy and k .- l .
wang for illuminating discussions .
this work received support from cic ( umsnh ) and conacyt grant nos . 4.10 , 4.22 , 46614-i and 128534 ; and u.s .
department of energy , office of nuclear physics , contract no . de - ac02 - 06ch11357 .
as practitioners will quickly learn , an iterative scheme is unlikely to find a gap equation solution with zeros .
we therefore employed a collocation method .
namely , upon introducing a quadrature rule , the integral equation is recast as a set of nonlinear algebraic equations .
( this generalises the procedure associated with eqs .
, in our manuscript . ) to solve that system , we employed the secant method and the broyden strategy , with a bounded trial solution .
( useful descriptions of solution strategies for nonlinear integral equations may be found in refs.@xcite . )
the two approaches produced solutions with zeros in those cases where such solutions were known to exist ; e.g. , eqs .
, above . as a step toward verifying this outcome , we then expanded the arguments of the integral equation in terms of a complete set of functions and solved for the expansion coefficients .
the solutions without zeros were recovered but no new solutions were found . in this approach we focused on the `` sinc '' and `` little sinc '' collocation methods @xcite . c. d. roberts and a. g. williams , prog .
. phys . * 33 * , 477 ( 1994 ) .
a. bashir _ et al .
_ , commun .
. phys . * 58 * , 79 ( 2012 ) .
r. fukuda and t. kugo , nucl .
b * 117 * , 250 ( 1976 ) . v. a. miransky , v. p. gusynin and y. a. sitenko , phys .
b * 100 * , 157 ( 1981 ) . v. a. miransky and p. i. fomin , phys .
b * 105 * , 387 ( 1981 ) .
i. fomin , v. p. gusynin , v. a. miransky and y. a. sitenko , riv .
nuovo cim .
* 6n5 * , 1 ( 1983 ) . v. a. miransky , i. a. shovkovy and l. c. r. wijewardhana , phys .
b * 468 * , 270 ( 1999 ) .
e. v. gorbar , v. p. gusynin , v. a. miransky and i. a. shovkovy , phys .
a * 313 * , 472 ( 2003 ) .
g. cheng and t .- k .
kuo , j. math .
phys . * 35 * , 6270 ( 1994 ) .
l. chang , y .- x .
liu , m. s. bhagwat , c. d. roberts and s. v. wright , phys .
c * 75 * , 015201 ( 2007 ) .
f. j. llanes - estrada , t. van cauteren and a. p. martin , eur .
j. c * 51 * , 945 ( 2007 ) .
r. williams , c. s. fischer and m. r. pennington , phys .
b * 645 * , 167 ( 2007 ) . c. s. fischer , d. nickel and r. williams , eur .
j. c * 60 * , 47 ( 2009 ) .
et al . _ ,
d * 86 * , 114001 ( 2012 ) .
m. gopfert and g. mack , commun .
82 * , 545 ( 1981 ) . c. j. burden , j. praschifka and c. d. roberts , phys .
d * 46 * , 2695 ( 1992 ) .
g. s. bali , h. neff , t. duessel , t. lippert and k. schilling , phys .
d * 71 * , 114513 ( 2005 ) .
m. bhagwat , m. pichowsky , c. roberts and p. tandy , phys .
c * 68 * , 015203 ( 2003 ) .
m. s. bhagwat and p. c. tandy , aip conf .
* 842 * , 225 ( 2006 ) . c. d. roberts , prog
. phys . * 61 * , 50 ( 2008 ) . , ( national academies press , 2012 ) . c. d. roberts , few body syst . * 52 * , 345 ( 2012 ) .
m. franz , z. tesanovic and o. vafek , phys .
b * 66 * , 054535 ( 2002 ) .
i. f. herbut , phys .
b * 66 * , 094504 ( 2002 ) .
i. o. thomas and s. hands , phys .
b * 75 * , 134516 ( 2007 ) .
k. novoselov _ et al .
_ , nature * 438 * , 197 ( 2005 ) .
v. gusynin , s. sharapov and j. carbotte , int .
j. mod .
b * 21 * , 4611 ( 2007 ) .
o. gamayun , e. gorbar and v. gusynin , phys . rev .
b * 81 * , 075429 ( 2010 ) .
h. pagels , phys .
d * 21 * , 2336 ( 1980 ) .
a. d. polyanin and a. v. manzhirov , ( campman & hall / crc , boca raton , florida , 2008 ) .
a. bashir , a. raya , s. snchez - madrigal and c. d. roberts , few body syst . *
46 * , 229 ( 2009 ) .
e. s. fradkin , zh .
. fiz . * 29 * , 258 ( 1955 ) .
l. d. landau and i. m. khalatnikov , zh .
* 29 * , 89 ( 1956 ) , [ sov .
jetp * 2 * , 69 ( 1956 ) ] .
k. johnson and b. zumino , phys .
lett . * 3 * , 351 ( 1959 ) .
b. zumino , j. math . phys . * 1 * , 1 ( 1960 ) . c. d. roberts and b. h. mckellar , phys .
d * 41 * , 672 ( 1990 ) .
feng , w .- m .
sun and h .- s .
zong , journal of modern physics * 4 * , 151 ( 2013 ) . j. s. ball and t .- w .
chiu , phys .
d * 22 * , 2542 ( 1980 ) .
d. c. curtis and m. r. pennington , phys .
d * 42 * , 4165 ( 1990 ) .
a. kizilers and m. r. pennington , phys . rev .
d * 79 * , 125020 ( 2009 ) .
l. chang , y .- x . liu and c. d. roberts , phys .
lett . * 106 * , 072001 ( 2011 ) .
a. bashir , a. raya and s. snchez - madrigal , phys .
d * 84 * , 036013 ( 2011 ) .
a. bashir , r. bermdez , l. chang and c. d. roberts , phys .
c * 85 * , 045205 ( 2012 ) .
qin , l. chang , y .- x .
liu , c. d. roberts and s. m. schmidt , phys .
b * 722 * , 384 ( 2013 ) .
m. a. krasnoselskii , ( pergamon , oxford , 1964 ) , translated from the russian edition ( moscow , 1956 ) by a. h. armstrong .
a. bashir , phys .
b * 491 * , 280 ( 2000 ) .
a. bashir and a. raya , phys .
d * 66 * , 105005 ( 2002 ) .
a. bashir and a. raya , nucl .
b * 709 * , 307 ( 2005 ) .
a. bashir and a. raya , few body syst . * 41 * , 185 ( 2007 ) .
j. c. r. bloch , hep - ph/0208074 , .
a. k. kaw and e. e. kalu , ( http://www.autarkaw.com/ ) . | theories that support dynamical generation of a fermion mass gap are of widespread interest .
the phenomenon is often studied via the dyson - schwinger equation ( dse ) for the fermion self energy ; i.e. , the gap equation .
when the rainbow truncation of that equation supports dynamical mass generation , it typically also possesses a countable infinity of simultaneous solutions for the dressed - fermion mass function , solutions which may be ordered by the number of zeros they exhibit .
these features can be understood via the theory of nonlinear hammerstein integral equations . using qed3 as an example
, we demonstrate the existence of a large class of gap equation truncations that possess solutions with damped oscillations .
we suggest that there is a larger class , quite probably including the exact theory , which does not .
the structure of the dressed - fermion gauge - boson vertex is an important factor in deciding the issue . |
the running of the strong coupling constant @xmath3 is parameterized by the @xmath4 parameter . in a recent letter
@xcite we have computed @xmath0 from the average plaquette and the scale parameter @xmath5 . an essential step in our calculation was the conversion of the boosted coupling @xmath6 to the renormalized coupling @xmath7 .
this was done in two - loop tadpole improved perturbation theory , which resulted in an ( estimated ) systematic uncertainty of the order of the statistical error .
we can do better now . in this talk
we shall apply pad techniques to approximate an unknown three - loop coefficient in the conversion formula .
this reduces the systematic uncertainty to an insignificant amount , as we shall see .
the calculation of @xmath0 proceeds in three steps : first we compute the average plaquette @xmath8 from which we obtain @xmath6 , @xmath9 where @xmath10 is the bare lattice coupling .
note that the non - perturbative contribution to the plaquette , i.e. the difference between the lattice number and the perturbative series , is less than @xmath11 @xcite . in the second step
we compute @xmath7 from @xmath6 .
this is the main task . to three loops we obtain , following the notation of @xcite , @xmath12 where @xmath13 .
we know all coefficients except for @xmath14 .
( see @xcite for references . )
the latter can be computed from the difference of the four - loop coefficients , @xmath15 , of the @xmath16 function @xmath17 where @xmath18 denotes the scheme .
we find @xmath19 we know @xmath20 , but we do not know @xmath21 . in the @xmath2 scheme
the four - loop @xmath16 function turns out to be very well reproduced by the pad - improved three - loop @xmath16 function @xmath22}^{\mathcal s}(g_{\mathcal s } ) = -g_{\mathcal s}^3 \frac{b_0b_1+(b_1 ^ 2-b_0b_2^{\mathcal s})g_{\mathcal s}^2}{b_1-b_2^{\mathcal s } g_{\mathcal s}^2}. \label{pade}\ ] ] in fig .
1 we compare ( [ pade ] ) with the four - loop expression .
we see that in the region @xmath23 , which is the region relevant to our calculation , the pad - improved three - loop result differs by less than @xmath24 from the four - loop result .
we expect that the same is true for the boosted @xmath16 function @xmath25 , which converges at a similar rate as @xmath26 .
thus we may approximate @xmath25 by @xmath27}^\square$ ] .
expanding ( [ pade ] ) gives us then @xmath28 we can improve the convergence of the series ( [ boxcon ] ) further by re - expressing it in terms of the tadpole improved coefficients @xmath29 this amounts , for example , to replacing every @xmath30 by @xmath31 we have already seen in @xcite that tadpole improvement reduces the coefficients of the two - loop fermionic contribution .
we find that tadpole improvement is equally successful in reducing the coefficients of the three - loop contributions .
in the final step we compute @xmath0 from @xmath32 using the renormalization group equation @xmath33 for @xmath26 we use the pad - improved four - loop @xmath16 function @xmath34}^{\overline{ms}}\,(g_{\overline{ms}})&= & - g^3_{\overline{ms } } \big(b_0\,(b_1 ^ 2-b_0b_2^{\overline{ms}})+(b_1 ^ 3 \nonumber\\ & - & 2b_0b_1b_2^{\overline{ms } } + b_0 ^ 2\,b_3^{\overline{ms}})\,g^2_{\overline{ms}}\big)\big/ \big(b_1 ^ 2 \nonumber \\ & - & b_0b_2^{\overline{ms } } + ( b_0b_3^{\overline{ms } } - b_1b_2)\,g^2_{\overline{ms } } \nonumber \\ & + & ( b_2^{\overline{ms}\ , 2 } -b_1b_3^{\overline{ms}})\ , g^4_{\overline{ms}}\big ) , \end{aligned}\ ] ] which is a better approximation than the four - loop @xmath16 function .
the result should be independent of @xmath35 if our approximations of @xmath26 in ( [ boxcon ] ) and ( [ conv ] ) are consistent .
we have varied @xmath35 between 4 and 10 gev and found indeed that @xmath0 changes by a fraction of a percent only .
finally we have chosen @xmath36 , which corresponds to @xmath37 gev .
we may estimate the error arising from our approximation ( [ b3 ] ) by replacing @xmath20 by @xmath38 as well .
we found that this changes our results by at most @xmath39 . [ cols="^,^,^,^,<,<,<,<",options="header " , ] to convert our numbers to a physical scale we have used the force parameter @xmath5 .
the results are presented in table 1 .
the numbers have increased by @xmath40 relative to our previous values @xcite , which is well inside the systematic error that was estimated in that paper .
we fit our data by @xmath41 ( for details of the fit see @xcite . ) in fig .
2 we show @xmath42 and @xmath43 which , if the fit is successful , should collapse the data points onto a single line each .
we see that this is indeed the case .
we also see that the largest uncertainty arises from the continuum extrapolation , while the data depend only weakly on the quark mass . in the chiral and continuum
limit our fit gives @xmath44 ( @xmath45 ) .
using @xmath46 fm , we finally obtain @xmath47 this is to be compared with our previous result @xmath48 .
our calculation faces two problems .
one is the conversion from the boosted to the @xmath2 scheme . the other is the extrapolation to the chiral and continuum limit .
we have solved the first problem , and with time and more data we hope to improve on the extrapolation . | we compute @xmath0 for two flavors of light dynamical quarks using non - perturbatively @xmath1 improved wilson fermions .
we improve on a recent calculation by employing pad - improved two - loop and three - loop perturbation theory to convert the lattice numbers to the @xmath2 scheme . |
in the minimal supersymmetric standard model ( mssm ) @xcite there are two scalar superpartners @xmath16 and @xmath17 , of the top quark which are the weak eigenstates .
the mass eigenstates the lighter top squark ( @xmath0 ) and the heavier top squark ( @xmath18 ) are linear combinations of the weak eigenstates . due to
mixing effects in the top squark mass matrix in the weak basis driven by the top quark mass ( @xmath19 ) there may be a significant mass difference between @xmath0 and @xmath18 .
in fact the former could very well be the next - to - lightest supersymmetric particle ( nlsp ) , the lightest neutralino ( @xmath4 ) being the lightest supersymmetric particle ( lsp ) by the standard assumption in r - parity conserving mssm .
this happens in a wide region of the mssm parameter space . in this scenario , henceforth referred to as the @xmath0 - nlsp scenario , the @xmath0 may be the only strongly interacting superpartner within the kinematic reach of tevatron run ii experiments with a relatively large production cross - section .
additional interest in the light top - squark scenario stems from the observation that the mssm can explain the baryon asymmetry of the universe via electroweak baryogenesis ( ewbg ) provided @xmath1 @xcite .
the search for @xmath0 is , therefore , a high priority program for the on going experiments at the tevatron .
the search for @xmath0-nlsp at tevatron run i and lep and , more recently , at tevatron run ii produced negative results and lower bounds on @xmath11 .
most of the analyses @xcite are based on the assumption that @xmath0 decays via the flavor changing neutral current ( fcnc ) induced loop decay , @xmath5 @xcite with 100 @xmath6 branching ratio ( br )
. we also employ this assumption which is by and large valid if tan@xmath20 where tan @xmath21 is the ratio of the vacuum expectation values for the two neutral higgs bosons present in the mssm@xcite . for lower values of this parameter
the four body decay of the @xmath0 may be a competing channel @xcite there are decay modes of the @xmath0 - nlsp other than the above two channels .
they are the tree - level two body decay , @xmath22 and the three body decay , @xmath23 .
the last modes are kinematically forbidden for small values of the mass difference @xmath24 which is the main concern of this paper . the search for @xmath0-nlsp are based on the jets plus missing @xmath25 channel@xcite .
some of the more recent works employed c - jet tagging by a lifetime based heavy flavour algorithm .
these jets become softer if @xmath7 is small . as a result the efficiency of the kinematical cuts for suppressing the background as well as that of c - jet tagging decreases .
this weakens the limit on @xmath11 from tevatron . at tevatron
run i the largest @xmath11 excluded was @xmath26 for @xmath27 .
the most recent analysis by the d0 collaboration at run ii @xcite with c - jet tagging obtained the limit @xmath28 @xmath29 for @xmath30for the most conservative cross - section after including the next to leading order ( nlo ) corrections @xcite . on the other hand
the lep lower - bounds on @xmath11 are restricted mainly due to kinematics and are around @xmath31 @xcite .
however , much smaller values of @xmath7 can be probed in the cleaner environment of an @xmath32 collider .
the prospect of @xmath0 - nlsp search via this decay channel at run ii was investigated in @xcite .
it was observed that a large region of the @xmath11 - @xmath3 parameter space corresponding to small @xmath7 is beyond the reach of run ii . for a given @xmath11
there is a minimum value of @xmath7 that can yield an observable signal . a modified strategy for @xmath0-nlsp searches in the limit of small @xmath7 is important in its own right .
the current interest in this search , however , is further strengthened by one of the cornerstones of the interface between particle physics and cosmology . a very attractive feature of the r - parity conserving mssm is that the lsp ( @xmath4 ) , is a very good candidate for the dark matter ( dm ) in the universe required , e.g , by the wilkinson microwave anisotropy probe ( wmap ) data @xcite .
the dm relic density depends on the annihilation cross - section ( thermally averaged ) of a lsp pair .
the coannihilation of the lsp with any other supersymmetric particle(sparticle)is another important mechanism for relic density production .
this mechanism is , however , efficient only when the two coannihilating particles have approximately the same mass .
thus in the small @xmath7 scenario @xmath0 - lsp coannihilation may indeed be an important mechanism for producing appropriate relic density @xcite .
the region of the parameter space of mssm consistent with the dm relic density is severely constrained by the wmap data .
nevertheless even in more restricted versions of the mssm like the minimal supergravity model(msugra ) @xcite one finds a narrow region of the parameter space where @xmath0 - lsp coannihilation is an important relic density producing mechanism@xcite .
the search for @xmath0-nlsp with a small @xmath7 is , therefore , important irrespective of the question of ewbg .
however , it is certainly worthwhile to check whether @xmath0 with mass in the quoted range preferred by ewbg can also produce an acceptable dm relic density .
this was investigated in @xcite .
it was found that in a significant region of the allowed parameter space @xmath7 is indeed small ( see figure 7 of @xcite ) .
the results of @xcite were illustrated by specific choices of other mssm parameters .
in particular ewbg in the mssm requires certain cp violating ( cpv ) phases . in a certain phase convention the relative phase(@xmath33 ) between the higgsino mass parameter @xmath34 and the su(2 ) gaugino mass @xmath35 is the most important one .
ewbg usually requires 0.05 @xmath36 1 . however , various uncertainties in the calculation does not rule out a much smaller magnitude of this phase .
thus calculations by neglecting this phase seems to be a reasonable approximation@xcite .
it should , however , be emphasized that the signal proposed by us is fairly model independent and does not depend on the cpv phase or many of the other mssm parameters at all as long as the br(@xmath5 ) is close to 100% . under this assumption
the size of the signal depends on @xmath11 through the production cross - section of the @xmath0 pair via the standard qcd processes and on @xmath3 through the efficiency of the kinematical cuts .
the starting point of our work is the observation that when mass difference @xmath7 is small , in most of the signal events , one of the @xmath37-quarks from @xmath0 pair decay is not energetic enough to produce a jet which may pass jet selection criteria of the experiments at tevatron .
it may be seen as an isolated energy deposit in the calorimeter coming from the decay of a heavy particle .
we call it _ isolated cluster _ ( _ ic _ ) .
thus the proposed signal consists of a c - jet of modest @xmath38 accompanied by missing energy and an _
isolated cluster_. in order to reduce the background we require another hard jet in the signal which in most cases comes from qcd radiation .
our simulations show that a set of selection criteria based on the above features of the signal can isolate it from the sm background . in this work
we do not consider the prospect of fully identifying the flavour of the heavy , isolated , decaying object because of the rather small statistics .
this leads to inevitable backgrounds from , e.g. , @xmath39 events and w / z + jets events .
however , we shall analyze at the generator level some important characteristics of this object which has the potential of reducing the sm backgrounds to a manageable level . at the same time we emphasize that this work is only suggestive of a new approach to @xmath0 search at the tevatron and needs detailed detector simulation for a more definitive statement and that is beyond the scope of this work .
we have used pythia ( v 6.206 ) @xcite for generation of both signal and background events which includes generation of the parton level events followed by the decay of the partons hadronization and decay of their daughters .
generation of both signal and background events take into account initial state radiation ( isr ) and final state radiation ( fsr ) .
the cross - sections @xmath43 and @xmath44 are very large and most of these events generated with low @xmath45 are not relevant for our analysis . to sample the @xmath39 and @xmath46 events better for our purpose and save computer time we have used a cut @xmath47 for generation of @xmath39 and @xmath46 events , where @xmath48 is defined in the cm frame of the colliding partons .
we have used the toy calorimeter simulation followed by jet formation in pythia ( pycell ) . 1 .
the calorimeter coverage is @xmath49 .
a cone algorithm with @xmath50 has been used for jet finding with @xmath51 gev and @xmath52 and jets have been ordered in @xmath38 .
we consider leptons @xmath53 with @xmath54 , @xmath55 .
the lepton should be isolated from jets ( @xmath56 ) .
4 . for charged particles
( @xmath57 , @xmath34 and charged hadrons ) , we have used their generator level momentum as _ track momentum _ when required .
5 . for jets containing a @xmath58 or a @xmath59 hadron
we have used their decay length information for determining the presence of a long lived particle .
a quark or a gluon from ( mainly ) fsr is seen as a jet and in most cases this jet appears to have the highest @xmath38 .
this prompts us to consider a rather unusual signature for the signal events as mentioned below .
the backgrounds , particularly @xmath39 and @xmath46 events have very large cross - sections and hence we need to generate a large number of events and retain only a small fraction of them which pass pre - selection for detailed analysis .
we have used the following pre - selection criteria : 1 .
event should have only two jets : @xmath60 ( see figure [ fig : njet ] ) 2 .
events with isolated leptons are rejected .
one of the jets should contain a long lived particle ( @xmath58 or , @xmath59 hadron ) and henceforth called _ matched - jet _ ( mj ) .
4 . event should have an _ isolated cluster _ resulting from the decay of a @xmath58 or @xmath59 hadron such that @xmath61 .
the direction of the _ isolated cluster _ is defined to be the direction of the decaying _ b _ or _ d _ hadron . in the final selection
this cluster has to be identified as the signature of a long lived particle , the criteria for which are discussed in detail later .
we assume that a _ b - jet _ with @xmath62 is tagged with a probability @xmath63 and for @xmath64 , @xmath65 where @xmath66 is the single @xmath67-jet tagging efficiency ( i.e. , the ratio of the number of tagged @xmath67-jets and the number of taggable @xmath67-jets ) .
the pre - selection efficiency for the signal events is rather small : for @xmath68 ( a ) only @xmath69 events survives pre - selection ; for @xmath70 ( b ) and @xmath71 ( c ) the rates are @xmath72 and @xmath73 respectively . for @xmath74 , @xmath39 , @xmath46 and _ v + jet _ events ( _ v = w , z _ ) rates are @xmath75 , @xmath76 , @xmath77 , @xmath78 and @xmath79 respectively ( see table [ tab : single ] for details )
. the leading order cross - section at @xmath80 , where @xmath81 is the qcd scale for @xmath82 production is from @xcite .
all background cross - sections have been computed by calchep ( version 2.3.7 ) @xcite at @xmath83 .
the largest background comes from _ z+jets _ production .
the variation of the cross - section of this process with the qcd scale is not very severe .
.efficiency table for the signal and sm backgrounds to pass various selection criteria .
column 3 shows the efficiency for pre - selection ; columns 4 - 8 show the efficiencies for each cut combined with the effect of pre - selection ; column 9 shows the efficiency for final selection .
( see text for details ) the last column shows the expected number of events to pass all selection(rejection ) criteria for @xmath84 .
[ cols="^,^,^,^,^,^,^,^,^,^ " , ] 1 . for the first jet we require @xmath85 and @xmath86 and for the second jet @xmath87 ( cut 1 ) .
2 . events should have @xmath88 ( cut 2 ) ( see figure [ fig : etmiss ] ) 3 .
the long lived particle in the _ matched - jet _ should have decay length @xmath89 mm .
( cut 3 ) 4
. the _ isolated cluster _ should be central and have a minimum @xmath38 : @xmath90 and @xmath91 .
it should have a decay length @xmath92 .
( cut 4 ) ( see figure [ fig : iso_pt ] ) 5 . in the signal
we expect the _ isolated cluster _ and the _ matched - jet _ to be approximately back - to - back in the transverse plane .
so , the cut @xmath93 ( see figure [ fig : phi_diff ] ) rejects background , particularly _ w+jets _ and _ z+jets _ events .
the _ matched - jet _ is most likely to be the leading jet in _ w+jets _ and _ z+jets _ events whereas it is the 2nd jet in the signal events .
we therefore select events whose @xmath94 ( see figure [ fig : pt_matched ] ) .
+ we also partially reconstruct invariant mass of the _ matched - jet _ ( @xmath95 ) using the charged tracks associated with it .
similarly @xmath96 is reconstructed for the _ isolated cluster_. the cuts @xmath97 and @xmath98 , reject the @xmath39 events and also reduce _
w+jets _ and _ z+jets _ backgrounds ( see figure [ fig : mass ] )
. the combined effect of these cuts ( cut 5 ) for the signal and the dominant backgrounds is shown in column 8 of table [ tab : single ] .
although a very high rejection factor of ( @xmath99 ) is achieved for the @xmath39 events and no event in the simulated sample survives , this may still be dangerous as @xmath43 is very large . since the signal events do not have spectacular signatures ,
it is not possible to apply more stringent criterion on any of the features and still retain a good signal .
it may be required to exploit the subtle features of the _ matched - jet _ and the _ isolated cluster _ to get rid of the @xmath39 events .
there are a few features which may be exploited to this end although it may be experimentally challenging : * number of charged tracks associated with the _
isolated cluster_. * upper cut on the lifetime related observables for the _ matched - jet _ , expected to be a _ c - jet _ and the _ isolated cluster _ , expected to come from the decay of a @xmath59 hadron in the signal . *
presence of a @xmath100 in the _ isolated cluster _ which carries a significant fraction of its @xmath101 .
* more reliable reconstruction of the invariant masses using full detector information .
since these observables are rather _ delicate _ , they may be simulated only using detailed detector modelling and estimation of the effects of criteria based on these observables is beyond the scope of this analysis .
however it may be said with some degree of confidence that @xmath39 and other background events may be further suppressed by using judicious choice of such criteria .
( left ) and @xmath96 ( right ) reconstructed using charged tracks associated with the _ matched - jet _ and the _ isolated cluster _ in signal ( @xmath41 , @xmath40 ) and @xmath39 events after pre - selection ( see text ) . ]
our final results are presented in table [ tab : signi ] . in view of
the lep limits @xcite we have looked into signals with @xmath102 and we have chosen @xmath103 ( in @xmath104 ) = 10 , 15 and 20 . the significance is defined as @xmath105 where @xmath106 ( @xmath107 ) is the number of signal ( background ) events passing the selection criteria for an integrated luminosity of @xmath84 . for @xmath108
we may hope for a _ discovery _ while beyond that it may be restricted to the level of a _
the parameter space probed by us contains a part of the region interesting in the context of ewbg .
again , it should be emphasized that this very preliminary analysis is designed to provoke the experimentalists to scan the region interesting from the point of view of dark matter relic density and ewbg . * acknowledgment * : ad , mm and nb acknowledge financial support from department of science and technology , government of india under the project no ( sr / s2/hep-18/2003 ) .
mm also acknowledges support from department of science and technology , government of india under the project no ( sp / s2/k-25/96-vi ) .
a part of this work was done when ad and nb were in the department of physics , jadavpur universiry , kolkata 700 032 , india . for reviews on supersymmetry , see , _
e.g. _ , h. p. nilles , ( 110,1984,1 ) ; h. e. haber and g. kane , ( 117,1985,75 ) ; j. wess and j. bagger , _ supersymmetry and supergravity _ , 2nd ed . , ( princeton , 1991 ) ; m. drees , p. roy and r. m. godbole , _ theory and phenomenology of sparticles _ , ( world scientific , singapore , 2005 ) .
m. carena , m. quiros and c.e .
wagner , ( 380,1996,81 ) , hep - ph/9603420 ; ( 503,1997,387 ) , hep - ph/9702409 ; ( 524,1998,3),hep - ph/9710401 ; d.delepine , j.m .
gerard , r.gonzalez felipe and j.weyers , ( 386,1996,183),hep - ph/960440 ; j. mcdonald , ( 413,1997,30),hep - ph/9707290 ; j.m .
cline and g.d .
moore , ( 81,1998,3315),hep - ph/9806354 .
lepsusywg , aleph , delphi , l3 and opal collaborations , note lepsusywg/04 - 02.1 ( http:/ lepsusy.web.cern.ch/lepsusy/welcome.html ) a. h. chamseddine , r. arnowitt and p. nath , ( 49,1982,970 ) ; r. barbieri , s. ferrara and c. a. savoy , ( 119,1982,343 ) ; l. j. hall , j. lykken and s. weinberg , ( 27,1983,2359 ) ; p. nath , r. arnowitt and a. h. chamseddine , ( 227,1983,121 ) ; n. ohta , ( 70,1983,542 ) .
see , e.g .
, a.pukhov , calchep@xmath109a package for evaluation of feynman diagrams and integration over multi - particle phase space ( hep - ph/9908288 ) .
for the more recent versions see : http://www.ifh.de/pukhov/calchep.html . | the search for the top squark ( @xmath0 ) within the kinematic reach of tevatron run ii is of great contemporary interest .
such a @xmath0 can explain the baryon asymmetry of the universe provided @xmath1 .
moreover if @xmath2 - @xmath3 is small , where @xmath4 is the lsp , the dark matter relic density as obtained from the wmap data may be explained via @xmath0- lsp coannihilation . in this scenario
the decay @xmath5 is likely to occur with 100 @xmath6 branching ratio but for small @xmath7 the conventional di - jet + @xmath8 signal becomes unobservable .
we propose a new search strategy based on the @xmath9 signature accompanied by an isolated cluster of energy which arises from a decaying heavy particle with characteristic decay length .
our preliminary simulations with pythia indicate that for @xmath10 this signal may be observable while somewhat lager @xmath11 may still provide hints of new physics .
-10 pt 42by .00025 truein .00025 truein 0.75 in 0.75 in 6.5 truein # 1 # 1 # 1 ( # 1,#2,#3)*#1 * ( # 2 ) # 3(#1,#2,#3)phys .
rev .
d ( # 1,#2,#3 ) ( # 1,#2,#3)phys .
lett .
b ( # 1,#2,#3 ) ( # 1,#2,#3)phys . rev .
lett .
( # 1,#2,#3 )
( # 1,#2,#3)phys . rep .
( # 1,#2,#3 ) ( # 1,#2,#3)nucl .
phys .
b ( # 1,#2,#3 ) ( # 1,#2,#3)j .
high energy phys .
( # 1,#2,#3 ) ( # 1,#2,#3)prog . theor .
phys .
( # 1,#2,#3 ) ( # 1,#2,#3)comp .
phys .
comm .
( # 1,#2,#3 ) ( # 1,#2,#3)astropart . phys .
( # 1,#2,#3 ) = 100000 iiser / hep/08/07 * search for top squarks at tevatron inspired by dark matter and electroweak baryogenesis * + nabanita bhattacharyya@xmath12 , amitava datta@xmath12 , and manas maity@xmath13 + @xmath14 _ indian institute of science education and research , kolkata , salt lake city , kolkata 700 106 , india .
_ + @xmath15 _ department of physics , visva - bharati , santiniketan 731235 , india _
+ pacs no : 11.30.pb , 13.85.-t , 12.60.jv , 14.80.ly |
quasar absorption line systems with strong lyman-@xmath0 lines are often divided into two classes : damped lyman-@xmath0 ( dlas , log n@xmath3@xmath8 20.3 ) and sub - damped lyman-@xmath0 ( sub - dla 19 @xmath9 log n@xmath3@xmath10 20.3 , @xcite ) which contain a major fraction of the neutral gas in the universe , while the majority of the baryons are thought to lie in the highly ionized and diffuse lyman-@xmath0 forest clouds with log n@xmath3@xmath9 15 in intergalactic space @xcite .
the lower threshold of log n@xmath3=20.3 for classification of dlas stems from previous 21 cm emission studies of nearby spirals , where the sensitivity limited column density of log n@[email protected] was seen to lie near the holmberg radius ( r@xmath11 ) of the galaxy @xcite .
nonetheless , the damping wings which can be used to accurately measure n@xmath3 in these systems do begin the sub - dla regime of log n@[email protected] . with their high gas content , the dla and sub - dla systems
are believed to be associated directly with galaxies at all redshifts in which they are seen . among the many elements often detected in qso absorber systems including c , n , o , mg , si , s , ca , ti , cr , mn , fe , ni , and zn , zn is the preferred tracer of the gas - phase metallicity as it is relatively undepleted in the galactic ism , especially when the fraction of h in molecular form is low , as is the case in most dlas .
zn also tracks the fe abundance in galactic stars ( e.g. , @xcite ) , and the lines of zn ii @xmath132026,2062 are relatively weak and typically unsaturated . these lines can be covered with ground based spectroscopy over a wide range of redshifts , 0.65 @xmath9 z @xmath9 3.5 , which covers a large portion ( @xmath14 ) of the history of the universe .
abundances of refractory elements such as cr and fe relative to zn also give a measure of the amount of dust depletion @xcite .
abundance ratios such as [ si / fe ] , [ o / fe ] and [ mn / fe ] shed light on the enrichment from the different types of supernovae , as the @xmath0-capture elements si and o for example are produces mainly in type ii explosions while the iron peak elements are produced mainly by type ia supernovae . in previous studies ,
dla systems have been the preferred systems for chemical abundance investigation owing to their high gas content @xcite .
most dlas however have been found to be metal poor , typically far below the solar level and below where models predict the mean metallicity should be at the corresponding redshifts at which they are seen ( e.g. , @xcite and references therein ) .
the sub - dla systems have until recently been largely ignored , with their contribution to the overall metal budget unknown .
evidence for the possibility of a non - negligible contribution from sub - dlas to the metal budget came from @xcite , who noted that based on fe ii lines the sub - dla systems have faster evolution of the fe abundance and higher abundances on average than dla systems .
this has also been validated by @xcite .
galactic chemical evolution is a slow process , and long timescales must be examined to search for the signs of the gradual chemical enrichment that models predict @xcite . although redshifts @xmath15 span 70@xmath16 of the age of the universe ( using a concordance cosmology of @xmath17 , h@xmath18=70 km s@xmath19mpc@xmath19 ) , few observations have been made of @xmath15 sub - dlas due to the lack of spectrographs with enough sensitivity in short wavelengths and the paucity of known sub - dlas in this redshift range .
with such a large fraction of the age of the universe covered in the redshift regime , it is clearly important for understanding the nature of sub - dla systems and galactic chemical evolution as well .
we have greatly increased the sample of sub - dlas in past several years with our vlt uves and magellan - ii mike spectra . in this paper
, we report on ten new systems observed with the mike spectrograph on the magellan - ii clay telescope .
the structure of this paper is as follows : in @xmath20 [ sec : obs ] , we discuss details of our observations and data reduction techniques . @xmath20
[ sec : objects ] gives details on the individual objects in these new observations .
@xmath20 [ sec : mnfe ] investigates the ratio of [ mn / fe ] in qso absorbers . in @xmath20
[ sec : abund ] we discuss the abundances of these absorbers and we give a brief discussion .
we also provide an appendix at the end of this paper , showing plots of the uv spectra with the fits to the lyman-@xmath0 lines , and tables containing the fit parameters for the individual systems .
the observations presented here were made with the 6.5 m magellan - ii clay telescope and the magellan inamori kyocera echelle ( mike ) spectrograph @xcite in 2007 sep .
this is a double sided spectrograph with both a blue and a red camera , providing for simultaneous wavelength coverage from @xmath73340 to @xmath79400 .
targets were observed in multiple exposures of 1800 to 2700 sec each to minimize cosmic ray defects .
the seeing was typically @xmath10 1@xmath21 , averaging @xmath7 0.7@xmath21 .
all of the target qsos were observed with the 1@xmath21x5@xmath21 slit and the spectra were binned 2x3 ( spatial by spectral ) during readout .
the resolving power of the mike spectrograph is @xmath719,000 and @xmath725,000 on the red and blue sides respectively with a 1@xmath21 slit .
table 1 gives a summary of the observations .
these spectra were reduced using the mike pipeline reduction code in idl developed by s. burles , j. x. prochaska , and r. bernstein .
wavelengths were calibrated using a th - ar comparison lamp taken after each exposure .
the data were first bias subtracted from the overscan region and flat - fielded .
the data were then sky - subtracted and the spectral orders were extracted using the traces from flat field images .
these extracted spectra were then corrected for heliocentric velocities and converted to vacuum wavelengths .
each individual order was then combined in iraf using rejection parameters to reduce the effects of cosmic rays .
these combined spectra were then normalized using a polynomial , typically of order five or less , or spline function to fit the continuum .
our new observations consists of 10 absorbers , nine sub - dlas and one dla z@xmath22@xmath9 1.5 .
see @xcite for a discussion of our selection criteria . throughout this paper
the qso names are given in j2000 coordinates , except in table 1 where the original name , based on j1950 coordinates is also given if applicable @xcite . | the damped and sub - damped lyman-@xmath0(dla and sub - dla ) systems seen in the spectra of qsos offer a unique way to study the interstellar medium of high redshift galaxies . in this paper
we report on new abundance determinations in a sample of 10 new systems , nine of the lesser studied sub - dlas and one dla , along the line of sight to seven qsos from spectra taken with the mike spectrograph .
lines of mg i , mg ii , al ii , al iii , ca ii , mn ii , fe ii , and zn ii were detected . here , we give the column densities and equivalent widths of the observed absorption lines , as well as the abundances determined for these systems .
zn , a relatively undepleted element in the local interstellar medium ( ism ) is detected in one system with a high metallicity of [ zn / h][email protected] .
in one other system , a high abundance based on the more depleted element fe is seen with [ fe / h]=@xmath2 , although zn is not detected .
the n@xmath3-weighted mean metallicity of these sub - dla systems based on fe is @xmath4[fe / h]@xmath5=@xmath6 , nearly @xmath70.7 dex higher ( a factor of 5 ) than what is seen in dlas in this redshift range .
the relative abundance of [ mn / fe ] is also investigated .
a clear trend is visible for these systems as well as systems from the literature , with [ mn / fe ] increasing with increasing metallicity in good agreement with with milky way stellar abundances .
[ firstpage ] quasars : absorption lines - ism : abundances |
consider the problem of motion of two point particles with masses @xmath0 , @xmath1 in the absence of external forces ( the two - body problem ) . here and in
what follows , all parameters belonging to the first particle will be labeled by index `` 1 '' and those belonging to the second particle , by `` 2 '' . in order to avoid unnecessary complication of our consideration , we assume the center of mass of the system to be at rest . if the motion of the system is uniform , one can readily come to the corresponding conclusions by introducing a moving coordinate frame .
we shall look for the solution to the problem in the approximation used , for instance , in monograph @xcite , i.e. , in the nonrelativistic approximation . there is a well known solution to this problem involving the reduced mass ( cf .
, e.g. , @xcite , 13 or @xcite , 12 ) . in this case ,
analysis of the motion of two particles is replaced by consideration of the motion of one fictitious @xmath2-particle with a mass @xmath3 ( reduced mass ) , with this @xmath2-particle assumed to move in a centrally symmetric field with a center at rest placed at the center of mass of the two particles .
the magnitude of this field is governed by the force of interaction of the two particles with one another .
on finding the law by which the @xmath2-particle moves , one will be able to readily reconstruct the motion of the real particles @xmath0 and @xmath1 .
one can offer the other kind of solution .
this solution assumes each particle moves in its own stationary potential well and for each body separately the energy and the angular momentum laws are valid . said otherwise , in place of an analysis of the motion of one @xmath2-particle in a gravitational potential field associated with particle interaction one may consider the motion of each real particle in the stationary gravitational potential field created by the other particle .
this stationary field differs , however from the gravitational field produced by the @xmath0 and @xmath1 masses .
our task is to fined this potential fields for every particle .
potential energy of interaction of two point masses at rest which are on the distance @xmath4 from one another is @xmath5 where @xmath6 is the gravitational constant , and @xmath7 is the vector from particle `` 1 '' to particle `` 2 '' .
this expression one can write in another form @xmath8 here @xmath9 and @xmath10 are gravitational potentials created by particles @xmath0 and @xmath1
respectively @xmath11 coordinate @xmath12 is the distance between the particles .
it is not convenient to use this coordinate when particles are moving .
it is more convenient to use coordinate @xmath13 which one can reckon from the center of mass .
taking it into account we write energies of the particles `` 1 '' and `` 2 '' , which are located at points @xmath14 and @xmath15 similar to formula ( [ 2 ] ) @xmath16 @xmath17 here @xmath18 and @xmath19 are the potentials which we must find , @xmath20 and @xmath21 are the coefficients connected this potentials with the gravitational potentials generated by the particles @xmath0 and @xmath1 : @xmath22 this formulation suggests that we are describing not the energy of interaction of particles with one another but rather the potential energy of each particle separately , which is referenced to a certain level . as is common practice , we take for the zero level the energy of interaction of a particle @xmath23 with other particle at infinity .
let us turn now to the above potentials .
we shall use for this purpose the relations derived for the well known solution involving the reduced mass .
we place the origin of coordinates at the center of mass of the system under consideration . in this case
@xmath24 introduce a vector defining the relative positions of the particles @xmath25 we have defined this vector by @xmath7 rather than by @xmath26 as this is done usually , in order not to confuse it with the running coordinate @xmath26 which we reckon from the origin .
vector @xmath7 is the same vector which enter in the ( [ 1 ] ) , ( [ 2 ] ) and ( [ 3 ] ) formulas .
the equalities ( [ 6 ] ) and ( [ 7 ] ) yield @xmath27 differentiating eqs .
( [ 6 ] ) ( [ 8 ] ) with respect to time , we come to similar relations for the point velocities @xmath28 @xmath29 @xmath30 here @xmath31 and @xmath32 are the velocities of the particles under consideration , and @xmath33 is the relative particle velocity .
recall that it is the relative quantities @xmath7 and @xmath33 that are invoked in dealing with the problem with the use of reduced mass . and it is the coordinate @xmath7 that is employed to describe the trajectory of the @xmath2-particle .
we can use now eq .
( [ 11 ] ) to express the kinetic energy of each particle in terms of the relative velocity : @xmath34 @xmath35 in eqs .
( [ 12 ] ) and ( [ 13 ] ) , the kinetic energy is expressed both through the velocities of individual particles and the relative velocity .
an analysis of eqs .
( [ 12 ] ) and ( [ 13 ] ) suggests that the expressions for the energy written in terms of the velocities of individual particles and through the relative velocity differ in characteristic factors : @xmath36 we are going now to express the potential energy of each particle through its coordinates ( see eqs .
( [ 4 ] ) and ( [ 5 ] ) ) , whereas up to now the potential energy of interaction of two particles was defined in terms of a relative parameter , namely , separation distance between the particles @xmath37 ( see eq . ( [ 1 ] ) ) .
this suggests that a transition from description of the energy through separation distance between the particles , a relative parameter , to that in terms of the energy of each particle separately should bring about the appearance of factors of the kind of eq .
( [ 14 ] ) .
furthermore , by the virial theorem ( see , e.g. , @xcite , 10 , or @xcite , 6 ) , for the coulomb potential well the kinetic energy is related to the potential energy through @xmath38 where the line above denotes averaging over time .
for circular motion , this relation is correct without averaging .
thus , the potential energy is proportional to the kinetic energy .
this all adds up to the following recasting of eqs .
( [ 4 ] ) and ( [ 5 ] ) : @xmath39 @xmath40 we obtain @xmath41 and potentials @xmath42 @xmath43 both potentials @xmath18 and @xmath19 form a stationary potential wells placed at the origin in which the real particles @xmath0 and @xmath1 move .
each particle , however , moves in its `` own '' potential well , i.e. , the particle @xmath0 moves in the potential well formed by the potential @xmath19 , and the particle @xmath1 , in the potential well created by the potential @xmath18 . the potentials @xmath18 and @xmath19 are naturally nothing more than conventional potentials . they have been introduced by convention to describe the motion of the real particles @xmath0 and @xmath1 .
the time has come now to check whether the expressions for the potentials @xmath18 and @xmath19 are correct .
how can one verify the correctness of these relations ?
first , the sum of the potential energies of both particles should be equal to the potential energy ( [ 1 ] ) of interaction of the both particles with one another .
summing eqs .
( [ 15 ] ) and ( [ 16 ] ) in conjunction with eq .
( [ 8 ] ) , we see clearly that this equality is upheld for all @xmath14 and @xmath15 at any moment of time .
because each particle moves in its fixed potential well , the laws of conservation of energy and momentum should be met for each particle separately .
let us check it .
we use eqs .
( [ 12 ] ) and ( [ 13 ] ) to find the ratio of the kinetic energies of the two particles @xmath44 whence @xmath45 next we take eqs .
( [ 15 ] ) , ( [ 16 ] ) , and ( [ 8 ] ) to find the ratio of the potential energies of the two particles @xmath46 whence @xmath47 summing eqs .
( [ 21 ] ) and ( [ 23 ] ) term by term , we come to the total energies @xmath48 and @xmath49 of each particle @xmath50 or , in a more concise form @xmath51 thus , we see that the ratio of the total energies of each particle obeys the equality similar to the relations ( [ 20 ] ) and ( [ 22 ] ) @xmath52 the sum of the total energies of each particle should yield the total energy of the whole system . using equalities ( [ 25 ] )
we come to @xmath53 @xmath54 the total @xmath55 is conserved as the energy of a closed system . hence , as follows from equalities ( [ 27 ] ) and ( [ 28 ] ) , the total energies of each of the particles are conserved , i.e. , are coordinate independent @xmath56 thus , each particle moves in its potential well with its own energy .
consider now the angular momentum of the system and of the particles separately .
we express the angular momentum of particles through the coordinates @xmath57 and the relative velocity @xmath58 . combining eqs .
( [ 8 ] ) and ( [ 11 ] ) , we write the angular momenta of the particles as follows @xmath59= m_1\frac{m_2 ^ 2}{(m_1+m_2)^2}[\mathbf{l}\times\mathbf{v}],\ ] ] @xmath60= m_2\frac{m_1 ^ 2}{(m_1+m_2)^2}[\mathbf{l}\times\mathbf{v}].\ ] ] as seen from eqs .
( [ 30 ] ) and ( [ 31 ] ) , the vectors @xmath61 $ ] , @xmath62 $ ] and @xmath63 $ ] are oriented in the same direction . from the ratio of the momenta
@xmath64 we immediately obtain @xmath65 the sum of the momenta of the two particles should give the total angular momentum of the system .
we sum the momenta to obtain @xmath66 @xmath67 the total angular momentum @xmath68 , as the momentum of a closed system , is conserved .
but then , as seen from the equalities ( [ 34 ] ) and ( [ 35 ] ) , the momenta of each of the particles are conserved , i.e. , they are coordinate independent : @xmath69 because for each particle both the total energy and the angular momentum are conserved , one may consider the motion of each particle in its potential well separately , and calculate the energies and angular momenta also separately for each particle .
also , there is no need to use the reduced mass ; indeed , one may use instead real masses of each particle .
relation of energies of the particles must obey eqs .
( [ 26 ] ) , ( [ 27 ] ) and ( [ 28 ] ) , and relation of angular momenta eqs .
( [ 32 ] ) , ( [ 34 ] ) and ( [ 35 ] ) .
if , however , one has to know the positions of particles with respect to one another , one will have to consult eqs .
( [ 6 ] ) and ( [ 9 ] ) .
we note in conclusion that this approach permits one to separate completely not only the kinetic but the potential energies as well . indeed , instead of expressing the potential energy through the relative coordinate @xmath7 ( see eq .
( [ 1 ] ) ) , it can be defined in terms of the coordinates of each particle separately @xmath70
because in each potential well particles move saving its energies and angular momenta one can study the behavior of particles with the use of the well known relations describing the motion of one particle in a potential well ( see , e.g. , @xcite , 15 ) . as follows from these relations ,
if the potential energy in a central field is inversely proportional to @xmath71 ( @xmath71 is the radial coordinate ) , i.e. , it can be written as ( showing explicitly the sign ) here @xmath55 is the total energy of the particle , and @xmath81 is the squared angular momentum .
if it is necessary to find the trajectory of the particle one must know quantities @xmath55 and @xmath81 .
it is known that the same relations can be used if one analyzes the relative motion of two particles . in this case
, one has to put for mass @xmath23 in these relations the reduced mass @xmath82 .
a fictitious particle with mass @xmath2 describes the motion of a system of two particles .
we are going in what follows to label all quantities related to the fictitious particle @xmath2 by index @xmath2 .
let us see how the trajectories of each particle are related to that of the fictitious particle @xmath2 describing the behavior of a system of two particles . and also , whether it is possible to trace the possible trajectories of each particle without resorting to the intermediate @xmath2 particle . if the axis from which the angle @xmath74 is reckoned passes through the point of the ellipse closest to the focus ( the perihelion ) , the constant of integration @xmath76 in eq .
( [ 38 ] ) is zero , and if this point is farthest from the focus of the ellipse , @xmath83 ( with either of the signs ) .
the perihelia of the orbits of particles `` 1 '' and `` 2 '' look , however , in opposite directions .
this is seen clearly already from equality ( [ 6 ] ) ; indeed , for this equality to hold , vectors @xmath84 and @xmath85 must reach the maximum and the minimum simultaneously .
the orbits of the particles are shown in fig . 1 . in fig . 1
, the center of mass of the system is at point @xmath86 ( the origin ) .
also shown are the orbits of particles `` 1''and `` 2 '' , vector @xmath87 and the @xmath2 orbit ( dashed ) along which the fictitious particle with mass @xmath2 moves . as seen from fig . 1 , the constant @xmath76 for the ellipse `` 1 '' should be equal to @xmath88 , and for the ellipse `` 2 '' , @xmath89 .
using eqs .
( [ 26 ] ) , ( [ 32 ] ) , and ( [ 42 ] ) , one can readily verify that @xmath92 , i.e. , the eccentricities for both particles are equal .
this implies that the shape of the trajectory ( shape of the ellipse ) is the same for the two particles . as for the @xmath77 parameters , i.e. , the dimensions of the ellipses , these parameters obey the relation express vector @xmath57 from eq .
( [ 7 ] ) through the ellipse parameters ( the coordinate @xmath57 describes the trajectory of the @xmath2 particle ) .
we can use the general relation ( [ 38 ] ) to express @xmath84 and @xmath85 through the parameters of the ellipses . because vectors @xmath84 and @xmath85 are antiparallel , in polar coordinates
the vectors will differ by @xmath95 .
for the coordinate @xmath71 of the trajectory of the @xmath2 particle we have latexmath:[$r=|\mathbf l|= coordinate of vector @xmath85 is @xmath97 , because it is for the `` 2 '' ellipse that the constant @xmath89 . for the vector @xmath84 , the angular coordinate @xmath98 .
apart from this , as already mentioned , for the ellipse `` 1 '' the constant @xmath88 . using eqs .
( [ 44 ] ) and recalling that the directions of the vectors differ by @xmath95 , we come to express the parameters of the ellipse @xmath2 in terms of those of ellipses `` 1 '' and `` 2 ''
. the angular momentum of the system should be the sum of momenta of the particles and equal to the momentum of the @xmath2-particle , because it is the latter that describes the behavior of the system : recasting expression ( [ 39 ] ) to the form @xmath106 , we substitute in expression ( [ 47 ] ) @xmath107 from relation ( [ 46 ] ) , and @xmath108 from relation ( [ 41 ] ) , and take into account that for system of two particles @xmath109 .
the end result is thus we have once again come to the well known result , namely , if we express the parameters of the ellipse along which the @xmath2-point moves through the parameters of motion of real particles , then for the mass we should take @xmath82 , i.e. , the reduced mass . summing up
, we have shown that in the `` two - body problem '' one can operate without using the concept of the reduced mass . in this case , however , one should use not the energy of particle interaction ( [ 37 ] ) but rather the energy of each particle , which is reckoned from the zero level : it might seem at first glance that this complicates solution of the problem .
this approach permits one , however , to separate the potential energies and consider separately the motion of each particle in its potential well .
we note in conclusion that using the standard approach with the reduced mass requires a two - step procedure , in which one first finds the solution for the reduced mass , and after that , derive from this solution the parameters of motion of the particles of interest .
application of the above technique permits one to calculate _ * directly * _ the possible trajectories of _ * each * _ particle .
such approach one can use in the case when particles have electric charges @xmath115 and @xmath116 ( in the case under consideration charges must have opposite signs ) .
such approach can be used only if velocities of the charges @xmath117 ( @xmath76 is the light velocity ) . in this case
we can neglect by the energy of the magnetic field which appears in moving charged particles in comparison with the energy of the electric field .
said otherwise , in this case we neglect by the vector potential @xmath118 and claim that particles move in scalar potential @xmath119 only .
99 l.d.landau , e.m.lifshitz , mechanics , third edition , butterworth - heinemann , ( translation from the russian ) , 1976 .
i.i.olkhovskii , theoretical mechanics for physicists ( in russian ) , second edition , publ .
moscow university , moscow , 1974 . | in contrast to the well - known solution of the two - body problem through the use of the concept of reduced mass , a solution is proposed involving separation of potentials .
it is shown that each of the two point bodies moves in its own stationary potential well generated by the other body , and the magnitudes of these potentials are calculated .
it is shown also that for each body separately the energy and the angular momentum laws are valid .
the knowledge of the potentials in which the bodies are moving permits calculation of the trajectories of each body without resorting to the reduced mass . * key words : * mechanics , two - body problem , gravitational potential , virial theorem .
pacs : 45.50.jf |
this paper , following @xcite , is the second in a series of three devoted to the analysis of the resolvent family and spectral measure for the laplacian on an asymptotically hyperbolic , nontrapping manifold .
the third paper , by the first author alone , will establish global - in - time strichartz estimates on such a manifold .
let @xmath6 be an asymptotically hyperbolic manifold of dimension @xmath7 ( see section [ subsec : ahm ] for the precise definition of ` asymptotically hyperbolic ' ) .
let @xmath0 be the positive laplacian on @xmath6 , which is essentially self - adjoint on @xmath8 .
it is well known that the spectrum of @xmath0 is absolutely continuous on @xmath9 @xcite with possibly finitely many eigenvalues ( of finite multiplicity ) in @xmath10 .
we write @xmath11 for the operator @xmath12 where the subscript @xmath13 indicates positive part thus , @xmath11 vanishes on the pure point eigenspaces . in this paper , we analyze the spectral measure @xmath14 of the operator @xmath11 , under the assumption that @xmath6 is nontrapping ( that is , every geodesic reaches infinity both forward and backward ) and that there is no resonance at the bottom of the continuous spectrum , @xmath15 . to do this ,
we express the spectral measure @xmath14 in terms of the boundary values of the resolvent @xmath16 just ` above ' and ` below ' the spectrum in @xmath17 .
we then use the construction of the resolvent given by mazzeo and melrose @xcite ( valid when the spectral parameter lies in a compact set ) , melrose , s barreto and vasy @xcite ( high energy estimates for a perturbation of the hyperbolic metric ) and the present authors @xcite ( and , independently , @xcite ) in the general high - energy case to get precise information about the schwartz kernel of the spectral measure . in particular , following the work of the second author with guillarmou and sikora @xcite in the asymptotically conic setting , this will allow us to obtain precise pointwise bounds on the schwartz kernel , when ( micro)localized near the diagonal in a certain sense .
we then apply these pointwise kernel bounds to prove operator norm estimates on the spectral measure @xmath14 , and on general functions @xmath18 of the operator @xmath11 , again following the general strategy of @xcite .
however , there are key differences in the results we prove here compared to the asymptotically conic case , which can be traced to the exponential , as opposed to polynomial , growth of the volume of large balls in the present setting . in the case of the restriction theorem , that is , an @xmath19 bound on the spectral measure , we prove more : we obtain an estimate for all @xmath20 , while in the asymptotically conic case , it is well known that such an estimate fails for @xmath21 , where @xmath22 is the dimension . in the case of the spectral multiplier result ,
that is , boundedness of @xmath18 , where we assume only a finite amount of sobolev regularity on @xmath5 , boundedness on @xmath23 spaces fails for @xmath24 due to results of clerc - stein @xcite and taylor @xcite .
instead , we obtain boundedness on @xmath25 for @xmath20 , provided @xmath1 is negatively curved .
consider functions of an abstract ( unbounded ) self - adjoint operator @xmath26 on a hilbert space @xmath27 .
these are defined by the spectral theorem for unbounded self - adjoint operators ( for example , see @xcite ) .
one standard version of this theorem says that there is a one - to - one correspondence between self - adjoint operators @xmath26 and increasing , right - continuous families of projections @xmath28 , @xmath29 , having the property that the strong limit of @xmath28 as @xmath30 is the zero operator and as @xmath31 is the identity .
the correspondence is given by @xmath32 if @xmath33 is a real - valued borel function on @xmath34 , then @xmath35 with domain @xmath36 is self - adjoint .
here the formula means @xmath37 which can be interpreted as a stieltjes integral since @xmath38 is a nondecreasing function of @xmath39 .
we call @xmath40 the spectral measure associated with the operator @xmath26
. in particular we can apply this when @xmath41 and @xmath42 .
we then write @xmath14 for the spectral measure of @xmath43 .
since @xmath43 is a positive operator , we only need to integrate over @xmath44 in this case . returning to the abstract operator @xmath26 , the resolvent family @xmath45 is a holomorphic family of bounded operators on @xmath27 for @xmath46 . in many cases , including in the present setting , the resolvent family extends continuously to the real axis as a bounded operator in a weaker sense , e.g. between weighted @xmath47 spaces , and is then differentiable in @xmath39 up to the real axis .
in that case , we find that @xmath28 is differentiable in @xmath39 and we have stone s formula @xmath48 in this case we write ( abusing notation somewhat ) @xmath40 for the derivative of @xmath28 with respect to @xmath39 .
stone s formula gives a mechanism for analyzing the spectral measure , namely we need to analyze the limit of the resolvent @xmath49 on the real axis . in the case of @xmath43
, we notice that the spectral measure @xmath14 for @xmath43 is @xmath50 times the spectral measure at @xmath51 for @xmath0 .
this gives us the formula @xmath52 stein @xcite and tomas @xcite proved estimates for the restriction of the fourier transform of an @xmath53 function to the sphere @xmath54 : @xmath55.\ ] ] alternatively , we may formulate the estimate in terms of the the restriction operator @xmath56 to the hypersphere , @xmath57 the stein - tomas theorem is equivalent to the boundedness of @xmath58 , which in turn is equivalent to the boundedness of @xmath59 the schwartz kernel of @xmath60 , @xmath61 is @xmath62 times the spectral measure @xmath63 for the square root of the flat laplacian on @xmath64 , since the spectral projection @xmath65 of @xmath66 is @xmath67 .
therefore , one may rewrite the restriction theorem as the following estimate : @xmath68 provided @xmath69.$ ] this naturally leads to the question : for which riemannian manifolds @xmath70 does the spectral measure for @xmath71 map @xmath72 to @xmath73 for some @xmath20 , and how does the norm depend in the spectral parameter ?
we refer to such an estimate as a ` restriction estimate ' or a ` restriction theorem ' .
such a result is a continuous spectral analogue of the well - known discrete restriction theorem of sogge ( * ? ?
* chapter 5 ) . as the present paper is inspired by work by the second author with guillarmou and sikora @xcite on asymptotically conic spaces , we review the results of @xcite here .
asymptotically conic spaces @xmath74 , of dimension @xmath75 , are modelled on spaces that at infinity look like the ` large end of a cone ' ; that is , have one end diffeomorphic to @xmath76 , where @xmath77 is a closed manifold of dimension @xmath78 , with a metric of the form @xmath79 where @xmath80 a metric on @xmath77 .
such spaces are euclidean - like at infinity , in the sense that the volume of balls of radius @xmath81 are uniformly bounded above and below by multiples of @xmath82 , and in the sense that the curvature tends to zero , and the local injectivity radius tends to infinity , at infinity .
if we add the condition that the manifold be nontrapping , then such spaces are also dynamically similar to euclidean space ( although they may have conjugate points ) .
consequently , the spectral analysis of such spaces behaves in many ways like euclidean space .
this is illustrated by the results from @xcite .
on @xmath83 , the spectral measure satisfies pointwise kernel bounds of the form @xmath84 and this estimate is essentially optimal , in the sense that neither exponent can be improved . in @xcite it was shown that , if @xmath74 is an asymptotically conic nontrapping manifold , and @xmath0 its laplacian , then there is a partition of unity @xmath85 , depending on @xmath39 , and @xmath86 such that @xmath87 for @xmath88 $ ] , where @xmath89 is the riemannian distance in @xcite , but in @xcite it was observed that the same construction gives the more general estimates in . ] .
the @xmath90 are semiclassical pseudodifferential operators ( with semiclassical parameter @xmath91 with small microsupport .
therefore , the operators @xmath92 can be considered to be the kernel of the spectral measure ( micro)localized near the diagonal . moreover , in the case where there are no conjugate points , then the estimate above is valid without the partition of unity . this estimate was shown to imply a global restriction estimate , that is , an @xmath93 operator norm bound on @xmath94 . in fact , this was proved at an abstract level : [ kernel2restriction ] let @xmath95 be a metric measure space , and @xmath26 an abstract positive self - adjoint operator on @xmath96 .
suppose that the spectral measure @xmath97 has a schwartz kernel satisfying ( with @xmath98 replaced by @xmath89 ) for @xmath99 , as well as for @xmath100 and @xmath101 if @xmath75 is even , or @xmath102 and @xmath103 if @xmath75 is odd .
then the operator norm estimate @xmath104 holds for all @xmath105 .
moreover , if the kernel estimates above hold for some range of @xmath39 , then holds for @xmath39 in the same range . finally , it was shown in @xcite that , at an abstract level , such a restriction estimate implies spectral multiplier estimates : [ restriction2multiplier ] let @xmath95 be a metric measure space , such that the volume of each ball of radius @xmath81 is comparable to @xmath82 .
suppose @xmath0 is a positive self - adjoint operator such that @xmath106 satisfies finite propagation speed on @xmath107 , and the restriction theorem @xmath108 holds uniformly with respect to @xmath105 for @xmath109 .
then there is a uniform operator norm bound on spectral multipliers on @xmath23 of the form @xmath110 where @xmath111 is an even function supported in @xmath112 $ ] , and @xmath113 .
in particular , one concludes and when @xmath1 is an asymptotically conic nontrapping manifold of dimension @xmath22 .
we next consider existing results on hyperbolic space .
we return to our convention where the dimension is @xmath7 . using explicit formulae for the schwartz kernel of functions of the operator @xmath114
, we deduce pointwise bounds @xmath115 for @xmath116 , and derivative estimates also , but we do not need such estimates in the low energy case . ]
@xmath117 when @xmath118 .
closely related pointwise bounds for the wave kernels @xmath119 and @xmath120 , the heat kernel @xmath121 and the schrdinger propagator @xmath122 on hyperbolic space have been exploited in various works ; see for example @xcite , @xcite , @xcite , @xcite . to the authors knowledge ,
the recent paper @xcite by huang and sogge is the only previous paper in which restriction estimates for hyperbolic space have been considered .
huang and sogge proved restriction estimates for @xmath123 in the same range @xmath124 $ ] as for euclidean space , using the exact expression for the hyperbolic resolvent , and complex interpolation , in the manner of stein s original proof of the stein - tomas restriction theorem @xcite ( this argument was presented in an abstract formulation in @xcite ) .
in fact , on hyperbolic space ( and , as we shall show , asymptotically hyperbolic nontrapping spaces ) , restriction estimates are valid for all @xmath125 see section [ sec : model ] for a very simple proof on @xmath126 .
spectral multiplier estimates on hyperbolic and asymptotically hyperbolic spaces on @xmath53 spaces ( much more general than those considered here ) have been well studied .
it was pointed out by clerc and stein @xcite for symmetric spaces and taylor @xcite for spaces with exponential volume growth and @xmath127 bounded geometry that a _ necessary _
condition for @xmath128 to be bounded is that @xmath5 admit an analytic continuation to a strip in the complex plane .
cheeger , gromov and taylor @xcite , and taylor @xcite showed that if @xmath74 has @xmath127 bounded geometry and injectivity radius bounded from below , then @xmath129 maps @xmath130 into itself for @xmath131 , provided that @xmath5 is holomorphic and even on the strip @xmath132 for some @xmath133 and satisfies symbol estimates @xmath134 on the strip . by constrast
, we want to consider the mapping properties of @xmath128 where @xmath5 has only finite sobolev regularity .
this is motivated by typical applications of spectral multipliers in harmonic analysis , such as riesz means , and in pde , in which one often wants to restrict to a dyadic frequency interval , that is , to the range of a spectral projector of the form @xmath135}(p)$ ] , or a smoothed version of this .
clearly , such a spectral multiplier can not have an analytic continuation to a strip . on the other hand ,
the work of clerc - stein and taylor shows that boundedness on @xmath53 , @xmath24 , can not be expected .
this motivates us to search for replacements for @xmath53 spaces , on which spectral multipliers are bounded .
the geometric setting in the present paper is that of asymptotically hyperbolic manifolds .
an asymptotically hyperbolic manifold @xmath6 is the interior of a compact manifold @xmath1 with boundary , such that the riemannian metric @xmath136 takes a specific degenerate form near the boundary of @xmath1 .
specifically , near each boundary point , there are local coordinates @xmath137 , where @xmath138 is a boundary defining function and @xmath139 restrict to local coordinates on @xmath140 , such that @xmath136 takes the form @xmath141 where @xmath142 is a family of metrics on @xmath140 , smoothly parametrized by @xmath138 . under the metric @xmath136 ,
the interior @xmath143 of @xmath1 is a complete riemannian manifold .
as is well known , @xmath7-dimensional hyperbolic space takes this form in the poincar ball model .
indeed , @xmath144 is given by the interior of the unit ball in @xmath145 , with the metric @xmath146 where @xmath147 are the standard coordinates on @xmath145 .
other examples include all convex co - compact hyperbolic manifolds , and compactly supported metric perturbations of these .
such spaces are termed asymptotically hyperbolic spaces as the sectional curvatures tend to @xmath148 at infinity @xcite .
analytically , they have many similarities to hyperbolic spaces .
consider the resolvent @xmath149 on @xmath144 , which is well - defined as a bounded operator on @xmath150 for @xmath151 .
notice that the axis @xmath152 corresponds to the spectrum of @xmath0 , and the point @xmath153 corresponds to the point @xmath154 in the spectrum of @xmath155 . on @xmath144 ,
the resolvent @xmath156 extends to a holomorphic function of @xmath157 when @xmath158 is even , and a meromorphic function with poles at @xmath159 when @xmath158 is odd . for asymptotically hyperbolic spaces , it is known from work of mazzeo - melrose @xcite and guillarmou @xcite that the resolvent @xmath160 extends to be a meromorphic function of @xmath161 on @xmath162 , and extends to be meromorphic on the whole of @xmath17 provided that @xmath136 is even in @xmath138 , that is , a smooth function of @xmath163 .
in addition , it is holomorphic in a neighbourhood of the spectral axis @xmath152 except possibly at the point @xmath164 itself , corresponding to the bottom of the continuous spectrum , which could be a simple pole @xcite . in the present article
, we shall _ assume _ that the resolvent is holomorphic at @xmath165 as well .
we point out that our estimates will certainly fail in the case of a zero - resonance , but weaker estimates will remain valid ; see @xcite , @xcite for an analysis of zero - resonances in the asymptotically euclidean case .
our first main result , analogous to , is that there is a partition of the identity , @xmath166 on @xmath107 such that the diagonal terms in the two - sided decomposition of @xmath14 satisfy the same type of pointwise bounds as are valid on hyperbolic space .
in fact , following @xcite , we prove a slightly stronger result , in which we retain information about the oscillatory nature of the kernel as @xmath167 . before stating the result
, we refer to section [ sec : ahm ] for the definition of the double space @xmath168 , the blow - up of @xmath169 at the boundary of the diagonal ; see figure [ fig ] .
this space has 3 boundary hypersurfaces : the lift to @xmath168 of the left and right boundaries in @xmath169 , denoted @xmath170 and @xmath171 , respectively , and the ` front face ' @xmath172 created by blowup .
we denote boundary defining functions for these boundary hypersurfaces by @xmath173 , @xmath174 and @xmath175 respectively .
[ thm : kernelbounds1 ] let @xmath6 be an asymptotically hyperbolic nontrapping manifold with no zero - resonance , and let @xmath11 be given by . then for low energies , @xmath116 , the schwartz kernel of the spectral measure @xmath176 takes the form @xmath177 where @xmath178_{{{\boldsymbol\lambda } } } \times x^2_0)$ ] . for high energies , @xmath118 , one can choose a finite pseudodifferential operator partition of the identity operator , @xmath179 such that the @xmath180 are bounded on @xmath53 , uniformly in @xmath181 , for each @xmath182 , and such that the microlocalized spectral measure , that is , any of the compositions @xmath183 , @xmath184 , takes the form @xmath185 where @xmath186 is in @xmath187_{{{\boldsymbol\lambda}}^{-1 } } \times x^2_0)$ ] and @xmath188 is in @xmath189_{{{\boldsymbol\lambda}}^{-1 } } \times x^2)$ ] , and the functions @xmath190 satisfy the following . for small distance , @xmath191 , we have @xmath192 for @xmath193 , @xmath190 is @xmath194 times a smooth function of @xmath195 , decaying to order @xmath164 at @xmath170 and @xmath171 : @xmath196_{{{\boldsymbol\lambda}}^{-1 } } \times x^2_0 ) .
\label{bpm2.5}\ ] ] moreover , if @xmath6 is in addition simply connected with nonpositive sectional curvatures , then the estimates above are true for the spectral meaure without microlocalization , i.e. in this case we can take @xmath197 to be the trivial partition of unity .
we can split the continuous spectrum of @xmath43 at any point @xmath198 to differentiate high and low energies .
using this structure theorem , we prove pointwise bounds on the microlocalized spectral measure : [ thm : kernelbounds ] let @xmath6 be as above .
then for low energies , @xmath116 , we have pointwise estimates on the spectral measure of the form @xmath199 for high energies , @xmath118 , one has , for sufficiently small @xmath86 and @xmath200 $ ] @xmath201 as before , if @xmath6 is in addition simply connected with nonpositive sectional curvatures , then the estimates above are true for the spectral meaure without microlocalization , i.e. in this case we can take @xmath202 to be the trivial partition of unity . using theorem [ thm : kernelbounds ] , we prove [ thm : restriction ] suppose @xmath203 is an @xmath204-dimensional non - trapping asymptotically hyperbolic manifold with no resonance at the bottom of the continuous spectrum
. then we have the following estimate for @xmath116 : @xmath205 for @xmath118 , we have the estimate @xmath206 the range of exponents @xmath123 is greater for a hyperbolic space than for a conic ( euclidean ) space .
indeed , it includes all @xmath207 , while on euclidean space @xmath64 , the well - known knapp example shows that the restriction estimate can not hold for @xmath208 .
( the knapp example does not apply to hyperbolic space as it relies on the dilation symmetry of @xmath64 . ) for high energies , @xmath118 , the exponent is the same as on @xmath64 for the range @xmath209 but again we get the full range of @xmath123 up to @xmath210 .
this surprising result is closely tied to a non - euclidean feature of hyperbolic space related to the kunze - stein phenomenon @xcite .
the kunze - stein phenomenon for semisimple lie groups is that there is a much larger set of exponents @xmath211 for which one has @xmath212 compared to euclidean space . since @xmath144 can be viewed as @xmath213 , this has consequences for convolution on @xmath144 .
anker and pierfelice @xcite , ( * ? ? ?
* section 4 ) showed that convolution with a radial kernel @xmath214 satisfies @xmath215 from this we see that if @xmath214 is smooth and decays as @xmath216 , then convolution with @xmath217 maps @xmath53 to @xmath218 for all @xmath20 .
additionally , this non - euclidean feature also affects the range of valid strichartz estimates on ( asymptotically ) hyperbolic manifolds see @xcite .
our result for spectral multipliers is restricted to the case where the manifold is , in addition , a cartan - hadamard manifold , i.e. simply connected with nonpositive sectional curvatures .
[ thm : sm ] suppose @xmath203 is an @xmath204-dimensional non - trapping asymptotically hyperbolic manifold with no resonance at the bottom of spectrum .
suppose in addition that @xmath1 is simply connected with nonpositive sectional curvatures .
then for any @xmath219 supported in @xmath112 $ ] with @xmath220 , and for all @xmath20 , @xmath221 is a bounded operator on @xmath222 uniformly with respect to parameter @xmath223 for @xmath224 , in the sense @xmath225 } \big\| f(\alpha p)\big\|_{l^p(x ) + l^2(x ) \longrightarrow l^p(x ) + l^2(x ) } < \infty.\ ] ] this is weaker than theorem [ restriction2multiplier ] , both because the function space is @xmath222 rather than @xmath53 , but also because we have strengthened the sobolev condition to @xmath226 for all @xmath123 . from the perspective of harmonic analysis
, it would be interesting to find a ` better ' function space , that is , more closely associated to the laplacian , to accommodate the boundedness of the spectral multiplier .
modern harmonic analysis ( caldern - zygmund theory ) is generally built on spaces with a doubling measure , which activates some kind of covering lemma and gives a simple structure of cube nets .
though some authors have investigated non - doubling spaces , the advances are mainly restricted to spaces of polynomial growth , which are semi - doubling " . in any case , the harmonic analysis on space of exponential growth is barely explored .
one recent work along these lines is due to bouclet @xcite , where it is shown that semiclassical spectral multipliers are bounded on appropriate _ weighted _
@xmath53 spaces in a setting with exponential volume growth .
the authors plan to pursue this question in future publications . in the third paper in this series , @xcite
, the first author will prove global - in - time strichartz type estimates without loss on non - trapping asymptotically hyperbolic manifolds .
namely , for solutions of the inhomogeneous schrdinger equation , @xmath227 with @xmath228 and @xmath5 orthogonal to eigenfunctions of @xmath0 on an @xmath204-dimensional asymptotically hyperbolic manifold @xmath1 , one has the estimate @xmath229 provided the pairs @xmath230 and @xmath231 are hyperbolic schrdinger admissible pairs of exponents .
the paper is organized as follows . in section [ sec :
model ] , we show how the main results in section [ subsec : mr ] follow in the simple case of hyperbolic 3-space @xmath126 . in section [ sec : ahm ] , we review the geometry and analysis of asymptotically hyperbolic manifolds , recalling the main results of @xcite and @xcite . in section [ sec : le ] we prove the restriction estimate , theorem [ thm : restriction ] , for low energy , which exploits , in some sense , the kunze - stein phenomenon on @xmath144 . in section [ sec : wf ] , in preparation for the high - energy estimates , we show how the microlocal support of the spectral measure may be localized by pre - and post - composition by pseudodifferential operators . in section
[ sec : he ] we prove theorem [ thm : restriction ] for high energy .
this uses , in a crucial way , the semiclassical lagrangian structure of the high - energy spectral measure proved in @xcite and @xcite .
finally , in section [ sec : sm ] , we prove the spectral multiplier result , theorem [ thm : sm ] .
the authors would like to thank c. guillarmou , a. mcintosh and a. sikora for various helpful discussions during working on this paper .
the authors gratefully acknowledge the support of the australian research council through discovery grant dp120102019 .
in this section we illustrate the results of theorems [ thm : kernelbounds ] , [ thm : restriction ] and [ thm : sm ] in the simple case of hyperbolic space .
we focus on the case of @xmath126 , in which the formulae are particularly simple . hyperbolic space can be defined in terms of the half space model @xmath233 equipped with the metric @xmath234 or in terms of the poincar disc model , as in . for odd dimensions , that is , when @xmath235 is even , the schwartz kernel of @xmath236 is given by the explicit formula @xmath237 where @xmath238 as before , and @xmath239 is geodesic distance on @xmath240 .
see @xcite for proof .
in particular , @xmath241 for @xmath242 is @xmath243 setting now @xmath244 , and applying stone s formula , we find that on @xmath126 , @xmath245 next , we deduce theorem [ thm : restriction ] for @xmath126 .
the estimate for low energy follows immediately from and .
the estimate for high energy and @xmath246 $ ] can be deduced from theorem [ kernel2restriction ] : @xmath247 maps @xmath248 to @xmath249 with a bound @xmath250 for all @xmath105 , provided @xmath251 .
we assert the kernel estimates of theorem [ kernel2restriction ] hold for this spectral measure , that is , @xmath252 in fact , one may see @xmath253 when @xmath254 ; @xmath255 when @xmath256 . on the other hand ,
it is clear that @xmath257 then applying theorem [ kernel2restriction ] proves the proposition . in the range @xmath258 and for high energy
, we again use complex interpolation , but rather than applying theorem [ kernel2restriction ] as a black box , we need to modify the proof slightly .
we observe that the spectral measure on @xmath126 satisfies @xmath259 we substitute this estimate in place of the kernel bounds of theorem [ kernel2restriction ] , and run the proof of ( * ? ? ?
* section 3 ) . as in that proof , we consider the analytic family of operators @xmath260 .
the proof works just the same ; for more details . ] in place of equation ( 3 - 7 ) of ( * ? ? ? * section 3 ) and the previous equation , we obtain @xmath261 on the line @xmath262 , for any @xmath263 .
let @xmath264 , and choose @xmath265 .
using the fact that the spectral measure is @xmath266 , and applying complex interpolation , we find that @xmath267 the hyperbolic space @xmath232 is a non - doubling space but rather has exponential volume growth , i.e. the volume of a ball with radius @xmath239 satisfies @xmath268 the lack of doubling means that we can not apply theorem [ restriction2multiplier ] directly .
nevertheless , we can decompose the kernel of a spectral multiplier @xmath18 into two parts , one supported where @xmath269 and one supported where @xmath270 , using a cutoff function @xmath271 , say , the characteristic function of @xmath272 on @xmath273 .
then the proof of theorem [ restriction2multiplier ] applies to @xmath274 , since all that is required for this proof to work is that doubling is valid for all balls of radius @xmath275 , which is certainly true .
we obtain [ smh3 ] for every even function @xmath111 supported in @xmath112 $ ] with @xmath276 , @xmath277 maps @xmath248 to itself with a uniform bound @xmath278 provided @xmath251 , where @xmath279 is the characteristic function of the set @xmath280 . in particular , if @xmath281 , then this is valid for @xmath282 , and thus by interpolation and duality for all @xmath283 $ ] . for the other part , supported where @xmath270 , we show boundedness from @xmath284 . by interpolation ,
it is enough to treat the case @xmath282 , since boundedness @xmath285 follows immediately from the boundedness of @xmath5 .
the @xmath286 operator norm of an integral operator @xmath287 is bounded by @xmath288 we express the kernel of @xmath289 using .
so we need to estimate @xmath290 write @xmath291 . for any @xmath292
, we get the estimate for @xmath293 : @xmath294 using the compact support of @xmath5 . combining this estimate with lemma [ smh3 ] ,
we have proved theorem [ thm : sm ] in the case of @xmath126 .
suppose @xmath6 is an @xmath296-dimensional asymptotically hyperbolic manifold .
let @xmath1 be the compactification .
we write @xmath138 for a boundary defining function , and use local coordinates @xmath297 near a boundary point of @xmath1 , where @xmath298 restrict to coordinates on @xmath140 , or @xmath147 in the interior of @xmath1 .
consider the space of smooth vector fields on the compactification , @xmath1 , that are of uniformly finite length . due to the factor @xmath299 in the metric , such vector fields take the form @xmath300 , where @xmath301 is a smooth vector field on @xmath1 .
such vector fields are called @xmath295-vector fields , spanned over @xmath302 near the boundary by the vector fields @xmath303 and @xmath304 , @xmath305 . as observed by mazzeo - melrose
, they are the space of sections of a vector bundle , known as the @xmath295-tangent bundle , @xmath306 . the dual bundle , known as the @xmath295-cotangent bundle and denoted @xmath307 , is spanned by local sections @xmath308 and @xmath309 near the boundary .
it follows that , near the boundary of @xmath1 , we can write points @xmath310 in the form @xmath311 this defines linear coordinates @xmath312 on each fibre of @xmath307 ( near the boundary ) , depending on the coordinate system @xmath137 .
the laplacian @xmath0 on @xmath1 is built out of an elliptic combination of @xmath295-vector fields .
in fact , in local coordinates @xmath137 near the boundary of @xmath1 , with @xmath136 taking the form , it takes the form @xmath313 where we use @xmath314 to denote differential operators of order @xmath315 generated over @xmath302 by @xmath295-vector fields .
we would like to understand the nature of the schwartz kernel of the resolvent @xmath316 , on @xmath317 .
following mazzeo - melrose , we use a compactification of the double space @xmath317 that reflects the geometry of @xmath6 , particularly near the diagonal .
this is important as we want to view the resolvent as some sort of pseudodifferential operator , which means that we need a precise notion of what it means for a distribution to be conormal to the diagonal , uniformly out to infinity .
compactifying @xmath143 to @xmath1 , we can initially view the resolvent kernel on @xmath169 .
however , on this space , the diagonal is not a p - submanifold where it meets the boundary .
that is , near the boundary of the diagonal in @xmath169 , there are no local coordinates of the form @xmath318 where @xmath138 , resp .
@xmath319 is a boundary defining function for the left , resp .
right , copy of @xmath1 and @xmath320 are the remaining coordinates , such that the diagonal is given by the vanishing of a subset of these coordinates . to give a workable definition of conormality to a submanifold , we require it to be a p - submanifold .
to remedy this , we blow up ( in the real sense ) the boundary of the diagonal .
this creates a manifold with corners , denoted @xmath168 , the ` @xmath295-double space ' , with three boundary hypersurfaces : the two original ones , @xmath170 ` left face ' and @xmath171 ` right face ' , corresponding to @xmath321 and @xmath322 in @xmath169 , and the new face @xmath172 , the ` front face ' , created by blowup see figure [ figure : x20 ] .
we denote a generic boundary defining function for @xmath323 or @xmath172 by @xmath324 and @xmath325 , respectively .
the @xmath295-blown - up double space @xmath326,scaledwidth=60.0% ] as in @xcite , we write down coordinate systems in various regions of @xmath168 , in terms of coordinates @xmath327 near the boundary of @xmath1 , or @xmath147 in the interior of @xmath1 .
the unprimed coordinates always indicate those lifted from the left factor of @xmath1 , while primed coordinates indicate those lifted from the right factor .
we label these different regions as follows : * region 1 : in the interior of @xmath168 .
here we use coordinates @xmath328 * region 2a : near @xmath170 and away from @xmath172 and @xmath171 . in this region , we use @xmath329 .
* region 2b : near @xmath171 and away from @xmath172 and @xmath170 .
symmetrically , we use @xmath330 .
* region 3 : near @xmath331 and away from @xmath172 .
here we use @xmath332 .
* region 4a : near @xmath172 and away from @xmath171 .
this is near the blowup .
in this region we can use @xmath333 for a boundary defining function for @xmath172 .
we use coordinate system @xmath334 * region 4b : near @xmath172 and away from @xmath170 .
symmetrically , we use @xmath335 * region 5 : near the triple corner @xmath336 . in this case
, a boundary defining function for @xmath172 is @xmath337 . by rotating the @xmath139 coordinates
, we can assume that @xmath338 in a neighbourhood of any given point in the triple corner .
assuming this , we use coordinates @xmath339 on @xmath168 , the lift of the diagonal , denoted @xmath340 , meets the boundary in the interior of the front face @xmath172 .
it has several good geometric properties : * @xmath341 is a p - submanifold disjoint from @xmath170 and @xmath171 ; * the @xmath295-vector fields @xmath303 , @xmath304 lift from the left and right factors of @xmath1 to be vector fields on @xmath168 that are non - tangential to @xmath340 , uniformly down to the boundary of @xmath340 .
moreover , these vector fields span the normal bundle of @xmath340 , again uniformly down to the boundary .
* the distance function @xmath342 is smooth in a deleted neighbourhood of @xmath340 , and its square is a quadratic defining function for the lifted diagonal , i.e. it is smooth and vanishes to precisely second order at @xmath340 .
taking advantage of the first and second geometric properties listed above , mazzeo and melrose ` microlocalized ' the space of @xmath295-differential operators to a calculus of @xmath295-pseudodifferential operators on @xmath1 .
the set of pseudodifferential operators of order @xmath75 on @xmath1 , denoted @xmath343 , is , by definition , the set of operators on half - densities , whose schwartz kernels are conormal of order @xmath75 to @xmath340 , and vanish to infinite order at @xmath170 and @xmath171 .
mazzeo and melrose @xcite showed that the resolvent @xmath344 takes the form @xmath345 for low energy , this description is precise enough to deduce kernel estimates for the spectral measure , restriction estimates , and spectral multiplier theorems .
however , as @xmath167 , we need a uniform description of the resolvent , and in particular we need to understand its oscillatory nature .
for this , we use the description by melrose - sa barreto - vasy @xcite , wang @xcite and the present authors @xcite ( in the first paper of this series ) of the high - energy resolvent as a semiclassical lagrangian distribution .
this is associated to the bicharacteristic relation on @xmath346 , that is , the submanifold of @xmath347 given by @xmath348 which is a smooth lagrangian submanifold provided that @xmath1 is nontrapping . by ` bicharacteristic '
we mean here the integral curves of the symbol of @xmath0 on the set where @xmath349 . in this case
these are precisely geodesics , viewed as living in the cotangent bundle .
the bicharacteristic relation splits into the forward and backward bicharacteristic relations , @xmath350 and @xmath351 , which was denoted @xmath352 in @xcite . ]
consist of those points @xmath353 for which @xmath354 is on the forward / backward half of the bicharacteristic relative to @xmath355 .
these two halves meet at @xmath356 , where @xmath357 denotes the conormal bundle of the diagonal , @xmath358 we wish to understand the way in which @xmath359 compactifies when viewed as living over the double space @xmath168 .
we consider the bundle @xmath360 , obtained by pulling back the bundle @xmath361 to @xmath168 by the blowdown map @xmath362 .
we denote the bundle projection maps by @xmath363 .
then , as explained in ( * ? ? ?
* section 3 ) , it is convenient to ` shift ' @xmath359 by the map @xmath364 defined by @xmath365 for some choice of boundary defining functions @xmath173 for @xmath170 and @xmath174 for @xmath171 ; that is , we consider @xmath366 .
it is convenient here to assume that @xmath173 and @xmath174 are both constant near @xmath340 , so that these two shifted lagrangian join smoothly at @xmath367 . in @xcite
we showed follow immediately . ]
[ prop : flowout - structure ] the bicharacteristic relation @xmath359 can be expressed as the union of two relatively open subsets @xmath368 , having the following properties .
* @xmath369 contains a neighbourhood of the intersection @xmath356 in @xmath359 , that is , the points @xmath370 .
* let @xmath371 denote the lift of @xmath369 to @xmath360 , together with its limit points lying over @xmath172 , @xmath170 and @xmath171 .
let @xmath372 denote the two halves of this submanifold , meeting at @xmath367 .
then @xmath373 are manifolds with codimension three corners , with the property that the interior of @xmath374 is the graph of the differential of the distance function on some deleted neighbourhood @xmath301 of @xmath375 , and the interior of @xmath376 is the graph of minus the differential of the distance function on @xmath301 .
thus the projection @xmath377 has full rank restricted to @xmath371 , except at @xmath378 , where the rank of the projection @xmath377 drops by @xmath158 .
the boundary hypersurfaces of @xmath371 are @xmath379 , lying over @xmath172 , @xmath380 , lying over @xmath170 and @xmath381 , lying over @xmath171 .
* the image @xmath382 of @xmath371 under the shift is a smooth lagrangian submanifold of @xmath383 ( nb : the standard cotangent bundle , not @xmath360 ) with codimension three corners .
the projection @xmath384 restricts to a map @xmath385 with full rank , except at @xmath386 , where the rank of the projection @xmath387 drops by @xmath158 .
* let @xmath388 denote the image of @xmath389 under the shift , and let @xmath390 denote the closure of @xmath388 in @xmath391 .
then @xmath390 is a smooth lagrangian submanifold of @xmath391 ( nb : the standard cotangent bundle , not the @xmath295-cotangent bundle ) with codimension two corners .
in terms of these lagrangian submanifolds we determined the semiclassical nature of the resolvent kernel in ( * ? ? ? * theorem 38 ) .
in view of stone s formula , , this has the ( almost ) immediate consequence for the spectral measure : [ thm : specmeas ] let @xmath6 be an asymptotically hyperbolic non - trapping manifold , with no resonance at the bottom of the continuous spectrum .
then the spectral measure @xmath14 with @xmath392 can be expressed as a sum of the following terms : * a semiclassical lagrangian distribution in @xmath393 , where @xmath371 is as in proposition [ prop : flowout - structure ] .
* an element of @xmath394 ; { { } ^0\omega^{1/2 } } ) + ( \rho_l \rho_r)^{n/2+i / h } h^\infty c^\infty(x^2_0\times [ 0 , h_0 ] ; { { } ^0\omega^{1/2}}),\ ] ] which can be regarded as an element of type ( i ) of order @xmath395 .
* a kernel lying in @xmath396 also associated to the bicharacteristic flowout , as above , but living on @xmath169 rather than @xmath168 ; * an element of @xmath397 ; { { } ^0\omega^{1/2 } } ) + ( xx')^{n/2+i / h } h^\infty c^\infty(x^2\times [ 0 , h_0 ] ; { { } ^0\omega^{1/2}}),\ ] ] which can be regarded as an element of type ( iii ) of order @xmath395 .
we first remark that the change in order from @xmath398 for the resolvent in ( * ? ? ?
* theorem 38 ) to @xmath399 for the spectral measure is simply due to the fact that the semiclassical resolvent in @xcite is @xmath400 times the resolvent in , together with the factor of @xmath401 in . in (
* theorem 38 ) it was shown that the resolvent kernel has a similar , but slightly more complicated structure : in place of the first term above , it consists of a semiclassical pseudodifferential operator , together with a semiclassical intersecting lagrangian distribution associated to @xmath367 together with the forward / backward half of the bicharacteristic relation ( for the outgoing / incoming resolvent ) .
we claim that when the incoming resolvent is subtracted from the outgoing , the pseudodifferential part cancels , and what is left is a lagrangian distribution associated to the full bicharacteristic relation .
this follows since the spectral measure satisfies an elliptic equation @xmath402 therefore , the spectral measure can have no semiclassical wavefront set outside the zero set of the symbol of @xmath403 .
this excludes all of @xmath367 except for that part contained in @xmath359 .
in addition , propagation of lagrangian regularity is an operator of real principal type , @xmath404 , and @xmath405 is a lagrangian distribution microlocally in some region @xmath301 of phase space , then @xmath405 is also lagrangian along the bicharacteristics passing through @xmath301 .
it follows from the parametrix construction for lagrangian solutions of operators of real principal type , and the propagation of singularities theorem . ]
shows that the spectral measure is a lagrangian distribution across @xmath367 ( given that we already know that it is lagrangian on both sides of @xmath367 corresponding to forward and backward flowout , and given that the hamilton vector field of the symbol does not vanish at @xmath406 ) .
this concludes the proof . the distance function on @xmath168 satisfies [ prop : dist ] on @xmath168 , the riemannian distance function @xmath407 is given by @xmath408 where @xmath409 is uniformly bounded on @xmath168 .
the result in the case that @xmath6 is a small perturbation of @xmath410 was shown by melrose ,
s barreto and vasy ( * ? ? ?
* ; * ? ? ?
* section 2 ) .
consider two points @xmath411 .
when @xmath412 are in a sufficiently small neighbourhood @xmath413 of the front face @xmath172 , say @xmath414 , @xmath415 with @xmath416 and @xmath417 ( taken with respect to the metric @xmath418 at the boundary ) , then the distance function parametrizes the lagrangian @xmath371 , and it follows from ( * ? ? ?
* proposition 20 ) that this takes the form @xmath419 in a neighbourhood of @xmath172 .
define @xmath420 to be the compact set @xmath421 .
let @xmath422 be the diameter of @xmath423 , that is , the maximum distance between two points of @xmath423
. now suppose that @xmath424 . in the complement of @xmath413
, we can take @xmath425 and @xmath426 .
if both @xmath123 and @xmath427 lie in @xmath423 , then the distance between @xmath123 and @xmath427 is at most @xmath74 , hence @xmath428 .
if one point , say @xmath123 , lies in @xmath423 and @xmath427 is not in @xmath423 , then a lower bound on @xmath429 is the distance from @xmath427 to the boundary of @xmath423 , which is exactly @xmath430 . on the other hand ,
an upper bound is the length of the path from @xmath427 to the closest point @xmath431 on @xmath432 , plus the distance from @xmath431 to @xmath123 .
this is at most @xmath433 .
if neither point lies in @xmath423 , then write @xmath414 and @xmath415 . due to the definition of @xmath413
, we must have @xmath434 .
we claim that any geodesic between @xmath123 and @xmath427 must enter @xmath423 .
it follows from this claim that a lower bound on the distance between @xmath123 and @xmath427 is the distance from @xmath123 to @xmath432 plus the distance between @xmath427 to @xmath432 , which is @xmath435 , that is , @xmath436 . also , an upper bound on the distance is clearly @xmath437 which is also @xmath436 .
thus , to complete the proof , it remains to establish the claim above .
consider any geodesic that lies wholly within the region @xmath438 .
parametrize the geodesic with arc length , such that the value of @xmath138 is maximal at @xmath439 say , equal to @xmath440 .
we recall the geodesic equations for @xmath441 where these are the @xmath295-cotangent variables as described in ( * ? ? ?
* section 2 ) : @xmath442 we also recall that @xmath443 along the geodesic , where @xmath444 .
we see that @xmath445 thus , we have @xmath446 for some @xmath447 slightly less than @xmath448 , which can be taken as close as desired to @xmath448 by choosing @xmath449 sufficiently small .
the initial condition @xmath450 arises as @xmath451 at @xmath439 .
we can integrate the differential inequality to obtain @xmath452 which yields @xmath453 plugging this into the equation for @xmath138 , we find that @xmath454 integrating this , we find that @xmath455 and with the help of the substitution @xmath456 , we obtain @xmath457 finally we turn to the equation for @xmath139 .
we have @xmath458 integrating the rhs from @xmath295 to @xmath459 at the value @xmath460 gives @xmath461 .
we get the same result for negative time , so that means that , along this geodesic , the maximum distance that @xmath139 can travel , with respect to the @xmath418 metric , is @xmath462 this is equal to @xmath463 when @xmath460 , and depends continuously on @xmath223 , hence is close to @xmath463 for @xmath223 close to @xmath448 , that is , when @xmath449 is sufficiently small . ] .
it follows that if @xmath434 , the geodesic between @xmath123 and @xmath427 must enter the region @xmath421 ( provided @xmath449 is sufficiently small ) .
this completes the proof of the proposition .
pointwise bounds on the spectral measure , and restriction estimates , are readily deduced from the regularity statement for the low energy resolvent .
the regularity statement for the resolvent , together with stones s formula , implies that the schwartz kernel of the low energy spectral measure @xmath14 takes the form @xmath464 where @xmath465 is a @xmath127 function on @xmath168 depending holomorphically on @xmath39 for small @xmath39 .
here we use our assumption that the resolvent is holomorphic in a neighbourhood of @xmath15 , the bottom of the essential spectrum ; on the other hand , the nontrapping assumption is irrelevant here .
we write the rhs as @xmath466 which implies that the kernel is bounded pointwise by @xmath467 using proposition [ prop : dist ] we may write @xmath468 . then estimating the sine factor by @xmath469 , we obtain the low energy ( @xmath116 ) estimate in theorem [ thm : kernelbounds ] . we have just seen that the spectral measure for low energy , @xmath116 , is bounded pointwise by @xmath470 times @xmath471 .
thus , to prove the low energy restriction estimate , it suffices to show that an integral operator , say @xmath472 , with kernel bounded pointwise by @xmath471 maps @xmath23 to @xmath473 for all @xmath20 . to do this
, we break up the kernel @xmath472 into pieces .
let @xmath413 be a neighbourhood of the front face @xmath172 in @xmath168 .
we consider @xmath474 and @xmath475 separately .
first consider @xmath475 . in this region , we may take @xmath425 and @xmath426 .
this part of the kernel is therefore bounded by @xmath476 .
thus , it is easy to check that @xmath475 is in @xmath477 , for any @xmath478 .
it therefore maps @xmath23 to @xmath473 for all @xmath20 .
now consider the remainder of the kernel , @xmath474 .
we may further decompose the set @xmath413 into subsets @xmath479 , where on each @xmath479 , we have @xmath480 and @xmath481 for some @xmath482 ( where the distance is measured with respect to the metric @xmath418 on @xmath140 ) .
choose local coordinates @xmath137 on @xmath1 , centred at @xmath483 , covering the set @xmath484 , and use these local coordinates to define a map @xmath485 from @xmath486 to a neighbourhood @xmath487 of @xmath488 in hyperbolic space @xmath144 using the upper half - space model ( such that the map is the identity in the given coordinates ) . the map @xmath485 induces a diffeomorphism @xmath489 from @xmath490 to a subset of @xmath491 , the double space for @xmath144 , covering the set @xmath492 in this space .
clearly , this map identifies @xmath173 and @xmath174 on @xmath479 with corresponding boundary defining functions for the left face and right face on @xmath491 .
we now consider the kernel @xmath493 as an integral operator on @xmath491 .
this kernel is bounded by @xmath494 , where @xmath239 is the geodesic distance on @xmath144 , since @xmath495 is comparable to @xmath496 on @xmath491 .
therefore , using , is bounded from @xmath497 to @xmath498 for every @xmath20 .
it is clear that @xmath485 are bounded , invertible maps from @xmath499 to @xmath500 .
this shows that the kernel @xmath501 is bounded from @xmath499 to @xmath502 for all @xmath20 .
this completes the proof of theorem [ thm : restriction ] in the case of low energy , @xmath116 .
according to theorem [ thm : specmeas ] , the spectral measure is a lagrangian distribution associated to the lagrangian submanifold @xmath371 ( on @xmath503 ) and to the lagrangian submanifold @xmath504 ( on @xmath505 ) .
we first define the notion of microlocal support , which is a closed subset of @xmath360 giving the essential support ` in phase space ' , for such distributions .
it is a special case of the notion of semiclassical wavefront set , defined for example in ( * ? ? ?
* section 8.4 ) .
we consider a local oscillatory integral expression for @xmath506 , where @xmath507 is a lagrangian submanifold of @xmath360 .
this is given by a local expression @xmath508 where @xmath509 , with @xmath510 smooth , and we use @xmath511 for local coordinates on @xmath168 , as described explicitly in regions 15 in section [ sec : ahm ] .
this requires that @xmath512 locally parametrizes @xmath507 ( nondegenerately ) , i.e. the map @xmath513 from @xmath514 , @xmath515 to @xmath507 , given by @xmath516 is a local diffeomorphism .
the microlocal support @xmath517 of is then contained in @xmath507 ( in general it can be any ) , and is determined by the support of the amplitude @xmath510 : @xmath518 it depends only on @xmath405 , not the particular form of .
we also recall that the schwartz kernel of a semiclassical @xmath295-pseudodifferential operator of order @xmath519 ( the first index is the semiclassical order , the second the differential order ) takes the form @xmath520 ( where @xmath521 is a symbol of order @xmath315 in @xmath522 ) near the diagonal and away from the boundary of @xmath168 , and @xmath523 ( where @xmath510 is a symbol of order @xmath315 in @xmath524 ) near the boundary of the diagonal in @xmath168 ; away from the diagonal , the kernel is smooth and @xmath525 .
we wish to show that by composing with pseudodifferential operators acting on @xmath1 , we can localize the microlocal support of @xmath506 .
more precisely , we shall establish [ prop : microsupport ] suppose that @xmath507 is a smooth lagrangian submanifold in @xmath360 , and let @xmath526 and @xmath527 .
then @xmath528 and we have @xmath529 here @xmath530 , @xmath531 is the left , resp .
right projection from @xmath532 , that is , the composite map @xmath533 where the first map is induced by the blow - down map @xmath362 , and the second is the left , resp .
right projection .
the second statement in follows from the first by switching the left and right variables .
so we only prove the first .
to do this , we write down local parametrizations of @xmath413 , and check the statement on each .
we use local coordinates valid in regions 15 as described in section [ sec : ahm ] .
@xmath534 region 1 . in this region
, @xmath413 has a local representation @xmath535 and @xmath536 has a representation .
the composition is given by an oscillatory integral @xmath537 we perform stationary phase in the variables @xmath538 .
we note that the hessian in these variables is non - degenerate , as the matrix of second derivatives takes the form @xmath539 which has nonzero determinant , irrespective of the top left entry .
the stationary phase expansion then shows that this expression can be simplified to @xmath540 where @xmath541 has an expansion @xmath542 where @xmath180 is a differential operator in the @xmath538 variables of degree @xmath543 .
this shows that @xmath528 and has microlocal support contained in @xmath544 ( since the amplitude @xmath541 is @xmath545 wherever @xmath546 ) . the microlocal support is also contained in the set @xmath547 which is to say that the microlocal support is contained in @xmath548 .
@xmath534 region 2a . in this region
, @xmath413 has a local representation @xmath549 and @xmath536 has a representation .
the composition is given by an oscillatory integral @xmath550 we change to coordinates @xmath551 and @xmath552 . in these coordinates
we have @xmath553 we then perform stationary phase in the variables @xmath554 .
there is a stationary point at @xmath555 we check that the hessian in these variables is non - degenerate at this critical point .
the matrix of second derivatives takes the form @xmath556 which has nonzero determinant when @xmath138 is small , irrespective of the starred entries .
the stationary phase expansion then shows that this expression can be simplified to @xmath557 where @xmath541 has an expansion @xmath558 where @xmath180 is a differential operator in @xmath559 of degree @xmath543 .
this shows that @xmath528 , and has microlocal support contained in @xmath544 ( since @xmath560 wherever @xmath546 ) . the microlocal support is also contained in @xmath561 which ( comparing with ) shows that the microlocal support of @xmath562 is also contained in @xmath563 .
@xmath534 region 2b . in this region , the calculation is similar to region 1 , so we omit the details .
@xmath534 region 3 . in this region , the calculation is similar to region 2a , so again we omit the details .
@xmath534 region 4a .
here we use the coordinates @xmath334 in this region , @xmath413 has a local representation @xmath564 and @xmath536 has a representation .
the composition is given by an oscillatory integral @xmath565 we introduce coordinates @xmath552 , @xmath566 .
the integral becomes @xmath567 we perform stationary phase in the variables @xmath568 .
it is straightforward to check that the hessian in these variables is nondegenerate at the stationary point @xmath569 we then get a stationary phase expansion , as in the previous regions , leading to the conclusion that @xmath562 has an expression @xmath570 such that @xmath541 is given in terms of @xmath571 and @xmath521 by a stationary phase expansion as in regions 1 or 2a above .
thus , @xmath562 is a lagrangian distribution in @xmath572 , and @xmath541 is @xmath545 wherever @xmath546 , and is supported where @xmath573 .
it follows ( using ) that @xmath574 is contained in @xmath575 .
@xmath534 region 4b .
this is given by a rather similar calculation to region 4a , which we omit .
@xmath534 region 5 .
here we use the coordinates @xmath576 in this region , @xmath413 has a local representation
@xmath577 writing @xmath578 and @xmath552 as before , the composition is given by an oscillatory integral @xmath579 we perform stationary phase in the variables @xmath568 .
there is a critical point at @xmath580 it is straightforward to check that the hessian in these variables is nondegenerate .
we then get a stationary phase expansion , as in the previous regions , leading to the conclusion that @xmath562 has an expression @xmath581 such that @xmath541 is given in terms of @xmath571 and @xmath521 by a stationary phase expansion .
thus , @xmath528 , and using the same reasoning as above , its microlocal support is contained in @xmath548 .
in this section , we prove theorem [ thm : kernelbounds1 ] for high energies , @xmath118 , which immediately implies also theorem [ thm : kernelbounds ] .
our first task is to choose an appropriate partition of the identity operator .
this is done in exactly the same way as was done in @xcite in the asymptotically conic case .
before getting into the details we explain the advantage of using a partition of the identity .
it is to microlocalize the spectral measure ( taking advantage of the microlocal support estimate , proposition [ prop : microsupport ] ) so that only the lagrangian @xmath371 is relevant , while the other part , @xmath504 , disappears .
this is important in our pointwise estimate , as the lagrangian @xmath371 locally projects diffeomorphically to the base manifold except at where it meets @xmath367 , i.e. the projection @xmath582 , restricted to @xmath371 has maximal rank except at the intersection with @xmath367 , which leads to the most favourable @xmath583 estimates .
( the drop in rank at the diagonal leads to the different form of the estimates for small @xmath342 in theorem [ thm : kernelbounds ] . ) by contrast , we can not control the rank of the projection from @xmath504 to the base ( except by making additional geometric assumptions , such as nonpositive curvature of @xmath143 , which we do in sections [ sec : restriction ] and [ sec : sm ] )
. our operators @xmath90 will be semiclassical @xmath295-pseudodifferential operators of order @xmath488 , where the first index denotes the semiclassical order and the second , the differential order , and the semiclassical parameter is @xmath584 . in fact , all but @xmath585 will have differential order @xmath395 .
first of all , we will choose @xmath586 of order @xmath488 microlocally supported away from the characteristic variety of @xmath403 , say in the region @xmath587 \cup [ 5/4 , \infty ) \}$ ] , and microlocally equal to the identity in a smaller region , say @xmath588 \cup [ 3/2 , \infty ) \}$ ] .
in light of the disjointness of semiclassical wavefront sets , the term @xmath589 has empty microlocal support , and is therefore @xmath545 . taking into account the behaviour at the boundary , we find that @xmath590 ) + h^\infty ( \rho_l \rho_r)^{n/2 + i / h } c^\infty(x^2_0 \times [ 0 , h_0 ] ) .
\label{q0term}\ ] ] this clearly satisfies theorem [ thm : kernelbounds1 ] .
we next choose a cutoff function @xmath591 , equal to @xmath448 for @xmath592 and @xmath295 for @xmath593 .
we decompose the remainder @xmath594 into @xmath595 and @xmath596 , and further decompose these two pieces in the following way .
we divide the interval @xmath597 $ ] into a union of intervals @xmath598 with overlapping interiors , and with diameter @xmath599 .
we then decompose @xmath595 into operators @xmath600 such that each operator @xmath90 has wavefront set contained in @xmath601 .
next , we decompose @xmath596 .
the idea is still to decompose this operator into pieces , so that on each piece the microlocal support is small .
let @xmath602 be the sasaki distance on @xmath603 .
we break up @xmath604 into a finite number of operators @xmath605 , each of which is such that the microlocal support has diameter @xmath606 with respect to the sasaki distance on @xmath603 .
this is possible since the microlocal support of @xmath596 is compact in @xmath603 .
we choose @xmath607 , where @xmath513 is the injectivity radius of @xmath6 .
we now prove a key property about the microlocal support of @xmath608 , when @xmath449 , @xmath609 and @xmath610 are sufficiently small .
[ prop : qi ] suppose that @xmath449 , @xmath609 and @xmath610 are sufficiently small .
then , provided that @xmath611 and @xmath39 satisfy @xmath88 $ ] for sufficiently small @xmath612 , the microlocal support of @xmath608 , @xmath613 , is a subset of @xmath371 . in the composition @xmath608
, we view all operators as semiclassical operators with parameter @xmath584 .
to do this for @xmath614 , we need to scale the fibre variables in the symbol by a factor of @xmath615 .
this is of little consequence as @xmath615 is close to @xmath448 by assumption .
recall that @xmath371 consists of a neighbourhood @xmath616 of @xmath617 in @xmath507 , together with a neighbourhood @xmath618 of @xmath619 in @xmath507 .
first suppose that @xmath620 . by proposition [ prop : microsupport ] ,
the microlocal support of @xmath621 is empty for sufficiently small @xmath612 , so the conclusion of proposition [ prop : qi ] trivially holds .
next suppose that @xmath622 .
we claim that if @xmath449 , @xmath609 and @xmath612 are sufficiently small , then the microsupport of @xmath608 is contained in @xmath616 .
let @xmath623 , i.e. a deleted neighbourhood of @xmath617 . since the microlocal support is always a closed set it suffices to show that @xmath624 by proposition [ prop : microsupport ] , this wavefront set is contained in @xmath625 we prove the claim by contradiction .
suppose that the claim were false .
choose sequences @xmath626 , @xmath627 and @xmath628 tending to zero as @xmath629 , and for each @xmath315 , a partition of the identity @xmath630 satisfying the conditions above relative to @xmath626 , @xmath627 and @xmath628 .
then , if the claim is false for all @xmath630 , there are sequences of pairs of points @xmath631 , @xmath632 in @xmath633 , lying on the same geodesic @xmath634 , with @xmath635 , @xmath636 , but the corresponding point @xmath637 not in @xmath638 . by compactness
we can take a convergent subsequence , with @xmath639 , @xmath640 , @xmath641 , @xmath642 , @xmath643 , @xmath644 , @xmath645 .
consider the limiting behaviour of the geodesic @xmath634 connecting @xmath631 and @xmath646 .
if @xmath647 then @xmath634 converges to a boundary bicharacteristic , that is , an integral curve of contained in @xmath321 , and therefore takes the form @xmath648 ( it is straightforward to check that this satisfies the geodesic equations in the parameter @xmath649 . ) therefore , @xmath650 and @xmath651 for some @xmath652 and @xmath653 . since @xmath654
, we have @xmath655 , and it then follows that @xmath656 .
this shows that the limiting point lies on @xmath657 , in fact over the fibre @xmath658 of @xmath172 lying over @xmath659 ( over which point on this fibre depends on the limiting values of @xmath660 and @xmath661 ) .
hence the sequence converging to it eventually lies in @xmath638 , which is our desired contradiction .
if @xmath662 then the limiting geodesic could be an interior bicharacteristic . in this case
we must have @xmath663 , i.e the points @xmath664 and @xmath665 are both an endpoint of this bicharacteristic .
however the condition that the difference @xmath666 along the sequence means that either both @xmath667 are @xmath668 or both are @xmath148 .
thus , the limiting points @xmath664 and @xmath665 are again equal in this case , with @xmath669 , @xmath670 , which shows that the limiting point lies on @xmath657 , hence the sequence converging to it eventually lies in @xmath638 , again producing a contradiction .
we next claim that if @xmath671 , then for @xmath610 sufficiently small , the wavefront set of @xmath672 is contained in @xmath618 .
the argument is similar .
choose a sequence @xmath673 tending to zero as @xmath674 , and for each @xmath315 , a partition of the identity @xmath630 satisfying the conditions above relative to @xmath673 .
then , if the claim is false for all @xmath630 , then we could find a sequence @xmath675 on the same geodesic , with @xmath676 , with each @xmath677 not in @xmath618 .
using compactness we can extract a convergent subsequence from the @xmath678 , converging to @xmath679 .
since @xmath680 the sequence @xmath681 also converges to @xmath682 .
but the point @xmath683 is in @xmath367 , and @xmath618 is a neighbourhood of @xmath684 in @xmath507 , so this gives us the contradiction .
we now assume that @xmath685 have been chosen small enough that the conclusion of proposition [ prop : qi ] is valid . in this section
we show that an element @xmath413 of @xmath686 satisfies and in the region @xmath191 . we divide this into the case where we work away from the boundary of @xmath340 , and near a point on the boundary of the diagonal
the first case , localizing away from the boundary of the diagonal , has been treated in ( * ? ? ?
* proposition 1.3 ) ( this was done in the context of asymptotically conic manifolds , but away from the boundary , one ` can not tell ' whether one is on an asymptotically conic or asymptotically hyperbolic manifold , so the argument applies directly ) . in @xcite , instead of @xmath7 , as here , when comparing ( * ? ? ?
* equation ( 1.13 ) ) with . ]
thus , it remains to deal with the case where @xmath14 is microlocalized to a neighbourhood of a point @xmath687 . in this case
, any parametrization of the lagrangian @xmath371 must have at least @xmath158 integrated variables , since the rank of the projection from @xmath371 to the base @xmath168 drops by @xmath158 at @xmath367 .
the following result is essentially taken from @xcite .
[ prop : param ] let @xmath688 be a point in @xmath689 , and let @xmath690 . then microlocally near @xmath688 , @xmath413 can be represented as an oscillatory integral of the form @xmath691 where the coordinates @xmath692 , @xmath693 define the boundary , that is , @xmath694 and the differentials @xmath695 and @xmath696 are linearly independent . here also , @xmath697 , and @xmath510 is smooth and compactly supported in all variables .
moreover , we may assume that @xmath698 has the properties @xmath699 in this region , @xmath700 furnish local coordinates on @xmath168 .
in these coordinates , the diagonal is defined by @xmath701 .
we will write @xmath692 for a suitable rotation of the coordinates @xmath702 .
let @xmath703 be the dual coordinates to @xmath704 .
we claim that , for some rotation @xmath692 of the @xmath702 coordinates , we have @xmath705 at @xmath688 .
this follows from the fact that @xmath706 projects to @xmath340 with @xmath158-dimensional fibres ( in fact @xmath706 is an @xmath707-bundle over @xmath340 ) , so the @xmath7 differentials @xmath708 span an @xmath158-dimensional space in @xmath709 for each @xmath710 .
secondly , we claim @xmath711 at @xmath688 . to see this
, we observe that there must be a vector @xmath712 tangent to @xmath507 but not tangent to @xmath367 .
therefore , @xmath301 must have a non - zero @xmath713 or @xmath714 component .
also , the vectors @xmath715 are tangent to @xmath507 at @xmath688 , since @xmath709 is the codimension 1 subspace of @xmath716 given by the vectors annihilated by @xmath717 .
thus , as @xmath507 is lagrangian , we have @xmath718 this implies that @xmath719 for each @xmath720 , which implies @xmath301 has a non - zero @xmath714 component , as claimed .
this shows that @xmath721 furnish coordinates on @xmath507 locally near @xmath688 .
thus , we can express the remaining coordinates , restricted to @xmath507 , as smooth functions of these : @xmath722 also , using the fact that @xmath507 is lagrangian , the form @xmath723 it follows that there is a function @xmath724 on @xmath507 , defined near @xmath688 , such that @xmath725 notice that @xmath726 , and at @xmath367 we have @xmath727 .
therefore , at @xmath728 , we have @xmath729 it follow that @xmath228 is constant when @xmath730 .
since @xmath228 is undetermined up to a constant , we may assume that @xmath731 when @xmath730 ; that is , @xmath732 .
we claim that the function @xmath733 locally parametrizes the lagrangian @xmath507 near @xmath688 , and satisfies properties ( a ) ( d ) above . to check that @xmath698 parametrizes @xmath507 , we set @xmath734 .
this implies that @xmath735 on the other hand , the 1-form identity shows that the functions @xmath736 and @xmath228 satisfy the identities @xmath737 the last of these identities shows that the rhs of vanishes .
we therefore find that the lagrangian parametrized by @xmath698 is @xmath738 it follows that , microlocally near @xmath688 , the spectral measure may be written as an oscillatory integral with phase function @xmath698 , as in , where the power of @xmath739 is given by @xmath740 where @xmath741 is the order of the lagrangian distribution , @xmath742 is the spatial dimension and @xmath743 is the number of integrated variables .
conditions ( a ) and ( b ) are easily verified , using the fact that @xmath133 and @xmath228 are @xmath744 . to check condition ( c ) ,
we write @xmath745 where @xmath536 is an @xmath746 matrix function of @xmath747 . we claim that @xmath536 is invertible at @xmath688 .
it suffices to check @xmath748 near @xmath688 .
on one hand , since @xmath698 is a phase function parametrizing @xmath507 nondegenerately in a neighbourhood of @xmath688 , then we have a local diffeomorphism @xmath749 the determinant of the differential of the map @xmath750 is thus equal to the determinant of the differential of the map @xmath751 which is simply @xmath752 on the other hand , it is obvious that the determinant of the differential of map ( [ eqn : projection map variant ] ) equals the determinant of the projection map @xmath753 from @xmath507 to @xmath168 .
since @xmath754 near @xmath688 ( * ? ? ?
* proposition 15 ) , we get @xmath755 which implies @xmath536 is invertible .
to check ( d ) , we notice that if @xmath730 and @xmath756 , then @xmath757 .
also , @xmath726 , so this is at the diagonal , i.e. @xmath758 , so this agrees with ( d ) . on the other hand , if @xmath759 , then condition ( c ) says that @xmath760 is nondegenerate . in this case
we can eliminate the extra @xmath761 variables , and reduce to a function of @xmath704 that parametrizes the lagrangian @xmath507 locally .
( this is an analytic reflection of the geometric fact that the projection @xmath762 has full rank restricted to @xmath507 , for @xmath759 . ) but then ( * ? ? ?
* proposition 16 ) implies that the value of @xmath698 , when @xmath734 , is equal to @xmath763 on the forward half of @xmath507 ( with respect to geodesic flow ) , and @xmath764 on the backward half of @xmath507 .
condition ( b ) implies that @xmath765 .
this completes the proof of proposition [ prop : param ] using this we now show [ prop : d<1 ] suppose @xmath690 .
then , @xmath413 can be written in the form with the amplitude functions @xmath766 satisfying in the region @xmath191 .
we estimate the integral by dividing into three cases , depending on the relative size of @xmath239 , @xmath767 and @xmath739
. * case 1 .
* @xmath768 . since @xmath728 at @xmath769 , and @xmath770 there ,
@xmath771 is comparable to @xmath407 .
so in this case , we have @xmath772 .
thus , we need to show ( comparing to ) @xmath773 for @xmath99 this is trivial .
consider @xmath774 .
we claim that this is also @xmath775 .
this differential operator is certainly harmless when applied to the amplitude , @xmath510 .
when applied to the exponential , it brings down a factor @xmath776 , which using ( b ) and @xmath777 , we write @xmath778 .
the term @xmath779 times the exponential @xmath780 is equal to @xmath781 .
we integrate by parts , shifting the @xmath782 derivative to the amplitude @xmath510 . in this way
we see that the result of applying @xmath783 to the expression is still @xmath775 .
a similar argument applies to repeated applications of @xmath783 .
thus this term takes the form .
* @xmath784 for some small constant @xmath541 . in this case
, there must be some @xmath785 such that @xmath786 .
then @xmath787 in a neighbourhood of @xmath688 , provided @xmath541 is sufficiently small .
we can then integrate by parts arbitrarily many times in @xmath788 , obtaining infinite order vanishing in @xmath739 .
the same is true for any number of @xmath783 derivatives applied to .
thus this term satisfies .
* case 3 . * @xmath789 and @xmath790 , with @xmath541 as in case 2 .
the idea for this region is to use a stationary phase estimate , as in ( * ? ? ?
* section 4 ) .
we follow this proof almost verbatim ; the changes required here are mostly notational . in this case
, we show a representation of the form , .
notice that if @xmath734 and @xmath789 then locally there are two sheets @xmath791 of @xmath507 above @xmath168 see proposition [ prop : flowout - structure ] .
on @xmath791 we divide by @xmath792 and show an estimate of the form .
the argument for each is the same , so we only describe the argument for @xmath793 .
thus , we define , with @xmath794 , @xmath795 and seek to prove the estimate @xmath796 since in case 3 , we have @xmath797 .
define @xmath798 and let @xmath799 .
notice that this function @xmath800 is @xmath127 in @xmath761 , and implies that all @xmath761-derivatives ( of all orders ) are uniformly bounded in the region @xmath789 and @xmath790 .
then the lhs of is @xmath801 therefore , we need to show that , for any @xmath609 , we have @xmath802 we now fix @xmath704 with @xmath803 .
we use a cutoff function @xmath804 to divide the @xmath761 integral into two parts : one on the support of @xmath804 , in which @xmath805 and the other on the support of @xmath806 , in which @xmath807 . on the support of @xmath804
, we integrate by parts in @xmath761 and gain any power of @xmath808 , proving . on the support of @xmath809
, we make a change of variable to @xmath810 coordinates : @xmath811 by property ( c ) of proposition [ prop : param ] , @xmath812 where @xmath813 is nondegenerate .
this shows that this change of variables is locally nonsingular , provided @xmath814 is sufficiently small .
thus , for each point @xmath761 in the support of @xmath806 , there is a neighbourhood in which we can make this change of variables . using the compactness of the support of @xmath510 in
, there are a finite number of neighbourhoods covering @xmath815 and the @xmath761-support of @xmath510 . on each such neighbourhood @xmath413
, we define @xmath816 . choose a @xmath127 function @xmath817 which is equal to @xmath448 when on the set @xmath818 and @xmath295 outside @xmath819 , and with derivatives bounded by @xmath820 here @xmath612 is a parameter which we will eventually choose to be @xmath821 ; however , for now we leave its value free . consider the integral after changing variables and with the cutoff function @xmath817 inserted ( where we stipulate @xmath822 , which means that that @xmath823 on @xmath824 ) : @xmath825 using property ( d ) of proposition [ prop : param ] , we see that @xmath826 when @xmath827 .
also , from , we have @xmath828 when @xmath829 .
hence @xmath830 .
hence @xmath831 it remains to treat the integral with the cutoff @xmath832 .
notice that @xmath833 is comparable to @xmath834 since @xmath828 when @xmath827 , and @xmath835 is nondegenerate when @xmath827 .
we define the differential operator @xmath26 by @xmath836 then the adjoint operator @xmath837 is given by @xmath838 we have chosen @xmath26 such that @xmath839 .
so we introduce @xmath840 factors of @xmath26 applied to the exponential @xmath841 and integrate by parts @xmath840 times to obtain @xmath842 inductively we find , using @xmath843 , that @xmath844 we choose @xmath845 to balance the two estimates and .
we finally obtain @xmath846 which proves as desired .
we express the spectral measure as a sum of 4 types of terms , ( i ) ( iv ) , as in theorem [ thm : specmeas ] .
then , using proposition [ prop : qi ] , we see that the microlocalized spectral measure @xmath847 has microsupport contained in @xmath371 , so the terms of type ( iii ) can be disregarded . clearly , terms of type ( ii ) and ( iv ) satisfy the conclusion of theorem [ thm : kernelbounds1 ] , so it is only necessary to consider the terms of type ( i )
. for terms of type ( i ) , proposition [ prop : d<1 ] shows that theorem [ thm : kernelbounds1 ] is satisfied . on the other hand , for large distance , we know from proposition [ prop : flowout - structure ] that @xmath848 and @xmath849 both project diffeomorphically to a open set @xmath850 under @xmath582 , and therefore , according to ( * ? ? ?
* proposition 20 ) , the lagrangian submanifolds @xmath851 are parametrized by the distance function , @xmath852 . the amplitude has a classical expansion in powers of @xmath584 , as shown by the construction of ( * ? ? ?
* section 4 ) , and the leading power of @xmath739 is @xmath853 , as claimed .
[ changed spectral measure estimates]let @xmath612 be a small positive number . for high energies , @xmath118 ,
one has @xmath854 provided @xmath611 is in @xmath855 $ ] .
we now prove the restriction theorem , theorem [ thm : restriction ] , at high energies , @xmath118 .
we apply complex interpolation to the analytic ( in the parameter @xmath856 ) family of operators @xmath857 here @xmath512 is a smooth function supported on @xmath858 $ ] and equal to @xmath448 on @xmath859 $ ] , whilst @xmath860 is an entire family of distributions , defined for @xmath861 by @xmath862 when @xmath863 , we have @xmath864 and using this identity , we extend @xmath860 to the entire complex @xmath510-plane .
since @xmath865 , we have @xmath866 , and more generally @xmath867 .
therefore , @xmath868\big ) \quad \mbox{and } \quad \chi_+^{-k}({{\boldsymbol\lambda}}- p ) = \big(\frac{d}{d{{\boldsymbol\lambda}}}\big)^{k - 1 } de_{p}({{\boldsymbol\lambda } } ) .
\label{chi - k}\ ] ] moreover , for any @xmath869 it is shown in @xcite that @xmath870 using this identity , and the fact that the @xmath39-derivatives of the spectral measure are well - defined and obey kernel estimates as in theorem [ thm : kernelbounds ] , we define , following @xcite , operators @xmath260 . for @xmath871 and @xmath872 , we define @xmath873 a standard application of stein s complex interpolation theorem @xcite yields [ prop : ci ] suppose that , for @xmath874 , we have @xmath875 and for some @xmath876 , @xmath877 then , the spectral measure @xmath878 is bounded from @xmath879 , @xmath880 , with an operator norm bound @xmath881 therefore , to prove theorem [ thm : restriction ] , for @xmath118 , we need to establish the estimates @xmath882 and for @xmath883 $ ] , we require @xmath884 while for @xmath885 , we require @xmath886 estimate follows immediately from the sup bound on the multiplier @xmath887 : @xmath888 for the remaining two estimates , we invoke ( * ? ? ? * lemma 3.3 ) , which we repeat here : [ infty ] suppose that @xmath889 , that @xmath890 and that @xmath891 . then there exists a constant @xmath892 such that for any @xmath893 function @xmath894 with compact support , one has @xmath895 for all @xmath896 . before proving and ,
we first rewrite @xmath897 as a convolution .
@xmath898 to prove , we rewrite the function @xmath899 as @xmath900 then we microlocalize the operator @xmath901 and apply lemma [ infty ] . in the following calculations ,
the operators ( that is , their schwartz kernels ) are evaluated at the point @xmath902 , which we do not always indicate in notation .
@xmath903 @xmath904 we now plug in corollary [ changed spectral measure estimates ] and get @xmath905 provided @xmath407 is small . on the other hand ,
if @xmath407 is large , @xmath906 when @xmath235 , the proof is identical , apart from using the other expression of the operator.@xmath907 we use the same argument and get @xmath908 to prove , we rewrite the function @xmath909 as @xmath910 then we again apply lemma [ infty ] to the microlocalized operator @xmath911 and get@xmath912 as in the first case , the spectral measure estimates give @xmath913 provided @xmath407 is small . on the other hand , if @xmath407 is large , @xmath914 notice that it is here that we gain an advantage by working on an asymptotically hyperbolic rather than conic space : the exponential decay @xmath915 in the large distance estimate kills the polynomial growth @xmath916 caused by @xmath720 differentiations of the phase function @xmath917 , so there is no limit to the number of differentiations that we can consider . on an asymptotically conic manifold , however , if we differentiate more than @xmath918 times , we get a growing kernel as @xmath919 , and no @xmath920 estimate is possible .
in this section we prove theorem [ thm : sm ] , assuming that @xmath6 is a cartan - hadamard manifold , as well as being asymptotically hyperbolic and nontrapping , with no resonance at the bottom of the continuous spectrum . in order to adapt the proof from section [ sec : model ] ,
we need to establish comparability between the riemannian measure on hyperbolic space , and the riemannian measure on @xmath6 , as expressed in polar coordinates . recall that on a cartan - hadamard manifold , the exponential map from @xmath921 , @xmath922 to @xmath1 is a diffeomorphism from @xmath923 to @xmath1 .
thus , the metric on @xmath1 can be expressed globally in polar normal coordinates based at @xmath123 .
let @xmath239 be the distance , and @xmath924 , be polar normal coordinates based at @xmath123 .
[ lem : measure ] let @xmath1 be an asymptotically hyperbolic cartan - hadamard manifold , and let @xmath922 be any point .
the riemannian measure on @xmath1 can be expressed in the form @xmath925 where @xmath926 is uniformly bounded on @xmath927 ( that is , uniform in @xmath123 as well as in @xmath928 ) .
this result can be extracted from the resolvent construction in @xcite .
recall that in that paper , the outgoing resolvent @xmath929 was shown to be a sum of terms , the principal one of which is a semiclassical intersecting lagrangian distribution @xmath930 . here
@xmath931 is the closure of the forward bicharacteristic relation , in a certain sense . in the case of a cartan - hadamard manifold
, the projection @xmath363 restricts to a diffeomorphism from @xmath932 to @xmath933 ; that is , except over the diagonal , @xmath793 projects diffeomorphically to the base @xmath168 .
we also point out that there is no need to decompose @xmath793 into pieces @xmath934 as was done in @xcite to deal with geodesics that might ` return ' to the front face @xmath172 ; this is not possible for cartan - hadamard manifolds .
the lagrangian @xmath793 can be given coordinates as follows : first , we use coordinates @xmath935 for @xmath936 , where @xmath937 is a coordinate in @xmath143 ( corresponding to the right variable in @xmath168 ) and @xmath924 is a coordinate on the unit tangent bundle in @xmath938 , with respect to the metric @xmath136 . then by definition @xmath793 is the flowout from @xmath936 by bicharacteristic flow , which coincides with geodesic flow in this case
let @xmath239 denote the function on @xmath793 equal to the time taken to flow to that point from @xmath936 by the left geodesic flow .
this gives us @xmath939 as coordinates on @xmath793 .
then , using the projection @xmath582 to the base , @xmath928 may be identified with polar normal coordinates based at @xmath937 . now consider the principal symbol at @xmath793 .
by @xcite , if we use coordinates @xmath902 arising from @xmath927 on @xmath793 ( away from @xmath367 ) , then the principal symbol is @xmath940 times @xmath941 , where @xmath942 ( @xmath943 ) indicate the riemannian measure in the left ( right ) variables , and we use the notation @xmath944 to mean that @xmath945 for some uniform @xmath892 .
next we recall from proposition [ prop : dist ] that the distance @xmath239 on @xmath168 is such that @xmath946 for @xmath270 .
it follows that the principal symbol is comparable to @xmath947 on @xmath793 , for @xmath270 . on the other hand ,
the principal symbol @xmath510 satisfies the transport equation @xmath948 in the coordinates @xmath939 .
since @xmath510 is a half - density , it must take the form @xmath949 we can compute @xmath950 by comparing with the symbol of the resolvent at @xmath357 . using coordinates @xmath951 , where @xmath652 is the norm on @xmath952 with respect to the metric @xmath136 ( that is , @xmath953 are polar coordinates in @xmath952 ) , this symbol is @xmath954 .
a simple calculation shows that @xmath955 . using ( * ? ? ?
* equation ( b.2 ) ) , we find that @xmath956 comparing and , we find that @xmath957 which completes the proof .
so , let @xmath958)$ ] , @xmath226 , be an even function .
we consider the operator @xmath221 , where @xmath959 $ ] . to analyze this operator
, we break the schwartz kernel into two pieces using the characteristic function @xmath960 .
the near - diagonal piece @xmath961 can be treated using the methods from @xcite ; this operator essentially satisfies theorem [ restriction2multiplier ] .
the far - from - diagonal piece , @xmath962 , can be treated rather like the case of hyperbolic space studied in section [ sec : model ] .
theorem [ restriction2multiplier ] does not apply directly in the current setting , since the volume of balls of radius @xmath81 are not comparable to @xmath963 for large @xmath81 on asymptotically hyperbolic manifolds ; instead , the volume grows as @xmath964 as @xmath965 .
however , it is certainly the case that the volume of balls of radius @xmath966 is comparable to @xmath963 .
this follows from the bishop - gromov inequality : if the sectional curvatures are between @xmath295 and @xmath967 , say , then the volume of any ball of radius @xmath81 is bounded by the volume in euclidean space , and the volume on a simply connected space of constant curvature @xmath967 .
[ lem : doubling ] suppose that @xmath95 is a metric measure space , with metric @xmath22 and doubling measure @xmath968 , such that the balls of radius @xmath966 have measure comparable to @xmath963 .
assume that @xmath969 is an integral operator , bounded from @xmath23 to @xmath970 for some @xmath971 .
let @xmath972 , be the integral operator given by the integral kernel of @xmath969 times the characteristic function of @xmath973 , for some @xmath974 .
then @xmath975 let @xmath95 be as in lemma [ lem : doubling ] .
suppose @xmath0 is a positive self - adjoint operator with finite propagation speed on @xmath107 .
if the restriction estimate @xmath976 holds for @xmath977 , then spectral multipliers localized near the diagonal are uniformly bounded in @xmath224 , in the sense @xmath978 where @xmath111 is an even function with @xmath979 supported in @xmath112 $ ] .
we follow the proof of ( * ? ? ? * section 2 ) .
suppose @xmath610 is an even smooth function compactly supported on @xmath980 , satisfying @xmath981 thus we take a partition of unity for @xmath982 , say @xmath983 , where @xmath984 by virtue of finite speed of propagation of @xmath985 @xcite , i.e. @xmath986 the kernel of @xmath987 is supported on @xmath988 as @xmath989 is supported on @xmath990 .
using restriction estimates as above , we have @xmath1002 where we used the fact that @xmath1003 on the support of @xmath1004 .
note that @xmath1005 and @xmath1006 where @xmath1007 for @xmath1008 and @xmath1009 . using this identity in and integrating by parts in @xmath649 yields @xmath1010 for any @xmath1011 .
taking @xmath840 sufficiently large , we obtain @xmath1012 combining and yields @xmath1013 and together with this proves the proposition .
it remains to treat the kernel @xmath1014 .
we will show that @xmath1014 maps @xmath222 to @xmath47 , with an operator norm uniform in @xmath959 $ ] .
it suffices to show that @xmath1014 maps @xmath285 , and @xmath286 , with operator norms uniform in @xmath223 .
the first statement follows from the fact that @xmath1015 , together with the result of section [ subsec : dist<1 ] .
in fact , we have proved @xmath1016 provided @xmath977 . noting the spectral multiplier is symmetric , we conclude that this operator is @xmath53 bounded for @xmath131 .
consequently , @xmath1017 is @xmath47 bounded .
so in the remainder of this subsection , we show boundedness from @xmath1018 to @xmath47 , with an operator norm uniform in @xmath223 . using , we expand the kernel of the spectral measure as follows : @xmath1025 } { { \boldsymbol\lambda}}^{n/2-j } b_{\pm , j}(z ' , r , \omega ) e^{-nr/2 } + c({{\boldsymbol\lambda } } , z ' , r , \omega ) e^{-nr/2 } \big ) \\
+ ( \rho_l \rho_r)^{n/2+i{{\boldsymbol\lambda } } } \ , a_+ + ( \rho_l \rho_r)^{n/2-i{{\boldsymbol\lambda } } } \ , a_- + ( xx')^{n/2+i{{\boldsymbol\lambda } } } \ , \tilde a_+ + ( xx')^{n/2-i{{\boldsymbol\lambda } } } \ , \tilde a_- \end{gathered } \label{sm - exp}\ ] ] where @xmath1026 and @xmath541 are bounded , and where @xmath1027 are as in theorem [ thm : kernelbounds1 ] . here
, @xmath541 is smooth in @xmath39 at @xmath1028 ( due to our assumption that the resolvent kernel is holomorphic at the bottom of the spectrum ) , and decays as @xmath1029 as @xmath167 for @xmath158 odd , or @xmath1030 as @xmath167 for @xmath158 even .
moreover , @xmath541 obeys symbolic estimates as @xmath167 , so @xmath1031 as @xmath167 when @xmath158 is odd , or @xmath1032 when @xmath158 is even .
we now consider a single term @xmath1033 in .
thus , we need to estimate @xmath1034 uniformly in @xmath223 and @xmath937 . arbitrarily choosing the sign @xmath13 , using the uniform boundedness of @xmath1035 , and simplifying , it is enough
to uniformly bound @xmath1036 we first observe that @xmath1043 .
indeed , since the function @xmath1044 is homogeneous of degree @xmath75 , the fourier transform is homogeneous of degree @xmath1045 , and hence is @xmath1046 as @xmath1047 .
the fourier transform @xmath1048 is therefore this homogeneous function convolved with @xmath1049 .
as @xmath1050 , @xmath1048 is @xmath583 , and still decays as @xmath1046 as @xmath1051 .
it therefore suffices to show that @xmath1052 we break the rhs into @xmath1053 using the inequality @xmath1054 , we estimate this by @xmath1055 the first of these terms we treat as follows .
we apply cauchy - schwarz to the inner integral , obtaining @xmath1056 the @xmath1057 integral just gives a constant . in the second integral , we change variable to @xmath1058 , and note that @xmath1059 .
the @xmath1060 integral again gives a constant , and we get an upper bound of the form @xmath1061 we can insert a factor @xmath1062 , since @xmath1063 and @xmath1064 .
this finally gives an estimate of the form @xmath1065 for the first term of .
we return to , which we write in the form @xmath1068 we change variables to @xmath1069 and @xmath1070 .
we also choose @xmath1071 to be identically @xmath448 on the support of @xmath5 , and write @xmath1072 .
the integral becomes @xmath1073 the @xmath1074 integral gives us @xmath1075 .
applying lemma [ lem : fg ] with @xmath1076 and @xmath1077 , we see that is bounded uniformly in @xmath223 , as required .
we next consider the terms involving @xmath541 , @xmath186 , and @xmath1078 .
the argument for all these terms is similar , so just consider @xmath541 . in this case , we need a uniform bound on @xmath1079 we use the identity @xmath1080 we integrate by parts .
this gives us @xmath1081 when the derivative falls on @xmath5 , we get @xmath1082 . since @xmath1083 , with compact support , the function @xmath1082 is @xmath1018 , with @xmath1018 norm uniformly bounded in @xmath39 .
since @xmath541 is uniformly bounded , this gives us a uniform bound on the @xmath39 integral in .
when the derivative falls on @xmath541 , using the symbol estimates , we find that @xmath1084 is integrable in @xmath39 , and then we can use the fact that @xmath1085 to see that in this case also , the @xmath39 integral in is uniformly bounded .
finally , the @xmath239 integral is convergent , so that establishes the uniform bound on .
u. bunke and m. olbrich , _ the wave kernel for the laplacian on the classical locally symmetric spaces of rank one , theta functions , trace formulas and the selberg zeta function , with an appendix by andreas juhl _ , ann .
global anal .
* 12 * ( 1994 ) , no .
4 , 357405 . j. cheeger , m. gromov and m. taylor , _ finite propagation speed , kernel estimates for functions of the laplace operator , and the geometry of complete riemannian manifolds _ , j. diff . geom . * 17*(1982 ) , 15 - 53 . | we consider the laplacian @xmath0 on an asymptotically hyperbolic manifold @xmath1 , as defined by mazzeo and melrose @xcite .
we give pointwise bounds on the schwartz kernel of the spectral measure for the operator @xmath2 on such manifolds , under the assumptions that @xmath1 is nontrapping and there is no resonance at the bottom of the spectrum .
this uses the construction of the resolvent given by mazzeo and melrose @xcite ( valid when the spectral parameter lies in a compact set ) , melrose , s barreto and vasy @xcite ( high energy estimates for a perturbation of the hyperbolic metric ) and the present authors @xcite ( see also @xcite ) in the general high - energy case .
we give two applications of the spectral measure estimates .
the first , following work due to guillarmou and sikora with the second author @xcite in the asymptotically conic case , is a restriction theorem , that is , a @xmath3 operator norm bound on the spectral measure .
the second is a spectral multiplier result under the additional assumption that @xmath1 has negative curvature everywhere , that is , a bound on functions @xmath4 of the square root of the laplacian , in terms of norms of the function @xmath5 .
compared to the asymptotically conic case , our spectral multiplier result is weaker , but the restriction estimate is stronger . in both cases ,
the difference can be traced to the exponential volume growth at infinity for asymptotically hyperbolic manifolds , as opposed to polynomial growth in the asymptotically conic setting .
the pointwise bounds on the spectral measure established here will also be applied to strichartz estimates in @xcite . |
following the hypothesis given in the abstract we select the following 4 p - wave nonets : @xmath5 .
they are displayed together with the pseudoscalar nonet in figure [ fig1 ] . as a consequence of the adopted selection criteria we exclude the following candidate states from the pdg listings @xcite : @xmath6 , @xmath7 , @xmath8 and resonances with a mass exceeding @xmath9 in particular @xmath10 and @xmath11 ,
see also @xcite .
we pose the following questions : - is the flavor mixing pattern of the @xmath12 and @xmath13 nonets similar , i.e. near singlet - octet ?
the two nonets would be parity doublets due to chiral symmetry , which is spontaneously and explicitely broken . - does the scalar gluonic meson distort through large mixing effects the scalar @xmath4 nonet beyond ( spectroscopic ) recognition ? in our analysis in @xcite we started from the hypothesis , that the answer to this question is no . inspecting the spectra of the p - wave nonets in fig .
1 we observe no strong distortion in the scalar sector indeed . - can the last question be resolved through direct observation of the scalar gluonic meson as a conventionally narrow resonance and how reliable are the mass estimates from purely gluonic lattice - qcd @xcite @xmath14 - can qcd sum rules including local gluonic operators @xcite shed light on the above mass estimate ?
the latter two questions are adressed in the accompanying paper @xcite . * the gell - mann - okubo square mass formula * according to our analysis @xmath15 ( 980 ) represents the su3 singlet ,
whereas @xmath16 ( 984.7 ) , @xmath17 ( 1412 ) , @xmath18 ( 1507 ) form the associated octet .
the gell - mann - okubo ( first order ) mass square relation then yields ( in @xmath19 units ) @xmath20 the deviation amounts to @xmath21 is numerically a refinement with respect to ref .
] there is no sign - yet - of any major distortion .
the degeneracy in mass of @xmath22 and @xmath16 , while not offending any basic principles , indicates further dynamic simplicity to be explained .
the aim is to study and eventually confirm the nonstrange versus strange @xmath4 flavor structure of the two singlet - octet assigned isoscalar scalars : @xmath23 the following phase convention shall be chosen in the flavor basis @xmath24 and in the ns - s basis we have @xmath25 so in the flavor basis we have @xmath26 as conjectured range of the singlet - octet angle @xmath27 we consider @xmath28 the corresponding range for @xmath29 becomes @xmath30 previously @xcite we have analysed the decays involving @xmath0 : @xmath31 , the radiative decays @xmath32 and @xmath33 and concluded on a large flavor mixing similar to @xmath34-@xmath35 in the pseudoscalar nonet .
now we extend our analysis to d decays .
our analysis of the mixing pattern shall focus on the two ratios of strange to nonstrange components @xmath36 if the premise of mainly singlet @xmath37 ( 980 ) is correct we infer the ranges @xmath38 and @xmath39 .
the relation @xmath40 follows from orthogonality .
the restricted ranges are @xmath41 we determine @xmath29 considering the three three body decays of the charmed mesons @xmath42 and @xmath43 @xcite : @xmath44 we consider the color favored amplitudes ( @xmath45 a ) which contribute to all processes in eq .
[ eq:10 ] , and color suppressed amplitudes ( @xmath45 @xmath46 ) obtained by @xmath47 ( see fig .
[ fig4 ] ) and obtain @xmath48 we consider the two ratios of partial decay widths @xmath49 and @xmath50 : @xmath51 in the approximation @xmath52 , @xmath53 with cabibbo angle @xmath54 , we find @xmath55 where @xmath56 denote the phase space in s - wave decays , proportional to the @xmath57 momentum in the decay resonance rest frame .
using the branching fractions established by the e791 collaboration and the pdg results we find numerically @xmath58 then we obtain as solutions two bands ( a and b ) due to the quadratic nature of the relations in eq .
[ eq:14 ] @xmath59 the angle @xmath29 is defined modulo @xmath60 .
the two solutions in eq .
[ eq:15 ] can be distinguished through the sign of the quantity @xmath61 ( or @xmath62 ) defined in eq .
[ eq:8 ] .
the phase ( + ) structure of @xmath63 , @xmath64 is determined from interference with other resonances in @xmath65 and @xmath66 decays into @xmath67 and @xmath68 @xcite .
the amplitudes a , b , c in eq .
[ eq:11 ] exhibit the + phase structure shown in table 1 .
@xmath69 [ tab : phases ]
it becomes clear from the results in table 1 and the assumed form of the amplitudes in eq .
[ eq:11 ] , that only the solution in band a ) in eq .
[ eq:15 ] is compatible with the data .
this implies @xmath70 confirming the near singlet quark flavor mixing of @xmath0 , the ideal singlet corresponds to @xmath71 .
the present analysis leaves the mixing with the scalar gluonic meson(s ) completely open .
here we refer to our present results on glueballs in ref .
future work will hopefully establish the full structure of the scalar nonet lowest in mass including the scalar glueball @xmath72 ) .
it is a pleasure to thank all organizers and participants of the qcd2002 conference , in particular stephan narison , for enabling the meeting to evolve in a lively and fruitful atmosphere .
aitala et al . , e791 collaboration , phys .
86 , ( 2001 ) 770 , hep - ex 0007028 , and phys .
86 , ( 2001 ) 765 , hep - ex 0007027 .
frabetti et al . , e687 collaboration , phys .
b407 ( 1997 ) 79 , and phys .
b351 ( 1995 ) 591 . | the hypothesis that there exists a nonet of scalars mainly composed of a valence quark - antiquark pair and mixed according to near singlet - octet separation : @xmath0 singlet , @xmath1 octet , is put to further tests from the three body decays @xmath2 with @xmath3 .
the analysis of decay phases supports the singlet nature of @xmath0 . |
consider the estimation of a random vector @xmath0 from the measurement model illustrated in figure [ fig : model ] .
the random vector @xmath2 , which is assumed to have independent and identically distributed ( i.i.d . )
components @xmath3 , is passed through a known linear transform that outputs @xmath4 .
the components of @xmath1 are generated by a componentwise transfer function @xmath5 .
this work addresses the cases where the distributions @xmath6 and @xmath5 have some unknown parameters , @xmath7 and @xmath8 , that must be learned in addition to the estimation of @xmath2 .
such joint estimation and learning problems with linear transforms and componentwise nonlinearities arise in a range of applications , including empirical bayesian approaches to inverse problems in signal processing , linear regression and classification @xcite , and , more recently , bayesian compressed sensing for estimation of sparse vectors @xmath2 from underdetermined measurements@xcite . also , since the parameters in the output transfer function @xmath5 can model unknown nonlinearities , this problem formulation can be applied to the identification of linear - nonlinear cascade models of dynamical systems , in particular for neural spike responses @xcite .
when the distributions @xmath6 and @xmath5 are known , or reasonably bounded , there are a number of methods available that can be used for the estimation of the unknown vector @xmath2 . in recent years
, there has been significant interest in so - called approximate message passing ( amp ) and related methods based on gaussian approximations of loopy belief propagation ( lbp ) @xcite .
these methods originate from cdma multiuser detection problems in @xcite , and have received considerable recent attention in the context of compressed sensing @xcite .
see , also the survey article @xcite .
the gaussian approximations used in amp are also closely related to standard expectation propagation techniques @xcite , but with additional simplifications that exploit the linear coupling between the variables @xmath2 and @xmath9 .
the key benefits of amp methods are their computational simplicity , large domain of application , and , for certain large random @xmath10 , their exact asymptotic performance characterizations with testable conditions for optimality @xcite .
this paper considers the so - called generalized amp ( gamp ) method of @xcite that extends the algorithm in @xcite to arbitrary output distributions @xmath5 ( many original formulations assumed additive white gaussian noise ( awgn ) measurements ) .
however , although the current formulation of amp and gamp methods is attractive conceptually , in practice , one often does not know the prior and noise distributions exactly . to overcome this limitation , vila and schniter @xcite and krzakala _ et al . _
@xcite have recently proposed extension of amp and gamp based on expectation maximization ( em ) that enable joint learning of the parameters @xmath11 along with the estimation of the vector @xmath2 . while simulations indicate excellent performance , the analysis of these methods is difficult .
this work provides a unifying analytic framework for such amp - based joint estimation and learning methods .
the main contributions of this paper are as follows : * generalization of the gamp method of @xcite to a class of algorithms we call _ adaptive gamp _ that enables joint estimation of the parameters @xmath7 and @xmath8 along with vector @xmath2
. the methods are computationally fast and general with potentially large domain of application .
in addition , the adaptive gamp methods include the em - gamp algorithms of @xcite as special cases .
* exact characterization of the asymptotic behavior of adaptive gamp .
we show that , similar to the analysis of the amp and gamp algorithms in @xcite , the componentwise asymptotic behavior of adaptive gamp can be described exactly by a simple scalar _ state evolution _ ( se ) equations .
* demonstration of asymptotic consistency of adaptive gamp with maximum - likelihood ( ml ) parameter estimation .
our main result shows that when the ml parameter estimation is computed exactly , the estimated parameters converge to the true values and the performance of adaptive gamp asymptotically coincides with the performance of the oracle gamp algorithm that knows the correct parameter values .
remarkably , this result applies to essentially arbitrary parameterizations of the unknown distributions @xmath6 and @xmath5 , thus enabling provably consistent estimation on non - convex and nonlinear problems .
* experimental evaluation of the algorithm for the problems of learning of sparse priors in compressed sensing and identification of linear - nonlinear cascade models in neural spiking processes .
our simulations illustrate the performance gain of adaptive gamp and its asymptotic consistency .
adaptive gamp thus provides a computationally - efficient method for a large class of joint estimation and learning problems with a simple , exact performance characterization and provable conditions for asymptotic consistency .
as mentioned above , the adaptive gamp method proposed here can be seen as a generalization of the em methods in @xcite . in @xcite , the prior @xmath6
is described by a generic @xmath12-term gaussian mixture ( gm ) whose parameters are identified by an em procedure @xcite . the expectation " or e - step is performed by gamp , which can approximately determine the marginal posterior distributions of the components @xmath13 given the observations @xmath14 and the current parameter estimates of the gm distribution @xmath6 .
a related em - gamp algorithm has also appeared in @xcite for the case of certain sparse priors and awgn outputs .
simulations in @xcite show remarkably good performance and computational speed for em - gamp over a wide class of distributions , particularly in the context of compressed sensing .
also , using arguments from statistical physics , @xcite presents state evolution ( se ) equations for the joint evolution of the parameters and vector estimates and confirms them numerically . as discussed in section [ sec : em - gamp ] , em - gamp is a special case of adaptive gamp with a particular choice of the adaptation functions . therefore , one contribution of this paper is to provide a rigorous theoretical justification of the em - gamp methodology . in particular , the current work provides a rigorous justification of the se analysis in @xcite along with extensions to more general input and output channels and adaptation methods .
however , the methodology in @xcite in other ways is more general in that it can also study seeded " or spatially - coupled " matrices as proposed in @xcite .
an interesting open question is whether the analysis methods in this paper can be extended to these scenarios as well .
an alternate method for joint learning and estimation has been presented in @xcite , which assumes that the distributions on the source and output channels are themselves described by graphical models with the parameters @xmath7 and @xmath8 appearing as unknown variables .
the method in @xcite , called hybrid - gamp , iteratively combines standard loopy bp with amp methods .
one avenue of future work is to see if methodology in this paper can be applied to analyze the hybrid - gamp methods as well .
finally , it should be pointed out that while simultaneous recovery of unknown parameters is appealing conceptually , it is not a strict requirement . an alternate solution to
the problem is to assume that the signal belongs to a known class of distributions and to minimize the maximal mean - squared error ( mse ) for the class .
this minimax approach @xcite was proposed for amp recovery of sparse signals in @xcite .
although minimax approach results in the estimators that are uniformly good over the entire class of distributions , there may be a significant gap between the mse achieved by the minimax approach and the oracle algorithm that knows the distribution exactly .
indeed , this gap was the main justification of the em - gamp methods in @xcite . due to its asymptotic consistency ,
adaptive gamp provably achieves the performance of the oracle algorithm .
the paper is organized as follows : in section [ sec : gampreview ] , we review the non - adaptive gamp and corresponding state evolution equations . in section [ sec :
gamp ] , we present adaptive gamp and describe ml parameter learning . in section [ sec : convprop ]
, we provide the main theorems characterizing asymptotic performance of adaptive gamp and demonstrating its consistency . in section [ sec : numex ] , we provide numerical experiments demonstrating the applications of the method . section [ sec : concl ] concludes the paper . a conference version of this paper has appeared in @xcite .
this paper contains all the proofs , more detailed descriptions and additional simulations .
before describing the adaptive gamp algorithm , it is useful to review the basic ( non - adaptive ) gamp algorithm of @xcite . consider the estimation problem in fig .
[ fig : model ] where the componentwise distributions on the inputs and outputs have some parametric form , [ eq : pxzlam ] p_x(x|_x ) , p_y|z(y|z , _ z ) , where @xmath15 and @xmath16 represent parameters of the distributions and @xmath17 and @xmath18 some parameter sets .
the gamp algorithm of @xcite can be seen as a class of methods for estimating the vectors @xmath2 and @xmath9 for the case when the parameters @xmath7 and @xmath8 are _ known_. in contrast , the adaptive gamp method that is discussed in section [ sec : gamp ] enables _ joint _ estimation of the parameters @xmath7 and @xmath8 along with the vectors @xmath2 and @xmath9 . in order that to understand how the adaptation works , it is best to describe the basic gamp algorithm as a special case of the more general adaptive gamp procedure .
the basic gamp algorithm corresponds to the special case of algorithm [ algo : gamp ] when the _ adaptation functions _ @xmath19 and @xmath20 output fixed values [ eq : hfix ] h^t_z(^t,,^t_p ) = ^t_z , h^t_x(^t,^t_r ) = ^t_x , for some _ pre - computed _ sequence of parameters @xmath21 and @xmath22 . by
pre - computed " , we mean that the values do not depend on the data through the vectors @xmath23 , @xmath24 , and @xmath25 . in the oracle scenario @xmath21 and @xmath22 are set to the true values of the parameters and do not change with the iteration number @xmath26 . the _ estimation functions _ @xmath27 , @xmath28 and @xmath29 determine the estimates for the vectors @xmath2 and @xmath9 , given the parameter values @xmath30 and @xmath31 .
as described in @xcite , there are two important sets of choices for the estimation functions , resulting in two variants of gamp : * _ sum - product gamp : _ in this case , the estimation functions are selected so that gamp provides a gaussian approximation of sum - product loopy bp .
the estimates @xmath32 and @xmath33 then represent approximations of the mmse estimates of the vectors @xmath2 and @xmath9 . *
_ max - sum gamp : _ in this case , the estimation functions are selected so that gamp provides a quadratic approximation of max - sum loopy bp and @xmath32 and @xmath33 represent approximations of the map estimates .
the estimation functions of the sum - product gamp are equivalent to scalar mmse estimation problems for the components of the vectors @xmath2 and @xmath9 observed in gaussian noise . for max - sum gamp , the estimation functions correspond to scalar map problems .
thus , for both versions , the gamp method reduces the vector - valued mmse and map estimation problems to a sequence of scalar awgn problems combined with linear transforms by @xmath10 and @xmath34 .
gamp is thus computationally simple , with each iteration involving only scalar nonlinear operations followed by linear transforms .
the operations are similar in form to separable and proximal minimization methods widely used for such problems @xcite .
appendix [ sec : gampdetails ] reviews the equations for the sum - product gamp .
more details , as well as the equations for max - sum gamp can be found in @xcite .
in addition to its computational simplicity and generality , a key motivation of the gamp algorithm is that its asymptotic behavior can be precisely characterized when @xmath10 is a large i.i.d .
gaussian transform .
the asymptotic behavior is described by what is known as a _ state evolution _ ( se ) analysis . by now
, there are a large number of se results for amp - related algorithms @xcite . here
, we review the particular se analysis from @xcite which is based on the framework in @xcite . [ as : gamp ] consider a sequence of random realizations of the gamp algorithm , indexed by the dimension @xmath35 , satisfying the following assumptions : * [ as : gauss ] for each @xmath35 , the matrix @xmath36 has i.i.d . components with @xmath37 and the dimension @xmath38 is a deterministic function of @xmath35 satisfying @xmath39 for some @xmath40 as @xmath41 . * [ as : px0 ] the input vectors @xmath2 and initial condition @xmath42 are deterministic sequences whose components converge empirically with bounded moments of order @xmath43 as [ eq : thetaxinit ] _
n ( , ^0 ) ( x,^0 ) , to some random vector @xmath44 for some @xmath45 . for the precise definition of this form of convergence .
* [ as : pyz ] the output vectors @xmath9 and @xmath1 are generated by [ eq : zax ] = , y_i = h(z_i , w_i ) i=1, ,m , for some scalar function @xmath46 and vector @xmath47 representing an output disturbance .
it is assumed that the output disturbance vector @xmath48 is deterministic , but empirically converges as [ eq : wlim ] _
n w , where @xmath43 is as in assumption [ as : gamp ] ( b ) and @xmath49 is some random variable .
we let @xmath5 denote the conditional distribution of the random variable @xmath50 . *
the estimation function @xmath51 and its derivative with respect to @xmath52 , is lipschitz continuous in @xmath52 at @xmath53 , where @xmath54 is a deterministic parameter from the se equations below .
a similar assumptions holds for @xmath55 .
* the adaptation functions @xmath19 and @xmath20 are set to for some deterministic sequence of parameters @xmath56 and @xmath57 .
also , in the estimation steps in lines [ line : zhat ] , [ line : shat ] and [ line : xhat ] of algorithm [ algo : gamp ] , the values of the @xmath58 and @xmath59 are replaced with the deterministic parameters @xmath60 and @xmath54 from the se equations defined below .
assumption [ algo : gamp](a ) simply states that we are considering large , gaussian i.i.d .
matrices @xmath10 . assumptions ( b ) and ( c ) state that the input vector @xmath2 and output disturbance @xmath48 are modeled as deterministic , but whose empirical distributions asymptotically appear as i.i.d .
this deterministic model is one of key features of bayati and montanari s analysis in @xcite .
assumption ( d ) is a mild continuity condition .
assumption ( e ) defines the restriction of adaptive gamp to the non - adaptive gamp algorithm .
we will remove this final assumption later .
note that , for now , there is no assumption that the true " distribution of @xmath61 or the true conditional distribution of @xmath62 given @xmath63 must belong to the class of distributions for any parameters @xmath7 and @xmath8 .
the analysis can thus model the effects of model mismatch .
now , given the above assumptions , define the sets of vectors [ eq : thetaxz ] _
x^t & : = & \{(x_j , r_j^t,_j^1 ) , j=1, ,n } , + _ z^t & : = & \{(z_i,_i^t , y_i , p_i^t ) , i=1, ,m } .
the first vector set , @xmath64 , represents the components of the the true , " but unknown , input vector @xmath2 , its gamp estimate @xmath32 as well as @xmath25 .
the second vector , @xmath65 , contains the components the true , " but unknown , output vector @xmath9 , its gamp estimate @xmath33 , as well as @xmath23 and the observed output @xmath14 .
the sets @xmath64 and @xmath65 are implicitly functions of the dimension @xmath35 .
the main result of @xcite shows that if we fix the iteration @xmath26 , and let @xmath66 , the asymptotic joint empirical distribution of the components of these two sets @xmath64 and @xmath65 converges to random vectors of the form [ eq : thetabarxz ] _ x^t : = ( x , r^t,^1 ) , _ z^t : = ( z,^t , y , p^t ) .
we precisely state the nature of convergence momentarily ( see theorem [ thm : stateevogamp ] ) . in
, @xmath61 is the random variable in assumption [ as : gamp](b ) , while @xmath67 and @xmath68 are given by [ eq : rvxt ] @xmath69 for some deterministic constants @xmath70 , @xmath71 , and @xmath54 that are defined below .
similarly , @xmath72 for some covariance matrix @xmath73 , and [ eq : yzpt ] y = h(z , w ) , ^t = g_z^t(p^t , y,_p^t,_z^t ) , where @xmath49 is the random variable in and @xmath73 contains deterministic constants .
the deterministic constants @xmath70 , @xmath71 , @xmath54 and @xmath73 represent parameters of the distributions of @xmath74 and @xmath75 and depend on both the distributions of the input and outputs as well as the choice of the estimation and adaptation functions .
the se equations provide a simple method for recursively computing these parameters .
the equations are best described algorithmically as shown in algorithm [ algo : se ] . in order that we do not repeat ourselves , in algorithm [ algo : se ] , we have written the se equations for adaptive gamp : for non - adaptive gamp , the updates and can be ignored as the values of @xmath57 and @xmath56 are pre - computed . with these definitions , we can state the main result from @xcite .
[ thm : stateevogamp ] consider the random vectors @xmath64 and @xmath65 generated by the outputs of gamp under assumption [ as : gamp ] .
let @xmath74 and @xmath75 be the random vectors in with the parameters determined by the se equations in algorithm [ algo : se ] .
then , for any fixed @xmath26 , the components of @xmath64 and @xmath65 converge empirically with bounded moments of order @xmath76 as [ eq : thetalimgamp ] _
n _ x^t _ x^t , _
z^t _ z^t . where @xmath74 and @xmath75 are given in .
in addition , for any @xmath26 , the limits [ eq : taulimgamp ] _ n
_ r^t = _ r^t , _
p^t = _ p^t , also hold almost surely .
the theorem shows that the behavior of any component of the vectors @xmath2 and @xmath9 and their gamp estimates @xmath32 and @xmath33 are distributed identically to a simple scalar equivalent system with random variables @xmath61 , @xmath63 , @xmath77 and @xmath78 .
this scalar equivalent model appears in several analyses and can be thought of as a _ single - letter characterization _
@xcite of the system . remarkably , this limiting property holds for essentially arbitrary distributions and estimation functions , even ones that arise from problems that are highly nonlinear or noncovex . from the single - letter characterization , one can compute the asymptotic value of essentially any componentwise performance metric such as mean - squared error or detection accuracy .
similar single - letter characterizations can also be derived by arguments from statistical physics @xcite .
[ algo : gamp ] [ line : taup ] [ line : phat ] [ line : lamz ] [ line : zhat ] [ line : shat ] [ line : taus ] [ line : taur ] [ line : rhat ] [ line : lamx ] [ line : xhat ] [ line : taux ] as described in the previous section , the standard gamp algorithm of @xcite considers the case when the parameters @xmath7 and @xmath8 in the distributions in are known .
the adaptive gamp method proposed in this paper , and shown in algorithm [ algo : gamp ] , is an extension of the standard gamp procedure that enables simultaneous identification of the parameters @xmath7 and @xmath8 along with estimation of the vectors @xmath2 and @xmath9 . the key modification is the introduction of the two _ adaptation functions _ : @xmath79 and @xmath80 . in each iteration , these functions output estimates , @xmath81 and @xmath82 of the parameters based on the data @xmath23 , @xmath14 , @xmath25 , @xmath58 and @xmath59 .
we saw the standard gamp method corresponds to the adaptation functions in which outputs fixed values @xmath57 and @xmath56 that do not depend on the data , and can be used when the true parameters are known . for the case when the true parameters are not known , we will see that a simple maximum likelihood ( ml ) can be used to estimate the parameters from the data . to understand how to estimate parameters via the adaptation functions ,
observe that from theorem [ thm : stateevogamp ] , we know that the distribution of the components of @xmath25 are distributed identically to the scalar @xmath67 in .
now , the distribution of @xmath67 only depends on three parameters @xmath70 , @xmath71 and @xmath7 .
it is thus natural to attempt to estimate these parameters from the empirical distribution of the components of @xmath25 . to this end , let @xmath83 be the log likelihood [ eq : phixll ] _
x(r,_x,_r,_r ) = p_r(r|_x,_r,_r ) , where the right - hand side is the probability density of a random variable @xmath84 with distribution @xmath85 then , at any iteration @xmath26 , we can attempt to perform a maximum - likelihood ( ml ) estimate + & & = _ _ x _ x _ ( _ r,_r ) s_x(_r^t ) \ { _ j=1^n _ x(r^t_j,_x,_r,_r)}. [ eq : hxml ] here , the set @xmath86 is a set of possible values for the parameters @xmath87 .
the set may depend on the measured variance @xmath59 .
we will see the precise role of this set below .
similarly , the joint distribution of the components of @xmath23 and @xmath14 are distributed according to the scalar @xmath88 which depend only on the parameters @xmath89 and @xmath8 .
thus , we can define the likelihood @xmath90 where the right - hand side is the joint probability density of @xmath91 with distribution @xmath92 then , we can attempt to estimate @xmath8 via the ml estimate @xmath93 again , the set @xmath94 is a set of possible covariance matrices @xmath89 .
it is useful to briefly compare the above ml parameter estimation with the em - gamp method proposed by vila and schniter in @xcite and krzakala _ et .
_ in @xcite .
both of these methods combine the bayesian amp @xcite or gamp algorithms @xcite with a standard em procedure @xcite as follows .
first , the algorithms use the sum - product version of the amp / gamp algorithms , so that the outputs can provide an estimate of the posterior distributions on the components of @xmath2 given the current parameter estimates .
specifically , at any iteration @xmath26 , define the distribution + & = & p_x(x_j|_x^1 ) .
[ eq : phatpost ] for the sum - product amp or gamp algorithms , it is shown in @xcite that the se equations simplify so that @xmath95 and @xmath96 , if the parameters were selected correctly . therefore , from theorem [ thm : stateevogamp ] , the conditional distribution @xmath97 should approximately match the distribution for large @xmath35 . if , in addition , we treat @xmath98 and @xmath59 as sufficient statistics for estimating @xmath13 given @xmath14 and @xmath10 , then @xmath99 can be treated as an approximation for the posterior distribution of @xmath13 given the current parameter estimate @xmath100 .
some justification for this last step can be found in @xcite . using the approximation
, we can approximately implement the em procedure to update the parameter estimate via a maximization + & : = & _ _ x _ x _ j=1^n , [ eq : lamem ] where the expectation is with respect to the distribution in . in @xcite ,
the parameter update is performed only once every few iterations to allow @xmath101 to converge to the approximation of the posterior distribution of @xmath13 given the current parameter estimates . in @xcite , the parameter estimate is updated every iteration . a similar procedure can be performed for the estimation of @xmath8 .
we thus see that the em - gamp procedures in @xcite and in @xcite are both special cases of the adaptive gamp algorithm in algorithm [ algo : gamp ] with particular choices of the adaptation functions @xmath102 and @xmath103 . as a result
, our analysis in theorem [ thm : stateevo ] below can be applied to these algorithms to provide rigorous asymptotic characterizations of the em - gamp performance . however , at the current time , we can only prove the asymptotic consistency result , theorem [ thm : consistent ] , for the ml adaptation functions and described above .
that being said , it should be pointed out that em - gamp update is generally computationally much simpler than the ml updates in and .
for example , when @xmath104 is an exponential family , the optimization in is convex .
also , the optimizations in and require searches over additional parameters such as @xmath105 and @xmath106 .
thus , an interesting avenue of future work is to apply the analysis result , theorem [ thm : consistent ] below , to see if the em - gamp method or some similarly computationally simple technique can be developed which also provides asymptotic consistency .
given the distributions in assumption [ as : gamp ] , compute the sequence of parameters as follows : * _ initialization _ set @xmath107 with [ eq : seinit ] _
x^0 = ( x , ^0 ) , _
x^0 = _ x^0 , where the expectation is over the random variables @xmath108 in assumption [ as : gamp](b ) and @xmath109 is the initial value in the gamp algorithm . * _ output node update : _ compute the variables associated with the output nodes + [ eq : outse ] compute the variables _ p^t & = & _ x^t , _ p^t = _ x^t , + _ z^t & = & h_z^t(p^t , y , _
p^t ) , [ eq : lambarzse ] + _ r^t & = & -^-1 , [ eq : taubarrse ] + _ r^t & = & ( _ r^t)^2 , + _ r^t & = & _ r^t , [ eq : alphase ] + where the expectations are over the random variables @xmath72 and @xmath62 is given in . * _ input node update : _ compute + [ eq : inse ] _
x^t & = & h_x^t(r^t,_r^t ) , [ eq : lambarxse ] + _ x^1 & = & _ r^t , + ^1_x & = & ( x,^1 ) , + where the expectations are over the random variables in .
before proving the asymptotic consistency of the adaptive gamp method with ml adaptation , we first prove the following more general convergence result . [ as : agamp ] consider the adaptive gamp algorithm running on a sequence of problems indexed by the dimension @xmath35 , satisfying the following assumptions : * same as assumption [ as : gamp](a ) to ( c ) with @xmath110 . * for every @xmath26 , * [ thm : stateevo ] consider the random vectors @xmath64 and @xmath65 generated by the outputs of the adaptive gamp under assumption [ as : agamp ] .
let @xmath74 and @xmath75 be the random vectors in with the parameters determined by the se equations in algorithm [ algo : se ] .
then , for any fixed @xmath26 , the components of @xmath64 and @xmath65 converge empirically with bounded moments of order @xmath110 as [ eq : thetalim ] _ n _ x^t _ x^t , _
z^t _ z^t . where @xmath74 and @xmath75 are given in .
in addition , for any @xmath26 , the limits [ eq : taulamlim ] @xmath111 also hold almost surely .
the result is a natural generalization of theorem [ thm : stateevogamp ] and provides a simple extension of the se analysis to incorporate the adaptation .
the se analysis applies to essentially arbitrary adaptation functions .
it particular , it can be used to analyze both the behavior of the adaptive gamp algorithm with either ml and em - gamp adaptation functions in the previous section .
the proof is straightforward and is based on a continuity argument also used in @xcite .
we cam now use theorem [ thm : stateevo ] to prove the asymptotic consistency of the adaptive gamp method with the ml parameter estimation described in section [ sec : mladapt ] .
the following two assumptions can be regarded as _ identifiability
_ conditions .
[ def : pxident ] consider a family of distributions , @xmath112 , a set @xmath113 of parameters @xmath114 of a gaussian channel and function @xmath83 .
we say that @xmath104 is _ identifiable with gaussian outputs _ with parameter set @xmath113 and function @xmath115 if : * the sets @xmath113 and @xmath17 are compact . * for any true " parameters @xmath116 , and @xmath117 , the maximization + & & , [ eq : lamxscore ] is well - defined , unique and returns the true value , @xmath118 .
the expectation in is with respect to @xmath119 and @xmath120 .
* for every @xmath7 and @xmath105 , @xmath106 , the function @xmath83 is pseudo - lipschitz continuous of order @xmath110 in @xmath52 .
[ def : pzident ] consider a family of conditional distributions , @xmath121 generated by the mapping @xmath122 where @xmath123 is some random variable and @xmath124 is a scalar function .
let @xmath125 be a set of covariance matrices @xmath89 and let @xmath126 be some function .
we say that conditional distribution family @xmath127 is _ identifiable with gaussian inputs _ with covariance set @xmath125 and function @xmath128 if : * the parameter sets @xmath125 and @xmath18 are compact . * for any true " parameter @xmath129 and true covariance @xmath130 ,
the maximization + & & , [ eq : lamzscore ] is well - defined , unique and returns the true value , @xmath131 , the expectation in is with respect to @xmath132 and @xmath133 .
* for every @xmath8 and @xmath89 , the function @xmath126 is pseudo - lipschitz continuous in @xmath134 of order @xmath110 .
in addition , it is continuous in @xmath135 uniformly over @xmath136 and @xmath137 . definitions [ def : pxident ] and [ def : pzident ] essentially require that the parameters @xmath7 and @xmath8 can be identified through a maximization . the functions @xmath115 and @xmath128 can be the log likelihood functions and , although we permit other functions as well , since the maximization may be computationally simpler .
such functions are sometimes called _ pseudo - likelihoods_. the existence of a such a function is a mild condition . indeed ,
if such a function does not exists , then the distributions on @xmath84 or @xmath138 must be the same for at least two different parameter values . in that case , one can not hope to identify the correct value from observations of the vectors @xmath25 or @xmath139 .
[ as : agamp - ml ] let @xmath104 and @xmath140 be families of distributions and consider the adaptive gamp algorithm , algorithm [ algo : gamp ] , run on a sequence of problems , indexed by the dimension @xmath35 satisfying the following assumptions : * same as assumption [ as : gamp](a ) to ( c ) with @xmath110 .
in addition , the distributions for the vector @xmath61 is given by @xmath141 for some true " parameter @xmath116 and the conditional distribution of @xmath62 given @xmath63 is given by @xmath142 for some true " parameter @xmath129 .
* same as assumption [ as : agamp](c ) . *
the adaptation functions are set to and .
[ thm : consistent ] consider the outputs of the adaptive gamp algorithm with ml adaptation as described in assumption [ as : agamp - ml ] .
then , for any fixed @xmath26 , * the components of @xmath64 and @xmath65 in converge empirically with bounded moments of order @xmath110 as in and the limits hold almost surely . * in addition , if @xmath143 , and the family of distributions @xmath144 , @xmath15 is identifiable in gaussian noise with parameter set @xmath86 and pseudo - likelihood @xmath115 ( see definition [ def : pxident ] ) , then [ eq : lamxcons ] _
n ^t_x = _ x^t = ^*_x almost surely . *
similarly , if @xmath145 for some @xmath26 , and the family of distributions @xmath146 , @xmath16 is identifiable with gaussian inputs with parameter set @xmath94 and pseudo - likelihood @xmath128 ( see definition [ def : pzident ] ) then [ eq : lamzcons ] _
n ^t_z = _ z^t = ^*_z almost surely .
_ proof : _ see appendix [ sec : consistentpf ] .
the theorem shows , remarkably , that for a very large class of the parameterized distributions , the adaptive gamp algorithm is able to asymptotically estimate the correct parameters .
moreover , there is asymptotically no performance loss between the adaptive gamp algorithm and a corresponding oracle gamp algorithm that knows the correct parameters in the sense that the empirical distributions of the algorithm outputs are described by the same se equations .
there are two key requirements : first , that the optimizations in and can be computed
. these optimizations may be non - convex .
secondly , that the optimizations can be performed are over sufficiently large sets of gaussian channel parameters @xmath113 and @xmath125 such that it can be guaranteed that the se equations eventually enter these sets . in the examples below , we will see ways to reduce the search space of gaussian channel parameters .
recent results suggest that there is considerable value in learning of priors @xmath6 in the context of compressed sensing @xcite , which considers the estimation of sparse vectors @xmath2 from underdetermined measurements ( @xmath147 ) .
it is known that estimators such as lasso offer certain optimal min - max performance over a large class of sparse distributions @xcite . however ,
for many particular distributions , there is a potentially large performance gap between lasso and mmse estimator with the correct prior .
this gap was the main motivation for @xcite which showed large gains of the em - gamp method due to its ability to learn the prior . here
, we illustrate the performance and asymptotic consistency of adaptive gamp in a simple compressed sensing example . specifically , we consider the estimation of a sparse vector @xmath0 from @xmath148 noisy measurements @xmath149 where the additive noise @xmath48 is random with i.i.d .
entries @xmath150 . here
, the output " channel is determined by the statistics on @xmath48 , which are assumed to be known to the estimator .
so , there are no unknown parameters @xmath8 . as a model for the sparse input vector @xmath2
, we assumed the components are i.i.d . with the gauss - bernoulli distribution , [ eq : gaussbern ] x_j ~\ { ll 0 &
= 1- , + ( 0,_x^2 ) & = . where @xmath151 represents the probability that the component is non - zero ( i.e.the vector s sparsity ratio ) and @xmath152 is the variance of the non - zero components .
the parameters @xmath153 are treated as unknown . in the adaptive gamp algorithm
, we use the estimation functions @xmath154 , @xmath155 , and @xmath156 corresponding to the sum - product gamp algorithm . as described in appendix [
sec : gampdetails ] , for the sum - produce gamp the se equations simplify so that @xmath157 and @xmath96 .
since the noise variance is known , the initial output noise variance @xmath158 obtained by adaptive gamp in algorithm [ algo : gamp ] exactly matches that of oracle gamp .
therefore , for @xmath159 , the parameters @xmath70 and @xmath71 do not need to be estimated , and conveniently simplifies to @xmath160 where @xmath161 \times [ 0 , + \infty)$ ] . for iteration @xmath162 ,
we rely on asymptotic consistency , and assume that the maximization yields the correct parameter estimates , so that @xmath163
. then , in principle , for @xmath162 adaptive gamp uses the correct parameter estimates and we expect it to match the performance of oracle gamp . in our implementation , we run em update until convergence to approximate the ml adaptation . fig .
[ fig : mseresults ] illustrates the performance of adaptive gamp on signals of length @xmath164 generated with the parameters @xmath165 .
the performance of adaptive gamp is compared to that of lasso with mse optimal regularization parameter , and oracle gamp that knows the parameters of the prior exactly . for generating the graphs
, we performed @xmath166 random trials by forming the measurement matrix @xmath10 from i.i.d .
zero - mean gaussian random variables of variance @xmath167 . in figure
[ fig : mseresults](a ) , we keep the variance of the noise fixed to @xmath168 and plot the average mse of the reconstruction against the measurement ratio @xmath169 . in figure
[ fig : mseresults](b ) , we keep the measurement ratio fixed to @xmath170 and plot the average mse of the reconstruction against the noise variance @xmath171 . for completeness , we also provide the asymptotic mse values computed via se recursion .
the results illustrate that gamp significantly outperforms lasso over the whole range of @xmath169 and @xmath171 .
moreover , the results corroborate the consistency of adaptive gamp which achieves nearly identical quality of reconstruction with oracle gamp .
the performance results indicate that adaptive gamp can be an effective method for estimation when the parameters of the problem are difficult to characterize and must be estimated from data . as a second example , we consider the estimation of the linear - nonlinear - poisson ( lnp ) cascade model @xcite .
the model has been successfully used to characterize neural spike responses in early sensory pathways of the visual system . in the context of lnp cascade model ,
the vector @xmath0 represents the linear filter , which models the linear receptive field of the neuron .
amp techniques combined with the parameter estimation have been recently proposed for neural receptive field estimation and connectivity detection in @xcite . as in section
[ sec : gaussbern ] , we model @xmath2 as a gauss - bernoulli vector of unknown parameters @xmath172 . to obtain the measurements @xmath14
, the vector @xmath4 is passed through a componentwise nonlinearity @xmath173 to result in u(z ) = .
let the function @xmath174 denote a vector ( z ) = ( 1 , u(z ) ,
, u(z)^r-1)^t .
then , the final measurement vector @xmath14 is generated by a measurement channel with a conditional density of the form [ eq : pyzexp ] p_y|z(y_i|z_i,_z ) = ^-f(z_i ) , where @xmath175 denotes the nonlinearity given by @xmath176 adaptive gamp can now be used to also estimate vector of polynomial coefficients @xmath8 , which together with @xmath2 , completely characterizes the lnp system . .
this plots illustrates near consistency of adaptive gamp for large @xmath35.,width=321 ] the estimation of @xmath8 is performed with ml estimator described in section [ sec : mladapt ] .
we assume that the mean and variance of the vector @xmath2 are known at iteration @xmath107 .
this implies that for sum - product gamp the covariance @xmath177 is initially known and the optimization simplifies to @xmath178 where @xmath179 .
the estimation of @xmath7 is performed as in section [ sec : gaussbern ] .
as before , for iteration @xmath162 , we assume that the maximizations and yield correct parameter estimates @xmath163 and @xmath180 , respectively .
thus we can conclude by induction that for @xmath162 the adaptive gamp algorithm should continue matching oracle gamp for large enough @xmath35 . in our simulations
, we implemented with a gradient ascend algorithm and run it until convergence . in fig .
[ fig : nlresults ] , we compare the reconstruction performance of adaptive gamp against the oracle version that knows the true parameters @xmath181 exactly .
we consider the vector @xmath2 generated with true parameters @xmath182 .
we consider the case @xmath183 and set the parameters of the output channel to @xmath184 $ ] .
to illustrate the asymptotic consistency of the adaptive algorithm , we consider the signals of length @xmath185 and @xmath186 .
we perform @xmath187 and @xmath188 random trials for long and short signals , respectively , and plot the average mse of the reconstruction against @xmath169 .
as expected , for large @xmath35 , the performance of adaptive gamp is nearly identical ( within @xmath189 ) to that of oracle gamp .
we have presented an adaptive gamp method for the estimation of i.i.d .
vectors @xmath2 observed through a known linear transforms followed by an arbitrary , componentwise random transform .
the procedure , which is a generalization of em - gamp methodology of @xcite that estimates both the vector @xmath2 as well as parameters in the source and componentwise output transform . in the case of large i.i.d .
gaussian transforms , it is shown that the adaptive gamp method is provably asymptotically consistent in that the parameter estimates converge to the true values .
this convergence result holds over a large class of models with essentially arbitrarily complex parameterizations .
moreover , the algorithm is computationally efficient since it reduces the vector - valued estimation problem to a sequence of scalar estimation problems in gaussian noise .
we believe that this method is applicable to a large class of linear - nonlinear models with provable guarantees can have applications in a wide range of problems .
we have mentioned the use of the method for learning sparse priors in compressed sensing .
future work will include learning of parameters of output functions as well as possible extensions to non - gaussian matrices .
as described in @xcite , the sum - product estimation can be implemented with the estimation functions [ eq : gsp ] g_x^t(r,_r,_x ) & : = & , [ eq : gxsp ] + g_z^t(p , y,_p,_z ) & : = & , [ eq : gzsp ] + g_s^t(p , y,_p,_z ) & : = & ( g_z^t(p , y,_p,_z ) - p ) , where the expectations are with respect to the scalar random variables [ eq : gspdist ] @xmath190 the estimation functions correspond to scalar estimates of random variables in additive white gaussian noise ( awgn ) .
a key result of @xcite is that , when the parameters are set to the true values ( i.e. @xmath191 ) , the outputs @xmath32 and @xmath33 can be interpreted as sum products estimates of the conditional expectations @xmath192 and @xmath193 .
the algorithm thus reduces the vector - valued estimation problem to a computationally simple sequence of scalar awgn estimation problems along with linear transforms .
moreover , the se equations in algorithm [ algo : se ] reduce to a particularly simple forms , where @xmath54 and @xmath71 in are given by @xmath194,\ ] ] where the expectations are over the random variables @xmath72 and @xmath62 is given in .
the covariance matrix @xmath73 has the form @xmath195
\beta\tau_{x0 } - \taubar_p^t & \beta\tau_{x0 } - \taubar_p^t \\[0.3em ] \end{bmatrix},\ ] ] where @xmath196 is the variance of @xmath61 and @xmath40 is the asymptotic measurement ratio ( see assumption 1 for details ) .
the scaling constant becomes @xmath157 .
the update rule for @xmath197 also simplifies to @xmath198,\ ] ] where the expectation is over the random variables in .
bayati and montanari s analysis in @xcite employs certain deterministic models on the vectors and then proves convergence properties of related empirical distributions . to apply the same analysis here
, we need to review some of their definitions .
we say a function @xmath199 is _ pseudo - lipschitz _ of order @xmath200 , if there exists an @xmath201 such for any @xmath2 , @xmath202 , @xmath203 now suppose that for each @xmath204 , @xmath205 is a set of vectors [ eq : vsetapp ] ^(n ) = \{_i(n ) , i=1,
,(n ) } , where each element @xmath206 and @xmath207 is the number of elements in the set .
thus , @xmath205 can itself be regarded as a vector with @xmath208 components .
we say that @xmath205 _ empirically converges with bounded moments of order @xmath76 _ as @xmath209 to a random vector @xmath210 on @xmath211 if : for all pseudo - lipschitz continuous functions , @xmath212 , of order @xmath76 , @xmath213 when the nature of convergence is clear , we may write ( with some abuse of notation ) @xmath214 or @xmath215 finally , let @xmath216 be the set of probability distributions on @xmath211 with bounded @xmath76th moments , and suppose that @xmath217 is a functional @xmath216 to some topological space @xmath218 . given a set @xmath205 as in ,
write @xmath219 for @xmath220 where @xmath221 is the empirical distribution on the components of @xmath222 . also , given a random vector @xmath210 with distribution @xmath223 write @xmath224 for @xmath225 . then
, we will say that the functional @xmath226 is _ weakly pseudo - lipschitz continuous _ of order @xmath76 if @xmath227 where the limit on the right hand side is in the topology of @xmath218 .
the proof follows along the adaptation argument of @xcite .
we use the tilde superscript on quantities such as @xmath228 , and @xmath229 to denote values generated via a non - adaptive version of the gamp .
the non - adaptive gamp algorithm has the same initial conditions as the adaptive algorithm ( i.e. @xmath230 ) , but with @xmath82 and @xmath81 replaced by their deterministic limits @xmath231 and @xmath232 , respectively .
that is , we replace lines [ line : zhat ] , [ line : shat ] and [ line : xhat ] with _ i^t & = & g_z^t(p_i^t , y_i,_p^t,_z^t ) , _
i^t = g_s^t(p_i^t , y_i,_p^t,_z^t ) , + _ j^1 & = & g_x^t(r_j^t , _ r^t,_x^t ) .
this non - adaptive algorithm is precisely the standard gamp method analyzed in @xcite .
the results in that paper show that the outputs of the non - adaptive algorithm satisfy all the required limits from the se analysis .
that is , @xmath233 where @xmath234 and @xmath235 are the sets generated by the non - adaptive gamp algorithm : @xmath236 the limits are now proven through a continuity argument that shows that the adaptive and non - adaptive quantities must asymptotically agree with one another . specifically , we will start by proving that the following limits holds almost surely for all @xmath237 the proof of the limits and is achieved by an induction on @xmath26 . although we only need to show the above limits for @xmath110 , most of the arguments hold for arbitrary @xmath45 .
we thus present the general derivation where possible . to begin the induction argument , first note that the non - adaptive algorithm has the same initial conditions as the adaptive algorithm .
thus the limits and hold for @xmath159 and @xmath240 , respectively .
we now proceed by induction .
suppose that @xmath241 and the limits and hold for some @xmath26 and @xmath242 , respectively . since @xmath10 has i.i.d .
components with zero mean and variance @xmath167 , it follows from the marenko - pastur theorem @xcite that that its @xmath243-norm operator norm is bounded .
that is , there exists a constant @xmath244 such that [ eq : matrixbnd ] _
n _ k c_a , _
n ^_k c_a .
this bound is the only part of the proof that specifically requires @xmath110 . from ,
we obtain @xmath245 where ( a ) is due to the norm inequality @xmath246 . since @xmath247 , we have that for any positive numbers @xmath248 and @xmath249 [ eq : pbndtwo ] ( a + b)^k 2^k(a^k + b^k ) . applying the inequality into
, we obtain @xmath250 now , since @xmath251 and @xmath252 are the outputs of the non - adaptive algorithm they satisfy the limits to prove , we first prove the empirical convergence of @xmath257 to @xmath88 . towards this end , let @xmath258 be any pseudo - lipschitz continuous function @xmath212 of order @xmath76 .
then @xmath259\right| } \nonumber \\ & \leq \frac{1}{m } \sum_{i = 1}^m \left|\phi(p_i^{t},y_i ) - \phi(\tilde{p}_i^{t},y_i ) \right| \nonumber\\ & + \left| \frac{1}{m } \sum_{i = 1}^m \phi(\tilde{p}_i^{t},y_i ) -\exp\left[\phi(p^t , y)\right]\right| \nonumber \\ & { { \overset{\mathrm{(a)}}{\leq } } } \frac{l}{m } \sum_{i = 1}^m \left(1 + |p_i^t|^{k-1 } + |\tilde{p}_i^t|^{k-1 } + |y_i|^{k-1 } \right)|p_i^t - \tilde{p}_i^t| \nonumber \\ & + \left| \frac{1}{m } \sum_{i = 1}^m \phi(\tilde{p}_i^{t},y_i ) -\exp\left[\phi(p^t , y)\right]\right| \nonumber \\ & { { \overset{\mathrm{(b)}}{\leq } } } lc \delta_p^t+\left| \frac{1}{m } \sum_{i = 1}^m \phi(\tilde{p}_i^{t},y_i ) -\exp\left[\phi(p^t , y)\right]\right| .
\label{eq : phipybnd}\end{aligned}\ ] ] in ( a ) we use the fact that @xmath212 is pseudo - lipschitz , and in ( b ) we use hlder s inequality @xmath260 with @xmath261 and define @xmath262 as @xmath263^{k/(k-1 ) } } \nonumber \\ & \leq \frac{1}{m}\sum_{i=1}^m \left ( 1 + |p_i^t|^{k-1 } + |\tilde{p}_i^t|^{k-1 } + |y_i|^{k-1}\right)^{k/(k-1 ) } \nonumber \\ & \leq \mbox{const}\x\left [ 1 + \left(\frac{1}{m}\left\|{\mathbf{p}}^t\right\|_k^k\right)^{\frac{k-1}{k } } \right . \nonumber\\ & \left.+ \left(\frac{1}{m}\left\|\porc^t\right\|_k^k\right)^{\frac{k-1}{k } } + \left(\frac{1}{m}\left\|{\mathbf{y}}\right\|_k^k\right)^{\frac{k-1}{k } } \right ] , \label{eq : cpbnddef}\end{aligned}\ ] ] where the first step is from jensen s inequality .
since @xmath264 satisfy the limits for the non - adaptive algorithm we have : [ eq : ptildeybnd ] @xmath265 < \infty \\ \lim_{n \rightarrow \infty } \frac{1}{m}\|{\mathbf{y}}\|_k^k & = \lim_{n \rightarrow \infty } \frac{1}{m}\sum_{i = 1}^m |y_i|^k = \exp\left[|y|^k\right ] < \infty\end{aligned}\ ] ] also , from the induction hypothesis , it follows that the adaptive output must satisfy the same limit @xmath266 < \infty.\ ] ] combining , , , , we conclude that for all @xmath237 [ eq : pyconv ] _
n ( ^t , ) ( p^t , y ) . the limit along with and the continuity condition on @xmath20 in assumption [ as : gamp](d ) prove the limit in .
the limit together with continuity conditions on @xmath28 in assumptions [ as : gamp ] show that , and hold for @xmath26 .
for example , to show , we consider the limit @xmath267 of the following expression @xmath268 where at ( a ) we used the lipschitz continuity assumption .
similar arguments can be used for and . to show , we proceed exactly as for . due to the continuity assumptions on @xmath269 ,
this limit in turn shows that holds almost surely .
then , and follow directly from the continuity of @xmath154 in assumptions [ as : gamp ] , together with and .
we have thus shown that if the limits and hold for some @xmath26 , they hold for @xmath270 .
thus , by induction they hold for all @xmath26 . finally , to show , let @xmath212 be any pseudo - lipschitz continuous function @xmath271 , and define @xmath272\right|,\end{aligned}\ ] ] which
, due to convergence of non - adaptive gamp , can be made arbitrarily small by choosing @xmath35 large enough .
then , consider @xmath273\right| \nonumber\\ & \leq \epsilon_n^t + \frac{1}{n } \sum_{j = 1}^n \left|\phi(x_j , \hat{r}_j^{t } , \xhat_j^{\tp1 } ) - \phi(x_j , \tilde{r}_j^{t } , \tilde{x}_j^{\tp1 } ) \right| \nonumber\\ & { { \overset{\mathrm{(a)}}{\leq } } } \epsilon_n^t + l\|{\mathbf{r}}^t - \rorc^t\|_1 + l\|\hat{{\mathbf{x}}}^{\tp1 } - \xorc^{\tp1}\|_1 \nonumber\\ & + \frac{l^\prime}{n}\sum_{j = 1}^n \left(|\hat{r}_j^t|^{k-1 } + |\tilde{r}_j^t|^{k-1 } \right)(|\hat{r}_j^t - \tilde{r}_j^t| + |\xhat_j^{\tp1 } - \tilde{x}_j^{\tp1}| ) \nonumber\\ & + \frac{l^\prime}{n}\sum_{j = 1}^n \left(|\xhat_j^{\tp1}|^{k-1 } + |\tilde{x}_j^{\tp1}|^{k-1}\right)(|\hat{r}_j^t - \tilde{r}_j^t| + |\xhat_j^{\tp1 } - \tilde{x}_j^{\tp1}| ) \nonumber\\ & { { \overset{\mathrm{(b)}}{\leq } } } \epsilon_n^t + l\left(\delta_r^t\right)^{\frac{1}{k } } + l\left(\delta_x^t\right)^{\frac{1}{k}}\nonumber\\ & + l^{\prime}\left(\delta_r^t\right)^{\frac{1}{k}}\left((\tilde{m}_x^{\tp1})^{\frac{k-1}{k}}+(\hat{m}_x^{\tp1})^{\frac{k-1}{k } } + ( \tilde{m}_r^{t})^{\frac{k-1}{k}}+(\hat{m}_r^{t})^{\frac{k-1}{k}}\right)\nonumber\\ & + l^{\prime}\left(\delta_x^t\right)^{\frac{1}{k } } \left((\tilde{m}_x^{\tp1})^{\frac{k-1}{k}}+(\hat{m}_x^{\tp1})^{\frac{k-1}{k } } + ( \tilde{m}_r^{t})^{\frac{k-1}{k}}+(\hat{m}_r^{t})^{\frac{k-1}{k}}\right ) \label{eq : xbound } \ ] ] where @xmath12 , @xmath274 are constants independent of @xmath35 and @xmath275 in ( a ) we use the fact that @xmath212 is pseudo - lipshitz , in ( b ) we use @xmath276-norm equivalence @xmath277 and hlder s inequality @xmath260 with @xmath261 . by applying of , and since
, @xmath278 , @xmath279 , @xmath280 , and @xmath281 converge to a finite value we can obtain the first equation of by taking @xmath41 .
the second equation in can be shown in a similar way .
this proves the limits . also , the first two limits in are a consequence of and .
the second two limits follow from continuity assumptions in assumption [ as : gamp](e ) and the convergence of the empirical distributions in .
this completes the proof .
part ( a ) of theorem [ thm : consistent ] is a direct application of the general result , theorem [ sec : stateevo ] . to apply the general result ,
first observe that assumptions [ as : agamp - ml](a ) and ( c ) immediately imply the corresponding items in assumptions [ as : agamp ] .
so , we only need to verify the continuity condition in assumption [ as : agamp](b ) for the adaptation functions in and .
we begin by proving the continuity of @xmath20 .
fix @xmath26 , and let @xmath282 be a sequence of vectors and @xmath283 be a sequence of scalars such that [ eq : yptaulim ]
_ n ( ^(n),^(n ) ) ( y , p^t ) _ n _ p^(n ) = ^t_p , where @xmath284 and @xmath285 are the outputs of the state evolution equations . for each @xmath35 , let [ eq : lamhatcons ] _ z^(n ) = h_z^t(^(n),^(n),_p^(n ) ) .
we wish to show that @xmath286 , the true parameter .
since @xmath287 and @xmath18 is compact , it suffices to show that , any limit point of any convergent subsequence is equal to @xmath288 .
so , suppose that @xmath289 to some limit point @xmath290 on some subsequence @xmath291 . from @xmath291 and the definition
it follows that @xmath292 now , since @xmath293 and @xmath289 , we can apply the continuity condition in definition [ def : pzident](c ) to obtain @xmath294 \geq 0.\end{aligned}\ ] ] also , the limit and the fact that @xmath128 is psuedo - lipschitz continuous of order @xmath76 implies that [ eq : phizcompc ] .
but , property ( b ) of definition [ def : pzident ] shows that @xmath288 is the maxima of the right - hand side , so [ eq : phizcompd ] = . since , by definition [ def : pzident](b ) , the maxima is unique , @xmath295 .
since this limit point is the same for all convergent subsequences , we see that @xmath296 over the entire sequence . we have thus shown that given limits , the outputs of the adaptation function converge as @xmath297 thus , the continuity condition on @xmath103 in assumption [ as : agamp](b ) is satisfied .
the analogous continuity condition on @xmath102 can be proven in a similar manner .
so , it remains to show parts ( b ) and ( c ) of theorem [ thm : consistent ] .
we will only prove ( b ) ; the proof of ( c ) is similar .
also , since we have already established , we only need to show that the output of the se equations matches the true parameter .
that is , we need to show @xmath298 .
this fact follows immediately from the selection of the adaptation functions : + & & _ _ x _ x _ ( _ r,_r ) s_x(_r^t ) + & & _ _ x _ x _ ( _ r,_r ) s_x(_r^t ) + & & + & & _ x^ * [ eq : lamxlim ] where ( a ) follows from the se equation ; ( b ) is the definition of the ml adaptation function @xmath299 when interpreted as a functional on a random variable @xmath67 ; ( c ) is the definition of the random variable @xmath67 in where @xmath300 ; and ( d ) follows from definition [ def : pxident](b ) and the hypothesis that @xmath301 .
thus , we have proven that @xmath298 , and this completes the proof of part ( b ) of theorem [ thm : consistent ] .
the proof of part ( c ) is similar . | we consider the estimation of an i.i.d .
( possibly non - gaussian ) vector @xmath0 from measurements @xmath1 obtained by a general cascade model consisting of a known linear transform followed by a probabilistic componentwise ( possibly nonlinear ) measurement channel .
a novel method , called adaptive generalized approximate message passing ( adaptive gamp ) , that enables joint learning of the statistics of the prior and measurement channel along with estimation of the unknown vector @xmath2 is presented .
the proposed algorithm is a generalization of a recently - developed em - gamp that uses expectation - maximization ( em ) iterations where the posteriors in the e - steps are computed via approximate message passing .
the methodology can be applied to a large class of learning problems including the learning of sparse priors in compressed sensing or identification of linear - nonlinear cascade models in dynamical systems and neural spiking processes .
we prove that for large i.i.d .
gaussian transform matrices the asymptotic componentwise behavior of the adaptive gamp algorithm is predicted by a simple set of scalar state evolution equations . in addition
, we show that when a certain maximum - likelihood estimation can be performed in each step , the adaptive gamp method can yield asymptotically consistent parameter estimates , which implies that the algorithm achieves a reconstruction quality equivalent to the oracle algorithm that knows the correct parameter values .
remarkably , this result applies to essentially arbitrary parametrizations of the unknown distributions , including ones that are nonlinear and non - gaussian .
the adaptive gamp methodology thus provides a systematic , general and computationally efficient method applicable to a large range of complex linear - nonlinear models with provable guarantees . |
the effect of disorder on different types of condensed matter orderings is nowadays a subject of considerable interest @xcite . for the case of disordered magnetic systems ,
random - field spin models have been systematically studied , not only for theoretical interests , but for some identifications with experimental realizations @xcite .
an interesting issue , is the study of how quenched randomness destroys some types of criticalities .
so , in what concerns the effect produced by random fields in low dimensions , it has been noticed @xcite that first - order transitions will be replaced by continuous transitions , so tricritical points and critical end points will be depressed in temperature , and a finite amount of disorder will suppresse them .
nevertheless , in two dimensions , an infinitesimal amount of field randomness seems to destroy any first - order transition @xcite .
interestingly , the simplest model exhibiting a tricritical phase diagram in the absence of randomness is the blume - capel model .
the blume - capel model @xcite is a regular ising model for spin-1 used to model @xmath6 mixtures@xcite .
the interesting feature is the existence of a tricritical point in the phase diagram represented in the plane temperature versus crystal field , as shown in figure 1 .
this phase diagram was firslty obtained in the mean - field approach , but the same qualitative properties were also observed in low dimensions .
the latter was confirmed through some approximation techniques as well as by monte carlo simulations @xcite . also , the tricritical behavior is still held in two dimensions @xcite .
nevertheless , in other models this situation is controversial .
for example , the random - field ising model in the mean - field approach @xcite also exhibits a tricritical point , but some monte carlo simulations @xcite in the cubic lattice suggest that this is only an artifact of the mean - field calculations . accordingly , this interesting fact in the blume - capel model motivated some authors to explore the richness of this model , within the mean - field approach , by introducing disorder in the crystal field @xcite as well as by adding and external random field @xcite . for the former case , it was obtained a variety of phase diagrams including different critical points with some similar topologies found for the random - field spin@xmath7 ising model @xcite . however , in those studies the fourth - order critical points , which limit the existence of tricritical points , were overlooked .
consequently , our aim in this work is to improve those previous studies by considering a more general probability distribution function for the crystal field , and bettering some results given in references @xcite .
the next section is dedicated to define the model and the special critical points produced by it . within the mean - field approach , where @xmath8 is the boltzman constant , @xmath9 is the temperature , @xmath10 is the coupling constant between each pair of spins , and @xmath11 is the crystal field ( also called the anisotropy field ) .
the black circle represents the tricritical point .
the ferromagnetic and paramagnetic phases are represented by @xmath12 and @xmath13 , respectively .
the full line represents the continuous or second - order critical frontier , and the dotted line is for the first - order frontier.,height=207 ]
the infinite - range - interaction blume - capel model is given by the following hamiltonian @xmath14 where @xmath15 , and @xmath16 is the number of spins .
the first sum runs over all distinct pairs of spins .
the coupling constant @xmath17 is divided by @xmath16 in order to maintain the extensivity .
the crystal fields are represented by quenched variables @xmath18 , obeying the probability distribution function ( pdf ) given by , @xmath19 + \frac{(1-p)}{\sqrt{2\pi } \,\sigma } \exp \left [ - \frac{{\delta_{i}}^{2}}{2{\sigma}^{2 } } \right ] ~ , \label{pdf}\ ] ] which consists of a superposition of two independent gaussian distributions with the same width @xmath1 , centered at @xmath20 and @xmath21 , with probabilities @xmath4 and @xmath5 , respectively .
for @xmath22 , we recover the bimodal distribution studied in references @xcite , and for @xmath23 , the simple gaussian one of reference @xcite . for @xmath24 and @xmath23 , we go back to the simple blume - capel model without randomness@xcite .
+ by standard procedures @xcite , we get the analytical expression for the free energy per spin ( @xmath25 ) , through which may be obtained a self - consistent equation for the magnetization @xmath26 .
thus , we have the following relations at the equilibrium , @xmath27 @xmath28}^{-1 } \right \ } ~ , \label{m}\ ] ] where the quenched average , represented by @xmath29 , is taken with respect to the pdf given in eq .
( [ pdf ] ) , and @xmath30 . to write conditions for locating tricritical and fourth - order critical points , we expand the right hand of eq .
( [ m ] ) in powers of @xmath26 ( landau s expansion , see @xcite ) . conveniently , we expand the magnetization up to seventh order in @xmath26 , so @xmath31 where @xmath32 @xmath33 @xmath34 @xmath35 and @xmath36 in order to obtain the continuous critical frontier one sets @xmath37 , provided that @xmath38 .
if a first - order critical frontier begins after the continuous one , the latter line ends at a tricritical point if @xmath39 , provided that @xmath40 .
the possibility of a fourth - order critical point is given for @xmath37 , @xmath39 , @xmath41 and @xmath42 .
thus , a fourth - order point may be regarded as the last tricritical point . + by taking @xmath43 ( @xmath44 ) , we get the asymptotic limit of eqs .
( [ f ] ) and ( [ m ] ) , so we have @xmath45 \right ) + \frac{\sigma}{\sqrt{2\pi } } \exp \left
[ - \frac{(jm-\delta)^{2}}{2 \ , { \sigma}^{2 } } \right ] \right ) \nonumber \\ & - & ( 1-p ) \left ( \frac{1}{2 } jm \left ( 1 + { \rm erf } \left[\frac{jm}{\sqrt{2 } \ , \sigma } \right ] \right ) + \frac{\sigma}{\sqrt{2\pi } } \exp \left
[ - \frac{j^{2 } m^{2}}{2 \ , { \sigma}^{2 } } \right ] \right ) ~ , \end{aligned}\ ] ] @xmath46 \right ) + \frac{(1-p)}{2 } \left ( 1 + { \rm erf } \left [ \frac{jm}{\sqrt{2}\ , \sigma } \right ] \right ) ~,\ ] ] where @xmath47 the critical frontiers , for a given pair @xmath48 , are obtained by solving a non - linear set of equations , which consist of equating the free energies for the corresponding phases ( maxwell s construction ) , and the respective magnetization equations based on the relations given in eqs .
( [ f ] ) , and ( [ m ] ) .
we must carefully verify that every numerical solution minimizes the free energy .
+ the symbols used to represent the different critical lines and points @xcite are as follows : * continuous or second - order critical frontier : continuous line ; * fist - order critical frontier : dotted line ; * tricritical point : located by a black circle ; * fourth - order critical point : located by an empty square ; * ordered critical point : located by an asterisk ; * critical end point : located by a black triangle . to clarify ,
we mean by a continuous critical frontier that which separates two distinct phases through which the order parameter changes continuously to pass from one phase to another , contrary to the case of the first - order transition , through which , the order parameter suffers a discontinuous change , so the two corresponding phases coexist at each critical point .
a tricritical point is basically the point in which a continuous line terminates to give rise a first - order critical line . a fourth - order critical point
is sometimes called a vestigial tricritical point , because it may be regarded as the last tricritical point .
an ordered critical point is the point , inside an ordered region , where a first - order critical line ends , above which the order parameter passes smoothly from one ordered phase to the other .
finally , a critical end point corresponds to the intersection of a continuous line that separates the paramagnetic from one of the ferromagnetic phases with a first- order line separating the paramagnetic and the other ferromagnetic phase . in following section
we make use of this definitions .
the distinct phase diagrams for the present model were numerically obtained by scanning the whole p - domain for each @xmath1-width .
so , distinct topologies belonging to different @xmath4-ranges were found for a given @xmath1 .
for instance , figure 2 shows the whole variety of them for a small @xmath49 , for each arbitrary representative @xmath4 .
note that for small values of @xmath4 , one only ferromagnetic order appears at low temperatures , as shown in figure 2(a ) for @xmath50 . we designate it as topology i. figures 2(b ) and
( c ) ( @xmath51 ) represent the same topology ( topology ii ) , which consists of one first - order critical line separating two different ferromagnetic phases @xmath52 and @xmath53 , and a continuous line remaining for @xmath54 .
figure 2(c ) , though qualitatively the same as in 2(b ) , is intended to show how the first - order line and the continuous line approach themselves as @xmath4 increases .
figure 2(d ) shows topology iii , for @xmath55 , so the preceding first - order line is now dividing the continuous line by two critical end points .
note that the upper continuous line terminates , following a reentrant path , at a critical end point where the phases @xmath52 , @xmath56 , and @xmath13 coexist .
so , at the lower critical end point , @xmath52 , @xmath53 , and @xmath13 coexist . above the ordered critical point , the order parameter passes smoothly from @xmath52 to @xmath56 ( see the inset there ) .
if we increase @xmath4 up to some @xmath57 , the upper continuous line and the first - order line will be met by a fourth - order point ( represented by a square ) as shown in figure 2(e ) .
thus , @xmath58 is the threshold for topology iv .
then , for @xmath59 , those lines will be met at a tricritical point , as noticed in figure 2(f ) .
conversely , tricritical points appear for @xmath60 , so the last one for @xmath57 .
the same types of phase diagrams are found in references @xcite .
nevertheless , we improve their results , not only bettering some of their numerical calculations , but in that we may now locate the regions of validity of these topologies in the plane @xmath61 . to this end , we start by locating the fourth - order points in the plane @xmath61 , as shown in figure 3 . . note that if @xmath62 , we recover the bimodal case studied in references @xcite , where we found @xmath63 , just in agreement with them . note that if @xmath23 , @xmath64 , so it is a @xmath1-limit for the tricritical behavior .
the dashed line is only a guide to the eyes.,height=226 ] note that @xmath65 is a cut - off for the tricritical behavior .
then , topology iv will no longer found for greater widths . on the other hand , we determine the threshold for topology iii , by estimating numerically which value of @xmath4 , for each @xmath66 , produces a situation like that presented in figure 4 ( case @xmath49 ) , where we see how topology iii emerges for a @xmath4 slightly greater than @xmath67 . for @xmath68
, we found this threshold for @xmath69 , which is smaller than that obtained in reference @xcite .
there , the authors suggested that topology iii disappears for @xmath70 .
however , figure 5 illustrates that this type of phase diagram is still present even for a smaller @xmath4 , as confirmed by the free energy evaluated at three disctinct ( @xmath71)-points along the first - order critical line , at which there are three types of coexistences , namely , @xmath52 with @xmath56 , @xmath52 with @xmath13 , and @xmath52 with @xmath53 .
we also noted another discrepancy with respect to a critical @xmath66 , found in reference @xcite , above which topology iii disappears for @xmath23 .
there , the authors affirmed that if @xmath72 , the paramagnetic - ferromagnetic transition becomes second order at all temperatures , but we noticed that it only happens for a greater width , namely , @xmath73 . + in order to obtain the frontier which separates topologies i and ii ( in the plane @xmath61 ) , we have to find the corresponding @xmath4 , for a given @xmath66 , that locates the one ordered critical point at @xmath74 . to this end , the next subsection is focused on zero temperature calculations . in order to perform zero temperature calculations we make use of the equations ( 11 ) and ( 12 ) .
consequently , for @xmath75 ( see reference @xcite ) , there are two ferromagnetic phases @xmath52 and @xmath53 coexisting at @xmath76 , having magnetizations @xmath77 and @xmath78 , respectively .
we observed that these relations still remain up to some finite @xmath1 , after which a @xmath1-dependency emerges .
so , for a greater width called @xmath79 , the ordered critical point ( that of topology iii ) must be found at @xmath74 . then , the first - order critical line is supressed and one only ferromagnetic order exists for any @xmath4 .
for instance , if we choose @xmath80 , we find @xmath81 , as illustrated in figure 6 . there , the zero temperature free energy versus the order parameter is plotted for three different values of @xmath66 , at the point where @xmath82 and @xmath83 coexist .
thus , in ( a ) , two minima are at the same level for @xmath84 . in ( b ) , it still happens for @xmath85 . nonetheless , in ( c ) , for @xmath86 , the ordered critical point is already at @xmath74 .
therefore , for this particular @xmath4 , there is only one ferromagnetic phase for @xmath87 .
for completeness , figure 7(a ) shows what figure 6(c ) illustrates by means of the free energy .
there , it is shown where the ordered critical point is located , that is , at @xmath74 . in figure 7(b ) , we see the line composed by the @xmath88-points .
this line separates phase diagrams containing two and one ferromagnetic phases .
particularly , for @xmath23 , @xmath89 , as obtained in reference @xcite and confirmed numerically by us . + we summarize the preceding analysis by showing , in figure 8 , the regions of validity for the four qualitatively distinct phase diagrams .
note that along the horizontal axis ( @xmath68 ) , regions ii and iii are separated by @xmath69 , and regions iii and iv by @xmath90 .
along the vertical axis ( at @xmath23 ) , regions iv and iii are separated by @xmath65 , regions iii and ii by @xmath73 , then , regions ii and i by @xmath91 . furthermore , the line separating topologies i and ii is the same as in figure 7(b ) .
the frontier separating topologies ii and iii consists of points estimated by the analysis illustrated by figure 4 .
finally , the line between topologies iii and iv is made of fourth - order critical points , i.e. , it is based upon the points in figure 3 . versus @xmath4 , associated with the topologies for the present model ( see also figure 2 ) .
the horizontal and the vertical axes represent the probability @xmath4 , and the width @xmath1 , respectively ( see eq .
( [ pdf ] ) ) .
the tricritical behavior belongs to the region iv .
the simplest topology belongs to region i , where only one ferromagnetic phase appears , whereas the rest topologies contain two ferromagnetic orders at low temperatures.,height=264 ]
we revisiting the study of the infinite - range - interaction spin-1 blume - capel model with quenched randomness , by considering a more general probability distribution function for the crystal field @xmath0 , which consists of two gaussian distributions centered at @xmath2 and @xmath3 , with probabilities @xmath4 and @xmath5 , respectively .
for @xmath24 , we recover the bimodal case studied in references @xcite , and for @xmath23 , the gaussian case studied in reference @xcite . for @xmath1-widths in @xmath92 ,
the system exhibits four distinct topologies according to the range in which @xmath4 belongs .
so , we designed them as topology i , ii , iii , and iv , in increasing order of @xmath4
. topology i contains one continuous critical line separating a ferromagnetic phase to the paramagnetic phase . in topology ii
, one first - order critical line separating two ferromagnetic phases is added .
this line terminates at an ordered critical point .
the most complex criticality belongs to topology iii , where the first - order line now divides the continuous critical line by two critical end points . in topology iv ,
the first - order line and the continuous line are met by a tricritical point .
accordingly , topology i presents one ferromagnetic phase , whereas the rest ones show two distinct ferromagnetic orders at low temperatures . on the other hand ,
the tricritical behavior manifested in topology iv emerges for @xmath60 , where @xmath58 denotes the probability for a given @xmath66 , where a fourth - order critical point is found .
this point may be regarded as the last tricritical point vanishing for @xmath93 , since @xmath64 leads to @xmath94 .
consequently , the tricritical behavior is no longer found for any @xmath4 .
topology iii disappears for @xmath95 , and topology ii is limited by @xmath96 , above which the first - order line separating the two ferromagnetic phases is suppressed for any @xmath4 .
after that , for @xmath97 , only the simplest topology survives .
+ + therefore , we show through this model how a complex magnetic criticality is reduced by the strength of the disorder ( see also @xcite ) .
nevertheless , the critical dimensions for these types of phase diagrams is still an open problem to be solved . | the multicritical behavior of the blume - capel model with infinite - range interactions is investigated by introducing quenched disorder in the crystal field @xmath0 , which is represented by a superposition of two gaussian distributions with the same width @xmath1 , centered at @xmath2 and @xmath3 , with probabilities @xmath4 and @xmath5 , respectively .
a rich variety of phase diagrams is presented , and their distinct topologies are shown for different values of @xmath1 and @xmath4 . the tricritical behavior is analyzed through the existence of fourth - order critical points , as well as how the complexity of the phase diagrams is reduced by the strength of the disorder .
keywords : random - field blume - capel model ; mean - field approach ; tricritical behavior . |
let @xmath2 be a commutative monoid ; we might be interested in natural operations @xmath3 .
here is an example of a natural operation @xmath4 : @xmath5 such an operation consists of adding up some copies of the things we started with .
we can regard this as a two - stage process : first we make copies , then we add .
so we can factor this operation as @xmath6 in general we can associate natural operations to maps of finite sets : * given a map @xmath7 of sets , we can produce a _ copying _ map @xmath8 via @xmath9 . * given a map @xmath10 of sets , we cap produce an _ addition _ map @xmath11 via @xmath12 . of course ,
we can compose these , and so given any diagram of finite sets @xmath13 we get an operation @xmath14 , which sends @xmath15 via @xmath16 we refer to a diagram of sets with this shape as a _
span diagram_. it certainly seems natural to suggest that span diagrams should give all the natural operations on a commutative monoid
. however , span diagrams @xmath17 and @xmath18 yield identical operations if they are isomorphic in the sense that @xmath19\ar[dd]^\wr&\\ x\ar[ur]\ar[dr]&&y\\ & y'\ar[ur]&}\ ] ] commutes .
moreover , we can compose span diagrams : we use pullbacks : @xmath20\ar[dr]&\\x&&y}}\right ) \circ \left(\vcenter{\xymatrix@r-8pt@c-16pt{&v\ar[dl]\ar[dr]&\\y&&z}}\right ) = \left(\vcenter{\xymatrix@r-8pt@c-16pt{&u{\mathop{\times}_{y}}v\ar[dl]\ar[dr]&\\x&&z}}\right).\ ] ] it can quickly be checked that this is the right thing to do , using the formula above for @xmath14 .
these span diagrams thus form a category @xmath21 .
hence it is reasonable to believe that this encodes the structure of commutative monoids precisely .
in fact , there is a classical result that it does : a commutative monoid is the same thing as a product - preserving functor from @xmath21 to sets .
the work of lawvere generalises this point of view considerably ; given a category @xmath22 generated under taking finite products by a single object , we say that @xmath22 is an _ algebraic theory _ and that a product - preserving functor from @xmath22 to sets is a _ model of @xmath22_. we might aim to apply this to homotopy theory : badzioch @xcite has shown that the models of @xmath21 in @xmath23 are exactly the generalised eilenberg
mac lane spaces .
his paper works in ordinary category theory , but a note observes that all results carry through in the world of simplicial categories
. we might hope for a different theory : frequently , a more useful notion of commutative monoid in @xmath23 is the notion of @xmath1-monoid @xcite .
this is a monoid which is commutative only up to coherent homotopy .
so we might ask , how might we change @xmath21 in order to get this more nuanced theory ? to find an answer , we must realise that we lose valuable information when we pass to isomorphism classes of span diagrams to form the category @xmath21 .
in order to preserve this information , we use _ quasicategories _ , as developed by joyal @xcite and lurie @xcite .
a quasicategory is a simplicial set obeying some extra axioms : these axioms are slightly less stringent than the axioms for a kan complex .
the philosophy is as follows : a category can be regarded as a simplicial set , via the standard nerve construction .
but an @xmath24-groupoid ( a higher categorical object consisting of objects , invertible maps , and all higher homotopies between them ) can also be regarded as a kan complex ( consisting of points , 1-simplices and higher simplices ) , and hence also inhabits the world of simplicial sets .
a quasicategory is a generalisation of both . in particular
, it has : * objects , represented by @xmath25-simplices ; * morphisms , represented by @xmath26-simplices , which need not be invertible ( as in an ordinary category ) ; * homotopies , represented by higher simplices , which are all invertible in some appropriate sense . because of the nature of this theory , we shall use the terms `` 0-simplex '' , `` point '' and `` object '' interchangeably , and similarly also use the terms `` 1-simplex '' , `` edge '' and `` morphism '' interchangeably ( depending perhaps on whether we are thinking of the object as a simplicial set , as a model for a space , or as a generalised category ) .
likewise , we find ourselves using `` functor '' and `` map '' interchangeably , when we do not wish to be precise about whether our objects or quasicategories or more general simplicial sets .
we start with some preliminaries ( section [ 2-cats - quasicats ] ) , giving general results on quasicategories , and on their relation to lower - dimensional category theory .
we then ( in section [ theories - generalities ] ) build a quasicategory @xmath27 which is the appropriate version of @xmath21 for @xmath1-objects in a quasicategory .
as in the category @xmath21 , objects are finite sets , and morphisms are span diagrams .
however , we build in higher simplices , which consist of more elaborate span - like diagrams in finite sets .
the quasicategory @xmath27 turns out to be equivalent to a @xmath0-category : all the cells of degree @xmath28 or more are effectively uninteresting .
this greatly facilitates technical checks of its properties . after setting up @xmath27 and its basic properties
, we spend some time proving that it does exactly what we want : we compare it to the classical theory of @xmath1-algebras and also to lurie s theory of commutative monoid objects .
it is worth mentioning some of the potential consequences of this approach to universal algebra in quasicategories that are not realised in this paper .
one perfectly general comment is that the approach is , in some senses , slightly more flexible than any operad - theoretic approach ( as in @xcite ) . in particular
, an operad can only describe structures subject to axioms which do not mention any particular variable twice in the same formula : the distributive law @xmath29 is not accessible by this approach , since the right - hand side mentions the variable @xmath30 twice .
so any operad - theoretic approach to ring objects must involve some indirection . and
it is indeed possible to give a description of the algebraic theory of @xmath1-ring ( or semiring ) objects which is built out of small diagrams , after the same fashion as spans .
we postpone this to a future paper , since some considerable extra machinery is required .
such a description is handy , as it places the monoidal structures corresponding to addition and to multiplication on the same footing : both are defined by way of span diagrams .
this means that constructions which require blurring the distinction between the two , such as the construction of the group of units @xmath31 of an @xmath1 ring space , become natural .
it is also possible to give a quick description of the group - completion of a theory , and the group - completion of a model of a theory .
we similarly defer discussion of these constructions to future papers . except for some of the material in section
[ comparison - theorems ] , all the results in this paper appeared in the author s sheffield phd thesis @xcite under the direction of neil strickland .
many people are thanked in that thesis , and those sentiments are as valid as before .
the first aim of this section is to compare various notions of 2-category , in order to match jacob lurie s definition of a 2-category @xcite with the classical notions .
there are several classical notions , with varying levels of strictness and laxity : as might be expected , it is simpler to construct the laxer versions , and simpler to use the stricter versions in constructions . at one end is the notion of a weak 2-category , and at the other is the notion of a strict 2-category @xcite .
there is little essential difference , insofar as the work of street and his coauthors @xcite ( proved also in @xcite ) says that any weak 2-category can be replaced with an equivalent strict 2-category .
a strict 2-category is exactly the same thing as a category enriched in categories , and we use this identification in what follows .
the second aim is to prove some basic results on quasicategories , which will be useful later on .
we use quasicategorical terminology without apology , even for arbitrary simplicial sets .
thus a 0-cell will often be called an _ object _ of a simplicial set , and a 1-cell
will often be called a _
morphism_. accordingly , when we use the word `` space '' , we mean kan complex .
let @xmath33 be a strict 2-category with all 2-cells invertible ( that is , a category enriched in groupoids ) .
we can define its nerve in two steps .
first we form a simplicial category @xmath34 with @xmath35 , and @xmath36 .
this is a category enriched in kan complexes , and is thus suitable for the coherent nerve construction described in @xcite , giving as our final definition that @xmath37 .
it is worth expanding this definition a little .
we recall the definition of the simplicial categories @xmath38 from @xcite*1.1.5 , and are used to define @xmath39 .
this lets us prove : an @xmath40-cell in @xmath41 consists of * an @xmath42-tuple @xmath43 of objects of @xmath33 , * morphisms @xmath44 of @xmath33 for all @xmath45 , * 2-cells @xmath46 of @xmath33 for all @xmath47 , such that for any @xmath48 there is an identity on 2-cells : @xmath49 as @xmath50 , a map of simplicial categories @xmath51 certainly distinguishes objects @xmath43 .
the @xmath25-simplices of homspaces of @xmath52 correspond to subsets of the interval @xmath53 containing both @xmath54 and @xmath55 , and composition is by disjoint union .
thus they are generated under composition by the minimal subsets @xmath56 .
these give us the morphisms @xmath44 .
the @xmath26-simplices of homspaces of @xmath52 correspond to ( the opposites of ) inclusions of pairs of subsets of @xmath53 containing both @xmath54 and @xmath55 .
these are generated by inclusions @xmath57 under horizontal and vertical composition , providing the maps @xmath46 of @xmath33 .
the interchange law for horizontal and vertical composition gives us the specified identity , arising from the agreement of the composite inclusions @xmath58 this identity generates all @xmath0-cells in @xmath52 , under composition .
as @xmath59 is the nerve of a groupoid , a map @xmath52 is uniquely specified by its effect on the 1-skeleton , so there are no further data or identities .
we refer to the identity as the _ compatibility condition _ , and since 2-cells are invertible we can write it graphically : it says that the pasting of the following diagram is the identity 2-cell .
@xmath60\ar@/^16pt/[rr]_{\uparrow}\ar@/_26pt/[rrr]^{\downarrow}\ar@/^26pt/[rrr]_{\uparrow}&x_1\ar[r]\ar@/_16pt/[rr]^{\downarrow}&x_2\ar[r]&x_3}\ ] ] we also get the following basic coherence result , which is obvious from the description above . a lax functor @xmath61 between bicategories with all 2-cells invertible yields ( via passing to strict 2-categories ) a map of quasicategories @xmath62 between their nerves .
we can replace @xmath63 with an equivalent functor of strict 2-categories , and then use the naturality of the nerve construction considered above .
also , this construction agrees with the construction of the nerve of a category .
let @xmath33 be a category , regarded as a bicategory with only identities for 2-cells
. then @xmath64 is the ordinary nerve of @xmath33 .
now , the nerve @xmath64 should be thought of as a model for @xmath33 in the world of quasicategories .
thus , we should expect it to be a ( 2,1)-category in the sense discussed above .
this means that all all extensions of maps @xmath65 to @xmath66-cells are unique for @xmath67 : its cells in degrees 3 and over are determined by those in lower degrees .
the facts support our intuition : [ nerves - of - bicats ] the nerve @xmath64 is a @xmath32-category .
suppose given an inner horn inclusion @xmath68 for @xmath67 , and @xmath69 .
we can recover all the 1-cells from this data : the 1-cell @xmath70 for @xmath45 will be given by the face numbered @xmath71 for any @xmath72 . if @xmath73 , then without loss of generality , @xmath74 ( as the case @xmath75 is dealt with in a symmetric manner ) .
we then have the following diagram : @xmath60\ar@/^16pt/[rr]_{\uparrow}\ar@/_26pt/[rrr]^{\downarrow}&x_1\ar[r]\ar@/_16pt/[rr]^{\downarrow}&x_2\ar[r]&x_3}\ ] ] this leaves us just missing the 2-cell @xmath76 .
but , since all 2-cells are invertible , we can take this to be the composite of all the 2-cells in the diagram above . in symbols ,
we define @xmath77 and this clearly fulfils the compatibility condition .
this choice is clearly forced , arising as it does by solving the compatibility condition for @xmath78 , and this means the extension is unique .
if @xmath79 , then all 2-cells are determined uniquely ( indeed , @xmath80 will be defined by face @xmath71 , for any @xmath81 ) . however , if @xmath82 there are some compatibility conditions which are not forced by the faces , and we must check that they hold . for calculations , we omit the identity parts of our 2-cells . then all composites are vertical composites , so we do not bother writing the @xmath83 .
there are five compatibility conditions coming from the faces : @xmath84 also , @xmath85 and @xmath86 commute .
we can see this using the interchange law : @xmath87 for horn inclusions @xmath88 , we have all coherence conditions except the one arising from face 1 , and must show that from the others .
but we have : @xmath89 for horn inclusions @xmath90 , we have all coherence conditions except the one from face 2 .
similarly , we have : @xmath91 horn inclusions @xmath92 can be dealt with by an argument symmetric to that used for horn inclusions @xmath93 .
the fact that all structure is determined means that the extension is unique . if @xmath94 , then nothing need be checked : the compatibility conditions on @xmath95 , @xmath96 , @xmath97 and @xmath98 will be fulfilled by face @xmath71 , for any @xmath99 . using this nerve construction , in the sequel we shall abuse terminology systematically , and confuse a strict 2-category with its nerve , a @xmath32-category . in this section
we prove some properties of lurie s model for @xmath100-categories , from @xcite*subsection 2.3.4 : these are those @xmath101-categories which admit all inner horn extensions @xmath102 uniquely where @xmath103 .
it follows immediately from the definition @xcite*2.3.4.9 that an @xmath100-category has at most one extension along @xmath104 for @xmath103 ; here s a strengthening of that statement : [ partial - delta - n - cat ] an @xmath100-category @xmath33 has unique liftings for @xmath104 where @xmath105 .
we can restrict the map @xmath106 to a map @xmath107 , and lift that uniquely to a map @xmath108 .
this is the only candidate for a lifting ; we must prove that it is compatible with the given map on all of @xmath109 : that is , show that it agrees on the first face .
but these two @xmath110-cells certainly agree on the boundary of the first face ( which is isomorphic to @xmath111 ) and thus agree . in a similar vein
is this : [ outer - horn - n - cat ] an @xmath100-category @xmath33 has unique liftings for outer horns @xmath112 and @xmath113 where @xmath114 .
we can uniquely extend a map @xmath115 to a map @xmath106 using proposition [ partial - delta - n - cat ] on the @xmath25th face . then we can uniquely extend that to a map @xmath108 using proposition [ partial - delta - n - cat ] again .
the case of @xmath116 is symmetrical .
the special case of ordinary categories will be of utility later : [ nerve - outer - horns ] the nerve of a category @xmath117 has unique liftings for outer horns @xmath118 and @xmath119 whenever @xmath79 . the following proposition reduces the work necessary to show that a map of @xmath100-categories is an acyclic kan fibration : [ acyclic - kan ] a functor @xmath120 of @xmath100-categories automatically has the right lifting property with respect to the maps @xmath104 for @xmath105 .
proposition [ partial - delta - n - cat ] gives a map @xmath108 , and by @xcite*2.3.4.9 , this is consistent with the given map @xmath121 .
we can say useful things about inner fibrations .
let @xmath61 be a functor between @xmath100-categories .
we have the following simple criterion for being an inner fibration : [ inner - fibs ] the functor @xmath63 is an inner fibration if and only if it has the right lifting property for inner horns @xmath122 for @xmath123 .
in particular , this gives the following simple criterion for @xmath32-categories : @xmath63 is an inner fibration if , for every pair of diagrams @xmath124^h&\\ x\ar[ur]^f&&z } } \text{in $ { \mathcal{c}}$,\hskip 1 cm and } \vcenter { \xymatrix{&y'\ar[dr]^{h'}{\ar@{=>}}^{k'}[d]&\\ x'\ar[ur]^{f'}\ar[rr]_{g'}&&z ' } } \text{in $ { \mathcal{d}}$,}\ ] ] such that @xmath125 and @xmath126 , there is a 1-cell @xmath127 and 2-cell @xmath128 such that @xmath129 and @xmath130 .
we now switch our attention to the more intricate notion of a cartesian fibration .
these are analogues of the classical notion of a grothendieck fibration of categories .
they are morphisms of simplicial sets which describe a family of quasicategories varying in a contravariant functorial manner over a base quasicategory . following lurie @xcite
, we make the following definition : [ defn - cartesian - fibration ] a _ cartesian fibration _
@xmath131 of quasicategories is a functor which is both an inner fibration and is such that , for every 1-morphism @xmath132 ( meaning a 1-cell @xmath133 with @xmath134 and @xmath135 ) and every lift @xmath136 of @xmath137 to @xmath33 ( meaning an 0-cell @xmath138 with @xmath139 ) , there is a @xmath140-cartesian morphism @xmath141 in @xmath33 which maps to @xmath142 under @xmath140 . in turn ,
a _ @xmath140-cartesian morphism _
@xmath143 is one such that the natural map @xmath144 where @xmath145 , is an acyclic kan fibration .
there is a dual notion of a _
cocartesian morphism _ and a _ cocartesian fibration _ :
a cocartesian fibration describes a family of quasicategories varying covariantly functorially over a base quasicategory .
given @xmath131 , a cocartesian morphism in @xmath33 is a cartesian morphism for @xmath146 , and @xmath140 is a cocartesian fibration if @xmath147 is a cartesian fibration .
lurie proves that overcategories of @xmath100-categories are @xmath100-categories @xcite*lemma 1.2.17.10 . the class of @xmath100-categories is not closed under fibre products . but
the following lemma does most of the work for us : the class of simplicial sets which are the coskeleton of their @xmath66-skeleton is closed under all limits .
the @xmath40-skeleton functor @xmath148 visibly preserves limits , and the @xmath40-coskeleton functor preserves limits since it is right adjoint to @xmath148 .
given this , this category is closed under limits . indeed , more is true : for any @xmath149 , the class of simplicial sets with unique liftings for maps @xmath150 is closed under limits . the class of such simplicial sets is closed under products since a lifting for the product is just the product of liftings of the factors ; we ll verify it for fibre products too .
so if @xmath151 , @xmath152 and @xmath153 are simplicial sets with unique extensions for the map @xmath150 , then i claim that @xmath154 has unique liftings for it too .
indeed , a map @xmath155 consists of maps @xmath156 whose composites with the maps from @xmath151 and @xmath152 to @xmath153 agree .
these extend uniquely to maps @xmath157 .
their composites with the maps to @xmath153 are both extensions of our map @xmath158 .
but such extensions are unique , and so they agree .
thus these maps assemble to a unique extension @xmath159 .
these combine to prove the following : let @xmath131 be a map between @xmath100-categories .
a morphism @xmath132 in @xmath160 is @xmath140-cartesian if and only if the morphism @xmath161 has the right lifting property for all maps @xmath162 for @xmath163 . by the results above ,
both sides are @xmath100-categories ; we thus apply proposition [ acyclic - kan ] to show the higher lifting conditions are automatic . in this section
we study the relationship between joyal s over construction ( described in @xcite*lemma 1.2.9.2 ) , and limits of simplicial sets .
the result is that taking overcategories commutes with taking limits of simplicial sets , in the following sense : [ limits - and - overs ] suppose we have an ( ordinary ) finite category @xmath164 , to be thought of as a diagram category , and a diagram @xmath165 .
suppose also that we have a cone on it : a simplicial set @xmath166 and a natural transformation @xmath167 to @xmath63 from the constant functor at @xmath166 .
we then get a map @xmath168 from @xmath166 to the limit of the diagram @xmath63 .
we then have that the over construction commutes with limits in the sense that @xmath169 we consider maps from a fixed simplicial set @xmath152 ; it s then just a straightforward check : @xmath170 we continue this analysis to derive a corresponding result for finite products and cartesian morphisms : [ prods - and - carts ] if @xmath171 and @xmath172 are maps of quasicategories , then , defining @xmath173 , the @xmath174-cartesian morphisms ( as defined in subsection [ defn - cartesian - fibration ] ) are exactly the products of @xmath175-cartesian morphisms and @xmath176-cartesian morphisms
. we must relate a pullback of overcategories of products to a product of pullbacks of overcategories .
the pullbacks and products commute , as usual ; proposition [ limits - and - overs ] provides that the formation of products and of overcategories commute .
lastly , we provide a handy alternative description of overcategories . given a quasicategory @xmath33 , consider the map @xmath177 induced by evaluation of the terminal vertex of @xmath178 .
[ alternative - overcat ] the fibres @xmath179 of @xmath140 are equivalent to @xmath180 .
maps @xmath181 are maps @xmath182 sending @xmath183 to @xmath184 , or equivalently are maps @xmath185 pointed at @xmath184 .
but there is a map @xmath186 , so there is a map @xmath187 . moreover , since this map @xmath186 is a strong deformation retract .
we immediately get equivalences of homspaces .
the required result then follows from the equivalence of simplicial categories and quasicategories . in this section ,
we show how quasicategorical limits in undercategories are related to limits in the original quasicategory .
[ forgetting - from - undercat - preserves - limits ] let @xmath33 be any quasicategory with limits , and let @xmath188 be any diagram in it . the `` forgetful '' map @xmath189 preserves limits .
suppose we have a diagram of shape @xmath166 in @xmath190 .
postcomposition with the forgetful map gives a diagram @xmath191 , which admits a limit @xmath192 ; and the map @xmath193 is acyclic kan .
our diagram @xmath194 is equivalent to a diagram @xmath195 . by the acyclic kan condition , this gives us a map @xmath196 , or equivalently , @xmath197 .
we must merely show that this is indeed a limit : that @xmath198 is acyclic kan .
however , given @xmath199 a cofibration , there is a bijective correspondence between squares of the two following sorts : @xmath200\ar[r]&j\ar[d]\\ { \mathcal{c}}_{d//(1\star k)}\ar[r]&{\mathcal{c}}_{d//k } } \qquad\text{and}\qquad \xymatrix{d\star i\ar[d]\ar[r]&d\star j\ar[d]\\ { \mathcal{c}}_{/(1\star k)}\ar[r]&{\mathcal{c}}_{/k}}.\ ] ] since @xmath193 is acyclic kan , we have a diagonal filler on the right , which gives us one on the left , too . by a straightforward dualisation ,
of course we also get : [ forgetting - from - overcat - preserves - colimits ] let @xmath33 be any quasicategory and @xmath188 any diagram in it .
then the forgetful map @xmath201 preserves colimits .
lastly , we have the following useful result : [ interchange - of - limits ] if @xmath33 is a complete quasicategory , and @xmath202 is a diagram in @xmath33 , then we have the usual interchange - of - limits isomorphisms @xmath203 this follows from @xcite*prop 4.2.2.7 .
this subsection serves two purposes .
firstly , it gives a straightforward construction of homotopy pullbacks in @xmath23 .
then it gives a couple of basic properties of colimits in the quasicategory of spaces ; they are both recognisable results in the discrete case , where they reduce to results about ordinary colimits in the ordinary category of sets .
[ homotopy - pullbacks - spaces ] we use the following a natural model for the quasicategorical pullback : we write @xmath204 for the standard contractible simplicial set on three points @xmath205 , @xmath206 and @xmath207 ( with one @xmath40-simplex for each @xmath208-tuple of vertices ) . then we define @xmath209 here the morphism @xmath210 is given by evaluation on @xmath205 and @xmath207 .
we also write down the structure maps of the limiting cone : @xmath211_{p_2}\ar[dr]^f\ar[r]^{p_1}&{\mathcal{c}}_1\ar[d]\\ { \mathcal{c}}_2\ar[r]&{\mathcal{e}}.}\ ] ] the maps @xmath212 and @xmath213 are the evident projections , and @xmath142 is induced by the map @xmath214 given by evaluation at @xmath206 .
the homotopies between the composites @xmath215 are induced by the equivalences @xmath216 and @xmath217 in @xmath204 .
this is of course merely a simplicial variant of the standard topological construction of the homotopy pullback @xcite*18.1.7 .
now we turn to colimits .
we start with an easy observation : [ colims - prods - commute ] colimits in @xmath23 commute with products .
we invoke @xcite*corollary 4.2.4.8 to show that it suffices to do this in the simplicial category @xmath218 , for arbitrary coproducts and homotopy pushouts .
both of these are easy checks .
indeed , @xmath219 also , since the formation of mapping cylinders commutes with products , homotopy coequalisers do also .
this allows us to prove : [ products - and - colimits ] let @xmath220 and @xmath221 be diagrams in the quasicategory of spaces
. then the diagram @xmath222 has colimit given by @xmath223 .
this is an easy calculation , using [ colims - prods - commute ] twice : @xmath224 as discussed above , cocartesian fibrations over @xmath33 are equivalent to functors @xmath225 . in the case where @xmath226 , we shall later have need of a direct way of replacing functors between quasicategories @xmath227 with cocartesian fibrations @xmath228 .
this approach is , in fact , a special case of lurie s _ relative nerve _ construction @xcite*3.2.5 ; however it may nevertheless be helpful to have a self - contained account of it .
firstly , we note that maps @xmath229 are classified by the preimages of the two vertices of @xmath178 .
thus we write @xmath230 for a simplex with the implied map to @xmath178 sending @xmath231 to @xmath25 and @xmath232 to @xmath26 .
we define our model @xmath233 by giving that @xmath234\ar[dr]&{\mathcal{e}}\ar[d]\\&\delta^1}}\right\ } = \left\{\text{diagrams}\quad\vcenter{\xymatrix{\delta^i\ar[d]\ar[r]^i&\delta^{i\sqcup j}\ar[d]\\{\mathcal{a}}\ar[r]_f&{\mathcal{b}}}}\right\},\ ] ] where the map @xmath54 is that induced by the evident inclusion @xmath235 .
we now show by parts that this serves for us . to start with , it is straightforward to check that the preimages of the vertices are isomorphic to @xmath236 and @xmath237 respectively .
[ mapping - cylinders-1 ] the map @xmath233 is an inner fibration .
we need to show that an inner horn @xmath238 extends to a full simplex @xmath239 .
since the preimages of both vertices are quasicategories , we need only concern ourselves with the case where @xmath231 and @xmath232 are both nonempty . in either case , the faces we have include a full map @xmath240 ; this merely leaves us with an inner horn extension @xmath241 , which is possible as since @xmath237 is a quasicategory .
[ mapping - cylinders-2 ] for any element @xmath242 , there is a @xmath140-cocartesian morphism of @xmath243 whose 0th vertex is @xmath244 , and which lies over the nontrivial 1-cell of @xmath178 .
we chose the 1-cell @xmath71 given by @xmath245 ^ 0\ar[d]_a&\delta^1\ar[d]^a\\ { \mathcal{a}}\ar[r]_f&{\mathcal{b}}.}\ ] ] now we go on to show that this is indeed @xmath140-cocartesian : that the morphism @xmath246 is acyclic kan .
a diagram @xmath247\ar[r]&\delta^n\ar[d]\\ { \mathcal{e}}_{\alpha/}\ar[r]&{\mathcal{e}}_{a/}{\mathop{\times}_{{\mathcal{e}}}}{\mathcal{b}}}\ ] ] unravels to give a diagram @xmath248\ar[dr]&\\ 1\star(1\star\partial\delta^n)\ar[dr]&&1\star(\emptyset\star\delta^n)\ar[dl]\\ & { \mathcal{e}}.&}\ ] ] since these agree on the part mapping to @xmath22 , the problem is to extend a map from @xmath249 to a map from @xmath250 .
however , the edge between the first two vertices is a degeneracy , so the given map factors through @xmath251 and can be extended to @xmath250 via the map @xmath252 which collapses the first two vertices .
we can now prove : [ mapping - cylinders ] the map @xmath233 is an cocartesian fibration .
we have just shown in proposition [ mapping - cylinders-1 ] that @xmath140 is an inner fibration ; there remains the question of cocartesian lifts .
we only need to provide cocartesian lifts over nonidentity cells of @xmath178 ; and proposition [ mapping - cylinders-2 ] does this .
in this section we generalise the notion of an algebraic theory , to the setting of quasicategories .
we base our account mostly on that given by borceux @xcite for the classical case .
we will have need of the quasicategory @xmath23 . by this
we mean the quasicategory obtained as the coherent nerve of the simplicial category of kan complexes ( together with their mapping complexes ) , as is used in @xcite*1.2.16.1
. however , there are other natural constructions of equivalent quasicategories , just as there are several natural model categories quillen equivalent to the standard model structure on topological spaces .
of course it will not matter which is used .
we regard @xmath253 as being the full subquasicategory on the discrete spaces ; this is evidently equivalent to the standard notion .
we shall use the adjective _ discrete _ frequently to describe phenomena which occur over @xmath253 rather than the whole of @xmath23 .
we should also say , once and for all , what we mean by this : [ full - subquasicategory ] a _ full subquasicategory _ of a quasicategory is a maximal subquasicategory with its set of @xmath25-cells .
we shall also call this _ @xmath26-full _ ; an @xmath40-full subquasicategory is a maximal subquasicategory with that particular set of @xmath66-cells for all @xmath254 . given a quasicategory @xmath33 and a subset @xmath255
, we say that _
@xmath33 is generated by finite products from @xmath256 _ if every object of @xmath33 is equivalent to a finite product of elements in @xmath256 .
if @xmath33 is generated by finite products from @xmath257 , we refer to @xmath258 as a _ finite product generator _ of @xmath33 .
[ theory - defn ] an _ algebraic theory _ is a quasicategory @xmath22 which admits finite products and which has a finite product generator @xmath258 .
a morphism of algebraic theories @xmath259 is a product - preserving functor @xmath259 which sends generators to generators . by product - preserving , i mean `` taking finite product diagrams to finite product diagrams '' ; i suppose this is the standard meaning , but i am departing from tradition in making this plain .
it follows immediately from the definition that a morphism of theories @xmath259 is essentially surjective .
we can define a quasicategory @xmath260 of algebraic theories to be the @xmath0-full subquasicategory of @xmath261 on the theories and morphisms of theories .
theories are not much use without introducing a notion of model : a _ model _ ( in spaces ) of a theory @xmath22 is a product - preserving functor @xmath262 .
the quasicategory of models of @xmath33 is the full subquasicategory @xmath263 of @xmath264 on the product - preserving objects . in exactly the same way , if @xmath265 is any category with all finite products , we define the quasicategory of _ models of @xmath22 in @xmath265 _ to be the quasicategory @xmath266 of product - preserving functors @xmath267 . by abuse of notation
, we often write @xmath26 for a choice of generator of a theory . given a model @xmath268
, we sometimes say that @xmath269 is the _ underlying object _ of the model , and that giving such a functor @xmath268 is equipping the space @xmath269 with a _
@xmath22-structure_. we now observe that this does indeed generalise lawvere s original notion ( which is discussed in @xcite ) .
to do this , we introduce terminology for this special case : an algebraic theory @xmath22 is _ discrete _ if @xmath22 is , in fact , an ordinary category .
if a theory @xmath22 is discrete , we say that a model @xmath2 of @xmath22 is _ discrete _ if it is valued in sets ( regarded as a subcategory of spaces ) .
the rationale is that , since the subquasicategory of simplicial sets on the discrete objects is equivalent to the ordinary category of sets , if @xmath22 is an ordinary category , then a functor ( of quasicategories ) @xmath270 is just an ordinary functor .
so an algebraic theory in the sense of lawvere , which is an ordinary category @xmath22 generated by finite products of a single object , is the same thing as a discrete algebraic theory in the sense defined here .
moreover , a model of an algebraic theory in the sense of lawvere is the same thing as a discrete model of the corresponding discrete algebraic theory .
a morphism @xmath271 of theories induces a functor @xmath272 by precomposition .
[ models - in - quasicategories ] once could study models of a theory @xmath22 in quasicategories simply by using @xmath273 as defined above .
usually , we will employ an equivalent but more easily manipulated definition : a _ model of @xmath22 in quasicategories _ is a cocartesian fibration over @xmath22 that is classified by a product - preserving functor .
we call such cocartesian fibrations _ productive_. we also need to deal with maps between models : we define the quasicategory @xmath274 of models in quasicategories of a theory @xmath22 .
this is the subquasicategory of the overcategory @xmath275 , consisting of all those cells whose vertices are productive cocartesian fibrations over @xmath22 , and whose edges are product - preserving functors taking cartesian morphisms to cartesian morphisms .
a morphism @xmath271 of theories induces a functor @xmath276 induced by pulling back the cocartesian fibrations . by the results of @xcite*section 2.4.2 and chapter 3
, there is an equivalence between @xmath274 and the previously defined notion @xmath273 .
occasionally , one has need to consider axioms for algebraic structures with several underlying objects , and maps between them .
accordingly , we define : let @xmath151 be a set .
multisorted theory with sorts indexed by @xmath151 _ consists of a quasicategory @xmath22 and a map @xmath277 , such that @xmath22 is generated by finite products from the image of @xmath151 .
we refer to multisorted theories with sorts indexed by @xmath278 as being _
@xmath40-sorted theories_. a _ model _ in @xmath33 of a multisorted theory @xmath22 with sorts indexed by @xmath151 is a product - preserving functor from @xmath22 to @xmath33 . by way of trivial example ,
if @xmath279 are theories with generators @xmath280 , then the product @xmath281 is the @xmath40-sorted theory ( with generators consisting of the elements of the form @xmath282 ) whose models are tuples consisting of a model of each of the theories @xmath283 : @xmath284 in the main , the basic results for single - sorted theories carry over to @xmath40-sorted theories as one would expect , and we shall not write them out
. we can regard all the categories of multisorted theories as forming subcategories of the quasicategory @xmath285 of quasicategories with all finite products , product - preserving functors , and homotopies between them .
in particular , @xmath286 .
[ models - has - an - adjoint ] fix a quasicategory @xmath265 with finite products
. the quasifunctor @xmath287 , which assigns to each theory its category of models , has a left adjoint .
our proof proceeds by exhibiting an adjunction in detail .
however , i consider that this motivating argument is considerably more enlightening .
the idea is that @xmath288 since products are computed pointwise .
this means that @xmath289 which is exactly the equivalence on homspaces required for an adjunction . by @xcite*section 5.2 ,
an adjunction is represented by a cartesian and cocartesian fibration over @xmath178 .
now the maps @xmath229 are described by the preimages of the vertices : they are equivalent to decompositions @xmath290 where @xmath291 and @xmath292 .
so we can define a simplicial set @xmath293 over @xmath178 by giving a compatible set of homsets @xmath294 .
we define @xmath293 by letting @xmath294 consist of maps @xmath295 and @xmath296 , together with a map @xmath297 , which are equipped with a natural equivalence with @xmath298 when restricted to @xmath299 , and which send the extra point @xmath26 to @xmath300 .
writing @xmath301 for the projection @xmath302 , we easily see that @xmath303 and @xmath304 .
we must show that @xmath301 is a bicartesian fibration , to show that it represents an adjunction .
we split this into two parts .
firstly we show that @xmath301 has the inner kan lifting property : the morphism @xmath301 is an inner fibration . of claim for greater flexibility , we index our simplices by finite linearly ordered sets in this argument .
so , given finite linearly ordered sets @xmath231 and @xmath232 , and @xmath66 some internal element of the concatenation @xmath305 , we must provide a lifting @xmath306\ar[r]&\delta^{i\sqcup j}\ar[d]{\ar@{-->}}[dl]\\ { \mathcal{d}}\ar[r]&\delta^1.}\ ] ] if either @xmath231 or @xmath232 have no elements , this clearly reduces to the statement that the preimages @xmath261 and @xmath307 of the endpoints of @xmath178 are both quasicategories .
supposing otherwise , we assume without loss of generality that @xmath308 ( the case @xmath309 is symmetrical ) . observing that @xmath310 we get that a morphism @xmath311 consists of maps @xmath312 , @xmath313 , and a map @xmath314 using @xcite*3.2.2 , we see that the inclusion @xmath315 is anodyne : it s a composite of horn extensions . tracing the argument carefully ( using that @xmath66 is not the initial object of @xmath231 ) we see that no horn extensions of shape @xmath316 are required , even in the case where @xmath66 is terminal in @xmath231 .
since we are doing that extension working over @xmath300 , only inner horn extensions are needed . and
now secondly we show the existence of cartesian and cocartesian lifts .
since there is only one nontrivial 1-cell @xmath317 , we must merely show : for any object @xmath318 , there is a cartesian morphism of @xmath293 over @xmath319 with target @xmath320 ; for any object @xmath321 , there is a cocartesian morphism of @xmath293 over @xmath319 with source @xmath322 . of claim given @xmath321
, we must give a cocartesian 1-cell in @xmath293 from it which lies over the nondegenerate 1-cell of @xmath178 ; we take the cell consisting of @xmath323 , @xmath324 , and @xmath325 representing the evaluation map @xmath326 .
similarly , given @xmath318 , the cartesian 1-cell in @xmath293 consists of @xmath327 , @xmath328 , and @xmath325 representing the evaluation map @xmath329 .
the proofs that these are indeed cocartesian and cartesian respectively are very similar .
we aim to show that the morphism @xmath330 is acyclic kan .
this unravels to the requirement that we can extend two compatible maps @xmath331 and @xmath332 to a map @xmath333 , with a requirement that all the maps we supply are product - preserving . that we can do so follows immediately from the adjunction ( in the quasicategorical sense ) of the functors @xmath334 and @xmath335 for @xmath336 , and is a quick check in the case @xmath337 .
this completes the proof . as an immediate corollary
, we get [ models - colimits - to - limits ] the `` models '' functor @xmath338 takes colimits of theories to limits of their quasicategories of models .
we shall show in proposition [ theories - cocomplete ] that colimits of theories exist ; and thus this will be a helpful tool .
[ finop - is - trivial ] let @xmath265 be any quasicategory with finite products . then @xmath339 in other words ,
a product - preserving functor from @xmath340 merely picks out an object of @xmath265 .
the inclusion @xmath341 induces a functor @xmath342 .
we claim that this is an equivalence .
moreover , we claim that an inverse equivalence is given by right kan extension .
we must first check that right kan extension does indeed define a functor @xmath343 .
the right kan extension @xmath63 of @xmath344 is given by @xmath345 now the category @xmath346 is the discrete category on the set of maps @xmath347 , so this limit is the @xmath151-fold product of @xmath348 .
hence this kan extension is indeed product - preserving .
morever , by this analysis , any extension of @xmath349 to a product - preserving functor @xmath350 is a right kan extension .
but there is a contractible space of such extensions .
hence the functor @xmath231 has all its preimages contractible , and is thus an equivalence .
these categories of models have good properties : [ models - limits ] if @xmath33 is a theory , then the quasicategory @xmath263 is complete , with limits computed pointwise .
the quasicategory @xmath351 is complete , with limits computed pointwise . by proposition
[ interchange - of - limits ] showing that limits can be interchanged , the limit of a diagram from @xmath263 is again in @xmath263 , and is thus the limit in @xmath263 .
[ models - functor - limits ] given a morphism of theories @xmath271 , the pullback functor @xmath352 preserves limits .
the pullback functor @xmath353 evidently preserves limits , since they re defined pointwise .
the result follows , since limits in @xmath263 are just limits in @xmath264 ( and the same for @xmath300 ) , and this pullback functor restricts to our desired one .
we recall from @xcite*section 5.3.1 the notion of a filtered simplicial set .
this is equivalent for having liftings for all maps @xmath354 , where @xmath320 is the nerve of a finite poset .
a filtered colimit is then just a colimit on a filtered diagram .
the category @xmath263 has filtered colimits , which are computed pointwise .
this is the same argument as [ models - limits ] , using @xcite*prop 5.3.3.3 , saying that filtered colimits commute with limits .
now we wish to study push - forwards of models , showing that taking left kan extensions provides a left adjoint to the pullback functor .
this will require some work ; we subdivide it into two major parts .
we show that this is plausible : [ left - kan - preserves - prods ] given a morphism @xmath271 of theories , and a model @xmath355 , the left kan extension of @xmath356 along @xmath142 preserves products and is thus a model of @xmath300 .
the left kan extension is given by @xmath357 we must show that @xmath358 and @xmath359 . in both cases
we show that there is a natural map from the colimit diagrams which define each side , which is cofinal ( in the sense of joyal , written up by lurie @xcite*4.1 ) , and thus there is an equivalence between them . in the first case
, we have @xmath360 it is easy to see that the inclusion of the terminal object @xmath361 into @xmath22 is cofinal . indeed , by joyal s characterisation of cofinal maps @xcite*4.1.3.1
, we must show that @xmath362 is weakly contractible .
this is clear : it has an initial object @xmath26 .
thus @xmath363 induces an isomorphism of colimits .
this terminal object is sent to @xmath364 and thence to @xmath365 .
this proves the first case . in the second case , @xmath366 is given by the colimit @xmath367 there is a functor @xmath368 which sends @xmath369 in the evident way .
according to @xcite*4.1.3.1 , to show this map is cofinal we need to show that , for any @xmath370 , the simplicial set @xmath371 is weakly contractible .
this simplicial set is isomorphic to @xmath372 which is the quasicategory of pairs @xmath373 equipped with maps @xmath374 , and 2-cells @xmath375 . but
this quasicategory has an evident terminal object @xmath376 and @xmath377 , which makes it weakly contractible .
and the colimit of @xmath378 is indeed @xmath379 , by proposition [ products - and - colimits ] .
this completes the proof .
now we can finish the job : [ models - pushforward ] given a morphism @xmath271 of theories , the pullback functor @xmath352 has a left adjoint @xmath380 , given by left kan extension .
we could do this simply by restricting the standard adjunction between @xmath351 and @xmath381 given by composition and left kan extension .
however , we build an adjunction by hand to make more of the structure visible
. first , we use @xmath382 to define a cocartesian fibration @xmath383 , as described in subsection [ mapping - cylinders - quasicategories ] ( it is a cocartesian fibration , as proved in proposition [ mapping - cylinders ] ) .
we need to show that it is also cartesian , so it represents an adjunction .
we have observed it to be an inner fibration already ( in proposition [ mapping - cylinders-1 ] ; we just need to demonstrate the existence of cartesian lifts for edges .
the simplicial set @xmath178 only has one degenerate 1-cell ; it is only over that cell that the problem is not vacuous .
given @xmath384 , we take a left kan extension of @xmath320 along @xmath142 , given by @xmath385 where we identify objects of @xmath22 and of @xmath300 for brevity .
this is product - preserving by proposition [ left - kan - preserves - prods ] .
our cartesian lift @xmath71 shall have this as its zero vertex , so we must exhibit a morphism @xmath386 .
this is provided by the universal property of the colimit .
we must now show that this 1-cell @xmath71 from @xmath387 to @xmath320 is cartesian .
that means showing that the projection @xmath388 is acyclic kan .
unpacking the definitions , this morphism is the evident projection @xmath389 we can show that this is acyclic kan by working pointwise and using the acyclic kan condition of the colimit .
note that this gives us a notion of a free model @xmath390 of a theory @xmath22 on a space @xmath151 : a space can be viewed as a model of the initial theory @xmath340 , and we can use the push - forward associated to the initial morphism of theories @xmath391 . a theory is said to be _ pointed _ if it has a zero object : an object @xmath25 which is both initial and terminal .
this is standard categorical terminology , and is also justified by the following proposition : if @xmath22 is a pointed theory , the terminal model @xmath26 ( the model given by the constant @xmath26 functor ) is a zero object in the category of models . in particular , any model @xmath267 factors through @xmath392 . limits and colimits in @xmath393 are computed pointwise , so @xmath26 is initial and terminal there .
the quasicategory @xmath394 is just a @xmath26-full subcategory , so @xmath26 is still a zero object
. it might be useful to have a quasicategory of pointed theories : this is the @xmath26-full subquasicategory of @xmath260 whose objects are the pointed theories .
[ finsop - property ] let @xmath265 be any quasicategory with finite products .
then there is an equivalence @xmath395 between product - preserving functors from @xmath396 and @xmath265 and the quasicategory of objects of @xmath265 equipped with maps from the terminal object .
the technique of proof is very similar to that of proposition [ finop - is - trivial ] .
we obtain a functor @xmath397 by a slightly contorted process .
first , choose a terminal object @xmath398 .
consider the morphism @xmath399 given by the map of pointed sets @xmath400 .
this map induces a functor @xmath401 .
since any product - preserving functor @xmath402 sends @xmath403 to a terminal object , we can modify @xmath231 by an equivalence to land in the full subcategory of @xmath404 on those objects sending the terminal edge to @xmath26 . by proposition [ alternative - overcat ] , we know that this is equivalent to @xmath392 . combining these constructions gives us the required functor @xmath397 . as in proposition
[ finop - is - trivial ] , it will suffice to show that any product - preserving functors @xmath402 are right kan extensions of their restrictions along @xmath405 to @xmath178 .
let @xmath406 be a morphism @xmath407 in @xmath265 .
since our map @xmath405 is a full subcategory inclusion , we can calculate the right kan extension @xmath356 of @xmath63 along @xmath405 as the limit @xmath408 the category @xmath409 consists of @xmath151 discrete points ( corresponding to the maps @xmath410 for each element of @xmath151 ) and a copy of @xmath178 ( corresponding to the zero maps @xmath410 and @xmath411 ) .
thus the limit is @xmath412 , as required . in cases where the terminal object is also initial
, this says that every object extends to a map from @xmath396 .
suppose @xmath265 has finite products and a zero object .
then @xmath413 this allows us to give a valuable structure result : [ finsop - is - universal ] the theory @xmath396 is the initial pointed theory .
moreover , a pointed theory can be regarded as a quasicategory with finite products @xmath22 and an essentially surjective , product - preserving functor @xmath414 . in this section
we show that the quasicategory of theories is complete ( proposition [ theories - complete ] ) , which is straightforward , and that it is cocomplete ( proposition [ theories - cocomplete ] ) , which is much harder . in order to prove the latter result
, we introduce a good deal of machinery .
intrinsic in this machinery is the ability to take the free theory on some fairly general collection of data , but we apply it only to take the free theory on a diagram consisting of other theories . thus we anticipate that the methods introduced here could be used to prove other theorems of this general character .
[ theories - complete ] the quasicategory of theories is complete . an @xmath231-shaped diagram in theories yields an underlying diagram @xmath415 .
we consider also the functors from the initial theory @xmath340 ; for each @xmath54 , we write @xmath416 for the map @xmath417 .
this is classified by a cartesian fibration @xmath418 .
lurie s model , which we recall from @xcite*3.3.3 , for the limit of this diagram ( in @xmath261 ) is the quasicategory of cartesian sections @xmath419 ( that is , the quasicategory of sections which take 1-cells to cartesian 1-cells ) .
we consider the full subquasicategory of this on the objects @xmath420 for which there is a finite set @xmath320 such that @xmath421 , that is , those which act diagonally on objects .
any cone over @xmath63 in @xmath260 acts diagonally on the objects , up to equivalence , since the maps commute with the structure maps .
hence the universal property of the product in @xmath261 gives us a universal property for this subobject in @xmath260 .
now we turn our attention to showing that theories have all colimits .
this will require some technical work , and we build up to the proof slowly .
the plan is as follows : lurie has proved that the quasicategory of quasicategories is cocomplete .
thus , for any simplicial set @xmath164 , the colimit @xmath422 of a diagram in theories factors uniquely through the colimit @xmath423 in the quasicategory of quasicategories .
indeed , we should expect it to be the universal quasicategory with a functor from @xmath423 such that the images of all the product cones in elements of @xmath164 are product cones .
consider the quasicategory @xmath424 of quasicategories with a map from @xmath425 .
we are interested in the full subcategory @xmath426 with objects the quasicategorical limit cones @xmath427 .
our first step is this : [ inclusion - of - limits - into - cones ] the inclusion functor @xmath428 preserves all limits . by @xcite*4.4.2.6 , it suffices to show it preserves all products and pullbacks . in this proof
we write @xmath429 for the terminal simplicial set , to avoid overuse of the symbol @xmath26 .
given a set of quasicategories and maps @xmath430 , all of them limit cones , then the diagonal map @xmath431 can easily be shown to be a limit cone .
now , we have to deal with pullbacks of quasicategories ; we recall the setup of definition [ homotopy - pullbacks - spaces ] . now , suppose we have limit cones @xmath432 .
we then have a diagonal map @xmath433 , and must show that this too is a limit cone .
suppose we have a cofibration @xmath199 ; we must show that there are liftings @xmath434\ar[d]&j\ar[d]{\ar@{-->}}[dl]\\ { \mathcal{c}}_{/(*\star d)}\ar[r]&{\mathcal{c}}_{/d},}\ ] ] or equivalently that there are extensions @xmath435\ar[dr]&(j\star * \star d){\ar@{-->}}[d]\\ & { \mathcal{c}},}\ ] ] provided that the restruction to @xmath436 is the given cone . unravelling using the definition of @xmath33 , we are demanding extensions @xmath437\ar[dd]\ar[dr]&&{\mathcal{c}}_1\ar[dd]\\ & j\star * \star d{\ar@{-->}}[ur]\ar[dd]&\\ ( ( e2\times i)\star * \cup(e2\times j)\star\emptyset)\star d\ar[rr]\ar[dr]&&{\mathcal{e}}\\ & ( e2\times j)\star * \star d{\ar@{-->}}[ur]&\\ ( i\star * \cup j\star\emptyset)\star d\ar[rr]\ar[dr]\ar[uu]&&{\mathcal{c}}_2\ar[uu]\\ & j\star * \star d{\ar@{-->}}[ur]\ar[uu]&}\ ] ] we can extend the top and bottom without difficulty , using that the maps @xmath438 are limit cones .
this leaves us with an extension problem @xmath439\ar[dr]&(e2\times j)\star * \star d{\ar@{-->}}[d]\\ & { \mathcal{e}}}\ ] ] which is readily checked to be a right lifting against a cofibration , and so follows from the fact that @xmath440 is a limit cone .
this completes the proof .
now we consider the diagram @xmath441^f\ar[dr]_{\sim}&({{\mathrm{cat}}_\infty})_{(1\star d)/}\ar[d]\\ & ( { { \mathrm{cat}}_\infty})_{d/}.}\ ] ] the diagonal map is an acyclic kan fibration , since every diagram naturally has a contractible space of limits .
note that @xmath261 is a presentable category @xcite*remark 1.2.11 , and @xcite*5.5.3.11 shows that undercategories of presentable categories are presentable .
thus all the categories in the diagram are presentable .
also , the proof of proposition [ inclusion - of - limits - into - cones ] demonstrates that colimits in @xmath442 and @xmath424 are computed in @xmath261 , and thus ( using that the diagonal map is an equivalence ) all three functors preserve colimits .
accordingly , we can apply lurie s adjoint functor theorem @xcite*5.5.2.9 to show the following : [ forcing - a - limit ] the functor @xmath443 admits a left and a right adjoint .
a straightforward consequence of the existence of a left adjoint is that , for every quasicategory @xmath33 and map @xmath444 , there is a universal quasicategory @xmath445 such that the composite @xmath446 is a limit cone , in the sense that @xmath447 is acyclic kan .
indeed , the morphism @xmath445 is just the unit of the adjunction . now , suppose we have a diagram @xmath448 .
we will construct a colimit .
firstly , the extension @xmath449 has a colimit @xmath450 .
transfinitely enumerate the finite product diagrams as @xmath451 . with this notation
, we prove the result we were aiming for : [ theories - cocomplete ] the quasicategory of theories is cocomplete .
we provide a colimit for any diagram @xmath452 .
firstly , we can obtain from our diagram @xmath63 a diagram @xmath453 , sending @xmath26 to the initial theory @xmath340 .
we take the colimit of that , using @xcite*3.3.4 .
we claim that the resulting colimit cocone has essentially surjective structure maps . indeed ,
any object in @xmath454 is in the essential image of @xmath455 for some @xmath456 : the structure maps are jointly essentially surjective .
( to prove this , it is quick to verify that the essential image of @xmath457 in @xmath454 satisfies the colimit property , and is thus all of it ) .
however , any @xmath458 is the essential image of some @xmath459 .
also , there is an equivalence ( induced by the image of the 2-cell @xmath460 in @xmath261 ) between the image of @xmath244 and the image of @xmath320 in @xmath454 .
so every structure map has the same essential .
we start with @xmath461 .
we can transfinitely enumerate the finite product diagrams in the quasicategory @xmath340 as @xmath462 for some ordinal @xmath463 , where @xmath464 is a discrete simplicial set , and @xmath465 .
we choose to do this with redundancy : we want each individual product diagram to appear infinitely many times and be cofinal in @xmath463
. one straightforward way to ensure this is to enumerate them without repetition with ordertype @xmath466 , then take @xmath467 and repeat our list @xmath468-many times .
our aim is to produce @xmath469 by starting from @xmath470 and extending all the maps @xmath471 to limit cones .
proposition [ forcing - a - limit ] supplies us with a quasicategory @xmath472 and a unit map @xmath348 such that the composite @xmath473 is a limit cone .
similarly , we produce @xmath474 from @xmath472 by using the adjunction to provide a quasicategory from which the map from @xmath475 is a limit cone
. then we proceed by transfinite induction , extending to limit ordinals by taking the filtered colimits ( in @xmath261 ) of the preceding quasicategories : @xmath476 the resulting quasicategory @xmath477 is our colimit .
we have several checks to make to show this to be the case .
firstly , it is necessary to show we havent enlarged our quasicategory in an unacceptable manner : all the structure maps @xmath478 for @xmath479 are essentially surjective . of claim it suffices to show both that the successor maps @xmath480 are essentially surjective , and also that a colimit of shape @xmath468 of essentially surjective maps is essentially surjective . a similar argument works for both .
we can show for each that a failure to be essentially surjective would violate the universal property : that the image of the morphism would provide a smaller object with the same property . indeed ,
if @xmath480 was not essentially surjective , the image @xmath481 would result in the nontrivial factorisation @xmath482 and the left - hand map can easily be checked to be a product cone .
this contradicts the universal property of the adjunction .
similarly , if a colimit @xmath483 of essentially surjective maps of quasicategories is not essentially surjective , then the essential image factors the structure maps of the colimit nontrivially , which contradicts the universal property of the colimit .
secondly , we need to show it is indeed a theory , and that the structure maps we ve defined are maps of theories .
we ve done essential surjectivity already , so we just need the following : the defined maps @xmath484 preserve all finite products . of claim given a limit cone @xmath485
, we must show that the composite @xmath486 is a limit cone too . given a lifting problem for a cofibration @xmath487 as follows : @xmath488\ar[d]&\delta^n\ar[d]{\ar@{-->}}[dl]\\ ( x_\kappa)_{/(1\star d)}\ar[r]&(x_\kappa)_{/d},}\ ] ]
we rewrite it as @xmath489\ar[dr]&\delta^n\star 1\star d{\ar@{-->}}[d]\\ & x_\kappa.}\ ] ] since the simplicial set @xmath490 is finite , the map from it to @xmath477 factors through some @xmath491 for which @xmath492 is a product cone ( since , by construction , the set of such ordinals @xmath466 is cofinal in @xmath463 ) .
the required extension exists in that @xmath491 and thus also in @xmath477 .
lastly , of course , we need to verify the universal property of a colimit . @xmath477 is universal among theories under @xmath63 . of claim
we need to show that the functor @xmath493 is acyclic kan .
suppose given a cofibration @xmath199 ; we have a lifting problem @xmath434\ar[d]&j\ar[d]\\ \operatorname{theories}_{(f\star 1)/}\ar[r]&\operatorname{theories}_{f/}.}\ ] ] it suffices to consider cofibrations @xmath487 .
we consider the @xmath337 and @xmath336 separately .
if @xmath337 , our cofibration is @xmath494 : we have a cone @xmath495 describing a theory @xmath22 under @xmath63 ; the aim is to factor it through @xmath477 .
since @xmath470 is the colimit of @xmath63 in @xmath261 , we have a diagram @xmath496 factoring our cone through @xmath470 .
working under @xmath63 , since the maps @xmath497 are product - preserving , we can factorise this successively through the @xmath491 to get an essentially surjective , product - preserving map @xmath498 under @xmath63 as required .
now , in case @xmath336 , we have compatible functors @xmath499 and @xmath500 , with the middle @xmath26 sent to @xmath477 .
equivalently , this is a diagram @xmath501 and we need to extend it to @xmath502 .
since @xmath470 is the colimit of @xmath63 , we can extend this the underlying diagram @xmath503 to a diagram @xmath504 , with the middle @xmath26 sent to @xmath470 . using the universal property of the adjunction and the colimiting property
, we can extend this to a map @xmath505 , where @xmath506 denotes the nerve of the ordinal @xmath507 viewed as a poset ( that is , as the poset of ordinals less than or equal to @xmath463 ) , and where the ordinal @xmath466 is sent to @xmath491 .
the terminal vertex of @xmath506 and the initial vertex of @xmath251 are both sent to @xmath477 . moreover , by construction , they are identical under @xmath63 , and so by construction the edge between them is the identity . restriction to @xmath508 thus gives us the required diagram in @xmath261 ; since all the edges were present already and were morphisms of @xmath260 , this is also a diagram in @xmath260 .
this completes the proof .
we can do similar things with this method : [ cinftypp - cocomplete ] the quasicategory @xmath285 of quasicategories with all finite products , and product - preserving functors between them , has colimits .
the proof of the preceding proposition generalises in a straightforward manner . since our diagrams are no longer cones of essentially surjective maps under @xmath340 , we need to consider products in all the diagrams and force their images to all be products ( whereas before it sufficed to consider only those in @xmath340 ) .
we no longer need to ensure essential surjectivity , but we do however need to provide limits for all the new objects introduced .
so we intersperse the operations which force cones to be product cones with operations that adjoin new products for the objects ( using the methods of @xcite*5.3.6 ) .
free models for a theory are shown to exist by proposition [ models - pushforward ] ; this section records a more explicit , less involved construction of free models on finite sets .
let @xmath22 be a theory .
we suppose given a functorial model @xmath509 for the homspaces in @xmath22 .
such models are shown to exist and are discussed further in @xcite*section 1.2.2 .
the map @xmath391 from the initial theory is equivalent to a functor @xmath510 , and we can compose this with the homspace functor to get a map @xmath511 in other words , we take @xmath512 . the functor @xmath513 is product - preserving since @xmath514 is , so we actually get a functor @xmath515 this behaves as we would hope : the functor @xmath513 is indeed the free model of @xmath22 on @xmath151 . for any model @xmath320 of @xmath22
, we have a natural equivalence @xmath516 this is a straightforward exercise using the quasicategorical yoneda lemma of @xcite*section 5.1 .
in particular , this agrees with the more general construction of proposition [ models - pushforward ] .
we can also prove : a theory @xmath22 is equivalent to the opposite of the full subquasicategory of @xmath263 on the free models on finite sets .
the yoneda embedding used above is full and faithful ; and the functor is evidently essentially surjective on objects .
throughout this subsection we assume given a canonical , functorial choice of pullbacks of finite sets .
we introduce a bicategory @xmath517 of span diagrams .
a 0-cell of @xmath517 is a finite set .
a 1-cell from @xmath470 to @xmath472 is a span diagram @xmath518 of finite sets .
a 2-cell between diagrams @xmath518 and @xmath519 is an isomorphism @xmath520 fitting into a diagram as follows : @xmath521\ar[dr]\ar[dd]_f^{\wr}&\\ x_0&&x_1\\ & y'.\ar[ul]\ar[ur]}\ ] ] 2-cells compose in the obvious way ; 1-cells compose by taking pullbacks : the composite of @xmath522 and @xmath523 is given by @xmath524 , where @xmath525 is the following pullback : @xmath526\ar[dr]&&\\ & x_{01}\ar[dl]\ar[dr]&&x_{12}\ar[dl]\ar[dr]&\\ x_0&&x_1&&x_2.}\ ] ] it is a simple exercise to show that this gives a bicategory .
we will later have cause to use a generalisation of this notion .
given a category @xmath33 which has all pullbacks , and a functorial choice of pullbacks , we can define the bicategory @xmath527 of spans in @xmath33 : 0-cells are objects of @xmath33 , 1-cells are span diagrams in @xmath33 , and 2-cells are isomorphisms of spans .
so , in this notation , our category @xmath517 is @xmath528 .
we work with the weak 2-category of spans @xmath527 , where @xmath33 is any category with pullbacks .
first we prove a more - or - less standard lemma of ordinary category theory : [ pb - se ] pullbacks of split epimorphisms are split epimorphisms .
suppose given a diagram @xmath529^k \ar[d]_h&b\ar[d]^g\\ c{\ar@{-{>>}}}[r]_f & d,}\ ] ] where the bottom morphism @xmath142 is a split epimorphism : a morphism such that there is @xmath530 with @xmath531 .
this affords us a map @xmath532 .
now , we have @xmath533 , and so , by the definition of the pullback , there is a map @xmath534 with @xmath535 , as required .
this allows us to prove an important structural result for span categories : [ span - isos ] objects @xmath536 are equivalent if and only if they are isomorphic as objects of @xmath33
. given two isomorphic objects in @xmath33 , any span of isomorphisms between them forms an equivalence in @xmath527 .
the data of an equivalence consists of 1-cells @xmath537 and @xmath538 , fitting into diagrams @xmath539^p\ar[dr]_q\ar@/_2ex/[ddll]_{=}\ar@/^2ex/[ddrr]^{=}&&\\ & u\ar[dl]^{a}\ar[dr]^{b}&&v\ar[dl]_{c}\ar[dr]_{d}&\\ x&&y&&x,}\ ] ] @xmath540^r\ar[dr]_s\ar@/_2ex/[ddll]_{=}\ar@/^2ex/[ddrr]^{=}&&\\ & v\ar[dl]^{c}\ar[dr]^{d}&&u\ar[dl]_{a}\ar[dr]_{b}&\\ y&&x&&y.}\ ] ] the maps @xmath244 , @xmath541 , @xmath30 and @xmath542 are split epimorphisms ( by inspecting the left and right composites in each diagram ) . but
this means that @xmath140 , @xmath174 , @xmath207 and @xmath543 are also split epimorphisms , by proposition [ pb - se ] .
however @xmath140 , @xmath174 , @xmath207 and @xmath543 are also split monomorphisms ( by inspection of the left and right composites ) , and thus isomorphisms ( since if a morphism is left and right invertible , the inverses agree ) .
this clearly means that @xmath244 , @xmath541 , @xmath30 and @xmath542 are isomorphisms , and thus that @xmath151 and @xmath152 are isomorphic .
continuing the analysis , we have the following proposition : [ span - auts ] let @xmath33 be a category .
the group of isomorphism classes of automorphisms of @xmath151 in @xmath527 is isomorphic to @xmath544 . by [ span - isos ]
, any automorphism of @xmath151 in @xmath527 looks like @xmath545 but such a span diagram is uniquely isomorphic to exactly one of the form @xmath546 this proves the claim .
we now define a quasicategory @xmath27 , one of the principal objects of study of this thesis , which is isomorphic to the nerve of the bicategory @xmath517 .
define @xmath547 to be the poset of nonempty subintervals @xmath548 in @xmath549=(0,\ldots , n)$ ] , equipped with the _ reverse _ inclusion ordering .
we regard @xmath547 as a category .
the poset of nonempty subintervals of a totally ordered set is a functorial construction , so the collection @xmath550 forms a cosimplicial object in categories as we vary over all finite totally ordered sets @xmath551 .
this enables us to define a simplicial set , which we shall soon prove ( in proposition [ span - structure ] ) to be a quasicategory : let @xmath33 be an ordinary category with pullbacks .
we define the _ span quasicategory _
@xmath552 to be the simplicial set whose @xmath40-cells @xmath553 are the collection of functors @xmath554 from @xmath547 to the category @xmath33 , with the condition that , if @xmath231 and @xmath232 are two nonempty intervals in @xmath549 $ ] with nonempty intersection , the diagram @xmath555\ar[d]&f(i)\ar[d]\\ f(j)\ar[r]&f(i\cap j)}\ ] ] is a pullback .
we refer to this condition later as the _ pullback property_. the collection @xmath552 is indeed a simplicial set , since @xmath322 is a cosimplicial category , and taking faces and degeneracies preserves the pullback property .
if @xmath556 is an @xmath40-cell of the quasicategory @xmath552 , then we will write @xmath557 for @xmath558 and @xmath559 for @xmath560 . if @xmath561 , then we write @xmath562 for the structure map @xmath563 induced by the inclusion .
now , we have our formal statement : [ span - structure ] suppose @xmath33 is any ordinary category with pullbacks .
then we have an isomorphism of simplicial sets @xmath564 .
thus @xmath552 is a @xmath32-category and in particular ( as suggested in the definition above ) a quasicategory .
we refer back to section [ two - one - categories ] for notation on bicategories .
given an @xmath40-cell @xmath565 , we associate an @xmath40-cell @xmath566 .
we take @xmath567 for all @xmath54 .
further , we take @xmath557 to be the middle part of the span 1-cell given by @xmath568 , so we have a diagram @xmath569 for all @xmath45 .
what is more , the 2-cell @xmath570 gives us a diagram as follows : @xmath571\ar[dddrr]\ar[d]^{\wr}&&\\ & & \bullet\pb{270}\ar[dl]\ar[dr]&&\\ & x_{ij}\ar[dl]\ar[dr]&&x_{jk}\ar[dl]\ar[dr]&\\ x_i&&x_j&&x_k,}\ ] ] for every @xmath47 .
however , such diagrams are in 1 - 1 correspondence with diagrams @xmath572\ar[dr]&&\\ & x_{ij}\ar[dl]\ar[dr]&&x_{jk}\ar[dl]\ar[dr]&\\ x_i&&x_j&&x_k,}\ ] ] where the composite @xmath573 is the given span diagram for @xmath574 .
the compatibility condition gives all the other pullbacks , and the functoriality of the maps @xmath575 and @xmath576 .
this construction is reversible ( and naturally commutes with faces and degeneracies ) so we get an isomorphism of simplicial sets . in a similar fashion ,
there are functors @xmath577 and @xmath578 , for any category @xmath33 .
they are evidently faithful , and according to proposition [ span - isos ] , if @xmath579 or @xmath580 is an equivalence then @xmath142 is an isomorphism .
we move on to considering products in the quasicategory @xmath27 .
the quasicategory @xmath27 has finite products .
the product of objects @xmath320 and @xmath581 is @xmath582 . recall the definition of limits in quasicategories : if @xmath583 is a morphism of simplicial sets , then a limit of @xmath142 is a terminal object of the over - category @xmath584 , given by @xmath585 for us , @xmath586 , with @xmath587 and @xmath588 .
thus @xmath589
we now specify the object @xmath590 of @xmath591 which we claim is the product : it consists of the object @xmath592 , with projection maps @xmath593 and @xmath594 . we need to show that it is a strongly final object in @xmath584 .
this means showing that any diagram @xmath595 with @xmath596 extends to a diagram @xmath597 .
this will be a straightforward , but notationally heavy , check .
define @xmath598 to be the poset of subintervals of @xmath599 that do not contain all of @xmath600 ( with the reverse inclusion order ) .
restricting to @xmath600 , we recover the poset @xmath601 of proper subintervals of @xmath600 defined in section [ span - quasicategory ] . the simplicial structure on @xmath602 guarantees that maps @xmath603 assemble to form diagrams @xmath604 with the pullback property , such that @xmath605 , @xmath606 , @xmath607 and @xmath608 . for example ,
if @xmath73 the diagram @xmath151 is as follows : @xmath609\ar[dr]&&&\\ & & x(0,2)\ar[dl]\ar[dr]&&x(1,3)\ar[dl]\ar[dr]&&x(2,4)\ar[dl]\ar[dr]&&\\ & x(0,1)\ar[dl]\ar[dr]&&x(1,2)\ar[dl]\ar[dr]&&x(2,3)\ar[dl]\ar[dr]&&a\ar[dl]\ar[dr]&\\ x(0)&&x(1)&&x(2)&&a\sqcup b&&a,}\ ] ] and the diagram @xmath152 is similar .
we can extend these to @xmath610 by defining @xmath611 we clearly have @xmath612 , have @xmath613 and @xmath614 by definition , and it is quick to check the pullback property . the same proof suffices to prove the following : [ prod - span - c ] for any category @xmath33 with finite coproducts and finite limits , finite products in @xmath552 exist , and are given on objects by coproducts in @xmath33 .
the inclusion maps are defined analogously to the case @xmath615 above . as an important corollary
, we have : the functor @xmath616 makes the category @xmath27 into an algebraic theory , as introduced in definition [ theory - defn ] . accordingly , since @xmath27 was motivated by the desire to produce a quasicategorical version of the theory of monoids , we define : [ lawvere - monoid - object ] let @xmath33 be a quasicategory with finite products .
a _ ( lawvere ) monoid object _ in @xmath33 is a model of @xmath27 in @xmath33 : a product - preserving functor @xmath617 .
also , since @xmath27 is self - opposite , we have the category @xmath552 has coproducts , which agree with products .
an immediate consequence of this is that the theory @xmath27 is pointed , as defined in section [ pointed - theories ] ; various important consequences of this are given there too . as we have seen , the category @xmath396 is equivalent to the quasicategory of spans whose right arm is a monomorphism .
proposition [ finsop - property ] tells us that models of @xmath396 are in some sense pointed objects .
so @xmath396 carries the part of the theory of monoids dealing with the unit , but not the product .
if a theory is required which has an associative but nonunital product , the natural choice is the quasicategory of spans whose right arm is an epimorphism ; this ensures that products are only taken over nonempty sets .
now we introduce a category @xmath618 . using the notation of subsection [ span - quasicategory ] , we define @xmath619 , where @xmath620 is the category of arrows in @xmath621 .
so an @xmath40-cell of @xmath618 is a pair of span diagrams @xmath622 with maps @xmath623 , which commute with all the structure maps .
equivalently , it s a natural transformation between functors @xmath624 , where both @xmath151 and @xmath152 have the pullback property .
there s a 2-functor @xmath625 coming from the functor @xmath626 which sends @xmath627 to @xmath152 . according to the description above ,
this sends a morphism of span diagrams to the codomain .
now , we want to study this functor .
first we find a good supply of @xmath140-cartesian morphisms ( as introduced in definition [ defn - cartesian - fibration ] above ) .
[ spant - span - cartesian - morphisms ] any 1-cell of @xmath618 of the form @xmath628&y_{01}\ar[l]_{=}\ar[r]\ar[d]\pb{315}&y_1\ar[d]\\
x_0 & x_{01}\ar[l ] \ar[r ] & x_1.}\ ] ] ie . which has the top left map the identity and right - hand square a pullback , is @xmath140-cartesian .
by proposition [ acyclic - kan ] , there are four checks to make on the functor @xmath629 to show that it is an acyclic kan fibration : we must check it has the right lifting property with respect to @xmath104 for @xmath630 .
we are using the notation @xmath137 for the 0-cell of @xmath618 given by @xmath631 .
firstly , we show the existence of liftings for @xmath632 . given a diagram like the following ,
which represents a @xmath25-cell of @xmath633 , @xmath634\ar[drdr]\ar[d]&&\\ & & x_{24}\pb{270}\ar[dl]\ar[dr]&&\\ y_2\ar[d]&x_{23}\ar[dl]\ar[dr]&&x_{34}\ar[dl]\ar[dr]&y_4\ar[d]\\ x_2&&x_3&&x_4,}\ ] ] we can fill it in to form a full 0-cell of @xmath635 as follows : @xmath636{\ar@{-->}}[dl]^{=}\ar@<2ex>[drdr]{\ar@{-->}}[dr]\ar[d]&&\\ & y_{24}{\ar@{-->}}[d]{\ar@{-->}}[dl]{\ar@{-->}}[dr]&x_{24}\pb{270}\ar[dl]\ar[dr]&y_{34}{\ar@{-->}}[d]{\ar@{-->}}[dl]^>>>{=}{\ar@{-->}}[dr]&\\ y_2\ar[d]&x_{23}\ar[dl]\ar[dr]&y_{34}{\ar@{-->}}[d]&x_{34}\ar[dl]\ar[dr]&y_4\ar[d]\\ x_2&&x_3&&x_4,}\ ] ] and this is the required lifting .
next , a diagram @xmath637\ar[d]&{{\mathrm{span}}^\times}_{/f}\ar[d]\\ \delta^1\ar[r]&({{\mathrm{span}}^\times}_{/y}){\mathop{\times}_{{\mathrm{span}}_{/py}}}({\mathrm{span}}_{/pf})}\ ] ] gives us a configuration of @xmath152 s as follows : @xmath638\ar[dr]\ar@<2ex>[drdr]\pb{270}&&&\\ & & y_{13}\ar[dl]\ar@<2ex>[drdr]&&y_{24}\ar[dl]\ar[dr]\pb{270}&&\\ & y_{12}\ar[dl]\ar[dr]&&y_{23}\ar[dl]\ar[dr]&&y_{34}\ar[dl]_{=}\ar[dr]&\\ y_1&&y_2&&y_3&&y_4,}\ ] ] where all squares commute and are pullbacks .
there is also a full diagram of @xmath639 s , and maps @xmath640 .
the parallel morphisms @xmath641 do not have to agree _ prima facie _ , but the composites @xmath642 do agree .
this maps to a complete span diagram of @xmath151 s in the obvious way . however , since @xmath643 is a pullback , the parallel morphisms into it do commute ( since the two composites into @xmath644 and @xmath645 do agree ) .
the maps @xmath646 and @xmath647 are isomorphisms , since they re pullbacks of an isomorphism .
this allows us to define a map @xmath648 , which makes the resulting top and left squares into pullbacks .
finally , the resulting parallel pair of morphisms @xmath649 agree , since they are isomorphic to the pair considered earlier .
now we brace ourselves and consider liftings for @xmath650 . here
the morphism @xmath651 gives us a diagram like @xmath652\ar[dr]\ar@<2ex>[drdr]\pb{270}&&&&\\ & & & y_{03}\ar[dl]\ar[dr]\ar@<2ex>[drdr]\pb{270}&&y_{14}\ar[dl]\ar[dr]\pb{270}&&&\\ & & y_{02}\ar[dl]\ar[dr]\ar@<2ex>[drdr]\pb{270}&&y_{13}\ar[dl]\ar[dr]\pb{270}&&y_{24}\ar[dl]\ar[dr]\pb{270}&&\\ & y_{01}\ar[dl]\ar[dr]&&y_{12}\ar[dl]\ar[dr]&&y_{23}\ar[dl]\ar[dr]&&y_{34}\ar[dl]\ar[dr]&\\ y_0&&y_1&&y_2&&y_3&&y_4}\ ] ] here all squares are pullbacks , but it is not given that the parallel pairs agree .
however , the morphism @xmath653 gives us exactly this necessary extra coherence data , completing this check .
lastly , it is straightforward to check that , given a lifting problem for @xmath654 , all data is given and is coherent : we get a complete span diagram .
[ span - cart - fib ] the map @xmath625 is a cartesian fibration .
firstly , we show that the map is an inner fibration . by proposition [ inner - fibs ] , we need only check horn extensions for @xmath655 .
this gives us the following diagram : @xmath656\ar[dr]\ar[d]&y_{02}\ar[dl]\ar[dr]\pb{270}&x_{12}\ar[dl]\ar[dr]\ar[d]&\\ x_0\ar[d]&y_{01}\ar[dl]\ar[dr]&x_1\ar[d]&y_{12}\ar[dl]\ar[dr]&x_2\ar[d]\\ y_0&&y_1&&y_2.}\ ] ] this can be filled in to a full map of span diagrams by taking @xmath525 to be the pullback of @xmath657 ; this maps to @xmath658 in an appropriate manner .
given a 1-cell @xmath659 of @xmath27 ( the numbering will make sense later ) and an 0-cell @xmath631 of @xmath618 , we need to find a @xmath140-cartesian morphism of @xmath618 which restricts to these two .
but we can define @xmath643 to form a 1-cell of @xmath618 as follows : @xmath660&y_{34}\ar[l]_{=}\ar[r]\ar[d]\pb{315}&y_4\ar[d]\\ x_3 & x_{34}\ar[l ] \ar[r ] & x_4.}\ ] ] this is @xmath140-cartesian by proposition [ spant - span - cartesian - morphisms ] above .
this construction is compatible with the construction by lurie @xcite*notation 2.4.1.2 of the cartesian fibration @xmath661 , in the following sense : there is a commuting diagram @xmath662^{l^\times}\ar[d]&{{\mathrm{span}}^\times}\ar[d]\\ { { { { \mathrm{finset}}}_*}}\ar[r]_{l } & { \mathrm{span}}.}\ ] ] in this section we classify all morphisms which are @xmath140-cartesian , where @xmath625 is the natural projection map . for convenience of notation , we will work with the equivalent notion in the opposite categories : classifying @xmath147-cocartesian morphisms where @xmath147 is the corresponding morphism @xmath664 . in the proof of proposition [ span - cart - fib ]
, we showed that a 1-cell @xmath665 given by @xmath666&x_{01}\ar[l]_{\lambda^x_{01}}\ar[r]^{\rho^x_{01}}\ar[d]&x_1\ar[d]\\ y_0 & y_{01}\ar[l]^{\lambda^y_{01}}\ar[r]_{\rho^y_{01 } } & y_1}\ ] ] is @xmath147-cocartesian if the morphism @xmath667 is an isomorphism , and if the right - hand square is a pullback square .
we write @xmath668 for @xmath669 .
the argument depends on the diagrams used in the proof of proposition [ span - cart - fib ] .
we will take to drawing the bottom part of a span upside - down : this will simplify the diagrams in practice .
[ cart - pb - surj ] if @xmath63 is @xmath147-cocartesian , then the natural map @xmath670 is surjective .
given an element @xmath671 , the solid arrows of the following diagram describe a cell @xmath672 : @xmath673_x{\ar@{-->}}[dl]\ar[ddrr]^=\ar@/^1.7ex/[ddddd]^y&&\\ & x_{01}\ar[ddd]\ar[dl]\ar[dr]&&&\\ x_0\ar[d]&&x_1\ar[d]&&1\ar[d]\\ y_0&&y_1&&1\\ & y_{01}\ar[ul]\ar[ur]&&y_1\ar[ul]\ar[ur]&\\ & & y_{01}\ar[ul]\ar[ur]&&}\ ] ] we are assuming that an extension to a cell @xmath674 exists ; this provides us with the dotted arrow @xmath675 : an element of @xmath676 which maps to @xmath677 .
this proves surjectivity .
[ cart - topmap - surj ] if @xmath63 is @xmath147-cocartesian , then the map @xmath678 is surjective .
suppose this is not the case : that @xmath679 has no preimage in @xmath676 .
we consider a lifting problem for @xmath680 along @xmath681 .
the data of such a situation is specified by solid arrows of the following diagram : @xmath682\ar[dr]&&&\\ & & 0\ar[dl]\ar[dr]&&1\ar@/^1.7ex/[ddll]{\ar@{-->}}[dl]\ar[dr]&&\\ & x_{01}\ar[dl]\ar[dr]\ar[ddd]&&0\ar[dl]\ar[dr]&&1\ar[dl]\ar[dr]&\\ x_0\ar[d]&&x_1\ar[d]&&1\ar[d]&&1\ar[d]\\ y_0&&y_1&&1&&1\\ & y_{01}\ar[ul]\ar[ur]&&y_1\ar[ul]\ar[ur]&&1\ar[ul]\ar[ur]&\\ & & y_{01}\ar[ul]\ar[ur]&&y_1\ar[ul]\ar[ur]&&\\ & & & y_{01}.\ar[ul]\ar[ur]&&&}\ ] ] by hypothesis , all the squares in each half are pullbacks .
since @xmath63 is assumed to be @xmath147-cocartesian , an extension exists along the dotted line : a contradiction .
[ cart - topmap - inj ] if @xmath63 is @xmath147-cocartesian , then the map @xmath678 is injective .
suppose not : that there is @xmath683 with @xmath684 a set of size at least 2 .
then there is a nontrivial automorphism @xmath71 of @xmath590 .
we now consider the following lifting problem for @xmath680 along @xmath685 , where @xmath54 is the inclusion @xmath686 , and the top parallel collection of morphisms need not commute : @xmath687^\alpha\ar@/_1.7ex/[ddll]_i\ar[dr]&&&\\ & & p\ar[dl]^i\ar[dr]&&1\ar@/^1.7ex/[ddll]^(.75){x}\ar[dr]&&\\ & x_{01}\ar[dl]\ar[dr]\ar[ddd]&&1\ar[dl]_x\ar[dr]&&1\ar[dl]\ar[dr]&\\ x_0\ar[d]&&x_1\ar[d]&&1\ar[d]&&1\ar[d]\\ y_0&&y_1&&1&&1\\ & y_{01}\ar[ul]\ar[ur]&&y_1\ar[ul]\ar[ur]&&1\ar[ul]\ar[ur]&\\ & & y_{01}\ar[ul]\ar[ur]&&y_1\ar[ul]\ar[ur]&&\\ & & & y_{01}.\ar[ul]\ar[ur]&&&}\ ] ] again , all squares are pullbacks . by assumption
this lifts to a complete diagram @xmath688 , meaning that @xmath689 , meaning that @xmath71 is trivial : a contradiction .
[ cart - pb - inj ] if @xmath63 is @xmath147-cocartesian , then the natural map @xmath690 is injective .
let @xmath677 be any element of @xmath691 , and let @xmath692 be two elements of the preimage .
we consider another lifting problem for @xmath680 along @xmath685 , where again the top parallel collection of morphisms need not commute : @xmath693\ar@/_1.7ex/[ddll]_{a'}\ar[dr]&&&\\ & & 1\ar[dl]^a\ar[dr]&&1\ar@/^1.7ex/[ddll]\ar[dr]&&\\ & x_{01}\ar[dl]\ar[dr]\ar[ddd]&&1\ar[dl]\ar[dr]&&1\ar[dl]\ar[dr]&\\ x_0\ar[d]&&x_1\ar[d]&&1\ar[d]&&1\ar[d]\\ y_0&&y_1&&1&&1\\ & y_{01}\ar[ul]\ar[ur]&&y_1\ar[ul]\ar[ur]&&1\ar[ul]\ar[ur]&\\ & & y_{01}\ar[ul]\ar[ur]&&y_1\ar[ul]\ar[ur]&&\\ & & & y_{01}.\ar[ul]\ar[ur]&&&}\ ] ] the fact that appropriate pullback squares exist follows from propositions [ cart - topmap - surj ] and [ cart - topmap - inj ] . since @xmath63 is assumed @xmath147-cocartesian , the lifting gives us that @xmath694 .
if @xmath63 is given by @xmath666&x_{01}\ar[l]_{\lambda^x_{01}}\ar[r]^{\rho^x_{01}}\ar[d]&x_1\ar[d]\\ y_0 & y_{01}\ar[l]^{\lambda^y_{01}}\ar[r]_{\rho^y_{01 } } & y_1,}\ ] ] then it is @xmath140-cartesian if and only if the right - hand square is a pullback and @xmath667 is an isomorphism .
one direction is proposition [ span - cart - fib ] , the other is jointly implied by propositions [ cart - pb - surj ] , [ cart - topmap - surj ] , [ cart - topmap - inj ] , and [ cart - pb - inj ] . we note that we have not used the lifting condition for @xmath650 , and deduce that it is automatically satisfied in the presence of the others : this is apparently not otherwise clear . given a quasicategory @xmath33 with cartesian products
, we shall produce a model of @xmath27 in quasicategories ( as defined in subsection [ models - in - quasicategories ] ) .
first we define an auxiliary category @xmath695 . for @xmath696 , we define @xmath695 to be the simplicial set represented by the following functor in @xmath166 : @xmath697 ( it is straightforward to check that this functor does indeed preserve colimits . ) this has the following important structural property : the projection @xmath698 is a cocartesian fibration .
we have @xmath699 the map @xmath625 was shown to be a cartesian fibration in proposition [ span - cart - fib ] .
since the map @xmath700 is evidently a cocartesian fibration , the projection @xmath701 is also a cocartesian fibration ( by @xcite , 2.3.2.3 ) .
these two maps satisfy the hypotheses for @xmath140 and @xmath174 respectively in @xcite*lemma 3.2.2.13 , and so the proposition is proved .
we can describe the fibre @xmath702 of @xmath695 over a finite set @xmath703 : [ fibre - description ] @xmath704 we have : @xmath705\ar[dr]_a&&{\tilde{{\mathcal{c}}}^\times}\ar[dl]\\&{\mathrm{span}}&}}\right\}\\ & = \operatorname{hom}(\delta^n{\mathop{\times}_{{\mathrm{span}}}}{{\mathrm{span}}^\times},{\mathcal{c}})\\ & = \operatorname{hom}(\delta^n\times{\mathrm{span}}({{\mathrm{finset}}}_{/a}),{\mathcal{c}}),\end{aligned}\ ] ] and so @xmath702 can be identified with the simplicial set of functors , from the category @xmath706 of spans of finite sets over @xmath320 , into @xmath33 .
but since @xmath707 , we get : @xmath708 this description allows us to analyse the @xmath709-cocartesian morphisms in @xmath695 : let @xmath71 be a morphism in @xmath695 , with image the 1-cell @xmath710 in @xmath27
. then @xmath71 is @xmath709-cocartesian if and only if , for every @xmath711 , the morphism @xmath71 takes the 1-cell @xmath712&g^*u\ar[l]\ar[d]\ar[r]&u\ar[d]\\ x & z \ar[l ] \ar[r]&y}\ ] ] of @xmath713 to an equivalence in @xmath33 .
first , we study the morphisms which are cocartesian with respect to the projection functor @xmath701 .
but proposition [ prods - and - carts ] makes it clear that these are the products of the morphisms which are cocartesian for @xmath700 and those which are cocartesian for the identity on @xmath27 . by @xcite*remark 2.3.1.4 ,
the former morphisms are the equivalences in @xmath33 , and the latter are all the morphisms in @xmath27 .
so the @xmath174-cocartesian morphisms are those which are an equivalence on the left factor . with these preliminaries
, the result follows by invoking @xcite*lemma 3.2.2.13 . since proposition [ fibre - description ] describes the fibre of @xmath714 over an object @xmath715 as the category of functors @xmath716 .
we can thus define @xmath717 to be the full subcategory whose objects are all the product - preserving functors @xmath716 for all @xmath703 .
we define @xmath718 to be the restriction of @xmath709 to @xmath717 , and begin to amass good properties of this functor .
[ cocart - from - prods ] the projection @xmath718 is a cocartesian fibration , with the same cocartesian morphisms as @xmath719 . the map @xmath140 is evidently an inner fibration , since it s a restriction of an inner fibration to a full subcategory .
so we just need to demonstrate that , given a span @xmath720 a 1-cell in @xmath27 , and an element @xmath63 of the fibre @xmath721 , there is cocartesian lift of @xmath71 in @xmath695 , whose left - hand vertex @xmath356 is an element of the fibre @xmath722 .
but , in terms of the description of the fibre we are given , if @xmath723 is product - preserving , then the vertex of the natural lift is given by @xmath724 which is clearly a product - preserving functor @xmath725 . [ lifting - prop ] assume that @xmath33 has finite products .
the projection @xmath718 has the property that , given a coproduct diagram in @xmath27 , consisting of @xmath726 and @xmath727 , the corresponding functors realise @xmath728 as the product of @xmath729 and @xmath730 .
the fibre @xmath731 over a finite set @xmath320 is the quasicategory of product - preserving functors @xmath716 . if @xmath33 has products , this is isomorphic to @xmath732 .
it is quick to check that the morphisms given realise the product structures correctly .
motivated by the above propositions , we make the following definition ( recalling the definition of a cocartesian fibration from @xcite*section 2.4 ) : a _ lawvere symmetric monoidal structure _ is a model of @xmath27 in quasicategories : a cocartesian fibration @xmath733 such that preimages of product diagrams in @xmath27 are product diagrams of categories .
so , the propositions above assemble to : an quasicategory with finite products gives a lawvere symmetric monoidal structure .
this is propositions [ cocart - from - prods ] and [ lifting - prop ] .
we need notions of functor too : a _ symmetric monoidal functor _ between two lawvere symmetric monoidal structures @xmath734 and @xmath735 is a map of models of @xmath27 in quasicategories : that is , a functor taking @xmath140-cocartesian morphisms to @xmath174-cocartesian morphisms .
a _ lax symmetric monoidal functor _ is one which takes @xmath140-cocartesian morphisms whose image in @xmath27 is collapsing to @xmath174-cocartesian morphisms .
now we seek to define an algebra object in a lawvere symmetric monoidal category .
we say that an 1-cell in the category @xmath27 is _ collapsing _ if it is of the form @xmath736\ar[r]^\sim&y},\ ] ] or equivalently if it s isomorphic to a diagram of the form @xmath737\ar[r]^=&a}.\ ] ] we can thus define a subquasicategory @xmath738 of @xmath27 containing all objects , all collapsing 1-cells and all higher cells all of whose edges are collapsing .
we define a morphism in @xmath739 to be collapsing if all the preimage of every element except the basepoint has size exactly one . by the same process
we can define a subquasicategory @xmath740 , and it is quick to check we have a pullback diagram @xmath741\ar[r]&{{\mathrm{span}}^{\textrm{coll}}}\ar[d]\\{{{{\mathrm{finset}}}_*}}\ar[r]_r&{\mathrm{span}}.}\ ] ] in other words : a morphism in @xmath739 is collapsing if and only if it yields a collapsing span diagram .
we define an algebra object in a style analogous to those found in @xcite : a _ lawvere commutative algebra object _ of a lawvere symmetric monoidal category @xmath742 consists of a section @xmath142 of @xmath140 which takes collapsing morphisms to @xmath140-cocartesian morphisms .
we can then define the quasicategory of lawvere commutative algebra objects @xmath743 to be the full subquasicategory of @xmath744 on the lawvere commutative algebra objects . when the monoidal structure is understood , which is most of the time , we write @xmath745 instead .
the purpose of this section is to show that lurie s definition of a symmetric monoidal category in @xcite is equivalent to the one advanced in the preceding section .
we begin with some reasonably lightweight sections , which sketch straightforward simple arguments why @xmath27 should be thought of as the theory for commutative monoids in the setting of quasicategories .
the purpose of this is to compare our theory @xmath27 ( both as a quasicategory , and via its models ) with more traditional ways of defining @xmath1 structures .
our comparison , ultimately , shall be with the barratt - eccles operad .
let @xmath746 be an operad in simplicial sets , and @xmath151 be a finite set .
the free @xmath746-algebra on @xmath151 , denoted @xmath747 $ ] has underlying space given by @xmath748| = \coprod_{n\geq0}({\mathcal{o}}(n)\times_{\sigma_n}x^n ) = \operatorname{colim}_{a\in{{{{\mathrm{finset}}}^{{\operatorname{\cong}}}}}}({\mathcal{o}}(a)\times x^a).\ ] ] thus @xmath749|^b$ ] is given by @xmath748|^b = \operatorname{colim}_{(a\stackrel{f}{\rightarrow}b)\in({{\mathrm{finset}}}_{/b})^{{\operatorname{\cong}}}}({\mathcal{o}}(f)\times x^a),\ ] ] where the colimit is taken in the category whose objects are finite sets with maps to @xmath581 , and whose morphisms are isomorphisms over @xmath581
. moreover , by the free - forgetful adjunction proved in the proposition above , a map @xmath753\rightarrow{\mathcal{o}}[x]$ ] is the same as a map of spaces @xmath754|$ ] , and so is equivalent to a span diagram @xmath755_f\ar[dr]^g&\\ x&&y}\ ] ] together with a point of @xmath756 .
the correspondence works as follows : taking preimages of elements in @xmath152 give a @xmath152-indexed family of sets @xmath757 , equipped with maps @xmath758 and elements of @xmath759 .
more is true : an @xmath40-simplex @xmath760,{\mathcal{o}}[x])_n$ ] is equivalent to a span diagram @xmath755_f\ar[dr]^g&\\ x&&y}\ ] ] together with an element of @xmath761 .
indeed , the various contributions from different span diagrams are disconnected from one another .
* objects of @xmath762 are finite sets . *
the homspace @xmath763 is the mapping space @xmath764,{\mathcal{o}}[x])$ ] , the full sub - kan - complex of @xmath765|,|{\mathcal{o}}[x]|)$ ] whose @xmath25-simplices are maps preserving operadic structure .
we define @xmath772 to be @xmath773 for any finite set @xmath151 . now , suppose given finite sets @xmath151 and @xmath152 and an @xmath40-simplex @xmath774 , which is the same as a map @xmath775|$ ] .
we must define an induced map @xmath776 . by our construction of @xmath749|$ ] as a disjoint union above , for each finite set @xmath153 we need a @xmath153-equivariant map @xmath779 however , such a map exists : we define it by first using the composition map @xmath780 , and then using the @xmath746-algebra structure on @xmath320 .
equivalently , regarding a map @xmath753\rightarrow{\mathcal{o}}[x]$ ] as a span diagram @xmath755_f\ar[dr]^g&\\ x&&y,}\ ] ] together with an element of @xmath756 , we find the corresponding map @xmath781 is the composition of the pullback map @xmath782 and an application of the operadic action map @xmath783 .
it is necessary to check that this assignment is indeed functorial : if we have finite sets @xmath151 , @xmath152 and @xmath153 , we must check that the diagram @xmath784,{\mathcal{o}}[y])\times\operatorname{map}({\mathcal{o}}[y],{\mathcal{o}}[x])\ar[r]\ar[d]&\operatorname{map}({\mathcal{o}}[z],{\mathcal{o}}[x])\ar[d]\\ \operatorname{map}(a^y , a^z)\times\operatorname{map}(a^x , a^y)\ar[r]&\operatorname{map}(a^x , a^z)}\ ] ] commutes . for 0-cells corresponding to span diagrams @xmath785_f\ar[dr]^g&\\ x&&y , } }
\qquad\text{and}\qquad \vcenter{\xymatrix{&v\ar[dl]_{f'}\ar[dr]^{g'}&\\ y&&z,}}\ ] ] it s immediate to check that both composites correspond to the composite span @xmath786_{f''}\ar[dr]^{g''}&&\\ & u\ar[dl]_f\ar[dr]^g&&v\ar[dl]_{f'}\ar[dr]^{g'}&\\ x&&y&&z,}\ ] ] with given element of @xmath787 obtained by pulling back the element of @xmath756 to obtain an element of @xmath788 and then using the operadic compition with the given element of @xmath789 .
another small piece of propaganda is provided by calculating the homspaces in the quasicategory @xmath27 ; this makes plausible much of the relationship between @xmath27 and various recognisably classical notions such as the barratt - eccles operad @xcite .
one should think of @xmath27 as a homotopy - theoretic elaboration of the theory of commutative monoids .
indeed , the theory of commutative monoids is the opposite of the full subcategory of @xmath796 on the objects @xmath797 , and there is a natural forgetful morphism @xmath798 .
writing @xmath803 for the category of finite sets and isomorphisms , we have @xmath804 ( where the sum on the second line is taken over a set of representatives of the isomorphism classes ) this is a famous model for the free @xmath1-monoid on one generator : it s that provided by the barratt - eccles operad .
moreover , we also have @xmath810 so we can think of the space of spans from @xmath151 to @xmath152 as being @xmath151-by-@xmath152 matrices with entries in the free @xmath1-monoid on one generator .
composition is then matrix multiplication , and it is easily seen that coproducts and products are given on homspaces by block sums of matrices .
we would like to be able to compare the weak @xmath0-categories @xmath814 and the groupoid - enriched category @xmath815 .
but this requires more machinery than we wish to develop .
so we do some hard work , and compare @xmath27 and @xmath816 themselves .
now , an element of @xmath819 consists of a morphism @xmath820\rightarrow{\mathrm{th}_{{\mathcal{e}}}}^\delta$ ] , in other words two finite sets @xmath152 and @xmath151 , and a point of @xmath821 , a point of @xmath822 by forgetting the @xmath823 ( which has only a single @xmath25-simplex anyway ) , we get a span diagram @xmath824 . an element @xmath825 gives us three such span diagrams by restricting to the three natural inclusions @xmath820\rightarrow{\mathfrak{c}}[2]$ ] , agreeing on objects thus : @xmath826\ar@/^3ex/[ddrr]&&\\ & a\ar[dl]\ar[dr]&&b\ar[dl]\ar[dr]&\\ x&&y&&z.}\ ] ] we also obtain a homotopy between the composite of the two short spans and the long one ; this gives an isomorphism from @xmath322 to some model of the pullback @xmath827 and thus provides a @xmath0-cell of @xmath27 . from a @xmath28-cell of
@xmath816 we get four @xmath0-cells of @xmath27 , which assemble to form a diagram as follows , @xmath828\ar[dr]\ar@/_5ex/[ddll]^{+}\ar@/^5ex/[ddrr]_{+}&&&\\ & & d\ar[dl]\ar[dr]\ar@/^5ex/[ddrr]&{+}&e\ar[dl]\ar[dr]\ar@/_5ex/[ddll]&&\\ & a\ar[dl]\ar[dr]&&b\ar[dl]\ar[dr]&&c\ar[dl]\ar[dr]&\\ w&&x&&y&&z}\ ] ] where everything is known to commute except the punctured polygons
. however , the map @xmath829\rightarrow{\mathrm{th}_{{\mathcal{e}}}}^\delta$ ] also provides a map @xmath830 , giving a commuting square @xmath831\ar[d]&d{\mathop{\times}_{y}}c\ar[d]\\ a{\mathop{\times}_{x}}e\ar[r]&a{\mathop{\times}_{x}}b{\mathop{\times}_{y}}c,}\ ] ] which makes the diagram commute as required .
now we must go on to show that it is an equivalence .
but this is not difficult : we shall show ( using the philosophy of @xcite*section 1.1.3 ) that @xmath818 is an equivalence since it s the identity on objects and induces weak equivalences on homspaces .
our calculation in subsection [ homs - in - span ] above , and the definition of @xmath816 show that both have homspaces @xmath832 these homspaces are classifying spaces of groupoids ; thus the definition of @xmath818 gives a weak equivalence immediately .
lurie shows @xcite*1.2.3.1 that the nerve functor ( of ordinary categories ) from ordinary categories to quasicategories has a right adjoint , denoted @xmath833 , the _
homotopy category_. we use this theory briefly to understand how @xmath27 really is a quasicategorical version of the algebraic theory of commutative monoids . since , as mentioned above , @xmath833 is the right adjoint of the nerve functor @xmath835 , for any quasicategory @xmath33 all product - preserving functors @xmath617 factor uniquely through the product - preserving functor @xmath836 .
we define @xmath838 to be the pullback
@xmath839\ar[d]&{{\mathcal{c}}^\times}\ar[d]^p\\ { { { { \mathrm{finset}}}_*}}\ar[r]_l & { \mathrm{span}}}\ ] ] the left - hand arrow is a cartesian fibration , by @xcite*lemma 2.4.2.3 .
the required product property is immediate , since unions in @xmath739 get sent to product diagrams in @xmath27 .
similarly , commutative algebra objects for @xmath717 yield commutative algebra objects in @xmath838 : given a section @xmath244 of @xmath733 , we can pull back to obtain a section of @xmath837 : @xmath841\ar[d]\pb{315}&{\mathrm{span}}\ar[d]\\ { { \mathcal{c}}^\otimes}\ar[r]\ar[d]\pb{315}&{{\mathcal{c}}^\times}\ar[d]\\ { { { { \mathrm{finset}}}_*}}\ar[r ] & { \mathrm{span}}.}\ ] ] this defines a functor @xmath842 .
similarly , for the present argument we define a _
lurie monoid object _ in @xmath33 to be a functor @xmath843 which takes the projection maps of disjoint unions to product diagrams in @xmath33 .
( this agrees with the definition in @xcite*remark 2.4.2.2 ) .
we can form quasicategories @xmath844 and @xmath845 of lawvere and lurie monoid objects respectively , as full subquasicategories of the functor categories @xmath846 and @xmath847 on the monoid objects .
the quasicategory of cocartesian fibrations over @xmath739 is equivalent to the quasicategory of functors @xmath852 ; and the disjoint union property for cocartesian fibrations is equivalent to taking disjoint unions to products . we can draw cells of @xmath856 as span diagrams equipped with a fenced - off sub - span diagram on the right - hand side , where the fenced - off part is in the essential image of @xmath858 : @xmath859\ar[dl]&&&&\\ & & & x_{03}\pb{270}\ar[dr]\ar[dl]&&x_{14}\pb{270}\ar[dr]\ar[dl]&{\ar@{--}}[dddlll]&&\\ & & x_{02}\pb{270}\ar[dr]\ar[dl]&&x_{13}\pb{270}\ar[dr]\ar[dl]&&x_{24}\pb{270}\ar[dr]{\ar@{^{(}->}}[dl]&&\\ & x_{01}\ar[dr]\ar[dl]&&x_{12}\ar[dr]\ar[dl]&&x_{23}\ar[dr]{\ar@{^{(}->}}[dl]&&x_{34}\ar[dr]{\ar@{^{(}->}}[dl]&\\ x_0&&x_1&&x_2&&x_3&&x_4}\ ] ] the full subcategory @xmath863 on the objects corresponding to @xmath864 consists of the contributions by @xmath865 , and this is a copy of @xmath27 ; the full subcategory @xmath866 on the objects corresponding to @xmath867 consists of the contributions by @xmath868 , and this is a copy of @xmath739 .
a. [ m - equivs ] for any set @xmath320 , the canonical 1-cell of @xmath856 defined by the constant maps @xmath873 and @xmath874 is sent by @xmath142 to an equivalence in @xmath33 , b. [ m - lawvere ] the restriction @xmath875 is a lawvere monoid object .
c. [ m - lurie ]
the restriction @xmath876 is a lurie monoid object we notice that , in the presence of condition ( [ m - equivs ] ) , conditions ( [ m - lawvere ] ) and ( [ m - lurie ] ) are equivalent to one another . indeed , both the sets of collapsing morphisms and the product properties match up under the given equivalences . indeed , to say that the following diagram is a kan extension diagram @xmath878^{f|_{{\mathcal{j}}_{\mathrm{span}}}}\ar[d]&{\mathcal{c}}\\ { \mathcal{j}}\ar[ur]^f&}\ ] ] is to say that , for every object @xmath879 , the diagram @xmath880^{f}\ar[d]&{\mathcal{c}}\\ ( { \mathcal{j}}_{\mathrm{span}})_{/x}\star 1\ar[ur]^f&}\ ] ] makes @xmath881 a colimit of @xmath882 .
here @xmath883 is notation for @xmath884 .
we can show that every map @xmath889 admits @xmath890 as a left kan extension to a map @xmath891 .
indeed , following the definition of a kan extension along an inclusion , this amounts to showing for @xmath892 that the diagram @xmath893\ar[rr]&&{\mathrm{span}}\ar[r]^{f_0}&{\mathcal{c}}\\ 1\star{\mathrm{span}}_{/x}\ar[r]&{\mathcal{j}}\ar[r]_{{\mathcal{l}}}&{\mathrm{span}}\ar[ur]_{f_0}&}\ ] ] is a colimit diagram of shape @xmath894 in @xmath33 .
however , @xmath894 has a terminal object given by the diagram of identities : @xmath792\ar[dr]&{\ar@{--}}[dl]\\ x&&x}\ ] ] thus the colimit is given by @xmath895 , with colimiting structure maps described by @xmath896 .
a. [ mak - a ] every @xmath902 admits a right kan extension @xmath142 , as shown fitting into the following diagram : @xmath841^{f_0}\ar[d]&{\mathcal{c}}\\ { \mathcal{j}}\ar[ur]_f&}\ ] ] b. [ mak - b ] given @xmath903 such that @xmath904 is a lurie monoid object , @xmath142 is a right kan extension of @xmath905 if and only if @xmath142 satisfies condition ( [ m - equivs ] ) above .
since there is an injection @xmath910 , there is an injection @xmath911 ; we thus write the objects of @xmath912 as morphisms @xmath913 of @xmath27 .
we let @xmath914 denote the full subcategory on the objects where @xmath541 is an isomorphism and @xmath244 an injection .
the inclusion @xmath915 has a right adjoint .
indeed , one choice of right adjoint sends the object @xmath913 to @xmath916 .
regarding an adjunction as a bicartesian fibration over @xmath178 , we need to provide cartesian lifts of the nontrivial 1-cell in @xmath178 ; a lift for @xmath917 is given by @xmath918\ar[dr]^=\pb{270}&&\\ & \operatorname{im}(a){\ar@{^{(}->}}[dl]\ar[dr]^=&&y\ar[dl]\ar[dr]&\\ k&&\operatorname{im}(a)&&z;}\ ] ] it is readily checked that this is indeed cartesian .
thus , using @xcite*lemma 4.3.2.7 , we merely need to show that @xmath925 has a limit in @xmath33 .
but by the product property , @xmath142 exhibits @xmath926 as a limit of @xmath925 , thus proving ( [ mak - a ] ) . a. [ d - equivs ] for any set @xmath320 , the canonical 1-cell of @xmath856 defined by the constant maps @xmath873 and @xmath874 is taken to an equivalence in @xmath717 , b. [ d - lawvere ] the restriction @xmath875 is a lawvere algebra object .
c. [ d - lurie ]
the restriction @xmath876 is a lurie algebra object , in the sense that the following diagram factors with a dotted arrow , as shown , to give one : @xmath932\ar[r]^\sim&{{{{\mathrm{finset}}}_*}}{\ar@{-->}}[r]&{{\mathcal{c}}^\otimes}\ar[d]\\ { \mathcal{j}}\ar[rr]&&{{\mathcal{c}}^\times}.}\ ] ] we notice that , in the presence of condition ( [ d - equivs ] ) , conditions ( [ d - lawvere ] ) and ( [ d - lurie ] ) are equivalent to one another . indeed , both the sets of collapsing morphisms and the product properties match up under the given equivalences . indeed , to say that the following diagram is a kan extension diagram @xmath878^{f|_{{\mathcal{j}}_{\mathrm{span}}}}\ar[d]&{{\mathcal{c}}^\times}\ar[d]^q\\ { \mathcal{j}}\ar[r]\ar[ur]^f&{\mathrm{span}}}\ ] ] is to say that , for every object @xmath879 , the diagram @xmath880^{f}\ar[d]&{{\mathcal{c}}^\times}\ar[d]^q\\ ( { \mathcal{j}}_{\mathrm{span}})_{/x}\star 1\ar[r]\ar[ur]^f&{\mathrm{span}}}\ ] ] makes @xmath881 a @xmath174-colimit of @xmath882 .
here @xmath883 is notation for @xmath884 .
a. [ ak - a ] every @xmath938 admits a @xmath174-kan extension @xmath142 , as shown fitting into the following diagram : @xmath841^{f_0}\ar[d]&{{\mathcal{c}}^\otimes}\ar[r]&{{\mathcal{c}}^\times}\ar[d]\\ { \mathcal{j}}\ar[rr]\ar[urr]_f&&{\mathrm{span}}}\ ] ] b. [ ak - b ] given @xmath939 such that @xmath904 is a lurie algebra object , @xmath142 is a @xmath174-right kan extension of @xmath905 if and only if @xmath142 satisfies condition ( [ d - equivs ] ) above .
since there is an injection @xmath910 , there is an injection @xmath911 ; we thus write the objects of @xmath912 as morphisms @xmath913 of @xmath27 .
we let @xmath914 denote the full subcategory on the objects where @xmath541 is an isomorphism and @xmath244 an injection . thus , using @xcite*lemma 4.3.2.7 , we merely need to show that @xmath925 has a @xmath174-limit in @xmath717 . but @xmath142 exhibits @xmath926 as a @xmath174-limit of @xmath925 , and that proves ( [ ak - a ] ) . | we adapt the notion of an algebraic theory to work in the setting of quasicategories developed recently by joyal and lurie .
we develop the general theory at some length .
we study one extended example in detail : the theory of commutative monoids ( which turns out to be essentially just a @xmath0-category ) .
this gives a straightforward , combinatorially explicit , and instructive notion of a commutative monoid .
we prove that this definition is equivalent ( in appropriate senses ) both to the classical concept of an @xmath1-monoid and to lurie s concept of a commutative algebra object . |
consider a uniform rectilinear displacement of a finite fluid slice of viscosity @xmath0 by another fluid of viscosity @xmath1 in 2d porous media or a hele - shaw cell as shown in fig .
[ fig : schematic ] .
the frontal interface becomes unstable if @xmath2 , otherwise it is the rear interface which features the fingering instability , for @xmath3 .
the viscosity of the fluids depends on the solute concentration @xmath4 , i.e. @xmath5 .
fluids are assumed to be incompressible and neutrally buoyant . with an additional condition of slow diffusion
, the above - mentioned flow problem can be described in terms of darcy - korteweg equations @xcite coupled with a convection - diffusion equation for the mass conservation of the solute concentration .
for the dimensionless formulation of the equations the diffusive length and time scales , @xmath6 and @xmath7 , are used as the respective characteristic scales .
characteristic pressure , velocity , concentration and viscosity are taken to be @xmath8 and @xmath1 , respectively . here
@xmath9 is the constant permeability of the homogeneous porous medium . for simplicity
we have assumed a constant isotropic diffusion of the solute concentration , characterized by the diffusion coefficient @xmath10 .
the dimensionless equations in a lagrangian frame of reference , moving with the speed @xmath11 are written as , @xmath12 here @xmath13 is the gap - averaged velocity having longitudinal and transverse components @xmath14 and @xmath15 , respectively , @xmath16 is the dynamic pressure , @xmath17 is the dimensionless korteweg stress constant @xcite and the operator @xmath18
. the governing equations ( [ eq : cont ] ) - ( [ eq : diff ] ) are associated with the following boundary conditions : at the longitudinal boundaries , @xmath19 ; and at the transverse boundaries , @xmath20 ( representing the constant pressure ) , @xmath21 , in the lagrangian frame of reference ( shown in fig .
[ fig : schematic ] ) . the initial velocity is considered to be @xmath22 , while the initial distribution of the solute concentration is , @xmath23 inside the finite slice and @xmath24 outside of that .
the relationship between the dynamic viscosity of the underlying fluids and the solute concentration is assumed to be of arrhenius type , @xmath25 where @xmath26 is the log - mobility ratio @xcite .
hence , the displacement of a less ( more ) viscous slice by a more ( less ) viscous ambient fluid is represented by @xmath27 ( @xmath28 ) .
direct numerical simulations of eqs .
( [ eq : cont])-([eq : diff ] ) in terms of the stream function , @xmath29 ( @xmath30 ) , @xmath31\cdot\hat{k } , \nonumber \\ & & \\ \label{eq : vs2 } & & c_t + \psi_yc_x - \psi_xc_y = \nabla^2 c,\end{aligned}\ ] ] are performed using a highly accurate pseudo - spectral method @xcite . here
@xmath32 represents a unit vector normal to the @xmath33-plane .
numerical simulations are performed in a rectangular domain of dimensionless width @xmath34 , and length @xmath35 , where @xmath36 , the ratio of the length to the width of the domain , is the aspect ratio
. a convergence analysis of the numerical method and grid independence have been performed .
nonlinear simulations are carried out in the above - mentioned computational domain with @xmath37 and @xmath38 spectra , and it has been observed that the maximum relative error between these two sets of simulations is of @xmath39
. further refinement of the spatial disretization points in either direction does not alter the dynamics of the fingering pattern .
hence , for optimal computational cost we discretize our computation domain of size @xmath40 \times [ 0 , 1024]$ ] using @xmath37 spectral points .
in the absence of the korteweg stress vf dynamics for both @xmath28 and @xmath27 are the same until the interaction between the two interfaces @xcite .
our aim in this letter is to determine the possible changes in the dynamics of such slices in the presence of the korteweg stress and relate it to the surface tension effect in immiscible slices . dns results for @xmath41 , @xmath42 are shown in fig .
[ fig : nonlinear1 ] , which depicts that at @xmath43 fingers appear at the rear interface of the slice for @xmath44 ( see fig .
[ fig : nonlinear1](a ) ) , whereas , for @xmath45 the dynamics of both the interfaces feature diffusive characteristic , but no finger formation at the frontal interface having the unstable viscosity gradient ( fig . [ fig : nonlinear1](b ) ) .
this signifies that the presence of the korteweg stresses induces different onset of finger formation for @xmath28 and @xmath27 .
moreover , the wavelength of the unstable modes is larger for @xmath27 than that corresponding to @xmath28 .
these differences between @xmath28 and @xmath27 persist in the long - time behavior of the fingering dynamics of the slices .
\(a ) ( b ) + when ( a ) @xmath46 , ( b ) @xmath47 . from top
to bottom @xmath48 . , title="fig:",width=307 ] when ( a ) @xmath46 , ( b ) @xmath47 . from top
to bottom @xmath48 . , title="fig:",width=307 ] it is customary to ask what causes such differences in the dynamics in the presence of the korteweg stress ?
conventionally , the @xmath49 relation , @xmath50 , keeps @xmath51 unchanged for both more and less viscous finite slices , but the dimensionless value of the less ( more ) viscous fluid changes with the sign of @xmath52 .
we notice that in the absence of transient interfacial tension eq .
( [ eq : vs1 ] ) is free from @xmath53 ( see eq .
( 9 ) in @xcite ) .
hence , the onset of fingering and the qualitative as well as the quantitative pattern of the fingers are identical for @xmath28 and @xmath27 in the absence of the korteweg stress .
\(a ) ( b ) ( top ) , @xmath54 ( bottom ) , in the lagrangian reference frame for @xmath55 : ( a ) @xmath56 , ( b ) @xmath57 . , title="fig:",width=307 ]
( top ) , @xmath54 ( bottom ) , in the lagrangian reference frame for @xmath55 : ( a ) @xmath56 , ( b ) @xmath57 . , title="fig:",width=307 ] choosing the displacing fluid viscosity , @xmath1 , as the characteristic scale for the dynamic viscosity in both the cases of more and less viscous slices we should be more careful about the dimensionless korteweg stress constant , @xmath58 , which is inversely proportional to the characteristic viscosity . as an example , consider the displacement of a glycerin - water mixture by pure water .
the dynamic viscosity of glycerin - water mixture with @xmath59 glycerin concentration is 19.4 cp ( @xmath60 cp ) and that of pure water is 1.002 cp ( @xmath61 cp ) .
therefore , in this case we have @xmath62 cp and @xmath63 cp , which can be written in terms of the dimensionless parameter @xmath44 . for this pair of miscible viscous fluids , a dimensional value @xmath64 n corresponds to the dimensionless value @xmath65 for an injection speed @xmath66 ms@xmath67 and diffusivity @xmath68 m@xmath69s@xmath67 . on the other hand , for the displacement of water by glycerin - water mixture ( @xmath70 ) @xmath71 cp and @xmath72 cp .
this makes the dimensionless value @xmath73 for the same dimensional value @xmath64 n with the same characteristic variables . in order to justify our claim
we perform numerical simulations for @xmath74 , and the results obtained are presented graphically in fig .
[ fig : nonlinear2](a ) , which depicts that the onset of instability and the finger pattern , until the interaction of the unstable and the stable interfaces , are identical to those for @xmath75 ( fig .
[ fig : nonlinear1](a ) ) .
this suggests , in order to compare the numerical results of two similar flow configurations one must choose the dimensionless parameters in such a way that they correspond to the same dimensional value .
subsequently , we ask the question whether there exists a suitable characteristic scale for the dynamic viscosity which automatically takes care of this fact . in particular we choose the less viscosity @xmath76 as the characteristic viscosity and @xmath77 as the characteristic pressure ( see table [ tab : viscosity ] ) .
hence , the dimensionless form of @xmath49 relation becomes , @xmath78 with this modifications we have @xmath28 for both the more and less viscous slices , and the initial dimensionless dynamic viscosity of the ambient fluid ( finite slice ) becomes @xmath79 and @xmath80 ( @xmath80 and @xmath79 ) , respectively ( see table [ tab : viscosity ] ) .
thus the magnitude of the characteristic viscosity , @xmath81 , remains the same for both the less and more viscous slices that results the same @xmath58 corresponding to every single dimensional value @xmath82 , in both the cases . since , the mathematical description of the displacement of a high viscous slice by a low viscous ambient fluid remains the same after the modification , in what follows is the discussion of the displacement of a low viscous slice by a high viscous ambient fluid to understand the consequences of such modifications on the vf instability of more and less viscous slices in the presence of the korteweg stresses .
[ fig : nonlinear2](b ) shows the concentration distributions at successive times for @xmath83 , @xmath84 , which are identical to those in fig .
[ fig : nonlinear2](a ) , hence depicts that the onset of vf and finger pattern are the same to those for a more viscous slice ( fig . [
fig : nonlinear1](a ) ) , until the interaction of the fingers with the respective stable interface .
[ tab : viscosity ] .dimensional and dimensionless ( superscripted ` @xmath85 ' ) values of the dynamics viscosity of the underlying fluids , characteristic viscosity , @xmath81 , and @xmath52 used in the computations . [ cols="<,^,^,^,^,^ " , ] for @xmath46 , ( @xmath86{1cm}{.4pt}}$ ] ) @xmath47 ( @xmath87{2mm}{.4pt}}\hspace{1mm}\textperiodcentered\hspace{1mm}{\rule[0.5ex]{2mm}{.4pt}}\hspace{1mm}\textperiodcentered\hspace{1mm}{\rule[0.5ex]{2mm}{.4pt}}}}$ ] ) and @xmath83 ( @xmath87{2mm}{.4pt}}\hspace{2mm}{\rule[0.5ex]{2mm}{.4pt}}\hspace{2mm}{\rule[0.5ex]{2mm}{.4pt}}}}$ ] ) .
the inset images show the concentration distribution approximately at time of interactions between the stable and unstable interfaces.,width=480 ] the onset of fingering instability and the interaction between the stable and unstable interfaces can be quantified appropriately from the temporal evolution of the interfacial length @xcite , @xmath88 . in the diffusion
dominated regime @xmath89 retains a constant value equal to the width of the domain , and increases with the growth of the fingers .
hence , the onset of fingering is marked as the instance when @xmath89 starts increasing from the constant value of the diffusive regime .
[ fig : nonlinear_il ] shows the temporal evolution of the interfacial length corresponding to the simulations of figs .
[ fig : nonlinear1 ] and [ fig : nonlinear2 ] .
it depicts that the displacements remain stable over the first two decades of time and the fingers form at @xmath90 , for the viscosity relations , @xmath91 and @xmath92 , while at @xmath93 for @xmath94 .
@xmath89 decreases with time after the interaction of the fingers with the respective stable interface .
inset images show the concentration distributions approximately at such instance of interaction and depict how this time differ from each other for different viscosity profiles .
hence , all the qualitative features of the dynamics observed in the absence of any gradient stress @xcite , can be seen even with the gradient stresses , but with different quantitative measures due to the influence of transient interfacial tension .
\(a ) ( b ) + for a more and less viscous finite slice of width , @xmath95 for @xmath34 and @xmath58 . , title="fig:",width=307 ] for a more and less viscous finite slice of width , @xmath95 for @xmath34 and @xmath58 . ,
title="fig:",width=307 ] vf enhances the concentration gradient and the area available for diffusive flux across the interface by stretching the fluid - fluid interface and hence modifies the rate at which mixing occurs .
it would be interesting to understand the effect of the korteweg stresses on the mixing of a more or less viscous finite slice .
the degree of mixing is defined as @xcite , @xmath96 , in terms of the global variance of the concentration field , @xmath97 , where @xmath98 represents the spatial averaging over the domain .
the degree of mixing of more and less viscous finite slices have been plotted in fig .
[ fig : degree_mixing ] corresponding to the simulations of figs .
[ fig : nonlinear1 ] - [ fig : nonlinear2 ] .
due to stronger instability for the case of @xmath75 than @xmath99 , the degree of mixing , @xmath100 , is higher in the former than the latter ( see fig . [
fig : degree_mixing](a ) ) . in fig .
[ fig : degree_mixing](b ) , @xmath100 has been shown for the three cases : ( i ) @xmath101 , ( ii ) @xmath102 and ( iii ) @xmath103 .
it depicts identical temporal evolution of @xmath100 for both more and less viscous slices until the interaction of the fingers with the respective stable interface .
afterwards , @xmath100 becomes higher in the latter case than the former , and it asymptotically saturates to the maximum value of @xmath79 , corresponding to the completely mixed state .
the coincidence of the three curves in the time interval @xmath104 confirms the results shown in figs .
[ fig : nonlinear1](a ) , [ fig : nonlinear2 ] .
here we discuss lsa of miscible slice in the presence of the korteweg stresses with the viscosity relations given by eqs .
( [ eq : visco_new1])-([eq : visco_new2 ] ) and compare the results obtained with those of dns .
finally , the obtained lsa results are compared with a lsa of immiscible slice which confirms that the korteweg stress and the surface tension have identical effects on the instability .
\(a ) ( b ) + , ( b ) temporal evolution of the growth rate of the most dangerous wave perturbation ( @xmath105 ) .
inset image shows the maximum growth , @xmath106 , from all possible wave number , @xmath107 , representing the onset of instability .
the parameter values are @xmath108 and @xmath109 . ,
title="fig:",width=307 ] , ( b ) temporal evolution of the growth rate of the most dangerous wave perturbation ( @xmath105 ) .
inset image shows the maximum growth , @xmath106 , from all possible wave number , @xmath107 , representing the onset of instability .
the parameter values are @xmath108 and @xmath109 .
, title="fig:",width=307 ] \(a ) ( b ) + .
ivc ( @xmath86{1cm}{.4pt}}$ ] ) and dns ( @xmath87{2mm}{.4pt}}\hspace{2mm}{\rule[0.5ex]{2mm}{.4pt}}\hspace{2mm}{\rule[0.5ex]{2mm}{.4pt}}}}$ ] ) .
inset image is magnified near the onset of instability .
( b ) dispersion curves at @xmath110 for @xmath109 with the self - similar lsa .
, title="fig:",width=278 ] .
ivc ( @xmath86{1cm}{.4pt}}$ ] ) and dns ( @xmath87{2mm}{.4pt}}\hspace{2mm}{\rule[0.5ex]{2mm}{.4pt}}\hspace{2mm}{\rule[0.5ex]{2mm}{.4pt}}}}$ ] ) .
inset image is magnified near the onset of instability .
( b ) dispersion curves at @xmath110 for @xmath109 with the self - similar lsa .
, title="fig:",width=307 ] the initial - boundary value problem described above possesses a self - similar diffusive decaying solution in a similarity transformation @xmath111-domain @xcite , @xmath112/2 $ ] , where @xmath113 is the similarity variable . here
, @xmath114 corresponds to the self - similar decay of a rectangular function of width @xmath115 and height unity .
we analyze how an infinitesimal perturbation to this reference solution grows ( decays ) with time . eliminating the pressure and transverse velocity component from the linearized form of the eqs .
( [ eq : cont])-([eq : diff ] ) , and using the @xmath49 relation from eqs .
( [ eq : visco_new1])-([eq : visco_new2 ] ) , a system of coupled partial differential equations is obtained .
these equations are homogeneous in @xmath116 and @xmath117 , hence we can decompose the perturbation quantities into normal modes as , @xmath118 . here
@xmath119 and @xmath120 are the amplitude of the perturbed concentration @xmath121 and axial velocity @xmath122 , respectively , with @xmath123 being the wave number and the growth rate of the perturbations @xcite .
hence , the linear stability problem can be written as the following eigenvalue problem , @xmath124 @xmath125 where @xmath126 . the system of coupled ordinary differential equations ( eqs .
( [ eq : coupledode1])-([eq : coupledode2 ] ) ) has been solved using a finite difference method @xcite to determine the instantaneous growth rate of the perturbations in terms of the wave number @xmath107 and the frozen diffusive time @xmath127 .
dispersion curves obtained from lsa for the displacement of more and less viscous fluid slice are presented in fig .
[ fig : lsa](a ) for the parameters @xmath128 at the frozen diffusive time @xmath110 .
it depicts that the two dispersion curves are almost indistinguishable for the unstable modes . also , the temporal evolution of the growth rate of the most unstable mode , @xmath129 ( fig .
[ fig : lsa](a ) ) , as well as the onset of instability for all possible modes coincide for @xmath46 and @xmath83 ( fig .
[ fig : lsa](b ) ) . linearizing the stream - function form of eqs .
( [ eq : cont])-([eq : diff ] ) with @xmath49 relation given by eqs .
( [ eq : visco_new1])-([eq : visco_new2 ] ) and using a pseudo - spectral method , an initial value calculation ( ivc ) is perfromed .
the growth rate of the perturbed concentration , @xmath121 , @xmath130 , where @xmath131 , obtained from ivc is compared with those obtained from dns @xcite .
the temporal evolution of the growth rates of ivc coincide with those obtained from dns for both @xmath46 and @xmath132 with @xmath133 ( fig .
[ fig : comparision](a ) ) . the numerical algorithm of the ivc will be discussed elsewhere in detail . fig .
[ fig : comparision](b ) depicts an excellent agreement between the dispersion curves obtained from self - similar lsa at @xmath110 corresponding to @xmath75 and @xmath74 , and confirms the observation from dns presented in fig .
[ fig : nonlinear2](a ) . for three sets of viscosities : @xmath134 ( @xmath86{1cm}{.4pt}}$ ] ) , @xmath135 ( @xmath87{2mm}{.4pt}}\hspace{2mm}{\rule[0.5ex]{2mm}{.4pt}}\hspace{2mm}{\rule[0.5ex]{2mm}{.4pt}}}}$ ] ) , @xmath136 ( @xmath137 ) .
, width=480 ] in what follows a comparison between an immiscible slice and a miscible slice with the korteweg stresses .
@xcite have shown that in a rotating hele - shaw cell both the qualitative and quantitative features of the korteweg stress in miscible fluids and surface tension in immiscible fluids are identical .
not only the stabilizing property of the korteweg stresses was confirmed , but also the number of fingers were measured to be the same for the miscible and immiscible fluids , both theoretically as well as experimentally .
our aim is to understand the relative importance of the viscous force to the surface tension force in three layer immiscible displacement , in which the growth rates of the perturbations at the two interfaces can be represented as ( see eq . ( 10 ) in @xcite ) , @xmath138 , where , @xmath139\xi - [ e^{kl}(\mu + \mu_r ) + e^{-kl}(\mu - \mu_r)]\tau , ~ c = \tau\xi(e^{kl } - e^{-kl } ) , ~ \xi = tk((\mu_r - \mu)u / t - k^2 ) , ~ \tau = sk(k^2 - ( \mu - \mu_l)u / s)$ ] . here
@xmath140 and @xmath141 correspond to the larger and smaller growth rates , respectively .
furthermore , @xmath142 and @xmath143 are the surface tension at the left and right interfaces , @xmath144 being the width of the middle layer , @xmath11 is the fluid velocity and @xmath145 are the viscosities of the left , middle and right fluid layers , respectively .
these show that the obtained analytic expressions for the dispersion relations at the two interfaces involve explicitly the dynamic viscosities of the three fluid layers , not their ratio . as a consequence ,
the dispersion curves corresponding to @xmath140 become different for two sets of viscosities although having the same viscosity ratio of two adjacent fluid layers .
for instance , we consider three sets of viscosities in such a way that the interface separating the middle layer from the right layer becomes unstable for the first two cases , while for the third , it is the interface between the left and the middle layer that features instability . corresponding dispersion curves ( @xmath140 vs. @xmath107 ) are shown in fig .
[ fig : daripaimmiscible ] for @xmath146 ; the solid line corresponds to the case of @xmath134 , dashed line corresponding to @xmath135 , and the circles representing the situation when @xmath136 .
the viscosities of the left and right layers are chosen to be equal so that it reduces to an analogous form of a finite miscible slice discussed above .
although the ratio of the high viscosity to the low viscosity remain to be the same , @xmath147 , the dispersion curve for the second case differs from the remaining two .
it signifies that the dispersion curves corresponding to a more and less viscous middle layer will overlap if the dynamic viscosities of the less and more viscous fluids are kept unchanged .
eventually , growth rate is always higher for the case with lager dynamic viscosity of the fluids .
another important observation is the variation of the cutoff wave number ( @xmath148 ) for the unstable mode and the most unstable wave number ( @xmath149 ) with the variation of the surface tension .
the qualitative changes of these two quantities are similar to those observed in miscible displacements under the influence of the korteweg stress .
both in miscible and immiscible fluid systems , @xmath148 and @xmath149 decrease with the increase of the magnitude of @xmath58 and the surface tension force , respectively .
we have theoretically investigated the vf instability of miscible slices with transient interfacial tension .
an appropriate choice of the dynamic viscosity of the fluid in the characteristic scales manifests its importance in the study of vf with transient interfacial tension .
our analysis is capable of reproducing the results existing in the literature .
the results obtained depict that the onset of vf and their dynamical pattern can become identical only under certain scaling analysis for both the more and less viscosity slices in the presence of korteweg stress , similar to the case when these stresses are absent @xcite .
synergetic mixing with vf and alternating injection @xcite will be the same with the present viscosity model irrespective of the choice of the viscosity of the fluid that fills the hele - shaw cell before the injection starts .
alike effect of the dynamic viscosity is also observed in immiscible fluids .
the findings of this letter will certainly help to understand multiphase viscous flows with different viscosity of each phase or when the viscosity depends non - monotonically on the solute concentration @xcite ; for instance , during the mixing process of chemical components , pollutant contamination in aquifers , co@xmath150 sequestration , etc .
miscible displacement of viscous fluids with korteweg stress has important applications in chemistry @xcite . in this letter we present a preliminary understanding of the relative importance of the korteweg stress to the viscous stress by incorporating one of the simplest situation of constant diffusivity , @xmath10 , of the solute concentration . a velocity induced or concentration dependent diffusivity should be studied for a more realistic situation .
more interesting and challenging problem of miscible vf in the presence of the korteweg stress induced by both the concentration and temperature gradients is the focus of the authors future work . | we perform linear stability analyses ( lsa ) and direct numerical simulations ( dns ) to investigate the influence of the dynamic viscosity on viscous fingering ( vf ) instability in miscible slices . selecting the characteristic scales appropriately
the importance of the magnitude of the dynamic viscosity of individual fluids on vf in miscible slice has been shown in the context of the transient interfacial tension .
further , we have confirmed this result for immiscible fluids and manifest the similarities between vf in immiscible and miscible slices with transient interfacial tension . in a more general setting , the findings of this pletter will be very useful for multiphase viscous flow , in which the momentum balance equation contains an additional stress term free from the dynamic viscosity .
displacement processes through porous rocks and mixing of two miscible fluids are active areas of research , having several industrial and environmental applications , such as enhanced oil recovery , hydrology and filtration , carbon capture and storage , etc .
vf , a hydrodynamic instability that occurs in both the immiscible and miscible fluids while displacing a more viscous fluid by a less viscous one , is inherent in such flow configuration @xcite . in immiscible fluids surface tension force at the interface acts against the instability @xcite . on the other hand , in miscible fluids , where a thermodynamically stable interface does not exist ,
a transition zone relaxes with time due to diffusion and acts against the finger growth .
experiments @xcite reveal , when the diffusion is slow , a steep gradient in the form of density , concentration or temperature between the underlying fluids gives rise to a weak transient interfacial tension that mimic surface tension effect .
this was first discussed by korteweg in 1901 @xcite , who introduced an additional stress term , known as the korteweg stress , in the equation of motion .
the existence of korteweg stress or transient surface tension is also observed in the experiments of colloidal suspensions @xcite and in the binary liquid system of isobuteric acid and water @xcite .
chen and wang @xcite analyzed the influence of vf instability on the spreading of a localized fluid slice having higher mobility than the surrounding fluid . on the other hand , de wit _
et al . _
@xcite studied the same problem in the context of separation in a chromatographic column when the viscosity of the sample is higher than the solvent .
mishra _
et al . _
@xcite have shown that the onset of vf instability and the subsequent finger pattern near the onset are identical for both the less and more viscous slice .
influence of the korteweg stresses on vf instability was investigated theoretically by joseph and his co - workers @xcite and chen _ et al . _
@xcite however , to the best of the authors knowledge the influence of such stresses on the nonlinear vf instability of more and less viscous miscible slices in a hele - shaw cell has never been addressed adequately . in particular
this letter addresses the question , what is the influence of the korteweg stress that describes volume forces arising because of the nonlocal molecular interactions on the vf at the rear and frontal interfaces of a localized slice ?
such a classical complex pattern dynamics has been investigated through a highly accurate fourier - spectral method based direct numerical simulations @xcite .
it has been proved theoretically that an appropriate choice of the dynamic viscosity of underlying fluids results the identical onset of fingering instability and the subsequent finger patterns are also identical for both more and less viscous slices in the presence of the korteweg stresses .
the dns results are found to be in excellent agreement with the corresponding lsa .
also , the similarities between the immiscible slices and miscible one with transient interfacial tension have been proved through lsa , which affirms the classical nature of this study . |
jet quenching , as observed in heavy ion collisions at rhic @xcite , has become an important probe of a deconfined state of matter with quasi - free quarks and gluons , the so - called quark gluon plasma ( qgp ) .
indeed , this observable allows inferences on the properties of the expanding medium .
especially jet quenching of heavy flavors charm @xmath0 and bottom @xmath1 could be an ideal probe of the qgp . due to the large massgev , @xmath2gev ] of these quarks , @xmath3 ,
they are produced in a very early stage of the collision and can later be considered as test particles in the ( equilibrated ) heat bath .
the quenching of light quarks is well understood as an effect of radiative energy loss when propagating through the medium @xcite . however , in the case of heavy quarks pure radiative energy loss seems not sufficient to explain experimental data ( mainly from electron yields ) on heavy flavor quenching @xcite , suggesting surprisingly strong attenuation .
there is an ongoing debate on the underlying reasons , which could be possibly clarified by analyzing spectra from heavy flavor decays .
an additional contribution of suppression , potentially important for the range of smaller ( to medium ) momenta , is the collisional energy loss by elastic processes as discussed in @xcite and with further modifications in @xcite .
it will turn out that in our approach a fairly large collisional energy loss is predicted , although some comparisons between collisional and radiative contributions @xcite indicate another picture where the collisional loss should be relatively small ( @xmath4 ) compared to the radiative one .
this conclusion was , however , based on a fixed coupling approach , and using a naive approximation for the required infrared regulator in the qcd cross section . in the following work both aspects will be considered in a more precise way , aiming at providing a computational framework to describe binary parton collision effects more reliably , which seems necessary as a baseline to understand quenching in heavy - ion experiments .
noted in this context is that certain observables might be more sensitive to the inclusion of binary collisions than the averaged energy loss , due to different probability distributions of radiative and collisional processes @xcite .
we start with a short overview of the collisional energy loss of a heavy quark in a static qgp with temperature @xmath5 .
the average loss per unit length @xmath6 has recently been recalculated @xcite , revising a previous result by braaten and thoma @xcite in the large - energy limit . in general , the mean energy loss of a heavy quark , with velocity @xmath7 , due to binary interactions with light / strange quarks and gluons ( @xmath8 ) from the plasma , = d , [ eqn:201 ] is the first moment of the differential transition rate @xmath9 with respect to the energy transfer @xmath10 in the interaction . to calculate @xmath6 beyond leading - log accuracy ( see eq .
below ) , it is convenient to decompose the phase space of the invariant momentum transfer @xmath11 into a soft and a hard part with respect to a separation scale @xmath12 . in the weak coupling limit , @xmath13 , @xmath12 can be chosen arbitrarily in the range @xmath14 . a sufficient separation of the ` soft ' screening scale , which is of the order of the debye mass @xmath15 , and the ` hard ' thermal scale , @xmath16 , seems to persist ( at least for the observables under consideration ) even for moderately large coupling , as will be discussed below .
the soft contribution to @xmath6 can be calculated from the imaginary part of the hard thermal loop ( htl ) dressed heavy quark self - energy @xcite .
the hard contribution can be obtained from the interaction rate evaluated with tree - level matrix elements @xmath17 , with a constrained momentum transfer @xmath18 , = & _ k_k_p ( |t|-|t^| ) + & ( 2)^4^(4)(p+k - p-k)_spin , color|m_i|^2 , [ eqn:202 ] where @xmath19 denotes the bose or fermi distribution even fairly close to the transition / crossover temperature .
] of the thermal target particle @xmath20 , similarly @xmath21 for the final states , and we use the short - hand notation @xmath22 . @xmath23 and @xmath24 ( @xmath25and @xmath26 ) denote the initial ( final ) 4-momenta of the collisional partners ; @xmath27 and @xmath25 are the heavy quark momenta .
the sum of the matrix elements @xmath28 runs over all initial and final states and is divided by the degeneracy factor @xmath29 of the heavy quark . in the large - energy limit , for @xmath30 , and to the level of accuracy we are interested in , eq .
can be simplified @xcite , = d_i_k_t_min^t^*dt ( -t ) .
[ eqn:203 ] here @xmath31 are the heavy quark born cross sections @xcite for the interaction with medium particles of type @xmath20 and degeneracy factor @xmath32 , and @xmath33 in terms of the invariant energy @xmath34 .
keeping sub - leading terms in the calculation of the soft and hard phase space contributions leads to the nll ( next - to - the leading energy logarithm ) formula @xcite = .
[ eqn:204 ] the constant @xmath35 depends on the number of active flavors , in the present context @xmath36 ( @xmath37 ) , c(n_f)&=an_f+b0.146n_f+0.050 , + a&=(2)-+c , + b&=(2)-+c , + c&=--1.147 , + & 0.577 [ eqn:205 ] the first ( coulomb ) logarithm in eq
. arises from @xmath11-channel interactions and was first obtained by bjorken @xcite .
the second logarithm , due to qcd compton scattering where a potential collinear divergence is screened by the heavy quark mass , was missed for some time in the literature .
we point out that formula assumes a _ fixed _ coupling @xmath38 . taking into account the momentum dependence of the qcd coupling amounts to evaluating prefactors of @xmath38 in at scales specific for each contribution : for the coulomb - log term @xmath39 @xcite , and @xmath40 in the compton logarithm term @xcite . in the numerical simulation of the next section
we will use differential transition rates calculated from the qcd born cross sections integrated over the whole available phase space , however with the singular @xmath11-channel contributions screened by a mass term , i.e. we replace @xmath41}/dt \sim t^{-2 } \;\to\ ; ( t-\mu^2)^{-2}$ ] .
the screening mass @xmath42 has often been chosen _
ad hoc _ to be equal to the debye mass although they have only the same _ parametric _ dependence on the coupling , @xmath43 .
we illustrate this point by showing in fig .
[ fig:201 ] that the resulting energy loss would not converge to the asymptotic formula .
in order to match this analytic result one actually has to introduce an effective screening mass of the form @xcite ^2(t)=4(1+)(t)t^2 , [ eqn:206 ] which @xmath44 depends on the momentum exchange via the running coupling , and @xmath45 also differs by a prefactor from the debye mass m_d^2(t ) = 4(1+)(m_d^2)t^2 .
[ eqn : md2 ] this prefactor @xmath46 has been determined @xcite , = 0.184 .
[ eqn:207 ] let us comment at this point on the separation of ` soft ' and ` hard ' scales , @xmath47 and @xmath48 respectively ( which underlies the derivation of equation ) , for larger coupling
. then prefactors become important ; for the typical hard scale the thermal average reads @xmath49 .
this value happens to coincide approximately with an upper bound , seen in non - perturbative lattice qcd calculations , sets the scale of the running coupling @xcite , is in quantitative agreement with lattice qcd calculations even fairly close to the transition temperature .
] for the squared debye mass , @xmath50 . therefore the facts that @xmath51 in the cut - off and the running coupling in being smaller than @xmath52 in the debye mass for harder momentum exchanges , suggest that a certain separation of scales persists even for moderately large coupling , and that our approach may provide reasonable numerical estimates . in comparison to the analytic nll - formula ; @xmath53 is the adjusted parameter , @xmath54 corresponds to screening with the debye mass.,width=321,height=230 ] we now generalize the mean energy loss formula to the case of a qgp heat bath with a collective flow . in this case
the thermal distribution functions @xmath55 have to be replaced by the jttner functions for bosons and fermions , which depend on the velocity @xmath56 of the medium , n_j ( k ) = 1e^(k - k ) /t 1 , [ eqn:208 ] where @xmath57 . noting that the rhs of , with jttner distributions in place , is a manifestly covariant expression
, we expect that the energy loss in a streaming qgp can be obtained from by simply replacing @xmath58 in the two logarithms , i.e. in is to be replaced as discussed below eq .
, however now with @xmath59 instead of @xmath60.[footnote : run coupling ] ] = , [ eqn:216 ] which we will prove in appendix [ appendix:100 ] for the coulomb term . by covariance we also expect that the constant @xmath35 is the same as in the static case ( where it is @xmath61-independent ) which we will justify numerically in section 3 ( a rigorous proof would require a more cumbersome evaluation of thermal jttner integrals both for the hard and the soft contributions , which is not investigated here ) .
gev charm quark as a function of the medium velocity @xmath62 for different flow directions of the qgp with respect to the charm momentum vector , in fixed - coupling approximation.,width=321,height=230 ] we illustrate in fig .
[ fig:202 ] that flow of the qgp , as expected in heavy - ion experiments , can have sizable effects on the mean collisional energy loss of heavy quarks .
the following simulations are based on the heavy quark double differential collision rate with respect to final momentum @xmath63 and scattering angle @xmath64 , @xmath65 .
the rates are ( pre-)computed in a thermalized qgp at rest ; flow of the medium is taken into account by lorentz boosts in the local rest frame . the divergence from the @xmath11-channel terms in the born cross sections
is screened by the effective screening mass @xmath66 , and the running of the qcd coupling is taken into account as described in section [ chap:200 ] .
discretizing @xmath63 and @xmath64 , and multiplying @xmath67 by a sufficiently small time interval @xmath68 yields , for given initial momentum @xmath69 , probability matrices describing transitions @xmath70 between momentum states with a deflection by @xmath64 .
these stochastic matrices allow to interpret the sequence of heavy quark interactions as a markov chain @xcite , and they form the basis of our monte carlo simulation which is an efficient alternative to other transport approaches such as @xcite . the resulting ( numerical ) mean energy loss is compared in fig . [ fig:301 ] to the analytic formula .
this comparison illustrates that sub - leading terms in @xmath61 , which are not captured by , are reasonably small already for intermediate charm energies .
it also shows that a potential dependence of the constant @xmath35 on the flow velocity @xmath62 is suppressed at large energy , as anticipated in section [ chap:200 ] . to the analytical formula .
the numerical calculations were done within our mc - simulation for charm quarks with @xmath71gev and @xmath72gev.,width=321,height=230 ] in our approach the heavy quarks are treated as test particles which do not , in turn , affect the medium . this approximation is justified because of the small number of heavy flavor pairs , about 4 comparing to 800 gluons in the mid - rapidity region of a central collision at rhic @xcite .
this allows us to use different approaches , such as hydrodynamics or bamps ( boltzmann approach of multi - parton scattering @xcite ) , to specify the temperature and flow profile of the qgp background in heavy - ion collisions .
we note in this context that we have successfully tested our approach for various background dynamic scenarios by comparing to original bamps calculations ( to that end we recalculated our interaction rates @xmath73 with the classical boltzmann distribution , as assumed in bamps ) .
the spacial initialization of heavy quark dijet events is done according to the glauber model @xcite and the initial momentum distributions are computed with the event generator @xmath74 at next - to - leading order precision @xcite .
shadowing and the cronin effect are neglected since the modification of heavy quark distributions should be small for momenta @xmath75gev @xcite .
an indirect comparison of the initial momentum distributions to experimental data can be found in @xcite .
when at a later stage in our simulation the qgp reaches the critical temperature @xmath76gev , we assume that the heavy quarks fragment into heavy mesons , described by the peterson function @xcite d_q^m(z)= , [ eqn:301 ] where @xmath77 is a normalization constant , @xmath78 the momentum fraction ( heavy meson to heavy quark ) and @xmath79 , @xmath80 are typical squared mass ratios for charm and bottom , respectively .
the following stage of heavy meson decays to electrons is calculated with pythia 8.1 @xcite .
in order to study thoroughly the mechanisms which affect the spectrum of electrons from heavy quarks in heavy - ion collisions energy loss at parton level , hadronization to mesons and their subsequent decay we consider first the simple bjorken model before investigating more realistic dynamics of the qgp background ( in so doing we generalize the exploratory study of @xcite , which considered only the charm sector at partonic level ) .
our observable is the nuclear modification factor , defined as the ratio of the final transverse momentum distribution to the distribution in a nucleon - nucleon collision scaled by the number of binary collision @xmath81 according to the glauber assumption , r_aa=. [ eqn:401 ] we improve the simple power - law momentum spectrum assumed in @xcite by the nlo initial momentum distribution .
besides that , we use the same parameter sets for the dynamics / geometry ( summarized in table [ tab:401 ] ) , except for the hadronization temperature which is set to a slightly larger value , @xmath82mev ( which somewhat reduces the life time @xmath83 of the qgp phase ) , for consistency with the more detailed simulations below . ,
@xmath84 ( peterson fragmentation ) and electron ( decay from @xmath84 with pythia ) quenching at mid - rapidity in the bjorken model to the observed @xmath85 from heavy flavor decays in the centrality class of @xmath86 @xcite.,width=321,height=230 ] fig .
[ fig:401 ] shows the combined modification factors of @xmath87 quarks , @xmath84 mesons and the resulting electron yields .
it is apparent ( as assumed in @xcite ) that the @xmath87 spectra are in good agreement with the electron spectra from heavy flavor decays for @xmath88gev . in this simplified framework
we capture the trend of the experimental data , with some room for effects due to heavy quark radiative energy loss .
however , since the radial expansion of the qgp will lead to shorter life times of the plasma phase , we now turn to more realistic dynamical models .
the local temperature @xmath89 and the flow velocity @xmath90 of the qgp background can be obtained from available parton transport codes or hydrodynamics .
we focus here on the bamps approach @xcite , which can reproduce short thermalization times of @xmath91fm / c as expected from experimental data @xcite ( and which were assumed for the parameter sets in table [ tab:401 ] ) .
we demonstrate with fig .
[ fig:402 ] that the quenching ratio is not very sensitive to the value of @xmath92 . in a peripheral nucleus - nucleus collision to different thermalization times
@xmath93fm / c , which are used as starting point of our mc simulation.,width=321,height=230 ] in order to control that the observables under consideration are not very sensitive to other details of the background dynamics , we also use temperature and flow profiles as computed with the ( 3 + 1)-dimensional ideal hydrodynamic solver shasta @xcite . here we set the initial energy density to that of bamps and assume an initialization time of @xmath94fm / c . figs .
[ fig:403 ] and [ fig:404 ] show our results for the nuclear modification factor in a central and in a peripheral collision with an impact parameter of @xmath95fm , which corresponds to the centrality class of @xmath96 . in both cases
the quenching due to collisional energy loss alone reproduces the trend of the data , again with some room for radiative effects , although somewhat less for more peripheral collisions . following a common attempt to estimate the relative importance of collisional vs. radiative energy loss we scale _ ad hoc _ the elastic cross section by a factor @xmath24 .
we note that this prescription can at best give indications since both the energy and the path length dependence of the two loss mechanisms differ markedly .
it turns out that the experimental data for both central and peripheral collisions lead to a constraint @xmath97 , from which we may conclude that collisional and radiative processes are of similar importance .
we note that our @xmath98 ( from peripheral to central collisions ) is somewhat smaller than @xmath99 found in @xcite , which can be traced back to bamps simulating only gluons distributed according to boltzmann statistics ( with the effect of the light - flavor degrees of freedom estimated by color - factor scaling ) . after the nuclear modification factor @xmath85 we now consider the elliptic flow , which is defined by the ensemble average v_2(p_t ) = .
[ eqn:402 ] the elliptic flow of heavy quarks forms at later stages of heavy - ion collisions by interactions with the medium ; consequently the sensitivity of the following results on the thermalization time @xmath92 is even less than for @xmath85 .
[ fig:405 ] shows the transverse momentum dependence of @xmath100 obtained in our approach in comparison with experimental data .
obviously , the data are underestimated considerably ; for a better agreement the binary cross sections would need to be multiplied by a factor of @xmath101 , which is slightly larger than what we found for @xmath85 , but again less than @xmath102 inferred in the bamps approach @xcite .
we emphasize here again that the crucial small value of the cut - off parameter @xmath46 in is justified only in the high - energy limit @xmath103 , i.e. a few gev for charm quarks @xcite .
lacking a similar calculation for a cut - off at smaller energies , we have extrapolated our approach to rather small @xmath104 values , therefore the results can only be considered as rough estimates . at mid - rapidity @xmath105 for a central @xmath106 collision with a center of mass energy of @xmath107gev ( rhic ) .
the curves are calculated within our transition matrix approach for the nlo initial momentum distribution of heavy quarks .
@xmath24 denotes an artificial scaling factor of the elastic cross section . for comparison , experimental data on electrons from heavy flavor decays in the centrality class of @xmath86 are shown @xcite.,width=321,height=230 ] but for a peripheral @xmath106 collision with an impact parameter of @xmath95fm . for comparison , experimental data on electrons from heavy flavor
decays in the centrality class of @xmath96 are shown @xcite.,width=321,height=230 ] but for elliptic flow.,width=321,height=230 ]
we have generalized the leading logarithmic result for the mean energy loss of energetic heavy quarks in the qgp to the situation of a constant flow of the medium .
the result has been tested numerically using our transition matrix approach , which incorporates running coupling and uses an improved infrared regulator , and which allows us compute efficiently the evolution of heavy quarks in any thermalized medium of quarks and gluons due to binary collisions . as an application we have calculated the heavy flavor nuclear modification factor and the elliptic flow in heavy - ion collisions . while binary scattering of heavy quarks in the qgp alone can not explain the experimental data , this mechanism contributes significantly , of a similar order as radiative processes .
this can be quantified phenomenologically by a parameter @xmath24 multiplying the elastic cross sections . in case of the nuclear modification factor
we would need an increase @xmath108 , whereas matching better the elliptic flow data requires @xmath101 ; both values are somewhat smaller than findings in other approaches , basically due to an improved infrared cut - off and using running coupling and full quantum statistics .
the fact that a simple upscaling of the binary cross sections seems not sufficient to reproduce @xmath85 and @xmath100 data simultaneously may help to understand better the energy loss due to strongly anisotropic gluon emission ( see @xcite ) and/or the initial conditions of the medium , which could modify its transverse expansion .
furthermore our results for the @xmath85 of central and peripheral collisions may be a direct indication of non - linear scaling in the pathlength of the radiative energy loss .
since the mean pathlength drops in non - central collisions this point of view explains naturally a small increase in @xmath24 from peripheral to central collisions in the case of @xmath85 .
a higher value of @xmath109 for the elliptic flow in comparison to @xmath24 of the nuclear modification factor could also be explained due to a different pathlength dependence of the radiative scenario , which amplifies the path asymmetry in and out of plane in a peripheral collision . on the other hand
, it seems also necessary to improve our understanding of elastic scattering processes in the low-@xmath104 region .
+ + + + * acknowledgments : * a. m. and j. u. acknowledge the kind hospitality at the university of cape town , where part of this work has been done .
part of the calculations was performed at the center for scientific computing of the goethe university frankfurt .
this work was supported by the helmholtz international center for fair within the framework of the loewe program launched by the state of hesse , and by the south african national research foundation .
we derive here the coulomb contribution in formula for the mean energy loss in the presence of a collective qgp flow ( neglecting the momentum dependence of the qcd coupling , see footnote [ footnote : run coupling ] ) . as for the medium at rest , the leading logarithm results from terms @xmath41}/dt \sim 1/t^{2}$ ] in the cross section , and we can extract it from the hard contribution with @xmath110 replaced by @xmath111 . for large energy , we can also simplify @xmath112 , hence ( omitting prefactors , which can be fixed by matching to the static case ) ~ _ k ( ) , [ eqn : a1 ] where @xmath113 denotes the jttner distribution function . to perform the @xmath114-integral , it is convenient to specify a coordinate system . in
spherical coordinates , with the heavy quark momentum @xmath115 in @xmath116-direction , the flow velocity @xmath117 in the @xmath118-plane , viz .
@xmath119 , and the orientation of the thermal momentum @xmath114 specified by the polar angle @xmath120 and azimuth @xmath121 , the argument of the jttner function reads = k = , [ eqn : a2 ] where we introduce an effective temperature . approximating @xmath122 with @xmath123 , we rewrite equation , ~ _ -1 ^ 1 dx_0 ^ 2 d_0^dk k n_j(k ) ( ) , [ eqn : a3 ] and integrate over @xmath114 , & & _ 0^dk k n_j(k ) + & & t_eff^2((1-x ) ) , where we have neglected terms which will not contribute to the logarithm in @xmath61 .
we proceed with the @xmath121-integral in , by writing the effective temperature introduced in as @xmath124 , with @xmath125 where @xmath126 and @xmath127 , & & _ 0 ^ 2d 1r^2 + & & = .
the auxiliary integrals @xmath128 and @xmath129 can be calculated in closed form , @xmath130\ , , \end{aligned}\ ] ] where @xmath131 . the final @xmath132-integration in can then be performed using @xmath133\,,\\ + \frac{l}{2\beta}\right]\,,\\ \int_{-1}^1 dx\,i_2&\;=\;2\gamma^2\left[1-\ln\left(\gamma\right)+\frac{l}{2\beta}\right ] , \end{aligned}\ ] ] where @xmath134 .
collecting all relevant terms contributing to the energy loss due to @xmath11-channel scattering at leading logarithmic order with flow , we arrive at @xmath135}}{dx } \sim t^2\ln\left(\frac{\gamma e\left(1-\beta\cos\psi\right ) t}{m_d^2}\right ) , % \label{eqn : a8}\ ] ] where we indeed recognize in the argument of the logarithm the 4-product @xmath136 , as anticipated in eq . .
the calculation of the collinear logarithmic contribution can be done along similar lines . 00 j. adams _ et al .
_ [ star collaboration ] , nucl .
phys . a * 757 * ( 2005 ) 102 [ arxiv : nucl - ex/0501009 ] .
k. adcox _ et al . _
[ phenix collaboration ] , nucl .
phys . a * 757 * ( 2005 ) 184 [ arxiv : nucl - ex/0410003 ] . i. arsene _ et al . _ [ brahms collaboration ] , nucl .
phys . a * 757 * ( 2005 ) 1 [ arxiv : nucl - ex/0410020 ] . b. b. back _ et al . _ , nucl . phys . a * 757 * ( 2005 ) 28 [ arxiv : nucl - ex/0410022 ] .
r. baier , y. l. dokshitzer , a. h. mueller , s. peigne and d. schiff , nucl .
b * 484 * ( 1997 ) 265 [ arxiv : hep - ph/9608322 ] .
m. gyulassy , p. levai and i. vitev , phys .
* 85 * ( 2000 )
5535 [ arxiv : nucl - th/0005032 ] . b. i. abelev _ et al .
_ [ star collaboration ] , phys .
* 98 * ( 2007 ) 192301 [ erratum - ibid .
* 106 * ( 2011 ) 159902 ] [ arxiv : nucl - ex/0607012 ] .
a. adare _ et al .
_ [ phenix collaboration ] , phys . rev . lett .
* 98 * ( 2007 ) 172301 [ arxiv : nucl - ex/0611018 ] . j. bielcik [ star collaboration ] , nucl
. phys .
a * 774 * ( 2006 ) 697 [ arxiv : nucl - ex/0511005 ] .
s. wicks , w. horowitz , m. djordjevic and m. gyulassy , nucl . phys .
a * 784 * ( 2007 ) 426 [ arxiv : nucl - th/0512076 ] .
m. g. mustafa and m. h. thoma , acta phys . hung . a * 22 * ( 2005 ) 93 [ arxiv : hep - ph/0311168 ] .
g. d. moore and d. teaney , phys .
c * 71 * ( 2005 ) 064904 [ arxiv : hep - ph/0412346 ] .
a. peshier , nucl .
a * 888 * ( 2012 ) 7 [ arxiv:0801.0595 [ hep - ph ] ] .
p. b. gossiaux and j. aichelin , phys .
c * 78 * ( 2008 ) 014904 [ arxiv:0802.2525 [ hep - ph ] ] .
b. gossiaux , r. bierkandt and j. aichelin , phys .
c * 79 * ( 2009 ) 044906 [ arxiv:0901.0946 [ hep - ph ] ] . h. van hees , v. greco and r. rapp , phys . rev .
c * 73 * ( 2006 ) 034913 [ nucl - th/0508055 ] .
h. van hees , m. mannarelli , v. greco and r. rapp , phys .
rev . lett .
* 100 * ( 2008 ) 192301 [ arxiv:0709.2884 [ hep - ph ] ] .
j. uphoff , o. fochler , z. xu and c. greiner , phys .
c * 84 * ( 2011 ) 024908 [ arxiv:1104.2295 [ hep - ph ] ] .
g. -y .
qin , j. ruppert , c. gale , s. jeon , g. d. moore and m. g. mustafa , phys .
* 100 * ( 2008 ) 072301 [ arxiv:0710.0605 [ hep - ph ] ] .
b. g. zakharov , jetp lett .
* 86 * ( 2007 ) 444 [ arxiv:0708.0816 [ hep - ph ] ] . s. peigne and a. peshier , phys .
d * 77 * ( 2008 ) 014015 [ arxiv:0710.1266 [ hep - ph ] ] .
s. peigne and a. peshier , phys .
d * 77 * ( 2008 ) 114017 [ arxiv:0802.4364 [ hep - ph ] ] .
e. braaten and m. h. thoma , phys . rev .
d * 44 * ( 1991 ) 1298 .
e. braaten and m. h. thoma , phys .
d * 44 * ( 1991 ) 2625 .
b. l. combridge , nucl .
b * 151 * ( 1979 ) 429 .
j. d. bjorken , fermilab - pub-82 - 059-thy .
a. peshier , phys .
* 97 * ( 2006 ) 212301 [ arxiv : hep - ph/0605294 ] .
a. peshier , hep - ph/0601119 .
j. uphoff , o. fochler , z. xu and c. greiner , phys .
c * 82 * ( 2010 ) 044906 [ arxiv:1003.4200 [ hep - ph ] ] .
z. xu and c. greiner , phys .
c * 71 * ( 2005 ) 064901 [ hep - ph/0406278 ] .
r. j. glauber , nucl .
a * 774 * ( 2006 ) 3 [ nucl - th/0604021 ] .
m. l. miller , k. reygers , s. j. sanders and p. steinberg , ann .
nucl . part .
* 57 * ( 2007 ) 205 [ nucl - ex/0701025 ] .
s. frixione and b. r. webber , jhep * 0206 * ( 2002 ) 029 [ hep - ph/0204244 ] .
s. frixione , p. nason and b. r. webber , jhep * 0308 * ( 2003 ) 007 [ hep - ph/0305252 ] .
n. armesto , m. cacciari , a. dainese , c. a. salgado and u. a. wiedemann , phys .
b * 637 * ( 2006 ) 362 [ hep - ph/0511257 ] . c. peterson , d. schlatter , i. schmitt and p. m. zerwas , phys .
d * 27 * ( 1983 ) 105 .
t. sjostrand , s. mrenna and p. z. skands , jhep * 0605 * ( 2006 ) 026 [ arxiv : hep - ph/0603175 ] .
t. sjostrand , s. mrenna and p. z. skands , comput .
commun .
* 178 * ( 2008 ) 852 [ arxiv:0710.3820 [ hep - ph ] ] .
d. h. rischke , s. bernard and j. a. maruhn , nucl .
a * 595 * ( 1995 ) 346 [ nucl - th/9504018 ] .
a. adare _ et al .
_ [ phenix collaboration ] , phys .
c * 84 * ( 2011 ) 044905 [ arxiv:1005.1627 [ nucl - ex ] ] .
u. w. heinz , aip conf .
* 739 * ( 2005 ) 163 [ nucl - th/0407067 ] .
r. abir , c. greiner , m. martinez , m. g. mustafa and j. uphoff , arxiv:1109.5539 [ hep - ph ] .
r. abir , u. jamil , m. g. mustafa and d. k. srivastava , arxiv:1203.5221 [ hep - ph ] . | we develop a transport approach for heavy quarks in a quark - gluon plasma , which is based on improved binary collision rates taking into account quantum statistics , the running of the qcd coupling and an effective screening mass adjusted to hard - thermal loop calculations .
we quantify the effects of in - medium collisions by calculating the heavy flavor nuclear modification factor and the elliptic flow for rhic energies , which are comparable to radiative effects .
we also derive an analytic formula for the mean collisional energy loss of an energetic heavy quark in a streaming quark gluon plasma . elastic energy loss , mean energy loss , heavy quarks , jet quenching , nuclear modification factor , elliptic flow |
let us recall here some basic facts about the relationship between 3-edge colorings ( also called tait colorings ) and hamiltonian cycles in cubic graphs ; the details can be found in @xcite . by a _
cubic graph _ we understand a connected 3-regular multi - graph that is allowed to have double edges , but no loops ( obviously , a graph with loops can not have either edge colorings or hamiltonian cycles ) . consider a set of three distinct elements called `` colors '' ( say , @xmath0 , where @xmath1 stands for `` red '' , @xmath2 for `` green '' , and @xmath3 for `` blue '' ) .
3-edge coloring _ , or _ tait coloring _ is an assignment of a color to every edge such that the edges incident with each vertex have distinct colors .
every 3-edge coloring of a cubic graph @xmath4 gives rise to three distinct 2-factors ( that is , 2-regular spanning subgraphs ) of @xmath4 called _ tait cycles _ : each tait cycle is the union of edges painted in two colors out of the three ( the complement to a tait cycle is a perfect matching the union of disjoint edges painted in the third color ) .
if a cubic graph @xmath4 has a hamiltonian cycle , then @xmath4 also admits a 3-edge coloring that is unique up to a permutation of colors : just paint the hamiltonian cycle ( which always has even length ) in two intermittent colors , and paint the complement perfect matching in the remaining third color .
the above connection between 3-edge colorings and hamiltonian cycles suggests the following method of enumerating ( and actually finding ) all the hamiltonian cycles in a cubic graph : 1 .
find all 3-edge colorings ( tait colorings ) of a given cubic graph up to permutations of colors ; 2 .
find all the corresponding 2-factors ( tait cycles ) ; 3 .
check for connected 2-factors ( hamiltonian cycles ) .
this procedure gives the complete list of hamiltonian cycles in a cubic graph .
for the enumeration of 3-edge colorings we use an exhaustive backtracking algorithm that works in two runs . during the first run it dynamically separates all the vertices into two types : `` rigid '' ones
( that admit a unique coloring of edges incident with them ) , and `` soft '' ones ( with exactly two possibilities of coloring the incident edges ) ; no backtracking is needed on this stage .
the second run is the actual painting of edges : after it successfully colored the graph or was unable to complete the coloring , it returns to the last visited soft vertex and tries a different possibility .
the details are explained in the next section .
first we partition the set of vertices @xmath5 , of the graph @xmath4 into two disjoint subsets @xmath6 of rigid ( @xmath7 ) and soft ( @xmath8 ) vertices .
we note that this partition is not canonical .
initially we put @xmath9 and dynamically change their content .
we also introduce a temporary set @xmath10 of _ unidentified _ vertices that we already visited , and an ordered list of colored edges @xmath11 .
we label the vertices of @xmath4 by integers @xmath12 . for the vertex with number @xmath13
we denote the numbers of adjacent vertices by @xmath14 . an edge connecting @xmath13 and @xmath15 we denote by @xmath16 $ ] .
we start at the vertex @xmath17 and add it to the set @xmath7 of rigid vertices .
we add the three edges @xmath18,\,[0,n^0_1],\,[0,n^0_2]$ ] incident with it to the list @xmath11 , and we add their endpoints @xmath19 to @xmath10 .
now we check if any of the vertices in @xmath10 are the endpoints of at least two edges in @xmath11 . if this is the case , we move all such vertices from @xmath10 to @xmath7 , and for every such vertex we also add the remaining third edge incident to it to the set @xmath11 ( if it is not already there ) .
we continue the above procedure until there is no vertex left in @xmath10 that is an endpoint of at least two edges in @xmath11 .
now pick the vertex from @xmath10 with the smallest number , say @xmath13 , and move it into @xmath8 .
note that @xmath13 is an endpoint of a single edge in the current set @xmath11 .
next , we append to @xmath11 the two remaining edges incident with @xmath13 , add their endpoints to @xmath10 , and again check if any of the vertices in @xmath10 bound at least two edges in @xmath11 . if they do , such vertices are moved to @xmath7 , and the missing edges incident with these vertices are added to @xmath11 .
otherwise , we pick a vertex from @xmath10 with the smallest number , move it to @xmath8 and repeat the procedure until @xmath10 becomes empty , or , equivalently , until @xmath11 coincides with @xmath20 . since @xmath4 is connected , this would mean that @xmath21 . setting @xmath22 and @xmath23
we obtain the required partition . the final list of edges @xmath11 provides the order in which we attempt to paint the edges of the graph .
as above , we start at the vertex 0 and paint the edges @xmath18,\,[0,n^0_1],\,[0,n^0_2]$ ] incident with it in colors @xmath24 respectively . if the next edge in @xmath11 has a rigid vertex as its endpoint , then there is a unique color left for the remaining third edge incident with that vertex .
if the next pair of edges is incident with a soft vertex , we choose one of the two options for painting these edges ( in case we visit this vertex for the first time ) . continuing this way we may successfully reach the end of the list @xmath11 and get a complete edge coloring that we save for our record .
it may as well happen that the procedure ends prematurely when two edges incident to the same vertex are painted in the same color . in both cases we return in @xmath11 to the previous soft vertex that was visited only once , and start over painting edges in a different way . at the end
we get the list of all possible tait colorings and check which of them produce connected tait cycles ( that is , hamiltonian cycles ) . let us now illustrate this algorithm on a simple example .
consider the graph @xmath4 on 8 vertices @xmath25 shown on fig .
1 . we start with @xmath26 .
the initial vertex @xmath17 is a rigid vertex , @xmath27 .
add the edges @xmath28,[0,3],[0,7]$ ] to @xmath11 and the vertices @xmath29 to @xmath10 .
none of these vertices is an endpoint of two edges in @xmath11 , so we remove the vertex 1 from @xmath10 , append it to @xmath8 , and add the edges @xmath30 $ ] and @xmath31 $ ] to @xmath11 . again , none of the vertices in @xmath10 is an endpoint of two edges in @xmath11 , so we move the vertex 2 to @xmath8 and add the edges @xmath32 $ ] and @xmath33 $ ] to @xmath11 . now
the vertex 3 bounds two edges in @xmath11 , namely , @xmath34 $ ] and @xmath32 $ ] . we move it to @xmath7 ,
add the edge @xmath35 $ ] to @xmath11 and the vertex 4 to @xmath7 .
none of the vertices @xmath36 in @xmath10 bounds at least two edges , so we move the vertex 4 to @xmath8 and add the edges @xmath37 $ ] and @xmath38 $ ] to @xmath11 .
we see that the vertices 5 and 7 become rigid , and so does the vertex 6 .
thus , we get @xmath39 , and the ordered set of edges is @xmath40,[0,3],[0,7],[1,2],[1,6],[2,3],[2,5],[3,4],[4,5],[4,7],[5,6],[6,7]\}$ ] . let us start coloring the graph .
we paint the edges @xmath28,[0,3],[0,7]$ ] in colors @xmath41 respectively .
there are two possible ways of coloring the edges incident with the first soft vertex 1 .
we first paint @xmath30 $ ] in color @xmath2 and @xmath31 $ ] in color @xmath3 . at the next soft vertex 2
we again have two options , and choose the first one of them paint @xmath32 $ ] in color @xmath1 , and @xmath33 $ ] in color @xmath3 .
the edge @xmath35 $ ] incident with the rigid vertex 3 necessarily have color @xmath3 . at the last soft vertex 4 both possibilities of coloring the edges @xmath37,[4,7]$ ] lead to complete edge colorings of @xmath4 .
now we return to the previous soft point 2 and paint@xmath32 $ ] in @xmath3 , and @xmath33 $ ] in @xmath1 .
this gives us one more tait coloring ( one of the options can not be completed ) .
finally , returning to the first soft point 1 and painting the edges @xmath30,[1,6]$ ] in colors @xmath42 respectively , we get the last coloring .
thus , there exist 4 distinct tait colorings of the graph @xmath4 ( up to permutations of colors ) listed in the following table : [ cols= " < , < , < , < , < " , ] the first one of these 4 edge colorings has no hamiltonian cycles associated with it , whereas the other three ones produce two hamiltonian cycles each . they are listed below ( we indicate in brackets the corresponding alternating colors ) : @xmath43
it is clear that the computational complexity of this algorithm is of order @xmath44 since the set of soft vertices @xmath45 depends on the ordering of vertices of @xmath4 , the complexity also depends on this ordering . to give an upper bound for @xmath46 for a simple graph without double edges we note that , every time we add two new edges to @xmath11 incident with a soft vertex , we encounter one of the three possibilities : 1 .
both new endpoints belong to @xmath10 ; 2 .
one new endpoint belongs to @xmath10 an one to @xmath7 ( the vertex from @xmath7 then gives rise to new vertices that are added to @xmath47 ) ; 3 .
both new endpoints belong to @xmath7 . in any case , every soft vertex gives rise to at least two new vertices in @xmath48 and the lower bound 2 is attained when the both endpoints of the edges incident with a soft vertex belong to @xmath49 thus , when the union @xmath50 becomes equal to @xmath5 for the first time , we have the inequality @xmath51 in particular , it implies that @xmath52 at this stage .
let @xmath2 be the girth of @xmath4 ( that is , the length of the shortest cycle in @xmath4 ) .
when we reach the stage @xmath53 , every new soft vertex gives rise to at least @xmath54 rigid vertices .
this means that no more than @xmath55 vertices will be added to @xmath8 .
therefore , in the case @xmath56 , i.e. , when @xmath4 is triangle free , the number of soft vertices @xmath57 as @xmath58 .
thus , the speed of our algoritm is on a par with the fastest algoritms that find a single edge coloring or a single hamiltonian cycle in a cubic graph , or even better ( cf .
, e.g. , @xcite ) .
the absence of short cycles makes the algorithm even faster with complexity bounded from above by @xmath59 .
( note that the presence of double edges does not slow down the algorithm because at least one of their two common endpoints is rigid . )
it is instructive to compare the above complexity estimate with the results of @xcite .
let @xmath60 $ ] denote the eigenvalues of the ( normalized ) adjacency matrix of a simple cubic graph @xmath4 .
consider the mean@xmath61 and the variance @xmath62 of the exponents @xmath63 .
for each fixed @xmath64 the points @xmath65 form clusters called _ filars _ that enjoy a fractal - like structure . from the results of @xcite combined with the above considerations it follows that the closer is the point @xmath65 to the origin , the faster works our algorithm for the corresponding graph .
this algorithm was implemented in c++ code and compiled on a windows x86 machine ( pentium iv 3.40 ghz processor with 1 gb of ram ) using ccg gnu compiler ( the program code is given in appendix ) .
a good benchmark for testing programs that search for a hamiltonian cycle is provided by the horton graph @xcite , displayed on fig .
2 . it is a cubic bipartite graph on 96 vertices without hamiltonian cycles , but with many `` long '' cycles ( that is , cycles of length close to 96 ) .
some programs choke when they reach such a long cycle , not being able to transform it into a hamiltonian one .
our program completed the search in @xmath66 steps ( so that the actual complexity is of order @xmath67 ) .
the process took 6336 sec .
of machine time , found 143982592 tait colorings and no hamiltonian cycles .
our program is an open source program and its ansi c++ code is available at the following address : @xmath68 ( we do not present it here because of its length ) .
the code does not use any platform specific header files , and with minor modifications it can be compiled with essentially any c++ compiler that is not mentally challenged .
* acknowledgement * we thank j. filar for his interest in this work .
the work of ve , sr was supported , in part , by the australian research council discovery grant dp0666632 .
the work of pz was partially supported by the president of russian federation grant nsh - u329.2006.1 and by the russian foundation for basic research grant 05 - 01 - 00899 .
tutte , w. t. _ graph theory_. cambridge univ . press , 1984 .
eppstein , d. improved algorithms for 3-coloring , 3-edge - coloring , and constraint satisfaction , 12th acm - siam symp .
discrete algorithms , washington , 2001 , 329337 .
bondy , j. a. and murty , u. s. r. , _ graph theory with applications_. ny , north holland , 1976 .
ejov , v. v. , filar , j. a. , lukas , s. k. and zograf , p.g .
clustering of spectra and fractals of regular graphs .
preprint math.co/0610742 ( to appear in jmaa ) . | we propose an effective algorithm that enumerates ( and actually finds ) all 3-edge colorings and hamiltonian cycles in a cubic graph .
the idea is to make a preliminary run that separates the vertices into two types : `` rigid '' ( such that the edges incident to them admit a unique coloring ) and `` soft '' ones ( such that the edges incident to them admit two distinct colorings ) , and then to perform the coloring .
the computational complexity of this algorithm is on a par with ( or even below ) the fastest known algorithms that find a single 3-edge coloring or a hamiltonian cycle for a cubic graph . |
graph invariants constitute an interesting tool in chemistry , communication or engineering @xcite . in mathematics ,
one of the topics for which graph invariants have revealed to play an important role is the classical problem of deciding whether two algebras are isomorphic .
this problem is usually dealt with by computing the reduced grbner basis of the system of polynomial equations that is uniquely related to the structure constants of both algebras .
this computation is , however , very sensitive to the number of variables @xcite and gives rise to distinct problems of computation time and memory usage even for low - dimensional algebras @xcite .
this paper deals with graph theory in order to reduce this cost of computation .
graph invariants have been proposed in the last years as an efficient alternative to study isomorphisms of distinct types of algebras @xcite .
nevertheless , the problem of identifying a functor that relates the category of algebras with that of graphs remains still open .
based on a proposal of mckay et al .
@xcite for identifying isotopisms of latin squares with isomorphisms of vertex - colored graphs , we describe in section 3 a pair of graphs that enable us to find faithful functors between finite - dimensional algebras over finite fields and these types of graphs .
these functors map isomorphic and isotopic algebras to isomorphic graphs .
reciprocally , any pair of isomorphic graphs is uniquely related to a pair of algebras so that there exists a multiplicative map between them .
the main advantage of our proposal , apart from the reduction of the mentioned cost of computation , is the feasibility of studying the possible isomorphism between two given finite - dimensional algebras defined over the same field , whatever the types of both algebras are . as an illustrative example , we focus in section 4 on the classification of partial quasigroup rings according to the known isotopism classes of partial latin squares on which they are based .
in this section we expose some basic concepts and results on graph theory , isotopisms of algebras , partial latin squares and computational algebraic geometry that we use throughout the paper . for more details about these topics
we refer , respectively , to the manuscripts @xcite .
a _ graph _ is a pair @xmath0 formed by a set @xmath1 of _ vertices _ and a set @xmath2 of @xmath3-subsets of @xmath1 called _ edges_. two vertices defining an edge are said to be _
the _ degree _ of a vertex @xmath4 is the number @xmath5 of edges containing @xmath4 .
the graph @xmath6 is _ vertex - colored _ if there exists a partition of @xmath1 into color sets . the color of a vertex @xmath4 is denoted as @xmath7 .
an _ isomorphism _ between two vertex - colored graphs @xmath6 and @xmath8 is any bijective map @xmath9 between their sets of vertices that preserves collinearity and color sets , that is , such that it maps edges to edges and @xmath10 , for all vertex @xmath4 in @xmath6 .
two algebras @xmath11 and @xmath12 over a field @xmath13 are said to be _ isotopic _ if there exist three non - singular linear transformations @xmath9 , @xmath14 and @xmath15 from @xmath11 to @xmath12 such that @xmath16 , for all @xmath17 .
the triple @xmath18 is an _ isotopism _ between @xmath11 and @xmath12 .
if @xmath19 , then this constitutes an _ isomorphism_. the _ structure constants _ of an @xmath20-dimensional algebra @xmath11 over a field @xmath13 of basis @xmath21 are the numbers @xmath22 such that @xmath23 , for all @xmath24 . if all of them are zeros , then @xmath11 is _ abelian_. in particular , the @xmath20-dimensional abelian algebra is not isotopic to any other @xmath20-dimensional algebra .
the _ left annihilator _ of a vector subspace @xmath25 of the algebra @xmath11 is the set @xmath26 .
right annihilator _ is the set @xmath27 .
the intersection of both sets is the _ annihilator _ @xmath28 .
[ lemm_annihilator ] let @xmath18 be an isotopism between two @xmath20-dimensional algebras @xmath11 and @xmath12 , and let @xmath25 be a vector subspace of @xmath11 .
then , a. @xmath29 .
b. @xmath30 .
c. @xmath31 let us prove assertion ( a ) .
assertion ( b ) follows similarly and assertion ( c ) is a consequence of ( a ) and ( b ) .
let @xmath32 and @xmath33 .
then , @xmath34 , because @xmath35 and @xmath36 .
hence , @xmath37 @xmath38 .
now , let @xmath39 and @xmath40 . from the regularity of @xmath9 , we have that @xmath41 .
the regularity of @xmath15 involves that @xmath42 .
thus , @xmath43 and hence , @xmath44 .
the _ derived algebra _ of @xmath11 is the subalgebra @xmath45 .
[ lemm_m1 ] let @xmath18 be an isotopism between two @xmath20-dimensional algebras @xmath11 and @xmath12 .
then , @xmath46 .
the regularity of @xmath9 and @xmath14 involves that @xmath47 and hence , @xmath48 .
let @xmath49 be a partial binary operation over the set @xmath50=\{1,\ldots , n\}$ ] .
the pair @xmath51,\cdot)$ ] is called a _ partial magma _ of _ order _ @xmath20 .
it is _ isotopic _ to a partial magma @xmath51,\circ)$ ] if there exist three permutations @xmath52 , @xmath53 and @xmath54 in the symmetric group @xmath55 such that @xmath56 , for all @xmath57 such that @xmath58 exists .
if @xmath59 , then the partial magmas are said to be _
the triple @xmath60 is an _ isotopism _ of partial magmas ( an _ isomorphism _ if @xmath59 ) . a _ partial magma algebra _ @xmath61 _ based on _ a partial magma @xmath51,\cdot)$ ] is an @xmath20-dimensional algebra over a field @xmath13 such that there exists a basis @xmath21 satisfying that , if @xmath58 exists for some pair of elements @xmath57 , then @xmath62 for some non - zero structure constant @xmath63 .
if all the structure constants are equal to @xmath64 , then this is called a _ partial magma ring_. [ lemm_partial_magma ] two partial magma rings are isotopic ( isomorphic , respectively ) if their respective partial magmas on which they are based are isotopic ( isomorphic , respectively ) .
let @xmath61 and @xmath65 be two partial magma rings based , respectively , on two isotopic partial magmas @xmath51,\cdot)$ ] and @xmath51,\circ)$ ] .
let @xmath21 and @xmath66 be the respective bases of these two algebras and let @xmath18 be an isotopism between their corresponding partial magmas . for each @xmath67 ,
let us define the map @xmath68 .
then , @xmath69 . from linearity ,
the triple @xmath70 determines an isotopism between @xmath61 and @xmath65 . if @xmath19 , then this constitutes an isomorphism .
the reciprocal of lemma [ lemm_partial_magma ] is not true in general .
thus , for instance , the two partial magmas @xmath71,\cdot)$ ] and @xmath71,\circ)$ ] that are respectively described by the non - zero products @xmath72 and @xmath73 are not isotopic . nevertheless , the partial magma rings @xmath61 and @xmath65 , with respective bases @xmath74 and @xmath75 , are isotopic by means of the isotopism @xmath76 , where the linear transformation @xmath9 is described by @xmath77 and @xmath78 .
a _ partial quasigroup _ is a partial magma @xmath51,\cdot)$ ] such that if the equations @xmath79 and @xmath80 , with @xmath81 $ ] , have solutions for @xmath82 and @xmath83 in @xmath50 $ ] , then these solutions are unique .
the concepts of _ partial quasigroup algebras _ and _ partial quasigroup rings _ arise similarly to those of partial magma algebras and rings .
lemma [ lemm_partial_magma ] also holds analogously for partial quasigroup rings .
every partial quasigroup of order @xmath20 constitutes the multiplication table of a _ partial latin square _ of order @xmath20 , that is , an @xmath84 array in which each cell is either empty or contains one element chosen from the set @xmath50 $ ] , such that each symbol occurs at most once in each row and in each column .
every isotopism of a partial quasigroup is uniquely related to a permutation of the rows , columns and symbols of the corresponding partial latin square .
the distribution of partial latin squares into isotopism classes is known for order up six @xcite . in this paper
we make use of graph invariants to study which ones of the known non - isotopic classes of partial latin squares of order @xmath85 give rise to isotopic classes of partial quasigroup rings over the finite fields @xmath86 and @xmath87 . in this regard
, it is straightforwardly verified that there exists only two one - dimensional partial quasigroup rings : the abelian and that one described by the product @xmath88 .
they constitute distinct isotopism classes .
let @xmath89 be a partial latin square of order @xmath20 without empty cells ( that is , a _ latin square _ ) .
mckay et al .
@xcite defined the vertex - colored graph @xmath90 with @xmath91 vertices @xmath92 , where each of the four subsets ( related to the rows ( @xmath93 ) , columns ( @xmath94 ) , symbols ( @xmath95 ) and cells ( @xmath96 ) of the latin square @xmath97 ) has a different color , and @xmath98 edges @xmath99 ( see figure [ fig_ls ] , where we have used distinct styles ( @xmath100 , @xmath101 , @xmath102 , @xmath103 and @xmath104 ) to represent the colors of the vertices ) .
two latin squares @xmath105 and @xmath106 of the same order are isotopic if and only if the graphs @xmath107 and @xmath108 are isomorphic ( see theorem 6 in @xcite ) .
@xmath109{ls.eps}\end{array}\end{array}$ ] let @xmath110 $ ] be a multivariate polynomial ring over a field @xmath13 .
the _ algebraic set _ defined by an ideal @xmath111 of @xmath110 $ ] is the set @xmath112 of common zeros of all the polynomials in @xmath111 . if this set is finite , then the ideal @xmath111 is _ zero - dimensional_. this is _ radical _ if every polynomial @xmath113 $ ] belongs to @xmath111 whenever there exists a natural number @xmath114 such that @xmath115 .
the largest monomial of a polynomial in @xmath111 with respect to a given monomial term ordering is its _
leading monomial_. the ideal generated by all the leading monomials of @xmath111 is its _
initial ideal_. a _ standard monomial _ of @xmath111 is any monomial that is not contained in its initial ideal . regardless of the monomial term ordering , if the ideal @xmath111 is zero - dimensional and radical , then the number of standard monomials in @xmath111 coincides with the krull dimension of the quotient ring @xmath110/i$ ] and with the number of points of the algebraic set @xmath112 .
this is computed from the reduced grbner basis of the ideal .
specifically , a _
grbner basis _ of the ideal @xmath111 is any subset @xmath6 of polynomials in @xmath111 whose leading monomials generate its initial ideal .
this is _ reduced _ if all its polynomials are monic and no monomial of a polynomial in @xmath6 is generated by the leading monomials of the rest of polynomials in the basis .
there exists only one reduced grbner basis , which can always be computed from buchberger s algorithm @xcite .
the computation that is required to this end is extremely sensitive to the number of variables .
[ gao ] let @xmath116 be a finite field , with @xmath117 a power prime .
the complexity time that buchberger s algorithm requires to compute the reduced grbner bases of an ideal @xmath118 defined over a polynomial ring @xmath119 $ ] , where @xmath120 are polynomials given in sparse form and have longest length @xmath121 , is @xmath122 . here
, sparsity refers to the number of monomials .
grbner bases can be used to determine the isomorphisms and isotopisms between two @xmath20-dimensional algebras @xmath11 and @xmath12 over a finite field @xmath116 , with @xmath117 a prime power , respective basis @xmath21 and @xmath66 , and respective structure constants @xmath123 and @xmath124 . to this end , let us define the sets of variables @xmath125 , @xmath126 and @xmath127 .
these variables play the respective role of the entries in the regular matrices related to a possible isotopism @xmath18 between the algebras @xmath11 and @xmath12 .
here , @xmath128 , for each @xmath67 . from the coefficients of each basis vector @xmath129 in the expression @xmath130 , we have that @xmath131 [ thm_cag_isom ] the next two assertions hold .
a. the isotopism group between the algebras @xmath11 and @xmath12 is identified with the algebraic set of the ideal @xmath132 of @xmath133 $ ] , which is defined as @xmath134 + where @xmath135 , @xmath6 and @xmath136 denote , respectively , the matrices of entries in @xmath137 , @xmath138 and @xmath139 . besides , @xmath140/ i^{\mathrm{isot}}_{a , a'})$ ] .
b. the isomorphism group between the algebras @xmath11 and @xmath12 is identified with the algebraic set of the ideal @xmath141 of @xmath142 $ ] , which is defined as @xmath143 where @xmath135 denotes the matrix of entries in @xmath137 . besides , @xmath144/ i^{\mathrm{isom}}_{a , a'})$ ] .
let us prove the second assertion , being analogous the reasoning for assertion ( a ) .
the generators of the ideal @xmath141 involve each zero @xmath145 @xmath146 of its algebraic set to constitute the entries of the regular matrix of an isomorphism @xmath9 between the algebras @xmath11 and @xmath12 .
the result follows from the fact of being this ideal zero - dimensional and radical .
particularly , the ideal @xmath141 is zero - dimensional because its algebraic set is a finite subset of @xmath147 .
besides , from proposition 2.7 of @xcite , the ideal @xmath111 is also radical , because , for each @xmath57 , the unique monic generator of @xmath148 $ ] is the polynomial @xmath149 , which is intrinsically included in each ideal of @xmath142 $ ] and is square - free .
[ coro_cag_isom ] the complexity times that buchberger s algorithm requires to compute the reduced grbner bases of the ideals @xmath132 and @xmath141 in theorem [ thm_cag_isom ] are , respectively , @xmath150 and @xmath151 .
we prove the result for the second ideal , being analogous the reasoning for the first one .
the result follows straightforwardly from theorem [ gao ] once we observe that all the generators of the ideal in theorem [ thm_cag_isom ] are sparse in @xmath142 $ ] .
more specifically , the number of variables is @xmath152 , the number of generators of the ideal under consideration that are not of the form @xmath149 is @xmath153 and the maximal length of these generators is @xmath154 .
theorem [ thm_cag_isom ] has been implemented as a procedure called _ isoalg _ in the open computer algebra system for polynomial computations singular @xcite .
this has been included in the library _ graphalg.lib_ , which is available online at ` http://personales.us.es/raufalgan/ls/graphalg.lib ` .
let us illustrate the use of this procedure with an example related to the distribution of the set @xmath155 of two - dimensional partial quasigroup rings over the finite field @xmath86 into isotopism and isomorphism classes .
all the computations that are exposed throughout this paper are implemented in a system with an _ intel core i7 - 2600 , with a 3.4 ghz processor and 16 gb of ram_. [ ejemplo_pq2 ] let us consider the pair of partial quasigroup rings in @xmath155 that are respectively related to the partial latin squares @xmath156 these two partial latin squares are not isotopic because isotopisms preserve the number of filled cells . nevertheless
, their related partial quasigroup rings over @xmath86 , with respective bases @xmath74 and @xmath75 , and which are respectively described by the products @xmath157 are isotopic .
specifically , by implementing the procedure _ isoalg _ , our system computes in @xmath158 seconds the existence of four isotopisms between these two partial quasigroup rings .
one of this isotopisms is , for instance , the isomorphism @xmath9 such that @xmath77 and @xmath159 .
the procedure _ isoalg _ also ensures us the existence of @xmath9 as the unique possible isomorphism .
@xmath160 in practice , in those cases in which the run time required for the computations involved in theorem [ thm_cag_isom ] becomes excessive , it is recommendable to eliminate the generators of the corresponding ideal that are referred to the determinants of the matrices @xmath135 , @xmath6 and @xmath136 .
this reduces the complexity time in corollary [ coro_cag_isom ] to @xmath161 and @xmath162 , respectively , and gives enough information to analyze a case study on which base the possible isomorphisms and isotopisms between two given algebras , whatever the base field is .
the next example illustrates this fact by focusing on the possible isotopisms that there exist over any field between the two partial quasigroup rings that appear in example [ ejemplo_pq2 ] .
[ ejemplo_pq2a ] the implementation of the procedure _ isoalg _ enables us to ensure that , whatever the base field is , the reduced grbner basis of the ideal @xmath132 in theorem [ thm_cag_isom ] related to the isotopism group between the two partial quasigroup rings of example [ ejemplo_pq2 ] holds that @xmath163 and @xmath164 .
if the characteristic of the base field is not two , then @xmath165 .
this involves @xmath136 to be singular and hence , these two partial quasigroup rings are not isotopic .
otherwise , it is straightforwardly verified that the linear transformation @xmath9 that is indicated in example [ ejemplo_pq2 ] constitutes an isomorphism between both rings for every base field of characteristic two .
based on the proposal of mckay et al .
@xcite for latin squares , we describe now a pair of graphs that are uniquely related to a finite - dimensional algebra @xmath11 over a finite field @xmath13 .
firstly , we define the vertex - colored graph @xmath166 with four maximal monochromatic subsets @xmath167 , @xmath168 , @xmath169 and @xmath170 , and edges @xmath171 . from this graph
we also define the vertex - colored graph @xmath172 by adding the edges @xmath173 . as an illustrative example ,
figure [ fig_1 ] shows the two graphs that are related to any @xmath20-dimensional algebra over the finite field @xmath86 , with basis @xmath21 , that is described as @xmath174 .
@xmath175{e1e2_e1_isot.eps } & \includegraphics[width=6cm]{e1e2_e1_isom.eps } \end{array}$ ] [ lemm_graph0 ] the next assertions hold .
a. if the algebra @xmath11 is abelian , then @xmath166 and @xmath172 have no vertices .
b. the graph @xmath166 does not contain triangles . c. in both graphs @xmath166 and @xmath172 , * the number of vertices is @xmath176 is @xmath177 d. in the graph @xmath166 , * @xmath178 , for all @xmath179 .
* @xmath180 , for all @xmath181 . * @xmath182 , for all @xmath183 , where @xmath184 is the _ adjoint action _ of @xmath4 in @xmath11 such that @xmath185 , for all @xmath186 .
e. let @xmath187 denotes the characteristic function .
then , in the graph @xmath172 , * @xmath188 , for all @xmath179 . * @xmath189 , for all @xmath181 . * @xmath190 , for all @xmath183 .
the result follows straightforwardly from the definition of the graphs @xmath166 and @xmath172 .
[ prop_graph0_a ] the next assertions hold .
a. the number of edges of the graph @xmath166 is @xmath191 @xmath192 .
b. the number of edges of the graph @xmath172 coincides with those of @xmath166 plus @xmath193 .
the result follows from the first theorem of graph theory @xcite , which enables us to ensure that the number of edges of a graph is the half of the summation of degrees of its vertices .
now , for each pair of vectors @xmath17 such that @xmath194 , the vertex @xmath195 is the only vertex in @xmath196 that is adjacent to the vertices @xmath197 , @xmath198 and @xmath199 .
they constitute indeed the three vertices related to the degree of @xmath200 that is indicated in assertion ( c ) of lemma [ lemm_graph0 ] . as a consequence ,
the summation of degrees of all the vertices in @xmath196 coincides with @xmath201 .
the result follows then from assertions ( d ) and ( e ) in lemma [ lemm_graph0 ] .
[ theo_graph0 ] let @xmath11 and @xmath12 be two finite - dimensional algebras over a finite field @xmath13 .
then , a. if both algebras are isotopic , then their corresponding graphs @xmath166 and @xmath202 are isomorphic .
reciprocally , if the graphs @xmath166 and @xmath202 are isomorphic , then there exist three bijective maps @xmath9 , @xmath14 and @xmath15 between @xmath11 and @xmath12 such that @xmath16 .
b. if both algebras are isomorphic , then their corresponding graphs @xmath172 and @xmath203 are also isomorphic .
reciprocally , if the graphs @xmath172 and @xmath203 are isomorphic , then there exists a multiplicative bijective map between the algebras @xmath11 and @xmath12 , that is , a bijective map @xmath204 so that @xmath205 , for all @xmath17 .
let @xmath18 be an isotopism between the algebras @xmath11 and @xmath12 .
we define the map @xmath52 between @xmath166 and @xmath202 such that @xmath206 the description of @xmath166 and @xmath202 , together with lemmas [ lemm_annihilator ] and [ lemm_m1 ] , and the regularity of @xmath9 , @xmath14 and @xmath15 , involves @xmath52 to be an isomorphism between these two graphs . the same map @xmath52 constitutes an isomorphism between the graphs @xmath172 and @xmath203 in case of being @xmath19 , that is , if the algebras @xmath11 and @xmath12 are isomorphic .
reciprocally , let @xmath52 be an isomorphism between the graphs @xmath166 and @xmath202 .
collinearity involves this isomorphism to be uniquely determined by its restriction to @xmath207 .
specifically , the image of each vertex @xmath195 by means of @xmath52 is uniquely determined by the corresponding images of @xmath208 , @xmath209 and @xmath210 .
let @xmath53 and @xmath211 be the respective bases of the algebras @xmath11 and @xmath12 and let @xmath212 be the natural map that preserves the components of each vector with respect to the mentioned bases .
that is , @xmath213 , for all @xmath214 .
let us define three maps @xmath9 , @xmath14 and @xmath15 from @xmath11 to @xmath12 such that @xmath215 @xmath216 @xmath217 from lemmas [ lemm_annihilator ] and [ lemm_m1 ] , these three maps are bijective .
let @xmath17 . if @xmath218 or @xmath219 , then there does not exist the vertex @xmath200 in the graph @xmath166 .
since @xmath52 preserves collinearity , there does not exist the vertex @xmath220 in the graph @xmath202 , which means that @xmath221 or @xmath222 . in any case
, we have that @xmath223 . finally , if @xmath224 and @xmath225 , then the vertex @xmath200 connects the vertices @xmath208 , @xmath209 and @xmath210 in the graph @xmath166 .
now , the isomorphism @xmath52 maps this vertex @xmath200 in @xmath166 to a vertex @xmath226 in @xmath172 that is connected to the vertices @xmath227 , @xmath228 and @xmath229 .
again , since @xmath52 preserves collinearity , it is @xmath230 , @xmath231 and , finally , @xmath232 . in case of being @xmath52 an isomorphism between the graphs @xmath172 and
@xmath203 it is enough to consider @xmath19 in the previous description .
this is well - defined because of the new edges that are included to the graphs @xmath166 and @xmath202 in order to define , respectively , the graphs @xmath172 and @xmath203 .
these edges involve the multiplicative character of the bijective map @xmath9 , that is , @xmath16 , for all @xmath17 .
theorem [ theo_graph0 ] enables us to ensure that graph invariants reduce the cost of computation that is required to distribute a set of finite - dimensional algebras over finite fields into isotopism and isomorphism classes .
it is only necessary to compute those reduced grbner bases in theorem [ thm_cag_isom ] that are related to a pair of algebras whose associated graphs have equal invariants .
the complexity to compute these invariants is always much less than that related to the calculus of a reduced grbner basis .
thus , for instance , the complexity to compute the number of vertices , edges and triangles of the graphs related to any @xmath20-dimensional algebra over the finite field @xmath116 is @xmath233 .
this corresponds to the computation of the adjacency matrices of both graphs by means of all the possible products among the @xmath234 distinct vectors of any such an algebra .
this enables us in particular to implement the formulas exposed in lemma [ lemm_graph0 ] and proposition [ prop_graph0_a ] . besides ,
recall that the trace of the adjacency matrix of a graph raised to the third power coincides with the number of triangles of such a graph .
all this computation has been implemented in the procedure _ isograph _ , which has been included in the mentioned library _ graphalg.lib_. in order to illustrate the efficiency of these invariants , we focus on the set @xmath235 of @xmath20-dimensional lie algebras over the finite field @xmath116 , with @xmath117 a power prime . recall that a _ lie algebra _ is an anti - commutative algebra @xmath11 that holds the _
jacobi identity _
@xmath236 , for all @xmath237 .
for @xmath238 , it is known @xcite that there are @xmath239 distinct lie algebras in @xmath240 , which are distributed into four isotopism classes and six isomorphism classes , and @xmath241 lie algebras in @xmath242 , which are distributed into four isotopism classes and seven isomorphism classes .
table [ table_0 ] shows the run time and memory usage that our computer system requires to determine the mentioned classification depending on whether graph invariants are considered ( _ graph _ ) or not ( _ alg _ ) .
further , tables [ table_1 ] and [ table_1a ] show , respectively , the invariants of the graphs @xmath243 and @xmath244 related to the isomorphism classes of @xmath245 , for @xmath246 .
the components of the @xmath247-tuples that are indicated in the corresponding columns of vertices in both tables refer , respectively , to the number of vertices in @xmath248 , @xmath249 , @xmath250 and @xmath196 .
let us finish with an illustrative example that focuses on those graphs @xmath243 and @xmath244 related to the set @xmath251 of @xmath20-dimensional partial quasigroup rings over a finite field @xmath13 that are based on the known distribution of partial latin squares of order @xmath85 into isotopism classes . in this
regard , table [ table_pq ] shows several graph invariants that are related to the isotopism classes of @xmath252 , for @xmath246 .
partial latin squares are written row after row in a single line , with empty cells represented by zeros .
[ theo_pq2a ] the set @xmath253 is distributed into seven isotopism classes , whatever the base field is . a computational case study based on a similar reasoning to that exposed in example [ ejemplo_pq2a ] enables us to ensure the result .
if the characteristic of the base field is distinct from two , then the seven isotopism classes under consideration are those related to the next partial latin squares of order @xmath3 @xmath254 otherwise , if the characteristic of the base field is two , then the isotopism classes related to the last two partial latin squares coincide . in this case , the next partial latin square corresponds to the seventh isotopism class @xmath255 if the characteristic of the base field is distinct from two , the partial quasigroup ring related to this partial latin square is isotopic to that related to the unique latin square of the previous list .
it is known @xcite that there are @xmath3 , @xmath256 and @xmath257 distinct isotopism classes of partial latin squares of respective orders @xmath64 to @xmath258 . in order to determine those distinct isotopism classes that give rise to isotopic partial quasigroup rings over the finite field @xmath86 ,
we have implemented the procedure _ isoalg _ in our previously mentioned computer system . with a total run time of @xmath259 seconds
, we have obtained that there exist @xmath3 , @xmath260 and @xmath261 distinct isotopism classes of partial quasigroup rings of respective dimensions @xmath64 to @xmath258 .
particularly , the existence of two classes for the one - dimensional case agrees with that exposed in subsection 2.3 . besides , the seven isotopism classes for the two - dimensional case agrees with theorem [ theo_pq2a ] . for the three - dimensional case ,
the next nine pairs of non - isotopic partial latin squares give rise to isotopic partial quasigroup rings @xmath262 the run time of 761 seconds that are required to determine the mentioned distribution of partial quasigroup rings reduces to only 30 seconds in the same computer system if the invariants that we have just exposed in table [ table_pq ] and those exposed in table [ table_pq3 ] are previously computed .
the new run time includes indeed the extra 9 seconds of computation that is required for computing such invariants .
we have described in this paper a pair of graphs that enable us to define faithful functors between finite - dimensional algebras over finite fields and these graphs .
the computation of related graph invariants plays a remarkable role in the distribution of distinct families of algebras into isotopism and isomorphism classes .
some preliminary results have been exposed in this regard , particularly on the distribution of partial quasigroup rings over finite fields .
based on the known classification of partial latin squares into isotopism classes , further work is required to determine completely this distribution . | in this paper we define a pair of faithful functors that map isomorphic and isotopic finite - dimensional algebras over finite fields to isomorphic graphs .
these functors reduce the cost of computation that is usually required to determine whether two algebras are isomorphic . in order to illustrate their efficiency ,
we determine explicitly the classification of two- and three - dimensional partial quasigroup rings .
graph theory , finite field , isomorphism , latin square .
05c25 , 05c30 , 05b15 . |
in perturbative qcd the states @xmath3 and @xmath2 are non - relativistic bound states of a charm and an anti - charm quark .
the decays of these states are expected to be dominated by the annihilation of the constituent @xmath4 into three gluons .
the partial width for the decays into an exclusive hadronic state , @xmath5 , is expected to be proportional to the square of the @xmath4 wave function overlap at the origin , which is well determined from the leptonic width @xcite .
since the strong coupling constant , @xmath6 , is not very different at the @xmath3 and @xmath2 masses , it is expected that for any state @xmath5 the @xmath3 and @xmath2 branching ratios are related by @xmath7 where @xmath8 denotes a branching fraction , @xmath5 is a particular hadronic final state , and the leptonic branching fractions are taken from the pdg @xcite .
this relation is sometimes called the 12% rule .
modest deviations from the rule are expected @xcite .
although the rule works well for some specific decay modes of the @xmath2 , it fails spectacularly for @xmath2 decays to final states consisting of one vector and one pseudoscalar meson ( vp ) , such as @xmath9 .
values of @xmath10 have been measured for a wide variety of final states @xcite .
a recent review @xcite of relevant theory and experiment concludes that current theoretical explanations are unsatisfactory .
clearly more experimental results are desirable .
this paper presents measurements of the following new decay modes of the @xmath2 : @xmath11 , @xmath9 , @xmath12 , @xmath13 , @xmath14 , @xmath15 , @xmath16 , @xmath17 , @xmath18 , @xmath19 , @xmath20 , @xmath21 , and more precise measurements of these previously measured modes : @xmath22 , @xmath23 , @xmath24 , @xmath25 , @xmath26 , @xmath27 , @xmath28 , @xmath29 , @xmath30 , @xmath31 , @xmath32 , @xmath33 , @xmath34 , @xmath35 , and @xmath36 .
we also obtain an improved upper limit for @xmath37 .
where applicable , the inclusion of charge conjugate modes is implied .
we select intermediate final states in the following decay modes : @xmath38 , @xmath39 , @xmath40 , @xmath41 , @xmath42 and @xmath11 , @xmath43 , @xmath44 and @xmath45 , @xmath46 and @xmath47 .
the data sample used in this analysis is obtained at the @xmath2 and the nearby continuum in @xmath1 collisions produced by the cornell electron storage ring ( cesr ) and acquired with the cleo detector .
the datasets consist of @xmath48=5.46 pb@xmath0 on the peak of the @xmath2 ( 2.57 pb@xmath0 for cleo iii @xcite , 2.89 pb@xmath0 for cleo - c @xcite ) and 20.5 pb@xmath0 at @xmath49=3.67 gev ( all cleo - c ) .
the nominal scale factor used to normalize continuum yields to @xmath2 data is @xmath50 , and is determined from the integrated luminosities of the data sets corrected for the @xmath51 dependence of the cross section , where the error is from the relative luminosity uncertainty .
the actual @xmath52 used for each mode also corrects for the small differences in efficiency between the @xmath2 and continuum data samples .
standard requirements are used to select charged particles reconstructed in the tracking system and photon candidates in the csi calorimeter .
charged particles are identified using a likelihood function that incorporates de / dx from the drift chamber and information from the rich .
the invariant mass of the decay products from intermediate resonant states such as @xmath53 , @xmath54 , @xmath55 , @xmath56 , @xmath57 , @xmath58 , @xmath59 and @xmath60 must lie within limits determined from monte carlo ( mc ) studies .
energy and momentum conservation requirements are imposed on the reconstructed final state hadrons , which have momentum @xmath61 and combined measured energy @xmath62 .
we require the measured scaled energy @xmath62/@xmath63 be consistent with unity within experimental resolution , which varies by final state .
we require @xmath64 .
together these requirements suppress backgrounds with missing energy or incorrect mass assignments .
the experimental resolutions are smaller than 1% in scaled energy and 2% in scaled momentum . for the modes with two or more charged pions , two @xmath65 or an @xmath54 , we reject @xmath66 ( @xmath67 , @xmath68 , or @xmath54 ) events in which the mass of any of the following falls within the range @xmath69 gev : the two highest momentum oppositely charged tracks , the recoil mass against the two lowest momentum oppositely charged tracks , or the mass recoiling against the 2@xmath65 or @xmath54 . for every final state , a signal selection range in @xmath62/@xmath63 is determined by mc simulation , and a sideband selection range is defined to measure background .
final states with intermediate particles must satisfy a scaled energy signal selection range requirement identical to the corresponding mode without the intermediate particle ; and the event yield is determined from signal and sideband selection ranges of the intermediate particle mass .
the number of events attributable to each @xmath2 decay mode is obtained by subtracting the number of events in the sideband region and the scaled continuum production from the number of events in the signal region .
interference between @xmath2 decay and continuum production of the same final state is neglected .
the efficiency for each final state is the average obtained from mc simulations for both detector configurations ; the two values are typically within a few percent ( relative ) to each other .
no initial state radiation is included in the mc , but final state radiation is accounted for .
we correct the number of signal for the efficiency and normalize to the number of @xmath70 , @xmath71 decays in the data , which has been determined previously to be @xmath72 @xcite .
the resulting relative branching ratios are obtained and then converted to the absolute branching ratios using @xmath73 @xcite and @xmath74 @xcite .
@xmath10 values are determined using the absolute @xmath2 branching ratios determined in this analysis and @xmath3 branching ratios from @xcite .
the branching ratios and @xmath10 values for 2-body modes are given in table [ tab:2body ] , where bes numbers are from @xcite ; preliminary results for multi - body modes @xcite are listed in table [ tab : multi ] . [ tab:2body ] [ tab : multi ]
we have presented first evidence for @xmath2 decays to @xmath11 , @xmath9 , @xmath12 , @xmath13 , and @xmath75 .
measurements for several other vp channels are also given .
the results suggest that , for vp final states , @xmath2 decays through three gluons are severely suppressed with respect to the 12% rule and the corresponding electromagnetic processes are not . the decay @xmath76 exhibits a @xmath38 signal but has a much larger component at higher @xmath77 mass .
we also present preliminary branching fractions for seven new multi - body decay modes of the @xmath2 , namely @xmath15 , @xmath16 , @xmath17 , @xmath18 , @xmath19 , @xmath20 , @xmath21 , and more precise measurements of nine previously measured modes , which are @xmath27 , @xmath28 , @xmath29 , @xmath30 , @xmath31 , @xmath32 , @xmath33 , @xmath34 , @xmath35 , and @xmath36 .
we also obtain an improved upper limit for @xmath37 . for multi - body decay modes
@xmath10 values are in general lower than , but , in most cases , consistent with the 12% rule .
we gratefully acknowledge the effort of the cesr staff in providing us with excellent luminosity and running conditions .
this work was supported by the national science foundation , the u.s .
department of energy , the research corporation , and the texas advanced research program .
cleo collaboration , y. kubota _ et al .
methods phys .
res . , sect .
a * 320 * , 66 ( 1992 ) ; d. peterson _ et al . _ ,
methods phys .
res . , sect .
a * 478 * , 142 ( 2002 ) ; m. artuso _ et al . _ ,
methods phys .
a * 502 * , 91 ( 2003 ) . | using 5.46 pb@xmath0 of @xmath1 annihilation data collected at the @xmath2 with the cleo detector we have observed a variety of new hadronic decay modes of the @xmath2 .
a comprehensive set of branching ratios and upper limits is presented . |
the observed masses and mixings of the quarks and leptons will provide a promising clue to a unified understanding of those fundamental particles .
for such the purpose , not only investigating a theoretical model , but also searching for phenomenological mass matrix seem to be still effective . as one of phenomenological mass matrix models of the quarks and leptons , the so - called yukawaon " model @xcite ( a kind of flavon " models @xcite ) has been proposed . here ,
the effective " yukawa coupling constants @xmath15 are given by @xmath16 @xmath17 is an energy scale of the effective theory , and @xmath18 are vacuum expectation value ( vev ) matrices of scalar fields @xmath19 with @xmath20 components .
( hereafter we call fields @xmath19 yukawaons " . )
the most characteristic point in the yukawaon model is that all vevs @xmath18 are described in terms of only one fundamental vev matrix @xmath21 .
for example , in a yukawaon model with an o(3 ) family symmetry @xcite , the charged lepton , neutrino , up - quark , and down - quark mass matrices @xmath1 , @xmath0 , @xmath2 and @xmath22 are given by @xmath23 @xmath24 @xmath25 @xmath26 @xmath27 @xmath28 @xmath29 where @xmath30 in the diagonal basis of @xmath1 , while @xmath31 is given by a form @xmath32 in the diagonal basis of @xmath2 . in this paper , we will find out a special form of the neutrino mass matrix which is compatible with the observed neutrino data in spite of having no adjustable parameters .
the form will be obtained along the lines of the yukawaon model by changing the structure of @xmath33 from eq.(1.2 ) .
the neutrino mass matrix is still given by the form of seesaw type , @xmath34 with @xmath35 , but the majorana neutrino mass matrix @xmath36 is changed into a simple form @xmath37 here , we assume that the charged lepton mass matrix @xmath1 is given by @xmath38 differently from eq.(1.2 ) , and we also redefine @xmath39 as @xmath40 where @xmath41 and @xmath42 note that there are no @xmath43 term , no @xmath44 term and no @xmath31 in eq.(1.11 ) , i.e. @xmath0 is simply given by @xmath45 in other words , there is no adjustable parameter in the present neutrino mass matrix , except for @xmath3 in @xmath1 and @xmath4 in @xmath46 ( i.e. @xmath39 ) .
the purpose of the present paper is not to derive the mass matrix forms ( 1.11 ) - ( 1.13 ) theoretically , but to demonstrate that the phenomenological neutrino mass matrix ( 1.15 ) with eqs.(1.11 ) - ( 1.13 ) can be compatible with the present neutrino data in spite of quite few parameters .
we predict @xmath5 , @xmath6 , @xmath7 , @xmath8 ev , and so on , by using observed values of @xmath9 , @xmath10 , and @xmath12 as input values . in this paper , for simplicity , we do not discuss the down - quark mass matrix @xmath22 and the cabibbo - maskawa- kobayashi @xcite ( ckm ) mixing . in the next section , superpotentials for yukawaons and the assignments of the fields in the present u(3 )
yukawaon model are investigated . in sec . 3 ,
numerical results of the model are discussed .
is devoted to the concluding remarks . in appendix
a , @xmath47-charge assignments are discussed . in appendix b
, we present a rotation matrix which transforms @xmath48 into @xmath49 .
in this section , we give superpotentials for the yukawaons and the assignments of the fields in the present u(3 ) yukawaon model . in the yukawaon model ,
the order of the fields is important .
therefore , in this paper , let us assume a u(3 ) family symmetry instead of o(3 ) and denote fields @xmath50 and @xmath51 of u(3 ) as @xmath52 and @xmath53 , respectively .
( therefore , it should be noted that a term @xmath54 is allowed , but @xmath55 and @xmath56 are forbidden . ) in the u(3 ) model , for example , the relation ( 1.6 ) is re - expressed as @xmath57 with @xmath58 . in order to distinguish each yukawaon from other yukawaons , although we assumed u(1)@xmath59 charge in the o(3 ) model @xcite , in this u(3 ) model , we assume only @xmath47 charge conservation instead of u(1)@xmath59 charge conservation . for the right handed neutrino sector ( @xmath60 ) , it should be noted that we can not add @xmath61 term to @xmath60 as in eq.(1.4 ) . in the old model , we assigned the u(1)@xmath59 charges @xmath62 only for gauge singlet fields , e.g. @xmath63 , i.e. @xmath64 . besides , we assumed @xmath65 in order to build a model without @xmath66 .
therefore , we could obtain @xmath67 in the old model .
however , in this u(3 ) model , we can not obtain @xmath68 .
besides , we can not introduce a @xmath43 term such as in eq.(1.5 ) .
we assume the following superpotential @xmath69 : @xmath70 @xmath71 @xmath72 + \frac{\lambda_e}{\lambda } { \rm tr } [ \bar{\phi}_0 ( e'+ a_e x_2 ) \bar{\phi}_0 \theta^e ] , \eqno(2.3)\ ] ] @xmath73 + \frac{\lambda_d}{\lambda } { \rm tr } [ \bar{\phi}_0 ( e + a_d e^{i\alpha_d } x_3 ) \bar{\phi}_0 { \theta}^d ] , \eqno(2.4)\ ] ] @xmath74 + \frac{\lambda^{\prime}_u}{\lambda } { \rm tr } [ \bar{\phi}_0 ( e + a_u e^{i\alpha_u } x_3 ) \bar{\phi}_0 \theta^{u \prime } ] , \eqno(2.5)\ ] ] @xmath75 + \frac{\lambda_{u}}{\lambda } { \rm tr}[\bar{\phi}_u { e}^u \bar{\phi}_{u } { \theta}_u ] + \frac{\lambda_{0u}}{\lambda^{5 } } { \rm tr } [ ( \bar{e}_u e^u)^{3 } \bar{e}_u \theta^u ] , \eqno(2.6)\ ] ] @xmath76 + \frac{\lambda_r}{\lambda}{\rm tr } [ \left ( \bar{\phi}_u e^u \bar{y}_e + \bar{y}_e { e}^u \bar{\phi}_u \right ) { \theta}^r ] , \eqno(2.7)\ ] ] where , in eq.(2.2 ) , @xmath77 and @xmath78 are su(2)@xmath79 doublet fields , and @xmath80 ( @xmath81 ) are su(2)@xmath79 singlet fields .
the other fields in eqs.(2.3)-(2.7 ) have quantum numbers defined in table 1 . .
assignments of @xmath47 charges , where @xmath82 , @xmath83 and @xmath84 .
the values in the third raw denote @xmath47 charge values in a special case under the assumptions ( a.11 ) and ( a.14 ) . for more details , see eqs.(a.1 ) - ( a.20 ) in appendix a. [ cols="^,^,^,^,^,^",options="header " , ]
in this paper we have found out a special form of the neutrino mass matrix @xmath0 based on a yukawaon model with u(3 ) family symmetry , which has quite few free parameters . with this form of @xmath0 ,
the @xmath0 can give reasonable predictions in spite of having no adjustable parameters , i.e. the @xmath0 is simply given by the form ( 1.15 ) , and the mass matrices @xmath1 and @xmath2 include parameters @xmath3 and @xmath4 , respectively , which are fixed by their observed mass ratios . in this yukawaon model , the yukawaon vev matrices are described in terms of a new fundamental vev matrix @xmath85 .
for example , the yukawaon vev matrix @xmath86 for the charged leptons is given by ( 2.10 ) which has the structure of @xmath87 with a new parameter @xmath3 .
this structure in @xmath86 has been chosen from a phenomenological point of view and there is no reason why @xmath86 takes such a form .
nevertheless if we accept the form ( 2.10 ) , then we can obtain a simple form of vev matrix @xmath88 for right - handed neutrinos without introducing the somewhat strange vev matrix @xmath89 and @xmath43 term that were introduced in the o(3 ) model to get the observed nearly tribimaximal neutrino mixing @xcite .
the new model has only four parameters @xmath90 as far as the up - quark and lepton sectors are concerned .
on the other hand , we have 12 observable quantities ( 2 up - quark mass ratios , 4 lepton mass ratios , and 4 + 2 lepton mixing parameters ) .
the parameter @xmath91 affects only the prediction of @xmath92 , so that we have fixed it by the observed value of @xmath93 .
the parameter @xmath94 is sensitive only to @xmath11 , so that we have fixed by the observed value of @xmath95 as seen in fig . 2 ( b ) .
the parameter @xmath3 is determined from the cross point of the predicted values of @xmath96 and @xmath97 in the @xmath3-@xmath4 plane .
note that we have used only the observed values of @xmath96 and @xmath97 in order to fix the three parameters @xmath98 .
although we have tacitly used @xmath99 , we have not used the observed value of @xmath100 explicitly . on the other hand , for the remaining 2 down - quark mass ratios and 4 ckm mixing parameters , we have additional 2 parameters ( @xmath101 and @xmath102 ) .
regrettably , we can not obtain reasonable predictions with the two parameters , although we can fit the values of down - quark mass ratios and @xmath103 .
the situation is the same as in the previous o(3 ) model .
we must introduce a phase matrix @xmath104 with two parameters @xmath105 and a common mass shift term @xmath106 .
then , five parameters can fit six observables barely .
therefore , the model is not so attractive for down - quark sector . in this paper
, we did not demonstrate the explicit numerical fitting for down - quark mass rations and ckm mixing parameters .
\(i ) the model predicts @xmath107 . in the previous o(3 ) model ,
the predicted value of @xmath108 was invisibly small , i.e. @xmath109 . the t2k experiment @xcite put a constraint @xmath110 ( @xmath111 c.l . ) for @xmath112 and a normal hierarchy .
our predicted value @xmath113 seems to be somewhat lower than the experimental lower bound .
however , as seen in table 3 , our prediction on @xmath114 gives @xmath115 , which decreases the lower bound @xmath116 of the t2k result to @xmath117 .
besides , the double chooz experiment @xcite has reported that @xmath118 at @xmath119 cl .
the lower value is @xmath120 .
therefore , we consider that the predicted value @xmath113 is yet not ruled out , although the status is considerably severe .
\(ii ) it also predicts a reasonable value of @xmath121 in contrast to the case of the o(3 ) model in which we could not predict @xmath122 .
( in the previous model , the value of @xmath122 needed to adjust the additional free parameter @xmath44 in eq.(1.4 ) . )
\(iii ) the present model gives approximately degenerate neutrino masses @xmath123 .
the predicted value for the effective majorana mass @xmath8 ev in the neutrinoless double beta decay will be within our reach of the future experiments . the big ansatz is the existence of the @xmath124 term in the charged lepton sector ( 1.12 ) . at present
, there is no idea on this term .
besides , it seems that the present lepton mass structure is ill matched with the charged lepton mass relation @xcite @xmath125 the purpose in the early stage of the yukawaon model was to predict the charged lepton mass relation ( 4.1 ) . the bilinear form ( 1.2 ) for
the charged lepton mass matrix was indispensable to predict @xcite the relation ( 4.1 ) .
if we adopt the present scenario , we must reconsider the origin of the charged lepton mass spectrum .
however , in this paper , we do not use the relation ( 4.1 ) , but only use the observed charged lepton mass values as input values . therefore the bilinear form such as ( 1.2 ) is not necessarily required in this paper .
nevertheless , the formula ( 4.1 ) is still attractive . on the other hand ,
it is also attractive that we can predict 12 observables ( 2 + 2 + 2 lepton and up - quark mass ratios , and 4 + 2 pmns mixing parameters ) under 4 adjustable parameters @xmath90 if once we accept this ansatz ( 1.12 ) .
it is a future task how to understand the existence of @xmath124 term . in conclusion , although the form @xmath0 is , at present , not one which is derived from a rigid theoretical ground , the form will offer a suggestive hint for a unification model of quark and lepton mass matrices . in the present model , as well as in the o(3 ) model , we construct a model without introducing a yukawaon @xmath66 by replacing @xmath66 by @xmath126 .
the simple way to guarantee that the yukawaon @xmath126 couples not only to the charged lepton sector but also to the dirac neutrino sector is to introduce the following @xmath47 charge assignment , @xmath127
@xmath128 the @xmath47 charge of @xmath129 is free parameter in the form ( 2.8 ) .
for simplicity , we take @xmath130 hereafter , we will denote @xmath131 and @xmath132 as @xmath133 and @xmath134 , respectively .
each yukawaon is distinguished from other yukawaons by the @xmath47 charges .
if we define a parameter @xmath135 as @xmath136 , \eqno(a.4)\ ] ] then , we can express the @xmath47 charges of the other fields from eq.(2.2 ) as follows : @xmath137 @xmath138 @xmath139 @xmath140 @xmath141 @xmath142 @xmath143 from eqs.(2.6 ) and ( 2.7 ) , we obtain @xmath144 @xmath145 respectively . from eqs.(a.12 ) and ( a.13 ) , we obtain a relation @xmath146 the relation ( a.14 ) leads to @xmath147 only when the value @xmath135 is a positive integer , eq.(a.15 ) means that an additional term @xmath148 , \eqno(a.16)\ ] ] can appear in the expression ( 2.6 ) .
note that if @xmath135 is not a positive integer , the factor @xmath149 does not have a physical meaning , because a term with @xmath150 can not appear in the superpotential terms .
therefore , the @xmath135 defined in eq.(a.4 ) is allowed only for @xmath151 . as we see in eqs.(a.5 ) - ( a.11 ) , these @xmath47 charges are described by four parameters @xmath152 , @xmath153 , @xmath133 and @xmath135 .
therefore , in order to fix these @xmath47 charge values , we have to assume four constraints for these @xmath47 charges . on the other hand ,
the fields @xmath154 , @xmath60 , @xmath155 , @xmath156 , @xmath157 , and @xmath158 are gauge singlets , so that they must be distinguished only by @xmath47 charges .
we can choose a suitable parameter set @xmath159 . here
, let us demonstrate an example of @xmath47 charge assignments , although it is not the purpose of the present paper to give such an explicit @xmath47 charge assignment . for example , we put the following working hypothesis : @xmath160 @xmath161 the constraint ( a.17 ) is an analogy that the yukawa coupling constants in the standard model do not have @xmath47 charges . the constraints ( a.17 ) and (
a.18 ) leads to the relation @xmath162 .
of course , since the yukawaon @xmath163 has been replaced by @xmath154 in the present model , the first constraint in eq.(a.17 ) reads as @xmath164 , and since @xmath165 in the model , the second constraint in eq.(a.6 ) reads as @xmath166 .
since @xmath167 is given by eq.(a.7 ) , the requirement @xmath164 together with @xmath166 requires @xmath168 .
thus , the constraints ( a.17 ) and ( a.18 ) fix the parameters @xmath159 as @xmath169 the explicit values of these @xmath47 values are listed in table 1 . since the @xmath47 charges of @xmath170 , @xmath171 and @xmath172 are still free parameters , we take @xmath173 for simplicity . as we see in table 1 , the fields @xmath154 , @xmath60 , @xmath155 , @xmath156 and @xmath158 can safely have different @xmath47 charges from each other . thus , the assumption can lead to plausible @xmath47 charge values ( a.19 ) , so that we consider that the assumption is reasonable .
now we have an additional term , @xmath174 , \eqno(a.20)\ ] ] which should be included in @xmath2 given in eq.(2.6 ) .
we define @xmath177 which are invariant under permutation symmetries s@xmath178 and @xmath176 , respectively .
a rotation matrix @xmath47 which transforms the matrix @xmath175 into @xmath176 has been discussed in ref.@xcite .
the rotation matrix @xmath47 is given by @xmath179 where @xmath47 is defined as follows : @xmath180 @xmath181 @xmath182 @xmath183 the matrix @xmath53 is known as a matrix which diagonalizes the matrix @xmath175 into @xmath184 and also @xmath185 in the expression ( b.9 ) of @xmath47 , when we define @xmath189 the rotation matrix @xmath47 is expressed as follows : @xmath190 here , @xmath191
( @xmath192 ) satisfies @xmath193 and we can choose @xmath191 such as @xmath194 suggested from the charged lepton mass relation ( 4.1 ) , if we choose @xmath191 as @xmath195 where @xmath196 , then , the matrix @xmath47 satisfies @xmath197 since @xmath198 where @xmath199 the following relation holds : @xmath200 where @xmath201 is defined by eq.(b.8 ) .
however , note that , from eqs.(b.8 ) and ( b.18 ) , we can not conclude @xmath202 .
thus , it seems the rotation matrix @xmath47 from @xmath175 into @xmath176 is deeply related to the charged lepton mass relation ( 4.1 ) , but it is not clear why the form @xmath176 appears in the charge lepton sector .
this is still an open question at present .
99 y. koide , phys . rev . * d 79 * , 033009 ( 2009 ) . c. d. froggatt and h. b. nielsen , nucl
* b 147 * , 277 ( 1979 ) . y. koide , phys . lett . * b 680 * , 76 ( 2009 ) .
h. nishiura and y. koide , phys .
* d 83 * , 035010 ( 2011 ) .
n. cabibbo , phys .
* 10 * , 531 ( 1963 ) ; m. kobayashi and t. maskawa , prog .
. phys . * 49 * , 652 ( 1973 ) .
h. fritzsch , phys . lett . *
73b * , 317 ( 1978 ) ; * 85b * , 81 ( 1979 ) ; nucl . phys . *
b155 * , 189 ( 1979 ) .
b. pontecorvo , zh . eksp .
fiz . * 33 * , 549 ( 1957 ) and * 34 * , 247 ( 1957 ) ; z. maki , m. nakagawa and s. sakata , prog .
* 28 * , 870 ( 1962 ) .
xing , h. zhang and s. zhou , phys .
d 77 * , 113016 ( 2008 ) . and also see , h. fusaoka and y. koide , phys . rev . *
d 57 * , 3986 ( 1998 ) .
particle data group , k. nakamura , _
et al_. , j. phys . *
g 37 * , 075021 ( 2010 ) .
m. doi , t. kotani , h. nishiura , k. okuda , and e. takasugi , phys .
* b103 * , 219 ( 1981 ) and * b113 * , 513 ( 1982 ) .
p. f. harrison , d. h. perkins and w. g. scott , phys .
* b 458 * , 79 ( 1999 ) ; phys .
* b 530 * , 167 ( 2002 ) ; z .- z .
xing , phys . lett . * b 533 * , 85 ( 2002 ) ; p. f. harrison and w. g. scott , phys . lett . * b 535 * , 163 ( 2002 ) ; phys .
* b 557 * , 76 ( 2003 ) ; e. ma , phys .
lett . * 90 * , 221802 ( 2003 ) ; c. i. low and r. r. volkas , phys . rev . *
d 68 * , 033007 ( 2003 ) .
k. abe , _ et al . _
( t2k collaboration ) , phys .
lett . * 107 * , 041801 ( 2011 ) .
h. de kerret , a talk at low nu , seoul , nov .
2011 : http://www.dchooz.org/docdb /0033/003393/003/dcatlownu11_kerret111109_official.pdf .
y. koide , lett .
nuovo cimento * 34 * , 201 ( 1982 ) ; phys .
b120 * , 161 ( 1983 ) ; phys . rev .
* d28 * , 252 ( 1983 ) . y. koide , mod . phys . lett .
* a5 * , 2319 ( 1990 ) .
y. koide and h. fusaoka , phys.rev .
* d 66 * , 113004 ( 2002 ) . | on the basis of the so - called yukawaon " model , we found out a special form of the neutrino mass matrix @xmath0 which gives reasonable predictions .
the @xmath0 is given by a multiplication form made of charged lepton mass matrix @xmath1 and up - quark mass matrix @xmath2 .
this @xmath0 has no adjustable parameters except for those in @xmath1 and @xmath2 . here , @xmath1 and @xmath2
are described by one parameter @xmath3 ( real ) and two parameters @xmath4 ( complex ) , respectively , and those parameters are constrained by their observed mass ratios . with this form of @xmath0 , in spite of having only three parameters , the @xmath0 can give reasonable predictions @xmath5 , @xmath6 , @xmath7 , @xmath8 ev , and so on , by using observed values of @xmath9 , @xmath10 , @xmath11 , and @xmath12 as input values . * neutrino mass matrix with no adjustable parameters * @xmath13 _ department of physics , osaka university , toyonaka , osaka 560 - 0043 , japan _ + _ e - mail address : [email protected]_ @xmath14 _ faculty of information science and technology , osaka institute of technology , hirakata , osaka 573 - 0196 , japan _
+ _ e - mail address : [email protected]_ pacs numbers : 14.60.pq , 11.30.hv , 12.60.-i , |
infrared - luminous galaxies ( lirgs ) , those objects with infrared ( ir , [ 81000 m ] ) luminosities above @xmath6 @xmath7 ( and the ultraluminous infrared galaxies , those objects with @xmath8 @xmath7 ) are an important population for understanding the cosmological evolution of galaxies . at low redshifts lirgs
are relatively rare , with only a few hundred known over the whole sky at @xmath9 @xcite . at higher redshifts
they are much more common ; at @xmath10 they reach a density on the sky of several hundred per square degree , and make a substantial , perhaps even dominant contribution to both the cosmic infrared background and the volume - averaged cosmic star formation rate @xcite .
theoretical and empirical studies have shown that many lirgs and virtually all ulirgs at low redshift are interacting systems @xcite .
the power source behind the ir emission was initially controversial , with both star formation and agn activity proposed as the sole driver , but in recent years a consensus has started to emerge , with most studies finding that both starbursts and agn power low - redshift lirgs and ulirgs , with the star formation usually dominating @xcite . at high redshifts
a broadly similar picture is thought to exist @xcite .
the ability to reliably distinguish between star formation and agn activity in high redshift lirgs is essential for tracing both the cosmic star formation history and the evolution of agn activity as a function of redshift . since these galaxies are dusty , the usual optical / near - infrared diagnostics are not always reliable . instead
, mid - infrared ( mir ) spectroscopy has proved itself the key tool for determining the dominant energy source in dusty galaxies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , but until recently the limited sensitivity of the infrared space observatory ( _ iso _ ) constrained the spectroscopic study of ulirgs to the local universe . with the infrared spectrograph ( irs , *
* ) onboard _ spitzer _ , wavelength coverage and sensitivity have improved enough to allow redshift determination and spectroscopic classification for ulirgs at all redshifts up to @xmath11 @xcite . in most cases , spectroscopic surveys of high - redshift sources with the irs have selected sources with a high infrared - to - optical flux ratio @xcite .
these turn out to be mostly obscured agn at high redshift , with spectra dominated by silicate absorption .
starburst ulirgs with strong pah emission are rare in these samples , but prolific if the selection criteria instead include sub - mm detections @xcite or prominent stellar ` bumps ' in the irac bands @xcite .
in contrast , the irs has undertaken relatively few surveys of sources selected solely on the basis of mir flux .
this means that moderately to lightly absorbed sources will be under represented in irs surveys , thus drawing an incomplete picture of the nature of the infrared galaxy population at high redshifts . in this work we present results on a sample of 70 lirgs and ulirgs selected in the mir with no constraints on the optical fluxes ( other than the requirement of detection in at least 3 optical bands in deep imaging and photometric or spectroscopic @xmath2 @xmath1 1 .
our objectives are : ( a ) to classify the sources into starburst - dominated ( sb ) , obscured agn - dominated ( agn2 ) and unobscured agn - dominated ( agn1 ) , based on their mir spectral properties , and compare this classification with that obtained from optical data .
( b ) to estimate the ir luminosity of the sources and the relative contribution of agn and star formation .
( c ) to obtain star formation rates for the obscured and unobscured agn .
the paper is structured as follows : 2 describes the sample selection ; 3 discusses the observations and data reduction ; 4 presents the measurements and diagnostics ; 5 compares the results to other high-@xmath2 mir - selected samples ; and 6 summarizes our conclusions .
we adopt the following cosmology : @xmath12 = 0.27 , @xmath13 = 0.73 , @xmath14 = 71 km s@xmath15 mpc@xmath15 .
-band flux as a function of @xmath16 magnitude for the sources in the elais - irs sample .
asterisks indicate sources classified as qsos from their optical spectrum or sed , while diamonds represent sources classified as galaxies .
triangles represent sources with uncertain or unknown optical class due to lack of enough optical data .
[ s15_vs_sr],width=321 ] the sources were selected from the european large area iso survey ( elais ; * ? ? ?
* ; * ? ? ?
* ) final band - merged catalog @xcite .
the selection criteria required detection in the isocam lw3 band @xcite with @xmath0 @xmath1 0.8 mjy , as well as detection in at least 3 optical bands and spectroscopic or photometric redshift @xmath2 @xmath17 1 . the 70 objects that matched were observed with the irs @xcite , resulting in one of the largest @xmath2 @xmath17 1 samples of mir spectra of infrared - luminous galaxies to date .
table [ registro_irs ] indicates the name and coordinates of these sources .
only the first 15 entries of each table are shown here .
the complete version of the tables is available as on - line material .
in addition to the mir spectra , there is also optical and ir photometry available for all sources , as well as optical spectroscopy for roughly half of them . for sources in the northern hemisphere ( elais n1 and n2 fields ) optical @xmath18 , @xmath19 , @xmath16 , @xmath20 ,
@xmath21 photometry comes from the isaac newton telescope wide field survey ( wfs ) , with @xmath22 limiting magnitudes 23.4 , 24.9 , 24.0 , 23.2 and 21.9 ( vega ) respectively @xcite . in the southern hemisphere ( elais s1 )
optical photometry was obtained down to @xmath23 , @xmath24 @xmath3 25 , @xmath25 @xmath3 24.5 @xcite .
eight sources ( six of them in elais s1 ) have near - infrared ( nir ) photometry from the two micron all sky survey ( 2mass ; * ? ? ?
* ) , and ten more are detected by the ukirt infrared deep sky survey ( ukidss ; * ? ? ?
* ) in the central region of elais n1 .
table [ fotometria - opt ] summarizes the optical and nir photometry for the elais - irs sources .
the three elais fields in which the elais - irs sample was selected were observed by spitzer in seven ir bands from 3.6 to 160 m , as part of the spitzer wide - area infrared extragalactic survey ( swire ; * ? ? ?
* ; * ? ? ?
we performed a source match between the elais - irs catalog and the latest working catalog from swire and found that 69 out of 70 elais - irs sources are detected in at least two irac bands and in the 24-m band from mips , while 33 are detected in the 70-m band and only 14 at 160 m ( see table [ fotometria - ir ] ) .
one source , eirs-65 , has no infrared counterpart in the swire catalog and is undetected in the irs spectrum . since the 15-m detection is strong , the 15-m source is probably real but was associated to the wrong optical counterpart .
photometric redshifts were calculated using optical and nir photometry for all sources in the elais bandmerge catalog @xcite .
these estimates were revised with posteriority to the sample selection using irac photometry from swire @xcite , leading in some cases to revised photometric @xmath2 @xmath26 1 .
at least four and up to nine optical and nir bands are used for photometric redshift estimation , except in four sources whose redshifts are obtained from just three bands and are thus very unreliable ( eirs-19,40,52,55 ) .
redshifts from optical spectroscopy are also available for 37 elais - irs sources , most of them from spectroscopic follow - ups of elais qsos @xcite or the sloan digital sky survey ( sdss ; * ? ? ?
since an optical photometric or spectroscopic redshift is required for source selection , the sample is biased against the most obscured sources ; and the lack of constraints in the ir - to - optical flux ratio provides a large number of unobscured agn which are absent in other high-@xmath2 mir - selected samples ( see [ otrasmuestras ] ) .
39 sources are classified as qsos by their optical spectral energy distribution ( sed ) or spectrum , while 28 are galaxies ( including starbursts and obscured agn ) and the remaining 3 are of undetermined type .
qsos and galaxies can be easily distinguished by their optical - to - mir flux ratios ( fig .
[ s15_vs_sr ] ) with little overlap between populations .
in order to provide a good reference sample for redshift estimation and calibrate the mir diagnostic tools , we have compiled a set of mir spectra and mir - to - fir photometry containing 137 well - known low - redshift sources ( the ` library ' ) , including qsos , seyfert 1 and seyfert 2 galaxies , ulirgs and starbursts .
the mir spectra are high- and low - resolution irs spectra published in the literature ( see table [ info_biblioteca ] ) , while redshifts , classification and ir photometry for the sources were extracted from the nasa extragalactic database ( ned ) .
all targets were observed in the two low - resolution modules of the irs , short - low ( sl ) and long - low ( ll ) , exposing the two sub - slits on each of them ( sl1 , sl2 , ll1 and ll2 ) and thus covering the 538 m range .
two nod positions were observed in each configuration in order to provide optimal background subtraction .
exposure times for the ll were defined using isocam 15-m flux densities in order to assure an adequate s / n at this wavelength .
integration times for the sl were calculated from the 15-m flux density assuming an m82-like sed .
the observations log is shown in table [ registro_irs ] .
raw irs data were reduced by the irs pipeline ( ver .
s15.3.0 ) at the spitzer science center .
this included ramp fitting , dark subtraction , flat - fielding and wavelength and flux calibration .
since our sources are faint , additional reductions were carried out starting from the basic calibrated data ( bcd ) stage .
images with the object centred in alternate sub - slits of the same module were subtracted ( i.e : sl1-sl2 , ll1-ll2 ) to remove sky background , and their corresponding uncertainty and mask images were , respectively , added and bit - wise xor combined .
one - dimensional spectra for both nod positions were extracted using the spitzer irs custom extraction ( spice ) version 1.4.1 .
the extraction aperture was set to the spice default , and the optimal calibration mode , in which the signal in each pixel is weighted with its uncertainty , was selected .
individual spectra were flux - calibrated with spice using standard calibration files for pipeline s15.3.0 , and then combined into a single spectrum using custom idl routines .
comparison with irac 8.0-m and mips 24-m photometry from swire and isocam 15-m from elais indicates that the relative flux calibration uncertainty is around 14 , 35 and 8 per cent at 8 , 15 and 24 m , respectively .
note that the uncertainty at 15 m increases sharply around 1 mjy as it approximates the detection limit of the elais survey ( fig .
[ calibracion - residuos ] ) .
the 8-m fluxes from irs data are in average 10 per cent lower than in the irac photometry , but we have preferred not to make any flux correction to the irs spectra since the 8-m band covers a somewhat noisy region of the short - low spectrum in the overlapping wings of the sl1 and sl2 transmission profiles . , width=321 ] wavelength calibration is performed by the irs pipeline , and is accurate to @xmath31/5 of the resolution element or @xmath27/@xmath28 @xmath3 0.001 . reduced and calibrated spectra for each of the 70 elais - irs sources
are shown in fig .
[ irsobserved ] .
spectroscopic redshifts ( @xmath29 ) are available for roughly half of the sample , all of them optical qsos except for one obscured agn ( eirs-7 ) .
the rest of the sources were selected based on photometric redshifts ( @xmath30 ) published in the elais bandmerged catalog @xcite , which were later revised with the addition of ir photometry from swire and ukidss after the elais - irs sample was selected @xcite , leading to revised @xmath2 @xmath26 1 in some cases . in many of the sources in the sample
a redshift estimate can also be obtained from the irs spectrum based on the pah features , the silicate profile and the shape of the continuum sed ( @xmath31 ) . to this aim
, we have developed an algorithm that compares every spectrum with those in the library and determines the best-@xmath2 estimate by a @xmath32-minimization procedure ( hernn - caballero et al . in prep . ) .
reliable @xmath31 are obtained for 26 sources , while for another 28 a @xmath2 estimate is obtained , but it is uncertain due to poor sed fitting or multiple solutions .
most sources with reliable @xmath31 are galaxies , because their pah - emission and/or silicate - absorption features produce a single sharp spike in @xmath32(@xmath2 ) , while in qsos a smoother spectrum allows for multiple , weaker solutions . . [ zphot_zbest],width=321 ] a comparison of optical- and mir - spectroscopic versus photometric redshifts ( fig .
[ zphot_zbest ] ) indicates that @xmath30 estimates have significant uncertainties , larger than those of the general elais sample but similar to the uncertainty found in the @xmath10 subsample @xcite .
photometric redshifts are significantly less accurate in optical qsos than in galaxies , due to aliasing problems , and may be also affected by variability issues , because the optical photometry from wfs was not obtained in the same epoch for all bands @xcite .
since @xmath31 estimates are stronger in galaxies and obscured agn , while @xmath29 is more appropriate for qsos , the combination of both techniques allows us to obtain reliable redshifts for 57 of 70 sources ( table [ origen_z ] ) .
the remaining 13 rely on @xmath30 ( eirs-47 , 60 and 65 ) or unreliable @xmath31 estimates ( 10 sources ) , and will be discarded in all redshift - sensitive analysis . or @xmath31.[zhistog],width=321 ] the redshift distribution of the elais - irs sample
is shown in fig .
[ zhistog ] .
starburst galaxies concentrate around @xmath2 @xmath3 1 because the selection at 15 m favours sources with a strong 7.7-m pah feature at this redshift , while simultaneously prevents selection of higher - redshift sources since a typical starburst sed dwindles quickly for @xmath28 @xmath26 6 m.
this behaviour is also observed in 24-m selected starburst galaxies , which tend to concentrate around @xmath2 @xmath3 1.7 @xcite .
the redshift distribution for qsos is smoother , with most of them in the @xmath33 range , but the number of obscured agn decreases sharply at @xmath2 @xmath1 1.2 because of their red sed shortwards of 6 m restframe .
@xmath2 @xmath26 2 , in which the model fitted in the spectral decomposition is extrapolated up to 16 m restframe ( _ top _ ) ; and @xmath34 sources , for which a linear interpolation between the end of the irs spectrum and the mips70 photometric point is used ( _ bottom_).[extiende],title="fig:",width=226 ] @xmath2 @xmath26 2 , in which the model fitted in the spectral decomposition is extrapolated up to 16 m restframe ( _ top _ ) ; and @xmath34 sources , for which a linear interpolation between the end of the irs spectrum and the mips70 photometric point is used ( _ bottom_).[extiende],title="fig:",width=226 ] we calculate restframe monochromatic fluxes ( f@xmath35 ) and luminosities @xmath36l@xmath35(@xmath28 ) at several wavelengths ( 2.2 , 5.5 , 7 , 10 and 15 m ) .
the procedure used depends on the observed and restframe wavelength : for the 2.2 m ( restframe @xmath37 band ) we interpolate the photometry in the irac bands ( useful for sources at 0.6 @xmath26 @xmath2 @xmath26 2.65 ) ; for 5.5 , 7 and 10 m the spectrum is averaged over a narrow range ( 5.35.8 , 6.67.4 and 9.510.5 m respectively ) ; while for the 15-m luminosity the procedure is more complex : in @xmath38 sources it is obtained by averaging the 1416 m spectrum ; in @xmath39 sources 15 m restframe is outside the spectral range observed by irs , but not too far , so the 15-m luminosity is obtained by extrapolation of the model spectrum fitted in the spectral decomposition analysis ( see [ decomposition ] ) . in @xmath34 sources
the 15-m luminosity is interpolated between the red end of the irs spectrum and the mips70 photometric point ( see fig .
[ extiende ] for an example of the 15-m flux measurement in the last 2 cases ) .
table [ tabla_continuo ] summarizes the results obtained . in the literature we find two main methods for measuring the flux in the aromatic bands , which differ in the way the continuum and the pah features are measured :
a ) the interpolation method , in which the continuum underlying the aromatic bands is estimated by linear ( e.g. * ? ? ? * ) or spline ( e.g. * ? ? ?
* ; * ? ? ?
* ) interpolation between pah - free regions of the spectrum , and the flux over the estimated continuum is attributed to the pah feature ; and b ) the lorentzian method , in which a lorentzian or drude profile is assumed for the pah features @xcite .
we implemented the later method using a procedure in which a double power - law for the continuum and lorentzian profiles for the pah features are fitted simultaneously .
the fitting algorithm is similar to that described in @xcite and @xcite and is implemented as follows : the continuum emission is modeled by a linear combination of two power - laws with spectral indices @xmath40 , @xmath41 ranging from -5 to 5 .
the pah features are modeled by lorentzian profiles @xmath42(@xmath43,@xmath44,@xmath45 ) for the five main features ( 6.2 , 7.7 , 8.6 , 11.3 and 12.7 m ) , while the rest are ignored .
the amplitude of the lorentzian ( @xmath43 ) must be non - negative , its width ( @xmath44 ) is constrained between 0.5 and 2 times the typical value found in the literature = 0.2 , 0.7 , 0.4 , 0.2 and 0.3 m for the features centred at 6.2 , 7.7 , 8.6 , 11.3 and 12.7 m , respectively @xcite . ] and the central wavelength ( @xmath45 ) can fluctuate up to 5 per cent from its expected value .
extinction in the continuum is implemented assuming a screen model , with opacity @xmath46(@xmath28 ) following the galactic centre extinction law ( gc ; * ? ? ?
* ) , which has been widely used in other ulirg samples with moderate or high silicate absorption @xcite .
thus we fit the restframe spectra in the 515 m range to the analytic expression : @xmath47 we quantify the intensity of the silicate feature using the ` silicate strength ' , @xmath48 ( e.g. * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ) , defined as : @xmath49 where @xmath50 is the extinction - corrected continuum at 9.7 m , and @xmath51 is the observed flux ( averaged in a narrow interval to minimize the effects of the noise in the spectrum ) .
@xmath48 is positive for sources with silicates in emission and negative in absorption . in two high redshift sources ( eirs-30 and eirs-63 )
@xmath48 could not be measured because 9.7 m restframe is outside of the observed wavelength range , and in three starburst sources ( eirs-13 , 25 and 32 ) the mean flux in the 9.59.9 m range is negative in their noisy spectra .
for these five sources we assume @xmath48 = -@xmath52 , where @xmath52 = @xmath53(@xmath28=9.7 m ) is the estimated apparent optical depth at 9.7 m from eq .
[ eq_pahfit ] . in most sources with silicates in emission ,
the silicate feature is wider and centred at longer wavelengths ( @xmath310.5 m ) than it is found when the feature appears in absorption ( see * ? ? ?
* for tentative explanations ) , and the extra flux in the silicate feature tends to increase the continuum level in the fit .
these effects lead to a systematical under estimation of the silicate strength in sources with silicates in emission . to compensate for this
, we run a second fitting loop in sources with @xmath52 @xmath3 0 and @xmath48 @xmath1 0 in which the 9.511 m range is excluded from the fit to avoid overestimation of the continuum , and @xmath48 is measured at 10.5 m. table [ tabla_pahfits ] summarizes the results obtained from the fitting .
fluxes of pah features with s / n below @xmath54 are considered as non - detections and shown as upper limits .
, width=321 ] ir luminosity ( @xmath55 ) is estimated by integrating in the 81000 m range the template sed that best fits the mir and far - infrared ( fir ) photometry of each source .
template seds are selected from well - known theoretical and semi - empirical models in the literature , including four fir models ( torus , cirrus , m82 and arp220 ) from @xcite ; one early - type , seven late - types , three starbursts , six agn and three sb+agn composites from @xcite ; a set of 64 theoretical galaxy seds ( from quiescent to active ) from @xcite ; theoretical models of normal and starburst galaxies as a function of @xmath55 from @xcite ; two empirical mean seds ( radio - loud and radio - quiet ) from a sample of 47 qsos @xcite ; and three mean seds of optically - selected qsos with strong , weak or no detection in the fir @xcite . to reduce the number of template seds , the best - fitting template for each source in the library was obtained by fitting its mid - to - far ir photometry ( 12850 m ) .
template seds that provided the best - fitting solution for less than two sources were discarded , and from very similar ones ( e.g. dale25 and dale26 from * ? ? ?
* ) only one of them was selected .
this reduced the sample of templates to 14 models ( see fig .
[ sedtemplates ] ) : m82 and arp220 from @xcite ; gl12 from @xcite ; dale26 from @xcite ; ngc6240 , mrk231 , i19254s , seyfert2 , qso - norm , qso - high and qso - low from @xcite ; strong fir qso , weak fir qso and no - fir qso from @xcite .
the best - fitting template for the elais - irs sources is determined by fitting the irac 3.6 , 4.5 , 5.8 and 8.0 m , isocam 15 m , and mips 24 , 70 and 160 m photometry , as well as synthetic photometry in the 1038 m range obtained from the irs spectrum .
the synthetic filters have gaussian transmission profiles centred at @xmath56 = 10.4 , 12.4 , 14.9 , 17.9 , 21.5 , 25.8 and 30.9 m , with fwhm = @xmath56/6 .
the photometric bands shortwards of 3.6 m are discarded because at @xmath2 @xmath17 1 the flux in these bands is dominated by the stellar population in starbursts and obscured agn , or by the accretion disc in qsos , and is poorly correlated to the fir emission from dust .
fitting results are shown in fig .
[ fullsed ] .
the fits look reasonable for most sources , except for a significant under estimation of the mir emission in some obscured agn with a steep slope in the irac bands ( e.g. eirs-12 , eirs-47 , eirs-55 ) .
these sources fit an m82 template , probably because of the lack of a suitable obscured - agn template . if we assume that the fir sed of the elais - irs sources is not very different to that found in the sources in the library , the uncertainties in @xmath55 for elais - irs can be estimated using the following procedure : + 1 ) calculate , for the sources in the library , the best - fitting sed and its associated ir luminosity ( @xmath57 ) , using the ir photometry up to 15 , 25 or 60 m ( which roughly corresponds to 38 , 70 and 160 m for sources at @xmath2 @xmath3 1.5 , see fig . [ correspondencia_bandas ] ) .
+ 2 ) compare @xmath57to the ` true ' ir luminosity ( @xmath55 ) obtained by fitting all the fir and sub - mm photometry available .
@xmath3 1.5 the mips70 and mips160 bands roughly correspond to iras 25 and 60 m photometry in local ulirgs .
the restframe spectral range common for both samples in irs spectroscopy ( 515 m ) is highlighted.[correspondencia_bandas],width=321 ] the mean value of the logarithmic error in @xmath55 , is around 0.05 when @xmath57is estimated using all the available photometry up to 60 m ( equivalent to detection up to 160 m in a @xmath58 elais - irs source ) , 0.15 when using the photometry up to 25 m ( equivalent to 70 m detected elais - irs sources ) , and 0.23 if only photometry up to 15 m is used ( equivalent to elais - irs sources undetected at 70 and 160 m ) .
in addition to these errors , uncertainty in the mips70 and mips160 photometry ( up to 50 per cent in sources near the detection limit ) and the bias introduced by the use of a discrete set of templates must also be taken into account .
considering these extra error sources , a factor 23 of uncertainty seems reasonable for most sources , significantly lower than the factor 510 estimated for extrapolations using only the mips24 photometry @xcite . for the elais - irs sources as calculated from sed fitting .
the dot - filled histogram represents the whole sample , while those of the subsamples with mips70 and mips160 detection are drawn in diagonal and horizontal lines , respectively.[histograma_lumir],width=321 ] most elais - irs sources have estimated ir luminosities in the ulirg range , with some extreme cases in the hyper - luminous range ( @xmath55@xmath1 10@xmath59 @xmath7 ) and 1/4 of the sample in that of lirgs . fig .
[ histograma_lumir ] shows a histogram of the luminosity distribution of the sample .
it is noteworthy that the sources detected at 70 and 160 m concentrate in the higher luminosity tail of the distribution .
this is an expected outcome given the relatively shallow depth of the mips70 and mips160 observations , and as a consequence , the most accurate @xmath55 estimations are obtained in the most luminous sources .
table [ lumir_sfr ] indicates the best - fitting sed and estimated ir luminosity for the elais - irs sources . for sources dominated by star formation
, the most robust measurement of the star formation rate is obtained from the integrated 81000 m ir luminosity ( @xmath55 ) . for starburst galaxies , @xcite finds the relation : @xmath60\ = \
1.72 \times 10^{-10}\ \textit{l}_{\rm{ir}}\ [ \rm{l_\odot}]\\ \label{sfr_kennicutt}\ ] ] unfortunately , @xmath55 has big uncertainties for individual sources , and in sources with agn activity there is not a straightforward method to isolate the starburst and agn contributions to @xmath55 .
many recent papers find a strong correlation between the luminosity of the pah bands ( @xmath61 ) and @xmath55 in starburst - dominated sources ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the ratio @xmath61/@xmath55 depends on the properties of the interstellar medium , and thus it varies with the luminosity of the sources : in normal spiral galaxies , with quiescent star formation , the mean ratio is highest @xcite , in starburst galaxies it is somewhat lower , and in local ulirgs with massive star formation , it is lowest @xcite . nevertheless , the dependency of @xmath61/@xmath55 with @xmath55 is small relative to the dispersion between sources of the same luminosity : @xcite find for the 7.7-m pah a mean value of 8.1 @xmath62 10@xmath63 in starburst galaxies and 5.5 @xmath62 10@xmath63 in starburst - dominated ulirgs .
@xcite find that the 6.2-m pah luminosity relates to @xmath55 trough a power - law of index 1.13 in a sample spanning 2 orders of magnitude in @xmath55 .
some authors use the 6.2-m pah because its measurement suffers little contamination from other mir features @xcite while others prefer the 7.7-m pah because it is the strongest pah band @xcite , or even use the combined flux of two or more features ( e.g. * ? ? ?
* ; * ? ? ?
the estimated luminosity in the pah features depends strongly on the measurement procedure : @xcite find discrepancies of a factor 1.7 at 6.2 m and 3.5 at 7.7 m depending on whether the continuum is calculated by spectral decomposition into lorentzian+continuum or it is interpolated between adjacent regions .
this suggests that the discrepancies in the value of @xmath61/@xmath55 found by different authors ( see table [ tabla_lpah_lir ] ) are mainly due to systematics .
in addition , the integrated infrared luminosity is not the same throughout the literature : some authors define @xmath55 in the range 11000 m ( e.g. * ? ? ?
* ) , while others use only the fir ( 42122 m , @xcite ; 40500 m , @xcite ) .
because of these issues , we obtain our own estimate of the average @xmath61/@xmath55 in the starburst - dominated elais - irs sources before using the pah luminosity to estimate sfr in the whole sample .
the mean values of @xmath61/@xmath55 found in the 7 starburst - dominated sources with no sign of significant agn emission ( eirs-21 , 25 , 37 and 41 excluded ) are : @xmath64/@xmath55 = 0.012 @xmath65 0.007 , @xmath66/@xmath55= 0.038 @xmath65 0.016 and @xmath67/@xmath55= 0.011 @xmath65 0.007 .
the ratio found for the 7.7-m pah is similar to that obtained by @xcite in a sample of starburst galaxies and @xcite in normal and starburst galaxies .
it agrees with that found by @xcite in starburst galaxies and ulirgs if we take into account the factor 3.5 decrement calculated by @xcite for fluxes measured by interpolation of the continuum using splines . for the 6.2-m feature ,
our ratio is in agreement with those of @xcite and @xcite but it is roughly double the ratio found by @xcite in a sample of normal and starburst galaxies even after correction for the differences in the continuum estimation procedure . , width=321 ] by applying the calibration in eq .
[ sfr_kennicutt ] , and assuming that the properties of the interstellar medium in which the pah emission originates are not significantly altered by the agn activity , the sfr in the whole sample can be calculated as : @xmath68 & = & 1.40 \times 10^{-8}\ l_{\rm{62}}\ [ \rm{l_\odot}]\label{sfr_eq1}\\ sfr\ [ \rm{m_{\odot } yr^{-1 } } ] & = & 4.56 \times 10^{-9}\ l_{\rm{77}}\ [ \rm{l_\odot}]\label{sfr_eq2}\\ sfr\ [ \rm{m_{\odot } yr^{-1 } } ] & = & 1.52 \times 10^{-9}\ l_{\rm{113}}\ [ \rm{l_\odot}]\label{sfr_eq3}\end{aligned}\ ] ] for sources in which two or three pah features are detected with s / n @xmath1 2 we estimate sfr from their combined luminosity , while if only one or no pah features are detected , we average estimates from individual upper limits . table [ lumir_sfr ] shows the estimated sfr for elais - irs sources .
they range from @xmath310 to @xmath31000 @xmath5 .
the stated uncertainties account for the uncertainty in the pah flux only ; if we consider the uncertainty in the ir luminosity and the dispersion in @xmath61/@xmath55 of the starburst - dominated sources , a more realistic estimate would be a factor 2 uncertainty in sfr .
figure [ histograma_sfr ] shows de distribution of sfr in the elais - irs sample .
most estimates are in the range 150600 @xmath5 , but if we consider upper limits for sources with undetected pahs , which account for more than half of the agn population , many of them have probably sfr @xmath26 100 @xmath5 .
a powerful tool for diagnostics is decomposition into several spectral components , which can provide considerable insight to the physics of the sources ( e.g. * ? ? ?
mir emission from active galaxies arises mostly from hii regions , photodissociation regions ( pdrs ) and agn @xcite , so we have used a simple model comprising the superposition of three spectral templates ( agn , pdr and hii ) obscured by a screen of dust .
the spectral decomposition is performed by fitting the 515 m restframe range of the spectrum to a parametrized @xmath69 of the form : @xmath70 where the free parameters ( @xmath43 , @xmath71 , @xmath72 , @xmath73 ) are calculated using a levenberg
marquardt @xmath32-minimization algorithm , and must have non - negative values . to keep the number of free parameters low , a single optical depth value is applied to the three spectral components , but for most sources the solution is almost identical when two optical depth values ( one for the agn and other for the pdr and hii components ) are used , with an increased @xmath32 value due to the extra free parameter .
@xmath46(@xmath28 ) is obtained from the gc extinction law , while @xmath74 is represented by the irs spectrum of the seyfert 1 galaxy ngc 3515 @xcite , @xmath75 by the isocam spectrum of a pdr in the reflection nebula ngc 7023 @xcite and @xmath76 by the isocam spectrum of m17 in the vicinity of ob stars @xcite .
we quantify the contribution of each spectral component to the mir spectrum as the ratio of its integrated luminosity to total luminosity in the 515 m restframe range , @xmath77 , @xmath78 and @xmath79 for the agn , hii and pdr components .
figure [ descomposicion_espectral ] shows the distribution of spectral components for the library and the elais - irs sample . ,
width=321 ] a number of mir diagnostic diagrams have been proposed to determine the nature of infrared galaxies ( e.g. * ? ? ?
? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
they rely mainly on the strength of the pah features to infer the dominant energy source ( agn or starburst ) , and on the slope of the continuum and sign of the silicate strength ( emission or absorption ) to distinguish obscured and unobscured sources .
+ there is a strong correlation between the optical and mir spectrum of agn : type 2 agn usually show a red mir spectrum , and the silicate feature in absorption , while most type 1 agn show a bluer mir spectrum , with stronger nir emission , and the silicate feature absent or in emission .
nevertheless , the match between populations is not perfect , and many mismatches ( most notably , type 1 agn with silicates in absorption and type 2 with silicates in emission ) have been reported ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
a silicate emission feature in type 2 agn can be reproduced by a low - si , low - tau smooth torus model @xcite , or by silicate emission being produced in an extended region , such as a dusty narrow line region @xcite .
silicate absorption in type 1 agn can be observed if the broad line region is partially obscured , maybe by dust in the galaxy .
a very remarkable case is mrk 231 @xcite , which shows a broad emission - line spectrum reddened with @xmath80 @xmath3 2 in the optical @xcite , and strong silicate absorption in the mir @xcite . since making assumptions on the seyfert type of agn based on their ir properties alone would be misleading , we will refrain from doing that and will instead classify the agn as obscured or unobscured in the ir .
using some well known diagnostics along with some others from our own we separate the mir spectra of the elais - irs sources into three main categories : starburst - dominated ( sb ) , unobscured agn - dominated ( agn1 ) and obscured agn - dominated ( agn2 ) .
every diagnostic criterion is calibrated using the known nature of sources in the low - redshift library , including seyfert type for agn , and a final mir classification is obtained as the most repeated result among all criteria .
measurements on the elais - irs and library spectra were obtained using exactly the same procedures , to prevent distortions in the diagnostic diagrams due to systematics .
limits between populations in the diagnostic diagrams are determined based solely on the library spectra , and are fine - tuned to minimize the number of misclassifications within the library .
figure [ diagnosticos ] shows an adaptation of three well known diagnostic diagrams by @xcite , @xcite and @xcite , as well as a new one that we propose . in the diagnostic diagram of @xcite the 6.2-m pah equivalent width ( @xmath81 )
easily separates starburst - dominated sources from the agn - dominated ones , while the ratio between 15 and 5.5 m restframe spectral luminosities ( @xmath82/@xmath83 ) distinguishes sources dominated by hot dust from an unobscured agn ( low @xmath82/@xmath83 ) or cold dust by an obscured agn or starburst ( high @xmath82/@xmath83 ) .
the value of @xmath81 that best separates starburst - dominated and agn - dominated sources from the library is @xmath3 0.2 , while in @xmath82/@xmath83 we find too much overlap between populations and prefer to drop it as a diagnostics criterion . in the diagram from @xcite the ratio between @xmath83 and @xmath55
is related to the relative amount of hot and cold dust in the galaxy , being higher for unobscured agn and lower for starbursts and deeply embedded sources .
the 6.2-m pah luminosity ( @xmath64 ) to @xmath55 ratio depends on the relative contribution of unobscured star formation to the ir luminosity , and is higher for unobscured starbursts or agn+sb composite sources , while lower for the dust enshrouded ones .
most elais - irs sources have too high upper limits on this ratio to be meaningful , so we also drop this parameter as diagnostic criterion .
the diagram from @xcite is based on @xmath81 and the strength of the silicate feature .
galaxies populate their diagram along two branches : one horizontal branch from agn - dominated to starburst - dominated spectra , and one diagonal branch from highly obscured nuclei to starburst - dominated spectra .
they hypothesize that the separation into two branches reflects two different configurations of the obscuring dust : clumpy in the horizontal and non - clumpy ( homogeneous shell ) in the diagonal . in our adaptation of the spoon diagram ( fig .
[ diagnosticos ] , bottom left ) it is noteworthy the lack of elais - irs sources in the high - extinction , weak - pah region of the diagram , where many local ulirgs are located .
the silicate strength is used to separate obscured and unobscured agn , but most sources have a small absolute value of @xmath48 , and its uncertainty can lead to a wrong classification in some cases .
we propose a new diagnostic diagram , inspired by the one from spoon but optimized for low s / n sources ( fig .
[ diagnosticos ] , bottom right ) . in our modified spoon diagram
, @xmath81 is replaced by the line ( peak ) to continuum ratio of the 7.7-m pah feature ( @xmath84/@xmath85 ) .
this parameter has been extensively used to distinguish starbursts from agn ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and works better in low s / n sources because the 7.7-m feature is stronger than the one at 6.2 m. the silicate strength is replaced by a more straight - forward and model - independent measurement : the ratio between @xmath82 and the spectral luminosity averaged in a narrow band centred at 10 m restframe ( @xmath86 ) .
this ratio separates type 1 and 2 agn from the library more efficiently than @xmath48 does , because it combines the two main distinctive factors : absorption / emission in the silicate band and slope of the continuum .
in addition , it is less sensitive than @xmath48 to noise in the spectrum .
its main drawback is that 15 m restframe is unobservable by irs for sources at @xmath2 @xmath1 1.5 , but it can be interpolated using the mips70 photometry as we showed in [ medidas_continuo ] .
type 1 and 2 agn from the library are better separated in this diagram than in the original from spoon in both axes , and only the type 1 agn with a strong starburst contribution overlap significantly with the type 2 ones .
this overlap is produced by a higher slope of the continuum and depleted silicate feature in type 1 agn with a starburst component . finally , the spectral decomposition diagram in fig .
[ descomposicion_espectral ] can also be used for diagnostics , since starburst - dominated sources locate near the left border , because of strong pdr and/or hii component , while agn - dominated sources concentrate in the lower - right corner .
from the five diagnostic diagrams described , we extract six criteria to classify infrared galaxies ( table [ tabla_criterios ] ) .
criteria c1 to c4 distinguish agn from starbursts based on the strength of the pah features and mir - to - fir luminosity ratios , while c5 and c6 rely on the silicate feature to separate obscured and unobscured agn . the final class assigned to each source
is the most repeated one among the six criteria .
this combination of multiple criteria allows for a more robust classification than that obtained using a single diagnostic diagram , since it is less likely to be altered by statistical uncertainties .
this is most significant in the lower s / n spectra , and also in obscured agn , which some criteria often classify as unobscured agn or starbursts because typical values for obscured agn lay in a narrow range between those of starbursts and unobscured agn . for sources in the library ,
the ir classification is consistent with the optical class for more than 90 per cent of them ( see table [ resumen_diagnostico_biblioteca ] ) .
an analysis of the results from individual criteria shows that mismatches are caused by two main factors : a ) strong star formation in type 1 agn , which leads to classification as obscured agn via high values of @xmath84/@xmath85 and @xmath82/@xmath86 .
b ) silicate emission in type 2 agn , which confuses the two criteria based on the silicate feature . in the elais - irs sample
we expect a somewhat higher mismatch rate because the lower s / n of the spectra increases the likelyhood of incorrect classification by some of the individual criteria . in all diagnostic diagrams the elais - irs sources are plotted with symbols representing their final ir class , so that mismatches between individual criteria and the final classification can be easily recognized .
based on the criteria presented above , 41 sources are classified as unobscured agn ( agn1 ) , and 17 are obscured agn ( agn2 ) .
the remaining 11 sources are classified as starbursts ( sb ) , including a few sources with strong continuum ( eirs-21 , eirs-25 , eirs-37 ) , which could host an obscured agn , and an agn1+sb composite ( eirs-41 ) .
table [ tabla_diagnostico ] shows the classification from individual criteria as well as the final ir class for the whole elais - irs sample .
a comparison of the optical versus ir classification for elais - irs sources ( table [ resumen_diagnostico_elais - irs ] ) shows 12 mismatches : seven are optical galaxies with unreliable redshifts classified in the ir as unobscured agn because of weak or absent pah features or presence of silicate emission .
some of them ( eirs-8 , eirs-38 ) present a rather flat slope in the irac bands , indicating low extinction , while others ( eirs-47 , eirs-50 , eirs-56 , eirs-60 ) show a steep slope and fir emission , suggesting a high level of extinction .
the other galaxy , eirs-27 , shows a very strong silicate emission feature , similar to that found in the seyfert 2 galaxy iras f01475 - 0740 @xcite .
the remaining five sources are optical qsos with reliable spectroscopic redshifts , which are classified as obscured agn in the ir due to intense pah emission and a small but negative value of @xmath48 . in these sources ,
the dust in star - forming regions may have depleted the silicate emission feature in the agn spectrum enough to make it appear in absorption . in three of these sources ( eirs-5 ,
eirs-39 and eirs-54 ) the criterion in @xmath82/@xmath86 predicts agn1 while @xmath48 @xmath26 0 .
a fourth source , eirs-63 , is at @xmath2 = 3.094 and the silicate feature moves out of the range observed by irs , so the ir diagnostic is inconclusive .
finally , eirs-58 has a deep silicate feature and is classified as agn2 by both @xmath82/@xmath86 and @xmath48 , but its optical sed is blue and shows a power - law spectrum in the optical and uv up to the lyman break .
the best fitting ir sed and mir classification are consistent with each other for most sources .
all sources classified as unobscured agn fit best to one of the seven qso seds , except eirs-47 , which depends on a photometric redshift and is obscured in the optical .
obscured agn also select mainly qso templates ( most significantly mrk 231 , a reddened type-1 qso with strong silicate absorption ) , but 1/3 of them fit best m82 or i19254s ( an obscured agn + starburst composite ) .
starburst sources favour pure - starburst ( gl12 and dale26 ) or composite ( seyfert 2 , i19254s and ngc 6240 ) templates , except for two sources with strong mir continuum ( eirs-41 and eirs-37 ) which fit best qso templates .
we have used the spectral classification in [ diagnostico ] to calculate average spectra of each population , which have better s / n than the spectra of individual sources . in the average spectrum
we only include sources with reliable redshift estimations , to prevent as much as possible the dilution of fine details by the dispersion in @xmath2 . to calculate the averaged spectrum , individual spectra
are shifted to the rest - frame , normalized at 7 m and interpolated in a common array of wavelengths uniformly distributed in log(@xmath28 ) . for each value of @xmath28
we discard the highest and lowest value of @xmath87 and the remaining ones are combined using a weighted average .
dispersion of individual spectra .
the dashed line in the bottom panel represents the average starburst spectrum after removal of sources with a strong continuum component ( eirs-21 , 25 , 37 and 41).[espectros_promedio],width=321 ] figure [ espectros_promedio ] shows the averaged spectra for starbursts , optical qsos and obscured agn ( optical galaxies classified as agn in the ir ) in the sample .
the quasar spectrum can be approximated by a power - law of index @xmath40 = -1 in the range 416 m ( fig .
[ espectro_qsos ] ) , which is slightly steeper than the index found in the averaged spectrum of a sample of local qsos @xcite .
in addition , the silicate feature appears in emission in the averaged spectrum , but with a very low profile .
these findings suggest the elais - irs quasars could be more obscured than average , nearby optically selected qsos .
see [ geometry ] for a discussion on the connection between the silicate feature and the mir spectral index .
@xmath3 2 from @xcite ( lutz+08 ) .
a power law @xmath88 @xmath89 @xmath90 with index @xmath91 is shown for reference ( dashed line).[espectro_qsos],width=321 ] the slope of the spectrum increases shortwards of 3 m towards the 1 m minimum of the sed , as has also been observed in other optically selected quasar samples ( e.g. * ? ? ?
this is interpreted within the torus model as thermal emission from graphite grains at temperatures up to @xmath31500 k. in the averaged spectrum there are tentative detections of some ionic lines of neon , [ ne ii ] 12.81 m and [ ne iii ] 15.56 m , but since they have low / medium excitation potentials , their emission can also be excited by star formation .
the pah features are weak but detected , with @xmath81 @xmath3 0.015 and @xmath92 @xmath3 0.05 . in [ sfr_in_agn ]
we estimate mean pah luminosities and star formation rates for these sources .
@xmath89 @xmath90 with index @xmath93 is shown for reference ( dashed lines ) .
[ agn2_highz],width=321 ] the averaged spectrum of the obscured agn shows a steep slope at @xmath28 @xmath26 8 m ( @xmath40 = -2 ) but it flattens longwards of the silicate trough at @xmath310 m ( fig . [ agn2_highz ] ) .
the depth of the silicate feature is relatively small ( @xmath52 @xmath3 0.7 ) , as has been observed in many seyfert 2 and qso2s ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
however , the requirement of optical spectroscopic or photometric redshift for selection may have reduced the average silicate absorption by biasing the sample against the most obscured sources .
emission lines for ionic species are very weak or absent , with the significant exception of [ ne ii ] 12.81 m. the 7.7-m pah feature seems rather strong , but it could be magnified by a peak in the underlying continuum emission at 8 m caused by the onset of the silicate absorption feature .
the other pah features are weak , with only the one at 11.3 m clearly detected .
the 6.2 m feature is missing even if it is detected in many individual spectra .
this may be caused by the cancellation effect produced by strong 6.0 m absorption bands by water ice or hydrocarbons in some sources @xcite . in the averaged spectrum of the starburst - dominated galaxies we find , in addition to the usual pah bands at 6.2 , 7.7 , 8.6 , 11.3 and 12.7 m , a few additional , fainter bands , centred at 5.2 , 5.8 , 14.2 and 17.0 m , which are also common in higher s / n spectra of low redshift galaxies .
the 3.3-m band , relatively strong in iso sws spectra of many local ulirgs , seems absent ; but the irs spectra of the elais - irs starbursts are very noisy in this range .
some ionic lines frequently observed in local starbursts also appear in the average spectrum , most significantly [ ar ii ] 6.99 m , [ ar iii ] 8.99 m , [ ne ii ] 12.81 m ( which overlaps with the 12.7-m pah ) and [ ne iii ] 15.56 m. it is noteworthy the absence of rotational transitions of the h@xmath94 molecule , with the possible exception of s(3 ) 9.67 m. the weak continuum shortwards of 6 m indicates that the relative contribution of agn or obscured star formation is small .
if the four starburst - dominated sources with suspected or known agn contribution are discarded , the average spectrum of the remaining starburst sources ( eirs-2 , 3 , 9 , 13 , 14 , 16 and 32 ) has a lower continuum emission at @xmath28 @xmath26 6 m , but remains unchanged at longer wavelengths while the pah features get significantly boosted .
before spitzer , the spectroscopic study of ulirgs was restricted to a few local sources due to the low sensitivity of previous instruments ; but in the last four years , thanks to the unprecedented sensitivity of irs , it has been possible to obtain mir spectra for several hundred ulirgs up to @xmath2 @xmath3 3 .
most of these sources were selected from wide - area mir surveys , and are usually heavily obscured , because only sources with high mir - to - optical flux ratios were selected for observation with irs ( e.g. * ? ? ?
* ; * ? ? ?
* ) . in some cases
, additional constraints were imposed to favour the selection of starburst or agn - dominated sources @xcite . by contrast
, the selection criteria for the elais - irs sample put no constraints on the mir - to - optical flux ratio of the sources , but required optical spectroscopic or photometric redshifts .
this introduced a bias that is not easy to quantify , but certainly favored sources with little or no obscuration at moderate redshifts .
in addition , selection at 15 m instead of 24 m also favours lower - redshift sources , because the pah features and the steep slope shortwards of 6 m in obscured agn enter the selection band at lower @xmath2 . in this sense ,
the elais - irs sample can be seen as complementary to other higher-@xmath2 spectroscopic samples .
the selection criteria leave their fingerprint in a colour - colour diagram like the one in fig .
[ s24_s8_sr ] .
indeed , the ratio between the flux density at 24 m and an optical band , used as selection criterion for all samples except elais - irs , determines to a large extent the nature of the sources selected .
the most extreme cases ( @xmath95/@xmath96 @xmath1 2000 ) dominate the samples from houck and polletta .
these sources are strongly obscured in the optical and show steep ir spectra , with strong silicate absorption , and tend to be at high redshift ( @xmath2 @xmath1 1.7 ) .
ratios of 300 @xmath26 @xmath95/@xmath96 @xmath26 2000 correspond to the obscured agn in elais - irs , the starburst galaxies from farrah and the composite sources from yan .
the distribution of @xmath2 goes down to @xmath30.5 with a gap around @xmath2 @xmath3 1.4 because of the silicate feature entering the mips24 band .
ulirg samples observed with the irs : @xcite ( x signs ) , @xcite ( diamonds ) , @xcite ( plus signs ) , @xcite ( triangles ) and elais - irs ( asterisks ) .
the dashed vertical line indicates the lower limit in @xmath95/@xmath96 for the sample of @xcite .
the continuous and dotted lines represent , respectively , tracks for m82 and arp220 redshifted from @xmath2 = 0 to @xmath2 = 3.[s24_s8_sr],width=321 ] at @xmath95/@xmath96 @xmath3 100
we find the elais - irs starburst at 0.6 @xmath26 @xmath2 @xmath26 1 , which are bluer in @xmath95/@xmath96 and @xmath95/@xmath97 than starburst sources from farrah or yan ; and at @xmath95/@xmath96 @xmath26 100 we find most of the elais - irs quasars , which are slightly or not obscured , spanning a wide range in redshift up to @xmath2 @xmath3 3 .
note that excluding the elais - irs sources , the rest are significantly redder in @xmath95/@xmath96 than would be m82 or arp220 at the same redshift , indicating that these sources are more obscured than many local ulirgs , and probably represent only the most extreme cases of the ulirg population at high redshift .
ulirg samples .
symbols are as in fig .
[ s24_s8_sr ] .
the eight elais - irs sources with lower 5.8-m luminosity that depart from the main trend are starbursts , while most of the other sources are agn or agn - dominated composite sources.[l - z_plane],width=321 ] since most of the sources in these samples have @xmath95 @xmath3 1 mjy , their distribution in the @xmath42@xmath2 plane covers a very narrow strip ( fig .
[ l - z_plane ] ) , indicating that , except for a few sources with ultra - deep irs spectra down to @xmath30.1 mjy @xcite , only the most luminous sources at each redshift have been probed so far
. a deeper insight into the dominant population of lirgs at @xmath2 @xmath1 1 will be possible once the next generation of mir spectrographs onboard _
jwst _ and _ spica _ become available .
we have classified 10 of the 27 optically - faint elais - irs sources as starburst - dominated galaxies , and there is also an optical quasar ( eirs-41 ) dominated by starburst features in the mir .
the selection criteria required @xmath30 @xmath1 1 for galaxies , but @xmath31 estimates are in the 0.61.2 range for all starburst galaxies , with @xmath98 = 0.8 .
this is caused by uncertainties in @xmath30 and by the selection at 15 m , which favours sources with strong pahs at @xmath2 @xmath3 1 , while at higher redshift requires a strong continuum at @xmath28 @xmath26 6 m , favoring agn . in spite of this bias towards pah - strong sources at @xmath2 @xmath3 1 , only 1/3 of the @xmath99 elais - irs sources are starbursts , while 2/3 are agn .
most significantly , if we exclude the qsos only half of the galaxies in this range are starbursts , the rest being obscured agn .
this indicates that the population of 15-m sources at @xmath2 @xmath3 1 is comprised mostly of agn for @xmath0 @xmath1 1 mjy .
the starbursts could still dominate the highly obscured galaxy population , since they are excluded from the elais - irs sample because they are too faint in the optical ; but this is probably not the case , because the agn also dominate the 24-m counts of sources brighter than 1 mjy at @xmath2 @xmath3 1.7 @xcite .
samples : elais - irs starbursts ( solid line ) , @xmath100 starbursts from @xcite ( dotted line ) , pah - strong sources from @xcite ( dashed line ) and @xmath100 starbursts from @xcite ( dot - dashed line).[highz_sb],width=321 ] the averaged spectrum of the elais - irs starbursts is similar to that found in higher redshift samples selected at 24 m ( fig .
[ highz_sb ] ) .
differences between them in the depth of the silicate feature and pah equivalent width can be interpreted as a combination of a ` pure starburst ' ( hii+pdr ) spectrum and an agn spectrum in variable proportions . in this case
, starburst galaxies from farrah and elais - irs would be the less contaminated by agn dust continuum , while in those of yan the continuum would dominate de mir spectrum .
starburst - dominated samples ( filled stars ) .
the open star represents the averaged spectrum of starburst - dominated elais - irs sources after removing the four sources with strong mir continuum .
positions for all sources in the library are also shown as a reference.[spoon_promedio],width=321 ] the spoon diagram for averaged spectra ( fig .
[ spoon_promedio ] ) indicates that the composite spectrum of the elais - irs starbursts is in the range of local starburst ulirgs .
if the starburst - dominated sources with strong continuum are removed from the composite , the resulting @xmath81 increases threefold and resembles that of lower luminosity local starbursts . in the diagram of pah luminosity ratios ( fig .
[ pah_ratios ] ) the composite spectrum of elais - irs starburst - dominated sources also situates among local starburst galaxies , with @xmath64/@xmath67 substantially higher than in typical starburst - dominated ulirgs .
this ratio is related to the ionization state of the pah molecules @xcite , a high ratio indicating abundance of cationic species , while a low ratio indicates most pah carriers are neutral .
@xmath64/@xmath66 shows anticorrelation with the extinction @xcite , and is usually lower in local ulirgs than in less luminous starburst galaxies @xcite . in the elais - irs average , it is somewhat higher than in most local ulirgs , suggesting a lower extinction level .
lower ionization an extinction values than in local ulirgs have also been found in higher-@xmath2 starburst - dominated galaxies @xcite , and could indicate that the star formation is extended over larger regions relative to their local counterparts , or , alternatively , that their gas - to - dust ratio is higher .
samples ( big solid stars ) .
the open star represents the averaged spectrum of starburst - dominated elais - irs sources after removing the four sources with strong mir continuum , while individual starburst - dominated sources from elais - irs are shown as asterisks .
sources from the library with a starburst component are also shown as a reference : starburst galaxies ( circles ) , starburst ulirgs ( squares ) , type 1 agn with starburst component ( triangles ) and type 2 agn with starburst ( inverted triangles).[pah_ratios],width=321 ] it has been stated that the sfr of seyfert galaxies correlates with the luminosity of the agn ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and that the sfr is systematically higher in type 2 seyferts relative to type 1 @xcite . for higher luminosity agn
there are contradictory results : most local qsos are hosted by elliptical galaxies , with very little star formation @xcite , and only 30 per cent of them show signs of interaction with other galaxies @xcite .
measurements in the [ o ii ] 3727-@xmath101 line indicate that the sfr is very low ( a few @xmath5 ) in optically selected qsos @xcite . in the ir
, @xcite find in irs spectra of palomar - green qsos that more than 30 per cent of the ir luminosity is due to star formation , but a later work from @xcite reduces the starburst contribution to @xmath325 per cent of the flux at 70 and 160 m which translates to @xmath310 per cent of @xmath55 if we assume templates like dale26 for the starburst and qso - norm for the agn component for a larger sample of optically selected qsos .
using mir and fir photometry from swire , @xcite estimate relative agn and sb contributions to @xmath55 in a sample of 70 sdss qsos with 70-m detection , and find the agn contribution varies from @xmath3100 to @xmath320 per cent of @xmath55 . optically selected qsos , however , represent less than half of the total qso population @xcite , the remaining being obscured quasars ( qso2s ) that show up in the infrared @xcite radio @xcite or x - rays @xcite .
conversely to what is found in local qsos , hst images of a sample of @xmath102 qso2s show signs of perturbations and intense star formation in the host galaxy @xcite , and in the mir there are signs of recent or ongoing star formation @xcite . at higher redshift there are samples of sub - mm selected qsos with very high sfr ( 10003000 @xmath5 ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , but most of the high-@xmath2 qsos are undetected in the sub - mm range ( @xmath370 per cent ; * ? ? ?
* ) , and the observed trend indicates that the starburst - to - agn luminosity ratio decreases with increasing agn luminosity @xcite , thus suggesting that the observed correlation betwen star formation and agn luminosity saturates at higher luminosities .
the same pattern seems to apply to qso2s : @xcite find that irs spectra of x - ray selected obscured agn show very weak or no pah bands , and @xcite estimate that the star formation contributes @xmath2620 per cent to the bolometric luminosity of a sample of obscured agn with @xmath103 @xmath1 10@xmath104 @xmath7 .
the elais - irs agn comprise a very heterogeneous population , since it covers a wide range of pah equivalent width and sfr . in some of the sources , the estimated starburst - induced ir luminosity ( @xmath105 ) calculated from the pah luminosity is higher than the total ir luminosity ( @xmath55 ) obtained from the sed fitting .
since it is unlikely that @xmath61/@xmath105 be lower for these sources ( indeed , pah destruction by the agn would lead to lower ratios ) the most likely interpretation is that the sed of these agn has a cold dust component which is not accounted for in the sed fitting effectively underestimating @xmath55 because of lack of deep enough photometry in the fir .
incidentally , mips70 and mips160 upper limits for many elais - irs agn can not rule out a significant cold dust component in the fir ( see fits in fig . [ fullsed ] ) .
the pah features are comparatively very weak in most elais - irs agn , and in many of them ( 35 sources , mostly unobscured agn ) only a loose upper limit of the ( pah - derived ) @xmath105 is available . to get a tighter estimate of its typical value , we measure the strength of the pah features in the averaged spectra of qso and agn2 sources .
assuming that the equivalent width of a pah feature in the averaged spectrum ( @xmath106 ) is similar to those of the individual sources , we can roughly estimate the typical pah luminosity of the sources by multiplying @xmath106 times the luminosity of the continuum underlying the pah feature .
since individual spectra are normalized at 7 m before averaging them , the mean pah luminosity may be overestimated because the lower luminosity sources tend to show higher pah equivalent widths .
to minimize this bias , we split the qso and agn2 populations in two subsamples of high and low luminosity sources , based on their restframe 5.5-m luminosity .
we place the limit between populations at @xmath83 = 4 @xmath62 10@xmath107 @xmath7 to include the same number of sources in both subsamples . in the lower luminosity population the 6.2
, 7.7 and 11.3 m pah features are detected in both qso and agn2 averages , while in the higher luminosity ones they all but disappear ( fig . [ espectros_agn_lum ] ) .
@xmath1 4 @xmath62 10@xmath107 @xmath7 .
( b ) optical qsos with @xmath108 @xmath26 4 @xmath62 10@xmath107 @xmath7 . ( c ) optical galaxies with @xmath108 @xmath1 4 @xmath62 10@xmath107 @xmath7 .
( d ) optical galaxies with @xmath108 @xmath26 4 @xmath62 10@xmath107 @xmath7 .
the dashed lines mark the position of the 6.2 , 7.7 and 11.3 m pah features .
[ espectros_agn_lum],width=321 ] table [ sfrtablepromedio ] shows the estimated average pah luminosity and associated @xmath105 and sfr for the averaged spectra of the elais - irs agn . in the lower luminosity
qsos the average pah luminosity corresponds to an ir luminosity in the lirg range , and accounts for 2025 per cent of @xmath55 as estimated from sed fitting , similar to the mean value found by @xcite in a sample of pg , 2mass and 3cr qsos at lower redshift .
the obscured agn in the same luminosity range have pah luminosities almost 3 times higher at 6.2 and 7.7 m ( which suggests star formation is responsible for about half of @xmath55 in these sources ) , but they are similar to that of qsos at 11.3 m. estimated sfr is @xmath350 @xmath5 for the lower - luminosity qsos and @xmath3100 @xmath5 for the obscured agn . in the higher - luminosity agn
the pah features are too weak to meassure them even in the composite spectrum .
their equivalent width is at least 3 times lower in the high luminosity qso composite than in the lower luminosity one , and the resulting upper limits in @xmath61 constraint the mean sfr to @xmath26 100 @xmath5 .
the implied ir luminosity from star formation accounts for @xmath2610 per cent of the estimated @xmath55 .
the high - luminosity obscured - agn composite contains only five sources and has lower s / n than the others .
upper limits for the 6.2 and 7.7 m features suggest sfr @xmath26 300 @xmath5 , and a starburst contribution to @xmath55 of 1015 per cent at most . to better constrain the amount of fir emission in the qsos ,
we have performed stacking analysis of the 70 and 160 m photometry .
fluxes for individual sources are estimated from the counts in the pixel containing the coordinates of the source , and their uncertainties are obtained from a fit to the histogram of ( finite - valued ) pixel values in the image .
a mean sky level is computed for the image and subtracted from the source flux in order to ensure that the mean value for random positions in the image is zero .
the mean flux @xmath109 and dispersion @xmath110 for the population are estimated from the individual measurements ( @xmath111 , @xmath112 ) as the maximum - likelihood solution for the distribution : @xmath113 a detailed description of the whole stacking procedure can be found in @xcite .
the mean fluxes obtained at 70 and 160 m are shown in table [ stacking ] .
if we assume the fir sed is dominated by a starburst component with an m82-like sed , the average ir luminosity needed to reproduce the observed mean fluxes is @xmath310@xmath104 @xmath7 for the @xmath114 sources , and @xmath35@xmath6210@xmath104 @xmath7 for those at @xmath2 @xmath1 1.5 .
similar results are obtained if we assume an arp220 sed instead of m82 for the starburst component .
the @xmath105 extrapolated from the observed pah luminosity is a factor 3 lower in the @xmath114 qsos ( and a factor @xmath115 lower in the @xmath116 qsos ) than that obtained from the stacking analysis of 70 and 160 m photometry , suggesting that the average pah - to - fir luminosity ratio is significantly lower than in starburst galaxies .
a likely interpretation is reduced pah emission in star - forming regions in the vicinity of an agn due to destruction of the pah carriers by x - ray and far - uv photons from the agn @xcite .
incidentally , the factor @xmath173 reduction in @xmath106 and in the pah - to - fir luminosity ratio between low- and high - luminosity qsos suggests higher depletion rates of the pah carriers in the most luminous agn .
alternatively , it could indicate a saturation of the correlation between the luminosities of the starburst and agn , as found in other samples of @xmath117 qsos @xcite and higher - luminosity , high-@xmath2 qsos @xcite , but results from the stacking analysis suggest the fir luminosity roughly scales with agn luminosity . estimated average luminosity of the 6.2 and 7.7 m pah features in the low - luminosity composites are about 23 times higher in the obscured agn than in the qsos .
this could indicate higher sfr in the obscured agn , or , alternatively , higher abundance of pah carriers due to absorption by the obscuring material of the energetic photons from the agn .
however , the statistical significance of this result is not strong due to the small number of sources contributing to the obscured agn average , and requires validation in larger samples .
a selection bias favouring obscured sources with strong pah emission could also be responsible for the increased pah luminosity , but we have not found any likely bias able to increase pah luminosity in obscured agn but not in qsos . in example , the pah features increase the observed flux in the selection band at @xmath2 @xmath3 1 , where we find roughly half of the obscured agn , but these sources are bright enough to be selected even without contribution from pah emission to the 15-m flux . since the pah features help to find a redshift estimate in many sources , this could play a role in the observed asymmetry , because most redshifts for agn2 sources rely on the irs spectrum , while most qsos have optical spectroscopic redshifts .
but in agn2 sources a reliable redshift estimate can usually be obtained from the ( absorption ) silicate feature alone , and indeed all sources with unreliable redshifts are classified as agn1 .
in addition , a lower accuracy in redshifts obtained from the silicate feature relative to those from optical spectroscopy imply that pah features may be blurred in the agn2 composite , leading to an underestimation of their real strength .
another possible interpretation could be lower mir luminosity in obscured agn than in qsos of the same intrinsic luminosity , as is expected if the putative torus is optically thick in the mir ; but this is not supported by measurements of the luminosity ratio between x - rays and mir @xcite .
a strong difference in sfr between optical type 1 and type 2 agn has been observed in a sample of qsos and qso2s from the sdss @xcite .
using estimates from the [ o ii ] 3727 line , they find a sfr an order of magnitude higher in type 2 sources , and this is probably a lower limit , due to higher extinction in the optical .
such a large difference between type 1 and 2 sources can hardly be explained within the agn unification scheme , and might indicate instead that the agn classification into type 1 or type 2 does not depend exclusively on the orientation of an agn torus .
recent papers suggest that a fraction of the obscuration in type 2 agn occurs in the host galaxy : @xcite find a correlation between silicate extinction and orientation of the plane of the galaxy , while @xcite suggest that foreground extinction by cold dust is required to reproduce both the silicate feature and nir continuum for almost half of the sources in a sample of mir selected obscured agn .
if we assume that the extinction law for dust in the torus is similar to that found in local ulirgs and in the galactic centre , it reaches a minimum in opacity at @xmath37 m and increases towards shorter wavelengths ; thus the optical depth is higher in the nir than at 7 m. if the silicate absorption feature originates in the torus , a decrease in the viewing angle of the torus should produce a deeper silicate feature and weaker and steeper nir continuum peaking at longer wavelengths ( e.g. * ? ? ?
* ; * ? ? ?
by contrast , if the silicate feature originates in the host galaxy , a deep silicate profile could be observed in conjunction with an intense and flat nir spectrum @xcite .
@xmath26 -0.5 ( solid line ) , -0.5 @xmath26 @xmath48 @xmath26 0 ( dotted line ) , 0 @xmath26 @xmath48 @xmath26 0.15 ( dashed line ) , and @xmath48 @xmath1 0.15 ( dot - dashed line).[compara_ssil],width=321 ] , width=321 ] fig .
[ agn2_highz ] compares the average spectrum of elais - irs agn2 sources with the average of sources that fit the torus model ( t ) or torus+cold absorber ( t+c ) from @xcite .
the nir slope and strength of the silicate feature are better reproduced by the t model .
figure [ compara_ssil ] shows the averaged spectrum of the elais - irs agn separated in four groups attending to their measure of @xmath48 . in sources with @xmath48
@xmath1 0 ( mostly qsos ) the nir slope is small and independent of @xmath48 , but in sources with @xmath48 @xmath26 0 the nir slope is substantially higher , particularly for the subsample with strongest silicate absorption .
this confirms the existence of a correlation between nir slope and silicate absorption , as expected if the obscuration originates mainly in the agn torus . to search for sources that depart from this trend , we calculated the spectral index @xmath40 of the continuum between 2.2 and 7 m restframe for individual spectra ( fig .
[ alpha_vs_ssil ] ) . most qsos yield @xmath40 @xmath26 1.5 , while most agn2 show @xmath40 @xmath1 1.5 .
there is a clear trend , even if the dispersion is high , towards a simultaneous increase of @xmath40 and @xmath48 in agn , but not in starbursts .
note that eirs-21 , which is suspected of harboring an obscured agn due to its strong mir continuum , departs strongly from the agn population .
this suggests that the observed extinction is better reproduced by a screen of cold dust instead of a thick agn torus .
incidentally , it is undetected down to @xmath118 erg cm@xmath15 s@xmath15 in [ 2 - 10 ] kev band images from xmm ( i. valtchanov priv . comm . )
, so the presence of a dust enshrouded agn is questionable .
the only agn2 source that separates slightly from the general trend is eirs-52 , but its spectrum is still better described by the torus model ; therefore , a dusty torus is favored as the source of the obscuration in the elais - irs agn .
the lack of t+c sources in elais - irs is probably caused by the selection criteria , which prevent selection of highly obscured agn as those from @xcite .
we have analyzed the ir properties of a sample of 70 mir selected lirgs and ulirgs at 0.4 @xmath26 @xmath2 @xmath26 3.2 .
the mir selection favours starburst - dominated sources in redshift intervals in which the strongest pah bands overlap with the passband of the selection filter , namely , @xmath2 @xmath3 1 at 15 m and @xmath2 @xmath3 1.8 at 24 m. for the agn - dominated sources , the redshift distribution is wider because their ir sed has smoother features .
the relatively shallow depth of the optical data in the elais - irs sample and the requirement of optical photometric or spectroscopic redshifts for source candidates favours the selection of weakly obscured sources , contrarily to other mir high-@xmath2 ulirg samples where highly obscured sources are preferred , and thus turns the elais - irs sample into a valuable complement for the study of the high-@xmath2 ulirg population .
redshift estimates for the elais - irs sources from their mir spectrum are reliable for obscured or star - forming sources , and unreliable in many unobscured agn .
this behaviour is the opposite to that obtained through optical spectroscopy , and thus guarantees that reliable redshift estimates can be obtained by one or other means for most ir sources .
a combination of mir diagnostic criteria efficiently separates starbursts , obscured and unobscured agn , with a good correlation to the optical classification .
the selection criteria of the elais - irs sample favour the selection of starburst - dominated sources at 0.6 @xmath26 @xmath2 @xmath26 1.2 , yet most galaxies selected in this range are agn - dominated .
this suggests that the population of @xmath120 @xmath1 1 mjy sources at @xmath2 @xmath17 1 is dominated by agn .
the averaged spectrum of elais - irs starbursts is similar to that of local ulirgs , but the observed depth of the silicate feature , strength of pah features and pah flux ratios are more alike to those found in lower - luminosity starburst galaxies .
this behaviour has already been observed in higher - redshift starburst samples and could be related to star formation extending over larger regions or a higher gas - to - dust ratio in the galaxy .
the correlation between mir and fir seds allows the estimation of ir luminosities from the mir spectrum with a factor 3 uncertainty . if mips70 and/or mips160 photometry is available the uncertainty reduces further and becomes dominated by uncertainties in the photometry and in the sed models .
the mean sfr in elais - irs qsos estimated from pah luminosity is 50 - 100 @xmath5 , but its implied fir emission is 310 times lower than that obtained from stacking analysis in the 70 and 160 m bands , thus suggesting the pah emission in substantially depleted in most qsos .
a lower pah - to - continuum luminosity ratio is found in the most luminous qsos relative to lower - luminosity ones , suggesting a saturation of the correlation between star formation and agn activity found in other samples of lower - luminosity sources or , alternatively , higher depletion rates of the pah carriers in the galaxies hosting the most powerful agn . in the obscured agn ,
the mean sfr is 23 times higher than in qsos of the same luminosity .
if confirmed in larger samples , this could indicate a connection between dust obscuration and star formation in the host galaxy .
however , the observed correlation between silicate absorption and the slope of the nir - to - mir continuum is consistent with the hypothesis of the obscuring dust being located in a dusty torus .
this work is based on observations made with the _ spitzer space telescope _ , which is operated by the jet propulsion laboratory , caltech under nasa contract 1407 .
999 afonso - luis a. et al . , 2004 ,
mnras , 354 , 961 alexander d. m. et al . , 2001 , apj , 554 , 18 armus l. et al . , 2007 , apj , 656 , 148 basilakos s. et al . , 2002 , mnras , 331 , 417 beelen a. , cox p. , benford d. j. , dowell c. d. , kovacs a. , bertoldi f. , omont a. , carilli c. l. , 2006 , apj , 642 , 694 berta s. et al . , 2006 ,
a&a , 451 , 881 berta s. et al . , 2007 , a&a , 467 ,
565 bertoldi f. , carilli c. l. , cox p. , fan x. , strauss m. a. , beelen a. , omont a. , zylka r. , 2003 , a&a , 406 , 55 boksenberg a. , carswell r. f. , allen d. , fosbury r. , penston m. , sargent w. , 1977 , mnras , 178 , 451 . borne k. d. et al .
, 1999 , apss , 266 , 137 brand k. et al . , 2007 , apj , 663 , 204 brandl , b. r. et al . , 2006 ,
apj , 653 , 1129 bridge , c. r. et al . , 2007 , apj , 659 , 931 buchanan c. , gallimore j. , odea c. , baum s. , axon d. , robinson a. , elitzur m. , elvis m. , 2006 , aj , 132 , 401 cesarsky d. , lequeux j. , abergel a.,perault m. , palazzi e. , madden s. , tran d. , 1996a , a&a , 315 , 305 cesarsky d. et al . , 1996b , a&a , 315 , l309 chapman s. c. , windhorst r. , odewahn s. , yan h. , conselice c. , 2003b , apj , 599 , 92 chiar j. e. , tielens a. g. g. m. , 2006 , apj , 637 , 774 clements d. l. , sutherland w. j. , mcmahon r. g. , saunders w. , 1996 , mnras , 279 , 477 crampton d. , cowley a. p. , hartwick f. d. a. , ko p. w. , 1992 , aj , 104 , 1706 cutri r. m. , nelson b. o. , kirkpatrick j. d. , huchra j. d. , smith p. j. 2001 , in asp conf
232 , the new era of wide field astronomy , ed .
r. g. clowes , a. j. adamson and g. e. bromage ( san francisco : asp ) , 78 dale d. , helou , g. , 2002 , apj , 576 , 159 dale d. et al .
, 2005 , apj , 633 , 857 deo r. p. , crenshaw d. m. , kraemer s. b. , dietrich m. , elitzur m. , teplitz h. , turner t. j. , 2007 , apj , 617 , 124 draine b. t. , li a. , 2001 , apj , 551 , 807 dunlop j. s. , mclure r. j. , kukula m. j. , baum s. a. , odea c. p. , hughes d. h. , 2003 , mnras , 340 , 1095 elvis m. et al . , 1994 , apjs , 95 , 1 efstathiou a. , rowan - robinson m. , 1995 , mnras , 273 , 649 farrah d. et al . , 2001 , mnras , 326 , 1333 farrah d. , verma a. , oliver s. , rowan - robinson m. , mcmahon r. , 2002 , mnras , 329 , 605 farrah d. et al . , 2003 , mnras , 343 , 585 farrah d. et al
, 2007 , apj , 667 , 149 farrah d. et al . , 2008 , apj , 677 , 957 frster schreiber n. m. , sauvage m. , charmandaris v. , laurent o. , gallais p. , mirabel i. f. , vigroux l. , 2003 , a&a , 399 , 833 fritz j. , franceschini a. , hatziminaoglou e. 2006 , mnras , 366 , 767 galliano f. , madden s. c. , tielens a. g. g. m. , peeters e. , jones a. p. , 2008 ,
apj , 679 , 310 gallimore j. f. , keel , w. c. , 1993 , aj , 106 , 1337 genzel r. et al
. , 1998 , apj , 498 , 579 gonzlez - solares e. a. et al . , 2005 , mnras , 358 , 333 guyon o. , sanders d. b. , stockton a. n. , 2006 , apjs , 166 , 89 haas m. et al . , 2003 , a&a , 402
, 87 hao l. , weedman d. w. , spoon h. w. w. , marshall j. a. , levenson n. a. , elitzur m. , houck j. r. , 2007 , apj , 655 , l77 hatziminaoglou e. et al . , 2005 ,
aj , 129 , 1198 hatziminaoglou e. et al . , 2008 , mnras , 386 ,
1252 hernn - caballero a. et al .
, 2008 , aspc , 381 , 367 ho l. c. , 2005 , apj , 629 , 680 horst h. , gandhi p. , smette a. , duschl w. j. , 2008 , a&a , 479 , 389 houck j. r. et al . , 2004 , apjs , 154 , 18 houck j. r. et al .
, 2005 , apj , 622 , 105l imanishi m. , dudley c. c. , maiolino r. , maloney p. r. , nakagawa t. , risaliti g. , 2007 , apjs , 171 , 72 kennicutt r. c. , 1998 , ara&a , 36 , 189 kim m. , ho l. c. , i m m. , 2006 , apj , 642 , 702 kimura h. et al . , 2002 , icarus , 159 , 529 lacy m. et al . , 2004 , apjs , 154 , 166 lacy m. et al . , 2005 , mem . s. a. it .
, 76 , 154 lacy m. et al .
, 2007 , apj , 669 , 61 la franca f. et al . , 2004 ,
aj , 127 , 3075 lagache g. , dole h. , puget j. l. , 2003 , mnras , 338 , 555 lagache g. , et al . , 2004 , apjs , 154 , 112 lagache g. , puget j. l. , dole h. , 2005 , ara&a , 43 , 727 laurent o. , mirabel i. f. , charmandaris v. , gallais p. , madden s. c. , sauvage m. , vigroux l. , cesarsky c. , 2000 , a&a , 359 , 887 lawrence a. et al .
, 2007 , mnras , 379 , 15599 le floch e. , mirabel i. f. , laurent o. , charmandaris v. , vigroux l. , cesarsky c. , 2001 , a&a , 367 , 487 li a. , draine b. t. , 2001 , apj , 554 , 778 lipari s. , colina l. , macheto f. , 1994 , apj , 427 , 174 .
lonsdale c. j. et al . , 2003a , pasp , 115 , 897 lonsdale c. j. , lonsdale c. j. , smith h. , diamond p. , 2003b
, apj , 592 , 804 lonsdale c. j. et al . , 2004 ,
apjs , 154 , 54 lonsdale c. j. , farrah d. , smith h. e. , 2006 , astrophysics update 2 - topical and timely reviews on astronomy and astrophysics .
john w. mason .
springer / praxis books ( astro - ph/0603031 ) lu n. et al . , 2003 , apj , 588 , 199 lutz d. spoon h. w. w. , rigopoulou d. , moorwood a. f. m. , genzel r. , 1998 , apj , 505 , 103l lutz d. , sturm e. , genzel r. , spoon h. w. w. , morwood a. f. m. , netzer h. , sternberg a. , 2003 , a&a , 409 , 867 lutz d. , maiolino r. , spoon h. w. w. , moorwood a. f. m. , 2004 , a&a , 418 , 465 lutz d. , valiante e. , sturm e. , genzel r. , tacconi l. j. , lehnert m. d. , sternberg a. , baker a. j. , 2005 , apj , 625 , 83 lutz d. et al .
, 2007 , apj , 661 , 25l lutz d. et al . , 2008 , apj , 684 , 853l maiolino r. , ruiz m. , rieke g. h. , keller l. d. , 1995 , apj , 446 , 561 maiolino r. , shemmer o. , imanishi m. , netzer h. , oliva e. , lutz d. , sturm e. , 2007 , a&a , 468 , 979 martnez - sansigre a. , rawlings s. , lacy m. , fadda d. , marleau f. r. , simpson c. , willott c. j. , jarvis m. j. , 2005 , nature , 436 , 666 morel t. et al .
, 2001 , mnras , 327 , 1187 netzer h. et al . , 2007 ,
apj , 666 , 806 norman c. et al .
, 2002 , apj , 571 , 2188 oliver s. et al . ,
2000 , mnras , 316 , 768 omont a. , beelen a. , bertoldi f. , cox p. , carilli c. l. , priddey r. s. , mcmahon r. g. , isaak k. g. , 2003 , a&a , 398 , 857 peeters e. , spoon h. w. w. , tielens a. g. g. m. , 2004 , apj , 613 , 986 polletta m. et al . , 2007 , apj , 663 , 81 polletta m. , weedman d. , hnig s. , lonsdale c. j. , smith h. e. , houck j. , 2008 , apj , 675 , 960 rapacioli m. , joblin c. , boissel p. , 2005 ,
a&a , 429 , 193 rigopoulou d. , spoon h. w. w. , genzel r. , lutz d. , moorwood a. f. m. , tran q. d. , 1999 , aj , 118 , 2625 rowan - robinson m. , 2001 , apj , 549 , 745 rowan - robinson m. , 2003 , mnras , 345 , 819 rowan - robinson m. et al . , 2004 , mnras , 351 , 1290 rowan - robinson m. et al .
, 2005 , aj , 129 , 1183 rowan - robinson m. et al
. , 2008 , mnras , 386 , 697 sajina a. , yan l. , armus l. , choi p. , fadda d. , helou g. , spoon h. w. w. , 2007 , apj , 664 , 713 sanders d. b. , mirabel i. f. , 1996 , ara&a , 34 , 725 schweitzer m. et al . , 2006 , apj , 649 , 79 serjeant s. , hatziminaoglou e. , 2009 , mnras , accepted ( astro - ph/0901.0552 ) shi y. et al . , 2006 , apj , 653 , 127 shi y. et al . , 2007 , apj , 669 , 841 siebenmorgen r. , krgel e. , spoon h. w. w. , 2004 , a&a , 414 , 123 siebenmorgen r. , haas m. , krgel e. , schulz b. , 2005 , a&a , 436 , 5l skrutskie m. f. et al
, 2006 , aj , 131 , 1163 smith j. d. t. et al . , 2007 , apj , 656 , 770 soifer , b. t. et al .
, 1984 , apjl , 278 , l71 spoon h. , keane j. , tielens a. , lutz d. , moorwood a. , laurent o. , 2002 , a&a , 385 , 1022 spoon h. w. w. et al .
, 2004 , apjs , 154 , 184 spoon h. w. w. , marshall j. a. , houck j. r. , elitzur m. , hao l. , armus l. , brandl b. r. , charmandaris v. , 2007 , apj , 654 , 49 stern d. et al .
, 2005 , apj , 631 , 163 sturm e. et al . , 2005 ,
apj , 629 , 21 sturm e. , hasinger g. , lehmann i. , mainieri v. , genzel r. , lehnert m. d. , lutz d. , tacconi l. j. , 2006 , apj , 642 , 81 surace j. a. et al . , 2005 , aas , 205 , 6303 swinbank a. m. , smail i. , chapman s. c. , blain a. w. , ivison r. , keel w. , 2004 , apjs , 617 , 64 takata t. , sekiguchi k. , smail i. , chapman s. c. , geach j. e. , swinbank a. m. , blain a. , ivison r. j. , 2006 , apj , 651 , 713 teplitz h. i. et al . , 2007 , apj , 659 , 941 tran q. d. et al .
, 2000 , apj , 552 , 527 vaccari m. et al . , 2005 , mnras , 358 , 397 vermeij r. , van der hulst j. m. , 2002 , a&a , 391 , 1081 voit g. m. , 1992 , mnras , 258 , 841 weedman d. w. et al . , 2005 , apj , 633 , 706 weedman d. w. et al . , 2006 , apj , 653 , 101 white r. l. , helfand d. j. , becker r. h. , gregg m. d. , postman m. , lauer t. r. , oegerle w. , 2003 , aj , 126 , 706 yan l. et al . , 2005 , apj , 628 , 604 yan l. et al . , 2007 , apj , 658 , 778 york d. g. et al . , 2000 , aj , 120 , 1579 llcccrcrrrr eirs-1 & elaisc15_j161401.0 + 544733 & 16h14m30.9s & + 54d52m15s & 0.79 @xmath65 0.10 & 147.1 & 2005 - 01 - 08 & 967 & 967 & 1950 & 1950 + eirs-2 & elaisc15_j161350.0 + 542631 & 16h14m19.3s & + 54d31m11s & 0.80 @xmath65 0.16 & 146.9 & 2005 - 01 - 08 & 1451 & 967 & 1950 & 1950 + eirs-3 & elaisc15_j163636.2 + 411049 & 16h36m42.9s & + 41d17m04s & 0.83 @xmath65 0.14 & 116.2 & 2005 - 02 - 14 & 967 & 967 & 1950 & 1950 + eirs-4 & elaisc15_j003640 - 433925 & 00h36m11.8s & -43d35m46s & 0.98 @xmath65 0.17 & 50.0 & 2005 - 06 - 08 & 967 & 967 & 1950 & 1950 + eirs-5 & elaisc15_j003408 - 431011 & 00h34m41.4s & -43d12m31s & 0.99 @xmath65 0.13 & 216.2 & 2004 - 11 - 16 & 967 & 967 & 1950 & 1950 + eirs-6 & elaisc15_j163721.3 + 411503 & 16h37m29.2s & + 41d21m13s & 1.00 @xmath65 0.14 & 118.4 & 2005 - 02 - 12 & 484 & 484 & 975 & 975 + eirs-7 & elaisc15_j163655.8 + 405909 & 16h37m03.2s & + 41d05m23s & 1.02 @xmath65 0.14 & 117.4 & 2005 - 02 - 13 & 484 & 484 & 975 & 975 + eirs-8 & elaisc15_j163455.0 + 412211 & 16h35m04.3s & + 41d28m16s & 1.04 @xmath65 0.15 & 120.9 & 2005 - 02 - 08 & 484 & 484 & 975 & 975 + eirs-9 & elaisc15_j160733.7 + 534749 & 16h08m01.6s & + 53d52m41s & 1.04 @xmath65 0.18 & 144.5 & 2005 - 01 - 09 & 484 & 484 & 975 & 975 + eirs-10 & elaisc15_j161511.2 + 550627 & 16h15m41.3s & + 55d11m06s & 1.06 @xmath65 0.12 & 147.5 & 2005 - 01 - 08 & 484 & 484 & 975 & 975 + eirs-11 & elaisc15_j163739.3 + 414348 & 16h37m47.5s & + 41d49m59s & 1.06 @xmath65 0.11 & 118.4 & 2005 - 02 - 12 & 484 & 484 & 975 & 975 + eirs-12 & elaisc15_j161229.0 + 542832 & 16h12m57.7s & + 54d33m19s & 1.06 @xmath65 0.17 & 145.8 & 2005 - 01 - 09 & 484 & 484 & 975 & 975 + eirs-13 & elaisc15_j163954.5 + 411109 & 16h40m02.3s & + 41d17m20s & 1.07 @xmath65 0.16 & 118.1 & 2005 - 02 - 13 & 484 & 484 & 975 & 975 + eirs-14 & elaisc15_j163536.6 + 404754 & 16h35m44.4s & + 40d54m05s & 1.12 @xmath65 0.12 & 118.1 & 2005 - 02 - 12 & 484 & 484 & 975 & 975 + eirs-15 & elaisc15_j161332.4 + 544830 & 16h14m02.5s & + 54d53m13s & 1.12 @xmath65 0.19 & 146.9 & 2005 - 01 - 08 & 484 & 484 & 975 & 975 + rcccccccc eirs-1 & 19.39 @xmath65 0.02 & 20.28 @xmath65 0.03 & 19.64 @xmath65 0.04 & 19.41 @xmath65 0.11 & 18.95 @xmath65 0.02 & 18.17 @xmath65 0.02 & & 16.42 @xmath65 0.01 + eirs-2 & 20.51 @xmath65 0.04 & 23.08 @xmath65 0.05 & 22.14 @xmath65 0.06 & 21.25 @xmath65 0.12 & 20.40 @xmath65 0.04 & 19.10 @xmath65 0.07 & & 17.11 @xmath65 0.03 + eirs-3 & & 23.22 @xmath65 0.10 & 22.27 @xmath65 0.05 & 21.01 @xmath65 0.10 & 20.40 @xmath65 0.05 & & & + eirs-4 & & & & & & & & + eirs-5 & & & & & & & & + eirs-6 & 20.51 @xmath65 0.03 & 20.77 @xmath65 0.02 & 20.59 @xmath65 0.04 & 20.14 @xmath65 0.06 & 19.97 @xmath65 0.04 & & & + eirs-7 & 22.63 @xmath65 0.10 & 22.87 @xmath65 0.04 & 22.92 @xmath65 0.08 & 23.02 @xmath65 0.23 & & & & + eirs-8 & 21.94 @xmath65 0.05 & 22.39 @xmath65 0.03 & 21.80 @xmath65 0.06 & 20.90 @xmath65 0.10 & 20.20 @xmath65 0.05 & & & + eirs-9 & & & 22.71 @xmath65 0.07 & 21.88 @xmath65 0.06 & 21.11 @xmath65 0.12 & & & + eirs-10 & 19.07 @xmath65 0.02 & 19.58 @xmath65 0.02 & 19.07 @xmath65 0.02 & 19.00 @xmath65 0.02 & 18.72 @xmath65 0.02 & & & + eirs-11 & 18.22 @xmath65 0.02 & 19.06 @xmath65 0.02 & 18.79 @xmath65 0.02 & 18.51 @xmath65 0.02 & 18.55 @xmath65 0.03 & & & + eirs-12 & 20.63 @xmath65 0.05 & 21.47 @xmath65 0.02 & 21.14 @xmath65 0.02 & 20.79 @xmath65 0.11 & 20.25 @xmath65 0.04 & 20.23 @xmath65 0.13 & & 19.46 @xmath65 0.15 + eirs-13 & & 23.57 @xmath65 0.07 & 22.76 @xmath65 0.06 & 22.10 @xmath65 0.11 & 20.49 @xmath65 0.06 & & & + eirs-14 & & & 23.46 @xmath65 0.14 & 21.78 @xmath65 0.12 & 21.18 @xmath65 0.11 & & & + eirs-15 & 22.49 @xmath65 0.10 & 22.72 @xmath65 0.04 & 21.74 @xmath65 0.05 & 20.49 @xmath65 0.11 & 19.68 @xmath65 0.02 & 18.08 @xmath65 0.02 & & 16.41 @xmath65 0.01 + rcccccccc eirs-1 & 297.0 @xmath65 1.9 & 366.0 @xmath65 2.4 & 471 @xmath65 6 & 789 @xmath65 5 & 0.79 @xmath65 0.10 & 2.48 @xmath65 0.02 & & + eirs-2 & 128.0 @xmath65 1.6 & 113.0 @xmath65 1.4 & 64 @xmath65 5 & 84 @xmath65 4 & 0.80 @xmath65 0.16 & 0.94 @xmath65 0.02 & 16.4 @xmath65 1.0 & 70.2 @xmath65 5.1 + eirs-3 & 63.4 @xmath65 1.1 & 40.7 @xmath65 1.2 & & & 0.83 @xmath65 0.14 & & & + eirs-4 & 347.0 @xmath65 2.0 & 524.0 @xmath65 2.6 & 745 @xmath65 7 & 1090 @xmath65 7 & 0.98 @xmath65 0.17 & 4.33 @xmath65 0.04 & 33.1 @xmath65 1.7 & + eirs-5 & 276.0 @xmath65 2.2 & 359.0 @xmath65 2.7 & 458 @xmath65 8 & 707 @xmath65 8 & 0.99 @xmath65 0.13 & 2.18 @xmath65 0.03 & & + eirs-6 & 31.2 @xmath65 0.9 & 41.6 @xmath65 1.2 & 56 @xmath65 4 & 140 @xmath65 4 & 1.00 @xmath65 0.14 & 0.48 @xmath65 0.02 & & + eirs-7 & 27.6 @xmath65 0.5 & 52.6 @xmath65 1.0 & 121 @xmath65 3 & 399 @xmath65 4 & 1.02 @xmath65 0.14 & 2.39 @xmath65 0.02 & 13.0 @xmath65 1.2 & + eirs-8 & 205.0 @xmath65 1.8 & 247.0 @xmath65 1.6 & 345 @xmath65 6 & 516 @xmath65 4 & 1.04 @xmath65 0.15 & 1.40 @xmath65 0.02 & & + eirs-9 & 98.6 @xmath65 1.3 & 77.7 @xmath65 1.5 & 96 @xmath65 5 & 163 @xmath65 5 & 1.04 @xmath65 0.18 & 1.91 @xmath65 0.02 & 36.1 @xmath65 1.3 & + eirs-10 & 560.0 @xmath65 2.9 & 722.0 @xmath65 3.6 & 823 @xmath65 7 & 1100 @xmath65 6 & 1.06 @xmath65 0.12 & 2.73 @xmath65 0.02 & & + eirs-11 & 171.0 @xmath65 1.4 & 267.0 @xmath65 1.7 & 412 @xmath65 5 & 693 @xmath65 4 & 1.06 @xmath65 0.11 & 2.34 @xmath65 0.02 & & + eirs-12 & 27.1 @xmath65 0.7 & 45.0 @xmath65 1.1 & 102 @xmath65 4 & 248 @xmath65 4 & 1.06 @xmath65 0.17 & 2.11 @xmath65 0.02 & & + eirs-13 & 99.4 @xmath65 1.2 & 81.2 @xmath65 1.2 & 68 @xmath65 4 & & 1.07 @xmath65 0.16 & 0.48 @xmath65 0.02 & & + eirs-14 & 107.0 @xmath65 1.0 & 82.8 @xmath65 1.0 & 109 @xmath65 3 & 133 @xmath65 3 & 1.12 @xmath65 0.12 & 1.30 @xmath65 0.02 & 22.7 @xmath65 1.4 & 66.1 @xmath65 5.5 + eirs-15 & 264.0 @xmath65 1.8 & 230.0 @xmath65 1.7 & 336 @xmath65 6 & 347 @xmath65 4 & 1.12 @xmath65 0.19 & 1.41 @xmath65 0.02 & & 47.3 @xmath65 5.9 + l l c l l c l l + + iras 05189 - 2524 & 0.043 & & iras 08572 + 3915 & 0.058 & & iras 12112 + 0305 & 0.073 + iras
14348 - 1445 & 0.083 & & iras 15250 + 3609 & 0.055 & & iras 22491 - 1808 & 0.078 + arp 220 & 0.018 & & mrk 231 & 0.042 & & mrk 273 & 0.038 + ugc 5101 & 0.039 & & & & & + + + + mrk 52 & 0.007 & & ngc 1222 & 0.008 & & ngc 2146 & 0.003 + ngc 3628 & 0.003 & & ngc 520 & 0.008 & & ngc 7252 & 0.016 + ngc 7714 & 0.009 & & & & & + l c c c c l eirs-1 & 0.387 & qso & zspec & a & prez - fournon ( priv . comm . ) + eirs-2 & 1.154 & galaxy & zirs & a & + eirs-3 & 0.676 & galaxy & zirs & a & + eirs-4 & 1.181 & qso & zspec & a & @xcite + eirs-5 & 1.065 & qso & zspec & a & @xcite + eirs-6 & 2.356 & qso & zspec & a & prez - fournon ( priv .
comm . ) + eirs-7 & 2.592 & galaxy & zspec & a & @xcite + eirs-8 & 0.884 & galaxy & zirs & b & + eirs-9 & 0.609 & galaxy & zirs & a & + eirs-10 & 1.265 & qso & zirs & b & + eirs-11 & 1.414 & qso & zspec & a & sdss + eirs-12 & 2.024 & galaxy & zirs & a & + eirs-13 & 1.091 & galaxy & zirs & a & + eirs-14 & 0.619 & galaxy & zirs & a & + eirs-15 & 0.827 & galaxy & zirs & a & + l r@ @xmath65 lr@ @xmath65 lr@ @xmath65 lr@ @xmath65 lr@ @xmath65 l r@ @xmath65 lr@ @xmath65 lr@ @xmath65 lr@ @xmath65 lr@ @xmath65 l eirs-1 & & 0.67 & 0.10 & 1.0 & 0.1 & 1.4 & 0.1 & 2.4 & 0.1 & & 3 & 0 & 4 & 0 & 3 & 0 & 4 & 0 + eirs-2 & 0.104 & 0.004 & 0.18 & 0.07 & 0.5 & 0.1 & 0.5 & 0.2 & 1.0 & 0.2 & 13.0 & 0.5 & 9 & 3 & 21 & 3 & 13 & 4 & 19 & 3 + eirs-3 & 0.061 & 0.002 & & & & & 2.6 & 0.1 & & & & + eirs-4 & 0.575 & 0.008 & 1.56 & 0.10 & 2.0 & 0.1 & 3.9 & 0.2 & 6.3 & 0.2 & 75.0 & 1.0 & 81 & 5 & 82 & 4 & 111 & 5 & 120 & 3 + eirs-5 & 0.362 & 0.006 & 1.08 & 0.09 & 1.2 & 0.1 & 1.6 & 0.2 & 3.0 & 0.2 & 38.5 & 0.6 & 45 & 3 & 41 & 3 & 37 & 4 & 47 & 2 + eirs-6 & 0.117 & 0.009 & & 0.3 & 0.2 & 0.9 & 0.4 & 1.0 & 0.2 & 55.1 & 4.2 & & 50 & 24 & 94 & 38 & 69 & 10 + eirs-7 & 0.387 & 0.009 & 2.33 & 0.25 & 2.7 & 0.2 & & 8.3 & 1.3 & 215.2 & 5.0 & 518 & 54 & 466 & 26 & & 680 & 102 + eirs-8 & 0.230 & 0.003 & 0.66 & 0.10 & 0.9 & 0.1 & 0.9 & 0.2 & 1.8 & 0.2 & 16.9 & 0.2 & 19 & 2 & 21 & 2 & 14 & 3 & 19 & 1 + eirs-9 & 0.099 & 0.001 & 0.35 & 0.10 & 0.9 & 0.1 & 0.4 & 0.1 & 1.8 & 0.1 & 3.4 & 0.0 & 4 & 1 & 9 & 0 & 3 & 1 & 8 & 0 + eirs-10 & 0.760 & 0.010 & 1.37 & 0.11 & 1.9 & 0.1 & 2.8 & 0.2 & 4.4 & 0.4 & 113.4 & 1.5 & 81 & 6 & 89 & 5 & 90 & 6 & 96 & 7 + eirs-11 & 0.357 & 0.009 & 1.12 & 0.17 & 1.5 & 0.1 & 2.9 & 0.2 & 4.7 & 0.7 & 66.0 & 1.7 & 82 & 12 & 88 & 8 & 119 & 6 & 126 & 19 + eirs-12 & 0.159 & 0.009 & 1.40 & 0.16 & 1.9 & 0.2 & 1.8 & 0.2 & 4.1 & 0.6 & 57.2 & 3.2 & 202 & 22 & 214 & 23 & 143 & 14 & 216 & 32 + eirs-13 & 0.080 & 0.003 & & 0.3 & 0.1 & & & 8.9 & 0.3 & & 10 & 4 & & + eirs-14 & 0.107 & 0.001 & & 0.7 & 0.1 & 0.3 & 0.1 & 1.1 & 0.1 & 3.8 & 0.0 & & 7 & 1 & 2 & 1 & 5 & 0 + eirs-15 & 0.248 & 0.004 & 0.51 & 0.10 & 0.6 & 0.1 & 0.8 & 0.2 & 1.6 & 0.1 & 15.9 & 0.3 & 12 & 2 & 12 & 2 & 11 & 2 & 14 & 1 + lr@ @xmath65 lr@ @xmath65 lr@ @xmath65 lr@ @xmath65 lr@ @xmath65 lr@ @xmath65 lr@ @xmath65 lr
eirs-1 & & & & 14.0 & 2.0 & 0.38 & 0.06 & 0.19 & 0.03 & 0.00 & 0.11 & 0.05 + eirs-2 & 9.6 & 0.9 & 1.43 & 0.14 & 1.89 & 0.18 & 40.0 & 1.5 & 5.70 & 0.21 & 7.88 & 0.29 & 0.61 & 0.61 & -0.58 + eirs-3 & 6.7 & 2.2 & 5.86 & 2.00 & 0.35 & 0.12 & 14.3 & 2.6 & 13.61 & 2.60 & 0.74 & 0.14 & -1.82 & 2.09 & 1.93 + eirs-4 & & & & & & & -0.23 & 0.08 & 0.17 + eirs-5 & 5.1 & 1.5 & 0.12 & 0.04 & 0.81 & 0.24 & 5.1 & 1.5 & 0.16 & 0.05 & 0.83 & 0.25 & 0.08 & 0.15 & -0.07 + eirs-6 & & & & & & & -0.22 & 0.50 & 0.28 + eirs-7 & & & & & & & 0.95 & 0.25 & -0.76 + eirs-8 & & & & & & & 0.48 & 0.25 & -0.51 + eirs-9 & 22.0 & 1.7 & 2.00 & 0.15 & 0.89 & 0.07 & 82.9 & 2.7 & 7.32 & 0.24 & 3.34 & 0.11 & 0.77 & 0.38 & -0.58 + eirs-10 & & & & & & & -0.05 & 0.11 & 0.13 + eirs-11 & 7.8 & 2.6 & 0.20 & 0.07 & 2.56 & 0.87 & 6.7 & 2.3 & 0.19 & 0.07 & 2.20 & 0.76 & -0.31 & 0.19 & 0.29 + eirs-12 & 5.9 & 2.4 & 0.17 & 0.07 & 4.74 & 1.88 & 13.4 & 3.1 & 0.42 & 0.10 & 10.70 & 2.48 & 0.95 & 0.21 & -0.77 + eirs-13 & 5.3 & 1.6 & 1.21 & 0.38 & 0.90 & 0.28 & 21.2 & 2.4 & 5.49 & 0.63 & 3.64 & 0.42 & 0.80 & 1.98 & -@xmath121 + eirs-14 & 25.0 & 2.9 & 3.11 & 0.37 & 1.05 & 0.12 & 64.5 & 3.5 & 8.47 & 0.47 & 2.71 & 0.15 & 0.84 & 0.60 & -0.38 + eirs-15 & 2.3 & 1.1 & 0.10 & 0.05 & 0.20 & 0.09 & 10.6 & 2.2 & 0.59 & 0.12 & 0.91 & 0.19 & 0.13 & 0.28 & -0.02 + c | c c c c c1 & @xmath81 & @xmath26 0.2 & @xmath26 0.2 & @xmath1 0.2 + c2 & l@xmath122/@xmath55 & @xmath1 0.1 & @xmath1 0.1 & @xmath26 0.1 + c3 & @xmath79 & @xmath26 15% & @xmath26 15% & @xmath1 15% + c4 & @xmath123 & @xmath26 0.15 & 0.151.0 & @xmath1 1.0 + c5 & @xmath48 & @xmath1 0 & @xmath26 0 & @xmath26 0 + c6 & l@xmath124/l@xmath125 & @xmath26 2 & 24 & @xmath1 4 + lcrcrcrcrcrcrcc eirs-1 & 0.387 & @xmath26 0.112 & 1,2 & 0.486 & 1,2 & 3.9 & 1,2 & 0.260 & 2 & 0.05 & 1 & 1.777 & 1 & 1 + eirs-2 & 1.154 & 1.434 & s & 0.030 & s & 57.1 & s & 8.010 & s & -0.58 & 2,s & 2.068 & 2 & s + eirs-3 & 0.676 & 5.861 & s & @xmath26 0.268 & s & 57.5 & s & 13.880 & s & 1.28 & 1 & 0.721 & 1 & s + eirs-4 & 1.181 & @xmath26 0.040 & 1,2 & 0.225 & 1,2 & 0.8 & 1,2 & @xmath26 0.046 & 1 & 0.16 & 1 & 1.614 & 1 & 1 + eirs-5 & 1.065 & 0.125 & 1,2 & 1.024 & 1,2 & 3.7 & 1,2 & 0.310 & 2 & -0.07 & 2,s & 1.875 & 1 & 2 + eirs-6 & 2.356 & @xmath26 0.306 & 1,2 & @xmath26 0.558 & s & 0.0 & 1,2 & @xmath26 0.340 & 1 & -0.26 & 2,s & 1.101 & 1 & 1 + eirs-7 & 2.592 & @xmath26 0.052 & 1,2 & 0.316 & 1,2 & 0.0 & 1,2 & @xmath26 0.052 & 1 & -0.76 & 2,s & & - & 2@xmath126 + eirs-8 & 0.884 & @xmath26 0.152 & 1,2 & 0.528 & 1,2 & 1.1 & 1,2 & @xmath26 0.274 & 1 & -0.51 & 2,s & 2.011 & 2 & 1 + eirs-9 & 0.609 & 2.004 & s & 0.046 & s & 46.0 & s & 6.560 & s & -0.58 & 2,s & 4.036 & s & s + eirs-10 & 1.265 & @xmath26 0.066 & 1,2 & 0.899 & 1,2 & 0.0 & 1,2 & @xmath26 0.066 & 1 & 0.20 & 1 & 1.588 & 1 & 1 + eirs-11 & 1.414 & 0.199 & 1,2 & 0.654 & 1,2 & 3.3 & 1,2 & 0.300 & 2 & 0.17 & 1 & 1.599 & 1 & 1 + eirs-12 & 2.024 & 0.165 & 1,2 & 0.117 & 1,2 & 4.9 & 1,2 & 0.260 & 2 & -0.77 & 2,s & 2.264 & 2 & 2 + eirs-13 & 1.091 & 1.208 & s & @xmath26 0.084 & s & 51.5 & s & 5.920 & s & -0.80 & 2,s & @xmath1 1.069 & s & s + eirs-14 & 0.619 & 3.113 & s & @xmath26 0.048 & s & 45.7 & s & 7.690 & s & -0.38 & 2,s & 3.486 & 2 & s + eirs-15 & 0.827 & 0.103 & 1,2 & 0.878 & 1,2 & 11.6 & 1,2 & 0.900 & 2 & -0.02 & 2,s & 1.893 & 1 & 2 + c c
c l rigopoulou ( 1999 ) & starburst & 7.7 m & @xmath66/@xmath55 = 0.0081 + & ulirg starburst & 7.7 m & l@xmath127/@xmath55 = 0.0055 + lutz ( 2003 ) & starburst & 7.7 m & l@xmath127/@xmath55 = 0.033 + spoon ( 2004 ) & normal and starburst & 6.2 m & l@xmath128/@xmath55 = 0.0034 + smith ( 2007 ) & normal and starburst & 6.2 m & l@xmath128/@xmath55 = 0.011 + & normal and starburst & 7.7 m & l@xmath127/@xmath55 = 0.041 + farrah ( 2008 ) & starburst & 6.2 + 11.3 m & l@xmath129/@xmath55 = 0.038 + this work & ulirg starburst & 6.2 m & l@xmath128/@xmath55
= 0.012 + & ulirg starburst & 7.7 m & l@xmath127/@xmath55
= 0.038 + & ulirg starburst & 11.3 m & l@xmath130/@xmath55 = 0.011 lr@ @xmath65 lr@ @xmath65 lr@ @xmath65 lccr@ @xmath65 lr@ @xmath65 l eirs-1 & & & 10.8 & 0.3 & qso high & 6.2 11.3 & & + eirs-2 & 16.4 & 1.0 & 70.2 & 5.1 & 12.4 & 0.1 & dale26 & 6.2 7.7 11.3 & 0.69 & 0.03 & 332 & 12 + eirs-3 & & & 11.0 & 0.3 & seyfert 2 & 6.2 7.7 11.3 & 2.03 & 0.34 & 40 & 6 + eirs-4 & 33.1 & 1.7 & & 12.5 & 0.5 & strong fir qso & 6.2 7.7 & & + eirs-5 & & & 11.6 & 0.3 & qso low & 6.2 7.7 11.3 & 1.47 & 0.19 & 115 & 15 + eirs-6 & & & 12.1 & 0.3 & qso high & 6.2 7.7 & & + eirs-7 & 13.0 & 1.2 & & 13.0 & 0.5 & mrk 231 & 6.2 7.7 & & + eirs-8 & & & 11.3 & 0.3 & qso low & 6.2 7.7 11.3 & & + eirs-9 & 36.1 & 1.3 & & 11.9 & 0.2 & ngc 6240 & 6.2 7.7 11.3 & 0.88 & 0.03 & 124 & 4 + eirs-10 & & & 12.1 & 0.3 & weak fir qso & 6.2 7.7 & & + eirs-11 & & & 12.0 & 0.3 & qso norm & 6.2 7.7 11.3 & 1.06 & 0.22 & 223 & 45 + eirs-12 & & & 13.1 & 0.3 & m82 & 6.2 7.7 & 0.25 & 0.06 & 576 & 143 + eirs-13 & & & 12.0 & 0.3 & gl12 & 6.2 7.7 11.3 & 0.99 & 0.10 & 173 & 17 + eirs-14 & 22.7 & 1.4 & 66.1 & 5.5 & 11.9 & 0.1 & dale26 & 6.2 7.7 11.3 & 0.86 & 0.05 & 119 & 6 + eirs-15 & & 47.3 & 5.9 & 11.3 & 0.3 & weak fir qso & 6.2 7.7 11.3 & 1.12 & 0.26 & 42 & 9 + c| r c r r z @xmath26 1.5 & 70 m & 11 @xmath65 2 & 8.5@xmath6210@xmath107 & 8.4@xmath6210@xmath107 + ( 22 sources ) & 160 m & 27 @xmath65 9 & 1.3@xmath6210@xmath104 & 8.9@xmath6210@xmath107 + + z @xmath1 1.5 & 70 m & 5.3 @xmath65 1.1 & 4.4@xmath6210@xmath104 & 8.3@xmath6210@xmath104 + ( 18 sources ) & 160 m & 28 @xmath65 7 & 6.3@xmath6210@xmath104 & 5.3@xmath6210@xmath104 + c c c c c
c qsos & 6.2 m & @xmath260.01 & @xmath267.1 @xmath62 10@xmath131 & @xmath265.9 @xmath62 10@xmath107 & @xmath26100 + high lum . &
7.7 m & @xmath260.03 & @xmath261.7 @xmath62 10@xmath132 & @xmath264.5 @xmath62 10@xmath107 & @xmath2680 + ( 18 sources ) & 11.3 m & @xmath260.01 & @xmath263.5 @xmath62 10@xmath131 & @xmath263.1 @xmath62 10@xmath107 & @xmath2653 + qsos & 6.2 m & 0.03 & 3.3 @xmath62 10@xmath131 & 2.7 @xmath62 10@xmath107 & 46 + low lum . & 7.7 m & 0.09 & 8.1 @xmath62 10@xmath131 & 2.1 @xmath62 10@xmath107 & 37 + ( 18 sources ) & 11.3 m & 0.05 & 3.4 @xmath62 10@xmath131 & 3.0 @xmath62 10@xmath107 & 52 + obsc .
agn & 6.2 m & @xmath260.04 & @xmath262.5 @xmath62 10@xmath132 & @xmath262.1 @xmath62 10@xmath104 & @xmath26352 + high lum . &
7.7 m & @xmath260.10 & @xmath265.6 @xmath62 10@xmath132 & @xmath261.5 @xmath62 10@xmath104 & @xmath26257 + ( 5 sources ) & 11.3 m & @xmath260.20 & @xmath263.8 @xmath62 10@xmath132 & @xmath263.5 @xmath62 10@xmath104 & @xmath26589 + obsc .
agn & 6.2 m & 0.10 & 8.1 @xmath62 10@xmath131 & 6.7 @xmath62 10@xmath107 & 114 + low lum . & 7.7 m & 0.34 & 2.6 @xmath62 10@xmath132 & 6.8 @xmath62 10@xmath107 & 121 + ( 6 sources ) & 11.3 m & 0.07 & 3.6 @xmath62 10@xmath131 & 3.3 @xmath62 10@xmath107 & 54 + | we present results on low - resolution mid - infrared ( mir ) spectra of 70 infrared - luminous galaxies obtained with the infrared spectrograph ( irs ) onboard _
spitzer_. we selected sources from the european large area infrared survey ( elais ) with @xmath0 @xmath1 0.8 mjy and photometric or spectroscopic @xmath2 @xmath1 1 .
about half of the sample are qsos in the optical , while the remaining sources are galaxies , comprising both obscured agn and starbursts .
redshifts were obtained from optical spectroscopy , photometric redshifts and from the irs spectra .
the later turn out to be reliable for obscured and/or star - forming sources , thus becoming an ideal complement to optical spectroscopy for redshift estimation .
we estimate monochromatic luminosities at several restframe wavelengths , equivalent widths and luminosities for the pah features , and strength of the silicate feature in individual spectra .
we also estimate integrated 81000 m infrared luminosities via spectral energy distribution fitting to mir and far - infrared ( fir ) photometry from the spitzer wide - area infrared extragalactic survey ( swire ) and the mir spectrum . based on these measurements ,
we classify the spectra using well - known infrared diagnostics , as well as a new one that we propose , into three types of source : those dominated by an unobscured agn , mostly corresponding to optical quasars ( qsos ) , those dominated by an obscured agn , and starburst - dominated sources .
starbursts concentrate at @xmath2 @xmath3 0.61.0 favored by the shift of the 7.7-m pah band into the selection 15-m band , while agn spread over the @xmath4 range
. star formation rates ( sfr ) are estimated for individual sources from the luminosity of the pah features .
an estimate of the average pah luminosity in qsos and obscured agn is obtained from the composite spectrum of all sources with reliable redshifts .
the estimated mean sfr in the qsos is 50100 @xmath5 , but the implied fir luminosity is 310 times lower than that obtained from stacking analysis of the fir photometry , suggesting destruction of the pah carriers by energetic photons from the agn .
the sfr estimated in obscured agn is 23 times higher than in qsos of similar mir luminosity
. this discrepancy might not be due to luminosity effects or selection bias alone , but could instead indicate a connection between obscuration and star formation .
however , the observed correlation between silicate absorption and the slope of the near- to mid - infrared spectrum is compatible with the obscuration of the agn emission in these sources being produced in a dust torus .
[ firstpage ] galaxies : starburst galaxies : high - redshift galaxies : active quasars : general |
single - top quarks are produced via the electroweak interaction and at leading order there are three different production mechanisms ; s - channel , t - channel , and @xmath4-channel , each of which are depicted in figure [ fig : singletop ] . the theoretical cross sections for the single - top processes are @xmath5
@xcite for the t - channel , @xmath6 @xcite for the s - channel , and @xmath7 @xcite for the @xmath4-channel .
note that each single - top process has a @xmath8 vertex .
only single - top events with an isolated and high-@xmath9 electron or muon in the final state are included in this study since events with only hadrons in the final state are difficult to distinguish from background .
ccc -channel single - top production.,width=113,height=113 ] -channel single - top production.,width=113,height=113 ] -channel single - top production.,width=113,height=113 ] the dominant background to single - top production is @xmath10 production , which has a cross section three times larger than the combined single - top production cross section . with a single high-@xmath9 lepton , two @xmath11-jets , and @xmath12 ,
semi - leptonic @xmath10 decay is most likely to mimic single - top events . due to the lack of leptons in the final state , fully - hadronic @xmath10 events are not a major contributor to single - top background .
@xmath13 + jets production constitutes a significant background since the cross - sections for these processes are several orders of magnitude greater than the single - top cross sections .
the background from di - boson events is minimal .
qcd will be estimated by data driven methods and is not considered in these studies .
the amount of qcd contamination depends on the specific selections used in the analyses .
the single - top processes were generated using pythia @xcite , with the matrix element calculated using acermc @xcite .
the @xmath10 sample , which included both di - lepton and semi - leptonic decays , was generated using herwig @xcite and the matrix element was calculated using mc@nlo @xcite .
the @xmath13 + jets sample was generated using pythia and the matrix element was calculated with alpgen @xcite .
mcfm @xcite was used to derive the k - factors needed to scale the lo processes calculated with alpgen to nlo .
the three single - top processes shared a common pre - selection . muons and electrons were required to satisfy reconstruction requirements of @xmath14 and @xmath15 and an isolation requirement of @xmath16 in a cone of radius 0.2 around the particle axis .
events were required to contain one muon or electron with @xmath17 and events with secondary leptons were removed to eliminate contamination from di - lepton @xmath10 events and to ensure the orthogonality of the muon and electron samples .
jet candidates were reconstructed using a cone algorithm with @xmath18 and were required to satisfy @xmath19 to be considered a jet . an event was required to have between 2 and 4 jets , with at least two of the jets having @xmath17 and at least one of the jets being @xmath11-tagged .
events were required to have @xmath20 , which corresponds to the energy of the non - detected neutrino present in leptonic @xmath13 decay .
the measurement of the single - top cross sections in the atlas detector @xcite will be obtained using the formula @xmath21 where @xmath22 is the total number of events in the data , @xmath23 is the number of expected background events , @xmath24 is the selection efficiency for single - top signal events , and @xmath25 is the luminosity .
cross section errors were estimated by randomly generating @xmath22 according to a poisson distribution and randomly varying @xmath23 and @xmath24 for every systematic quantity by an amount determined by a gaussian distribution around the central value of that quantity .
the sources of experimental uncertainty were jet energy scale ( jes ) , @xmath11-tagging likelihood , and luminosity .
the sources of theoretical uncertainties were background cross sections , initial state radiation ( isr ) and final state radiation ( fsr ) , pdfs , and @xmath11 quark fragmentation .
cut - based and multivariate analyses of the cross section measurements were performed for each of the three single - top channels in order to understand the size of the statistical and systematic errors and the amount of integrated luminosity needed to obtain evidence and achieve discovery of the single - top quark . a full description of the single - top analyses can be found in @xcite .
the cut - based analysis for the t - channel was performed by requiring , in addition to the event pre - selection , a @xmath11-tagged jet with @xmath26 in order to remove low-@xmath9
@xmath13 + jets background and @xmath27 for the hardest light jet to remove @xmath10 contamination . after this selection
, there were 1,460 signal and 3,906 background events for a 1@xmath3 sample .
the quantity of @xmath10 background remaining after the t - channel specific event selection necessitated the use of a boosted decision tree ( bdt ) .
variables providing a good signal to background separation were used as input to the bdt and a cut of 0.6 on the bdt output minimized the total uncertainty on the cross section and corresponded to a signal over background ratio of 1.3 .
the bdt output is shown on the left side of figure [ fig : t - channel : bdt ] and the leptonic top mass distribution using a cut of 0.6 on the bdt output is shown on the right side of figure [ fig : t - channel : bdt ] .
ccc table [ tbl : t - channel : errors ] lists the uncertainties for 1@xmath3 and 10@xmath3 of integrated luminosity . for both analysis methods and both integrated luminosity samples ,
the systematic uncertainties far outweigh the statistical uncertainties .
the main systematic uncertainties on the t - channel cross section measurement were jes , isr and fsr , and luminosity .
the systematics achieved using the bdt were approximately half of those obtained with the cut based analysis .
.uncertainties on the t - channel cross section measurement for 1@xmath3 and 10@xmath3 of integrated luminosity . [ cols="^,^,^,^,^",options="header " , ]
for a 5@xmath28 single - top quark discovery in the t - channel , 1@xmath3 of integrated luminosity is needed while 30 @xmath3 is needed for 3@xmath28 evidence in the s - channel . 1 @xmath3
is needed for evidence and 10 @xmath3 is needed for discovery in the @xmath4-channel .
systematic uncertainties are the limiting factor in the single - top cross section measurement , except for the s - channel where statistics also play a limiting role .
the systematics in the current analysis have a strong dependence on monte carlo and the use of data - driven analysis techniques should help minimize them . | the single - top production cross section is one third that of the top - pair production cross section at the lhc . during a year of data - taking , assuming an average luminosity of @xmath0 and a cms energy of @xmath1 , the determination of the major contributions to the total single - top cross section should be achievable .
comparisons between the measured cross sections and the theoretical predictions will provide a crucial test of the standard model .
these measurements should also lead to the first direct measurement of @xmath2 , with a precision at the level of a few percent .
in addition , they will probe for new physics via the search for evidence of anomalous couplings to the top quark and the measurements of additional bosonic contributions to single - top production .
methods developed to optimize the selection of single - top events in the three production channels are presented and the potential for the cross section measurements with 1@xmath3 and 30@xmath3 of integrated luminosity is established . |
[ [ a .- toeplitz - determinant - for - chikappa . ] ] a. toeplitz determinant for @xmath74. + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the fcs generating function ( [ chi ] ) can be written as a trace in the multi - particle space and then re - expressed as a determinant in the single - particle space @xcite : @xmath76\ , , \label{determinant-1}\ ] ] where @xmath77 is the fermi occupation number and @xmath78 is the operator of multiplication by @xmath8 .
the operator @xmath77 is diagonal in the momentum space , while @xmath78 is diagonal in the coordinate space . in this paper
, we consider the zero - temperature case , so @xmath77 is a projector onto the occupied states ( below @xmath2 ) . in the limit of no smoothing ( @xmath12 )
, @xmath78 is also a projector ( in the coordinate space ) , and the determinant is symmetric with respect to exchanging coordinate and momentum : @xmath79\ , .\ ] ] this operator is of toeplitz ( or wiener hopf ) form in both coordinate and momentum representations . at @xmath46 , the term with @xmath24 in the decomposition ( [ expansion - a ] ) does not oscillate with @xmath9 .
it therefore arises from the end points of the segment of measurement ( @xmath80 and @xmath81 ) .
this allows us to calculate @xmath28 directly from a single - step contribution : @xmath82\ , , \label{det - one - step}\ ] ] where @xmath83 is the operator of multiplication by the single - step counterpart of @xmath8 : @xmath84 note that while @xmath85 has different limits at @xmath86 , the exponent @xmath87 tends to 1 in both limits , and the determinant ( [ det - one - step ] ) is well defined . for rc and ga models ( the same as in fig .
[ fig-2]d ) as calculated numerically , together with the analytic approximations given by ( [ c1-asympt1 ] ) . *
( b ) * the difference between the approximate and exact values of @xmath28 ( upper curve : ga model , lower curve : rc model ) . ] for numerical calculations of @xmath29 and @xmath28 , we approximate the integral operators ( [ determinant-1 ] ) and ( [ det - one - step ] ) by finite - dimensional matrices by discretizing the momentum space ( which is equivalent to considering the system on a circle ) . in the momentum space ,
these matrices are of toeplitz form .
the determinants of those matrices are calculated numerically for several matrix sizes and then extrapolated to the infinite matrix size in order to obtain @xmath29 and @xmath28 , respectively . for numerical calculations of the determinants ,
we use the lapack library @xcite , and fourier transformations necessary for computing the matrix elements were done with the help of the gnu scientific library @xcite . in our calculations ,
determinants of matrices of the linear size up to 600 were calculated , which allowed us to achieve good precision for the considered ranges of parameters .
the values of @xmath88 were then fitted to the expansion ( [ expansion - a ] ) , and the coefficients @xmath26 and @xmath24 were extracted ( the fit also included terms up to @xmath89 in the exponents ) . as an additional check of this fitting procedure , we have verified that the values of @xmath28 obtained from these fits agree with those calculated independently using the one - step method described above . as a function of @xmath21 in the rc model at @xmath90 for selected values of @xmath25 .
periodicity in @xmath21 is lost with increasing @xmath25 . *
( b ) * @xmath91 as a function of @xmath38 in the rc model at @xmath46 and @xmath92 . for large @xmath38 ,
the second term in ( [ expansion - a ] ) dominates . ]
the approximation ( [ c1-asympt1 ] ) , while derived under the assumption @xmath94 , is remarkably accurate even at small @xmath25 . in the lo model , where @xmath28 is given exactly by ( [ c1-lo - exact ] )
, the approximation ( [ c1-asympt1 ] ) reproduces the same exact result for all @xmath25 . in the rc and ga models
, the difference between ( [ c1-asympt1 ] ) and @xmath28 tends to zero as @xmath95 , but remains finite at finite @xmath10 .
however , this difference is numerically small for these two models , even at @xmath12 .
we plot this difference as a function of @xmath25 in fig .
[ one_tail_analytic ] .
we illustrate the crossover described in the main body of the paper by two plots . in fig .
[ two - plots]a , we show how the periodicity of @xmath29 in @xmath21 is gradually lost as @xmath25 increases .
figure [ two - plots]b illustrates how the second term in ( [ expansion - a ] ) dominates at large @xmath9 ( in rc model , for @xmath46 , this happens at @xmath96 ) . m. stone , _ bosonization _ ( world scientific , 1994 ) ; + a. o. gogolin , a. a. nersesyan , and a. m. tsvelik , _ bosonization in strongly correlated systems _ , ( university press , cambridge 1998 ) ; + t. giamarchi , _ quantum physics in one dimension _ , ( claverdon press oxford , 2004 ) .
i. p. levkivskyi and e. v. sukhorukov , phys .
103 * , 036801 ( 2009 ) .
+ _ noise - induced phase transition in the electronic mach - zehnder interferometer . _ + a. helzel , l. v. litvin , i. p. levkivskyi , e. v. sukhorukov , w. wegscheider , and c. strunk , phys .
b * 91 * , 245419 ( 2015 ) .
+ _ counting statistics and dephasing transition in an electronic mach zehnder interferometer .
_ d. b. gutman , y. gefen , and a. d. mirlin , europhys . lett . * 90 * , 37003 ( 2010 ) .
+ _ bosonization out of equilibrium . _ + phys .
b * 81 * , 085436 ( 2010 ) .
+ _ bosonization of one - dimensional fermions out of equilibrium . _ n. kitanine , k. k. kozlowski , j. m. maillet , n. a. slavnov , and v. terras , comm . math .
phys . * 291 * , 691 ( 2009 ) .
+ _ riemann hilbert approach to a generalized sine kernel and applications . _ + k. k. kozlowski , arxiv:0805.3902 .
+ _ truncated wiener
hopf operators with fisher hartwig singularities .
_ d. a. ivanov , a. g. abanov , and v. v. cheianov , j. phys .
a : math . theor .
* 46 * , 085003 ( 2013 ) . + _ counting free fermions on a line : a fisher hartwig asymptotic expansion for the toeplitz determinant in the double - scaling limit .
_ d. a. ivanov and a. g. abanov , j. phys . a : math .
theor . * 46 * , 375005 ( 2013 ) .
+ _ fisher hartwig expansion for toeplitz determinants and the spectrum of a single - particle reduced density matrix for one - dimensional free fermions . _ k. s. tikhonov and m. v. feigelman , phys . rev .
lett . * 105 * , 067207 ( 2010 ) .
+ _ quantum spin metal state on a decorated honeycomb lattice . _ + k. s. tikhonov , m. v. feigelman , and a. yu . kitaev , phys . rev .
lett . * 106 * , 067203 ( 2011 ) .
+ _ power - law spin correlations in a perturbed spin model on a honeycomb lattice . _ | we revisit the problem of full counting statistics of particles on a segment of a one - dimensional gas of free fermions . using a combination of analytical and numerical methods , we study the crossover between the counting of discrete particles and of the continuous particle density as a function of smoothing in the counting procedure . in the discrete - particle limit
, the result is given by the fisher hartwig expansion for toeplitz determinants , while in the continuous limit we recover the bosonization results .
this example of full counting statistics with smoothing is also related to orthogonality catastrophe , fermi - edge singularity and non - equilibrium bosonization .
[ [ introduction . ] ] introduction. + + + + + + + + + + + + + + full counting statistics ( fcs ) , in theoretical - condensed - matter context , refers to a class of problems involving the probability distribution of a quantum observable ( usually the number of electrons found in a certain region of space or transported through the system over a certain time ) with a particular focus on quantum behavior .
examples of fcs problems are the anti - bunching of electrons in a one - dimensional conductor due to their fermionic statistics @xcite and a single - electron emitter , first proposed theoretically @xcite and recently realized experimentally @xcite .
the simplest fcs problems assume non - interacting fermions ( electrons ) , so that the resulting counting statistics can be expressed in terms of a determinant taking into account antisymmetrization of the relevant multi - particle processes @xcite .
an alternative approach is based on the bosonization technique ( using the equivalence between bosons and fermions in one dimension ) @xcite .
bosonization methods for fcs can be extended to include interactions , but they usually do not fully take into account the discreteness of particles @xcite .
it is therefore important to understand limitations of the bosonization approach to fcs and its connection to the exact calculation with discrete fermions . in this letter
, we address this problem by studying fcs in a one - dimensional free - fermion model , where an uncertainty ( smoothing ) in the counting procedure is introduced , so that the particle number is no longer quantized . in this model
, we can study in full detail a crossover between the discrete - particle and bosonization results as the uncertainty increases . while the details of this crossover depend on the profile of the introduced uncertainty , the qualitative description of the crossover is found to be universal : fcs respects the discreteness of particles if the uncertainty region is much narrower than the ( average ) inter - particle distance , but crosses over to the bosonization result when the uncertainty region is much wider than the inter - particle distance .
mathematically , this problem amounts to studying an evolution of the asymptotic behavior of a toeplitz ( or , more precisely , wiener
hopf ) determinant with a fisher hartwig singularity @xcite as the singularity is smoothed in a certain way . in the case of a sharp singularity ,
the corresponding determinant is given by a double asymptotic series of a fisher hartwig type fully respecting the particle discreteness @xcite . as the singularity is smoothed , the secondary branches of this expansion get suppressed , and the remaining leading branch reproduces the bosonization result . for @xmath0 and @xmath1 for the three models of smoothing . ]
[ [ model . ] ] model. + + + + + + + we consider the one - dimensional model of spinless free fermions on a line ( both coordinate and momentum are continuous ) at zero temperature .
the system is in its ground state , which is the slater determinant of plane waves characterized by the fermi wave vector @xmath2 : the states with the wave vectors smaller than @xmath2 ( in absolute value ) are occupied , and the states with the wave vectors larger than @xmath2 are empty .
we introduce the fcs generating function as @xcite @xmath3 where the average is taken over the ground state . the particle - counting operator @xmath4 is defined as @xmath5 where @xmath6 and @xmath7 are the fermionic creation and annihilation operators and @xmath8 is the weight function for the particle counting .
the weight function depends on the two parameters : the length of the interval @xmath9 and the smoothing length scale @xmath10 , fig .
[ fig-1 ] .
we always assume @xmath11 . in the original formulation of the fcs problem , the smoothing is absent ( @xmath12 ) , and
the weight function is the characteristic function of a line segment , @xmath13 in this case , @xmath4 takes integer values , and the generating function ( [ chi ] ) is periodic in the counting variable , @xmath14 .
it can be expressed as a toeplitz determinant ( see appendix ) , which has been subject to extensive studies @xcite .
the periodicity of @xmath15 is reflected in a fisher hartwig - type asymptotic series ( conjectured and numerically verified to a very high order ) @xcite : @xmath16 @xmath17\ , .
\label{expansion-2}\end{gathered}\ ] ] [ expantot ] the coefficients in this double expansion can be explicitly calculated : @xmath18 is expressed in terms of barnes g functions as @xmath19 , and @xmath20 are polynomials in @xmath21 , which can be computed iteratively , order by order @xcite ( the same coefficients are denoted @xmath22 in @xcite ) .
@xmath15 depends only on one parameter : the average particle number on the segment , @xmath23 and @xmath24 as a function of @xmath21 for several values of @xmath25 , as computed from numerical fits .
@xmath26 are shown by solid lines , @xmath24 by dashed lines .
different symbols correspond to different values of @xmath25 as shown in the legends . in panel
( c ) , the dash - dotted line is the bosonization result ( [ bosonization - asymptotics ] ) at @xmath27 . *
( d ) * : coefficient @xmath28 as a function of @xmath25 for the three models .
the plot for lo model is the exact analytical result ( [ c1-lo - exact ] ) .
the dotted line is the leading asymptotics ( [ c1-rc - asympt2 ] ) for rc model . ] on the other hand , a bosonization approach proposed for the generating function @xmath29 @xcite leads to @xmath30 \big)\ , , \label{bosonization - general}\ ] ] where we have defined the average number of particles in the smoothing region @xmath31 and the constant at the logarithm depends on the details of the smoothing .
the bosonization expression ( [ bosonization - general ] ) corresponds to the leading terms in the leading branch ( @xmath32 ) of the double expansion ( [ expantot ] ) .
we expect ( and confirm it the present work ) that this result applies provided the smoothing @xmath10 is sufficiently large .
it is not periodic in @xmath21 , which is consistent with a continuous spectrum of the operator @xmath4 at a finite @xmath10 .
the goal of the present study is to examine in detail the crossover from the discrete particle counting given by the expansion ( [ expantot ] ) to the bosonization result ( [ bosonization - general ] ) as a function of the smoothing parameter @xmath10 .
some of the results are derived analytically , while a more detailed picture of the crossover is obtained numerically .
smoothing is introduced as a convolution @xmath33 where @xmath34 is a smoothing function for the scaled coordinate @xmath35 .
we always assume the normalization @xmath36 .
to be specific , we consider three models of smoothing corresponding to the three choices of the smoothing function : rc ( this type of smoothing may be relevant for particle - counting in time , where the counter has a rc - type response function ) , lorentzian ( lo ) and gaussian ( ga ) : @xmath37^{-1 } & \text{lo , } \\
\exp(-\tilde{x}^2/2)/\sqrt{2\pi } & \text{ga } \end{cases}\ ] ] ( see fig . [ fig-1 ] ) . in any of these models ,
the fcs generating function @xmath29 depends on @xmath9 and @xmath10 via the two dimensionless parameters @xmath38 and @xmath25 .
[ [ results . ] ] results. + + + + + + + + + we conjecture ( and support this conjecture by analytical and numerical results ) that , at a finite @xmath10 , the fcs generating function admits the same structure of the expansion ( [ expantot ] ) as at @xmath12 , except that the expansion coefficients are not periodic in the branch index @xmath39 and depend on the smoothing parameter @xmath10 .
if we limit the range of @xmath21 to the interval @xmath40 $ ] , we can write the leading terms of the expansion as @xmath41 \\ + \exp \big [ 2 \pi i ( \kappa-1 ) n_l - 2 ( \kappa-1)^2 \ln n_l + c_1(\kappa , a ) \big ] \ , . \label{expansion - a}\end{gathered}\ ] ] at @xmath12 , in agreement with the full expansion ( [ expantot ] ) , the coefficients @xmath26 and @xmath24 are given by @xmath42 we expect that , as @xmath10 increases , @xmath43 also increases to reproduce the bosonization result ( [ bosonization - general ] ) , while @xmath44 decreases to suppress the second term in the expansion ( [ expansion - a ] ) .
this behavior is indeed confirmed by numerical studies .
we report the details of our numerics in the appendix , an here we only plot in fig .
[ fig-2]a - c the results of the numerical fits for the coefficients @xmath26 and @xmath24 for the three models of the smoothing .
figure [ fig-2]d shows more detailed results for the coefficient @xmath45 representing the suppression of the secondary fisher hartwig branch at @xmath46 .
this coefficient was , in fact , computed as a doubled contribution from a single `` phase slip '' ( at the beginning or at the end of the interval ) in the weight function @xmath8 ( note that at @xmath46 the values @xmath47 and @xmath48 are equivalent , see appendix ) .
an agreement between the results computed with this method ( at @xmath46 ) and the fitting procedure for the full function @xmath29 serves as an independent check of our numerical scheme . as expected , while the actual values of the coefficients @xmath26 and @xmath24 depend on the chosen smoothing model , qualitatively the crossover between the discrete particle counting and a continuous bosonization description occurs at @xmath49 in all the three models .
furthermore , some of the numerical results presented above can be verified by analytical means .
first , the constant in the bosonization formula ( [ bosonization - general ] ) can be computed from the corresponding toeplitz determinant using the strong szeg theorem @xcite .
this results in the asymptotic behavior for @xmath26 : @xmath50 where @xmath51 \ , , \label{upsilon - def}\ ] ] in the case of a general smoothing model ( [ smoothing ] ) . here
@xmath52 is the euler mascheroni constant and @xmath53 are the fourier components of the smoothing function , @xmath54 for the three models considered in our paper , a calculation gives @xmath55 these analytical results perfectly agree with our numerical fits : for the rc and lorentzian smoothing , the computed values of @xmath26 for @xmath56 and @xmath57 , respectively , are indistinguishable in the plots of fig .
[ fig-2]a , b from the bosonization asymptotics ( [ bosonization - asymptotics ] ) , ( [ upsilon - values ] ) . for the gaussian smoothing , the difference between @xmath26 and the bosonization asymptotics is visible for @xmath27 , but not for @xmath56 ( fig .
[ fig-2]c ) .
second , the asymptotic behavior of @xmath45 as @xmath58 can also be calculated analytically using an asymptotic formula for the toeplitz determinant with a non - zero winding number @xcite .
a calculation results in @xmath59\ , , \label{c1-asympt1}\end{gathered}\ ] ] where @xmath60 is given by ( [ upsilon - def ] ) and @xmath61 is defined as @xmath62\ , .
\label{xi - def}\ ] ] an explicit calculation gives @xmath63 remarkably , the asymptotics ( [ c1-asympt1 ] ) , while not exactly coinciding with @xmath45 , gives a very good approximation in the whole range of the values of @xmath10 ( in the plot in fig .
[ fig-2]d it would be indistinguishable from the exact values , see details in appendix ) .
the analytic approximation ( [ c1-asympt1 ] ) also allows us to extract the leading asymptotic behavior .
for the rc model , we find @xmath64 i.e. , the secondary branch of @xmath65 decays as @xmath66 .
for the lorenzian smoothing , the toeplitz determinant for a single `` phase slip '' @xmath67 may be computed exactly by using the analyticity of the function @xmath68 $ ] in one of the half - planes . an argument in the spirit of @xcite then leads to the exact result : @xmath69 finally , for the gaussian smoothing , the asymptotic behavior is @xmath70 however , this expression by itself [ unlike the integral formula ( [ c1-asympt1 ] ) ] does not provide a good approximation for @xmath45 , since the next - order corrections are not @xmath71 .
remarkably , even though the coefficient at the secondary branch is reduced with increasing @xmath25 , the dependence on @xmath38 does not change . in particular , for @xmath72 , the bosonization ( first ) term in ( [ expansion - a ] ) always decreases with @xmath38 faster than the non - bosonization ( second ) one .
this implies that the non - bosonization corrections arising from the discreteness of particles should be visible , in this range of @xmath21 , at sufficiently large @xmath9 ( see appendix ) .
the corresponding values of @xmath9 depend on the particular choice of smoothing and can be deduced from comparing the two terms in ( [ expansion - a ] ) using the asymptotics at large @xmath25 derived above .
a practical observation of this effect may be limited at large @xmath25 if both terms in ( [ expansion - a ] ) become too small .
[ [ discussion . ] ] discussion. + + + + + + + + + + + + in this letter , we have discussed the crossover from discrete to continuous fcs in the model of free one - dimensional fermions , as a function of smoothing .
this model may serve as an illustration of a relation between discrete and continuous descriptions in a wide spectrum of similar problems , including fcs with temporal measurements , fermi - edge singularity , orthogonality catastrophe , and non - equilibrium bosonization . by fcs with temporal measurements
we understand a fcs setup where the counting is performed over a certain time interval , e.g. , by opening and closing an electric contact or by applying a time - dependent voltage pulse @xcite .
this formulation of fcs used in the original fcs papers differs from our approach where the measurement is extended in space instead of time .
this difference implies the necessity of regularizing the contribution of the fermi sea in temporal fcs : a large current of left - movers is compensated by a similarly large current of right - movers , so the average charge transfer is determined by the vicinity of the fermi level , while the quantum noise depends on the ultraviolet cutoff defined by the fermi energy @xmath73 ( arising from the bottom of the fermi sea ) . in our spatial formulation , there is no cutoff introduced by the bottom of the fermi sea , and both the total number of particles and its fluctuations depend on the same scale @xmath2 . as a consequence of this difference in ultraviolet cutoffs ,
our results can not be literally translated to the case of temporal fcs .
however , we expect that the main qualitative conclusion will hold also in the temporal case : the generating function @xmath74 loses periodicity in @xmath21 when the smoothing time scale exceeds the typical time between particles ( i.e. , @xmath75 ) .
determinants similar to the generating function @xmath74 also appear in problems related to the orthogonality catastrophe and to the fermi - edge singularity ( fes ) . in those cases ,
the counterparts of the secondary branches of @xmath74 are secondary singularities ( cusps or peaks ) in the frequency - dependent response function ( the closed - loop contribution in the fes context ) .
such secondary cusps and peaks were studied , e.g. , in the recent work @xcite ( and , in the case of a bound - state contribution , earlier in @xcite ) . according to our predictions ,
such cusps and peaks should be most visible in case of instant switching of the scattering potential , but get suppressed if the switching time of the scattering potential exceeds @xmath75 .
we should also remark that most studies of fes involve an artificial regularization of the fermi - sea contribution at energy scales smaller than @xmath73 , so that secondary singularities in the response function are neglected @xcite .
such secondary singularities ( separated from the main peak by @xmath73 ) are probably not experimentally relevant in physical metals , where @xmath73 is a large energy scale , but may be of interest in other models with fes physics where @xmath73 is , for some reason , small ( see , e.g. , ref .
@xcite for an example from spin - liquid theory ) .
finally , we also mention recent works on non - equilibrium bosonization where determinants similar to ours appear ( again , in the temporal form ) @xcite . similarly to the fes problem discussed above , these works assume a regularization of the fermi - sea contribution , which is equivalent to neglecting the secondary branches of the fisher hartwig expansion related to the bottom of the fermi sea . [ [ acknowledgment . ] ] acknowledgment. + + + + + + + + + + + + + + + + the authors are grateful to a. g. abanov , e. demler , m. v. feigelman , l. glazman , l. levitov , a. mirlin , and e. v. sukhorukov for helpful discussions .
the work of d.a.i . was supported by the swiss national foundation through the nccr qsit .
i.p.l . was supported by marie curie actions cofund program . |
according to the ability to conduct electric currents , materials can usually be classified into insulators and metals .
recently , researchers discovered a new class of materials , of which their interior can not conduct electric currents , similar to insulators , but their surfaces have conducting channels .
although the surface states can also exist in ordinary insulators , the uniqueness of this class of materials lies in the fact that the existence of surface states is related to its bulk property , which can be described by the mathematical theory , topology .
thus , these materials are dubbed `` topological insulators ( tis ) '' @xcite .
the quantum hall effect @xcite , discovered in 1980 , can be viewed as the first example of topologically non - trivial states . under a strong magnetic field ,
two - dimensional ( 2d ) electron gases form landau levels and get localized , while at the edge , there are gapless conducting modes propagating along one direction , known as the `` chiral edge state . ''
a recent progress in this field is the discovery of time - reversal ( tr ) invariant tis @xcite , of which the gapless edge / surface states are protected by tr symmetry .
different from the chiral edge states in the quantum hall effect , the edge / surface states of tr invariant tis contain two counter - propagating modes with opposite spins , thus dubbed `` helical edge / surface states . ''
two branches of helical edge / surface states are related to each other by tr symmetry , forming the kramers pairs . according to the kramers theorem
, the energy of one branch at the momentum @xmath16 must be the same as that of the other branch at @xmath17 .
therefore , at the momenta invariant under tr , such as @xmath18 , two branches of surface states are degenerate and form a dirac - cone - like energy dispersion .
tr invariant tis have been confirmed in various materials , such as hgte quantum wells @xcite , bisb @xcite , bi@xmath19se@xmath20 family of materials @xcite , etc.@xcite , by different experimental methods , including angular - resolved photon emission spectroscopy ( arpes ) @xcite , scanning tunneling microscopy ( stm ) @xcite , and transport measurements @xcite .
since the topologically non - trivial edge / surface states require the protection from tr symmetry , these systems are also regarded as symmetry - protected topological states @xcite in free - fermions systems .
based on the above discussion , we can see that degeneracy due to symmetry plays an essential role in topological states of free - fermion systems that are protected by symmetry .
it is well known that in crystals , there are different types of degeneracy , including accidental degeneracy and essential degeneracy .
the essential degeneracy arises from the presence of certain symmetries and is related to the representation theory of symmetry groups .
therefore , one may ask the following questions : ( 1 ) whether or not the concept of tr invariant tis can be generalized to other types of symmetries , leading to new classes of topological states ; and ( 2 ) if one can use representation theory to classify topological states that are protected by other types of symmetries .
it was recently realized that crystalline symmetry can indeed protect new types of topological states , which are dubbed `` topological crystalline insulators ( tcis ) '' @xcite . in particular ,
snte systems are found to be a mirror - symmetry - protected topological state , which has recently been confirmed by the arpes measurement @xcite .
various physical properties , including mass acquisition @xcite , unconventional orbital texture @xcite , spin - filtered edge states @xcite , quasi - particle interference @xcite , and interface superconductivity @xcite , have been considered . in this paper , we will systematically study different types of tcis based on the representation theory of crystalline symmetry groups . instead of directly studying three - dimensional ( 3d ) bulk system
, we consider a 3d semi - infinite system with one surface and investigate non - trivial gapless surface states which can not be gapped once the crystalline symmetry is preserved .
since a 3d semi - infinite crystal is classified by 2d space group , this approach allows us to classify different types of non - trivial surface states based on the representation theory of 2d space group . after classifying 2d topologically non - trivial surface states , we can further investigate topological property of the 3d bulk system and identify the corresponding 3d bulk topological invariants .
the details of our approach will be discussed in the following , but we would like first to illustrate our main results .
our main results are summarized in table [ tab1 ] for 17 2d space groups .
the first column of the table is the group name and the second column gives the shape of 2d surface brillouin zone ( bz ) , in which `` re '' is for rectanglar , `` h '' for hexagonal , `` rh '' for rhombic , and `` s '' for square .
different 2d surface bzs are shown in fig .
[ fig:1 ] .
the next two columns list high symmetry momentum points ( hsps ) and high symmetry momentum lines ( hsls ) , which are essential for determining topologically non - trivial surface states .
topological classifications of tcis for the spinless and spinful cases are shown in the last two columns . here
the symbols @xmath9 and @xmath21 indicate whether the space of ground states for the insulating phases are partitioned into topological sectors labeled by an integer or a @xmath21 quantity , respectively .
@xmath22 ( @xmath23 ) means that we can define @xmath24 different integer topological invariants .
we would like to emphasize that for several space groups , including @xmath3 , @xmath4 , @xmath5 , @xmath14 , @xmath7 and @xmath15 , the classification depends on which irreducible representations ( irreps ) of the states near the band gap belong to .
therefore , for these cases , we list the required condition of irreps in the bracket after the classification .
below , we will discuss our results of tcis for 17 2d space groups in several categories .
\(1 ) for the groups @xmath0 , @xmath11 and @xmath10 ( sec .
[ sec : pm ] ) , we have integer number of copies of topologically non - trivial surface states , which are characterized by mirror chern numbers ( mcns ) , belonging to a @xmath9 classification , for both the spinless and spinful cases .
two independent mcns ( @xmath8 ) can be defined for the @xmath0 group , but only one can be defined for the @xmath11 and @xmath10 groups .
\(2 ) for the groups @xmath1 and @xmath12 ( sec .
[ sec : pmm ] ) , mcns are not allowed for the spinless case .
in contrast , four ( two ) independent mcns can be defined for the spinful case of @xmath1 ( @xmath12 ) group , leading to a @xmath13 ( @xmath8 ) classification .
\(4 ) for the @xmath3 , @xmath4 and @xmath5 ( sec .
[ sec : p4 m ] ) , the classification depends on the irreps of the states near the fermi energy , and we need to illustrate the classification for each group separately . for the spinless case of the @xmath3 group , when all the states at @xmath25 and @xmath26 belong to the doublet ( or 2d ) irrep ,
we can define two integer topological invariants ( halved mirror chirality ( hmc ) ) , giving rise to @xmath8 classification .
however , if any state at @xmath25 or @xmath26 belongs to one - dimensional ( 1d ) irreps , no topological invariant can be defined . for the spinful case of the @xmath3 group ,
three topological invariants of hmc type ( @xmath27 ) can be defined , irrespective of any condition . for the @xmath4 group ,
there are three hsps @xmath25 , @xmath28 , and @xmath29 with the point group @xmath30 and three hsls with mirror symmetry .
when the states at @xmath24 hsps ( @xmath31 ) belong to doublet irreps , the topological classification is @xmath22 with @xmath24 distinct hmcs for both the spinful and spinless cases .
if there is only one hsp or no hsp with the states in doublet irreps , one mcn ( @xmath9 ) can be defined .
for the @xmath5 group , we also have three hsps , in which @xmath25 belongs to the @xmath32 group , @xmath28 to the @xmath30 group , and @xmath26 to the @xmath33 group , and also three hsls with mirror symmetry . for the spinless case , there are two doublet irreps at @xmath25 and one doublet irrep at @xmath28 .
one hmc ( @xmath9 classification ) can be defined when all the states at @xmath25 and @xmath28 belong to doublet irreps .
the @xmath33 group at @xmath26 has four 1d irreps @xmath34 and @xmath35 ( @xmath36 ) .
please refer to table .
[ tab:2 ] for the definition of @xmath34 and @xmath35 .
the characters of the mirror symmetry @xmath37 and @xmath38 are the same in irreps @xmath34 , while opposite in irreps @xmath35 .
the topological classification on @xmath25-@xmath26-@xmath28 depends on the irreps of the bands near the fermi surface at @xmath26 .
if all the states at @xmath26 belong to @xmath34 ( or @xmath35 ) , another hmc can be defined on @xmath25-@xmath26-@xmath28 . a detailed discussion about representation dependent classification is given in section .
[ sec : pmg ] . for the spinful case ,
all the states at @xmath25 and @xmath26 belong to the doublet irrep , while both doublet and singlet irreps are possible for the states at @xmath28 .
thus , three independent hmcs ( @xmath27 ) can be defined on @xmath25-@xmath28 , @xmath28-@xmath26 , and @xmath26-@xmath25 if the states at @xmath28 belong to doublet irreps . otherwise , there are two hmcs that can be defined on @xmath25-@xmath26 and @xmath25-@xmath28-@xmath26 .
\(5 ) for three non - symmorphic symmetry groups @xmath7 , @xmath14 , and @xmath15 ( sec . [ sec : pmg ] ) , the classification also depends on irreps at the hsps @xmath25 , @xmath39 , @xmath40 and @xmath26 . for the spinless case of the @xmath14 group ,
all the states at @xmath39 and @xmath40 are doubly degenerate ( doublet irreps ) .
the classification depends on 1d irreps at @xmath25 and @xmath26 .
there are four types of different 1d irreps at @xmath25 ( @xmath26 ) , denoted as @xmath34 and @xmath35 ( @xmath36 ) .
if all the states at @xmath25 ( @xmath26 ) belong to the @xmath34 or @xmath35 irreps , one integer topological invariant ( @xmath9 classification ) , dubbed halved glide chirality ( hgc ) , in analogy to hmc , can be defined along the hsl @xmath39-@xmath25-@xmath40 ( @xmath39-@xmath26-@xmath40 ) . if both the @xmath34 and @xmath35 states exist at @xmath25 and @xmath26 near the band gap , one can no longer define hgc along these momentum lines .
however , a @xmath21 topological invariant can still be defined .
for the spinful case , the classification is similar to that of the spinless case .
the only difference is that now all the states at @xmath25 and @xmath26 are doubly degenerate and it depends on the 1d irreps at @xmath39 and @xmath40 if a @xmath21 or @xmath9 topological invariant can be defined along the momentum line @xmath25-@xmath39-@xmath26 ( @xmath25-@xmath40-@xmath26 ) . for the @xmath7 group , all the states at @xmath39 and @xmath26 are doubly degenerate , and thus a hmc
can always be defined along the line @xmath39-@xmath26 ( @xmath9 classification ) .
another independent @xmath9 topological invariant can be defined along the line @xmath39-@xmath25-@xmath40-@xmath26 if all the states at @xmath25 and @xmath40 belong to either @xmath34 or @xmath35 irreps . for the spinful case of the @xmath7 group
, the classification is similar and only the role of @xmath39 , @xmath26 and @xmath40 , @xmath25 is switched . for the @xmath15 group , in the spinless case ,
when the states at @xmath25 and @xmath26 belong to doublet irreps , three independent hgcs can be defined on @xmath25-@xmath39 , @xmath39-@xmath26 and @xmath26-@xmath25 .
if only the states at @xmath26 ( @xmath25 ) belong to doublet irrep , and all the states at @xmath25 ( @xmath26 ) belong to the 1d irreps @xmath34 or @xmath35 of point group @xmath41 , two hgcs can be defined on @xmath39-@xmath26 ( @xmath39-@xmath25 ) and @xmath39-@xmath25-@xmath26 ( @xmath39-@xmath26-@xmath25 ) . if both the @xmath34 and @xmath35 states exist at @xmath25 ( @xmath26 ) , the hgc can only be defined on @xmath39-@xmath26 ( @xmath39-@xmath25 ) .
if the states at both @xmath25 and @xmath26 belong to singlet irreps , no topologically non - trivial phase can be realized . in the spinful case
, all the states are doubly degenerate at @xmath25 and @xmath26 , and thus one hgc can be defined on @xmath25-@xmath26 .
when the states at @xmath39 belong to @xmath34 ( or @xmath35 ) of the group @xmath33 , another one hgc can be defined on @xmath25-@xmath39-@xmath26 . | topological crystalline insulators define a new class of topological insulator phases with gapless surface states protected by crystalline symmetries . in this work
, we present a general theory to classify topological crystalline insulator phases based on the representation theory of space groups .
our approach is to directly identify possible nontrivial surface states in a semi - infinite system with a specific surface , of which the symmetry property can be described by 17 two - dimensional space groups .
we reproduce the existing results of topological crystalline insulators , such as mirror chern insulators in the @xmath0 or @xmath1 groups , @xmath2 topological insulators in the @xmath3 , @xmath4 and @xmath5 groups , and topological nonsymmorphic crystalline insulators in the @xmath6 and @xmath7 groups . aside from these existing results
, we also obtain the following new results : ( 1 ) there are two integer mirror chern numbers ( @xmath8 ) in the @xmath0 group but only one ( @xmath9 ) in the @xmath10 or @xmath11 group for both the spinless and spinful cases ; ( 2 ) for the @xmath1 ( @xmath12 ) groups , there is no topological classification in the spinless case but @xmath13 ( @xmath8 ) classifications in the spinful case ; ( 3 ) we show how topological crystalline insulator phase in the @xmath6 group is related to that in the @xmath0 group ; ( 4 ) we identify topological classification of the @xmath3 , @xmath4 , and @xmath5 for the spinful case ; ( 5 ) we find topological non - symmorphic crystalline insulators also existing in @xmath14 and @xmath15 groups , which exhibit new features compared to those in @xmath6 and @xmath7 groups .
we emphasize the importance of the irreducible representations for the states at some specific high - symmetry momenta in the classification of topological crystalline phases .
our theory can serve as a guide for the search of topological crystalline insulator phases in realistic materials . |
massive stars play a key role in the evolution of the galaxy .
they are the principal source of heavy elements and uv radiation . through the combination of winds , massive outflows ,
expanding hii regions and supernova explosions , they provide an important source of mixing turbulence in the interstellar medium ( ism ) .
hii regions are particularly interesting because they can trigger the formation of a new generation of stars .
processes such ` collect and collapse ' ( e.g. @xcite ) and radiative driven implosion ( rdi ) ( e.g. @xcite ) have been convincingly demonstrated that occur around hii regions leading to formation of stars ( e.g. @xcite ) .
these processes generate dense layers of gas and dust at the interface between the hii regions and their parent molecular cloud , elongated columns ( pillars ) of gas pointing towards the ionizing sources , and cometary globules of dense gas @xcite . understanding the interplay between the ionizing radiation and
the dense surrounding gas is very important to explain the origin of these peculiar structures , and hence to characterize triggered star formation .
g46.5 - 0.2 ( hereafter g46 ) is a poorly studied hii region about 8 in size located at ra @xmath6 191726 , dec @xmath6 115554 ( j2000 ) , on the border of the molecular cloud catalogued as grsmc g046.34 - 00.21 @xcite .
@xcite observed a recombination line at v@xmath7 57.2 towards this region .
@xcite detected hi absorption features up to 56.6 and based on kinematical considerations proposed for this hii region the near distance of 3.8 kpc , while @xcite , using the same techniques , favored the far distance of 7.8 kpc .
finally , @xcite resolved the kinematic distance ambiguity for g46 based on existing hi and @xmath0co sky surveys , confirming a near distance of 4 kpc , which will be adopted in what follows .
figure [ present ] shows a two - color composite image of g46 field where the _ spitzer_-irac 8 @xmath5 m emission distribution is displayed in cyan and the _ spitzer_-mipsgal at 24 @xmath5 m in red .
the emission displayed in cyan mainly corresponds to radiation originated in polycyclic aromatic hydrocarbon ( pah ) molecules .
the far - uv photons leaking from the hii region excite the pahs , which then emit in the mid - infrared .
this emission encircles the 24 @xmath5 m emission ( in red ) that arises from very small size dust grains which are heated but not destroyed by the uv photons .
the emission at 8 @xmath5 m has a horse - shoe morphology , being thicker and more intense towards the east - northeast direction and open to the southwest , where it dilutes into the surrounding medium . precisely in the southwest direction ,
about 10 away from g46 there are pillar - like features shining at 8 @xmath5 m . in this paper
we explore the interrelationship among the hii region g46 , the environmental conditions and the young stellar objects in the region , underscoring evidence of triggered stellar formation in the immediate vicinity of g46 and farther .
the j=10 data were extracted from the galactic ring survey ( grs ) .
the survey , performed by the boston university and the five college radio astronomy observatory , maps the galactic ring in the mentioned molecular line with an angular and spectral resolutions of 46 and 0.2 , respectively ( see @xcite ) .
the observations were performed in both position - switching and on - the - fly mapping modes , achieving an angular sampling of 22 .
data are presented in main beam temperature ( @xmath8 ) .
additionally we used , , c@xmath2o j=32 , and , hcn j=43 data extracted from the james clerk maxwell telescope ( jcmt ) data archive .
these observations ( proposal i.d m10ah02 , p.i .
jonathan williams ) were carried out with the jcmt in mauna kea , hawaii , using the harp - acsis instrument .
the angular and spectral resolutions are about 15 and 0.05 .
we used the reduced data ( cal .
level 2 ) , which were reduced using the standard orac - dr pipelines .
data are presented in units of corrected antenna temperature @xmath9 , which is related to the main beam temperature ( @xmath8 ) using @xmath10 . following @xcite
a value of @xmath11 was used .
the zenith opacity was between 0.04 and 0.06 for all the observed lines .
even though we extracted reduced data , in the case of c@xmath2o j=32 , , and hcn j=43 data we applied a hanning smoothing to improve the signal - to - noise ratio which altered the spectral resolution in no more than a factor of 2 , and minor polynomials were used for baseline corrections using xspec . in this work ,
archived infrared data are also used to reveal star formation in g46 .
these data come from the glimpse and mipsgal surveys .
glimpse ( galactic legacy infared mid - plane survey extraodinaire , * ? ? ?
* ) using the irac ( using the infrared array camera , * ? ? ?
* ) on board the _ spitzer space telescope _ @xcite surveyed the inner 130 degrees of the galactic plane at 3.6 , 4.5 , 5.8 , and 8.0 bands .
the 5 @xmath12 sensitivities of the four bands are 0.2 , 0.2 , 0.4 , 0.4 mjy , respectively .
in addition to images , glimpse survey performed point - source photometry .
photometric data at the
_ j _ ( 1.25 @xmath5 m ) , _ h _ ( 1.65 @xmath5 m ) , and _ ks _ ( 2.17 @xmath5 m ) bands from the two micron all sky point source catalog ( 2mass psc , * ? ? ?
* ) are provided in the glimpse catalog to build - up a 7-bands photometric system .
both the images and the catalog of glimpse are publicly available at the infrared science archive ( irsa ) , where we have retrieved cutouts in the four irac bands and point sources in a @xmath13 region centered at @xmath14 , @xmath15@xmath16 .
we restricted the extracted catalog to be a more reliable data set with the following criteria : a ) only sources with photometric errors no larger than 0.2 mag in the irac bands are taken into account ; b ) for 2mass bands , a threshold of 0.1 mag photometric error is used to forsake unreliable photometric values .
the photometric errors for irac and 2mass sources have been adopted following @xcite .
mipsgal @xcite is a complement to the glimpse legacy survey .
this survey using the mips ( multiband infrared photometer for _ spitzer _ , * ? ? ? * ) instrument on board the _ spitzer space telescope _ surveyed an area comparable to that of glimpse . the version 3.0 of mipsgal data includes mosaics only at 24 @xmath5 m with sky coverage of @xmath17 for @xmath18@xmath19 , and @xmath20 for @xmath21@xmath22 . the spatial resolution and @xmath23 sensitivity at 24 @xmath5 m are @xmath24 and 1.7 mjy . from the irsa server ,
we have extracted a cutout of a region same as that of glimpse cutouts .
we conducted point - source extraction and aperture photometry of point sources in the mips 24 image using the psf fitting capability of iraf / daophot @xcite .
the psf was determined to be about 6 by fitting the profiles of ten bright point sources in the investigated field .
the standard deviation ( @xmath12 ) of the sky was estimated to be about @xmath25 .
daofind was used to extract candidates of point sources with a threshold of 10 @xmath12 .
the final sources were determined by visual inspection using the tvmark task .
sources affected by ghosts , diffraction spikes , halos from bright sources and artifacts residing in bright extended emission were rejected .
for aperture photometry of the extracted sources , radii of the apertures and inner and outer limits of the sky annuli were selected to be 4.5 , 15 and 21 , respectively .
magnitudes of the extracted sources were determined using the magnitude zero point of 7.17 jy provided in the mips instrument handbook .
finally , we cross - matched the 24 sources with the glimpse catalog using a cone radius of 2 .
additionally , we have queried the mipsgal point source catalog given by @xcite .
a total of 309 sources are found in the region of interest . among them , 305 have been covered by our 24 @xmath5 m point source list .
the other four sources are weaker than 11.5 mag .
a comparison of the fluxes indicates that our photometry is consistent with that given in @xcite , the mean difference is smaller than 5% in magnitude .
the completeness of our catalog was estimated by counting the number of sources as a function of magnitude .
a histogram plot of source magnitudes at each band has been plotted and carefully inspected .
the sources magnitudes are exponentially distributed .
we plot the number of sources in log scale and the magnitude in linear scale , then a straight line can be fitted .
the magnitude at which a deviation emerges has been considered as the completeness limit .
we found that our catalog is complete to a magnitude of 14.0 at 3.6 @xmath5 m , 13.5 at 4.5 @xmath5 m , 12.3 at 5.8 @xmath5 m , 12.0 at 8.0 @xmath5 m , and 8.5 at 24 @xmath5 m .
for the central region with diffuse mid - infrared emission , our catalog is complete to a magnitude of 14.0 at 3.6 @xmath5 m , 13.5 at 4.5 @xmath5 m , 12.0 at 5.8 @xmath5 m , 11.5 at 8.0 @xmath5 m , and 8.0 at 24 @xmath5 m .
in what follows we separately analyze the physical properties of the hii region g46 ( sec . 3.1 ) , of the ambient molecular gas ( sect .
3.2 ) and of the young stellar objects ( ysos ) near g46 ( sec .
3.3 ) , with the aim of investigating the impact of the ionizing radiation on the surrounding medium and the likelihood of triggered star formation . a search for the exciting star of g46 in the available ob - type stars catalogues @xcite , was unsuccessful .
we can , anyway , conjecture the earliest spectral type of a probable single ionizing star from the radio continuum emission of the hii region .
the number of uv ionizing photons needed to keep an hii region ionized is given by n@xmath26 @xcite , where @xmath27 is the electron temperature in units of 10@xmath28 k , @xmath29 the distance in kpc , @xmath30 the frequency in ghz , and s@xmath31 the measured total flux density in jy . assuming a typical electron temperature of t = 10@xmath28 k
, a distance of 4 kpc , and a total flux density of 2.7 jy as measured from the magpis image at 1.4 ghz , the total amount of ionizing photons needed to keep this source ionized turns out to be about n@xmath32 . assuming errors of about 10% in both the distance and the radio continuum flux density , we conclude that the exciting star should be an o7v @xcite . in a rough attempt to identify the star(s ) candidate(s ) responsible for the ionized gas in the region we performed an optical and infrared photometric study of the point sources in the area ( see fig.[excitingstar ] ) based on the astrometric ucac3 catalog @xcite . only sources with detection in the optical b and r bands , and in the three near - infrared ( nir )
_ j , h , k _ bands extracted from the two micron all sky survey ( 2mass ) , were considered .
we found eight sources inside a circle of size 4 centered at the central coordinate of the hii region .
their locations are shown in fig.[excitingstar ] . from the optical and infrared magnitudes we constructed their spectral energy distribution ( sed ) .
we fit the available magnitudes ( optical b and r bands , the three _ jhk _ 2mass bands , and the four _ spitzer_-irac bands ) using the kurucz photospheric models @xcite included in the tool developed by @xcite to obtain the effective temperature , t@xmath33 , of each source .
the fitting tool requires the assumption of the visual extinction , a@xmath34 , and the distance .
we adopt a distance between 3.5 and 4.5 kpc . regarding the a@xmath34 , we derived it for each source from their _ ( j@xmath35h ) _ and _ ( h@xmath35k ) _ colors .
we assumed the interstellar reddening law of @xcite ( a@xmath36 /a@xmath37 = 0.282 , a@xmath38 /a@xmath37
= 0.175 , and a@xmath39 /a@xmath37 = 0.112 ) and the intrinsic colors _
( j@xmath35h)@xmath40 _ and _ ( h@xmath35k)@xmath40 _ obtained by @xcite . among the eight stars found in the region ,
only three are compatible with a massive star located at the distance of g46 .
the effective temperatures of the massive candidates stars obtained from the fitting are shown in table [ tablestar ] .
stars # 6 and # 8 have a t@xmath33 of about 30000 k and 35000 k , respectively , which agree with the temperatures of o9.5v and o7v stars , respectively @xcite .
the models predict an effective temperature of @xmath41 26000 k for source # 3 , suggesting that this star would be of a spectral type later than b0 .
the fitted seds for sources # 6 and # 8 are shown in fig.[sed - kurucz ] . for the remaining sources # 1 , # 2 , # 4 , # 5 and # 7
, the kurucz s models fail to fit with confidence their optical and nir magnitudes as a massive star at about 4 kpc .
these sources are probably less massive foreground stars not related with g46 .
thus , stars # 6 and # 8 are the most probable candidates to be the exciting stars of the hii region .
finally , the location of the star # 8 with respect to the radio continuum emission ( red in fig.[excitingstar ] ) , which looks like an incomplete ring , seems to better explain the curved morphology observed in the ionized gas distribution .
the hii region g46 is located in projection onto a border of the molecular cloud grsmc g046.34 - 00.21 catalogued by @xcite with an associated v@xmath42 of @xmath43 .
@xcite established a kinematical distance of about 4 kpc for this molecular cloud , the same as for the hii region g46 , supporting that both objects are linked .
figure[13cogrs ] ( left ) displays the j=10 integrated emission distribution in the 40 62 velocity range .
it can be seen that the hii region seems to be evolving on the northeastern edge of the cloud .
the molecular gas appears located in projection onto the diffuse emission at 8 @xmath5 m observed towards the southwest of the open border of g46 .
it is noteworthy that there is no evidence of related molecular emission towards the east - northeast edge of the hii region , where its associated pahs emission is thicker and more intense .
the numerous studies of hii regions and their surroundings usually show that the associated pahs emission is more intense towards the borders of the hii regions with presence of molecular material .
in other words , it is expected that an ionized bubble opens in the direction away from the cloud edge because the ionized gas would be less confined and could stream - out into the lower density ism , forming what is known as a `` blister - type '' hii region @xcite . in the g46 case , however , the ring opens to the southwest , facing the elongated molecular cloud where apparently the density is higher .
a possible explanation can be found from the morphology of the g46 radio continuum emission , which suggests a stalling effect of the ionized gas against the photodissociation region ( pdr ) towards the northeast ( see fig.[excitingstar ] ) .
this fact suggests that although there is no j=10 detected towards this region , the density of the photodissociated gas is enough to confine the ionized gas .
the atypical thickness of the pdr towards this edge could be due to the presence of relatively low density gas which is expected towards the edge of a molecular cloud whereby the far - uv photons would have greater penetration in the material . by the other hand
, the non detection of radio continuum emission towards the open border of g46 suggests that the ionized gas escapes from g46 and dilutes in the ism .
the presence of diffuse emission at 8 @xmath5 m towards the southwest of g46 supports this scenario . as above mentioned , about 10 towards the southwest of the hii region ( @xmath4112 pc at the assumed distance of 4 kpc ) it can be seen some pillar - like features emitting at 8 @xmath5 m , which are pointing to the ionizing source of g46 and seem to be embedded in the molecular cloud grsmc g046.34 - 00.21 .
even though it seems to be a large distance , several pillars were found as far from the ionizing source as those analyzed in this work ( e.g in the vulpecula rift ; @xcite ) .
what it is really remarkable in this case is that these pillars - like features , resembling heaps and corrugations , are well outside of the hii region , in contrary as is usually found : bubble - like structures with pillars inside or over their boundaries . taking into account that the pillars point to the g46 open border , it is suggested that they were produced by uv photons escaped from the hii region . by inspecting the j=10 towards the pillar - like features ( fig.[13cogrs]-right )
, we find that there are two molecular structures associated with them .
the easternmost ir pillar - like feature has associated molecular gas between 54.0 and 56.5 ( white contours in fig.[13cogrs]-right ) , while the western one is related to a molecular structure ranging between 56.5 and 59.0 ( yellow contours in fig.[13cogrs]-right ) .
the different velocity ranges for these molecular structures may either indicate that both features are located at slightly different positions along the line of sight or that they have a somewhat different kinematics .
it is remarkable the good agreement between the morphology of the molecular gas and the pillar - like features as seen in ir .
moreover , the molecular structures present a morphology consisting of a dense head with a less dense tail as is usually found and predicted by models and observations towards this kind of structures in the surroundings of hii regions @xcite .
however , following the work of @xcite , the appearance of these structures resembles a previous evolutive stage in the formation of pillars because it can not be appreciated a well - formed tail behind their heads . in what
follow we characterize the interaction between the ionization radiation escaping from g46 and the pillar - like features . in order to study the radiation influence over the tips of the pillar - like features and the possibilities of triggered star formation via radiative driven implosion ( rdi , e.g. @xcite , @xcite , @xcite ) we evaluate the pressure balance between the ionized gas stalling at the head of the pillars and the neutral gas of their interiors .
assuming that the ionizing photons came from an o7v - type star located at the center of g46 ( see sect.[star ] ) , we use the predicted ionizing photon flux for an o7v star from @xcite and the projected distance between the star and the pillars to roughly estimate the amount of uv photons arriving at the surface of the pillars in @xmath44 @xmath45 s@xmath46 .
this value represents an upper limit due to the fact that the projected distance between the star and the pillars is a lower limit to the actual distance between them .
our predicted ionizing photon flux is similar to those measured towards several bright - rimmed clouds @xcite , which like the pillars , are molecular clouds sculpted by the radiation leaking from hii regions , and in many cases they have the same large scale morphology as the pillars @xcite . using the above obtained @xmath47 and following @xcite we estimate an upper limit for the electron density of about 64 @xmath48 for the expected ionized boundary layer ( ibl ) at the tip of both pillar - like structures .
this is almost 3 times greater than the critical value of @xmath49 @xmath48 above which an ibl is able to develop around a cloud @xcite . to obtain this value
we consider that the pillar heads have a radius of 0.5 and assume an effective thickness of the ionized boundary layer of @xmath50 .
then , using a typical sound speed of the ionized gas of 11.4 , we obtain that the pillars tips are supporting an external pressure of p@xmath51 @xmath48 k. it is important to note that the obtained p@xmath52 represents strictly an upper limit because the predicted @xmath53 is an upper limit and due to the used @xmath54 ( see @xcite ) . on the other side , integrating the j=10 emission over the area that contain each pillar head and using the typical lte formulae to derive the and h@xmath55 column densities by assuming a t@xmath56 k and [ h@xmath55/ ] @xmath57 ( see e.g. @xcite ) , we estimate the mass of each pillar head in about 450 and 300 for the eastern and western one , respectively .
the area of each pillar head was assumed by taking into account the radius of their tips curvature , this is a circular area with radius of 0.5 for both pillar heads .
thus , assuming spherical shapes , we estimate the molecular densities in about @xmath58 and @xmath59 @xmath48 , respectively .
then , considering the velocity interval in which each pillar structure extends , we obtain the velocity dispersion @xmath60 , which is in agreement with those predicted by models of pillars formation @xcite . finally , using the obtained densities and @xmath61 we derive the internal pressure for each pillar head in p@xmath62 , and @xmath63 @xmath48 k for the eastern and western one , respectively . as @xcite state ,
these pressure values are very likely underestimated because the j=10 line underestimates the true density and this molecule may be depleted by selective photodissociation at the boundary of the clouds .
following these authors , the internal pressure is likely underestimated by no more than a factor of 15 , thus we conclude that p@xmath64 should be between @xmath65 and @xmath66 @xmath48 k. in conclusion , we obtain p@xmath67 @xmath68 p@xmath52 , which suggests that the diluted ionization front stalls at the pillar heads , probably until the effects of mass evaporation and increasing recombination within the ibl raises the ionized gas pressure to equilibrium with the interior pressure @xcite .
this result shows that it is unlikely that a shock is propagating farther into the molecular gas , discarding that the rdi mechanism is on going in the pillars interior . with the purpose of investigating the molecular material towards the open border of g46 we used higher - angular resolution data of several molecular species extracted from the jcmt database acquired towards this region ( see the rectangle in fig.[13cogrs]-left ) .
figure [ cojc ] shows the integrated emission of the , , and c@xmath2o j=32 , while fig.[hcnco+ ] displays the and hcn j=43 integrated emission .
the and j=32 emission show clumpy , elongated molecular features , with a high density filament extending towards the southwest of the mapped region . towards the open border of g46 do not appear any considerable molecular emission , suggesting the presence of a pre - existing region with scarce molecular gas or that the uv - photons have carved the molecular cloud .
in any case this can be the path followed by the uv - photons escaped from g46 to reach the farther pillar - like structures .
the c@xmath2o j=32 , and hcn j=43 emissions are concentrated in a small compact clump related to the infrared dark cloud irdc 046.424 - 0.237 @xcite and to the millimeter continuum source bgps g046.427 - 00.237 ( see the bolocam galactic plane survey v2 ; @xcite ) , suggesting that the emission lines emanate from a deeply embedded clump .
moreover , this structure is associated with the cold high - mass clump g46.43 - 0.24 @xcite in which the ( 1,1 ) and ( 2,2 ) ammonia lines were detected at v@xmath7 52.3 and 52.9 , respectively .
the detection of these molecular species , mainly the hcn j=43 line that its critical density can be between some @xmath69 and @xmath70 @xmath48 @xcite , indicates the presence of very high density gas .
taking into account that this clump is a potential site of star formation , in what follows we analyze it .
the line parameters of the observed molecular transitions towards the center of the clump are given in table[lines ] as derived from gaussian fits from the spectra shown in fig.[spectra ] .
the j=32 was fitted using three gaussians , while the parameters of the others lines were obtained from single - components gaussian fits which coincide in velocity with the main component .
the different velocity components observed in the spectrum are present almost in the whole region .
thus , we conclude that they correspond to different molecular components seen along the line of sight , reflecting the clumpiness in the region . in order to have a rough estimate of the molecular clump mass
we assume local thermodynamic equilibrium ( lte ) .
we calculate the excitation temperature from @xmath71 } \label{eq1}\ ] ] where @xmath72 ) is the @xmath1co peak temperature towards the clump center at @xmath73 , obtaining t@xmath74 18 k. we derive the and c@xmath2o optical depths @xmath75 and @xmath76 from ( e.g. @xcite ) : @xmath77 where @xmath0t@xmath78 and @xmath2t@xmath78 are the peak temperatures of the and c@xmath2o j=32 line at the center of the region , and @xmath79 is the assumed isotope abundance ratio [ /c@xmath2o ] @xcite , obtaining @xmath80 and @xmath81 , which indicate that the c@xmath2o j=32 line appears moderately optically thin .
thus , we estimate its column density from : @xmath82 with @xmath83 to obtain the molecular hydrogen column density n(h@xmath55 ) we assume an abundance ratio of [ h@xmath55/c@xmath2o ] @xmath84 @xcite . finally the mass was derived from : @xmath85 } , \label{eq5}\ ] ] where @xmath86 is the solid angle subtended by the beam size
, @xmath87 is the hydrogen mass , @xmath5 , the mean molecular weight , is assumed to be 2.8 by taking into account a relative helium abundance of 25 % , and @xmath88 is the distance .
summation was performed over all beam positions on the molecular structure observed in c@xmath2o displayed in contours in fig.[cojc ] .
the obtained mass is about 375 .
assuming an ellipsoidal volume with semi - axis of 45 , 20 , and 20 we derive a density of n @xmath89 @xmath48 . the density should be higher in the innermost region of the clump where the hcn j=43 line emanates , showing a density gradient . using the deconvolved radius of about 0.5 pc calculated from : @xmath90 where @xmath91 is the area inside the clump , we appreciate that the above obtained mass value is slightly over the mass size threshold for massive star formation in irdcs presented in @xcite , suggesting that massive ysos can be formed within this clump . using the derived and hcn j=43 parameters listed in table[lines ] we performed a non - lte study of these molecular species with the code radex , which uses the mean escape probability approximation for the radiative transfer equation @xcite . using the measured @xmath92v we ran the code to fit t@xmath93 and estimate the column densities . taking into account that @xcite measured from the ammonia lines a kinetic temperature of t@xmath94 k towards this clump
, we fix this parameter in 20 k and assume densities between 10@xmath95 and 10@xmath96 @xmath48 to obtain the values presented in table[tradex ] . from the obtained column densities we observe abundance hcn/ ratios of about 4.4 , 3.8 , and 1.2 for n@xmath97 , 10@xmath98 , and 10@xmath96 @xmath48 , respectively . in all cases
we obtain a n(hcn)/n ( ) ratio larger than unity as found towards several clumps in the irdc g48.66 - 0.22 @xcite and the active star - forming region w49a @xcite .
as these authors point out , the steady - state chemical models for molecular species in gas - phase predict hcn @xmath68 only for t@xmath99 25 k , with a density of n@xmath100 @xmath48 , which is consistent with our results .
finally , given that the scenario is very favorable to initiate the formation of new generations of stars , we analyze in this section the existence and properties of young stellar objects in all the investigated area .
young stellar objects ( ysos ) always show an excess in the infrared emission .
the level of excess in the infrared can be effectively used for discriminating ysos from field stars and distinguishing different evolutionary stages . at a considerably early evolutionary stage , protostars are mostly embedded in dust envelopes , they exhibit large excess of infrared emission and an infrared spectral index @xmath101 indicative of flat or ascending spectral energy distribution ( sed ) at wavelength longward of 2 @xmath5 m @xcite . for pre - main sequence ( pms ) stars which possess optically thick disks , the seds tend to descend and the infrared spectral indices are in the range @xmath102 .
finally `` transition disk '' sources are more evolved ysos , where the inner part of the disks have been cleared by photoevaporation of the central stars or by planet forming processes . for such ysos
it is expected to detect ir excess at wavelengths longer than 16 @xmath5 m @xcite .
these properties of ysos make photometric observations in the near- to mid - infrared plausible to discriminate them from field stars . different color - based source identification and classification schemes have been developed and verified in the practice . in this section , we identify potential ysos following the scheme proposed by @xcite .
the resulting ysos are classified into class i ( protostars , including class 0 , class i , and `` flat spectrum '' sources ) , class ii , and `` transition disk '' sources .
there are several kinds of contaminants that could be misidentified as ysos in our original sample .
extragalactic contaminations could stem from star - forming galaxies and broad - line active galactic nuclei ( agn ) , which show pah feature emission yielding very red 5.8 and 8.0 @xmath5 m colors @xcite . in our own galaxy ,
unresolved knots of shock emission and resolved pah emission are often detected in the irac bands , yielding additional contaminations .
following the techniques presented in @xcite we exclude contaminants , and then identify and categorize potential ysos as described as follows . in the first phase ,
only sources with valid detections in all four irac bands have been considered .
any source fulfilling color criteria of @xmath103-[4.5]>0.7 $ ] and @xmath104-[5.8]>0.7 $ ] is regarded as a class i yso . in the remaining pool ,
class ii sources have been picked out based on the constraints of _ i _ ) @xmath103-[4.5]-\sigma_1>0.15 $ ] , _ ii _ ) @xmath103-[5.8]-\sigma_2>0.35 $ ] , _ iii _ ) @xmath104-[8.0]-\sigma_3>0.5 $ ] , and _ iv _ ) @xmath103-[5.8]+\sigma_2\le\frac{0.14}{0.04}\times(([4.5]-[8.0]-\sigma_3)-0.5)+0.5 $ ] . here ,
@xmath105-[4.5])$ ] , @xmath106-[5.8])$ ] , and @xmath107-[8.0])$ ] are combined errors , added in quadrature . in the second phase , sources with 24 data in the remaining pool are reexamined .
sources with colors of @xmath108-[24]>2.5 $ ] or @xmath104-[24]>2.5 $ ] are classified to be `` transition disks '' . with the potential contaminants and ysos identified
above excluded , there still remain some sources with bright 24 emission . for sources lacking valid photometric data in one or more of the irac bands ,
they are picked out as additional class i type ysos once fulfilling @xmath109>7 $ ] and @xmath110-[24]>4.5 $ ] mag , where @xmath110 $ ] is the longest wavelength irac detection that we have .
the classification scheme of @xcite includes three phases , using 2mass data to identify more ysos in addition to the aforementioned methods .
however , in our case as g46 is much farther than the sources in @xcite , 2mass photometry would be severely affected by foreground interstellar extinction and the use of these near - ir data to identify ysos would induce heavy contamination from foreground field stars .
thus , we do not use the near - ir bands in our ysos search .
the above identification procedures have resulted in 106 ysos . and they are classified into 22 class i , 60 class ii , and 24 `` transition disks '' objects . a summary of the results is given in table [ tb - count ] .
we note that the color based classification scheme would lead to misidentification . a class ii yso viewed at high inclination would show features resembling that of a class i source .
an edge - on class i yso can have similar infrared color of a class 0 source .
thus , all class i , ii , and `` transition disk '' sources identified in this paper are yso candidates .
the distribution of these yso candidates on distinct color - color spaces are presented in figs .
[ fig - cc - irac ] and [ fig - cc - mips ] .
it is important to note that we also use the point sources extracted from @xcite to identify yso candidates and reached identical results .
figure [ fig : yso - dis](a ) shows the large scale spatial distribution of the 106 ysos identified in the region . among the 82 class i and
class ii ysos more than 40 ( about 50 percent ) are located in projection onto the molecular cloud grsmc g046.34 - 00.21 ( see fig .
[ 13cogrs ] ) which fills only the 20 percent of the whole survey field .
such a concentration of ysos is suggestive of active star formation taking place in the cloud . by the other hand , from the 24 `` transition disk '' found in the whole field
, no one is detected in association with the molecular gas , indicating the relatively young nature of the ysos embedded in this cloud .
interestingly all ysos associated with the molecular cloud are placed in projection between the open border of g46 and the head of the pillars . moreover , there are not young sources in the farthest part of the cloud behind the heads of these pillars ( see fig .
[ fig : yso - dis](b ) ) . from figure
[ fig : yso - dis](a ) it can be appreciated only two concentrations of young objects over the whole field , one centered at 191630 , @xmath111115100 ( j2000 ) , consisting of 15 class ii type sources and the other one centered at 191700 , @xmath111114700 ( j2000 ) mostly composed by class i type ysos ( 9 sources ) , which are located just ahead the pillars - like features .
a closeup view of the pillar - like features region is shown in figure [ fig : yso - dis](b ) .
there is a point source revealed at 24 @xmath5 m which is too faint at irac bands and not listed in the glimpse catalog .
we have marked this 24 @xmath5 m point source using a red circle in figure [ fig : yso - dis](b ) .
weak at short wavelengths , this point source would be younger than other ones and could be a class 0 candidate .
the fact that the class ii concentration is located closer to the open border of g46 than the class i group strongly suggests an age gradient in the ysos distribution . by the other hand
, the absence of young sources inside and behind the pillar - like structures could show that the hii region influence has not reached the molecular material behind the pillars , which is in agreement with the result of section [ pillars ] .
we suggest a scenario where the propagation of ionization radiation escaping from the hii region has triggered the star formation observed in the molecular cloud through rdi mechanism and then has stalled at the surface of the pillars heads .
we can not discard that a growing density of the ibl reaches equilibrium with that of the cloud and thus , the shock front will continue their propagation into the cloud @xcite .
additionally , in the region surveyed with the jcmt ( see yellow rectangle in fig.[13cogrs ] ) there are 5 yso candidates .
one of them , a class i source , lies exactly at the center of the dense molecular clump mapped with the and hcn j=43 ( see fig.[hcnco+ ] ) , which is in agreement with the molecular gas conditions studied in sec.[southclump ] .
we have analyzed the hii region g46.5 - 0.2 , its molecular environment and the young stellar objects placed in a wide region around it , searching for evidence of their physical connection and the possibility of induced star formation .
we found a rich combination of mutual influences , underscoring the important role of hii regions in favoring the formation of new stars not only on their immediate vicinity but also farther than a distance equivalent to its own radius .
we characterized that the hii region , located at about 4 kpc and with a radius of @xmath112 pc , is probably excited by an o7v star .
its shape resembles a horse - shoe , with the bright , thick border towards the east - northeast and the open portion towards the southwest side . when mapping the distribution of the molecular gas associated with g46 we found that the hii region is located close to the edge of the grsmc g046.34 - 00.21 molecular cloud and , curiously , g46 instead of opening in the direction of lower ambient density it does exactly in the opposite direction , towards the cloud .
filamentary structure in the molecular cloud is observed , particularly in the @xmath1co j=32 and @xmath0co j=32 maps . towards the open border of g46
do not appear any considerable molecular emission , suggesting the presence of a pre - existing region with scarce molecular gas or that the uv - photons have carved the molecular cloud .
besides , close to the end of the observed filamentary structures in the cloud , the infrared images reveal the existence of pillar - like structures pointing towards the hii region open border , which are associated with some narrow molecular filaments characterized using the @xmath0co j=10 line . from a pressure balance study we found that the internal pressure of the neutral gas in the pillar - like features heads are larger than the external pressure due to the ionized gas stalling at their tips , discarding that the rdi mechanism is on going .
in addition , we analyzed a compact molecular clump located over the south border of g46 , obtaining a mass of @xmath41 375 and a density of about @xmath113 @xmath48 , with a higher density towards its center , where the hcn j=43 line is detected .
taking into account that the expansion of the hii region itself , and the injection of extra energy into the molecular cloud can drive turbulence and trigger star formation we searched for candidates of ysos and classified them according to their evolutionary stage .
we identified two main concentrations of young objects over the whole region , one , closer to the g46 open border consisting of class ii type sources , and other one mostly composed by class i type ysos located just ahead the pillars - like features strongly suggesting an age gradient in the ysos distribution .
we wish to thank the anonymous referee for her / his useful comments and corrections .
, a.p . , and e.g. are members of the _ carrera del investigador cientfico _ of conicet , argentina .
g.d . acknowledges hospitality at the national astronomical observatories ( chinese academy of sciences ) in beijing .
this work was partially supported by argentinian grants awarded by conicet , anpcyt and uba ( ubacyt ) .
this work is partially supported by the china ministry of science and technology through grant of 2010dfa02710 ( the china - argentina radio telescope project ) and nsfc through grants of 110732027 , 11373009 , 11433008 , and 11403040 .
y and y. f. h are supported by the young research grant of naoc .
m = cyan and 20 cm = red ) .
the symbols represent the only eight point sources with _
jhk _ near - infrared and rb optical bands measurements found in the region ( dashed circle ) .
the diamonds ( sources # 6 and # 8) indicate the location of the most likely candidates to be the exciting stars of g46 .
the beam of the 20 cm emission from magpis is included in the top right corner.,width=279 ] m emission with contours of the j=10 emission integrated between 40 and 62 , with levels of 10 , 15 , 20 , and 25 k .
the yellow circle is the beam of the observations .
the rms noise is about 1 k . the rectangle represents the region studied with the jcmt data . right : integrated j=10 from the grs towards the pillar - like features observed in 8 @xmath5 m .
the white contours represent the emission integrated between 54.0 and 56.5 with levels of 4.5 , 5.0 , and 5.5 k , while the yellow ones represent the integration between 56.5 and 59.0 with levels of 4 and 5 k .
the beam of the molecular data is shown in the top - right corner.,title="fig:",width=302 ] m emission with contours of the j=10 emission integrated between 40 and 62 , with levels of 10 , 15 , 20 , and 25 k .
the yellow circle is the beam of the observations .
the rms noise is about 1 k . the rectangle represents the region studied with the jcmt data . right : integrated j=10 from the grs towards the pillar - like features observed in 8 @xmath5 m .
the white contours represent the emission integrated between 54.0 and 56.5 with levels of 4.5 , 5.0 , and 5.5 k , while the yellow ones represent the integration between 56.5 and 59.0 with levels of 4 and 5 k .
the beam of the molecular data is shown in the top - right corner.,title="fig:",width=302 ] o j=32 emission integrated between 40 and 60 displayed in contours over the _ spitzer_-irac 8 @xmath5 m emission .
the contours levels are 45 , 58 , and 72 k ; 13 , 20 , and 27 k ; and 3.3 , 5.0 , and 6.7 k , for the , , and c@xmath2o , respectively . the rms noises are about 2.5 , 0.8 , and 0.5 k , respectively .
the beam of the observations is shown at the bottom left corner of the surveyed region.,width=566 ] m emission with contours of the j=43 transition of the ( left ) and hcn ( right ) integrated between 40 and 65 .
the contours levels are 2.3 , 3.3 , 5.0 , and 8.3 k for the , and 1.2 , 2.0 , and 2.8 k for the hcn .
the rms noise is about 0.3 k for both emissions .
the yellow rectangle represents the same area as shown in previous figures.,width=321 ] ccccccccccc # & 2mass designation & @xmath114 & @xmath115 & @xmath116 & b & r & a@xmath34[mag ] & t@xmath33[k ] & @xmath117 & spectral type + 3&19172025 + 1156369&12.508&12.130&11.942&16.80&14.70&4 - 5&@xmath4126000&1.08&later than b0 + 6&19172053 + 1154438&12.498&12.108&11.9815&15.70&14.10&3.5 - 4.5&@xmath4130000&1.28&@xmath41o9.5v + 8&19172239 + 1155353&10.094&9.353 & 8.914&18.40&14.80&4 - 5&@xmath4135000&1.16&@xmath41[o7-o7.5]v + lccc molecular line & t@xmath93 peak & v@xmath42 & @xmath92v ( fwhm ) + & ( k ) & ( ) & ( ) + ( 32 ) & [email protected] & [email protected] & [email protected] + & [email protected] & [email protected] & [email protected] + & [email protected] & [email protected] & [email protected] + ( 32 ) & [email protected] & [email protected] & [email protected] + c@xmath2o ( 32 ) & [email protected] & [email protected] & [email protected] + hco@xmath3 ( 43 ) & [email protected] & [email protected] & [email protected] + hcn ( 43 ) & [email protected] & [email protected] & [email protected] + cccccc input n@xmath119 & n ( ) & @xmath120 & n(hcn ) & @xmath121 & @xmath122 + ( @xmath48 ) & ( @xmath45 ) & & ( @xmath45 ) & & + 10@xmath95 & 1.9 @xmath123 & 7.50 & 8.4 @xmath123 & 4.50 & 4.42 + 10@xmath98 & 1.7 @xmath124 & 0.70 & 6.5 @xmath124 & 0.37 & 3.82 + 10@xmath96 & 5.4 @xmath125 & 0.14 & 6.8 @xmath125 & 0.09 & 1.25 + lr & + in glimpse catalog & 19210 + valid in all four irac bands & 4491 + with 24 data & 338 + rejected contaminants & 883 + class i type & 22 + class ii type & 60 + transition disks & 24 + total yso candidates & 106 + | hii regions are particularly interesting because they can generate dense layers of gas and dust , elongated columns or pillars of gas pointing towards the ionizing sources , and cometary globules of dense gas , where triggered star formation can occur .
understanding the interplay between the ionizing radiation and the dense surrounding gas is very important to explain the origin of these peculiar structures , and hence to characterize triggered star formation .
g46.5 - 0.2 ( g46 ) , a poorly studied galactic hii region located at about 4 kpc , is an excellent target to perform this kind of studies . using public molecular data extracted from the galactic ring survey ( @xmath0co j=10 ) and from the james clerk maxwell telescope data archive ( @xmath1co , @xmath0co , c@xmath2o j=32 , hco@xmath3 and hcn j=43 ) , and infrared data from the glimpse and mipsgal surveys , we perform a complete study of g46 , its molecular environment and the young stellar objects placed around it .
we found that g46 , probably excited by an o7v star , is located close to the edge of the grsmc g046.34 - 00.21 molecular cloud .
it presents a horse - shoe morphology opening in direction of the cloud .
we observed a filamentary structure in the molecular gas likely related to g46 and not considerable molecular emission towards its open border .
we found that about 10@xmath4 towards the southwest of g46 there are some pillar - like features , shining at 8 @xmath5 m and pointing towards the hii region open border .
we propose that the pillar - like features were carved and sculpted by the ionizing flux from g46 .
we found several young stellar objects likely embedded in the molecular cloud grouped in two main concentrations : one , closer to the g46 open border consisting of class ii type sources , and other one mostly composed by class i type ysos located just ahead the pillars - like features , strongly suggesting an age gradient in the ysos distribution . |
m31 is the nearest galaxy with a normal bulge where there is good evidence for the existence of a supermassive black hole ( sbh ) of @xmath10 ( e.g. kormendy & bender 1999 ; kormendy 1987 ; dressler & richstone 1988 ; richstone , bower , & dressler 1990 ) .
early stratosphere ii balloon observations ( light et al . 1974 ) saw that m31 appeared to have a dense nucleus embedded on top a galaxy bulge .
nieto et al . ( 1986 ) showed that the nucleus was off centered slightly from the bulge and asymmetric , which was hinted at in light et al .
it was not until _ hst ( hubble space telescope ) _
wfpc ( wide field and planetary camera ) and foc ( faint object camera ) observations , that the nucleus was resolved into double components ( lauer et al .
1993 [ l93 ] ; king , stanford , & crane 1995 [ k95 ] ; lauer et al .
1998 [ l98 ] ) .
figure 1 shows a grey scale @xmath1-band @xmath11 image of the nucleus from _ hst _
wfpc2 in @xmath1 , which was deconvolved and kindly provided by t. lauer ( see l98 ) .
the brighter of the double peak in the @xmath1-filter is designated as p1 , and it is separated from the fainter p2 by @xmath12 , or 1.8 pc ( l93 ) . in the far - uv ,
however , k95 discover that p2 is actually more luminous than p1 .
there is a striking uv peak ( upturn ) on top of p2 which stands apart in color from both p2 and p1 . in literature ,
p2 is synonymous with the uv peak , but here i make the distinction because the peak is not centered on p2 , as shown later . in the infrared ,
the contrast between the two nucleus is not quite as strong as in the optical , and the peak seen in uv - light all but disappears , either because it is intrinsically faint or is smeared away by the psf ( corbin , oneil , & rieke 2001 ) .
dust is ruled out as being the cause for the appearance and asymmetry of the double nucleus based on color images of the optical , far - uv , and near - infrared h - k ( corbin , oneil , & rieke 2001 ; mould et al .
1989 ) , as well as a two - dimensional decomposition by k95 , who impose the assumption that both p1 and p2 components are smooth . despite the prominence of p1 in optical images , l93 and l98
show that it is the uv peak which corresponds more closely to the center of the galaxy bulge , using isophotal centroids .
they also suggest it is the location of the sbh . from rotation curves ,
kormendy & bender ( 1999 [ kb99 ] ) , bacon et al .
( 2001 [ ba01 ] ) , and statler et al . ( 1999 [ s99 ] ) find that the uv peak of p2 also lies close to the dynamical center .
the most precise measurements put the zero - velocity center at @xmath13 ( kb99 ) and @xmath14 ( ba01 ) from the uv peak of p2 , in the direction of p1 .
however , the peak of the velocity dispersion is @xmath15 from p2 , _ away _ from p1 ( kb99 , ba01 ) .
kb99 locate the sbh near the center of the uv peak .
k95 speculate that the strong uv peak may be due to a non - thermal , low - level , active galactic nucleus ( agn ) , coincident with x - ray and radio emissions , but acknowledge it is unclear whether it is truly resolved in pre - costar _
_ images . moreover , it has the energy output of only a single post asymptotic giant star ( k95 ) .
l98 find that the uv peak is resolved by _ hst _
wfpc2 , leading them to hypothesize that the bluish color may not be an agn , but instead is caused by a population of late b to early a - type stars in a cluster , possibly in a region with high metallicity abundance ( e.g. burstein et al .
1988 , and k95 ) .
forming such a cluster may be difficult , however , because of the tidal influence of a nearby @xmath16 black hole .
the agn speculation about the uv peak has not been clarified despite chandra observations .
garcia et al . ( 2000 ) discover x - ray emission in proximity to p2 with an unusual spectrum indicating possible relation to an active nucleus . the error circle , however , is 1(twice the separation between p1 and p2 ) , and
is not alone there are a number of other x - ray sources farther away around the bulge .
if the uv peak is in fact agn related , it is curious that it does not have a counterpart in the near - ir where the emissions are often seen more clearly than in the optical due to a typical rise in the agn spectrum and diminished extinction ( quillen et al .
2001 ) .
several empirical evidences indicate that the nucleus and the bulge have a different formation history . within @xmath17 of p1 ,
l98 show that the color is redder in @xmath18 ( @xmath19 ) than the surrounding bulge ( @xmath20 ) .
but in @xmath21 , p1 ( = 2.29 ) and its anti - p2 direction ( @xmath21=2.18 ) appear to be bluer than the bulge ( @xmath21=2.39 ) .
this is in contrast to ba01 who find both p1 and anti - p2 to also be redder .
nonetheless , both their findings indicate that the nucleus has a stellar population , metallicity , and perhaps origins , different from that of stars in the bulge .
silchenko , burenkov , & vlasyuk ( 1998 [ sbv98 ] ) show that the nucleus is more metal rich than the surrounding .
in the two regions , the spectral indices differ by , @xmath22 and @xmath23 .
kb99 confirm that p1 and p2 have a similar stellar population , have higher metal line strengths , and are more similar to each other than to the bulge or to any globular cluster .
sbv98 also use @xmath24 $ ] to show that stars in the nucleus are younger by roughly @xmath25 gyrs .
the compactness of the p1 and p2 peaks is particularly interesting and challenging to explain from the standpoint of galaxy formation and dynamical evolution . kb99
determine the combined mass of the double nucleus to be @xmath26 , comparable to the mass of the sbh .
that such a configuration exists at all with such close separation ( 1.7 pc ) is surprising , and a natural interpretation is that p1 is either a merger remnant or a captured star cluster .
emsellem & combes ( 1997 ) model this scenario with n - body simulations in a sbh potential and find that , although they can reproduce the geometry , the lifetime is on the order of 0.5 myr . if the cluster is massive enough to survive disruption , in most cases , dynamical friction would cause p1 to spiral in and coalesce with p2 within several dynamical times , unless somehow p1 co - rotates with the bulge ( k95 ) .
although the short life span of an accreting scenario is not in itself conclusive proof , the discovery of other similar such systems ( e.g. ngc 4486b , lauer et al .
1996 ) favors a scenario which has a much longer lifespan .
moreover , s99 find that the sinking star cluster of emsellem & combes ( 1997 ) model does not reproduce either the rotation curve or the dispersion profile .
tremaine ( 1995 ) proposes an eccentric disk model to explain the nuclear configuration which currently stands as the favorite model .
the thick eccentric disk , nearly keplerian , is composed of stars in ring orbits around a black hole , and p1 is caused by a bottleneck of stars slowing down at the turnaround radius ( apocenter ) .
this model also predicts that the velocity center is displaced by @xmath27 from p2 in the direction of p1 .
moreover , it predicts that the stars in p1 and p2 should have a similar stellar population because they belong to the same system , which might not be the case if p1 is a merger remnant . with spectroscopic data , kb99 show that the shift of the rotation center is in the right sense of the prediction , even if the amount is somewhat lower than prediction .
moreover , p1 and p2 have more similar stellar populations to each other than to the bulge in accord with the model prediction .
however , the missing element in the tremaine model is an explanation for how the ring orbits maintain alignment under precession : the original model has no self gravity .
since then , several workers have proposed enhancements . among them , salow & statler ( 2001 ) propose a semi - analytic eccentric disk with self - gravity around a black hole , where the disk is made up of a superposing set of keplerian orbits dispersed in eccentricity and orientation according to a certain distribution function .
it predicts that the line - of - sight velocity distributions of the disk near the black hole should have a distinctive double peak , which provides a further observational constraint to test the eccentric disk model . using a different formalism , with integrable models , jalai & rafiee ( 2001 )
show that the double nucleus geometry can be sustained with four general types of regular orbits in a stckel potential , even in the absence of a sbh . however , the requirement that the nuclei be cuspy , with surface density @xmath28 is a stringent requirement , and is not seen in m31 .
furthermore , it is not clear why the nucleus would be asymmetric .
bekki ( 2000 [ be00 ] ) , ba01 , and sambhus & sridhar ( 2001 [ ss01 ] ) , propose scenarios for how such an eccentric disk might be formed and sustained under precession with self - gravity . in be00
n - body simulation , an inward bound @xmath29 star cluster is completely disrupted by the tidal shear of a @xmath30 black hole , consequently forming a thick , eccentric , stellar system .
the alternative scenario by ba01 involves a circular disk , already present in the bulge , that gets excited into an eccentric disk by a natural @xmath0 perturbation .
the excitation may either be caused by a passing giant molecular cloud or globular cluster , which can induce the lopsidedness in the disk ( seen in fig .
1 ) that persists for @xmath31 years in their simulations
. however , for the non - axisymmetric waves to develop , the disk needs to be thin .
they furthered considered a model in which a black hole was shifted from the center of the potential , and another where its velocity was perturbed .
too many particles escaped in the simulations for them to be viable .
the model proposed by sambhus & sridhar ( 2001 ) extends tremaine s ( 1995 ) eccentric disk to simulate a larger disk mass ( @xmath32 ) which involves a shredding of @xmath33 globular cluster near the vicinity of the sbh .
subsequently , the cluster stars merge into a pre - existing disk of few @xmath34 , orbiting the central black hole on both prograde and retrograde , quasi - periodic , loop orbits .
the @xmath0 instability which causes the large eccentricity is induced by resonant response to the counter - rotating orbits . in this model ,
lopsidedness of the eccentric disk geometry is created in response to the presence of retrograde orbits .
all three models roughly reproduce the double nucleus geometry and the dynamics , such as the offset in the velocity dispersion profile from p2 by @xmath35 , to varying degrees of accuracy . however , one important difference between be00 and ba01 + ss01 scenarios is that the be00 model requires the stellar cluster be much less massive than ( @xmath36 ) the black hole so that it can be disrupted .
by contrast , the ba01 and ss01 models are constructed to have a much larger disk , about @xmath37 of the total central mass concentration .
these two scenarios can be directly tested if a mass estimate of the disk can be robustly measured , and the uncertainty can be quantified .
the ambiguity in the bulge decomposition is reflected in the available photometry of p1 , which can significantly differ in brightness from different studies .
furthermore , thus far it is not clear how much of p2 is part of the bulge or the eccentric disk .
in addition to the eccentric disk , the bulge structure of m31 and how it fits into the developing picture of galaxy formation are interesting on their own merits .
however , their studies have been complicated by the double nucleus . the steepness of galaxy nuclei and correlations with other structural parameters ( e.g. faber et al .
1997 ) reflect the manner by which bulges are formed .
numerical simulations show that black hole mergers can flatten galaxy cores by ejecting stars from the center ( e.g. ebisuzaki , makino , & okumura 1991 ; nakano & makino 1999 ; and milosavljevi & merritt 2001 ) .
this scenario appears promising for explaining the correlation found between large bulges and low central surface brightness . in this paper ,
i study the detailed properties of the m31 bulge by decomposing optical images to provide new structural parameters . with detailed
decomposition one can address the following issues : are there subtle structures in the bulge that are not obvious in full light ?
what are the relative contributions of the bulge , p1 , and p2 components ?
what are their shapes ?
what is the bulge profile and how sharply is it peaked ?
finally , i discuss what the new photometry of p1 and p2 reveal about the two competing scenarios that explain the formation of the eccentric disk .
although some of these questions can be addressed from other data in previous studies , this new analysis provides a unique look at the double nucleus based on a more flexible range of assumptions than foregoing studies . in the sections to follow ,
section 2 discusses the data , section 3 briefly describes the analysis algorithm used to deblend the bulge .
section 4 discusses the decomposition , followed by the environment of the bulge in section 5 .
section 6 compares the eccentric disk formation models .
conclusions follow in section 7 . throughout the discussion
i assume that the distance to m31 is @xmath38 kpc , following kb99 .
i also assume that the galactic extinction is @xmath39 ( burstein & heiles 1984 ) , which is similar to @xmath40 determined by schlegel et al .
kormendy ( 1988 ) and kb99 dynamical models also suggest that @xmath41 for the bulge stars .
this is similar to star formation models of bell & de jong ( 2001 ) , from which one can derive @xmath42 based on the @xmath18 color of the bulge .
i obtain @xmath1-band ( @xmath43 ) data ( go 5236 , westphal ) from the _ hst _ archive , as well as the deconvolved version from t. lauer .
the two sets of images are used for different purposes .
the deconvolved version has been trimmed to a smaller field of view ( fov ) of @xmath44 . while ideal for studying the nucleus , there is not enough realty for measuring the bulge profile and its shape .
therefore i also create a mosaic image with a fov of @xmath45 , with a missing echelon in the planetary camera ( pc ) quadrant .
the net exposure time is 300s , which has sufficiently high signal - to - noise ( s / n ) to measure the bulge profile .
in addition to the deconvolved and mosaic images , i create a dithered image of the pc chip ( fov = @xmath46 ) by combining four exposures of 300 second images . this will be used later for comparisons with the deconvolved image .
i use a general galaxy fitting program called galfit to do the 2-d decomposition .
detailed information on the software and how it is implemented are found in a companion paper by peng , ho , impey , & rix ( 2002 ) , but here i describe it briefly .
one of the design capabilities of galfit is to accurately decompose nearby galaxies that are highly resolved to study them closely and uncover or extract galaxy sub - structures , such as nuclear disks , bars , and double nuclei . to be highly flexible ,
galfit allows simultaneous fitting of standard function types such as srsic ( 1968 ) , gaussian , exponential , and nuker .
the program has the option to either convolve the models with the psf to simulate the seeing , or not convolve if the image has already been deconvolved .
the number of components to fit is not limited a priori .
galfit minimizes @xmath47 residuals using a down - hill gradient / parabolic expansion method called levenberg - marquardt ( press et al .
1997 ) by iteratively creating model images , convolving them with the psf , and subtracting them from the data .
even though the gradient method is not as `` smart '' as alternative simulated annealing or metropolis algorithms , it has the virtue of being fast . as an example , fitting 6 components with 41 free parameters over the entire @xmath48 pixel image , while doing convolution , takes roughly one minute per iteration on a pentium 450 mhz computer , and converges in @xmath49 iterations .
speed is desirable because the @xmath47 topology of the fit is complex ; to adequately explore it and find an optimal fit thus requires testing various model combinations and initial parameters . the merit function to minimize is the @xmath47 , or in normalized form , @xmath50 , defined as : @xmath51 where @xmath52 is the uncertainty , or weight , at each pixel ; @xmath53 is the degree of freedom ; @xmath54 and @xmath55 are the dimensions of the image in @xmath56 and @xmath57 direction .
@xmath58 is the sum of all the model components fitted .
traditionally , 2-d models use purely elliptical azimuthal shapes , but most galaxies are somewhat disky or boxy depending on whether they are more rotationally or pressure supported , respectively .
galfit can fit them by replacing the ellipses with a shape of , @xmath59 which can be rotated to any position angle ( pa ) .
the component shape is boxy if @xmath60 , disky if @xmath61 , and pure ellipse when @xmath62 .
this azimuthal function was motivated originally by athanassoula et al .
( 1990 ) to characterize the shapes of galaxy bars .
the parameter @xmath63 is the axis ratio of each component . in the radial direction ,
the 1-d srsic light profile is defined as , @xmath64}\ ] ] where @xmath65 is the effective radius of the galaxy , @xmath66 is the surface brightness at @xmath65 , @xmath67 is the power - law index , and @xmath68 is coupled to @xmath67 such that half of the total flux is within @xmath65 .
the original de vaucouleurs ( 1948 ) profile is the special case with @xmath69 and @xmath70 .
the usefulness of the srsic function is that it can fit a continuum of profiles ranging from a gaussian , to exponential , to a de vaucouleurs , by smoothly varying the exponent : @xmath71 for a gaussian , @xmath72 for an exponential disk , and @xmath69 for de vaucouleurs . in galfit ,
the flux parameter that is fitted is the total magnitude , instead of @xmath66 .
the translation between the total magnitude of a component and @xmath66 is given in peng et al .
( 2002 ) .
bulges that are modeled with the srsic profile have flat cores , but most galaxies have profile rise that extend into the resolution limit ( e.g. lauer et al .
1995 , carollo et al .
1997 , rest et al .
2001 , and ravindranath et al 2001 ) .
a function which can describe many double power - law and core type galaxy bulges in 1-d was introduced by lauer et al .
( 1995 ) , known as the nuker law : @xmath73 ^{\frac{\gamma-\beta}{\alpha}}}\ ] ] there are five adjustable profile shape parameters : @xmath74 and @xmath75 . taken to the limits of large and small radii ,
the parameter @xmath75 is the slope of the inner power law , and @xmath76 is the slope of the outer power law .
the break radius @xmath77 is the location where the profile changes slope , @xmath78 is the surface brightness at @xmath77 , and @xmath79 describes how sharply the two power laws connect .
galfit fits the surface brightness magnitude @xmath80 instead of the intensity @xmath78 .
i generalize this profile to 2-d , hence there are 5 additional parameters : @xmath81 , pa , and @xmath82 . to account for the psf seeing when fitting models to the dithered image , galfit convolves a doubly oversampled psf , created using tiny tim ( krist & hook 1997 ) , with the models . the tiny tim software can reproduce the core of wfpc2 psfs well , even if the diffraction spikes are harder to reproduce . for convolution purposes this should suffice . later in the paper
, i will compare these results with fits to the deconvolved image of l98 .
locally , the bigger of the two contributors to the root - mean - square ( rms ) fluctuation of the data image is the surface brightness fluctuations ( sbf ) of star clusters resolved by the _ hst _ , rather than the poisson noise .
unlike poisson noise where the noise - to - signal ratio decreases as @xmath83 , the ratio of the sbf noise to the galaxy background asymptotically approaches a constant value .
thus , in this regime , the derived uncertainties in the fitting parameters remains constant with increasing exposure time .
moreover , while the poisson noise correlates on scale of @xmath84 pixel in the dithered image , the sbf noise correlates on scales of a few pixels . to get realistic error bars from model
fits thus requires having a reasonable estimate of the pixel weights , @xmath52 , that account for the sbf in a steeply rising galaxy bulge .
a poisson map of the image bears an imprint of the sbf variance because the same star clusters that produce the poisson signal are responsible for the sbf .
i assume that the sbf rms can be approximated by a scaling of the poisson noise .
to do so , i compute the ratio of the total rms to the poisson noise in @xmath85 pixel boxes sampled randomly around the bulge in the dithered image
. then , multiplying this factor ( 1.9 ) to the poisson map of the image , i obtain an approximate sbf noise image to use as pixel weights .
the exact scale factor does not affect the estimated uncertainties because they are renormalized by @xmath86 ( see below ) .
this scale factor appears reasonable , and for visually good fits , @xmath87 comes out close to unity .
it is worth noting that the intrinsic galaxy sbf amplitude is higher than in the observed image because of psf convolution .
moreover , because the sbfs are spatially resolved , psf smearing correlates the sbf noise so that the pixels are not statistically independent .
however , in scaling the pixel weights as described , the psf correlation of the sbf is implicitly taken into account in the @xmath87 statistics .
to first order , the psf smearing effectively reduces the @xmath53 in the observed image by a factor @xmath88 .
this factor is @xmath89 , where @xmath90 is the intrinsic variance ( i.e. without psf filtering ) and @xmath91 the observed variance ( i.e. psf - filtered ) .
furthermore , the effective pixel weights @xmath92 have the same factor @xmath88 , which cancels out of equation [ 1 ] in calculating @xmath87 .
once galfit converges on a solution , it estimates the fitting uncertainties by using the covariance matrix of the parameters . for a more detailed discussion , see peng et al .
( 2002 ) , but here i give a brief description . since @xmath47 is a function of the fitting parameters , the surface of constant @xmath93 centered around the best fit @xmath94 can be roughly approximated as an _
n_-dimensional ellipsoid , where _
n _ is the number of free parameters in the fit .
defined as the local @xmath47 curvature with respect to all pairs of parameter combinations @xmath95 and @xmath96 , i.e. @xmath97 , the covariance matrix of the fit provides a handle on the correlation between parameters . to estimate parameter uncertainties accounting for correlation
, i consider an ellipsoid which encloses a region up to @xmath98 , which is the 68% confidence region for a single degree of freedom .
the ellipsoid semi - major axes and their lengths are calculated from the eigenvectors and eigenvalues of the covariance matrix .
i determine the uncertainty @xmath99 for parameter @xmath95 by projecting the vector sum of all the major axes vectors onto the unit coordinate axis @xmath100 . in cases
when the fit does not produce a reduced @xmath101 , i also scale the uncertainties by @xmath86 to avoid under - estimating them .
strictly speaking , while this method of estimating uncertainties is fair , it may produce uncertainties that are equal to or greater than the true uncertainties .
this over - estimate is probably not a serious concern because , ultimately , the limitation is that we do not know how well the local @xmath47 topology may be represented by an ellipse .
furthermore , the parameter values may not be normally distributed around the mean . in the discussions to follow ,
i quote @xmath87 and parameter values based on fits to dithered images rather than deconvolved images .
although the image of m31 , with a very high s / n , has been deconvolved accurately , the process introduces an additional correlation by the psf , which is hard to interpret in the presence of sbf . in comparison ,
the noise property of the unrestored image is easier to interpret .
while the signal in individual , unrestored , images are correlated by the psf from pixel to pixel , the poisson noise is _
uncorrelated_. combining dithered images introduces at most a sub - pixel correlation which is small compared to the sbf . the correlation in the signal
is accounted for in the fitting process by convolving the models with the psf .
my attempt at decomposing the double nucleus is not the first of its kind , but is motivated by previous studies of l93 , k95 , and ba01 , who perform 2-d image decomposition using iterative techniques . k95
fit the @xmath102-band image by first subtracting a smooth model of p2 from the data . then , fitting isophotes to p1 and obtaining a smooth model , they use it to get a better estimate of p2 . the process is then repeated .
ba01 decompose the wfpc2 @xmath103-band image by assuming the bulge can be fitted by three gaussians after masking out the central @xmath104 of the image . in so doing , there is an implicit assumption that the galaxy core flattens out .
the fit is then applied to @xmath1 and @xmath102-band images after proper normalization .
l93 use a nonlinear least - squares routine to deblend the nuclear region simultaneously with two components , assuming that p2 is indistinguishable from the bulge .
each component is concentric , with otherwise arbitrary position , brightness , ellipticity , and position angle .
my procedure is similar to l93 study , but with a more general set of assumptions : i relax the restriction that p1 and p2 are symmetric , and the number of components needed to fit them . the steepness of the bulge is a free parameter .
m31 , being an sb - type galaxy , has a large galactic disk component that spans @xmath105 in diameter , with a bulge that has an effective radius of @xmath106 ( kb99 ) .
although the pc chip has twice the angular resolution to resolve the double nucleus compared to the wide field chips , the combined wfpc2 mosaic , with @xmath107 fov , is better suited for measuring the large bulge . to take advantage of the capabilities of both pc and wfpc2
, i perform fitting in two separate steps .
first i fit the bulge with a nuker profile using the entire @xmath1-band mosaic image .
to do so , i mask out the echelon - shaped region outside the pc , the dust lanes in the mosaic image , as well as inner @xmath108 of the nucleus . masking the nucleus
does not seriously bias the bulge measurement because , in the second step , i take the fitted parameters for the bulge , and optimize it simultaneously along with other components for the nucleus in the pc chip . in this step
the bulge position and inner slope parameter , @xmath75 , can vary freely , but all other _ bulge _ parameters are fixed at their optimal values .
all the other _ nuclear _ components can change freely without constraints .
the results are discussed in
4.3 , and summarized in table 1 . as an independent check on the decomposition , i perform a similar fit by replacing the nuker bulge with a srsic bulge , presented in 4.4 .
i note that even though m31 has a large disk component , using 1-d decomposition , kb99 show that the extrapolation of the disk into the center , and within the wfpc2 fov , is nearly constant at @xmath109 mag / arcsec@xmath110 .
henceforth i adopt this value as a sky level in all fits .
i determine the number of components to fit iteratively by starting with the bare - minimum assumption , then increase that number to see how the fit improves @xmath47 .
l93 has already found that two components failed to fit the m31 nucleus .
therefore , my first trial starts with three , representing p1 , p2 , and a bulge . for a three component fit ,
the @xmath87 is 1.57 , with @xmath111 . in comparison , an excellent fit to the bulge in the pc image , after masking out the double nucleus , produces a significantly lower @xmath112 .
the reason for the bad fit is that p1 is asymmetric ( l93 , k95 , l98 , and ba01 ) , moreover , the uv peak is not centered on p2 ( ba01 ) .
this can be gleaned from figure 1 where the bright uv peak is displaced to slightly south - east of p2 , and p1 is not purely ellipsoidal .
to model asymmetry , each nucleus requires at least another component nearby , totaling five .
i use the srsic profiles for most components , except for the bulge . for the bulge ,
i choose from either a srsic or nuker component .
fitting five components to just the pc image and not the mosaic , allowing _ all _ parameter to optimize ( 40 - 42 parameters , depending on whether a srsic or nuker function is used ) , one can obtain a good fit with a @xmath113 .
a closer examination reveals that the dominant component , i.e. the bulge , in this trial has an axis ratio @xmath114 with uncertainty @xmath115 .
the fit might initially be deemed acceptable except for the fact that the bulge is noticeably elliptical in the mosaic image , with an axis ratio @xmath116 , and the inner nucleus is even more flattened .
therefore , at an intermediate radius , there must be a round component that dominates , and has significantly different properties than the larger scale bulge .
i can derive meaningful bulge parameters from the mosaic image as described previously in 4.1.1 , and apply it fixed just to the pc image alone , using five components . here again , the component with @xmath117 reappears at a highly significant level , at the expense of the uv peak component .
inevitably , in a realization using five components , the uv peak is poorly fitted because it makes up only 0.1% , compared to the round bulge component , which makes up as much as 16% , of the flux in the inner few arcseconds .
the round @xmath117 component can in principle be a stellar disk seen face on , but this is unlikely .
it has a half light radius @xmath118 to @xmath108 , roughly five times the double nucleus separation .
hence , from here on i group this component with that of the bulge , but designate it as the `` spherical '' bulge ( fifth component in table 1 ) to distinguish from the large scale bulge component .
i will return to a more detailed discussion of the spherical component in section 4.6 .
the best fit , therefore , requires six components , with @xmath119 : two for each nucleus , and two bulge - like components , to be discussed shortly . to fit p1 and p2
, srsic models do a fine job without having to use a more flexible nuker function .
table 1 summarizes the properties of all the components for two different trials . in the first trial ,
i fit the dominant galaxy bulge with a 2-d nuker model , and in the second , with a srsic model .
i discuss these two scenarios with more detail independently in 4.3 and 4.4 . to fit the double nucleus accurately i ultimately resort to fitting the image with six components , with 40 - 43 free parameters , which might seem bewilderingly large and ill - constrained
i emphasize however , that they may not all be physically distinct components .
in particular two components are used each for the bulge and p1 because they are not simple to parameterize .
therefore , it may be more meaningful to interpret the sum of each pair as a single component rather than each as separate entities . with a high number of components
, several issues can be raised regarding the uniqueness and validity of decompositions ( a. quillen and t. lauer , private communication ) .
first , if the nuclei are related to an eccentric disk , its light distribution may be irregular and may not be parameterizable .
so even if the decomposition may look acceptable , the model will not be physically meaningful .
another objection is that the presence of the disk may perturb the bulge enough kinematically to invalidate the assumption that the bulge is smooth at the center .
therefore , extrapolating the bulge profile into the center based on outer regions may be unreliable .
i argue that there is evidence that the bulge profile is not disrupted much by the process of the double nucleus formation : often , kinematics can reveal subtle physical attributes ( e.g. kinematically decoupled core ) that are not obvious in the starlight .
published rotation curves and velocity contours show that the nucleus has a regular solid body rotation out to @xmath120 , which becomes keplerian beyond @xmath121 ( kb99 , ba01 ) .
the kinematic curves are otherwise featureless apart from the velocity dispersion peak being offset @xmath15 from the uv peak away from p1 .
therefore , either the formation of the disk has not been sufficiently violent to disarrange the nuclear dynamics , or the perturbations have relaxed into the bulge . hence i believe the bulge can be parameterized by smooth functions down to small physical scales .
furthermore , i believe that the degree to which the double nucleus is parameterizable and regular may yield insights into the eccentric disk and formation model predictions on the morphology . being a complicated system , irregularities in the nuclei would likely show up in the residuals because the functions i fit have a high degree of symmetry .
recognizing , however , that inferences about the morphology of the double nucleus is intimately tied to the bulge , i will check how assumptions about the bulge can affect the decomposition . as a prelude to discussions to come
, i find that , throughout various trials for the bulge , the sum of the parts for p1 , p2 , and the bulge , result in fundamentally similar components , even if individual sub - components may not be unique .
for these reasons i believe the decomposition to be valid and robust . but ultimately , as with k95 i let the smoothness and the pattern of the residuals be an empirical guide to what is a reasonable solution . in their fit to a 1-d bulge profile ,
kb99 obtain the nuker parameters : @xmath122 , @xmath123 , @xmath124 , @xmath125 , @xmath126 , based on a fit to a region @xmath127 , excluding the double nucleus .
the large uncertainties are due to parameter coupling in the nuker profile . as initial parameters to my fit of the bulge ,
i use the values from kb99 but allow them to vary .
the best fit model i obtain , according to @xmath47 , has a bulge that is represented by a nuker model , with an overall @xmath119 . in all
, six components are used , and they fit the entire bulge region well , as shown in figure 2 .
figure 2b shows the net residuals ; fig .
2c shows the model + residual for p1 , and similarly for p2 in fig .
2d . the images shown in fig .
2 are fits to deconvolution restored image of l98 , but the numbers presented in table 1 , including @xmath87 are from the unrestored image for reasons in 3.2 . in comparison to kb99 , my fits to the mosaic and dithered pc images produce parameters very similar to that obtained by kb99 , to within their uncertainties : @xmath128 , @xmath129 , @xmath130 , @xmath131 , @xmath132 . my central cusp measurement ,
@xmath132 , is slightly shallower than kb99 , @xmath133 , in this more detailed fit to the double nucleus . in this model ,
the centroid of the bulge component is displaced from the uv peak , by @xmath134 , which is a direction away from the dust patches in the bulge .
such a large shift is most likely due to disturbance by dust around the nucleus ; forcing the bulge to be centered on the uv peak increases the @xmath87 by 0.003 . in light of the unknown effect by dust
, this small difference may not be statistically significant . the residual error in fig .
2b is 6% on average at the location of the worst fit , which is comparable to the sbf .
the parameters obtained using the dithered and deconvolved images agree very well to within the uncertainties .
figure 3 shows the residuals near the double nucleus divided by the rms image the calculation of which was discussed in 3.2 .
the worst fit within the 1 circle is near p1 at ( @xmath135 , 004 ) , where one pixel deviates by @xmath136rms .
indeed , even though the fit is overall quite good , there are differences between the model and the data . in figure 4
, i show the model image by summing p2 ( without the uv peak ) and the two components of p1 , together making up the eccentric disk .
the contour spacing is 0.1 magnitude and the circle at ( 0,0 ) marks the location of the uv peak . comparing fig . 4 and fig .
1 shows that although the asymmetry of p1 is well modeled , p1 in the data is slightly rounder than the model , causing an oversubtraction near ( @xmath135 , 004 ) . simultaneously fitting a 7th component to the image
does reduce the ellipticity of p1 somewhat , and reduces the @xmath87 by 0.02 .
however , as the superficial improvement does not affect the results below , i choose to present a simpler model using 6 components .
the major axis orientation of the eccentric disk shown in fig .
4 is at pa=@xmath137 . assuming the disk is thin and circular at outer isophotes , its inclination is @xmath138 to our line of sight .
the sum of the fluxes from p1+p2 is @xmath139 , or @xmath140 after correcting for galactic extinction and distance .
the mass of p1+p2 is @xmath141 . in isolation , p1 has a brightness of @xmath142 mag .
figure 5 shows the radial surface brightness profile of the components resulting from the decomposition ( shifted to a common center ) , as well as shape parameters from isophote fitting the entire wfpc2 mosaic _
v_-band image . in that figure ,
i represent the individual components used in the fit as dashed lines . because p1 is asymmetric ,
i only show the brighter of the two sub - component , p1a .
the uv peak is not plotted .
the solid data points shown with errorbars are the intrinsic profile of the bulge in the absence of the double nucleus .
exterior to @xmath143 , those points are measured from isophote fits to the raw image , but interior , they have to be extrapolated from the analytic model fits because of the double nucleus .
to do so , i create an image of the bulge by summing the nuker and spherical bulge components , then fit it with isophotes . for completeness , i also show the shape parameters of the bulge out to 100 ; the @xmath144 panel is a measure of the diskiness ( @xmath145 ) and boxiness ( @xmath146 ) of a given isophote .
the bulge is only slightly boxy , evidenced in @xmath147 and from 2-d fits ( @xmath148 ) .
the presence of the spherical bulge component is manifest in the dip in the ellipticity profile at @xmath149 .
although the bulge profile in general appears to be smooth with a gentle curvature farther out , with no apparent breaks ( see kb99 ) , in the decomposition , there appears to be a break at a radius of @xmath150 .
this kind of behavior is sometimes seen in large bulges or spheroidal systems ( e.g. lauer et al .
1995 , peng et al . 2002 ) .
i plot a 1-d nuker function having the parameters shown in the figure to guide the eye ; it is not strictly a fit that minimizes the @xmath47 .
i will discuss this break in more detail in 5 .
for the remainder of this sub - section , i test the extent to which p1 and p2 are affected by details of the bulge decomposition . to put a limit on the central cusp of the bulge i remove the uv peak model and instead fit the bulge to that location and allow all the parameters to re - optimize .
the profile of the uv peak differs significantly from that of the bulge , and the @xmath87 increases by @xmath151 to 1.31 , with @xmath152 . in this model ,
the bulge component is centered on the uv peak .
however , p2 remains significantly off - centered from the uv peak by @xmath153 .
i test to see if this is real or is caused by fitting degeneracy by forcing p2 to the uv peak position .
i refit both the deconvolved image of l98 and my own dithered image .
both fits produce residuals much worse than the asymmetric models
. in particular , in the unrestored image the region in between p1 and p2 is significantly over - subtracted . to constrain the luminosities of p1 and p2 , hence the eccentric disk itself
, i increase the nuclear slope parameter @xmath75 from the best fit value .
if @xmath75 is forced up to 0.40 , the fit significantly oversubtracts the bulge , which can be safely ruled out .
this fit puts a lower limit on the brightness of p1 to be 0.15 mag , and p2 to be 0.4 mag , fainter than the optimal fit . in their decomposition
, k95 find that p1 has a brightness of @xmath154 , and in l93 , they find 14.5 .
my numbers are significantly brighter than both by about 0.9 to 1.2 magnitude
. the differences may be in my assumptions about the true shape of p1 and p2 , as well as the relative contribution of the bulge .
l93 assume that both p1 and p2 are individually symmetric and concentric , with the bulge centered on p2 .
part of the remaining differences might also be in the criteria of smoothness .
my criterion is formally more strict than theirs .
l98 also find the blue uv peak to have @xmath155 , in good agreement to my finding of @xmath156 , which is represented by a srsic profile .
both the large uncertainty and the large boxiness , @xmath157 , are caused by the sbf . to test how different assumptions about the galaxy bulge can affect the decomposition of p1 and p2 , i redo the above fit by replacing the nuker bulge component with a srsic model .
figure 6 shows the surface brightness profile and shape parameter plots similar to figure 5 . the most prominent difference between the two is the extent of the bulge profile slope inner to a radius of @xmath158 . in terms of @xmath87 ,
this fit is degenerate with that presented in figure 5 .
another notable difference between this decomposition and the previous is the strength of the spherical bulge component , which now has a brightness of @xmath159 .
although this component is somewhat fainter than the previous fit ( @xmath160 ) , it is still a significant component in the bulge relative to p1 and p2 . in this realization ,
the light from the two components p1 and p2 adds up to @xmath161 , which is nearly identical to the nominal fit with a nuker bulge above .
p1 has a magnitude @xmath162 , again , nearly identical to before .
this decomposition illustrates the main point of these different exercises that throughout various trials , although the sub - components of p1 or p2 , and even the bulge contribution , may not be unique , their summed flux values and shapes are insensitive to model assumptions and initial parameter values . in the discussions below ,
i formally adopt the parameters from the nuker bulge decomposition because it gives a slightly better fit statistically , and is a more general fit to the bulge , whereas the srsic profile has , by definition , a flat core . in the previous two sub - sections
, i preluded the presence of an underlying spherical component embedded inside a much larger bulge , and is distinct from the double nucleus . with
the bulge accurately removed , this extra component has an axis ratio @xmath3 , an effective radius of 12 pc ( 32 ) , and close to an exponential disk profile . in comparison
, the bulge has an axis ratio of @xmath4 .
the spherical component half light radius is roughly 5 times the separation of the double nucleus , moreover , contains 16% of the flux within 2@xmath163 ( table 1 ) .
it is essential for a good fit , both qualitatively and quantitatively , as shown throughout various trials above .
the spherical bulge component is not likely to be caused by dust in the bulge .
figure 7 shows a bulge subtracted residual map of the entire wfpc2 mosaic in the _ v_-band .
superimposed on the residual map are contours from the image prior to subtraction .
the compass arrows show the major axis orientation of the large - scale galactic disk and bulge . apparently , the bulge isophotes are elliptical well into the central few arcseconds , where there is little sign of dust .
in fact , figure 8 shows a @xmath18 color map of the wfpc2 fov , revealing strong color differences in areas affected by dust .
remarkably , within @xmath164 of the center , the presence of dust actually diminishes . rather than due to dust , the existence of the spherical component most likely results from a small but significant departure from the nuker profile and shape of the bulge .
quillen , bower , & stritzinger ( 2000 ) , using _ hst _
nicmos images , discover that many galaxies , especially luminous core - types with boxy isophotes , have rounder isophotes at small radii .
peng et al .
( 2002 ) also show that a spherical component can be accurately extracted from the centers of some other galaxies .
there are two theoretical models that predict the decrease in ellipticity and boxiness .
one involves the mixing of stochastic orbits by the presence and growth of the central sbh ( norman , may , & van albada 1985 ; gerhard & binney 1985 ; and merritt & valuri 1996 ) , which steepens the central cusp .
the other involves a dissipationless merging of binary black holes that scatter stars from the center ( milosavljevic & merritt 2001 ) , thereby flattening the nuclear cusp , with a break radius that can extend well beyond the sphere of influence of the sbh . in milosavljevic & merritt ( 2001 ) model , binary black holes of similar mass
can eject stars amounting to the combined mass of the black holes . consequently , these randomized stellar orbits , after redistributing in phase space , could conceivably have produced a spherical bulge component , whose mass would then be roughly that of the coalesced black holes .
the discovery of the spherical component , in addition to the shallowness of the bulge cusp , suggest that the binary merger scenario is a more natural explanation for the bulge formation .
furthermore , the mass of the spherical component inferred from its luminosity of @xmath165 is @xmath5 a mass surprisingly similar to the sbh of @xmath6 .
it remains to be seen whether this agreement is a mere coincidence , or a confirmation that the bulge had been scoured by merging binary black holes .
if the spherical bulge component resulted from a binary black hole merger scenario , then the eccentric disk was formed more recently than the bulge .
the mixing of stellar orbits after a binary sbh merger would have erased any organized sub - structure that previously resided at the nucleus .
this hypothesis is consistent with the finding of sbv98 , who discover that the nuclear stars are significantly younger than stars in the bulge .
kb99 show a 1-d profile of the bulge , assembled piece - wise from several studies , out to 6300 .
their plot shows that the bulge has a gentle curvature which has no meaningful break radius . on the other hand , in my representations of figures 5 and 6 , the bulge appears to have a triple power law with a distinct break in the bulge profile near 2 which has not been seen before because of the complications caused by the double nucleus below @xmath166 . from
4.3 , i obtain 2-d nuker parameters that are very similar to those obtained by kb99 , which is seen as a dashed line in figure 5 , and which indeed does have a large break radius .
however , including the spherical bulge reveals there to be a significant break at @xmath167 .
whereas before , inferring from the nuker fit alone , m31 bulge falls outside of the correlation between @xmath77 with galaxy luminosity and @xmath80 found by faber et al . ( 1997 ) , my bulge decomposition shows that the correlation now holds very well .
my new estimates on the luminosity density ( @xmath168 ) , slope ( @xmath75 ) , and the corresponding mass density ( @xmath169 ) , and slope ( @xmath170 ) values are listed in table 2 .
they , too , now fall within correlations with galaxy luminosity shown in faber et al .
( 1997 ) .
the decomposition of the mosaic @xmath1-band image ( figure 7 ) shows a beautiful dust lane that runs from the lower - left of the image towards , but stopping near , the center of the nucleus , and reappearing on the other side .
this feature has previously been noted in ground based studies of wirth , smarr , & bruno ( 1985 ) and sbv98 . in sbv98 , the spiral dust arms extend out to roughly 30 on both sides of the nucleus .
the color map of figure 8 reveals a significant gradient between the nucleus and the bulge further out , toward the south .
it also highlights the winding dust lanes towards the nucleus .
however , it is curious that despite the extensive structure , the dust arms appear to stop within @xmath164-@xmath171 of the center .
this will be discussed further in 7 .
the color difference between the nucleus and its immediate surrounding noted by l98 and ba01 is evident .
the mass of p1+p2 is instrumental for deciding between competing models that form the double nucleus , in particular between the natural @xmath0 models ( ba01 and ss01 ) , and the star cluster disruption model of be00 .
all three produce an eccentric disk as proposed by tremaine , and broadly satisfy key dynamical and morphological constraints . in be00 model ,
the key requirement is that the globular cluster gets completely disrupted and scattered into a thick disk by the sbh .
a massive cluster with strong self - gravity must approach the black hole at a low impact parameter , which translates into a disk with high ellipticity and a short life time . to satisfy the morphological and dynamical constraints ,
the cluster size is limited to @xmath172 . on the other hand
, the ba01 and ss01 models can be considerably more massive , in the range 0.7-@xmath173 .
my nominal integrated brightness for p1+p2 is @xmath139 which translates to a mass @xmath174 . in comparison to globular clusters in m31 (
barmby , huchra , & brodie 2001 ) , p1 alone ( @xmath175 , @xmath176 ) would at least be in the 90 percentile of the most luminous globular clusters .
table 1 lists two sets of masses for apertures @xmath7 and @xmath149 centered on the uv peak , for all the components .
i find that within @xmath7 of the uv peak , the masses of the components are @xmath177 , @xmath178 , and @xmath179 .
the sum p1+p2 is comparable to the black hole mass of @xmath6 . for convenience ,
i also provide masses within @xmath104 aperture of the uv peak in table 1 .
the total mass of p1+p2 determined here is about an order of magnitude too high for the globular cluster disruption scenario to work .
i note that p1 alone has sufficient mass to rule out this scenario , if p2 is at least in part considered as the bulge .
the eccentric disk masses determined by l93 and k95 can rule out the globular cluster disruption scenario , as well .
ba01 also conclude the disk is massive , @xmath180 , based on modeling the bulge with multiple gaussians .
my mass determination for the eccentric disk , being higher than the first two studies , provides additional buffer to rule out reasonable uncertainties in the m / l ratio by factors of a few . on the other hand ,
my p1+p2 flux is somewhat fainter ( by 0.57 mag ) than the measurement of @xmath181 by kb99 .
all estimates of the disk mass are that it is massive , while section 4 has shown that the uncertainty due to the bulge decomposition amounts to about 6% .
therefore , the formation model favored is that of ba01 or ss01 . in their models ,
an initial circular disk of stars was formed in the bulge , but became more eccentric after a brief encounter with a giant molecular gas cloud or a globular cluster , which need not be disrupted .
the discovery of a spherical bulge component confirms that the bulge potential is spherical around the double nucleus , which might be important for sustaining a thin eccentric disk . in a triaxial potential ,
an uniformly precessing eccentric disk , held together by self - gravity , is subject to tidal torque exerted by the bulge , which can cause misalignments in the orbital configuration , or cause the orbits to diffuse into the bulge through phase mixing .
i decompose m31 with galfit to accurately extract the double nucleus .
the large scale bulge is made up of two components : a small , spherical structure , embedded at the center of a large , and moderately elliptical component .
i then use the decomposition to study the structural parameters of the bulge , finding that the break radius , @xmath77 , correlates with other galaxy parameters that are found in faber et al .
( 1997 ) for a large sample of early - type galaxies .
the spherical component extracted has a mass of @xmath182 , which is surprisingly similar to the mass of m31 s supermassive black hole .
it remains to be seen whether this is the result of binary black hole mergers , as predicted by milosavljevic & merritt ( 2001 ) n - body simulations , or a mere coincidence .
the inferred mass of p1 and p2 combined is comparable to the black hole mass , and the bulge mass contained in a small region out to the radius of @xmath121 from uv peak , with relative masses @xmath183 .
the inferred mass of p1+p2 is insensitive to the bulge parameters , as others ( e.g. ba01 ) had also found .
coupled with the large impact parameter suggested by the disk , it seems a cluster that resulted in the eccentric disk of p1+p2 ( @xmath184 ) could not have been disrupted enough by the black hole ( @xmath185 ) to form an eccentric disk around it .
in such a scenario , the progenitor of p1+p2 would have to be still larger than my mass estimate .
the large mass of the disk is consistent with a scenario in which a stellar disk was formed in the nucleus , then became more eccentric after an @xmath0 perturbation , by , for example , passing giant molecular clouds or globular clusters ( ba01 ) .
another possibility is that an approaching globular cluster of mass @xmath186 might have come within , and was shredded by , the sbh , then was integrated into the pre - existing disk ( ss01 ) . despite how well the current n - body simulations reproduce the kinematics and morphologies , given that all the components within @xmath104 of the bulge are comparable in mass , a perturbation that would excite the @xmath0 mode in the disk might also perturb the bulge .
thus it is unclear whether their mutual tidal interactions need to be further considered in n - body simulations , or might a smooth bulge potential suffice .
the n - body simulations , coupled with a hypothetical shroud of dust at the nucleus , might explain both the eccentric disk geometry as well as the slight color difference between the bulge and the nucleus .
however , it can not be the whole story : kb99 show that p1 and p2 have metal line strengths stronger than any globular cluster , hence they are unlikely to be the digested remnant of a globular cluster or an elliptical galaxy .
a more complete picture also needs to account for the differing bulge and nuclear stellar ages inferred by sbv98 .
one viable scenario as suggested by l98 is that the disk formation was a separate event that occurred well after the formation of the bulge .
this may be the case because the bulge formation event would have disrupted any pre - existing sub - structure at the center . to grow a disk to @xmath173 ,
one scenario is by the disruption of several globular clusters .
however , this possibility is remote because angular momentum conservation and scattering by the sbh would tend to produce a more spheroidal geometry .
a more likely explanation , in private communication with l. c. ho , is that a significant amount of gas and dust roughly @xmath174 had accreted into the center .
indeed , extended dust structures in figures 7 and 8 , as well as sbv98 , strongly suggest it was a possibility . through gravitational settling ,
a circular gas+dust disk enriched with reprocessed material could have grown steadily , out of which young stars would then form .
this might explain the metallicity enhancement and the younger stellar population in the nucleus compared to the bulge .
the disk might have subsequently experienced a kick from a passing gmc or a globular cluster that increased the eccentricity .
however , it is unclear at the moment how long ago or how such a massive @xmath187 disk could have settled into the center . assuming that the accretion occurred in the form of gas , the rate is limited by the near absence of non - thermal agn activity at the m31 nucleus .
perhaps a significant amount of wind from star formation near the center might quench the agn activity by blowing fuel away from the central engine ( c. d. impey , private communication ) .
the accretion + wind scenario is attractive for explaining the possible presence of a uv star cluster at the center , and for the absence of dust lanes in the immediate vicinity of the double nucleus , despite there being a large and extended dust structure mere @xmath188 away ( figs . 7 and 8) .
* acknowledgments . *
i thank peter strittmatter for financial support .
i also thank luis ho and chris impey for comments and reading the original draft ; alice quillen , and dennis zaritsky , and roelof de jong , for general discussions ; daniel eisenstein for assistance in the error propagation ; and tod lauer for providing the deconvolved images of m31 , and for discussions about the sbf and fitting of the double nucleus .
i also thank the referee for helpful and insightful comments .
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 .
lauer , t. r. , faber , s. m. , groth , e. j. , shaya , e. j. , campbell , b. , code , a. , currie , d. g. , baum , w. a. , ewald , s. p. , hester , j. j. , holtzman , j. a. , kristian , j. , light , r. m. , lynds , c. r. , oneil , & e. j. jr .
1993 , , 106 , 1436 .
[ l93 ] llrrrrrrrrrrrrrl solution 1 : & srsic & @xmath189 & @xmath190 & 0.004 & 0.002 & 19.07 & 0.08 & 0.42 & & & 0.63 & 70.4 & 0.67 & 1.30 & uv peak + nuker bulge & srsic & @xmath191 & @xmath192 & 0.24 & 0.15 & 14.17 & 0.87 & 0.98 & & & 0.66 & 54.3 & 0.05 & & p2 + & srsic & 0.31 & 0.42 & 0.26 & 0.16 & 13.88 & 1.00 & 1.31 & & & 0.62 & 64.8 & @xmath191 & & p1a + 13.10 , 12.20 mag & srsic & 0.49 & 0.37 & 0.08 & 0.05 & 15.46 & 0.37 & 0.69 & & & 0.92 & @xmath193 & 0.57 & & p1b + & srsic & 0.13 & @xmath194 & 0.11 & 0.16 & 12.76 & 3.21 & 0.83 & & & 0.97 & 56.0 & 0.14 & & spherical bulge + @xmath195 , & nuker & @xmath194 & @xmath192 & 0.29 & 0.45 & 17.76 & 66.48 & 1.10 & 1.99 & 0.17 & 0.81 & 50.1 & 0.05 & & bulge + @xmath196 & offset & & & 0.01 & 0.01 & 20.4 & & & & & & & & & mag / arcsec@xmath110 + solution 2 : & srsic & @xmath190 & @xmath190 & 0.005 & 0.002 & 18.86 & 0.09 & 0.41 & & & 0.67 & 60.7 & 1.99 & 1.30 & uv peak + srsic bulge & srsic & @xmath197 & @xmath198 & 0.27 & 0.18 & 13.98 & 0.90 & 1.00 & & & 0.69 & 55.6 & 0.01 & & p2 + & srsic & 0.33 & 0.43 & 0.28 & 0.20 & 13.83 & 0.90 & 1.20 & & & 0.66 & 64.9 & @xmath199 & & p1a + 13.10 , 12.20 mag & srsic & 0.51 & 0.38 & 0.07 & 0.04 & 15.77 & 0.36 & 0.68 & & & 0.88 & @xmath200 & 0.83 & & p1b + & srsic & 0.00 & 0.16 & 0.08 & 0.11 & 13.51 & 2.40 & 0.44 & & & 0.96 & 56.6 & 0.08 & & spherical bulge + @xmath195 , & srsic & 0.10 & @xmath201 & 0.29 & 0.47 & 4.97 & 214.71 & 2.00 & & & 0.81 & 52.0 & 0.06 & & bulge + @xmath196 & offset & & & 0.01 & 0.01 & 20.4 & & & & & & & & & mag / arcsec@xmath110 + uncertainties & & 0.02 & 0.02 & & & 0.2 & 2% & 0.03 & 0.02 & 0.02 & 0.03 & 2 & 0.02 & & + | the double nucleus geometry of m31 is currently best explained by the eccentric disk hypothesis of tremaine , but whether the eccentric disk resulted from the tidal disruption of an inbounding star cluster by a nuclear black hole , or by an @xmath0 perturbation of a native nuclear disk , remains debatable .
i perform detailed 2-d decomposition of the m31 double nucleus in the _ hubble space telescope _
@xmath1-band to study the bulge structure and to address competing formation scenarios of the eccentric disk .
i deblend the double nucleus ( p1 and p2 ) and the bulge simultaneously using five srsic and one nuker components . p1 and
p2 appear to be embedded inside an intermediate component ( @xmath2 ) that is nearly spherical ( @xmath3 ) , while the main galaxy bulge is more elliptical ( @xmath4 ) .
the spherical bulge mass ( @xmath5 ) , being coincident with the supermassive black hole mass ( @xmath6 ) , conjoined with a shallow bulge cusp , are consistent with the scenario that the bulge was scoured by spiraling binary supermassive black holes . in the 2-d decomposition , the bulge is consistent with being centered near the uv peak of p2 , but the exact position is difficult to pinpoint because of dust in the bulge .
p1 and p2 are comparable in mass . within a radius @xmath7 of p2 ,
the relative mass fraction of the nuclear components is @xmath8 , assuming the luminous components have a common mass - to - light ratio of 5.7 .
the eccentric disk as a whole ( p1+p2 ) is massive , @xmath9 , comparable to the black hole and the local bulge mass . as such
, the eccentric disk could not have been formed entirely out of stars that were stripped from an inbounding star cluster .
hence , the more favored scenario is that of a disk formed in situ by an @xmath0 perturbation , caused possibly by the passing of a giant molecular cloud , or the passing / accretion of a small globular cluster . |
the riemann surface associated to our problem is the double cover of the sphere ramified over the endpoints of the segments from @xmath0 .
this is a genus @xmath9 compact surface @xmath10 with its affine part given by the equation [ x ] w^2=_j=1 ^ 2g+2(x - e_j ) .
the surface admits the hyperelliptic involution @xmath11 with @xmath12 fixed points @xmath13 as well as an anticonformal involution ( reflection ) @xmath14 .
fixed points of the reflection @xmath15 make up the _ real ovals _ of the surface ; fixed points of another anticonformal involution @xmath16 are known as coreal ovals .
coreal ovals cover the set @xmath0 ; real ovals cover the compliment @xmath17 , in particular two points @xmath18 from the same real oval cover the infinity .
we introduce a symplectic basis in the homologies of @xmath10 as shown in the fig .
[ basis ] .
dual basis of holomorphic differentials satisfies the normalization conditions [ dunorm ] _
a_jdu_s:=_js , and generates the period matrix [ pi ] _ b_jdu_s=:_js .
one easily checks that the introduced bases of cycles and differentials behave as follows under the reflection : |ja_s = a_s ; |jb_s =- b_s ; [ du ] _ even _ and those changing sign are called _
odd_. differentials with the property ( [ du ] ) are usually called _ real _ and their periods along even / odd cycles are real / pure imaginary respectively , see @xcite for more details . in paricular , our period matrix @xmath19 is purely imaginary .
standard facts are the symmetry of this matrix and positive definiteness of the imaginary part @xcite .
( 150,55 ) ( 0,0 ) given the period matrix , we define the complex @xmath9-dimensinal torus known as a jacobian of the complex curve @xmath10 : jac(x):=^g / l ( ) , l():=^g+^g .
[ jac ] the abel - jacobi ( aj ) map embeds the curve to its jacobian [ ajmap ] u(p):=_p_1^p du l ( ) jac(x ) , du:=(du_1,du_2, ,du_g)^t .
it is convenient to represent points @xmath20 as the so called theta characteristics , i.e. couple of real g - vector columns @xmath21 : u=(+)/2 .
the points of jacobian in this notation correspond to two vectors with real entries modulo 2 .
second order points of jacobian are @xmath22 matrices with binary entries . in particular , the images of the hyperelliptic involution fixed points @xmath23 the under the aj map ( [ ajmap ] ) are as follows : [ cols="^,^,^",options="header " , ] here @xmath24 and @xmath25 are the columns of the identity matrix and the period matrix respectively .
please note that we use a nonstandard notation of theta characteristic as two column vectors written one after another . usually the transposed matrix is used .
the following absolutely convergent series is known as theta function : [ thetadef ] ( u , ) : = _ m^g ( 2i m^tu+i m^tm ) , u^g;=^t^gg ; i m > 0 .
often it is convenient to consider theta function with charachteristics , a slight modification of the latter : [ thetachardef ] ( u , ) : = _ m^g ( 2i(m+)^t(u+)+i ( m+)^t(m+ ) ) @xmath26 matrix argument @xmath19 of theta function may be omitted if it does not lead to a confusion .
vector argument @xmath27 may also be ommited which mean that we set @xmath28 , the value of the theta function is called the theta constant in this case ( note , there remains the dependence on the period matrix ) this function ( [ thetadef ] ) has the following easily checked quasi - periodicity properties with respect to the lattice @xmath29 : [ quasiperiod ] ( u+m+m ; ) = ( -im^tm-2im^tu)(u ; ) , m , m^g .
theta functions with charachteristics have similar transformation rules @xcite which may be easily derived from the above formula and we omit them for brevity . [ remtheta ] ( i ) theta function with integer characteristics @xmath30 $ ] is either even or odd depending on the parity of the inner product @xmath31
. in particular , all odd theta constants are zeroes .
\(ii ) adding matrix with even entries to integer theta characteristics can at most spoil the sign of the theta function .
hence , the binary arithmetic plays a great role in the calculus of theta functions .
theta function may be considered as a multivalued function in the jacobian or as a section of a certain line bundle .
the zero set of theta function
the theta divisor is well defined in the jacobian since the exponential factor in the right hand side of ( [ quasiperiod ] ) does not vanish .
the theta divisor is described by so called riemann vanishing theorems @xcite .
one of the important ingredients in those theorem is a vector of riemann s constants @xmath32 which depends on the choice of homology basis and the initial point in aj map . in
the above setting the vector of riemann s constants may be found by a straightforward computation @xcite or by some combinatorial argument @xcite and corresponds to a characteristic @xmath33^t.\ ] ] zeroes of theta function transferred to the surface by the aj map are described by the following let @xmath34 be a degree @xmath9 positive non - special divisor on the curve @xmath10 , then the function @xmath35 of the argument @xmath36 has exactly @xmath9 zeroes at the points of @xmath34 counting their multiplicities .
[ thr ] here and in what follows we use the standard notions of the function theory on riemann surfaces @xcite .
_ divisor _ is a finite set of points on a surface taken with integer weights : @xmath37 , @xmath38 , @xmath39 .
the _ degree of a divisor _ is the sum of all weights of participating points : @xmath40 .
_ positive divisor _ is the divisor with positive weights only .
the notion of divisor is very natural and useful in e.g. description of zeroes / poles of meromorphic functions or differentials on a surface . _
_ i__ndex of speciality @xmath41 of a positive divisor @xmath42 is the dimension of the space of holomorphic differentials vanishing at the points of @xmath42 ( counting multiplicities ) .
divisors with index @xmath43 are called _ _
s__pecial , which means that their points are not in generic positions .
two divisors are _ _
l__inearly equivalent iff their difference is a set of zeroes and poles ( latter should be taken with negative multiplicities ) of a meromorphic function .
riemann - roch theorem @xcite says that class of a positive divisor contains a unique element iff this divisor is not special .
the goal of this section is to give a closed expression for the green s function for the collection @xmath0 of aligned segments .
there exists a unique 3rd kind abelian differential @xmath44 with simple poles at two distinguished points @xmath45 of the curve @xmath10 with residues @xmath46 respectively and purely imaginary periods .
normalization conditions for this differential imply that it is real and therefore it s @xmath47 periods ( along even cycles ) vanish @xcite .
green s function for @xmath0 has the representation : [ green ] g_e(x)=|re(_q^p d)| , where the upper limit @xmath48 covers the argument @xmath49 of green s function , @xmath50 and the lower limit @xmath51 lies on the coreal oval , @xmath52
. proof .
we lift the compliment @xmath53 to the riemann surface ( [ x ] ) so that infinity is mapped to the distinguished point @xmath54 of @xmath10 , and call it the top sheet of the surface .
other lifting we call the lower sheet . the function @xmath7 on the top sheet of the surface and @xmath55 on the lower sheet make up a harmonic function with two log poles at @xmath45 .
it is a real part of some 3rd kind abelian integral which has the same normalizations as the introduced above integral @xmath56 .
it remains to check that integral of @xmath44 along any coreal oval is purely imaginary .
this holds since the distinguished differential is real .
the explicit formula for the 3rd kind abelian integrals in terms of theta functions was given yet by riemann @xcite : @xmath57(u(p)+u(\infty_+))}{\theta[\epsilon,\epsilon'](u(p)-u(\infty_+ ) ) } \right| \right|,\ ] ] where @xmath50 and @xmath58 $ ] is the integer ( odd non - singular ) theta charachteristic corresponding to half - period @xmath59 where @xmath60 is a sum of any @xmath61 different branch points : @xmath62 , @xmath63 ; @xmath64 .
proof is based in riemann s theorem on the zeros of theta functions : the divisors @xmath65 are non - special , therefore the fraction @xmath66 as a multivalued function of @xmath67 has a unique zero at @xmath68 and a unique pole at @xmath69 .
same zero and pole has the function @xmath70 , moreover both functions acquire the same factors when the argument @xmath48 goes around the cycles of the surface .
latter claim follows from the transformation rules of theta fuctions ( [ quasiperiod ] ) and riemann bilinear relations @xcite . from liouville s theorem in now follows that both functions differ by a constant nonzero factor .
the usage of theta charachteristics allows us to simplify the expression for the abelian integral .
formula ( [ cape ] ) for the capacity contains yet another ingredient along with the green s function : the independent variable @xmath49 .
the representation for the hyperelliptic projection of the surface ( [ x ] ) to the sphere in terms of variables of jacobian is well known @xcite : @xmath71(u(p))}{\prod_\pm\theta[\epsilon,\epsilon'](u(p)\pm u(\infty_+ ) ) } \frac{\theta^2[\xi,\xi'](u(\infty_+))}{\theta^2[\xi,\xi']}\ ] ] where @xmath58 $ ] is again the integer theta charachteristic corresponding to half - period @xmath59 and @xmath60 is a positive divisor of any @xmath61 different branch points with the exception of first and last ones ; @xmath72=[\epsilon,\epsilon']+[(0\dots0)^t,(1\dots1)^t ] \bmod 2 $ ] proof .
the function @xmath73 on the surface @xmath10 has double pole at @xmath74 and two poles at @xmath75 , same as the fraction @xmath76 it follows from riemann s theorem on zeroes of thetas .
the latter expression is single -valued on the surface and therefore differs by a constant factor from the projection @xmath73 .
this factor may be found if we evaluate both functions at @xmath77 .
the resulting expression may be simplified by the usage of theta charachteristics .
let us assemble the previously obtained formulas : @xmath78(u(p)\mp u(\infty_+))}{\theta[\epsilon,\epsilon'](u(p)\pm u(\infty_+))}\right|=\ ] ] [ main ] =| |^2 . where @xmath58 $ ] is the integer theta charachteristic corresponding to half - period @xmath59 with @xmath60 a positive divisor of any @xmath61 different branch points , except of the first and the last one ; @xmath72=[\epsilon,\epsilon']+[(0\dots0)^t,(1\dots1)^t ] \mod 2 $ ] .
to make this formula computationally effective one has to calculate the period matrix @xmath19 and the image of infinity in the jacobian @xmath79 .
however , there is a way to avoid usage of quadrature rules
see e.g. techniques developed in @xcite .
the latter approach is in particular useful in the cases close to degenerateones when the segments of @xmath0 tend to vanish or merge .
for the verification of the above formula ( [ main ] ) we need the good stock of the sets @xmath0 with analytically known capacities .
we take the inverse polynomial images of segments as such sets .
indeed , let @xmath80 $ ] , with @xmath81 being a degree @xmath82 polynomial .
then the green s function has a simple appearance @xmath83 wherefrom it immediately follows that @xmath84 .
consider chebyshev polynomials @xmath85 of even degrees @xmath86 .
we compute the capacity of the sets @xmath87=[-1 ; \cos\frac{\pi}{4n } ] \cup \bigcup_{s=1}^{n-1 } [ \cos\frac{(4s-1)\pi}{4n } ; \cos\frac{(4s+1)\pi}{4n } ] \cup [ \cos\frac{(4n-1)\pi}{4n } ; 1].\ ] ] by formula ( [ main ] ) and compare it to the analytic value [ capen ] cap(e_n)=2 ^ -1/n .
below we provide the following data for each @xmath88 : period matrix of the associated curve , image of infinity under the abel - jacobi map ( up to the apparent accuracy ) , capacity @xmath89 calculated by formula ( [ main ] ) and its comparison with the value ( [ capen ] ) . | * in this note we present a universal formula in terms of theta functions for the log- capacity of several segments on a line . the case of two segments was studied by n.i.akhiezer ( 1930 ) ; three segments were considered by a.sebbar and t.falliero ( 2001 ) . * for physicist
the capacity is a coefficient relating the charge and the voltage of a condenser , or equivalently the energy of an electric field inside the condenser under the unit voltage . for mathematician
the capacity is a certain property of a set which shows how massive is it .
this concept turned out to be very usefull in approximation theory , geometric function theory @xcite , partial differential equations , potential theory to name a few . for a compact set in the complex plane its capacity coinsides with the chebyshev constant , the transfinite diameter , conformal radius ( for the simply connected sets ) and a simple formula links it to the robin constant of the set .
let @xmath0 be a collection of @xmath1 segments on a real line : [ edef ] e:=_j=0^g [ e_2j+1,e_2j+2 ] with strictly increasing sequence of endpoints @xmath2 .
we know that the capacity of a set is independent of its translations and homogenious with respect to dilations , so without loss of generality we set the most left point of @xmath0 being @xmath3 and the most right point being @xmath4 : @xmath5 the capacity @xmath6 is defined in terms of the asymptotic of the green s function @xmath7 of this set at infinity : [ cape ] g_e(x)=|x| -c + o(1 ) , x. we are going to deduce a closed formula for @xmath8 in terms of riemann s theta functions so we have to introduce some necessary ( but standard ) constructions related to riemann surfaces . |
the hamiltonian describing the free motion of fermions in bilayer graphene in the presence of magnetic field reads as @xmath184 @xmath185 is the single particle zeeman coupling of electrons spin with the magnetic field , set perpendicular to the bilayer graphene plane .
the orbital effect of the magnetic field is captured via minimal substitution @xmath26 , with @xmath186 and strength of the magnetic field reads as @xmath25 .
the eight component fermionic field is defined as @xmath187^\top$ ] , where @xmath188,\ ] ] and @xmath189 are the projections of electrons spin along the z - direction .
this representation is spin rotationally invariant and therefore our formalism can be extended easily even when the field is tilted .
a generic four fermion interactions in bilayer graphene is described by the interacting hamiltonian@xcite @xmath190 + \frac{v_{2k}}{2a_{uc } } \bigg [ \left(\psi^{\dagger}\mathcal{m}^{(b)}_j\psi\right ) \left(\psi^{\dagger}{\mathcal{m}^{(b)}_j}^{t}\psi\right ) \nonumber \\ & + & \left(\psi^{\dagger}{\mathcal{m}^{(b)}_j}^t\psi\right ) \left(\psi^{\dagger}\mathcal{m}^{(b)}_j\psi\right)\bigg ] \bigg\}. \end{aligned}\ ] ] @xmath91 and @xmath92 respectively corresponds to the strength of forward and back scattering interactions .
onsite hubbard model is also described by @xmath191 , with a constraint @xmath176.@xcite various matrices appearing in @xmath191 are defined as @xmath192 to perform the variational mean field calculation , i add and subtract the layer anti - ferromagnet ( laf ) and the ferromagnet ( fm ) order parameters ( source terms ) , @xmath193 \psi(r ) \quad + m \int d^2 r \ :
\psi^\dagger(r ) \left [ \sigma_3 \otimes i_4 \right ] \psi(r)=\int d^2 r \ :
\psi^\dagger(r ) \hat{h}_{op } \psi(r).\ ] ] for now , i keep the orientation of the anti - ferromagnet order parameter ( @xmath16 ) arbitrary , but restrict the ferromagnet order parameter ( @xmath58 ) only along the applied magnetic field .
next i compute the energy spectrum of the auxiliary hamiltonian @xmath194 to diagonalize the auxiliary hamiltonian , @xmath195 , it is worth noticing that two valleys remain decoupled , even in the presence of layer anti - ferromagnet and ferromagnet orders .
one can therefore , bring @xmath195 in block diagonal .
it can be achieved by exchanging the 2nd and the 3rd @xmath196 block of @xmath195 , yielding @xmath197 , where @xmath198 however , both @xmath199 are unitarily equivalent to a generic hamiltonian , @xmath200 explicitly , @xmath201 where @xmath202 and @xmath203 with @xmath204 , but @xmath205 . in a similar way one can also diagonalize the effective single - particle hamiltonian for single layer graphene when a nel order develops at the charge - neutrality point in the presence of magnetic fields , originally shown in ref . . however , the structure of two matrices @xmath206 and @xmath207 are slightly different for monolayer and bilayer graphene .
energy spectrum can then be immediately computed yielding a set of landau levels at @xmath38 , where @xmath208^{1/2 } + \sigma \left ( \lambda + m \right ) \right)^2 \right]^{1/2},\ ] ] with degeneracy per unit area @xmath43 for @xmath69 , and @xmath41 for @xmath44 , where @xmath64 is the magnetic length , @xmath209 and @xmath28 is the cyclotron frequency .
next i wish to find the orientation of the anti - ferromagnet order that minimizes the ground state energy of filled fermi sea . at half - filling all the states at negative energies
are completely filled , while those at positive energies are completely empty .
therefore , the hartree - fock ground state energy of the single particle auxiliary hamiltonian reads as @xmath49= - \ ; \frac{1}{2 \pi l^2_b } \sum_{\sigma=\pm } \left ( e_{0 \sigma } + e_{1 \sigma } + 2 \sum_{n \geq 2 } e_{n \sigma } \right).\ ] ] to find the configuration of @xmath16 to minimize the hartree - fock ground state energy , i choose @xmath210 and @xmath211 as independent variables .
then the energy minimization condition @xmath212}{\partial n_3 } = 0 \ : \rightarrow \ : \sum_{\sigma=\pm } \sigma \left ( \lambda + m\right ) \left [ \sum_{n=0,1}\frac{1}{e_{n \sigma } } + \sum_{n \geq 2 } \frac{2 n_3}{e_{n \sigma } \sqrt{n^2_3 + n(n-1 ) \omega^2_c } } \right ] = 0.\ ] ] the left hand side of this equation is a negative definite function of @xmath213 and vanishes only for @xmath214 .
therefore , in the presence of the zeeman coupling the anti - ferromagnet order is projected in a plane perpendicular to the direction of the magnetic field , the spin - easy - plane .
one can as well check that such configuration corresponds to the minima of the energy .
from now i , therefore set @xmath214 , and @xmath215 , for notational simplicity .
the landau level wave functions for @xmath44 , however localized near @xmath216 valley at energy @xmath217 reads as @xmath218 , \quad \bigg| - e_0 \bigg\rangle = \left [ \begin{array}{c } -a_0 \chi_n \\ 0 \\ b_0 \chi_n \\ 0 \end{array}\right],\ ] ] while those residing near the valley at @xmath219 assume the form @xmath220 , \quad \bigg| - e_0 \bigg\rangle = \left [ \begin{array}{c } 0 \\ a_0 \chi_n \\ 0 \\ b_0 \chi_n \\ \end{array}\right].\ ] ] with @xmath221 as total magnetization , i have @xmath222 the wave - function of the landau levels at energies @xmath223 for @xmath69 , localized in the vicinity of @xmath224 valley are @xmath225 , \ : \ : \bigg| - e_{n,- } \bigg\rangle = \left [ \begin{array}{c } - b_n \chi_n \\ b_n \chi_{n-2 } \\ a_n \chi_n \\ a_n \chi_{n-2 } \\ \end{array}\right ] , \ : \ : \bigg| + e_{n,+ } \bigg\rangle = \left [ \begin{array}{c } -c_n \chi_n \\ - c_n \chi_{n-2 } \\
-d_n \chi_n \\ d_n \chi_{n-2 } \end{array}\right ] , \ : \ :
\bigg| - e_{n,+ } \bigg\rangle = \left [ \begin{array}{c } d_n \chi_n \\ d_n \chi_{n-2 } \\
-c_n \chi_n \\ c_n \chi_{n-2 }
\\ \end{array}\right],\ ] ] where @xmath226 and @xmath227 .
the wave functions of the landau levels at @xmath223 for @xmath69 in the vicinity of @xmath228 are otherwise identical .
next i evaluate the variation ground state energy of the total hamiltonian , @xmath229 where @xmath230 , in the presence of quantizing magnetic field , which quenches the spectrum of the quasi - particle into a set of landau levels , obtained from the diagonalization of the auxiliary hamiltonian @xmath231 .
it is then worth to rewrite the fermionic field as @xmath232,\ ] ] where @xmath76 is the wavenumber , @xmath77 is the landau level index , @xmath78 is the valley index , and @xmath79 is the energies of the landau levels with finite zeeman coupling and @xmath189 shown in eq .
( [ spectrum ] ) , after setting @xmath233 . in the above expression the term with annihilation operator @xmath234
gets summed over all the empty states at positive energies , while the other one with creation operator @xmath235 gets summed over all the filled states at negative energies . in the presence of the zeeman coupling and layer anti - ferromagnet order ,
spin is no longer a good quantum number .
hence , instead of electrons spin , i identify a new _ effective _ quantum number @xmath79 , the energies of the landau levels of @xmath195 to complete the landau level basis .
the auxiliary hamiltonian , @xmath195 in the terms of the landau level creation and annihilation operators reads as @xmath236 the ground state @xmath237 is chosen such that both @xmath234 and @xmath238 annihilates @xmath237 .
the ground state expectation value of @xmath195 is @xmath239 where @xmath240 .
next i compute the ground state energy of the interacting part of the total hamiltonian .
let us start our discussion with the quadratic piece of the interacting hamiltonian @xmath241 .
the ground state expectation value of @xmath242 reads as @xmath243 + n^2 \bigg [ \frac{1}{e_{n,+ } } + \frac{1}{e_{n,- } } + \frac{2}{e_{0 } } \bigg ] \bigg\ } \times \ ; 2 d.\end{aligned}\ ] ] finally i compute the ground state expectation value of the quartic interaction @xmath191 , which reads as @xmath244 \left[\psi^\dagger(r ) { \cal o}_2 \psi(r ) \right ] | 0 \rangle = \sum_{n<0,p>0 } \left[\psi^\dagger_{n}(r ) { \cal o}_1 \psi_{p } ( r ) \right ] \left [ \psi^\dagger_p ( r ) { \cal o}_2 \psi_n ( r ) \right ] + \sum_{n , p<0 } \left[\psi^\dagger_{n}(r ) { \cal o}_1 \psi_n ( r ) \right ] \left [ \psi^\dagger_p ( r ) { \cal o}_2 \psi_p ( r ) \right],\ ] ] where @xmath77 and @xmath245 are landau level indices . after a long and tedious calculation
, one can compactly write the ground state expectation value of the quartic interaction terms as @xmath246 \bigg[\frac{1}{4 } \bigg ( \frac{1}{e_{p,+ } } + \frac{1}{e_{p,- } } \bigg ) + \frac{1}{e_0}\bigg ] + \bigg [ 2 + \sigma \frac{e_n- m_t}{e_{n,-}}- \sigma \frac{e_n+ m_t}{e_{n,+}}\bigg ] \nonumber \\ & \times&\bigg [ \frac{1}{4 } \bigg(2- \sigma \frac{e_p- m_t}{e_{p,- } } + \sigma \frac{e_p+ m_t}{e_{p,+ } } \bigg ) + \bigg(1 + \sigma \frac{m_t}{e_0 } \bigg ) \bigg ] \bigg\ } + 4 v_0 \sum_{n , p } \bigg\ { \bigg ( 1 + \frac{m_t}{e_0}\bigg ) + \frac{1}{4 } \bigg(2- \frac{e_p - m_t}{e_{p,- } } + \frac{e_p+m_t}{e_{p,+}}\bigg ) \bigg\ } \nonumber \\ & \times & \bigg\ { \bigg ( 1- \frac{m_t}{e_0}\bigg ) + \frac{1}{4 } \bigg(2 + \frac{e_n - m_t}{e_{n,- } } - \frac{e_n+m_t}{e_{n,+}}\bigg ) \bigg\ } - n^2 \bigg[\frac{1}{e_0}+\frac{1}{4 } \bigg(\frac{1}{e_{n,+ } } + \frac{1}{e_{n,-}}\bigg ) \bigg ] \ ; \bigg[\frac{1}{e_0}+\frac{1}{4 } \bigg(\frac{1}{e_{p,+ } } + \frac{1}{e_{p,-}}\bigg ) \bigg ] \nonumber \\ & + & v_0 \sum_{n , p } \bigg\ { -\frac{n^2}{4 } \bigg[\frac{1}{e_{n,+}}+ \frac{1}{e_{n,-}}\bigg ] \bigg[\frac{1}{e_{p,+}}+ \frac{1}{e_{p,-}}\bigg ] + \frac{1}{4 } \bigg [ 2 + \frac{e_n - m_t}{e_{n,- } } -\frac{e_n+m_t}{e_{n,+ } } \bigg ] \ ; \bigg [ 2 + \frac{e_p - m_t}{e_{p,- } } -\frac{e_p+m_t}{e_{p,+ } } \bigg].\end{aligned}\ ] ] next i minimize the total variation energy @xmath247 with respect to the ferromagnet ( @xmath58 ) and the easy plane layer anti - ferromagnet order ( @xmath248 ) order parameters to arrive at the gap equations , which are @xmath249&=&0 , \\
1- \bigg(\frac{v_0+v_{2k}}{4 \pi l^2_b}\bigg ) \bigg[\sum_{n \geq 2 } \bigg ( \frac{1}{e_{n,+ } } + \frac{1}{e_{n,- } } \bigg ) + \frac{2}{e_0}\bigg]&=&0,\end{aligned}\ ] ] as shown in the main part of the paper , with @xmath250 in the first gap equation , and @xmath251 in the second one .
i next compare the the single excitation gap of the auxiliary hamiltonian @xmath195 , namely @xmath252 , with the one i obtain from variational approach . to extract the excitation gap in this approach
let us construct an excited state @xmath253 out of the ground state @xmath254 , as @xmath255 , where @xmath256 , @xmath257 contains all the four indices to complete the basis , and for example @xmath258 .
let us first compute the expectation value of the non - interacting piece of the hamiltonian the excitation energy @xmath259 in the excited state . a straight forward computation yields @xmath260 next i evaluate the expectation value of the quadratic piece of the interacting hamiltonian in the excited state @xmath261 .
this quantity depends on the index @xmath256 and @xmath257 , and is given by a simple expression @xmath262 where @xmath263 and @xmath264 are the respective the landau level wave functions with positive and negative energies , shown in sec .
i. next i compute the expectationvalue of the four - fermion interactions in the excited state . after completing some standard algebra ,
i obtain @xmath265 \left[\psi^\dagger(r ) { \cal o}_2 \psi(r ) \right ] | e \rangle - \langle 0 | \left[\psi^\dagger(r ) { \cal o}_1 \psi(r ) \right ] \left[\psi^\dagger(r ) { \cal o}_2 \psi(r ) \right ] | 0 \rangle\ ] ] @xmath266 \ ; \left [ \psi^\dagger_\eta { \cal o}_2 \psi_\alpha \right ] -\sum_{\eta>0 } \left [ \psi^\dagger_\beta { \cal o}_1 \psi_\eta \right ] \ ;
\left [ \psi^\dagger_\eta { \cal o}_2 \psi_\beta \right ] -\sum_{\eta<0 } \left [ \psi^\dagger_\eta { \cal o}_1 \psi_\alpha \right ] \ ; \left [ \psi^\dagger_\alpha { \cal o}_2 \psi_\eta \right ] + \sum_{\eta<0 } \left [ \psi^\dagger_\eta { \cal o}_1 \psi_\beta \right ] \ ; \left [ \psi^\dagger_\beta { \cal o}_2 \psi_\eta \right ] \nonumber \\ & + & \left [ \psi^\dagger_\alpha { \cal o}_1 \psi_\alpha \right ] \ ; \sum_{\eta<0 } \left [ \psi^\dagger_\eta { \cal o}_2 \psi_\eta \right ] + \left [ \psi^\dagger_\alpha { \cal o}_2 \psi_\alpha \right ] \ ; \sum_{\eta<0 } \left [ \psi^\dagger_\eta { \cal o}_1 \psi_\eta \right ] -\left [ \psi^\dagger_\beta { \cal o}_1 \psi_\beta \right ] \ ; \sum_{\eta<0 } \left [ \psi^\dagger_\eta { \cal o}_2 \psi_\eta \right ] -\left [ \psi^\dagger_\beta { \cal o}_2 \psi_\beta \right ] \ ; \sum_{\eta<0 } \left [ \psi^\dagger_\eta { \cal o}_1 \psi_\eta \right ] \nonumber \\ & - & \left [ \psi^\dagger_\alpha { \cal o}_1 \psi_\alpha \right ] \ ; \left [ \psi^\dagger_\beta { \cal o}_2 \psi_\beta \right ] -\left [ \psi^\dagger_\alpha { \cal o}_2 \psi_\alpha \right ] \ ; \left [ \psi^\dagger_\beta { \cal o}_1 \psi_\beta \right ] + \left [ \psi^\dagger_\beta { \cal o}_1 \psi_\alpha \right ] \ ; \left [ \psi^\dagger_\alpha { \cal o}_2 \psi_\beta \right ] + \left [ \psi^\dagger_\alpha { \cal o}_1 \psi_\beta \right ] \ ; \left [ \psi^\dagger_\beta { \cal o}_2 \psi_\alpha \right].\end{aligned}\ ] ] after a lengthy and tedious computation , i obtain the expression for the many - body excitation gap for the @xmath0 hall state as @xmath267 as shown in the main part of the paper .
the term proportional to @xmath268 possibly captures the effect of quantum hall ferromagnet .
however , excellent agreement with number of experiments is achieved for @xmath269 , when the layer - anti ferromagnet order is accompanied by a sizable interaction induced ferromagnetic order ( @xmath58 ) . also notice that within the framework of hubbard model at the lattice scale @xmath176 .
i here provide the detail comparison of self - consistently calculated gap for the @xmath0 hall state with various experiments@xcite . in my numerical calculation
i do not evaluate the gap at zero magnetic field , rather use it as an input to calculate the gap at finite magnetic fields .
i am extremely thankful and in debt to markus weiss and cristian sch@xmath174nenberger for providing number of data from their papers@xcite .
.i here extract the value of the observed gap in ref . substituting @xmath98 in the fitting function
proposed in ref .
@xmath270 , with @xmath271 mev , and @xmath9 mev / t . [ cols="^,^,^,^",options="header " , ]
i here provide some additional numerical results for the self - consistent solution of the laf and fm orders , when the blg is subject to a tilted magnetic field . all together , the results presented in the main part of the paper , and the ones shown in fig . 1 , may strengthen my claim that there exists no direct transition between the easy - plane laf state to a pure fm state , as one increases the the parallel component of the magnetic field , while keeping its perpendicular component fixed . 99 r. e. throckmorton , and o. vafek , phys . rev .
b * 86 * , 115447 ( 2012 ) .
j. velasco jr . , l. jing , w. bao , y. lee , p. kratz , v. aji , m. bockrath , c.n .
lau , c. varma , r. stillwell , d. smirnov , fan zhang , j. jung , a.h .
macdonald , nat .
nano . * 7 * , 156 ( 2012 ) .
f. freitag , m. weiss , r. maurand , j. trbovic , and c. sch@xmath174nenberger , solid state commun .
* 152 * , 2053 ( 2012 ) .
f. freitag , w. weiss , r. maurand , j. trbovic , and c. sch@xmath174nenberger , phys .
b * 87 * , 161402(r ) ( 2013 ) .
o. vafek , phys .
b * 82 * , 205106 ( 2010 ) .
i. f. herbut , phys . rev .
b * 76 * , 085432 ( 2007 ) . | a variational ground state for insulating bilayer graphene ( blg ) , subject to quantizing magnetic fields , is proposed . due to the zeeman coupling ,
the layer anti - ferromagnet ( laf ) order parameter in fully gapped blg gets projected onto the spin easy plane , and simultaneously a ferromagnet order , which can further be enhanced by exchange interaction , develops in the direction of the magnetic field .
the activation gap for the @xmath0 hall state then displays a crossover from quadratic to linear scaling with the magnetic field , as it gets stronger , and i obtain excellent agreement with a number of recent experiments with realistic strengths for the ferromagnetic interaction .
a component of the laf order , parallel to the external magnetic field , gives birth to additional incompressible hall states at filling @xmath1 , whereas the remote hopping in blg yields @xmath2 hall states .
evolution of the laf order in tilted magnetic fields , scaling of the gap at @xmath3 , the effect of external electric fields on various hall plateaus , and different possible hierarchies of fractional quantum hall states are highlighted .
two dimensional chiral electron gas in single and bilayer graphene respectively discerns anomalous quantization of hall conductivity at fillings @xmath4 and @xmath5 in weak magnetic fields , where @xmath6@xcite . while the valley and the spin degrees of freedom of the chiral quasi - particles stand responsible for the four fold degeneracy of the landau levels ( lls ) , the particle - hole symmetric quantization of the hall conductivity reflects the dirac or dirac - like vacuum structure in these materials@xcite .
additional twofold orbital degeneracy of the zeroth ll ( zll ) in bilayer graphene ( blg ) arises from the parabolic dispersion at low energies @xcite , yielding a constant electronic density of states at the charge - neutrality point ( cnp ) in the absence of magnetic fields , which in turn enhances the effect of electron - electron interactions .
interestingly , a number of recent experiments strongly suggesting the possibility of broken - symmetry phases in blg even without external magnetic and/or electric fields @xcite .
on the other hand , dirac fermions in monolayer graphene continue to find themselves in a robust semi - metallic phase , and ordering possibly takes place only in the presence of magnetic fields . among numerous possibilities@xcite ,
some promising candidates for the underlying ordered phases in pristine blg are gapless _ _
nematic__@xcite , and fully gapped _
layer antiferromagnet _ ( laf )
states@xcite .
while the former one breaks the threefold rotational symmetry of the honeycomb lattice , the laf order corresponds to a staggered pattern of fermion spin among the layers , which , for example , can be favored by on - site hubbard interaction @xcite .
splitting of the topologically protected zll in graphene - based systems necessarily requires the electron - electron interaction and/or zeeman coupling of electrons spin with the magnetic field .
the existence of completely filled valence band lls , which , in principle , can get renormalized due to an ordering in the vicinity of the cnp @xcite , places the quantum hall physics in carbon based layered materials in a different paradigm than that in regular non - relativistic two - dimensional electron gases@xcite .
therefore , the fully gapped states or masses " , such as layer - polarized state , corresponding to an imbalance of average electronic density between two layers @xcite , and laf in blg or nel order in monolayer graphene @xcite , optimally lowers the ground state energy by mixing non - interacting electron- and hole - like lls , and thereby pushing further down all the filled lls , placed below the chemical potential .
however , due to single - particle zeeman coupling , the laf order parameter(op ) in blg gets projected onto the easy plane , in a direction perpendicular to the applied magnetic field , and a ferromagnetic order develops in its direction , resembling in this regard the situation in monolayer graphene with nel order @xcite .
i name this ground state easy - plane laf(eplaf ) .
although various experiments have suggested the existence of insulating blg @xcite , the nature of the broken symmetry phase remained puzzling for a while . recently
a well - resolved gap ( @xmath7 ) in pristine blg has been observed , which increases _ monotonically _ with the magnetic field ( b ) , conforming to a closed form @xmath8 where @xmath9 mev / t and @xmath10 mev @xcite .
softening of this gap in a weak perpendicular electric field , and negligibly small two - terminal conductance ( @xmath11s @xmath12 ) at zero magnetic field , respectively excludes the possibility of an underlying layer - polarized state and topological quantum spin hall insulators ( qshi)/anomalous hall insulators in blg .
thus far the laf state appears to be the most promising ground state in half - filled blg .
observation of insulating behavior in blg has also been reported in ref . , and for @xmath13 samples in ref . .
the zero magnetic field gap in these samples is @xmath14 mev , which increases linearly with the field as @xmath15mev / t .
here , i address the evolution of the laf state in blg under the influence of quantizing magnetic fields , and show that with the underlying eplaf state one finds reasonably good agreement with the observed scaling of the gap in various experiments at neutral and finite fillings @xcite .
the free energy in the presence of a uniform background of the electronic density , laf ( @xmath16 ) order , and magnetization ( @xmath17 ) reads as @xcite @xmath18 . \label{totalhf}\ ] ] @xmath19 $ ] is the ground state energy per unit area of the effective single - particle hamiltonian @xmath20 where @xmath21 , with @xmath22/(2 m^\ast).\ ] ] the effective mass of the parabolic dispersion in blg is @xmath23 , where @xmath24 is the electronic mass@xcite .
the magnetic field @xmath25 is set to be perpendicular to the blg plane , so is magnetization , and @xmath26 .
the zeeman coupling in blg reads as @xmath27 , where @xmath28 is the cyclotron frequency .
the @xmath29 matrices read as @xmath30 , @xmath31 , @xmath32 , @xmath33 , @xmath34 , where @xmath35 are the two dimensional unity and pauli matrices , respectively , and @xmath36 @xcite .
the spectrum of @xmath37 is composed of a set of lls at well separated energies @xmath38 , where for @xmath39 @xmath40 ^ 2 \right]^{1/2 } , \label{energy}\ ] ] with degeneracies per unit area @xmath41 for @xmath42 and @xmath43 for @xmath44 . here @xmath45 is the total magnetization , @xmath46 is the _ magnetic length _ , and @xmath47 , @xmath48 .
at half filling , lls at negative ( positive ) energies are filled ( empty ) , and therefore @xmath49= - \ ; \frac{1}{2 \pi l^2_b } \sum_{\sigma=\pm } \big ( e_{0 \sigma } + e_{1 \sigma } + 2 \sum_{n \geq 2 } e_{n \sigma } \big ) .
\label{hfenergy}\ ] ] with @xmath50 and @xmath51 as independent variables , the free energy optimization condition @xmath52/\partial n_\parallel = 0 $ ] yields @xmath53 = 0.\ ] ] the left - hand side of this equation is a negative definite function of @xmath51 for any nontrivial zeeman coupling , and vanishes only for @xmath54 .
therefore , in the presence of magnetic field , laf order gets projected onto the easy plane ( @xmath55 ) due to the zeeman coupling , yielding the eplaf state .
this configuration also corresponds to the _ minima _ of the energy . placed in a magnetic field , an identical ground state , easy - plane nel order , can also be realized in monolayer graphene @xcite . with @xmath56 , minimizing @xmath57 with respect to @xmath58 and @xmath59 , we respectively obtain the coupled gap equations @xmath60 \nonumber \end{aligned}\ ] ] @xmath61.\ ] ] within the framework of a microscopic density - density interaction , such as the on - site hubbard model , @xmath62 at the lattice scale ( @xmath63mev in blg ) .
however , the magnetic field introduces a new length scale in the system , magnetic length @xmath64 ( thus a new effective cut - off @xmath65 ) , and generically @xmath66 .
we redefine the couplings as @xmath67 , for @xmath68 .
besides splitting the half - filled zll(@xmath44 ) , laf and ferromagnet ( fm ) ops , respectively , pushes down and splits all the filled lls ( @xmath69 ) . as a result , the first gap equation is devoid of any divergences , while the second one exhibits an _
ultraviolet logarithmic divergence _ , which , however , can be regularized by substituting @xmath70 , ensuring the cutoff independence of the laf op in magnetic fields . here
@xmath71 stands for the zero magnetic field laf gap in blg .
the same set of gap equations can also be obtained in a variation approach , developed in ref .
@xcite , where the fermionic field operators are expressed as @xmath72.\end{aligned}\ ] ] @xmath73 is the ll wave functions of @xmath37 at energies @xmath74 , where @xmath75 , with @xmath76 as the wave number , and @xmath77 , @xmath78 as the ll and valley index , respectively .
@xmath79 is as in eq .
( [ llenergy ] ) , but with @xmath56 @xcite .
the variational ground state energy is @xmath80 , where @xmath81 @xmath82 @xmath83 @xmath84 , and the ground state @xmath85 is chosen such that @xmath86 . here @xmath87 is a generic four - fermion density - density interaction at the lattice scale @xcite , and the above gap equations are obtained by minimizing @xmath88 , with respect to @xmath58 and @xmath89 , where @xmath90 .
@xmath91 and @xmath92 respectively represent the forward and back - scattering interactions @xcite .
the magnetization ( @xmath58 ) increases monotonically with @xmath93 , and it scales linearly with the magnetic field , when @xmath94 t .
in this regime , @xmath95 , but coefficient of @xmath96 decreases with increasing @xmath93 , however , very softly . for stronger magnetic fields ( @xmath97 t )
@xmath58 scales non - linearly with @xmath98 for a given @xmath93 , and it becomes challenging to track the scaling of @xmath58 with @xmath98 for a fixed @xmath93 .
instead we search for the self - consistent solutions of @xmath59 and @xmath58 , yielding reasonable agreements with the recently observed scalings of the gap at cnp @xcite , which here reads as @xmath99 .
results are shown in fig . 1 ( left ) , exhibiting excellent agreement with various experiments when the laf order is accompanied by a sizable fm order ( second and third panels of fig . 1 ) ,
see also sec . iv of ref . .
notice that quadratic scaling of the gap at low fields in ref . crosses over to a linear one for @xmath100 t , and the scaling of @xmath101 with an underlying eplaf state yields excellent description of these two scaling regimes . on the other hand , in refs .
, gaps at the cnp have been measured for @xmath102 t , where it scales quite linearly with @xmath98 , and the scaling of @xmath101 is in good agreement with these observations as well .
it is interesting to note that at minimal cost of laf order , blg can develop a large fm order ; compare second and the third panels of fig . 1 .
such peculiar behavior has a root in the fact that by depleting the laf order system looses a significant amount of condensation energy , since the laf order pushes down all the filled lls below the chemical potential .
the compensating fm order , which , on the other hand , lowers the ground state energy only by enhancing the splitting of the zll , therefore needs to be large , in agreement with the results obtained from the self - consistent calculations .
the fm op(m ) scales quite linearly ( third panel of fig . 1 ) and the dimensionless ferromagnetic interaction ( @xmath93 ) decreases monotonically ( fourth panel of fig . 1 ) , with increasing @xmath98 along all the curves in fig .
1(left ) , observed experimentally .
hence , @xmath93 exhibits _ universal flow _ towards its bare value @xmath103 , which is @xmath104 in refs . as @xmath105 . at such strong magnetic fields , the two - band continuum description of blg [ eqs .
( [ twoband2 ] ) ] completely breaks down , and finite - size effects of the system become important@xcite .
however , at intermediate strength of the magnetic field blg can be properly described by a _
four - band _ model ( including the split - off bands ) in the continuum limit .
such crossover roughly takes place around @xmath106 t , when only few lls ( say @xmath107 ) are placed within the cutoff @xmath63 mev for two band model @xcite .
the scaling of the gap at and near the cnp in blg beyond @xmath108 becomes qualitative similar to the one in monolayer graphene @xcite , about which in a moment .
t ( red ) , @xmath109 t ( black ) , @xmath110 t ( blue ) , @xmath111 t ( green ) , and @xmath112 , as a function of @xmath93.,title="fig:",width=158,height=122 ] t ( red ) , @xmath109 t ( black ) , @xmath110 t ( blue ) , @xmath111 t ( green ) , and @xmath112 , as a function of @xmath93.,title="fig:",width=158,height=122 ] on the other hand , the excitation spectrum in the variational approach @xmath113 , where @xmath114 is the excited state @xcite , reads as @xmath115 where @xmath116 , @xmath117 , and @xmath118 corresponds to the ll index of @xmath119 @xcite .
therefore , the excitation spectrum depends on two parameters @xmath93 and @xmath120 . the interaction @xmath120 possibly captures the effect of quantum hall ferromagnet order @xcite .
however , for @xmath121@xcite the lowest energy excitation always occurs by creating particle - hole pair within the zll , and i obtain excellent agreements with different experiment with @xmath122 .
therefore , it appears that near the cnp hall ferromagnet order plays a very minor role in the quantum hall regime of blg .
however , the relative importance of these ops in blg can only be settled through future experiments .
it is also interesting to investigate the evolution of the eplaf state in tilted magnetic fields .
the laf and the interaction - driven fm ( @xmath58 ) ops scale only with the perpendicular component of magnetic field ( @xmath123 ) , while the zeeman term couples with the total magnetic field ( @xmath124 ) . performing the same set of self - consistent calculations , however ,
in the presence of tilted magnetic fields , i can not see any indication of a phase transition from eplaf to a pure fm state , even for fixed @xmath125 t , and a parallel component of the field @xmath126 as high as @xmath127 t and for @xmath128 , see fig .
2@xcite .
the existence of laf order even without a magnetic field possibly provides such robustness to the laf state in blg , placed in tilted magnetic fields , which has also been demonstrated in a recent experiment @xcite , where the gap at cnp is found to decrease as @xmath129ev / t @xmath130 with a _
perfectly _ parallel magnetic field @xcite .
nevertheless , interactions in blg can be weak enough , such that ordering possibly happens only in the presence of a perpendicular magnetic field , similar to what happens in monolayer graphene @xcite .
it is then possible to realize a transition from eplaf to a pure fm phase , at least when @xmath131 @xcite .
a pure fm state in blg yields a two - terminal hall conductance @xmath132 @xcite , since fm and qshi leads to identical splitting of the zll , which in turn supports four counter - propagating edge states ( @xmath133 in monolayer graphene @xcite ) . on the other hand ,
the edge states in pure laf / eplaf state are fully gapped , leading to @xmath134 at the cnp . recently ,
quantized two - terminal conductance of @xmath132 at the cnp , when @xmath135 t , and @xmath136 t , has been observed in a metallic blg @xcite .
however , quantized ( @xmath137 ) two - terminal conductance in an insulating blg , placed in parallel magnetic field , remains to be observed .
placing the chemical potential close to the first excited state at @xmath138 , additional incompressible hall states at filling @xmath1 can be formed by developing a third component of the laf order @xmath139 , in the direction of the applied magnetic field . to the leading order in @xmath51 the activation gap for @xmath1 hall states reads as @xmath140 a similar mechanism can be responsible for the formation of @xmath2 hall states in monolayer graphene @xcite . with an underlying eplaf ordering at @xmath0
, @xmath51 receives contributions only from half of the zll , and hence @xmath141 , but @xmath142 .
hence , @xmath143 is smaller than the gap for the @xmath0 hall state , and that is possibly why @xmath1 hall states are resolved only for @xmath144 t @xcite . for @xmath144 t
the gap at @xmath0 scales linearly with @xmath98 @xcite , and thus the gap at @xmath3 should also scale linearly with @xmath98 , in qualitative agreement with recent experimental observations@xcite .
since , in the presence of perpendicular electric field @xmath101 decreases @xcite , resultantly @xmath143 should increase .
strong electric field induced enhancement of the gap for @xmath3 hall state has already been observed experimentally @xcite .
however , a finite @xmath51 at filling @xmath1 also causes simultaneous layer - polarization of average electronic density .
hence , @xmath143 can either increase or decrease with a weak electric field , depending on the relative sign of @xmath51 and electric field induced layer polarizations , which may serve as a litmus test of the proposed scenario . at stronger magnetic fields , the linear scaling of the @xmath145 hall states
is expected to cross over to a @xmath146 scaling , similar to the one in monolayer graphene @xcite .
recently observed linear scaling of the gap at @xmath147 for @xmath148t@xmath149 t @xcite and a @xmath146 scaling of the @xmath3 hall state for @xmath150 t @xcite possibly bears the signatures of such crossover scaling in blg . due to the enhanced interaction effect in blg
such crossover can take place within accessible range of magnetic fields .
interaction - driven orders can not lift the orbital degeneracy ( @xmath151 ) of the zll in blg .
however , the remote hopping between the sites on two layers , represented by @xmath152 gives rise to a non interacting gap between the zlls with @xmath153 and @xmath148 , since in the presence of magnetic fields , @xmath154 , where @xmath155 is the ll number operator , yielding @xmath2 hall states . here @xmath156 and @xmath157 for @xmath158 .
@xmath159 are respectively the intralayer and interlayer ( next-)nearest - neighbor hopping amplitudes @xcite . the activation gap for the @xmath2 hall state , @xmath160 should scale linearly with the magnetic field , and with currently estimated strength for various band - parameters @xmath161 mev / t @xcite .
similar splitting can also be achieved by applying an electric field between the layers@xcite , and so far in an insulating blg @xmath162 hall plateau has only been observed in the presence of perpendicular electric fields@xcite .
nevertheless , in metallic blg the @xmath162 hall state has been observed at strong magnetic fields , and the gap is found to scale as @xmath163mev / t @xcite and @xmath164 k / t@xcite .
formation of fractional quantum hall states in the zll depends on its degeneracy lifting at integer fillings@xcite . at weak magnetic fields ( and without any electric field ) , when the orbital degeneracy of the zll is protected , but @xmath165 hall plateaus are well resolved , plateaus are expected to appear at fractional fillings @xmath166 for @xmath167 , where @xmath168 , and @xmath169 is the standard jain s sequence@xcite .
the additional factor of 2 in the numerator arises from the residual orbital degeneracy of the zll . at stronger magnetic fields , when the orbital degeneracy of the zll is lifted and plateaus at fillings @xmath170 are well resolved , blg should discern standard jain s sequences at fillings @xmath171 and @xmath172 .
a detailed study of the fractional quantum hall effect in blg is quite rich , and i leave it for future investigation .
nevertheless , recently there have been suggestive signatures for the @xmath173 fractional hall plateau in blg , where @xmath162 hall state has also been resolved @xcite .
_ acknowledgement : _ i would like thank o. vafek for suggesting this work , many useful discussions and his continued interest in this work .
author is in debt to igor .
f. herbut for number on interesting discussions and valuable comments on this rapid communication .
it is the author s pleasure to acknowledge fruitful discussion with c. n. lau and j. velasco , jr .
the author is very thankful to m. weiss and c. sch@xmath174nenberger for providing many data from refs . , .
this work was supported at national high magnetic field laboratory by nsf cooperative agreement no .
dmr-0654118 , the state of florida , and the u. s. department of energy .
i am thankful to ecole de physique , les houches for hospitality during the summer school `` strongly interacting quantum systems out of equilibrium '' where a part of this work was finalized .
99 k. s. novoselov , a. k. geim , s. v. morozov , d. jiang , m. i. katsnelson , i. v. grigorieva , s. v. dubonos , and a. a. firsov , nature ( london ) * 438 * , 197 ( 2005 ) .
y. zhang , y .- w . tan , h. l. stormer , and p. kim , nature ( london ) * 438 * , 201 ( 2005 ) . v.
p. gusynin , and s. g. sharapov , phys .
rev .
lett . * 95 * , 146801 ( 2005 ) .
e. mccann , and v. i. falko , phys .
rev .
lett . * 96 * , 086805 ( 2006 ) .
r. t. weitz , m. t. allen , b. e. feldman , j. martin , a. jacoby , science * 330 * , 812 ( 2010 ) .
f. freitag , j. trbovic , m. weiss , c. sch@xmath174nenberger , phys .
rev .
lett . * 108 * , 076602 ( 2012 ) .
j. velasco jr . , l. jing , w. bao , y. lee , p. kratz , v. aji , m. bockrath , c.n .
lau , c. varma , r. stillwell , d. smirnov , fan zhang , j. jung , a.h .
macdonald , nat .
nano .
* 7 * , 156 ( 2012 ) .
f. freitag , w. weiss , r. maurand , j. trbovic , c. sch@xmath174nenberger , phys .
rev .
b * 87 * , 161402(r ) ( 2013 ) .
f. freitag , m. weiss , r. maurand , j. trbovic , and c. sch@xmath174nenberger , solid state commun .
* 152 * , 2053 ( 2012 ) .
b. roy , phys .
rev .
b * 88 * , 075415 ( 2013 ) .
r. nandkishore , and l. levitov , phys .
rev .
lett . * 104 * , 156803 ( 2010 ) ; arxiv:1002.1966 . o. vafek and k. yang , phys . rev .
b * 81 * , 041401(r ) ( 2010 ) . o. vafek , phys . rev .
b * 82 * , 205106(2010 ) . r. e. throckmorton and o. vafek , phys .
rev .
b * 86 * , 115447 ( 2012 ) .
d. v. kveshchenko , phys .
. lett .
* 87 * , 246802 ( 2001 ) .
v.p .
gusynin , v.a .
miransky , s.g .
sharapov , i.a .
shovkovy , phys .
rev .
b * 76 * , 195429 ( 2006 ) .
i. f. herbut , phys . rev .
b * 75 * , 165411 ( 2007 ) .
b. roy , phys .
rev .
b * 84 * , 035458 ( 2011 ) .
e. v. gorbar , v. p. gusynin , and v. a. miransky , phys .
rev .
b * 81 * , 155451 ( 2010 ) ; e. v. gorbar , v. p. gusynin , j. jia and v. a. miransky , _ ibid _ * 84 * , 235449 ( 2012 ) ; e. v. gorbar , v. p. gusynin , v. a. miransky , and i. a. shovkovy , _ ibid_. * 85 * , 235460 ( 2012 ) . for quantum hall physics in regular 2deg see _ quantum hall effect _ 2nd edition , edited by r. e. prange and s. m. girvin , ( springer - verlag , new york , 1989 ) .
i. f. herbut , phys .
rev .
b * 76 * , 085432 ( 2007 ) .
f. zhang , h. min , m. polini , a.h .
macdonald , phys .
rev .
b , * 81 * , 041402 ( r ) ( 2010 ) .
j. martin , b. e. feldman , r. t. weitz , m. t. allen , and a. yacoby , phys .
rev .
lett . * 105 * , 256806 ( 2010 ) .
y. zhao , p. cadden - zimansky , z. jiang , and p. kim , phys .
rev .
lett . * 104 * , 066801 ( 2010 ) .
j. velasco jr . , y. lee , z. zhao , l. jing , p. kratz , m. bockrath , c. n. lau , arxiv:1303.3649
. i. herbut _ a modern approach to critical phenomena _ , ( cambridge university press , cambridge , 2007 ) .
i. f. herbut , v. jurii , and b. roy , phys .
rev .
b * 79 * , 085116 ( 2009 ) ; b. roy , and i. f. herbut , phys . rev .
b * 82 * , 035429 ( 2010 ) .
see supplementary material for detail of ll wave - functions , spectrum of @xmath37 ; the variational calculation of gap equations , exciation spectrum ; and additional numerical results . c. yannouleas , i. romanovsky , and u. landman , phys .
rev .
b * 82 * , 125419 ( 2010 ) .
cut - off in a four - band model for blg is @xmath175ev .
m. kharitonov , phys .
rev .
b * 86 * , 195435 ( 2012 ) . within the framework of continuum description of the hubbard model in blg @xmath176 , and
thus @xmath121 at the scale @xmath177ev .
v.p .
gusynin , v.a .
miransky , s.g .
sharapov , i.a .
shovkovy , phys .
rev .
b * 77 * , 205409 ( 2008 ) .
d. a. abanin , k. s. novoselov , u. zeitler , p. a. lee , a. k. geim , l. s. levitov , phys .
rev .
lett . * 98 * , 196806 ( 2007 ) .
p. maher , c. r. dean , a. f. young , t. taniguchi , k. watanabe , k. l. shepard , j. hone , and p. kim , nat . phys .
* 9 * , 154 ( 2013 ) .
i. f. herbut , and b. roy , phys .
rev .
b * 77 * , 245438 ( 2008 ) , b. roy , and i. f. herbut , _ ibid _ * 83 * , 195422 ( 2011 ) ; _ ibid _ * 88 * , 045425 ( 2013 ) . in literature
sometime different notations are used for various hopping parameters , where @xmath178 , @xmath179 , and @xmath180 .
see for example ref ..
l. m. zhang , z. q. li , d. n. basov , m. m. fogler , z. hao , and m. c. martin , phys
. rev .
b * 78 * , 235408 ( 2008 ) .
r. cote , j. lambert , y. barlas , and a. h. macdonald , phys .
rev .
b * 82 * , 035445 ( 2010 ) .
d. v. khveshchenko , phys rev .
b * 75 * , 153405 ( 2007 )
. j. k. jain , _ composite fermions _ ,
( new york : cambridge university press , 2007 ) .
w. bao , z. zhao , h. zhang , g. liu , p. kratz , l. jing , j. velasco jr . ,
d. smirnov , c. n. lau , phys .
rev .
lett . * 105 * , 246601 ( 2010 ) . * supplementary material of `` theory of integer quantum hall effect in insulating bilayer graphene '' * bitan roy@xmath181 @xmath182 national high magnetic field laboratory , florida state university , florida 32306 , usa + @xmath183 condensed matter theory center , department of physics , university of maryland , college park , md 20742 , usa i here present details of the diagonalization of effective single - particle hamiltonian @xmath37 , derivation of gap equations using variational approach@xcite , and the computation of the excitation spectrum for @xmath0 quantum hall state in insulating bilayer graphene with an underlying easy - plane layer anti - ferromagnet order .
moreover , i also present some details on comparison of the gap at charge neutrality point @xmath101 in hartree - fock approach , with the measured gaps in various experiments@xcite , and evolution of the layer anti - ferromagnet order in tilted magnetic fields .
let us first show some detail of the how one can arrives at the gap equations , i have presented in the main part of the paper , with only the easy - plane component of the anti - ferromagnet and the easy axis ferromagnetic order , in the presence of a magnetic field . |
connecting the bench of experiment and the blackboard of theory offers physicists opportunities for creativity that we propose to make explicit .
traditional views underplay the physicist s role in making these connections .
although physicists have wished for mathematics that would connect directly to experiments on the bench , the equations of quantum mechanics express quantum states and operators not directly visible in spectrometers or other devices . here
we look into quantum mechanics as mathematical language used to model behaviors of devices arranged on the laboratory bench . after separating models as mathematical objects from any assertion that a certain model describes a given experiment with devices ,
we ask : given a certain form of model , which models , if any , fit the behavior of some particular devices on the bench ?
in contrast to any hope for a seamless , unique blackboard description of devices on a laboratory bench , we argue , based on mathematical proofs presented in sec . 3 , that no matter what experimental trials are made , if a quantum model generates calculated probabilities that match given experimentally determined relative frequencies , there are other quantum models that match as well but that differ in their predictions for experiments not yet performed .
the proofs demonstrate what before could only be suspected : between the two pillars of calculation and measurement must stand a third pillar of choice making , involving personal creativity beyond logic , so there can be no reason to expect or demand that any two people choose alike . what does recognizing choice mean for physicists ? in physics , as in artistic work , pleasure and joy come from the choices one makes that lead to something interesting . looking back
, physicists can hardly help noticing that their proudest accomplishments , whether theoretical or experimental , have involved choices made by reaching beyond logic on the basis of intuition , hunches , analogies some kind of guess , perhaps inspired , but still outside of logic . to understand the proofs is to see opportunities for making guesses .
although hunches and guesses and intuition can be as personal as dreams , the recognition of guesswork as a permanent pillar of physics has more than personal impact : 1 . describing device behavior
will be recognized in sec . 4 as a bi - lingual enterprise , with a language of wrenches and lenses for experimental trials on the bench and a different language of states and operators for the blackboard , linked by metaphors as guesswork .
we show how freedom to choose particles as constituents of models of devices both helps in modeling devices and allows us to replace a widespread but questionable postulate of `` state reductions '' by a theorem .
the need for bi - lingual descriptions bridging bench and blackboard gives local color to certain concepts . in sec .
5 we develop a notion of _ force _ in the context of light detection that gives meaning both at the blackboard and the bench not only to expectation values of light forces , but also to higher - order statistics associated with them , with application to models and detecting devices by which to better distinguish one source of weak light from another .
3 . in sec . 6 we uncover a symmetry , pertaining to entanglement , to make vivid the way judgments about how to model light are interdependent with judgments about how to model light - detectors .
experimental records can hold : ( 1 ) numerals interpreted as the settings of knobs that control an experiment , and ( 2 ) numerals interpreted as experimental outcomes , thought of as the clicks and flashes and electronically tallied pulses by which the devices used in the laboratory respond to and measure whatever is at issue . as an abstraction by which one can model experimental outcomes
, quantum theory offers what we shall call _ theoretical outcomes _ ( referred to in the literature variously as _ outcomes , results , the finding of particles _ and _ the finding of states _ ) .
probabilities of theoretical outcomes are expressed in terms of states and operators by what we shall call _ quantum models_. we discuss these first , and then distinguish the probabilities expressed by models from relative frequencies of experimental outcomes . the propositions about linking models to devices ( sec .
3 ) can be proved using any formulation of quantum mechanics that includes probabilities of theoretical outcomes .
here is a standard formulation taken from dirac @xcite and von neumann @xcite , bulwarked by a little measure theory @xcite ; however , as discussed in sec.[sec:4 ] , we invoke no postulate of state reductions .
let @xmath0 be a hilbert space , let @xmath1 be any self - adjoint operator of unit trace on @xmath0 ( otherwise known as a density operator ) , and let @xmath2 be a @xmath3-algebra of subsets of a set @xmath4 of possible
_ theoretical outcomes_. by a _ theoretical outcome _ we mean a number or a list of numbers as a mathematical object , in contrast to an experimental outcome extracted from an experimental record .
let @xmath5 be any projective resolution on @xmath2 of the identity operator on @xmath0 ( which implies that for any @xmath6 , @xmath7 is a self - adjoint projection @xcite ) .
let @xmath8 be a unitary time - evolution operator ( typically defined by a schrdinger equation or one of its relativistic generalizations ) .
these mathematical objects can be combined to define a probability distribution @xmath9 on @xmath2 , parameterized by @xmath10 : @xmath11 , \label{eq : mu}\ ] ] where @xmath12 is the probability of an outcome in the subset @xmath13 of @xmath4 , for the parameter value @xmath10 . for a probability @xmath12 to be compared to anything experimental , one needs to make explicit the dependence of the @xmath1 and @xmath5 that generate it on the experimentally controllable parameters .
it is convenient to think of these parameters as the settings of various knobs ; to express them we let @xmath14 and @xmath15 be mathematically arbitrary sets ( interpreted as sets of knob settings ) .
let @xmath16 be the set of functions from @xmath14 to density operators acting on @xmath0 .
let @xmath17 be the set of functions from @xmath15 to projective resolutions of the identity on @xmath2 of the identity operator on @xmath0 .
then what we shall call _ a specific quantum model _ is a triple of functions @xmath18 together with @xmath19 , @xmath20 , and @xmath8 a unitary evolution operator . by the basic rule of quantum mechanics , such a specific quantum model generates a probability - distribution as a function of knob settings and time : @xmath21 & & \qquad\mu(a , b , t ) ( \omega)=\mathrm{tr}[u(t ) \rho(a)u^\dag(t)e(b)(\omega ) ] .
\label{eq : muab } \end{aligned}\ ] ] often one needs something less specific than a triple @xmath18 . by a _ model _
we shall mean a set of properties of that limit but need not fully specify @xmath1 , @xmath22 , and @xmath5 .
for example , in modeling entangled light , we might construct a model in this sense by specifying relevant symmetry properties of @xmath5 , @xmath22 , and @xmath1 , leaving many fine points unspecified . * remarks * 1
. an element @xmath23 can be a list : @xmath24 ; similarly an element @xmath25 can be a list .
2 . by the _ domain _ of a model ,
we mean the cartesian product of the sets @xmath26 .
models of a given domain can differ as to the functions @xmath1 , @xmath5 , and @xmath22 defined on the sets @xmath14 , @xmath27 , and @xmath28 , respectively , so that different models with a given domain can differ in the states , operators , and probabilities that they assert .
3 . in defining the domain of a model
, we view @xmath9 as a function from @xmath29 $ ] , but in writing the left side of eq .
( [ eq : muab ] ) we take the alternative view of @xmath9 as a function @xmath30 .
different specific models can be distinguished by labels : a model @xmath31 consisting of @xmath32 generates a probability function @xmath33 . because @xmath33 is invariant under any unitary transformation applied to all three of @xmath34 , @xmath35 , and @xmath36 , an important feature of the states asserted by a specific model @xmath31 is the unitary - invariant overlap between pairs of them .
the measure of overlap convenient to sec .
[ sec:3 ] is @xmath37 .
\label{eq : overlap}\ ] ] we express the difference in state preparations asserted by specific models @xmath31 and @xmath38 having the same domain for elements @xmath39 by @xmath40 .
we call a model @xmath38 a _ restriction _ of a model @xmath31 if the domain of model @xmath38 is a subset of the domain of model @xmath31 .
so much for quantum models as mathematical objects ; how do we compare probabilities from these models with results of an experiment with lasers and lenses and other devices ? first one contrives to view the experiment as consisting of trials , each for certain settings of some knobs , yielding at each trial one of several possible experimental outcomes .
by tallying the experimental outcomes for various knob settings , one extracts from the experimental record the relative frequencies of experimental outcomes as a function on a domain of experimental knob settings and outcomes . to compare experimental relative frequencies with probabilities calculated from a model , both viewed as functions on domains of knob settings and outcome bins ,
it is necessary to identify the experimental domain as a subset of the model domain .
this entails associating to each experimental outcome @xmath41 some model outcome @xmath42 .
for the experimental relative - frequency of outcome @xmath41 for each setting of knobs @xmath43 in the experimentally covered subset of @xmath44 we write @xmath45 ; this is the ratio of the number of trials with knob settings and time @xmath43 and an experimental outcome in @xmath41 to the number of trials with knob settings and time @xmath43 regardless of the outcome . letting @xmath46 denote the set of experimental outcomes
, one has @xmath47 by virtue of the mapping @xmath48 , one can compare the experimental relative - frequency function @xmath49 with the probability function @xmath50 asserted by any model @xmath31 having a domain containing a subset identified with the domain of the experimental relative frequencies .
because of this need for identification , a choice of model domain constrains the design , or at least the interpretation , of experimental records to which models of that domain can be compared . in compensation , committing oneself to thinking about an experimental endeavor in terms of a particular model domain makes it possible to : 1 .
organize experiments to generate data that can be compared with models having that domain ; 2 .
express the results of an experiment mathematically without having to assert that the results fit any particular model ; 3 .
pose the question of whether the experimental data fit one model having that domain better than they fit another model @xcite .
what can relative frequencies of outcomes of experimental trials of devices tell us about how to model those devices ? at best , from relative frequencies of experimental outcomes as a function of knob settings , as in eq .
( [ eq : nudef ] ) , one abstracts an approximation to a probability - distribution function of knob settings . by ignoring statistical and other error
, we picture an ideal case of arriving at some @xmath51 , but without any further information concerning the states or operators that generate it .
this raises a question inverse to the text - book task of calculating probabilities from states and operators : given the probability function @xmath51 , what states and operators generate it via the rule of eq .
( [ eq : muab ] ) ? put another way , what are the constraints and freedoms on a density - operator function @xmath1 , an evolution operator @xmath8 , and a measurement - function @xmath5 if these are to generate a given probability function @xmath51 via the rule of eq .
( [ eq : muab ] ) ? here are some answers .
if for some values @xmath52 , @xmath53 , @xmath54 , and @xmath13 one has @xmath55 large and @xmath56 small , then eq . ( [ eq : muab ] ) implies that @xmath57 is significantly different from @xmath58 .
this can be quantified in terms of the _ overlap _ of two density operators @xmath57 and @xmath58 defined in eq .
( [ eq : overlap ] ) .
* proposition 1 * : for a specific model @xmath18 to be consistent with a probability function @xmath9 in the sense of eq.([eq : muab ] ) , the overlap of density operators for distinct knob settings @xmath52 and @xmath53 has an upper bound given by @xmath59 & & \mathrm{tr}[\rho(a_1)^{1/2}\rho(a_2)^{1/2}]\leq \min_{b , t,\omega } \{[\mu(a_2,b , t)(\omega)]^{1/2 } + [ 1-\mu(a_1,b , t)(\omega)]^{1/2}\}.\nonumber\\ \label{eq : upper}\end{aligned}\ ] ] _ proof _ : for purposes of the proof , abbreviate @xmath57 by @xmath52 , @xmath58 by @xmath53 , and @xmath60 by @xmath5 . because @xmath5 is a projection , @xmath61 . then the schwarz inequality and @xmath62 for which the traces exist , @xmath63^{1/2 } [ \mathrm{tr}(gg^\dag)]^{1/2}$ ] . ] and a little algebra implies @xmath64 & \leq & |\mathrm{tr}(a_1^{1/2}ea_2^{1/2})| + |\mathrm{tr}(a_1^{1/2}(1-e)a_2^{1/2})| \nonumber \\[3pt ] \noalign{\goodbreak } & \leq & ( \mathrm{tr}\,a_1)^{1/2 } [ \mathrm{tr}(ea_2^{1/2}a_2^{1/2}e)]^{1/2}\nonumber\\[3pt ] & & \mbox { } + [ \mathrm{tr}(a_1^{1/2}(1-e)(1-e)a_1^{1/2})]^{1/2 } ( \mathrm{tr}\,a_2)^{1/2 } \nonumber \\[3pt ] & = & [ \mathrm{tr}(ea_2^{1/2}a_2^{1/2}e)]^{1/2 } + [ \mathrm{tr}(a_1^{1/2}(1-e)(1-e)a_1^{1/2})]^{1/2 } \nonumber \\[3pt ] & = & [ \mathrm{tr}(a_2e)]^{1/2 } + [ 1 - \mathrm{tr}(a_1e)]^{1/2 } .
\end{aligned}\ ] ] expanding the notation , we have @xmath65 & & \qquad\mathrm{tr}[\rho(a_1)^{1/2}\rho(a_2)^{1/2}]\leq \min_{b , t,\omega}\{[\mathrm{tr}(\rho(a_2)u^\dag(t)e(b)(\omega)u(t))]^{1/2 } \nonumber \\[3pt ] & & \hskip1.85in\mbox { } + [ 1-\mathrm{tr}(\rho(a_1)u^\dag(t)e(b)(\omega)u(t))]^{1/2}\},\end{aligned}\ ] ] which , with eq .
( [ eq : muab ] ) , completes the proof .
_ example _ : for @xmath67 , if for some @xmath54 and @xmath13 , @xmath68 and @xmath55 @xmath69 , then it follows that @xmath70 \leq \epsilon^{1/2 } + \delta^{1/2}$ ] .
if , in addition , @xmath71 and @xmath72 , then we have @xmath73 .
the preceding proof of an upper bound invites the question : is there a corresponding positive lower bound ?
the answer turns out to be `` no . ''
* proposition 2 * : for any specific model @xmath74 and any knob settings @xmath75 , regardless of overlap@xmath76 there is a specific model @xmath77 with @xmath78 and overlap@xmath79 . _
proof by construction _ : let @xmath80 be the direct sum of three hilbert spaces @xmath81 , @xmath82 , and @xmath83 , each a copy of the hilbert space @xmath84 of model @xmath31 : @xmath85 .
let @xmath86 be the direct sum of three copies of @xmath87 , one for each of the @xmath88 ; similarly , let @xmath89 be the direct sum of three copies of @xmath36 . define @xmath90 by @xmath91 this defines a model @xmath38 of the form of eq .
( [ eq : muab ] ) for which we have @xmath92 but for the overlap as defined in eq .
( [ eq : overlap ] ) , @xmath93 , regardless of the value of @xmath94 .
@xmath66 this proof shows the impossibility of establishing by experiment a positive lower bound on state overlap without reaching outside of logic to make an assumption , or , to put it baldly , to _ guess _ @xcite .
the need for a guess , no matter how educated , has the following interesting implication .
any experimental demonstration of quantum superposition depends on showing that two different settings of the @xmath14-knob produce states that have a positive overlap .
for example , a superposition @xmath95 has a positive overlap with state @xmath96 .
because , by proposition 2 , no positive overlap is experimentally demonstrable without guesswork , we have the following : * corollary to proposition 2 * : experimental demonstration of the superposition of states requires resort to guesswork .
much the same story of constraint and freedom holds for resolutions of the identity . for the norm of an operator @xmath14 we take @xmath97 , where @xmath98 ranges over all unit vectors
then we have : * proposition 3 * : in order for a specific model @xmath18 to generate a given probability function @xmath9 , the resolution of the identity must satisfy the constraint @xmath99 _ proof _ : replacing unit vectors by density operators in the definition of the norm results in the same norm , from which we have @xmath100\\[3pt ] & & \mbox { } - \mathrm{tr}[\rho(a)e(b_2)(\omega_c)]|\\[5pt ] & = & \max_{a \in
a,\,t \in i}|\mu(a , b_1)(c ) - \mu(a , b_2)(c)|.\\ & & \hskip2.9 in \box\end{aligned}\ ] ] the difference of any two commuting projections has norm less than or equal to 1 .
can requiring a model @xmath31 to generate a given probability function @xmath9 impose any upper bound less than 1 on @xmath101 ? * proposition 4 * : for any specific model @xmath74 and any knob settings @xmath102 and @xmath103 , regardless of @xmath104 there is a specific model @xmath77 with @xmath105 and @xmath106 .
_ proof by construction _ : let the hilbert space for model @xmath38 be @xmath107 where @xmath108 is a space orthogonal to @xmath84 .
let @xmath109 and @xmath110 , where @xmath111 is the unit operator on @xmath108 and @xmath112 is the zero operator on @xmath108 .
let @xmath113 .
then , @xmath114@xmath115 .
@xmath66 * remark * : suppose , as displayed in the proof of proposition 2 , models @xmath116 and @xmath38 have the same domain and @xmath33 = @xmath117 , but the @xmath31-states have overlaps differing from those of the @xmath38-states
. then there is always a resolution of the identity @xmath118 outside the range of @xmath35 and @xmath119 , such that for some @xmath120 , @xmath121 \ne \mbox{tr}[\rho_\alpha(a)e']$ ] . in this sense
the models @xmath31 and @xmath38 conflict concerning their predictions @xcite .
the connection of any specific quantum model to experiments is via a probability function .
this and the proofs of propositions 2 and 4 show something that experiments can not show , namely that modeling an experiment takes guesswork , and that a model , once guessed , is subject to surprises arising in experiments not yet performed .
some guesses get tested ( one speaks of _ hypotheses _ ) , but testing a guess requires other guesses not tested . by way of example , to guide the choice of a density operator by which to model the light emitted by a laser , one sets up the laser , filters , and a detector on a bench to produce experimental outcomes . but to arrive at any but the coarsest properties of a density operator one needs , in addition to these outcomes , a model of the detector , and concerning this model , there must always be room for doubt ; we can try to characterize the detector better , but for that we have to assume a model for one or more sources of light . when we link bench and blackboard , we work in the high branches of a tree of assumptions , holding on by metaphors , where we can let go of one assumption only by taking hold of others . because of the guesswork needed to bridge between models and experiments , describing device behavior is forever a bi - lingual enterprise , with a language of wrenches and lenses for the bench and a different language of states and operators for the blackboard
. we will show how some words work as metaphors , straddling bench and blackboard , where by ` words ' we mean to include whatever mathematical symbols are used to describe devices .
we consider the mathematics of quantum mechanics not in contrast to words but as blackboard language , words of which are sometimes borrowed for use at the bench to describe devices . by showing some choices of metaphorical uses of the words _ state , operator , spacetime , outcome , and particle _ , we promote freedom to invent particles as needed to describe interesting features of device behavior .
recognizing choices in word use reflects back on how we formulate quantum mechanics : the notion of repeated measurements ` of a state ' will be revealed as neither necessary nor sensible , and the so - called postulate of state reductions will evaporate , leaving in its place a theorem .
we start by looking at several related but distinct uses of _
spacetime coordinates_. in the laboratory one uses clocks and rulers to assign coordinates to acts of setting knobs , transmitting signals , recording detections , _ etc _ , and one thinks of these experimentally generated coordinates as points of a spacetime something mathematical .
we call this a ` linear spacetime ' to distinguish it from a second spacetime , that we call ` cyclic , ' onto which the linear spacetime is folded , like a thread wound around a circle , so that experimental outcomes for different trials can be tallied in bins labeled by coordinates .
distinct from linear and cyclic spacetimes , any quantum - mechanical model involves a third spacetime on which are defined solutions of a schrdinger equation ( or one of its relativistic generalizations ) , and it is with reference to this spacetime that particles as theoretical constructs are defined .
any quantum model written in terms of particles generates probabilities , and if the probabilities of the model fit the relative frequencies of experimental outcomes well enough , one is tempted to say that one has `` seen the particles '' ; however , because particles in their mathematical sense are creatures of models , and multiple , conflicting models are consistent with any given experimental data , this `` seeing of particles '' stands on guesswork and metaphor , needed , for example , to bind the _ electron _ as a solution of the dirac equation defined on a model spacetime to a flash from a phosphor on a screen .
this metaphorical role of _ electron , photon , etc .
_ , though habitual and easily overlooked , can be noticed when a surprise prompts one to make a change in the use of the word _ electron _ at the blackboard while leaving the use at the bench untouched , or _
vice versa_.=-1 next we address notions of ( a ) components of a theoretical outcome , ( b ) a distinction between signal particles and probe particles , and ( c ) various measurement times . we take these in order .
the term _ theoretical outcome _ pertains to a vector space of multi - particle wave functions defined on a model spacetime .
this vector space is a tensor product of factors , one factor for each particle . for a resolution of the identity that factors
accordingly , we shall view each theoretical outcome for this resolution as consisting of a list of components , one component for each of the factors .
a probability density for such multi - component outcomes can be viewed as a joint probability density for the component parts of the outcome , modeling the joint statistics of the detection of many particles .
we who model are always free to shift the boundary between states ( as modeled by density operators ) and measuring devices ( as modeled by resolutions of the identity ) so as to include more of the measuring devices within the scope of the density - operator part of the model @xcite .
consider for example a coarse model @xmath31 that portrays a detecting device by a resolution of the identity . while a resolution of the identity has no innards ,
detecting devices do .
to model , say , a photo - diode and its accompanying circuitry , we can replace model @xmath31 by a more detailed model @xmath38 : the quantum state asserted by model @xmath31 becomes what we call a _ signal _ state , a factor in a tensor product ( or more generally a sum of tensor products ) accompanied by factors for one or more _ probe - particle states_. according to this model @xmath38 , the signal state is measured only indirectly , via a probe state with which it has interacted , followed by a measurement of the probe states , as modeled by `` a resolution of the identity '' that works on the probe factor , not the signal factor .
recognizing probe - states as free choices in modeling clarifies a variety of times relevant to quantum measurements .
for any quantum model @xmath31 , the form of eq .
( [ eq : muab ] ) links an outcome ( whether single- or multi - component ) to some point time @xmath122 ; however , the use of such a model is to describe an actual or anticipated experiment , and for this , as described above , one is always free to choose a more detailed model @xmath38 , in which the state of model @xmath31 appears as the _ signal _ state that interacts with a probe state , followed by a measurement of the probe state at some time @xmath123 after the interaction . thus model @xmath38 replaces the point time @xmath122 by a time stretch during which the signal and probe states interact , thus separating the time during which the signal state interacts with the probe from the time at which the probe state is measured . in more complex models involving more probe states ,
a succession of `` times of measurement '' in the sense of interactions can be expressed by a single resolution of the identity .
finally , in modeling spatially dispersed signal states that interact with entangled probe particles @xcite , one can notice a prior `` probe - interaction time '' during which the probe particles must interact with one another , in order to have become entangled .
recognizing choice in modeling allows one to sidestep a long - troubling issue in formulating quantum language .
in logical conflict with the schrdinger equation as the means of describing time evolution @xcite , dirac and other authors introduce _ state reductions _ by a special postulate that asserts an effect on a quantum state of a resolution of the identity ; allegedly needed to express repeated measurements of a system .
once we recognize the modeling freedom to make signal - probe interactions explicit , we can always replace any story about devices involving a `` state to be measured repeatedly '' by a model in which a _ signal _ state interacts with a succession of probe states , followed by a simultaneous measurement of all the probe states , as expressed by a single resolution of the identity and a composite state that incorporates both signal and probe states .
thus any apparent need for a postulate to do with `` repeated measurements '' evaporates , and with it the unfortunate appearance of state reductions in a postulate .
although inconsistent as a postulate , state reduction still works in many cases as a trick of calculation , as justified by the following theorem . * theorem * : assume any specific model of the form eq .
( [ eq : muab ] ) , and assume @xmath124 is an outcome with components @xmath125 and @xmath126 . if @xmath127 is a tensor product @xmath128 , then for any density - operator function @xmath129 , the joint probability distribution @xmath130 induces a conditional probability distribution for @xmath126 given @xmath125 that matches the quantum probability of @xmath126 obtained using a `` reduced density operator '' obtained by the usual rule for state reduction applied to @xmath1 .
_ proof _ : streamline notation by suppressing the dependence on @xmath120 and @xmath54 , and incorporate @xmath22 into @xmath1 , so that the relevant form is that of eq .
( [ eq : mu ] ) . for any state @xmath1 it follows from eq .
( [ eq : mu ] ) that @xmath131 $ ] . the conditional probability of @xmath126 given @xmath125 is defined by bayes rule @xcite : @xmath132}{\mbox{tr}[\rho ( e_a(j)\otimes \sum_{k'}e_b(k'))]}.\end{aligned}\ ] ] by the definition of a resolution of the identity
, we have @xmath133 ; recalling that @xmath134 is a projection that commutes with @xmath135 , one then has @xmath136,\ ] ] for an operator @xmath137}.\ ] ] this @xmath138 matches the ` reduced density operator ' obtained by the usual rule of state reduction . q.e.d . * remarks * : 1 . in case @xmath139
is a pure state , then @xmath140 , where the reduced state @xmath141 which is one form of the usual rule for state reduction , but here obtained by calculation with no need for any postulate .
2 . either of the outcome
components can be a composite , so the theorem applies to cases involving more than two outcome components .
3 . in relativistic formulations of quantum mechanics , detections at spatially separated locations
@xmath14 and @xmath15 can be modeled by projections of the form assumed .
of interest in particle physics , astrophysics , and emerging practical applications , sources of weak light are characterized experimentally by the experimental outcomes of detectors @xcite . because detector outcomes are statistical
, trial - to - trial differences in outcomes can arise both from trial - to - trial irregularity in the sources and from quantum indeterminacy in their detection . as we shall see , detecting devices work in two parts , one of which balances a light - induced force against some reference .
by taking advantage of the freedom to invent probe particles when we model particle detectors , we are led to a quantum mechanical expression of force in the context of balancing devices , with application to models and detecting devices by which to better distinguish one source of weak light from another . in newtonian physics , the word _
force _ is used both on the blackboard and with balancing devices . in quantum physics ,
_ force _ , as used at the blackboard , gets re - defined in terms of the expectation values pertaining to dynamics of wave functions @xcite
. we will find useful a concept of force in characterizing light .
because of its employment in experimental work , our concept of _ force _ necessarily takes on local coloring special to one or another experimental bench ; we develop a notion of _ force _ in the context of light detection that gives meaning both at the blackboard and the bench not only to expectation values of forces , but also to higher - order statistics associated with them .
these higher - order statistics allow the expression , within quantum mechanics , of the teetering of a balance that happens when forces are nearly equal .
we begin by reviewing some details of detector behavior . under circumstances to be explored , particle detectors employed to decide among possible quantum states produce unambiguous experimental outcomes .
seen up close , a detecting device consists of two components .
the first is a _
transducer _ such as a photo - diode that responds to light by generating a small current pulse . to tally a transducer response as corresponding to one or another theoretical outcome in the sense of quantum mechanics
, one has to _ classify _ the response using some chosen criterion . as phrased in the engineering language of inputs and outputs , the response of the transducer
is fed as an input to a second component of the detector , in effect an unstable balance implemented as a flip flop ( made of transistors organized into a cross - coupled differential amplifier ) .
the flip - flop produces an out_put _ intended to announce a decision between two possible experimental out_comes _ , say 0 and 1 .
if we think classically , we picture the flip - flop as a ball and two bins , one bin for each possible outcome , separated by a barrier , the height of which can be adjusted , as shown in fig .
[ fig:1 ] .
the ball , starting in bin 0 is kicked by the transducer ; an outcome of 1 is recorded if and only if the balance is tipped and the ball rolls past the barrier into bin 1 .
this ball - and - bin technique avoids ambiguity by virtue of a convention that gives the record a certain leeway : it does not matter if the ball is a little off center in its bin , so long as the ball does not teeter on the barrier between bins .
although usually producing an unambiguous outcome , the flip - flop , seen up close , can teeter in its balancing , perhaps for a long time , before slipping into one bin or the other @xcite .
absent some special intervention , two parties ( people or machines ) to which a teetering output fans out can differ in how they classify this output as an outcome : one finds a 0 , the other finds a 1 . to reduce the risk of disagreement , the two parties have to delay their reading of the output , hopefully until the ball slips into one or the other bin .
ugly in this classical cartoon are two related features : ( 1 ) the ball can teeter forever , so that waiting is no help , and ( 2 ) the mean time for teetering to end is entirely dependent on some _ ad hoc _ assumption about `` noise '' @xcite . although we have described the flip - flop classically , it is built of silicon and glass presumably amenable to quantum modeling .
to see what quantum models offer us , the first step is to recall that a quantum model implies probabilities to be related to an experiment , so that inventing a quantum model and choosing an experimental design go hand in hand .
thirty years ago , thinking not in quantum but in circuit terms , we designed and carried out an experiment to measure teetering of the output of a flip - flop .
as we recognized only recently , the record of this experiment is compatible with some quantum models .
a quantum model to be offered shortly describes the experiment already performed and serves as a guide for designing future experiments to exploit what can be called a statistical texture of force , previously obscured by the `` noise '' invoked in classical analysis . to experiment with the teetering of a detecting device comprised of a transducer @xmath142 ( for diode ) connected to a flip - flop @xmath143 , shown in fig .
[ fig:2 ] , we replace the transducer @xmath142 by a laboratory generator of weak electrical pulses to drive @xmath143 .
, one could replace one transistor of flip - flop @xmath143 by a photo - transistor . ]
putting the flip - flop @xmath143 into a teetering state takes very sensitive adjustment of the pulse generator , achieved by feedback from a running average of the outcomes produced by @xmath144 and @xmath145 to the pulse generator .
the output of @xmath143 is made to fan out , as shown in fig .
[ fig:3 ] , to a matched pair of flip - flops , @xmath144 and @xmath145 , each of which acts as an auxiliary detector , not of the incoming light but of the output of @xmath143 .
the flip - flops @xmath144 and @xmath145 are clocked at a time @xmath146 later than is @xmath143 .
the experimental outcome consists of two binary components , one from @xmath144 and the other from @xmath145 .
if after the waiting time @xmath146 the output of @xmath143 is still teetering , the flip - flops @xmath144 and @xmath145 can differ , one registering a 1 while the other registers a 0 ; the disagreement between the two flip - flops registers the teetering of the output of @xmath143 . the measured relative frequency of disagreements between @xmath144 and @xmath145 is shown in fig .
[ fig:4 ] .
we want to model the teetering statistics that we will use to discriminate among various sources of weak signals .
traditional analyses of solid - state detecting devices and their flip - flops invoke quantum mechanics only to determine parameters for classical stories involving voltage and current .
analyzed that way , teetering in a photo - diode - based detector that employs a flip - flop made of transistors arises in two ways : first , there can be teetering in the entry of electrons and holes into the conduction band of the photo - diode ; second , there can be teetering in the response of @xmath143 to whatever amplified pulse comes from the photo - diode .
although both these teeterings involve electrons and holes going into a conduction band , the statistical spread of outputs for a given state is blamed on _ noise _ unconnected with the signal , and known analyses of a flip - flop invoke _ noise _ to evade the embarrassment of a possible infinite hesitation . avoiding the invocation of ` noise , ' we picture @xmath143 quantum mechanically as a pair of probe particles .
light acting via the transducer applies a force to the two probe particles .
this scattering process transforms an initially prepared state of the light and probes to an out - state consisting of a sum ( or integral ) of products , each of which has a factor for the light and a factor for the two probe particles of @xmath143 .
after some waiting time of evolution @xmath146 , the two probe particles are measured , as expressed by a resolution of the identity that ignores the signal state ; thus the probabilities of theoretical outcomes of the detection after the interaction are expressible by a reduced density operator obtained by tracing over the signal states . in this view , teetering shows up in the probability of detecting the two probe particles on different sides of a reference ; we will show how this probability depends on both the signal detected and a waiting time @xmath146 , and how in this dependence planck s constant enters .
in the model presented here , we simplify the effect of the signal state to that of preparing , at time 0 , a pair of probe - particle wave functions . for simplicity ,
the probe - particle wave functions have only one space dimension .
let @xmath147 be the space coordinate for one probe particle and @xmath148 be the space coordinate for the other .
the difference between one possible signal state and another is reflected by concentrating the initial probe - state wave functions slightly to one side or the other of an energy hump centered at the origin @xmath149 , @xmath150 .
( as might be expected , teetering is most pronounced in the borderline case of a signal state that puts the initial probe - particle wave functions evenly over the energy hump . )
we model the flip - flop @xmath144 as a resolution of the identity that has an theoretical outcome of 1 for @xmath151 and 0 for @xmath152 ; similarly @xmath145 is modeled as a resolution of the identity for @xmath148 .
thus the two - component theoretical outcomes are 00 for @xmath153 ; 01 for @xmath154 , @xmath155 ; 10 for @xmath156 , @xmath157 ; and 11 for @xmath158 . by assuming a coupling between the two probe particles
, we will model how increasing the waiting until time @xmath146 to detect the probe particles decreases the probability of disagreement between @xmath144 and @xmath145 , _ i.e._diminishes the probability of @xmath147 and @xmath148 being measured with different signs .
thinking of the @xmath147-probe particle as a wave - function concentrated near an energy hump , assuming that the long - time behavior of the particle depends only on the hump curvature , and for the moment neglecting coupling between the two probe particles , we express the dynamics of the @xmath147-particle by the schrdinger equation for an unstable oscillator : @xmath159 where the instability comes from the minus sign in the term proportional to @xmath160 .
we express the @xmath148 probe particle similarly . in order to produce growth over time in the correlation of the detection probabilities
, we put in the coupling term @xmath161 .
this produces the following two - particle schrdinger equation which is the heart of our model : @xmath162 the natural time parameter for this equation is @xmath163 defined here by @xmath164@xmath165 ; similarly there is a natural distance parameter @xmath166 . for the initial condition
, we will explore a wave packet of the form : @xmath167\exp[-(y - c)^2/2b^2 ] .
\label{eq : initxy}\ ] ] for @xmath168 , this puts the recording device exactly on edge , while positive or negative values of @xmath41 bias the recording device toward 1 or 0 , respectively .
as discussed in appendix a , the solution to this model is @xmath169 with @xmath170 , \nonumber \\ b_2 ^ 2(t ) & = & b^2\left[1 + \left(\frac{\hbar^2}{\omega^2 m^2 b^4(\lambda-1)}-1\right ) \sin^2 \sqrt{\lambda-1}\,t\right ] .
\label{eq : bphysdef}\end{aligned}\ ] ] the probability of two detections disagreeing is the integral of this density , @xmath171 , over the second and fourth quadrants of the @xmath172-plane . for the especially interesting case of @xmath173 ,
this integral can be evaluated explicitly as shown in appendix a : @xmath174\sin^2\sqrt{\lambda-1}\,\omega t}{\displaystyle 1 + \left(\frac{\hbar^2}{\omega^2 m^2 b^4}+1\right)\sinh^2 \omega t}\right)^{1/2}. \label{eq : edge1phys}\end{aligned}\ ] ] this formula works for all real @xmath175 . for @xmath176 , it shows an oscillation , as illustrated in fig .
[ fig:4 ] . for the case
@xmath177 , the numerator takes on the same form as the denominator , but with a slower growth with time and lacking the oscillation , so that the probability of disagreement still decreases with time , but more slowly . picking values of @xmath54 and @xmath175 to fit the experimental record
, we get the theoretical curve of fig .
[ fig:4 ] , shown in comparison with the relative frequencies ( dashed curve ) taken from the experimental record . for the curve shown , @xmath178 and @xmath179 times the characteristic distance @xmath180 . according to this model @xmath31 , a design to decrease the half - life of disagreement
calls for making both @xmath181 and @xmath175 large .
raising @xmath175 above 1 has the consequence of the oscillation , which can be stronger than that shown in fig .
[ fig:4 ] .
when the oscillation is pronounced , the probability of disagreement , while decreasing with the waiting time @xmath10 , is not monotonic , so in some cases judging sooner has less risk of disagreement than judging later . as proposition 2 of sec .
[ sec:3 ] suggests , one can construct alternatives to the above model @xmath31 of a flip - flop . instead of initial probe states specified by `` locating blobs , '' expressed in the choice of the value of @xmath41 in eq .
( [ eq : initxy ] ) , a model @xmath38 can employ initial probe states specified by momenta . in this `` shooting of probe particles at an energy hump , ''
the initial wave functions are concentrated in a region @xmath182 and propagate toward the energy saddle at @xmath183 . writing a 0
is expressed by an expectation momentum for the initial state less than that for the initial state that corresponds to writing a 1 .
hints for this approach are in the paper of barton @xcite , which contains a careful discussion of the energy eigenfunctions for the single inverted oscillator of eq .
( [ eq : phiprob ] ) , as well as of wave packets constructed from these eigenfunctions . such a model @xmath38 based on an energy distinction emphasizes the role of a flip - flop as a decision device : it `` decides '' whether a signal is above or below the energy threshold . for finite @xmath54 , the limit of eq .
( [ eq : edge1phys ] ) as @xmath185 is @xmath186 this classical limit of model @xmath31 contrasts with the quantum - mechanical eq .
( [ eq : edge1phys ] ) in how the disagreement probability depends on @xmath175 .
quantum behavior is also evident in entanglement exhibited by the quantum - mechanical model . at @xmath187
the wave function is the unentangled product state of eq.([eq : initxy ] ) .
although it remains in a product state when viewed in @xmath188-coordinates discussed in appendix a , as a function of @xmath172-coordinates it becomes more and more entangled with time , as it must to exhibit correlations in detection probabilities for the @xmath147- and @xmath148-particles . by virtue of a time - growing entanglement and the stark contrast between eq .
( [ eq : edge1phys ] ) and its classical limit , the behavior of the 1-bit recording device exhibits quantum - mechanical effects significantly different from any classical description .
the alternative model @xmath38 based on energy differences can be expected to depend on a _ sojourn time _ with its interesting dependence on planck s constant , as discussed by barton @xcite . both models @xmath31 and @xmath38 thus bring planck s constant into the description of decision and recording devices , not by building up the devices atom by atom , so to speak , but by tying quantum mechanics directly to the experimentally recorded relative frequencies of outcomes of uses of the devices .
detection of teetering in a detector of weak light pulses allows finer distinctions by which to characterize a source of that light . without teetering , a first measure of a weak light source
is its mean intensity , expressed operationally as the fraction of 1 s detected in a run of trials in which it illuminates a detector .
now comes a refinement that draws on teetering . if the detector s balancing flip - flop @xmath143 fans out to auxiliary flip - flops @xmath144 and @xmath145 , two sources @xmath14 and @xmath15 that produce the same fraction of 1 s can be tested for a finer - grained distinction as follows
for each source , using feedback to stabilize the relation between the source and @xmath143 , so that the running fraction of 1 s detected is held nearly steady , conduct one run of trials for a fixed waiting time @xmath189 , a second run of trials for a fixed waiting time @xmath190 , _ etc .
_ let @xmath191 denote the fraction of trials of source @xmath14 for which the outcome components from @xmath144 and @xmath145 disagree in the run of trials with waiting time @xmath192 ; similarly , @xmath193 denotes this fraction for source @xmath15 .
these additional data express additional statistical `` texture '' by which to compare source @xmath14 against source @xmath15 .
even when they produce the same overall fractions of 1 s , they are still measurably distinguishable if they consistently show strong differences , for some @xmath192 , between @xmath191 and @xmath193 .
for example , if source @xmath15 has more classical jitter than source @xmath14 , so that the @xmath15 quantum state bounces up and down in expectation energy from pulse to pulse , then source @xmath15 is more apt than source @xmath14 to push the probe particles both over the hill or neither over the hill . in other words ,
source @xmath14 will produce more teeterings , and hence we would find @xmath191 significantly greater than @xmath193 .
with the the freedom to invoke probe particles as developed in secs . 4 and 5 , we can show a striking additional freedom of choice in modeling , resolvable only by going beyond the application of logic to experimental data .
this freedom pertains to entanglement .
suppose that experimental trials yield outcomes consistent with a model , according to which a source of entangled weak light illuminates a pair of unentangled detectors at two separate locations , @xmath14 and @xmath15 .
models of detectors detailed enough to invoke probe - particle states , as in sec . 5 , must specify how these probes are prepared . as discussed in a different context long ago
@xcite , there is the possibility of entangling the probe particles , amounting to entangling the detectors .
this possibility of an entangled pair of detectors points to a symmetry relation between entanglement of signals and entanglement of probes . to see how this works ,
consider first modeling a single detector involving a probe state prepared by choosing some @xmath194 , where @xmath195 is a parameter such as the expectation momentum for this state .
consider also a set of possible signal states @xmath196 .
assume that the detector model calls for detecting the probe particle after its interaction with the signal particle , as expressed by a measurement operator @xmath5 acting on the probe alone .
then the probability of outcome @xmath125 resulting from an initial signal @xmath196 interacting with a probe @xmath197 has the form of the square of a complex amplitude that depends on both @xmath198 that labels the signal state and @xmath199 that labels the probe state : @xmath200 with @xmath201 here @xmath22 , acting on the combined signal - probe space of wave functions , expresses the signal - probe interaction . assume for the moment that the probability in eq .
( [ eq : pj ] ) is symmetric under interchange of signal and probe : @xmath202 which implies that for some real - valued phase - function @xmath203 , @xmath204 to see how this symmetry impacts on modeling a pair of detectors , consider two such detectors , one at location @xmath14 , the other at @xmath15 , having identical initial probe states @xmath205 and @xmath206 , respectively .
for unentangled signals having initial states @xmath207 and @xmath208 , the amplitude for the joint outcome @xmath209 at @xmath14 and @xmath210 at @xmath15 is then @xmath211 & & [ ( 1_{sa}\otimes e_a(j))\otimes(1_{sb}\otimes e_b(k))](u_a\otimes u_b ) \nonumber \\[8pt ] & & \quad [ ( |s_a(q_1)\rangle |p_a(q_0)\rangle)\otimes ( |s_b(q_2)\rangle \label{eq : joe } \end{aligned}\ ] ] ( which can be written as a product of @xmath14- and @xmath15-factors , expressing the lack of correlation when there is no entanglement ) .
now consider the same pair of detectors responding to an entangled signal state @xmath212 , \label{eq : entstate}\ ] ] where @xmath213 is a normalizing constant , dependent on @xmath195 and @xmath214 , and we have condensed the notation by writing @xmath215 for @xmath216 , _ etc_. the combined signal - probe state written with the tensor products in the order assumed in eq .
( [ eq : joe ] ) is then @xmath217 ; \label{eq : entstate3}\ ] ] thus the joint probability of outcomes for the entangled signal state is @xmath218,p_{a0 } p_{b0 } ) = } \hspace{1.0 in } \nonumber \\[8pt ] & & \|\mathcal{n}[(1_{sa}\otimes e_a(j))\otimes(1_{sb}\otimes e_b(k))](u_a\otimes u_b)\nonumber \\[8pt ] & & \;\;[s_{a1}p_{a0}s_{b2}p_{b0 } + e^{i\theta } s_{a2}p_{a0 } s_{b1}p_{b0}]\|^2 .
\label{eq : joe1}\end{aligned}\ ] ] assuming the invariance up to phase of eq .
( [ eq : syma ] ) , the exchange of signal and probe parameters results only in phases that leave the probability unaffected , leading to the relation @xmath219,p_{a0 } p_{b0})}\quad\nonumber\\[8pt ] & = & \pr(j_a , j_b|s_{a0}s_{b0},\mathcal{n}[p_{a1}p_{b2}+ e^{i\theta}p_{a2}p_{b1 } ] ) , \end{aligned}\ ] ] so that outcome probabilities are the same for an entangled state measured by untangled detectors as they are for an unentangled state measured by entangled detectors . without assuming eqs .
( [ eq : symp ] ) , ( [ eq : syma ] ) , one can still ask : given a model @xmath31 that ascribes probabilities of outcomes to an entangled signal measured via an unentangled probe state @xmath220 , is there an alternative model @xmath38 that ascribes the same probabilities to an unentangled signal state measured via an entangled probe state ? we conjecture that the answer is `` yes , '' in which case no experiment can distinguish between a model of it that asserts entangled signal states measured via unentangled probes and a model that asserts unentangled signal states measured via entangled probes .
tai tsun wu called our attention to the conflict between the schrdinger equation and state reductions .
we thank howard e. brandt for discussions of quantum mechanics .
we thank dionisios margetis for analytic help .
this work was supported in part by the air force research laboratory and darpa under contract f30602 - 01-c-0170 with bbn technologies .
starting with eq . ( [ eq : pairphys ] ) , and writing @xmath10 as the time parameter times a dimensionless `` @xmath10 '' and @xmath147 and @xmath148 as the distance parameter times dimensionless `` @xmath147 '' and `` @xmath148 , '' respectively , we obtain @xmath221 this equation is solved by introducing a non - local coordinate change : @xmath222 with this change of variable , eq .
( [ eq : pair ] ) becomes @xmath223 for which separation of variables is immediate , so the general solution is a sum of products , each of the form @xmath224 the function @xmath203 satisfies its own schrdinger equation , @xmath225 which is the quantum - mechanical equation for an unstable harmonic oscillator , while @xmath226 satisfies @xmath227 which varies in its interpretation according to the value of @xmath175 , as follows : ( a ) for @xmath228 , it expresses an unstable harmonic oscillator ; ( b ) for @xmath229 , it expresses a free particle ; and ( c ) for @xmath176 , it expresses a stable harmonic oscillator .
the last case will be of interest when we compare behavior of the model with an experimental record . by translating eq .
( [ eq : initxy ] ) into @xmath188-coordinates , one obtains initial conditions @xmath230 the solution to eq .
( [ eq : phieq ] ) with these initial conditions is given by barton @xcite ; we deal with @xmath203 and @xmath226 in order . from ( 5.3 ) of ref .
@xcite , one has @xmath231 where @xmath232 .
\label{eq : bdef}\ ] ] similarly , integrating the green s function for the stable oscillator ( @xmath233 over the initial condition for @xmath226 yields @xmath234 where @xmath235 .
\label{eq : b2def}\ ] ] multiplying these and changing back to @xmath172-coordinates yield the joint probability density @xmath236 the probability of two detections disagreeing is the integral of this density over the second and fourth quadrants of the @xmath172-plane .
this is most conveniently carried out in @xmath188-coordinates .
for the especially interesting case of @xmath173 ( and @xmath233 , this integral can be transformed into @xmath237 & = & \frac{1}{\pi b_1(t ) b_2(t)}\int^{\infty}_{-\infty } dv \int^v_{-v}du\ , \exp\left\{\frac{-u^2}{b_1 ^ 2(t ) } - \frac{v^2}{b_2 ^ 2(t)}\right\ } \nonumber \\[6pt ] & = & \frac{4}{\pi}\int_0^{\infty}e^{-v^2}\ , dv\int_0^{b_2(t ) v / b_1(t)}du\ , e^{-u^2 } \nonumber \\[6pt ] \noalign{\goodbreak } & = & \frac{2}{\pi}\tan^{-1}\left(\frac{b_2(t)}{b_1(t)}\right ) \nonumber \\[6pt]\noalign{\goodbreak } & = & \frac{2}{\pi}\tan^{-1}\left ( \frac{\displaystyle 1 + \left[\frac{1}{b^4(\lambda-1)}-1\right]\sin^2\sqrt{\lambda-1}\ , t}{\displaystyle 1 + \left(\frac{1}{b^4}+1\right)\sinh^2 t}\right)^{1/2}.\qquad \label{eq : edge1}\end{aligned}\ ] ]
it is easy to check that this formula works not only when @xmath238 but also for the case @xmath228 . for @xmath239 ,
the numerator takes on the same form as the denominator , but with a slower growth with time , so that the probability of disagreement still decreases with time exponentially , but more slowly . converting eq .
( [ eq : jointxy ] ) from dimensionless back to physical time and distance variables results in eq .
( [ eq : jointxyphys ] ) with eqs .
( [ eq : bphysdef ] ) , and similarly eq .
( [ eq : edge1 ] ) leads to eq .
( [ eq : edge1phys ] ) .
j. m. myers and f. h. madjid , in : quantum computation and information , s. j. lomonaco , jr.and h. e. brandt ( eds . ) , contemporary mathematics series , vol .
305 , american mathematical society , providence , 2002 , pp . 221244 .
a. verevkin , g. n. goltsman , and r. sobolewski , in : opto - canada : spie regional meeting on optoelectronic , photonics , and imaging , technical digest of spie , vol .
td01 , spie , bellingham , wa , 2002 , pp.3940 . | although quantum states nicely express interference effects , outcomes of experimental trials show no states directly ; they indicate properties of probability distributions for outcomes .
we prove categorically that probability distributions leave open a choice of quantum states and operators and particles , resolvable only by a move beyond logic , which , inspired or not , can be characterized as a guess . by recognizing guesswork as inescapable in choosing quantum states and particles , we free up the use of particles as theoretical inventions by which to describe experiments with devices , and thereby replace the postulate of state reductions by a theorem . by using the freedom to invent probe particles in modeling light detection
, we develop a quantum model of the balancing of a light - induced force , with application to models and detecting devices by which to better distinguish one source of weak light from another .
finally , we uncover a symmetry between entangled states and entangled detectors , a dramatic example of how the judgment about what light state is generated by a source depends on choosing how to model the detector of that light .
and quantum mechanics , modeling , detection , metastability , agreement 03.65.ta , 03.65.nk , 84.30.sk |
materials with two order parameters ( op ) have long been known .
for example , superconductivity in stoichiometric ternary compounds errh@xmath11b@xmath11 and m@xmath12mo@xmath13s@xmath14 ( with m meaning ho , dy , or er and @xmath15 or @xmath16 ) coexists with a helical magnetic order .
@xcite in the last decades interest in the systems with two ops has increased drastically in connection with the discovery of high-@xmath17 superconductors ( see , e.g. , ) .
very recently , it has been experimentally established that , in cuprates , the superconducting op @xmath18 coexists with a state with a charge modulation ( see recent papers refs . and references therein ) . in another class of recently discovered superconductors the so - called fe - based pnictides
superconductivity may coexist with a spin density wave ( sdw ) ( for a review see refs . ) .
coexistence of ops of different types results in several interesting phenomena .
one can mention the enhancement of the london penetration depth@xcite or a peak in the specific heat jump@xcite at the doping level at which the sdw is formed , the peculiar dynamics of the ops ( see recent papers refs . and references therein ) .
nonhomogeneous states in systems with two ops are also very interesting and unusual . for example , a cdw arises in the center of vortices in cuprates.@xcite nonhomogeneous states in superconductors may arise even in the absence of magnetic field .
for example , fulde - ferrel - larkin - ovchinnikov ( fflo ) states may appear in superconductors in the presence of an exchange field@xcite and the so called amplitude solitons can be energetically favorable in conductors with cdw or sdw .
the latter have been predicted in refs . and observed in ref . in systems with a single op in quasi - one - dimensional conductors with a cdw .
amplitude solitons in systems with a cdw mean that the amplitude @xmath7 of the cdw drops to zero at some point , and a local energy level @xmath19 arises in the system.@xcite for example , if @xmath20 , the phase @xmath21 of @xmath1 changes from @xmath22 at @xmath23 to @xmath24 at @xmath25 .
the energy level @xmath19 may vary in time when a sufficiently high current passes through the system . in this case , the stationary state is unstable and one deals with a dynamical amplitude soliton with @xmath26.@xcite another case of a non - stationary ( moving ) soliton was studied in a recent work.@xcite in both cases , the structure of the amplitude soliton can be found analytically from a solution of microscopic equations.@xcite it is relevant to mention the stripes in high-@xmath17 superconductors@xcite that are higher dimensional relatives of the solutions of refs . .
fulde - ferrel - larkin - ovchinnikov states in superconductors coexisting with other ops , such as cdw or sdw , were studied recently in several works.@xcite in particular , it has been pointed out in ref . that the states similar to the fflo state are possible in superconductors competing with cdw or sdw .
note that another type of nonhomogeneous states in systems with two ops has been studied in refs .
using the so - called brazoskii - type model.@xcite in particular , it has been shown that a `` glassy '' phase may arise in these systems . however , the fflo - like state may arise , e.g. , in fe - based pnictides , only at low enough temperatures @xmath27 when the dependence of @xmath1 on the curvature @xmath4 is a multivalued function@xcite which is in a full analogy with superconductors with an exchange field @xmath28 .
in the latter case , the dependence @xmath29 at low @xmath27 is a multivalued function in a certain interval of @xmath28.@xcite spatial dependence of @xmath0 and @xmath1 can be described by a generalized eilenberger equation complemented by self - consistency relations@xcite but their analytical study is not simple . in this paper , we analyze nonuniform states of the ops on the basis of the ginzburg landau equations that are considerably more transparent than the eilenberger equation .
we are interested in nonuniform states corresponding to topological defects .
we concentrate here on one - dimensional structures and consider the dependence only on one coordinate @xmath30 .
this means that we consider a situation when the superconducting op @xmath0 changes its phase from @xmath2 to @xmath3 across this defect , while the amplitude of the cdw ( or sdw ) @xmath7 is localized at the center of the defect decaying to zero away from this point .
of course , the opposite situation is also possible , i.e. , when the function @xmath7 changes sign having opposite values at @xmath31 and @xmath32 and @xmath10 is a localized function . using a system of coupled ginzburg
landau equations we show that , while the superconducting op may vary in space as @xmath33 , the form of the amplitude @xmath7 depends on temperature or doping .
it is described by the gross pitaevskii equation resulting in a peculiar quantization of the solutions for @xmath7 .
the function @xmath34 corresponding to the ground state has the form of a soliton whereas the functions @xmath9 corresponding to excited states have nodes . in the opposite case , when a solution @xmath35 is brought about , the superconducting op @xmath0 is localized near the point @xmath36 and changes its form with variation of temperature in a rather unusual way . the nonuniform states considered in this paper may arise in the bulk , near the surface or in heterostructures consisting of materials with two ops . in the latter cases such states appear necessarily because of boundary or matching conditions imposed on the ops . note that the interesting and novel phenomena arising at the surface or in heterostructures are already known .
a new type of superconductivity ( triplet odd - frequency ) in superconductor / ferromagnet bilayer in the ferromagnet@xcite and the appearance of bound edge states with possible formation of majorana fermions at the surface of superconductors@xcite are remarkable examples of these phenomena .
we consider a model which is described by the g
l equations . as has been shown in ref .
, it is applicable to quasi - one - dimensional superconductors with a cdw and to two - band superconductors with an sdw .
the latter model has been developed in detail in refs .
for fe - based pnictides .
after certain modification , this model can be applied also to cuprates.@xcite on its basis , one can derive g
l equations for the ops @xmath0 and @xmath1 .
we neglect space variations of the phases of @xmath0 and @xmath1 and consider these ops as real quantities . to make the physical meaning of the coefficients in the g
l expansion more transparent , we write the g
l equations first for the case of superconductors with a cdw ( or sdw ) . in the notation of refs . ,
these equations have the form @xmath37 & = 0 \ , , \label{1 } \\ -\xi_{w}^{2 } \nabla^{2 } w + w \big [ \langle 2 \mu^{2 } s_{1 m } \rangle + w^{2 } s_{3 m } + \delta^{2 } s_{2 m } - \ln(t_{w}/t ) \big ] & = 0 \ , , \label{1'}\end{aligned}\ ] ] where @xmath38 are the coherence lengths ( at low temperatures ) for @xmath0 and @xmath1 , respectively , and @xmath39 are , respectively , the critical temperatures for the transition into the pure superconducting state or into a state with a cdw or an sdw only . in other words ,
@xmath40 is the critical temperature for the transition into the charge - ordered state in absence of @xmath0 and @xmath4 , while @xmath41 is the superconducting transition temperature in absence of @xmath1 .
the angle brackets mean the angle averaging ( in fe - based pnictides ) or integration along the sheets of the fermi surfaces in quasi - one - dimensional superconductors .
the functions @xmath42 , @xmath43 , etc .
, are functions of the normalized curvature ( see appendix ) @xmath44 and @xmath45}$ ] is a curvature in quasi - one - dimensional superconductors with a doping - dependent value of @xmath46 .
it is assumed that the fermi surface of these superconductors consists of two slightly curved sheets which are perpendicular to the @xmath30 axis.@xcite in the case of fe - based pnictides , @xmath47 is a quantity that describes an elliptic ( @xmath48 ) and circular ( @xmath49 ) fermi surfaces of electron and hole bands.@xcite all quantities@xmath0 , @xmath1 and @xmath50are measured in units of @xmath51 .
the expressions for the coefficients in the g
l expansion with account for impurity scattering have been calculated in ref .
( see also ref . )
. replacing the derivative @xmath52 , one can use eqs . and to describe vortices in superconductors with a cdw @xcite , where @xmath53 is the magnetic flux quantum and @xmath54the modulus of the vector potential of a magnetic field . as it is seen from eq .
, the critical temperature @xmath40 depends on doping , i.e. , on the parameter @xmath4 .
we choose this parameter @xmath55 in such a way that @xmath56 .
this means that at @xmath57 , the quantities @xmath58 , and , thus , @xmath59 obeys the equation @xmath60 where @xmath61 and @xmath59 is a function of two parameters , i.e. , @xmath62 .
then , we expand the function @xmath63 in the deviations @xmath64 = \mu^{2 } - \mu_{c}^{2}}$ ] and @xmath65 , thus obtaining @xmath66 \rangle}$ ] , and use eq .
( [ 2 ] ) to obtain equations in a general standard form ( assuming that all the functions depend only on one coordinate @xmath30 ) , @xmath67 & = 0 \ , , \label{3 } \\
-\xi_{w}^{2 } w^{\prime \prime } + w \big [ -a_{w } + b_{w } w^{2 } + \gamma \delta^{2 } \big ] & = 0 \ , , \label{3'}\end{aligned}\ ] ] with @xmath68 and @xmath69 as well as @xmath70 and @xmath71 denoting the first and second derivatives with respect to @xmath30 , respectively .
these equations determine extrema of the free energy functional @xmath72 with respect to @xmath0 and @xmath1 , and the corresponding coefficients of the g
l expansion are related to variables in eqs .
( [ 1 ] ) and ( [ 1 ] ) via @xmath73 , @xmath74 , @xmath75 \rangle } $ ] , @xmath76 , @xmath77 , where @xmath78 . the expressions for @xmath79 are given in the appendix .
the coupled g
( [ 3 ] ) and ( [ 3 ] ) are , of course , not new and have been used long time ago in , e.g. , ref .
@xcite for studying competition between superconductivity and cdw in the presence of disorder or commensurability .
our aim now is to find new non - trivial inhomogeneous solutions of eqs .
( [ 3 ] ) and ( [ 3 ] ) . for simplicity
, we consider the case when the last term in eq .
( [ 3 ] ) can be neglected , which is legitimate when the coupling constant @xmath80 or a small amplitude @xmath1 is small ( we will see that at temperatures @xmath27 or doping level @xmath4 near some critical values @xmath81 and @xmath82 the amplitude @xmath1 is indeed small ) . in the zero - order approximation
we obtain for @xmath10 @xmath83 = 0\ , .
\label{5}\ ] ] equation ( [ 5 ] ) has the well - known nonuniform solution ( see for example ref . ) , @xmath84 where @xmath85 and @xmath86 .
this equation describes , for instance , the behavior of @xmath10 in a vicinity of s / n interface at the superconductor side , where n is a normal metal with a strong depairing .
we consider this solution in an infinite superconductor
. substituting this expression into eq .
( [ 3 ] ) , we obtain an equation for the amplitude of the cdw or sdw @xmath87 = g w^{3 } \ , , \label{7}\ ] ] where @xmath88 , @xmath89 , @xmath90 , and @xmath91 .
these quantities may be written in notations used for quasi - one - dimensional supercondcutors and fe - based pnictides as @xmath92}$ ] , @xmath93 , @xmath94 .
equation for spatial variation of the cdw amplitude @xmath1 has a form of the well known gross
pitaevskii equation.@xcite solutions of this equation can be written rather easily in limiting cases .
we consider the simplest situation when the rhs of eq .
( [ 7 ] ) is small , i.e. , @xmath95 .
we are interested in solutions with @xmath10 given by eq .
( [ 6 ] ) and @xmath7 decaying to zero at @xmath96 .
in particular , the solution for @xmath7 may have the form of a soliton .
such a state with a finite @xmath0 and zero @xmath1 at infinity is stable if the conditions @xmath97 and @xmath98 at @xmath99 and @xmath100 are satisfied .
one can see that @xmath101 is always positive and @xmath102 is positive if the quantity @xmath103 is negative .
we will see that just at negative @xmath103 , eq .
( [ 7 ] ) has a solution in the form of a soliton . in zero - order approximation
we obtain for @xmath8 @xmath104 = 0 \ , .
\label{8}\ ] ] this equation is integrable and its solutions @xmath105 corresponding to a discrete spectrum of @xmath106 are expressed in terms of hypergeometric functions.@xcite in our notations , the `` energy '' levels of discrete spectrum are given by@xcite @xmath107^{2 } \
, , \label{9'}\ ] ] and their maximal number @xmath108 is determined by @xmath109 .
note that in fe - based pnictides , @xmath110 in the ballistic case .
we expand the correction @xmath111 to the zero - order solution @xmath8 in terms of the normalized eigenfunctions @xmath105 of the operator @xmath112 .
these functions obey the equation @xmath113 solutions of eq .
( [ 7 ] ) can be written explicitly if the quantity @xmath114)}$ ] is close to a certain `` energy '' level @xmath106 , say to @xmath115 , such that @xmath116)}$ ] ( in the language of the original electronic model , the `` temperature '' @xmath117 or doping @xmath118 $ ] should be chosen properly ) .
we write eq .
( [ 7 ] ) in the form @xmath119 with @xmath120 and represent @xmath1 as @xmath121 , where @xmath122 , and the summation runs over all @xmath123 except the term @xmath124 .
we substitute this @xmath7 into eq .
( [ 12 ] ) and multiply this equation first by @xmath125 and then by @xmath105 with @xmath126 , then integrating the obtained result each time over @xmath30 .
thus , taking into account the orthogonality of different eigenfunctions , we find the coefficients @xmath127 @xmath128 where @xmath129 , where the double angle brackets are used to distinguish this operation from the averaging over the angles introduced in the appendix . obviously , in eq .
( [ 15 ] ) , @xmath105 and @xmath125 have to have same parity ( both even or both odd ) .
the obtained expressions are valid provided the condition @xmath130 is satisfied .
this condition means that if the `` temperature '' @xmath117 or doping @xmath118 $ ] is chosen in such a way that the quantity @xmath131)$ ] is close to @xmath115 , i.e. , the difference on the lhs of this condition is smaller than the difference between any energy level @xmath106 and @xmath115 , the spatial dependence of @xmath7 is given by the leading order while the second term , @xmath132 , gives a small correction . since we assumed that the rhs of eq .
( [ 7 ] ) is small compared to the term @xmath133 , the condition @xmath134 should be also satisfied . the ground state is realized if at some `` temperature '' @xmath135 the quantity @xmath136 is close to @xmath137 . in this case
, @xmath34 has the form of a soliton .
if @xmath138 is close to @xmath139 , the amplitude of the cdw is an odd function of @xmath30 . for the ground state ,
( [ 9 ] ) yields @xmath140 b_{2 } \big\rangle } { \eta_{0 } } \bigg ) = s_{3 } \frac{s_{3m}}{8 r^{2 } } \bigg [ \sqrt { 1 + \frac{8 r^{2 } s_{2m}}{s_{3 m } s_{3 } } } - 1 \bigg]^{2 } \ , , \label{17}\ ] ]
where @xmath141 and @xmath142 . at a given @xmath64 = \mu^{2 } - \mu_{c}^{2}}$
] , that can be both positive and negative , this equation determines the `` temperature '' @xmath143 at which the solution of the gross pitaevskii equation has a soliton - like solution @xmath8 .
similarly , setting @xmath144 in eq .
( [ 9 ] ) , one can find a `` temperature '' @xmath145 corresponding to the first excited state with an odd function of the op @xmath146 etc .
in fig . [
fig : w_n ] we plot the spatial dependence of the cdw amplitude @xmath7 for @xmath147 , @xmath148 , @xmath149 and @xmath150 .
note that if @xmath151 , only a single soliton - like solution exists . in eq .
( [ eq : eta_on_mu ] ) on @xmath152 for @xmath153 .
the critical doping @xmath152 is calculated from eq .
( [ 2 ] ) and represents a line in the @xmath154 plane , which is shown in the upper part of the figure . in the lower part , the value of @xmath155 along the obtained @xmath152-line is presented as function of @xmath156 ( inserting the corresponding value of @xmath157 ) .
the coefficient @xmath155 is negative , thus , as @xmath158 due to conditions of realization of the superconducting state far from the topological defect , from eq .
( [ eq : eta_on_mu ] ) it is seen that @xmath159 < 0}$ ] . ] as an example , we calculate for the ground state the dependence of @xmath160 on @xmath161 $ ] .
it follows from eq .
( [ 17 ] ) that , assuming @xmath161 $ ] independent on @xmath162 , @xmath163 \ , , \label{eq : eta_on_mu}\ ] ] where the coefficient @xmath155 is given by @xmath164^{-1 } \,.\ ] ] the coefficient @xmath155 depends on @xmath152 defined by eq .
( [ 2 ] ) .
it is negative and , thus , since @xmath158 , @xmath161 $ ] should also be negative .
we plot the dependence of @xmath155 on @xmath152 in fig .
[ fig : eta_on_mu ] .
more precisely , the critical doping @xmath152 is calculated from eq .
( [ 2 ] ) and represents a line in the @xmath154 plane ( the upper part of fig .
[ fig : eta_on_mu ] ) . projecting this line onto the @xmath156 axis and inserting the corresponding values of @xmath157 we obtain the plot of @xmath165 presented in the lower part of fig .
[ fig : eta_on_mu ] .
consider the temperature interval where the soliton - like solution for @xmath7 exists . as follows from eq .
( [ 15 ] ) , the difference @xmath166 must be positive if the constant @xmath90 is positive .
this implies that the difference @xmath167 has to be positive as well ( at @xmath168 ) .
therefore , at @xmath169 , no @xmath1 appears at the topological defect , but at @xmath170 , a soliton - like solution for @xmath7 arises with the amplitude @xmath171 . on the other hand , as follows from eq .
( [ 17 ] ) , the temperature @xmath172 is less than the temperature @xmath173 , where @xmath174}$ ] .
this means that the soliton - like solution for @xmath1 as well as solutions corresponding to excited states exist in the interval @xmath175 ( red ) , @xmath176 ( black ) , @xmath177 ( green ) and @xmath178 ( blue ) near the corresponding `` energy '' levels @xmath179 for the case when the superconducting state is favored far from the defect at @xmath36 .
exactly at @xmath179 , as follows from eqs .
( [ 15 ] ) and ( [ 15 ] ) , @xmath100 .
note that in the opposite case when the cdw or the sdw state is more favorable at @xmath180 , one needs to make the exchange @xmath181 and , correspondingly , @xmath182 and @xmath183 , and the shown curves will describe the dependence @xmath10 while @xmath184 . ]
the solutions found above are valid if the free energy @xmath185 of the superconducting state at @xmath96 is lower than @xmath186 for a state with @xmath187 .
this is possible if @xmath188 < 0}$ ] .
this condition determines a temperature interval in which our considerations are valid . if the difference @xmath189 is positive , then the same procedure of finding solutions of g
l equations can be repeated with an exchange @xmath181 adapting correspondingly @xmath106 and other quantities . in particular , @xmath184 , with @xmath190 and @xmath191 , and the superconducting op @xmath10 is expressed in terms of hypergeometric functions , i.e. , it is localized at @xmath36 .
consider , for example , an n / s system where in the superconductor s there exists not only the superconducting op @xmath0 , but also a density wave @xmath1 , and n is a normal metal with a strong suppression of @xmath1 ( for example , with a strong interband impurity scattering which suppresses the op @xmath1 @xcite ) .
then , at the s - side , the dependence @xmath7 is determined by the above written expression , and at a certain temperature which may be even higher than @xmath41 , at the n / s interface superconductivity may arise spreading over a distance @xmath192 from the interface .
note that the found nonhomogeneous solutions for @xmath10 and @xmath7 are energetically favorable in comparison with uniform solutions , @xmath193 and @xmath194 , provided the energy loss [ due to the gradient of @xmath10 ] @xmath196 \sim \delta_{\infty}^{4 } \xi_{s } / \sqrt{a_{s}}}$ ] is less than the energy gain @xmath197 ( due to the appearance of @xmath1 ) @xmath198 .
this can not occur in the considered case of small @xmath1 . however , in heterostructures , like an n / s system , the solution eq .
( [ 6 ] ) ( at @xmath199 ) is dictated by a boundary condition in case of strong depairing in the n metal and , therefore , there is no energy loss in the superconducting part of the free energy .
thus , the considered states may be realized in heterostructures .
the case of uniform superconductors with a not small op @xmath1 requires a separate consideration .
on the basis of ginzburg
landau equations we studied a possibility of nonhomogeneous states in systems with two ops .
materials , where the superconducting op @xmath0 and the op @xmath1 related to a cdw ( or an sdw ) may exist , belong to this class of systems .
in the situation when the superconducting state is more favorable , the ginzburg
landau equations have nonhomogeneous solutions which describe @xmath10 in the form of a topological defect , eq .
( [ 6 ] ) , and @xmath7in the form of a function localized near the center of the defect , @xmath36 . the form of @xmath7 is described by the gross pitaevskii equation and depends essentially on the proximity of the function @xmath131)$ ] to the eigenvalues @xmath115 of the linearized gross pitaevskii equation .
if @xmath200 ) = \mathcal{e}_{n}}$ ] at some temperature @xmath201 and doping @xmath202 $ ] , then the amplitude of the function @xmath7 turns to zero and increases as @xmath203)|}}$ ] when @xmath117 or @xmath118 $ ] deviate from @xmath204 and @xmath202 $ ] . at a given temperature @xmath27 in the interval eq .
( [ 17a ] ) , there are , generally speaking , several solutions for @xmath7 .
the most stable one is the solution which corresponds to the ground state ( soliton - like solution ) .
therefore , in the equilibrium case one can observe only this solution for @xmath7 .
other solutions may affect the response of the system to the influence of fluctuations or of external perturbations . on the other hand ,
if the state with @xmath187 and @xmath205 corresponds to a minimum of the free energy , then nonhomogeneous solutions are possible with @xmath7 determined by eq .
( [ 6 ] ) ( correspondingly adapted as @xmath206 , @xmath207 ) and @xmath10 is localized near the point @xmath36 . in principle
, such solutions may arise in the bulk ( especially near some defects ) and in heterostructures of type n / n@xmath208 , where n@xmath208 is a material under consideration in which @xmath0 and/or @xmath1 may exist , and n is a material with a strong depairing towards the ops @xmath0 and @xmath1 .
for example , in an n / n@xmath209 heterostructure , the op in the vicinity of the interface has inevitably the form of eq .
( [ 6 ] ) and a localized @xmath210 arises at the interface . in this case , one deals with a localized interfacial superconductivity .
this type of superconductivity has been studied very actively in recent years and has been observed in different materials including cuprates and fe - based pnictides ( see recent papers refs . and references therein ) .
several proposals have been made to explain this phenomenon , but most experimental observations remain unexplained .
the mechanism considered here may be responsible for interfacial superconductivity in systems with two ops , but applicability of this mechanism to real materials deserves a separate consideration .
we appreciate the financial support from the dfg via the projekt ef 11/8 - 1 ; k. b. e. gratefully acknowledges the financial support of the ministry of education and science of the russian federation in the framework of increase competitiveness program of nust `` misis '' ( nr . k2 - 2014 - 015 ) .
the free energy has the form ( see also refs . ) @xmath211 where @xmath212 , @xmath213 and @xmath214 is the matsubara frequency with a cut - off @xmath215 ; @xmath216 and @xmath217 are the interaction constants of the superconductivity and spin- or charge - density wave , respectively . expanding this expression in @xmath0 and @xmath1 and performing variation with respect to these variables , we come to eqs .
( [ 1 ] ) and ( [ 1 ] ) with the coefficients defined as @xmath218^{-1 } \ , , \\ s_{2 m } & = \sum_{n=0}^{\infty } \big\langle \big [ ( 2 n + 1)^2 - m^2 \big ] ( 2 n + 1)^{-1 } \big [ ( 2 n + 1)^2 + m^2 \big]^{-2 } \big\rangle \ , , \\ s_{3 m } & = \sum_{n=0}^{\infty } \big\langle ( 2 n + 1 ) \big [ ( 2 n + 1)^2 - 3 m^2 \big ] \big [ ( 2
n + 1)^2 + m^2 \big]^{-3 } \big\rangle \ , , \\
\beta_1 & = \sum_{n=0}^{\infty } \big\langle 4 m^2 ( 2 n + 1 ) \big [ ( 2 n + 1)^2 + m^2 \big]^{-2 } \big\rangle \ , , \\
\beta_2 & = \sum_{n=0}^{\infty } 2 ( 2
n + 1)^{-1 } \big [ ( 2 n + 1)^2 + m^2 \big]^{-1 } \,,\end{aligned}\ ] ] where @xmath219 and the angle brackets denote the angle averaging ( in fe - based pnictides ) or integration along the sheets of the fermi surfaces in quasi - one - dimensional superconductors .
74ifxundefined [ 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 10.1007/bf00115620 [ * * , ( ) ] link:\doibase 10.1080/00018738500101741 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.75.1201 [ * * , ( ) ] link:\doibase 10.1080/00018730903122242 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.075128 [ * * , ( ) ] link:\doibase 10.1038/nphys2641 [ * * , ( ) ] link:\doibase 10.1126/science.1246310 [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.195115 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.035149 [ * * , ( ) ] @noop ( ) , link:\doibase 10.1126/science.1223532 [ * * , ( ) ] link:\doibase 10.1038/nphys2456 [ * * , ( ) ] link:\doibase 10.1126/science.1243479 [ * * , ( ) ] link:\doibase 10.1126/science.1242996 [ * * , ( ) ] link:\doibase 10.1038/nature10345 [ * * , ( ) ] link:\doibase 10.1038/ncomms3113 [ * * , ( ) ] link:\doibase 10.1038/nphys2502 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.054506 [ * * , ( ) ] link:\doibase 10.1038/nphys2805 [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.224513 [ * * , ( ) ] http://stacks.iop.org/0034-4885/74/i=12/a=124508 [ * * , ( ) ] link:\doibase 10.1146/annurev - conmatphys-020911 - 125055 [ * * , ( ) ] link:\doibase 10.1126/science.1219821 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.177003 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.111.157004 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.257002 [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.144503 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.024506 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.024511 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.054511 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrev.135.a550 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.42.1698 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.108.096801 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.60.781 [ * * , ( ) ] @noop * * , ( ) @noop ( ) , link:\doibase 10.1103/physrevb.67.125420 [ * * , ( ) ] link:\doibase 10.1103/physrevb.40.7391 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/0921-4534(89)90316-x [ * * , ( ) ] link:\doibase 10.1103/physrevb.39.9749 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.80.1272 [ * * , ( ) ] link:\doibase 10.1038/375561a0 [ * * , ( ) ] link:\doibase 10.1038/nphys999 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.020510 [ * * , ( ) ] link:\doibase 10.1103/physrevb.69.014208 [ * * , ( ) ] link:\doibase 10.1103/physrevb.70.024207 [ * * , ( ) ] link:\doibase 10.1103/physrevb.79.165122 [ * * , ( ) ] @noop * * , ( ) http://stacks.iop.org/0953-2048/23/i=5/a=054011 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.174538 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.014521 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.134524 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.86.4096 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.77.1321 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.82.3045 [ * * , ( ) ] link:\doibase 10.1143/jpsj.81.011013 [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.214504 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.214521 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.020506 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop _ _ , advanced books classics series ( , , ) link:\doibase 10.1007/bf02731494 [ * * , ( ) ] @noop * * , ( ) @noop _ _ , ed . , , vol .
( , , ) link:\doibase 10.1103/physrevb.79.140507 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.140512 [ * * , ( ) ] link:\doibase 10.1038/nature12494 [ * * , ( ) ] link:\doibase 10.1038/501474a [ * * , ( ) ] link:\doibase 10.1038/nmat3719 [ * * , ( ) ] | on the basis of coupled ginzburg
landau equations we study nonhomogeneous states in systems with two order parameters ( op ) . superconductors with superconducting op @xmath0 , and charge- or spin - density wave ( cdw or sdw ) with amplitude @xmath1 are examples of such systems .
when one of op , say @xmath0 , has a form of a topological defect , like , e.g. , vortex or domain wall between the domains with the phases @xmath2 and @xmath3 , the other op @xmath1 is determined by the gross pitaevskii equation and is localized at the center of the defect .
we consider in detail the domain wall defect for @xmath0 and show that the shape of the associated solution for @xmath1 depends on temperature and doping ( or on the curvature of the fermi surface ) @xmath4 .
it turns out that , provided temperature or doping level are close to some discrete values @xmath5 and @xmath6 , the spacial dependence of the function @xmath7 is determined by the form of the eigenfunctions of the linearized gross pitaevskii equation .
the spacial dependence of @xmath8 corresponding to the ground state has the form of a soliton , while other possible solutions @xmath9 have nodes .
inverse situation when @xmath7 has the form of a topological defect and @xmath10 is localized at the center of this defect is also possible . in particular
, we predict a surface or interfacial superconductivity in a system where a superconductor is in contact with a material that suppresses @xmath1 .
this superconductivity should have rather unusual temperature dependence existing only in certain intervals of temperature .
possible experimental realizations of such non - homogeneous states of ops are discussed . |
in the usual models of genome rearrangement , duplications and losses of genes are not taken into account .
there were attempts to incorporate them to the classical models , but the consecutive combinatorial complexity of the models so obtained made their study quite difficult .
following @xcite , we focus on the duplication - loss problem by considering the _ tandem duplication - random loss model _ of genome rearrangement in which genomes are modified _ only _ by duplications and losses of genes . one _ step _ of tandem duplication - random loss , or duplication - loss for short , consists in @xmath13 the tandem duplication of a contiguous fragment of the genome , _
i.e. , _ the duplicated fragment is inserted immediately after the original fragment , and @xmath14 the loss of one of the two copies of every duplicated gene .
we assume that the loss occurs immediately after the duplication of genes , which is , on an evolutionary time - scale , a good approximation to reality .
the _ width _ of a step is the number of duplicated genes .
see figure [ ex : duplication - loss ] for an example .
@xmath15 from a formal point of view , a genome consisting of @xmath12 genes is modelled by a permutation @xmath16 of the set of integers @xmath17 . in @xcite ,
the authors define the cost of a duplication - loss step of width @xmath0 to be @xmath1 , @xmath18 being a constant parameter .
they suggest that other cost functions can be considered , and in particular affine functions . in this paper
, we consider a _ piecewise constant _ cost function : the cost of a step of width @xmath0 is @xmath19 if @xmath4 and is infinite for @xmath6 , for some fixed parameter @xmath7 .
obviously , for this model to be meaningful , we assume that @xmath20 .
we also consider the possibility that @xmath21 is dependent on the size @xmath12 of the permutation on which the duplication - loss operations are performed .
both models are generalizations of the _ whole genome duplication - random loss model _ : it corresponds to the case @xmath22 in the model of @xcite , @xmath23 or @xmath24 in our model .
many models of evolution of permutations are inspired by computational biology issues : see @xcite , @xcite , @xcite , @xcite for examples in the literature .
our model of evolution of permutations can be viewed in the framework of _ permuting machines _ defined in @xcite .
such a machine takes a permutation in input , and transforms it into an output permutation , the transformation being subject to satisfy the two properties of independence with respect to the values and of stability with respect to pattern - involvement ( see @xcite for more details ) .
the important point is that the duplication - loss transformation satisfies these two properties .
thus , one duplication - loss step ( in one of the models defined above ) corresponds to running an adequate permuting machine once .
when we will consider permutations obtained after a sequence of duplication - loss steps , it will correspond to permutations obtained in the output of a combination in series of identical permuting machines .
for ease of exposition in some proofs , we will sometimes use a graphical representation of permutations , as shown in figure [ fig : graphical - representation ] .
( -4,0)(11,11 ) ( 0,0)(8,8 ) ( 0.5,-0.5)@xmath19 ( 1.5,-0.5)@xmath25 ( 2.5,-0.5)@xmath26 ( 3.5,-0.5)@xmath27 ( 4.5,-0.5)@xmath28 ( 5.5,-0.5)@xmath29 ( 6.5,-0.5)@xmath30 ( 7.5,-0.5)@xmath31 ( -0.5,0.5)@xmath19 ( -0.5,1.5)@xmath25 ( -0.5,2.5)@xmath26 ( -0.5,3.5)@xmath27 ( -0.5,4.5)@xmath28 ( -0.5,5.5)@xmath29 ( -0.5,6.5)@xmath30 ( -0.5,7.5)@xmath31 ( 0.5,5.5)0.3 ( 1.5,7.5)0.3 ( 2.5,0.5)0.3 ( 3.5,2.5)0.3 ( 4.5,4.5)0.3 ( 5.5,3.5)0.3 ( 6.5,1.5)0.3 ( 7.5,6.5)0.3 though not appearing clearly for the moment , there exist strong links between the duplication - loss model and some pattern - avoiding classes of permutations .
hence , we need to recall a few definitions concerning those classes .
a permutation @xmath32 is a bijective map from @xmath33 $ ] to itself .
the integer @xmath12 is called the _ size _ of @xmath34 , denoted @xmath35 .
we denote by @xmath36 the image of @xmath37 under @xmath34 .
a permutation can be seen as a word @xmath38 containing exactly once each letter @xmath39 $ ] . for each entry @xmath36 of a permutation @xmath34 , we call @xmath37 its _ position _ and @xmath36 its
_ value_. a permutation @xmath40 is a _ pattern _ of a permutation @xmath32 if there is a subsequence of @xmath34 which is order - isomorphic to @xmath41 ; in other words , if there is a subsequence @xmath42 of @xmath34 ( with @xmath43 ) such that @xmath44 whenever @xmath45 .
+ we also say that @xmath41 is _ involved _ in @xmath34 and call @xmath42 an _ occurrence _ of @xmath41 in @xmath34 .
[ def : pattern ] we write @xmath46 to denote that @xmath41 is a pattern of @xmath34 . a permutation @xmath34 that does not contain @xmath41 as a pattern is said to _
avoid _ @xmath41 . the class of all permutations avoiding the patterns @xmath47 is denoted @xmath48 , and @xmath49 denotes the set of permutations of size @xmath12 avoiding @xmath50 .
we say that @xmath48 is a class of pattern - avoiding permutations of _ basis _ @xmath51 .
for example @xmath52 contains the pattern @xmath53 , and @xmath54 , @xmath55 , @xmath56 and @xmath57 are the occurrences of this pattern in @xmath34 . but @xmath58 : @xmath34 avoids the pattern @xmath59 as no subsequence of size @xmath26 of @xmath34 is isomorphic to @xmath59 , _
i.e. _ , is decreasing . [
ex : pattern ] in the tandem duplication - random loss model described above , we will focus on two kinds of problems .
first , as hinted before , we will consider permutations obtained after a certain number of duplication - loss steps , that is to say permutations in output of a combination in series of a certain number of permuting machines .
for this , we define the class @xmath60 as follows : the class @xmath60 denotes the class of all permutations obtained from @xmath61 ( for any @xmath12 ) after @xmath8 duplication - loss steps of width at most @xmath9 , for some constant parameters @xmath8 and @xmath9 .
we do not consider the case @xmath11 here .
be careful that the duplication - loss steps are not reversible , as noticed in @xcite , and that consequently @xmath60 is _ not _ the class of permutations that can be _ sorted to _
@xmath61 in @xmath8 steps of duplication - loss of width at most @xmath9 . like for the various classes of permutations obtained after a combination in series of permuting machines considered in @xcite
, we obtained combinatorial properties of @xmath60 in terms of pattern - avoidance .
namely , we show that @xmath60 is a class of pattern - avoiding permutations . in the case
@xmath62 ( section [ section : one - step ] ) , we give a precise description of the basis @xmath63 of excluded patterns : @xmath64 , @xmath65 being the set of all permutations of @xmath66 that do not start with @xmath19 nor end with @xmath67 , and containing exactly one descent . in particular , @xmath63 is of cardinality @xmath68 and contains patterns of size at most @xmath67 . for the general case ( section [ section : p - steps ] ) , we can not get such a precise result but only a bound on the size of the excluded patterns : we show that @xmath60 is a class of pattern - avoiding permutations whose basis contains patterns of size at most @xmath69 .
a second point of view is to examine how many steps of a given width are necessary to obtain any permutation of @xmath10 starting from @xmath61 .
namely in section [ section : algo ] we fix a width @xmath9 ( constant , or @xmath11 ) and a size @xmath12 and search for the number @xmath8 such that any permutation of @xmath10 can be obtained from @xmath61 in at most @xmath8 duplication - loss steps of width at most @xmath9 .
we describe an algorithm computing a possible scenario of duplications and losses for any @xmath16 , this scenario involving @xmath70 duplication - loss steps in the worst case and on average .
we also show that @xmath71 steps are necessary ( in the worst case and on average ) to obtain any permutation of @xmath10 from @xmath61 .
these upper and lower bounds coincide in most cases .
before focusing on the classes @xmath72 and @xmath60 defined for our model , we will get back to the simpler whole genome duplication - random loss model ( corresponding to @xmath73 in our model , but defined previously by other authors )
. we will not prove new theorems , but will interprete the existing results from the pattern - avoidance point of view .
let us recall that in the whole genome duplication - random loss model , any duplication - loss step has cost @xmath19 , so that we can consider _
w.l.o.g _ that the duplicated fragment is the whole permutation at any step .
the cost of obtaining a permutation @xmath32 from the identity is just the minimal number of steps of a duplication - loss scenario transforming @xmath61 into @xmath34 .
a statistics of permutations that matters for our purpose is their number of _
descents_. given a permutation @xmath34 of size @xmath12 , we say that there is a _ descent _
_ ascent _ ) at position @xmath37 , @xmath74 , if @xmath75 ( resp .
@xmath76 ) .
we write @xmath77 the number of descents of the permutation @xmath34 .
[ def : descent ] for example , @xmath78 has @xmath26 descents , namely at positions @xmath19 , @xmath26 and @xmath27 .
[ ex : descent ] a permutation @xmath34 of size @xmath12 has at most @xmath79 descents , the case of @xmath79 descents exactly corresponding to the reversed identity permutation @xmath80 .
it is also of common knowledge that the average number of descents among permutations of size @xmath12 is @xmath81 . in @xcite ,
the authors prove the following theorem .
let @xmath32 . in the whole genome duplication - random loss model ,
@xmath82 steps are necessary and sufficient to obtain @xmath34 from @xmath61 .
[ thm : whole - genome ] it is equivalent to say that the permutations that can be obtained in at most @xmath8 steps in the whole genome duplication - random loss model are exactly those whose number of descents is at most @xmath83 .
now , we can notice that the property of being obtainable in at most @xmath8 steps is stable for the pattern - involvement relation @xmath84 : if @xmath34 can be obtained in at most @xmath8 steps , and if @xmath46 , then @xmath41 can also be obtained in at most @xmath8 steps .
indeed , it is enough to perform the same duplication - loss scenario on @xmath34 , keeping track only of the elements of @xmath34 that form an occurrence of @xmath41 .
this stability for @xmath84 implies that the class of permutations obtainable in at most @xmath8 steps is a class of pattern - avoiding permutations , whose excluded patterns are the minimal ( again in the sense of @xmath84 ) permutations that can not be obtained in @xmath8 steps .
then , by theorem [ thm : whole - genome ] , the excluded patterns are the minimal permutations with @xmath85 descents .
we initiated a study of the minimal permutations with @xmath86 descents in @xcite .
however , it is simple to notice that a permutation with @xmath86 descents and minimal for this criterion has size at most @xmath87 , since it does not contain to consecutive ascents by minimality .
an immediate consequence is that the number of excluded patterns is finite .
this allows us to state the following version of theorem [ thm : whole - genome ] : the permutations that can be obtained in at most @xmath8 steps in the whole genome duplication - random loss model form a class of pattern - avoiding permutations .
the excluded patterns are the permutations with exactly @xmath85 descents that are minimal ( in the sense of @xmath84 ) for this criterion .
these excluded patterns are in finite number . in @xcite , we will give a simpler description and some properties of these minimal permutations with @xmath86 descents .
as an introduction to the study of @xmath60 , we deal in this section with the simpler case of the class @xmath88 of permutations obtained from @xmath89 in one duplication - loss step of width at most @xmath9 .
assume in this section that the parameter @xmath20 is fixed . throughout this section ,
when referring to a duplication - loss step , we always mean duplication - loss step of width @xmath9 , except when otherwise explicitly stated .
it is easily noticed that any permutation of @xmath90 can not have more than one descent .
conversly , any permutation of size at most @xmath9 having exactly one descent belongs to @xmath90 .
although it is a technical point of importance in the proof of theorem [ thm : one - step ] , the following proposition comes straightforward : the permutations of size @xmath67 that do not belong to @xmath90 and having exactly one descent are exactly those of @xmath66 with one descent that do not start with @xmath19 nor end with @xmath67 .
[ prop : one - step - sizek+1 ] let @xmath91 be a permutation of size @xmath67 that does not belong to @xmath90 but has exactly one descent .
now , if @xmath92 , then @xmath93 is a permutation ( of @xmath94 ) of size @xmath9 having one descent , and therefore @xmath95 can be obtained from @xmath96 in one duplication - loss step .
applying the same transformation to @xmath97 will then produce @xmath34 , contradicting that @xmath98 .
the same reasoning holds when @xmath99 .
so @xmath34 does not start with @xmath19 nor end with @xmath67 .
now if @xmath34 is a permutation of size @xmath67 having exactly one descent , that does not start with @xmath19 nor end with @xmath67 , we claim that @xmath34 can not be obtained from @xmath100 in one duplication - loss step .
this is because no duplication - loss step of width @xmath9 can move both @xmath19 and @xmath67 in @xmath100 .
the class @xmath90 of permutations obtained from @xmath101 ( for some @xmath102 ) in one duplication - loss step of width @xmath9 is a class @xmath103 of pattern - avoiding permutations whose basis @xmath63 is composed of @xmath68 patterns of size at most @xmath67 .
namely @xmath64 , @xmath65 being the set of all permutations of @xmath66 that do not start with @xmath19 nor end with @xmath67 , and containing exactly one descent .
[ thm : one - step ] @xmath104 + @xmath105 [ ex : thm - one - step ] we prove the reversed statement : @xmath106 if and only if @xmath34 can not be obtained from an identity permutation in one duplication - loss step of width @xmath9 .
assume @xmath106 .
then there exists @xmath107 such that @xmath108 . if @xmath109 , @xmath110 or @xmath111 , then @xmath34 has at least @xmath25 descents and can not be obtained in one duplication - loss step
otherwise , using proposition [ prop : one - step - sizek+1 ] , there exists @xmath112 such that @xmath113 and @xmath114 .
now if @xmath34 could be obtained in one duplication - loss step , then so would be @xmath115 , yielding a contradiction .
so @xmath98 .
conversly , assume that @xmath98 .
if @xmath34 contains at least @xmath25 descents , then @xmath34 contains an occurrence of @xmath116 or @xmath110 or @xmath111 , since these three are the minimal permutations ( in the sense of the relation @xmath84 ) with @xmath25 descents . and
consequently , @xmath106 .
thus we may assume that @xmath34 has exactly one descent .
we decompose @xmath32 into @xmath117 , where @xmath118 is a permutation of the set @xmath119 that does not start with @xmath120 nor end with @xmath121 , and contains exactly one descent .
this decomposition is shown in figure [ fig : proof - thm - one - step ] .
we denote by @xmath122 the size of @xmath118 . since @xmath98 ,
necessarily @xmath123 or we would get a contradiction .
if @xmath124 , we get that @xmath118 is an occurrence of some pattern of @xmath125 in @xmath34 . as a consequence , @xmath106
what is left to prove is that this extends to the case @xmath126 .
we just need to show that we can remove elements in @xmath118 without violating any of the properties below : * the permutation does not start with its smallest element * the permutation does not end with its greatest element * the permutation has exactly one descent until we get a permutation of size @xmath67 .
at that point @xmath118 contains an occurrence of a pattern in @xmath65 , and so does @xmath34 , and we get that @xmath106 .
now , because of the conditions on @xmath118 , the only descent in @xmath118 necessarily goes from the greatest to the smallest element in @xmath118 , ensuring that it is possible to remove elements without violating any of the properties above ( see figure [ fig : proof - thm - one - step ] ) .
( -4,0)(25,11 ) ( 4,10)decomposition of @xmath34 ( 0,0)(8,0 ) ( 0,0)(0,8 ) ( 8,8)(8,0 ) ( 8,8)(0,8 ) ( 0,3.5)(8,3.5 ) ( 0,5.5)(8,5.5 ) ( 3.5,0)(3.5,8 ) ( 5.5,0)(5.5,8 ) ( 0.5,0.5)0.3 ( 1,1)0.3 ( 1.5,1.5 ) .
( 2.5,2.5 ) .
( 3,3)0.3 ( 0.5,-0.5)@xmath19 ( 1,-0.5)@xmath25 ( 1.8,-0.5)@xmath127 ( 3,-0.5)@xmath128 ( 4.2,5)0.3 ( 4.7,4)0.3 ( 3.6,3.6)(4.2,5 ) ( 4.2,5)(4.7,4 ) ( 4.7,4)(5.4,5.4 ) ( 4.5,-0.5)@xmath118 ( 6,6)0.3 ( 6.5,6.5 ) .
( 7.5,7.5)0.3 ( 6,-0.5)@xmath129 ( 6.9,-0.5)@xmath127 ( 7.7,-0.5)@xmath12 ( 19,10)shape of @xmath118 ( 16,1)(22,1 ) ( 16,1)(16,7 ) ( 22,7)(22,1 ) ( 22,7)(16,7 ) ( 18,1)(18,7 ) ( 18.8,1)(18.8,7 ) ( 19.6,1)(19.6,7 ) ( 16,1.8)(22,1.8 ) ( 16,6.2)(22,6.2 ) ( 19.2,1.4)0.3 ( 18.4,6.6)0.3 ( 16.1,2.1)(17.9,5.9 ) ( 19.2,1.4)(18.4,6.6 ) ( 19.7,2.1)(21.9,5.9 ) as for the case of @xmath72 in section [ section : one - step ] , we prove ( theorem [ thm : p - steps ] ) in this section that the class @xmath60 of all permutations obtained from an identity permutation after @xmath8 duplication - loss steps of width at most @xmath9 is a class of pattern - avoiding permutations .
however , we do not get a precise description of the basis of this class , but only an upper bound on the size of the excluded patterns . as in the previous section ,
when referring to a duplication - loss step , we always mean duplication - loss step of width @xmath9 , except when otherwise explicitely stated . to prove the announced result
, we will need a few more notations and technical lemmas .
the _ vector _ from @xmath37 to @xmath130 in a permutation @xmath34 consists of all elements whose positions lie between the positions of @xmath37 and @xmath130 , @xmath37 and @xmath130 being included .
the _ size _ of a vector is the number of elements in it .
for example , the vector from @xmath30 to @xmath25 in the permutation @xmath131 is @xmath132 , and has size @xmath27 .
let @xmath34 be a permutation of @xmath10 .
the _ value - position vector associated with @xmath39 $ ] _ ( _ @xmath133-vector _ for short ) is the vector of @xmath34 going from @xmath37 to @xmath36 , if @xmath37 is not a fixpoint of @xmath34 . in the case
@xmath134 , the @xmath133-vector associated with @xmath37 is empty .
[ def : vp - vector ] it should appear in this definition that the @xmath133-vector associated with @xmath37 , going from the element of @xmath34 which has value @xmath37 to the element of @xmath34 at position @xmath37 , represents the necessary move for @xmath37 to reach its position in the sorted permutation @xmath61 . as it can be seen on figure [ fig : vp - vectors ] , on the graphical representation of permutations used throughout the paper , the @xmath133-vector associated with @xmath37 is an arrow going horizontally from the element at ordinate @xmath37 to the diagonal . we can also notice that a non - empty @xmath133-vector contains at least two elements . to take into account all the moves necessary to sort @xmath34 to @xmath101
, it is convenient to introduce the _ value - position domain _ : let @xmath34 be a permutation of @xmath10 .
the _ value - position domain _ of @xmath34 ( _ @xmath133-domain _ for short ) is composed of all elements of @xmath34 appearing in at least one @xmath133-vector . [ def : vp - domain ] these two definitions are illustrated on figure [ fig : vp - vectors ] .
( -4,0)(23,10 ) ( -1,7.5)@xmath135 ( 1,7.5)@xmath27 ( 2,7.5)@xmath19 ( 3,7.5)@xmath25 ( 4,7.5)@xmath26 ( 5,7.5)@xmath28 ( 6,7.5)@xmath30 ( 7,7.5)@xmath29 ( 2.2,6.5)(0.8,6.5 ) ( 3.2,6)(1.8,6 ) ( 4.2,5.5)(2.8,5.5 ) ( 0.8,5)(4.2,5 ) ( 7.2,6.5)(5.8,6.5 ) ( 5.8,6)(7.2,6 ) ( 3,3)@xmath133-domain of @xmath136 ( 4,1.5)@xmath137 ( 15,1)(22,8 ) ( 15,1)(16,2 ) ( 16,2)(17,3 ) ( 17,3)(18,4 ) ( 18,4)(19,5 ) ( 19,5)(20,6 ) ( 20,6)(21,7 ) ( 21,7)(22,8 ) ( 16.5,1.5)(15.2,1.5 ) ( 17.5,2.5)(16.2,2.5 ) ( 18.5,3.5)(17.2,3.5 ) ( 15.5,4.5)(18.8,4.5 ) ( 21.5,6.5)(20.2,6.5 ) ( 20.5,7.5)(21.8,7.5 ) ( 15.5,4.5)0.3 ( 16.5,1.5)0.3 ( 17.5,2.5)0.3 ( 18.5,3.5)0.3 ( 19.5,5.5)0.3 ( 20.5,7.5)0.3 ( 21.5,6.5)0.3 now , observe that for any permutation , the @xmath133-vectors are reversible in the sense that reversing all the arrows will give a set of vectors that represent the moves of elements that are necessary to `` unsort '' @xmath101 into @xmath34 .
it is easily seen from definitions [ def : vp - vector ] and [ def : vp - domain ] and this remark that for any permutation @xmath138 , any element belonging to the @xmath133-domain of @xmath34 also belongs to at least one of the duplication - loss steps used to obtain @xmath34 from @xmath101 .
consequently , the @xmath133-domain of @xmath34 contains at most @xmath139 elements .
consider a permutation @xmath34 , and the permutation @xmath140 obtained from @xmath34 by the removal of some element @xmath130 .
then for any element @xmath141 such that @xmath142 , either this element becomes a fixpoint in @xmath140 or the size of the @xmath133-vector associated with this element in @xmath140 remains constant , is increased of 1 or is diminished of 1 with respect to the size of the @xmath133-vector associated with @xmath37 in @xmath34 .
[ lem : vp - vector - unchanged ] it is easily seen on the graphical representation of @xmath34 .
any element that does not lie just above or just below the diagonal can not become a fixpoint when removing an element @xmath130 . for
elements that do not becom fixpoints , the horizontal distance to the diagonal can only change of @xmath143 , @xmath19 or @xmath144 when removing some element @xmath130 ( see figure [ fig : vp - vector - unchanged ] ) .
( -2,-2)(40,12 ) ( 0,0)(8,8 ) ( 0,0)(1,1 ) ( 1,1)(2,2 ) ( 2,2)(3,3 ) ( 3,3)(4,4 ) ( 4,4)(5,5 ) ( 5,5)(6,6 ) ( 6,6)(7,7 ) ( 7,7)(8,8 ) ( 2,-1)(2,9 ) ( 3,-1)(3,9 ) ( -1,6)(9,6 ) ( -1,7)(9,7 ) ( 2.5,6.5)0.3 ( -1.5,6.5)@xmath130 ( 3,2)(4,3 ) ( 4,3)(5,4 ) ( 5,4)(6,5 ) ( 6,5)(7,6 ) ( 3,4)(4,5 ) ( 3,5)(4,6 ) ( 4,5)(5,6 ) ( 0,7)(1,8 ) ( 1,7)(2,8 ) ( 3,0)(4,1 ) ( 4,0)(5,1 ) ( 5,0)(6,1 ) ( 6,0)(7,1 ) ( 7,0)(8,1 ) ( 3,1)(4,2 ) ( 4,1)(5,2 ) ( 5,1)(6,2 ) ( 6,1)(7,2 ) ( 7,1)(8,2 ) ( 4,2)(5,3 ) ( 5,2)(6,3 ) ( 6,2)(7,3 ) ( 7,2)(8,3 ) ( 5,3)(6,4 ) ( 6,3)(7,4 ) ( 7,3)(8,4 ) ( 6,4)(7,5 ) ( 7,4)(8,5 ) ( 7,5)(8,6 ) ( 6.5,2.5)(5.5,2.5 ) ( 4.5,5.5)(3.5,5.5 ) ( 0.5,7.5)(0.5,6.5 ) ( 5.5,7.5)(4.5,6.5 ) ( 1.5,0.5)0.15 ( 0.5,3.5)0.15 ( 14,0)(22,8 ) ( 14,0)(15,1 ) ( 15,1)(16,2 ) ( 16,2)(17,3 ) ( 17,3)(18,4 ) ( 18,4)(19,5 ) ( 19,5)(20,6 ) ( 20,6)(21,7 ) ( 21,7)(22,8 ) ( 19,-1)(19,9 ) ( 20,-1)(20,9 ) ( 13,2)(23,2 ) ( 13,3)(23,3 ) ( 19.5,2.5)0.3 ( 12.5,2.5)@xmath130 ( 16,3)(17,4 ) ( 17,4)(18,5 ) ( 18,5)(19,6 ) ( 18,3)(19,4 ) ( 20,0)(21,1 ) ( 21,0)(22,1 ) ( 20,1)(21,2 ) ( 21,1)(22,2 ) ( 14,7)(15,8 ) ( 15,7)(16,8 ) ( 16,7)(17,8 ) ( 17,7)(18,8 ) ( 18,7)(19,8 ) ( 14,6)(15,7 ) ( 15,6)(16,7 ) ( 16,6)(17,7 ) ( 17,6)(18,7 ) ( 18,6)(19,7 ) ( 14,5)(15,6 ) ( 15,5)(16,6 ) ( 16,5)(17,6 ) ( 17,5)(18,6 ) ( 14,4)(15,5 ) ( 15,4)(16,5 ) ( 16,4)(17,5 ) ( 14,3)(15,4 ) ( 15,3)(16,4 ) ( 21.5,0.5)(20.5,0.5 ) ( 18.5,3.5)(18.5,2.5 ) ( 15.5,6.5)(15.5,5.5 ) ( 21.5,5.5)(20.5,4.5 ) ( 20.5,7.5)(19.5,6.5 ) ( 14.5,1.5)0.15 ( 16.5,0.5)0.15 ( 26,8)(27,9 ) ( 30.9,8.3)diagonal ( 26,6)(27,7 ) ( 33.7,6.3)candidate fixpoints ( 26,5)(27,5 ) ( 26.5,4.3)0.15 ( 30.8,4.8)changes ( 33,3.8)in the @xmath133-vectors ( 33,2.1)variation of the distance ( 33,1.1)to the diagonal ( 26,-0.5)(27,0.5 ) ( 28.2,-0.2)@xmath143 ( 29.7,-0.5)(30.7,0.5 ) ( 32.2,-0.2)@xmath145 ( 33.7,-0.5)(34.7,0.5 ) ( 36.2,-0.2)@xmath144 for any permutation @xmath34 , there is at least one element @xmath130 such that the permutation @xmath140 obtained from @xmath34 by the removal of @xmath130 contains at most one more fixpoint than @xmath34 .
[ lem : one - fixpoint ] it is convenient to introduce the _ quasi - diagonal elements _ of @xmath34 , defined as follows .
@xmath37 is a quasi - diagonal element of @xmath34 if @xmath146 or @xmath147 .
these two cases correspond respectively to elements of @xmath34 lying just above or just below the diagonal in the graphical representation of @xmath34 .
any element of @xmath34 that may become a fixpoint in @xmath140 is necessarily a quasi - diagonal element .
if there is no quasi - diagonal element , then we can remove any element @xmath130 to obtain a permutation @xmath140 that does not have more fixpoints than @xmath34 .
if there are some , then we pick @xmath130 among the quasi - diagonal elements . we claim that at most one fixpoint is create while removing @xmath130 .
the argument is simple .
suppose @xmath130 is such that @xmath148 , the other case being similar .
then the only fixpoint that may appear is @xmath149 , if @xmath150 .
this should appear clearly on figure [ fig : one - fixpoint ] .
( -2,-2)(30,7 ) ( 0,0)(5,5 ) ( 0,0)(1,1 ) ( 1,1)(2,2 ) ( 2,2)(3,3 ) ( 3,3)(4,4 ) ( 4,4)(5,5 ) ( 2.5,3.5)0.3 ( 2,-1)(2,6 ) ( 3,-1)(3,6 ) ( -1,3)(6,3 ) ( -1,4)(6,4 ) ( 3,2)(4,3 ) ( 8,0)(13,5 ) ( 8,0)(9,1 ) ( 9,1)(10,2 ) ( 10,2)(11,3 ) ( 11,3)(12,4 ) ( 12,4)(13,5 ) ( 10.5,1.5)0.3 ( 10,-1)(10,6 ) ( 11,-1)(11,6 ) ( 7,1)(14,1 ) ( 7,2)(14,2 ) ( 9,2)(10,3 ) ( 16,4)(17,5 ) ( 20.9,4.3)diagonal ( 16.5,2.5)0.3 ( 23.4,2.3)removed element ( 16,0)(17,1 ) ( 23.7,0.3)candidate fixpoint consider a permutation @xmath151 such that for any strict pattern @xmath140 of @xmath34 , @xmath152 . then the @xmath133-domain of @xmath34 is of size at most @xmath153 .
[ lem : bounded - vp - domain ] by lemma [ lem : one - fixpoint ] , we can choose some @xmath154 with @xmath155 and such that @xmath140 has at most one more fixpoint than @xmath34 .
call @xmath130 the element deleted in @xmath34 to obtain @xmath140 . by a previous remark , since @xmath152 ,
the @xmath133-domain of @xmath140 is of size at most @xmath139 , and is therefore composed of at most @xmath139 @xmath133-vectors .
each of these @xmath133-vectors in @xmath140 yields a @xmath133-vector in @xmath34 , whose size is smaller or equal or possibly increased by @xmath19 .
let us denote by @xmath156 the set of @xmath133-vectors of @xmath34 obtained from a @xmath133-vector of @xmath140 .
then the number of elements of @xmath34 that belong to a @xmath133-vector of @xmath156 is at most @xmath157 .
however @xmath156 is not yet the @xmath133-domain of @xmath34 .
we must complete it with up to two @xmath133-vectors : the one associated with the element @xmath130 deleted , and the one associated with the fixpoint of @xmath140 that was not a fixpoint in @xmath34 , if such a point exists .
if such an element exists , then it is a quasi - diagonal element in @xmath34 and its @xmath133-vector ( denoted @xmath158 ) in @xmath34 is necessarily of size @xmath25 , so that @xmath159 has total size at most @xmath153 .
now it is easily observed that any element of @xmath34 belonging to one @xmath133-vector necessarily belongs to at least two @xmath133-vectors ( this can be seen as a `` balance condition '' ) .
consequently , all the elements of the @xmath133-vector associated with @xmath130 are already covered by a vector of @xmath159 , so that the @xmath133-domain of @xmath34 is exactly the set of elements covered by @xmath159 .
therefore , its size is at most @xmath153 . consider a permutation @xmath151 of size @xmath160 such that for any strict pattern @xmath140 of @xmath34 , @xmath152
. then @xmath34 is of the form @xmath161 with @xmath162 a permutation of @xmath163 $ ] and @xmath164 a permutation of @xmath165 $ ] .
it is possible that @xmath162 or @xmath164 is empty .
[ lem : nearly - sorted - pi ] by lemma [ lem : bounded - vp - domain ] , the @xmath133-domain of @xmath34 is of size at most @xmath153 .
we can decompose @xmath34 into _ free windows _ of consecutive elements outside the @xmath133-domain of @xmath34 , separated by windows of consecutive elements of the @xmath133-domain .
now , there are at most @xmath166 windows of consecutive elements of the @xmath133-domain , and consequently , there are at most @xmath167 free windows in @xmath34 . since @xmath34 is of size @xmath168 , at least one of the free windows of @xmath34 has size strictly greater than @xmath139 , _
i.e. , _ contains at least @xmath166 elements . by definition
, these elements do not belong to the @xmath133-domain of @xmath34 , and hence they allow the decomposition of @xmath34 into @xmath161 with @xmath162 a permutation of @xmath163 $ ] and @xmath164 a permutation of @xmath165 $ ] .
figure [ fig : nearly - sorted - pi ] represent the decomposition of @xmath34 used in this proof .
( -1.2,-3)(16,3 ) ( -1.2,0.3)@xmath135 ( 1.5,0)(2,1 ) ( 4,0)(7,1 ) ( 8.5,0)(9.5,1 ) ( 12.5,0)(14.5,1 ) ( 0,0)(15,1 ) ( 1.75,1.1)(7.35,2.3 ) ( 5.5,1.1)(7.45,2.3 ) ( 9,1.1)(7.55,2.3 ) ( 13.5,1.1)(7.65,2.3 ) ( 7.5,3)at most @xmath166 @xmath133-windows ( 0.75,-0.1)(7.3,-2.3 ) ( 3,-0.1)(7.4,-2.3 ) ( 7.75,-0.1)(7.5,-2.3 ) ( 11,-0.1)(7.6,-2.3 ) ( 14.75,-0.1)(7.7,-2.3 ) ( 7.5,-3)at most @xmath167 free - windows consider a permutation @xmath169 where @xmath170 is a permutation of @xmath171 $ ] . if @xmath34 is obtainable after @xmath8 duplication steps of size at most @xmath9 then @xmath34 is obtainable after @xmath8 duplication steps of size at most @xmath9 such that the duplicated window for each step does not intersect [ lem : localisation ] the key idea is to consider the first sequence @xmath172 of duplication - loss steps and create a new sequence @xmath173 such that : * each step @xmath174 concerns only elements of @xmath175 $ ] . * after every step @xmath174 , the elements @xmath176 are in the same order than after performing steps @xmath177 .
then the proof is by induction on the number of steps .
if there is only one step then the proof is straighforward .
suppose now that the above statement is true until @xmath178 steps .
then for the last step , we use our hypothesis for @xmath178 so that we have operations @xmath179 respecting the above conditions . for @xmath180 , only notice that the elements of @xmath171 $ ] involved in @xmath181 are also in a window of size @xmath9 in the permutation obtained after @xmath182 and in the same relative order by our induction hypothesis which proves the existence of @xmath180 . using these lemmas , we state and
prove a key proposition : consider a permutation @xmath151 .
then either @xmath34 is of size at most @xmath69 , or there exists a strict pattern @xmath140 of @xmath34 that does not belong to @xmath60 .
[ prop : p - steps ] consider a permutation @xmath151 such that any strict pattern @xmath140 of @xmath34 belongs to @xmath60 .
we want to show that @xmath34 is of size @xmath183 .
let us assume the contrary . by lemma [ lem : nearly - sorted - pi ] , there exist @xmath39 $ ] , @xmath162 a permutation of @xmath163 $ ] and @xmath164 a permutation of @xmath165 $ ] such that @xmath161 .
let us denote @xmath118 the permutation @xmath184 , where @xmath185 is the permutation of @xmath186 $ ] obtained from @xmath164 by subtracting @xmath19 to every element of @xmath164 .
@xmath118 is a strict pattern of @xmath34 , hence @xmath187 .
consider a shortest sequence of duplication - loss steps of width at most @xmath9 that produces @xmath118 from @xmath188 .
this sequence has at most @xmath8 steps , each of width at most @xmath9 .
it implies that the total distance crossed by the elements that are duplicated is at most @xmath139 .
consequently , it is not possible to bring an element of @xmath162 and an element of @xmath189 in two consecutive positions .
so it is necessary that the duplication - loss steps of the scenario we consider are _ internal _ to @xmath162 and @xmath189 .
we can reproduce these steps in @xmath162 and @xmath164 to obtain @xmath34 from @xmath61 in at most @xmath8 duplication - loss steps of width at most @xmath9 , contradicting that @xmath151 .
it is then quite easy to prove theorem [ thm : p - steps ] : the class @xmath60 of all permutations obtained from an identity permutation after @xmath8 duplication - loss steps of width at most @xmath9 is a class of pattern - avoiding permutations whose basis is finite and contains only patterns of size at most @xmath69 .
[ thm : p - steps ] we set @xmath190 and show that @xmath191 .
consider @xmath151 .
if @xmath192 , then @xmath193 and @xmath106 . otherwise , if @xmath194 , then by proposition [ prop : p - steps ] , there exists a strict pattern @xmath140 of @xmath34 that does not belong to @xmath60 .
reasoning by induction on the size of the permutations , we deduce from @xmath195 that @xmath196 .
a direct consequence is that @xmath106 .
this proves that @xmath197 .
conversely , consider @xmath138 .
then any pattern @xmath140 of @xmath34 is also obtainable from an identity permutation in at most @xmath8 steps of width at most @xmath9 ( using the sequence of duplication - loss steps associated with @xmath34 ) , _
i.e. , _ @xmath152 . then @xmath34 does not contain an occurrence of any pattern of @xmath63 , _
i.e. , _ @xmath198 .
this shows that @xmath199 , ending the proof of the theorem .
the _ whole genome duplication - random loss _ model is studied in @xcite , and the authors describe a method to compute an optimal duplication - loss scenario , _
i.e. , _ a scenario of duplications ( of the whole genome in this case ) and losses whose number of steps is minimal .
our model with bounded size duplication operations reduces to the _ whole genome duplication - random loss _
case when @xmath200 and thus to a radix - sort algorithm as shown in @xcite and to a bubble - sort when @xmath201 .
thus we give some algorithm whose complexity matches the two extremal cases and shows some continuity between the two sorting algorithms .
it is worth noticing that any scenario in our model can be viewed as a whole genome duplication - random loss scenario .
consequently , the number of steps of an optimal whole genome duplication - random loss scenario is a lower bound to the number of steps of an optimal scenario in our duplication - loss model .
it is also easy to see that , when considering permutations of size at most @xmath9 , our model and the whole genome duplication - random loss model coincide .
indeed , we will use for our purpose the procedure of @xcite , which is given in algorithm [ alg : binary ] .
we omit the proof of correctness and optimality of this algorithm .
see @xcite for details . in order to examine every bit of the labels given to the elements of @xmath202 $ ] , the number of steps in the loop on line 4 is @xmath203number of maximal increasing substrings of @xmath204 .
a consequence is that the number of steps in an optimal whole genome duplication - random loss scenario from @xmath101 to @xmath34 is @xmath205 in the worst case and on average ( see equation ( [ eq : log - average ] ) for the average case ) .
note that the same algorithm can be used to compute an optimal whole genome duplication - random loss scenario from @xmath206 , with @xmath4 and @xmath207 , to any permutation of @xmath208 . in this section ,
we provide an algorithm that computes , for any permutation @xmath32 in input , a possible scenario of duplications and losses to obtain @xmath34 from @xmath101 .
we will restrict ourselves to duplication - loss steps of width at most @xmath9 , so that the number of duplication - loss steps corresponds to the cost of the scenario in our cost model .
we are interested in the number of duplication - loss steps of the scenario produced by the algorithm , in the worst case , and on average .
it provides an upper bound on the number of duplication - loss steps that are necessary to obtain a permutation .
the algorithm we use is described in algorithm [ alg : bucket ] .
a few keys to understand algorithm [ alg : bucket ] are the following remarks .
the set @xmath209 of values defined at line 3 represents the rightmost @xmath210 elements of @xmath34 not yet examined .
the algorithm consists in two different loops , the first one corresponding to lines @xmath25 to @xmath28 of the algorithm and the second one from line @xmath29 to @xmath31 . at the end of the first loop ( line 5 )
, @xmath41 is decomposed into windows of width @xmath211 ( except the leftmost one which is of width at most @xmath9 ) ; and each of these windows is an increasing sequence containing exactly the same elements as the window of @xmath34 corresponding to the same indices . in the second loop , we consider these windows from right to left and since there are of width less than @xmath9 , we can call algorithm [ alg : binary ] ( that implements whole genome duplication - random loss ) on each window successively to transform @xmath41 into @xmath34 .
an example is given with @xmath212 and @xmath213 .
we first cut @xmath34 in chunks of size @xmath214 and obtain @xmath215 .
then the first loop of the algorithm ( step @xmath25 to @xmath28 ) starts from @xmath216 and takes the elements in increasing order to the same chunk the belong to in @xmath34 .
this gives @xmath217 .
then the second loop sorts each chunk separately to obtain @xmath34 using the radix sort algorithm [ alg : binary ] introduced in @xcite .
notice here that we use in the second loop ( except for the leftmost window ) only duplication - loss steps of width @xmath211 .
an improvement we considered is to use whole genome duplication - random loss scenarios on windows of width @xmath9 , that are nonetheless increasing sequences .
unfortunately , we were not able to analyse how many duplication - loss steps there are in a scenario produced by such an algorithm .
we now analyse the number of steps of the scenario produced by algorithm [ alg : bucket ] .
the number of duplication - loss steps of a scenario produced by algorithm [ alg : bucket ] on a permutation of size @xmath12 is at most @xmath70 asymptotically . [
prop : worst - case - upper - bound ] suppose we are at iteration @xmath37 of the first loop .
we have to move the @xmath211 elements of @xmath209 to their positions ( from @xmath218 to @xmath219 ) by duplication - loss steps of width at most @xmath9 .
the worst situation is when the elements of @xmath209 are at the begining of @xmath41 .
but in this case , we can move to the right the elements of @xmath209 by @xmath220 positions at every duplication - loss step , until they reach their position .
the total number of duplication - loss steps in this first process is then at most @xmath221 consider now the second loop of algorithm [ alg : bucket ] . in each window of size
@xmath211 , it performs at most @xmath222 duplication - loss steps ( line 7 ) and in the leftmost window ( line 9 ) , at most @xmath223 by the result of @xcite .
consequently the number of duplication - loss steps produced by the second loop is @xmath224 we finally get that the total number of duplication - loss steps in a scenario produced by algorithm [ alg : bucket ] is at most @xmath70 asymptotically in the worst case .
it is easily noticed that this worst case corresponds to the _ reversed identity _
permutation @xmath225 .
this corresponds to our intuition of a worst case situation in this context .
we can also notice that @xmath226 for `` small '' values of @xmath9 , namely as long as @xmath227 .
if on the contrary @xmath228 then @xmath229 .
when @xmath230 , the two terms are of the same order .
we can also compute the average number of duplication - loss steps of a scenario produced by algorithm [ alg : bucket ] .
the number of duplication - loss steps of a scenario produced by algorithm [ alg : bucket ] on a permutation of size @xmath12 is on average @xmath231 asymptotically .
[ prop : average - upper - bound ] first , we introduce a few notations .
consider @xmath34 a permutation of size @xmath12 , and decompose it from right to left into @xmath232 windows of width @xmath211 , except the leftmost one , whose width is @xmath233 .
we denote @xmath234 this decomposition .
now , let us denote @xmath235 the number of duplication - loss steps produced in the first loop of algorithm [ alg : bucket ] on @xmath34 . and in particular , we denote @xmath236 the number of such steps produced by the first iteration of this loop , _ i.e. _ , the number of steps to move the elements of @xmath237 at the end of the permutation . for computing the average number of such steps , we consider @xmath238 .
it is simple to conceive that @xmath239 let us focus on @xmath240 .
figure [ fig : min ] should convince the reader that @xmath241 ( 0,-3)(30,5 ) ( 1,0)@xmath135 ( 3,-0.5)(18,-0.5 ) ( 3,1)(18,1 ) ( 3,-0.5)(3,1 ) ( 18,-0.5)(18,1 ) ( 15,-0.5)(15,1 ) ( 15.5,1.5)(9,1.5 ) ( 16.5,2)(7,2 ) ( 17.5,2.5)(13,2.5 ) ( 12,3)@xmath133-vectors ( 7,0.2)0.3 ( 0,-1.5)positions : ( 3.2,-1.5)@xmath19 ( 4.,-1.5)@xmath25 ( 5.5,-1.5)@xmath127 ( 8.5,-1.5)@xmath242 ( 12,-1.5)@xmath127 ( 17.5,-1.5)@xmath12 ( 15,-0.5)(18,1 ) ( 25,-2.5)@xmath211 rightmost positions ( 22.5,-1.5)(16,0.5 ) ( 7,4)(18,4 ) ( 12,4.5)@xmath243 elements ( 7,4)(7,0.5 ) ( 18,4)(18,0.5 ) now , we can notice that the number of permutations @xmath34 of size @xmath12 such that @xmath244 is @xmath245 . this yields @xmath246 consequently , @xmath247 giving after a few computations @xmath248 therefore , we consider two sequences @xmath249 and @xmath250 satisfying the relations @xmath251 and @xmath252 respectively if @xmath253 , and @xmath254 for any @xmath255
. then we have @xmath256 .
we can solve the recurrence equations for @xmath257 and @xmath258 ; and if we write @xmath259 ( then @xmath260 ) , we get : @xmath261 and @xmath262 consequently , the average number of duplication - loss steps produced by the first loop of algorithm [ alg : bucket ] on permutations of size @xmath12 is @xmath263 .
what is left to compute is the average number of duplication - loss steps produced by the second loop of algorithm [ alg : bucket ] on permutations of size @xmath12 .
this number is given by @xmath264 since @xmath265 , we deduce that the average number of duplication - loss steps produced by the second loop of algorithm [ alg : bucket ] on permutations of size @xmath12 is @xmath266 .
hence we focus on the computation of @xmath267 for @xmath268 . by concavity of the @xmath269 function , since @xmath270 , we get that @xmath271 moreover , it is clear that @xmath272 so that we deduce that @xmath273 consequently , the average number of duplication - loss steps produced by the second loop of algorithm [ alg : bucket ] on permutations of size @xmath12 is @xmath274 .
finally , we end the proof concluding that the total number of duplication - loss steps in a scenario produced by algorithm [ alg : bucket ] on a permutation of size @xmath12 is @xmath275 on average .
it is possible to provide very simple lower bounds on the number of duplication - loss steps necessary to obtain a permutation .
these lower bounds are given and proved in propositions [ prop : worst - case - lower - bound ] and [ prop : average - lower - bound ] below .
they are tight in most cases , however not in any case .
indeed the upper and lower bounds coincide up to a constant factor whenever @xmath9 is a constant , or when @xmath21 , except when @xmath276 . in the worst case , @xmath71 duplication - loss steps of width @xmath9
are necessary to obtain a permutation of @xmath10 from @xmath277 .
[ prop : worst - case - lower - bound ] let us consider first the number of inversions in a permutation that can create a duplication - loss step @xmath278 of width @xmath9 .
it is easily seen that these new inversions can only involve two elements of @xmath278 .
call @xmath37 the number of elements of @xmath278 that are kept in the first copy
. then the maximum number of inversions that can be created by @xmath278 is @xmath279 .
now , a permutation @xmath32 has up to @xmath280 inversions , so that at least @xmath281 duplication - loss steps are necessary to transform @xmath277 into @xmath34 . to get the other term of the lower bound
, we just refer to the result of @xcite recalled at the beginning of this section , namely that @xmath282 steps are necessary in the worst case in the whole genome duplication - random loss model , in which duplication - loss operation are less restricted .
finally , we get a lower bound of @xmath283 necessary duplication - loss steps to obtain a permutation of @xmath10 from @xmath284 in the worst case
. on average , @xmath283 duplication - loss steps of width @xmath9 are necessary to obtain a permutation of @xmath10 from @xmath277 . [
prop : average - lower - bound ] as before , a duplication - loss step can create at most @xmath285 inversions in a permutation .
but the average number of inversions in a permutation of @xmath10 is @xmath286 , so that on average at least @xmath287 duplication - loss steps are necessary to transform @xmath277 into @xmath32 .
again , @xcite provides use with the @xmath288 lower bound , referring to the whole genome duplication - random loss model which is more general than ours , so that this bound applies in our context .
we conclude that a lower bound on the average number of duplication - loss steps necessary to obtain a permutation of @xmath10 from @xmath284 is @xmath283 .
we discuss the results of section [ section : algo ] on the average ( or worst case ) number of steps of width @xmath9 to obtain a permutation of size @xmath12 .
it appears that we could not provide lower bounds that coincide with the upper bounds given by our algorithm , but we claim that they are tight in many cases .
indeed , whenever @xmath289 , we get that @xmath290 , and consequently the upper bound can be rewritten as @xmath291 , which coincide up to a constant factor with the lower bound @xmath292 . for the case @xmath230 , the same argument holds , but the constant factor between the lower and the upper bound might be much greater . finally , if @xmath293 , then @xmath294 and @xmath295 , so that upper and lower bounds coincide again . on the contrary ,
when @xmath296 , the upper and lower bounds provided do not coincide .
we leave as an open question the problem of finding an algorithm that computes a duplication - loss scenario whose number of steps is optimal ( on average and in the worst case ) up to a constant factor , when the width @xmath9 of the duplicated windows satisfies @xmath296 .
several other questions are still open .
first of all neither of our algorithms is optimal for a specific permutation and our results are only optimal asymptotically in average and/or in the worst case
. it could be interesting to provide algorithms that produce optimal duplication - loss scenarios on any permutation @xmath34 , for @xmath21 in order to provide some continuity between the bubble sort ( corresponding to @xmath201 ) and the radix sort ( corresponding to @xmath297 ) .
albert , r.e.l .
aldred , m.d .
atkinson , h.p .
van ditmarsch , c.c .
handley , d.a .
hotlon , and d.j .
compositions of pattern restricted sets of permutations . technical report ,
university of otago , 2004 .
technical report number oucs-2004 - 12 .
m.c . chen and r.c.t .
lee . sorting by transpositions based on the first increasing substring concept . in _
bibe 04 : proceedings of the 4th ieee symposium on bioinformatics and bioengineering _
, page 553 , washington , dc , usa , 2004 .
ieee computer society .
anthony labarre .
a new tight upper bound on the transposition distance . in rita casadio and gene myers , editors , _ algorithms in bioinformatics , 5th international workshop , wabi 2005 , mallorca , spain , october 3 - 6 , 2005 , proceedings _ ,
volume 3692 of _ lecture notes in computer science _ , pages 216227 .
springer , 2005 . | in @xcite , chaudhuri , chen , mihaescu and rao study algorithmic properties of the _ tandem duplication - random loss model _ of genome rearrangement , well - known in evolutionary biology . in their model ,
the cost of one step of duplication - loss of width @xmath0 is @xmath1 for @xmath2 or @xmath3 . in this paper , we study a variant of this model , where the cost of one step of width @xmath0 is 1 if @xmath4 and @xmath5 if @xmath6 , for any value of the parameter @xmath7 .
we first show that permutations obtained after @xmath8 steps of width @xmath9 define classes of pattern - avoiding permutations .
we also compute the numbers of duplication - loss steps of width @xmath9 necessary and sufficient to obtain any permutation of @xmath10 , in the worst case and on average . in this second part
, we may also consider the case @xmath11 , a function of the size @xmath12 of the permutation on which the duplication - loss operations are performed . sorting , permutations , pattern |
observations of in intervening quasar absorption line systems are vital to the study of the interstellar medium of low - redshift galaxies and their environments , as lies in the optical from @xmath7 , and serves as a probe of low - ionization gas . through photoionization models ,
it is possible to derive many properties of the absorbing gas .
this includes the line - of - sight extent , density , temperature , and molecular content , which are constrained by absorption in the different ionization states of various chemical elements .
@xcite and @xcite found that roughly one - third of weak absorbers at redshifts @xmath8 , with @xmath9 , are in multiple - cloud systems .
they propose that many of the weak absorbers contain multi - phase structures with a range of levels of ionization .
such absorbers are generally thought to arise in sub - lyman limit system environments , with @xmath10 , and metallicities log@xmath11 > -1 $ ] @xcite .
strong absorbers ( @xmath12 ) , in contrast , are almost always associated with lyman limit systems ( @xmath13 ) , and many with damped absorbers ( dlas ; with @xmath14x@xmath15 ) .
@xcite found that @xmath16 of mgii absorbers with @xmath17 and @xmath18 were dlas in an hst survey for @xmath19 systems . in that sample ,
the average @xmath20 was @xmath21 x @xmath22 for @xmath23 , and @xmath24 x @xmath25 for @xmath26 .
most dlas at low @xmath27 are thought to be associated with galaxies with a variety of morphological types , from @xmath28 galaxies to low surface brightness galaxies @xcite .
@xcite find molecular hydrogen in 13 - 20% of dla systems at high redshift , but note that there is no correlation between the detection of molecular hydrogen and column density . despite this lack of correlation
, @xcite find molecular hydrogen in 9 out of 18 high metallicity systems ( [ x / h]@xmath29 ) at high redshift . in this paper , we study a multiple - cloud weak system toward he0001 - 2340 ( @xmath30 , @xcite ) at @xmath31 .
the four weak clouds are spread over a velocity range of @xmath32 . and , which are generally only detected in the very strongest absorbers , are found in one of the four clouds comprising this system .
, , and are also detected in this cloud ; these neutral states have not been reported to exist in any other extragalactic environment , even in most dlas , and have only been found in several dense galactic molecular clouds @xcite .
@xcite notes that the ratios of @xmath33 and @xmath34 in this system are orders of magnitude higher than in other absorbers , implying a very low ionization state .
she also observes that there is an extreme underabundance of mg with respect to fe , which can not be explained by nucleosynthesis or dust depletion , and can not be reproduced by photoionization models .
the metallicity of the system can not be directly determined from due to the absorption from a lyman limit system at @xmath35 @xcite . due to a lack of metallicity constraints
, @xcite assumes dla - regime column densities , noting that the observed @xmath36 of the system is comparable to the sample of 11 dla by @xcite .
she further constrains her parameters by comparing to the local cold interstellar clouds of @xcite , finding that systems with similar amounts of @xmath37 have metallicities @xmath38<-2.78 $ ] . with such a low assumed metallicity and high @xmath39 column density , @xcite is unable to reproduce the observed ratios of @xmath40 and @xmath41 in photoionization models , and is thus unable to draw concrete conclusions about the properties of this cloud .
we propose that these noted differences suggest that this unusual cloud is part of class of systems unrelated to previously observed dlas , which @xcite notes have @xmath42 two orders of magnitude higher than this system , as well as no previous , , and detections .
we therefore explore a range of metallicities and @xmath43 in our modeling process .
as it is difficult to reproduce the line ratios in this unusual cloud , we also expand the consideration of parameter space to explore the possibility of unresolved saturation .
the effect of partial covering of the background quasar is also explored , since photoionization models of the system suggest the cloud is compact enough to partially cover the broad emission line region of the quasar , with high densities ( @xmath44 ) , cold temperatures ( @xmath45k ) , and a molecular hydrogen fraction larger than @xmath46 .
partial covering has only been observed once before in an intervening quasar absorption line system . in the lensed quasar apm 08279 + 5255 at @xmath47 @xcite ,
one - third of the components of a strong absorber were not detected toward the second image of the qso , while fits to the components suggested @xmath48 in the other image , constraining the absorber size to be as small as @xmath49 pc .
we begin , in [ sec:2 ] , with a description of the vlt / uves spectrum of he @xmath50 , and display and quantify the observed properties of the @xmath51 multiple cloud weak system .
[ sec:3 ] details the voigt profile fit performed on this four - cloud system , including covering factor analysis of the first cloud .
it also describes the photoionization modeling method used to constrain the ionization parameters / densities of the four clouds .
[ sec:4 ] gives modeling results for each cloud , while
[ sec:5 ] discusses the implications of the cloud models on the origin of the gas , while
[ sec:6 ] summarizes the findings and considers the properties of the absorption system in the context of broader questions relating to galaxy environments .
a spectrum of he0001 - 2340 , taken in 2001 , was procured from the eso archive , having been obtained as part of the eso - vlt large programme `` the cosmic evolution of the igm '' @xcite .
this @xmath30 quasar , he0001 - 2340 ( @xmath52 ) , was observed with the ultraviolet and visual echelle spectrograph ( uves ) on the very large telescope ( vlt ) , as detailed in @xcite .
the data were reduced as described in @xcite .
the spectrum has a resolution r @xmath53 ( @xmath54 ) and covers a range of 3050 - 10070 .
breaks in wavelength coverage occurred at 5750 - 5840 and 8500 - 8660 .
the spectrum is of extremely high quality , with @xmath55 per pixel over most of the wavelength coverage .
continuum fitting ( with a cubic spline function ) was performed using the irafsfit task as described in @xcite . in the @xmath51 system , we detect four distinct @xmath562796 subsystems at @xmath57 levels in the vlt spectrum , and all are confirmed by @xmath562803 . fig .
[ fig : clouds_labeled ] shows the location of these features , with a velocity scale centered so that half of the system optical depth lies blueward of @xmath58 , at @xmath59 .
detections may be noted at @xmath60 , @xmath58 , @xmath61 , and @xmath62 in , at @xmath60 and @xmath58 in , and at @xmath60 in .
the fourth cloud is very weak , with @xmath562803 detected at just over @xmath63 ; this detection is made possible by the very high s / n of the spectrum .
the first cloud also has detected , , , , and , as seen in figure [ fig : system_plot_unresolved ] .
all transitions blueward of @xmath64 ( @xmath65 in the rest frame of the @xmath51 system ) are potentially contaminated by forest lines .
the only two detected features displayed in fig . [ fig : system_plot_unresolved ] that are not from the @xmath51 system , and are outside of the forest region , occur in the @xmath564228 panel , at @xmath66 , and in the @xmath563022 panel , at @xmath67 . the former feature is @xmath562374 from a system at @xmath68 , while the latter is @xmath561548 from a system at @xmath69 .
lyman series lines for the @xmath51 system are unavailable , due to a full lyman limit break from a system at @xmath70 @xcite , allowing no direct constraints on the metallicity of the system .
rest frame equivalent widths for all transitions detected at @xmath71 in the spectra , as determined by gaussian fits to the unresolved line profiles , are given in table [ tab : ew ] for the strongest transitions .
the column densities and doppler parameters , using voigt profile fitting , assuming full coverage , are listed in table [ tab : nb ] .
cloud 2 was resolved into two blended components by our fitting procedure .
table [ tab : ew2 ] shows equivalent widths for additional transitions detected in cloud 1 .
upper limits are given for blended detections .
since the detection of neutral transitions in a weak absorber is quite surprising , we must consider whether the apparent , , , and detections for cloud 1 are valid , and not chance superpositions . in table
[ tab : ew2 ] , the oscillator strengths of the nine detected transitions are listed , along with their equivalent widths . only three of these detections , @xmath562484 , @xmath562502 , and @xmath562524 , are in the forest .
the oscillator strengths of the various transitions are roughly consistent with the relative equivalent widths , however , a detailed analysis suggests that partial covering or unresolved saturation affects the line profiles , as discussed further in [ sec:3 ] . both and are located outside of the forest , and are detected at @xmath72 significance .
the @xmath562515 detection is within the forest , and can not be confirmed by other transitions .
however , we expect that it is valid because of the precise alignment with the @xmath73=-69 cloud .
we thus conclude that there is more than sufficient evidence that absorption in neutral species is detected , making cloud 1 in this system unique among other weak systems @xcite .
the ratio of the @xmath56 2803 to @xmath56 2796 equivalent width in cloud 1 of the vlt spectrum ( fig .
[ fig : clouds_labeled ] ) is not @xmath74 as expected @xcite , and as seen in clouds 2 - 4 , but is @xmath75 ; the weaker member is considerably stronger than would be expected .
this ratio implies either unresolved saturation of the line profile , or partial covering of the quasar broad emission line region ( belr ) .
if the profile were unresolved and saturated , profile fits to would severely underestimate the column density and overestimate the doppler parameters of fits to the lines .
we consider this possibility in defining a range of models to be considered .
we initially perform voigt profile fitting on the using minfit @xcite .
we choose to optimize on ( to require cloudy models to reproduce exactly the observed value ) because it is the only ion detected in all four clouds , and is the strongest ion detected for this @xmath51 absorber outside of the forest .
the doppler parameters ( @xmath76 ) and column densities ( @xmath77 ) given from the fit for clouds 2 - 4 are in table [ tab : nb ] , along with the redshifts of each cloud .
the profile for cloud 2 is asymmetric , and we find that a two component fit provides a significantly better fit than does one component .
these components are denoted as 2a and 2b in table [ tab : nb ] .
we note that the resolution of the spectrum , @xmath54 , is greater than the value of @xmath76 for four out of the five voigt profile components , implying that the clouds may not be fully resolved .
the best fit to the cloud 1 vlt spectrum is provided by a model with column density @xmath78 $ ] and doppler parameter @xmath79 , however this fit overproduces @xmath562796 and underproduces @xmath562803 , as expected by the difference in @xmath80 and the ratio of the oscillator strengths of the two transitions
. fits to the other clouds were adequate .
because of the likelihood of unresolved saturation , we increase @xmath77 until the ratio of equivalent widths of the synthetic profile matches that of the observed profile .
we also consider smaller doppler parameters in our modeling process , and for these values we adjust @xmath77 accordingly .
we also consider the possibility of partial covering of the belr of he0001 - 2340 to explain the observed equivalent widths .
the size of the broad emission line region of he0001 - 2340 can not be calculated directly , as the spectrum is not flux calibrated .
however , via comparison to a quasar of similar redshift and magnitude with a flux calibrated spectrum , we may approximate the belr size . with @xmath81 and @xmath52 , he0001 - 2340
may be compared to s5 0836 + 71 with @xmath82 and @xmath83 .
@xcite gives the following relation for belr size : @xmath84=-22.198 + 0.539 \log \lambda l_{\lambda } ( 5100 \rm{\aa})\ ] ] with @xmath85 , we estimate that the belr size of he0001 - 2340 is @xmath86 pc . therefore
, partial coverage of the belr requires a very small cloud size .
we explore this possibility as an alternative to unresolved saturation , and perform vp fitting to the doublet for cloud 1 , using a modified version of minfit @xcite in which the covering factor , @xmath87 , is allowed to vary along with @xmath77 and @xmath76 . a covering factor of @xmath88 best fits the profile , as determined by the @xmath89 minimization technique described in @xcite , and applied to a large doublet sample in @xcite .
the @xmath42 and @xmath90 values for this @xmath91 are @xmath92 and 3.1 , respectively . since we would not expect to find evidence for partial covering in an absorber that is not intrinsic to a quasar , we examine how significant an improvement @xmath93 provides as compared to other possible values of the covering factor .
we force @xmath91 to have various other values and in each case find the best @xmath77 and @xmath76 and compute the @xmath89 comparing the best model to the data .
figure [ fig : cf ] shows that a clear minimum in @xmath89 occurs at @xmath93 .
the @xmath91 measured from vp fitting is an `` effective covering factor '' , representing the absorption of the total flux at that wavelength , which is a combination of flux from the quasar continuum source and belr @xcite .
the different transitions for the intervening @xmath51 system fall at different positions relative to the quasar emission lines , and will therefore absorb different relative fractions of continuum and belr flux . in general , @xmath94 , where @xmath95 is the covering factor of the continuum , @xmath96 is the covering factor of the belr , and @xmath91 is the total covering factor @xcite .
the value of w , the ratio of the broad emission - line flux to the continuum flux at the wavelength of the narrow intervening absorption line ( @xmath97 ) , can be determined for each transition using a low resolution spectrum @xcite . however , in order to calculate the effective covering factor , @xmath91 , we must make an assumption about the relative covering factors of the continuum and belr . for , the value @xmath98 can correspond to a range of possible @xmath95 , @xmath96 pairs .
however , we know that @xmath95 can not be very small , as we see detections of many transitions that are not superimposed on an emission line . since the continuum source is likely to be considerably smaller than the belr , it is most straightforward to assume that it is fully covered and that the belr is partially covered . for the w value measured for ( @xmath99 ) , and @xmath100 , we then find @xmath101 .
@xmath96 values should be the same for all transitions , if their absorption arises from the same cloud . with this assumption , using the w values of each transition
, we can compute the effective covering factors as @xmath102 .
table [ tab : ew2 ] lists the @xmath103 values for all transitions detected from cloud 1 .
many of the detected neutral transitions have @xmath103 values that are close to @xmath104 , rendering their absorption stronger relative to the , which only partially absorbs a significant fraction of the incident flux due to its position on a broad emission line . for each model ,
we begin with a column density , doppler parameter , and covering factor for . for cloud 1
, we consider a range of @xmath42 , @xmath90 pairs that are consistent with the data , including fits affected by unresolved saturation . starting with the column density as a constraint
, we used the code cloudy ( version 07.02.01 ) to conduct photoionization models @xcite . for each cloud , we assume a plane - parallel slab with the given @xmath42 and illuminate it with a haardt - madau extragalactic background spectrum , including quasars and star - forming galaxies ( with an escape fraction of 0.1 ) at @xmath51 @xcite . given that the absorption is so weak in , it seems unlikely that this absorber is housed in the midst of a galaxy where the local stellar radiation field would be significant .
we initially run a grid of models for ionization parameters @xmath105 to @xmath106 , and metallicities @xmath107 to @xmath108 .
all models assume a solar abundance pattern , unless otherwise stated .
the cloudy output includes model column densities for all transitions , as well as an equilibrium temperature , @xmath109 , for the cloud .
the turbulent component of the doppler parameter ( @xmath110 ; where the thermal component @xmath111 ) is calculated from the equilibrium temperature and observed @xmath112 . given @xmath113 and @xmath114 ,
the doppler parameter can be computed for all other elements .
after running cloudy for all components from the voigt profile fit , the @xmath76 parameters and @xmath91 values are combined with the model column densities to create a synthetic spectrum .
this model spectrum is compared to the data in order to constrain the @xmath115 and @xmath116 values .
the modeling method is the same as that used by @xcite , @xcite , and @xcite .
cloud 2 can be fit with @xmath117 ; lower values overproduce , while higher values underproduce .
although there are no lyman series lines to use for direct constraints on the metallicity , we find that if log@xmath118 , the is overproduced .
the temperature of the cloud is found to be @xmath119k for @xmath120 , with a size of tens of parsecs , and a density @xmath121x@xmath122 .
model parameters are listed in table [ tab : models ] , where clouds @xmath123 and @xmath124 represent the two components of cloud 2 , while the best model is shown in figure [ fig : system_plot_unresolved ] .
clouds 3 and 4 are quite weak , and are detected in the vlt spectrum only because it has such a high s / n . for clouds 3 and 4 , a wide range of metallicities and ionization parameters provide an adequate fit to the transitions covered by the spectrum .
low ionization species other than are not detected at this velocity , but because of the weakness of the lines , are also not predicted by models with any reasonable ionization parameter , @xmath125 .
coverage of higher ionization states would be needed to further constrain the ionization parameter of these clouds .
constraints for these clouds are summarized in table [ tab : models ] , where parameters are given for two acceptable values of @xmath115 and log@xmath126 for each cloud .
we run a grid of models with varying column densities , metallicities , and ionization parameters , as described in section
[ sec:3 ] , for @xmath127 , the doppler parameter given by a voigt profile fit , which assumes the line is resolved , unsaturated , and fully covered .
we find that detectable amounts of are not produced in any model , and that is always underproduced . for smaller @xmath76 parameters we would expect that @xmath128 and @xmath129 would be larger because and would be on the linear parts of their curves of growth where the corresponding larger @xmath77 would affect their equivalent widths .
in contrast , the would be on the flat part of its curve of growth so that the increased @xmath77 would have little effect on its equivalent width
. we therefore considered @xmath42 , @xmath130 pairs with @xmath131 ( giving @xmath132 ) and @xmath133 .
an example model , with log@xmath134 , @xmath135 , and @xmath136 is shown in fig .
[ fig : system_plot_unresolved ] . due to the seemingly cold nature of this cloud , we opt to add dust grains to the cloudy @xcite simulations
. the primary flaw of this model is the over - production of @xmath562853 by @xmath137 dex in column density ; the apparent underproduction of some and transitions may be attributed to forest contamination of the observed profiles .
greater metallicities further overproduce , and a small fe abundance enhancement of @xmath138 dex is needed to account for observed and profiles .
constraints for this model are given in table [ tab : models ] , under cloud @xmath139 .
we find a size of 0.01 - 0.6 pc , t@xmath3 k , and @xmath140 [ ] . with a neutral hydrogen column density of @xmath141 ( sub - dla to dla range ) , we find @xmath142 . @xmath143 and @xmath144 .
we similarly find that a small @xmath76 ( @xmath145 ) is necessary to reproduce the observed with partial coverage models .
an example of such a model , with @xmath146 , @xmath147 , and @xmath148 is given in fig .
[ fig : system_plot_partial ] . for this example model ,
the observed , , , , , and are adequately reproduced within the uncertainties .
the absorption at the position of @xmath562484 and @xmath562524 is not fully produced , however the location of these transitions in the forest makes contamination fairly likely .
the only discrepancy of this model is the over - production of @xmath562853 by @xmath137 dex in column density .
in addition to this sample model , a range of ionization parameters and metallicities provide a similar fit .
for @xmath148 and @xmath149 the cloud size / thickness is 0.2 pc , a size comparable to that of the quasar belr .
the range of models that provide an adequate fit to this cloud have extreme properties .
the ionization parameters for successful models range from @xmath150 to @xmath151 , implying densities of @xmath152 . for the range of possible metallicities , the equilibrium temperatures are low , @xmath153 k. the neutral hydrogen column densities , @xmath154 ) , are in the sub - dla range , while @xmath155 .
cloud properties for acceptable models are summarized in in table [ tab : models ] under cloud @xmath156 , while a sample model is plotted in figure [ fig : system_plot_partial ] .
the @xmath51 system toward he0001 - 2340 , shown in figures [ fig : clouds_labeled ] and [ fig : system_plot_unresolved ] , significantly differs from typical weak absorption line systems , due to the detection of low ionization states in one of four main absorption components .
we consider the environments of these clouds in the context of known structures that could produce these absorption signatures .
cloud 2 has conditions similar to those of the 100 weak systems modeled by @xcite , with @xmath157 , @xmath158 , a size of tens of parsecs , a density of @xmath159 , and a temperature of @xmath119k ( table [ tab : models ] ) .
clouds 3 and 4 also fall within the range of properties exhibited in the @xcite sample , and although their properties are not well - constrained by models , they appear to be of a temperature and density typical of the weakest absorbers studied to date .
detection of such low @xmath42 absorbers is limited to very high s / n spectra , suggesting that such clouds may often go undetected near what are perceived as single - cloud weak absorbers .
@xcite divides weak absorbers at @xmath160=-0.3 $ ] into iron - rich and iron - poor subcategories .
while it is not possible to classify clouds 3 and 4 , cloud 2 , with log(n(feii)/n(mgii))=-0.62 , falls into the `` iron - poor '' subcategory , with corresponding @xmath161 .
clouds associated with superwind condensations could be responsible for this iron - poor environment , as type ii supernovae driving the wind will lead to @xmath162-enhancements , as they build up high metallicities .
such systems may be predecessors to the high metallicity ( @xmath163 ) absorbers of @xcite , which have sizes @xmath164 pc .
with radii less than the jean s length for self - gravitating clouds , these high metallicity clouds are likely to be short - lived , with lifetimes on the order of the time - scale for free expansion , @xmath165 years . in comparison ,
cloud 2 would have a lifetime on the order of @xmath166 years in its present state , which could precede its cloud phase .
based solely upon its detected absorption features , cloud 1 is an anomaly .
has only been detected along a fraction of galactic sightlines passing through molecular clouds @xcite .
these sightlines pass through clouds that are very cold ( @xmath3k ) , and have densities ranging from @xmath167 , on par with the results given by photoionization models of cloud 1 .
partial covering and unresolved saturation models of cloud 1 in the vlt spectrum provide similar constraints to its properties as those dense galactic clouds .
both require a narrow 2796 profile , with @xmath168 , to reproduce the observed , suggesting that regardless of whether the system is partially covering the belr , it is unresolved .
we note that the other properties of cloud 1 given by cloudy are similar in both models , as shown in table [ tab : models ] .
such models are not mutually exclusive , as it would be impossible to distinguish partial covering from unresolved saturation based on a comparison of the members of the doublet . with either model ,
the size of the absorber is likely to be on the order of the qso belr size at the distance of the absorber ( parsec - scale ) , thus partial covering is not unlikely .
it is important to note that the covering factors of different transitions depends upon their position on broad emission lines , a concept that is key to intrinsic absorption line studies , as partial covering by clouds in the immediate vicinity of the quasar is common . in a sample of 81 vlt qsos , with a redshift path length of @xmath169 @xcite ,
we have found only one system with detected .
however , considering the small size of this object , even this one detection is significant , suggesting that there may be a significant cross section covered by these objects , which lie in the outskirts of galaxies and remain undetected . assuming the same average redshift pathlength per quasar , @xmath170 , and a poisson distribution for the detections
, we find that there is a @xmath171 chance of one or more such detections in a sample of 56 more quasars , and a @xmath172 chance of one or more detections in a sample of 243 more quasars ( which corresponds to 332 more weak absorption line systems ) .
the one system that was detected in a pathlength of @xmath173 yields @xmath174 , and there is a probability of 95% that @xmath175 ( based on one detection in a pathlength of @xmath176 ) .
@xcite finds that , for a range of schechter luminosity function parameters , a statistical absorption radius of @xmath177 kpc would explain the observed @xmath178 of strong absorbers ( @xmath12 ) at @xmath179 . to consider the radius specifically at the redshift of the @xmath51 absorber , we scale those results considering that @xcite find @xmath180 for strong absorbers at @xmath1 .
this implies a corresponding absorber halo radius of 24 - 49 kpc , assuming a covering factor of order unity . from the estimated @xmath181 of absorbers , we estimate that approximately @xmath182 of the area covered by strong absorbers ( @xmath183-@xmath184 kpc@xmath185 ) is covered by such small molecular clouds .
these clouds are likely to take the form of filaments or thin sheets around many galaxies .
it is therefore of interest to consider what the covering factor would be of molecules in the milky way halo , if we were to observe it from outside the halo .
the covering factor of 21-cm high velocity clouds , to a limit of @xmath186 , was measured to be 37% , looking out from our vantage point .
although molecules have been detected in several hvcs @xcite , the fraction of hvc sightlines with detected h@xmath187 is small , e.g. 1/19 in the fuse survey of @xcite .
nonetheless , looking through the milky way from outside , we might expect roughly @xmath188%@xmath189% as the covering factor for molecular absorption , consistent with the 2% halo covering factor that we estimated for absorbers at @xmath1 .
an alternative to a large fraction of galaxies producing absorbers with small individual covering factors is a small population of galaxies that which have larger individual covering factors .
one possibility is starburst galaxies with superwinds , since the neutral species is commonly detected in absorption through these objects @xcite .
nearby starbursts are found to have strong absorption over regions @xmath104-@xmath190 kpcs in extent .
weaker absorption , consistent with its expected strength for cloud 1 in the @xmath51 absorber , would be detected over larger areas .
also , starbursts are more common at @xmath191 than at the present .
these factors could combine to lead to a significant covering factor of ( and therefore ) absorption from starburst winds at the redshift of the @xmath51 absorber .
the model gives neutral hydrogen column densities of the absorber that suggest a sub - dla or dla environment , with @xmath192 .
although h@xmath187 is frequently detected in higher redshift systems , the molecular hydrogen fraction is only @xmath193 in @xmath194 of the high - z dlas surveyed by @xcite , and does not rise above @xmath195 for sub - dlas and dlas .
@xcite perform simulations of high - z dlas , and find that the area of a region that would produce dla absorption that contains molecules is a very small fraction of its overall area .
they note that in high ultraviolet background , the molecular fraction can be @xmath196 , which is large enough to fuel star formation .
small , dense ( @xmath197 ) , cold ( @xmath3k ) absorbers , with significant molecular fractions , like the @xmath51 absorber toward he0001 - 2340 , may occur over a small fraction of structures that give rise to sub - dlas and dlas , but are rarely intersected by qsos due to their relatively smaller area .
while extremely high molecular hydrogen fractions seem to be rare in sub - dlas and dlas , up to @xmath198 of total hydrogen mass was found to be molecular in a survey of @xcite , which examined co emission from 18 nearby galaxies . in these detections
, the h@xmath187 mass can be as high as 0.1 times the stellar mass .
the ratio of co @xmath199 intensity to co @xmath200 intensity suggests that the emitting gas is optically thick with an excitation temperature of @xmath201k .
although the exponential scale lengths of the survey targets range only from 0.8 - 3.2 kpc , they note that with increased sensitivity , objects at larger distances could be detected .
similar high resolution co observations of structures likely belonging to the milky way by @xcite show fractalized clumpuscules at high galactic latitudes , with sizes of @xmath164 au , masses @xmath202 , @xmath203h@xmath204 , and column densities of @xmath205 ; he notes that such small structure would not be visible in the inner galaxy , due to overcrowding in normal molecular gas .
a previous study of high galactic latitude molecular clouds via stellar absorption lines by @xcite likewise found high molecular hydrogen fractions , high densities , and small sizes .
@xcite and @xcite suggest such small structures may make up a significant percentage of halo matter , and propose that such structures may account for a fraction of missing baryonic dark matter .
these structures are believed to be formed by jeans unstable and isothermal clouds , which fragment into smaller cloudlets .
an alternative to this fractalized ism model is explored through simulations by @xcite .
they study h@xmath187 formation in turbulent gas , predicting that h@xmath187 should rapidly form in any dense turbulent atomic cloud ( within 0.05 myr ) , regardless of whether or not the cloud is gravitationally bound or whether it is magnetic subcritical or supercritical ; up to @xmath206 of the initial atomic hydrogen may be converted to h@xmath187 in this process .
they find that regions of low density ( @xmath207 ) contain more h@xmath187 than expected in gas in photodissociation equilibrium due to turbulent transport of h@xmath187 from higher density ( @xmath208 ) areas .
hydrodynamic simulations by @xcite of the blowout of starburst - driven superbubbles from the molecular disk of ulirgs also show similar small - scale clumps .
thin dense shells are created as the bubbles sweep up dense disk gas ; rayleigh - taylor instabilities then cause these shells to fragment .
clumpiness is seen down to the limits of the simulation resolution of 0.1 pc .
these clumps contain high mass , and have @xmath209 . as small molecular structures have been observed both in the galactic neighborhood ( on the order of hundreds of au),and in hydrodynamic simulations ( to the limit of their resolutions ) one would expect there to exist such structures beyond the local universe , at higher redshifts . because these structures are far too small to detect via imaging of extragalactic environments , absorption lines serve as the only probe of small - scale structure of distant galaxies .
the @xmath51 absorber may be a first glimpse of such a cold , dense cloud at an earlier epoch .
cloud 1 , with @xmath210(/)@xmath211 , is an extreme case of an iron - rich absorber .
iron - rich systems discussed in @xcite were generally found to have high metallicities ( @xmath212 ) , @xmath213 ( @xmath214 ) , and small sizes ( @xmath77()+@xmath77()@xmath215 , @xmath216 pc ) .
solar abundances , and not @xmath217-enhancement , are needed to explain their similar and column densities , which implies type ia event - enrichment of the gas @xcite .
type ia enrichment requires a @xmath137 billion year delay from the onset of star formation before the elements produced enter the interstellar medium , which could be consistent with gas trapped in potential wells of dwarf galaxies , or with intergalactic star - forming structures in the cosmic web @xcite .
rapid condensation of turbulent gas from a supernova could explain the small size , high molecular content , and high iron abundance observed in cloud 1 .
weak absorption systems are not often found within 30 kpc of luminous star - forming galaxies , with @xmath218 @xcite ; the astrophysical origins of such systems have not been identified , although they may include extragalactic high velocity clouds @xcite , dwarf galaxies @xcite , material expelled by superwinds in dwarf galaxies @xcite , and/or massive starburst galaxies and metal - enriched gas in intergalactic star clusters @xcite .
@xcite suggest that at intermediate redshifts ( @xmath219 ) , weak clouds may have been ejected from starburst supernova - driven winds during an intermediate phase of free expansion , prior to achieving equilibrium with the igm .
supernova - driven winds are believed to have a multi - phase structure , with a cold component ( @xmath220k ) detected in @xcite . a warm neutral phase ( @xmath221k )
may surround this component @xcite .
this gas will likely fragment through hydrodynamical instabilities as it moves through the halo of the galaxy , and would appear initially as weak absorption , and later as weak absorption associated with lines in the forest @xcite . to account for the differing
/ observed in the different clouds of the same system , it seems necessary to propose different stellar populations in the vicinity of the absorbing structure .
cloud 1 , due to its iron - rich nature , must be enhanced by a type ia supernova , implying the gas does not originate from a very young stellar population .
the density and temperature of this gas , comparable to that of milky way molecular clouds , suggests that it is a potential site of star formation .
cloud 2 , in contrast , is likely a remnant of a supernova from a massive star recently formed in its vicinity , or in a superwind driven by a collection of massive stars , as suggested by its @xmath162-enhancement .
it is not surprising that we would see a mix of different processes and stellar populations in an absorption line system .
some combination of dwarf galaxies , high velocity clouds , superwind and supernova shell fragments , and tidal debris could give rise to such variety .
in such a scenario , clouds like cloud 2 would commonly be grouped together into systems , sometimes with fe - rich clouds , but something as extreme as cloud 1 would be rare .
although detections are extremely rare in extragalactic absorbers , photoionization models of cloud 1 suggest that its small size ( @xmath222pc ) may cause many such absorbers to go undetected ; we estimate that two percent of the areas of @xmath191 halos ( 24 - 49 kpc in radius ) should be covered by such objects . with cold temperatures ( @xmath3k ) , high densities ( @xmath223 ) , and a large molecular hydrogen fraction ( @xmath224 ) , the properties of cloud 1 are similar to the dense , small milky way clouds at high galactic latitudes observed in co by @xcite and @xcite , suggesting that pockets of gas like cloud 1 may be analogs of these milky way clouds in the halos of other galaxies .
small scale clumps , with sizes of hundreds of au , are predicted to exist both as part of the fractalized ism in the halos of galaxies @xcite and as condensates in turbulent gas @xcite .
a type - ia supernova may have been responsible for both the turbulence and observed iron abundance in cloud 1 .
in contrast , the @xmath162-enhancement observed in cloud 2 , with its low - fe content , is suggestive of an origin in a type ii supernova - driven superwind .
multiple - cloud weak absorption line systems are generally thought to originate in lines of sight that pass through more than one dwarf galaxies , through sparse regions of luminous galaxies at high impact parameters , or in gas - poor galaxies . in this case
, one of the four components happens to be a relatively rare type of cold and dense weak absorber . via imaging of the he0001 - 2340 field
it may be possible to further constrain the absorption origin , perhaps by identifying a host galaxy or galaxies , or by finding a nearby galaxy whose environment is being sampled by these absorbers .
it is also of interest to find another similar cloud , by searching large numbers of high-@xmath225 , high - resolution quasar spectra .
perhaps in another case , the sightline would be clear enough to allow access to the lyman series lines , so that its metallicity could be directly measured , or to higher ionization absorption lines so that any surrounding , lower density gas phases could be identified .
ccccccc + @xmath104 & 0.452061 & @xmath228 & @xmath229 & 0.028 @xmath230 0.007 & 0.0392 @xmath230 0.0006 & 0.0128 @xmath230 0.0006 + @xmath231 & 0.452387 & 0.0761 @xmath230 0.0004 & @xmath232 & @xmath233 & @xmath234 & @xmath235 + @xmath236 & 0.452622 & 0.019 @xmath230 0.001 & @xmath237 & @xmath238 & @xmath238 & @xmath239 + @xmath240 & 0.454864 & 0.009 @xmath230 0.001 & @xmath241 & @xmath239 & @xmath242 & @xmath239 @xmath139 & @xmath243 & @xmath244 & 500 - 1100 & 0.01 - 0.6 & @xmath3 & @xmath245 & @xmath246 & @xmath247 & @xmath248 & 0.72 - 0.91 & @xmath249 & @xmath250 & @xmath251 & @xmath252 & @xmath253 + @xmath254 & @xmath255 & @xmath256 & 30 - 1100 & 0.08 - 0.19 & @xmath257 & @xmath258 & @xmath259 & @xmath260 & @xmath261 & 0.76 - 0.94 & @xmath262 & @xmath263 & @xmath264 & @xmath265 & @xmath253 + @xmath123 & @xmath266 & @xmath267 & 0.004 & 15 & 9200 & @xmath268 & @xmath269 & @xmath270 & @xmath269 & 0 & @xmath271 & @xmath272 & @xmath273 & @xmath274 & 0.4 + @xmath124 & -0.3 & @xmath267 & 0.004 & 57 & 9100 & @xmath275 & @xmath276 & @xmath277 & @xmath278 & 0 & @xmath279 & @xmath280 & @xmath281 & @xmath282 & 2.77 + @xmath283 & @xmath284 & @xmath285 & 110 & 0.0004 & @xmath286 & @xmath287 & @xmath288 & @xmath289 & @xmath290 & 0 & @xmath291 & @xmath292 & @xmath293 & @xmath294 & @xmath295 + @xmath296 & @xmath297 & @xmath267 & 0.004 & 110 & 9200 & @xmath298 & @xmath299 & @xmath300 & @xmath301 & 0 & @xmath302 & @xmath303 & @xmath304 & @xmath305 & 14.53 + @xmath306 & @xmath284 & @xmath285 & 110 & 0.0002 & 3250 & @xmath307 & @xmath308 & @xmath309 & @xmath310 & 0 & @xmath311 & @xmath312 & @xmath313 & @xmath314 & 6.00 + @xmath315 & @xmath297 & @xmath267 & 0.004 & 5 & 9200 & @xmath316 & @xmath310 & @xmath317 & @xmath310 & 0 & @xmath318 & @xmath319 & @xmath320 & @xmath321 & 6.00 + + | several neutral species ( , , , ) have been detected in a weak absorption line system ( @xmath0 ) at @xmath1 along the sightline toward he0001 - 2340 .
these observations require extreme physical conditions , as noted in @xcite .
we place further constraints on the properties of this system by running a wide grid of photoionization models , determining that the absorbing cloud that produces the neutral absorption is extremely dense ( @xmath2 ) , cold ( @xmath3k ) , and has significant molecular content ( @xmath4 ) .
structures of this size and temperature have been detected in milky way co surveys , and have been predicted in hydrodynamic simulations of turbulent gas . in order to explain the observed line profiles in all neutral and singly ionized chemical transitions , the lines must suffer from unresolved saturation and/or the absorber must partially cover the broad emission line region of the background quasar .
in addition to this highly unusual cloud , three other ordinary weak clouds ( within densities of @xmath5 and temperatures of @xmath6 k ) lie within 500 along the same sightline .
we suggest that the `` bare molecular cloud '' , which appears to reside outside of a galaxy disk , may have had in situ star formation and may evolve into an ordinary weak absorbing cloud . |
many elementary problems in geometric combinatorics ask how often a particular type of point configuration can occur in subsets of some ambient space .
one of the most famous is the erds single distance problem , which asks how often any fixed distance can occur in a large finite set of points in the plane .
the conjecture is that for a set of @xmath3 points , no distance can occur more than @xmath4 times , for some constant @xmath5 , independent of @xmath3 , and any @xmath6 .
the best known estimate of this is due to spencer , szemer ` edi , and trotter , in @xcite , who have shown that @xmath7 is an upper bound .
a closely related problem , the erds distinct distances problem , asks for a lower bound on the number of distinct distances determined by point pairs from a large finite point set .
this was resolved in the plane by guth and katz , in @xcite .
analogous questions have been studied for dot products ; see @xcite , @xcite , and @xcite . here , we consider triples of points which determine a pair of dot products in large finite point sets . in the settings of vector spaces over various finite rings
, there has been activity on the special case of zero dot products by the second listed author , and iosevich @xcite , as well as pham and vinh , in @xcite .
information about the dot products determined by a point set finds applications in varied areas such as coding theory , @xcite , graph theory , @xcite , and frame theory , @xcite .
we now fix some notation . in
what follows , if two quantities , @xmath8 and @xmath9 , vary with respect to some natural number parameter , @xmath3 , then we write @xmath10 if there exist constants , @xmath5 and @xmath11 , both independent of @xmath3 , such that for all @xmath12 , we have @xmath13 .
if @xmath10 and @xmath14 , we write @xmath15 given a set of points , @xmath16 ^ 2 $ ] , let @xmath17 denote the number of distinct triples of points , that determine a given pair of dot products .
that is , for real numbers @xmath18 and @xmath19 , @xmath20 we will typically restrict @xmath18 and @xmath19 to be positive nonzero , because zero dot products behave differently , as demonstrated by proposition [ zerodps ] .
our first main result applies to any point set in the plane .
[ general ] given a set , @xmath21 , of @xmath3 points in @xmath22 , and fixed @xmath23 , @xmath24 in general , for a set of @xmath3 points , @xmath0 , one can not expect to get an upper bound better than theorem [ general ] , as shown in an explicit construction below , proposition [ sharp ] . in cases where the points are evenly - distributed , such as applications with sensor placement or code construction , we have tighter bounds on @xmath25 .
our second main result is for point sets with a minimum separation between points .
[ separation ] let @xmath26 ^ 2 $ ] be a set of @xmath3 points that obeys the following separation condition @xmath27 for @xmath28 , and fixed @xmath23 , we have @xmath29 notice that if @xmath30 is chosen to be too small , theorem [ separation ] is outdone by theorem [ general ] . similarly ,
if @xmath30 is close to 1 , there can not be many points in the unit square . keeping this in mind , the range in which theorem [ separation ] is most useful is @xmath31 .
this range lines up with other results on a wide class of finite point sets called @xmath32-adaptable sets .
we now introduce the notion of @xmath32-adaptability .
this should be viewed as a measure for how well - distributed the points are .
this property has been used to study many types of geometric point configuration problems .
see @xcite , @xcite , and @xcite , for example .
families of point sets which are @xmath32-adaptable can be used to transfer results between discrete point sets and sets with positive hausdorff dimension . a large , finite point set @xmath26 ^ 2 $ ] ,
is said to be _
@xmath32-adaptable _ if the following two conditions hold : @xmath33 by setting @xmath34 , and appealing to the definition of @xmath32-adaptability given here , we get the following estimate as a corollary .
[ main ] let @xmath26 ^ 2 $ ] be a set of @xmath3 points that is s - adaptable . for @xmath35 , and fixed @xmath36 , @xmath37 first , we construct examples of point sets that illustrating the sharpness of theorem [ general ] as well as an illustration of why we assume the restriction of @xmath38 .
next , we prove theorem [ general ] and theorem [ separation ] in section [ proofs ] .
section [ lemmaproofs ] contains the proofs of two technical lemmas , included for completion .
[ sharp ] given a natural number @xmath3 , and real numbers @xmath39 , there is a set , @xmath21 , of @xmath3 points in @xmath40 ^ 2 $ ] for which @xmath41 let @xmath42 be the point with coordinates @xmath43 .
now , staying within the unit square , distribute @xmath44 points along the line @xmath45 , and distribute the remaining @xmath46 points along the line @xmath47 .
clearly , there are @xmath48 pairs of points @xmath49 , where @xmath50 is chosen from the first line , and @xmath51 is chosen from the second .
notice that @xmath42 contributes a triple to @xmath25 for each such pair , giving us @xmath41 [ zerodps ] there exists a set , @xmath21 , of @xmath3 points in @xmath40 ^ 2 $ ] for which @xmath53 arrange @xmath54 along the @xmath55-axis , and @xmath54 points along the @xmath56-axis .
now , for each of the @xmath54 points on the @xmath55-axis , there are @xmath57 pairs of points on the @xmath56-axis .
notice that any point chosen from the @xmath55-axis will have dot product zero with each point from the pair chosen from the @xmath56-axis .
therefore , each of these @xmath58 triples will contribute to @xmath59 .
we can get just as many triples that contribute to @xmath59 by taking single points from the @xmath56-axis , and pairs of points from the @xmath55-axis . in total
, we get @xmath60
the basic idea is to estimate the number of triples , @xmath61 , by considering pairs of points , @xmath62 , and bounding the number of possible points , @xmath63 , which could contribute to @xmath25 .
call any line through the origin a _ radial line_. let @xmath63 have coordinates @xmath64 .
we define the _ @xmath18-line _ for a point , @xmath42 , to be the set of points that have dot product @xmath18 with the point @xmath42 . holding @xmath42 fixed
, this set will be a line of points @xmath50 , with coordinates @xmath65 satisfying the equation @xmath66
+ we can define a _
@xmath19-line _ similarly . by solving for @xmath67
, we can see that the slope of the @xmath18-line of a point , @xmath42 , will be equal to the slope of the @xmath19-line of the point @xmath42 . moreover , these lines will be perpendicular to the radial line through @xmath42 .
we let @xmath68 and @xmath69 denote the set of all points of @xmath21 incident to the @xmath18-line or @xmath19-line for the given point @xmath42 .
draw @xmath18-lines and @xmath19-lines for any given pair of @xmath70
. we will refer to the set of point pairs with distinct @xmath18-lines and @xmath19-lines as @xmath71 and pairs with a shared line as @xmath72 . to be precise :
@xmath73 @xmath74 we first consider triples in @xmath25 of the form @xmath75 where @xmath76 .
notice that the @xmath18-lines and @xmath19-lines of a pair of points , @xmath76 , can intersect at most four times , by definition of the set @xmath71 .
so for every pair of points in @xmath71 , there are at most four possible locations for a point in @xmath21 which would contribute a triple to @xmath25 . as @xmath77
, we see that @xmath78 . from this , we see that pairs in @xmath71 can not add more than @xmath79 triples to @xmath25 . //
+ we now turn our attention to the set @xmath72 . without loss of generality ,
suppose that the @xmath18-line of a point , @xmath80 , coincides with the @xmath19-line of a point , @xmath81 , then we appeal to the following lemma .
[ setb ] for any pair of points , @xmath82 , as defined above , the following hold : 1 .
both @xmath50 and @xmath51 must lay along the same radial line .
the ratio of the distances from @xmath50 and @xmath51 to the origin must equal the ratio between @xmath19 and @xmath18 .
lemma [ setb ] is proved in section [ lemmaproofs ] . with these conditions in tow , we see that the pairs of points in @xmath72 are quite rare .
fix any radial line , @xmath83 .
each point from @xmath84 can have its @xmath18-line overlap with at most one @xmath19-line .
similarly , each point from @xmath84 can have its @xmath19-line overlap with at most one @xmath18-line .
so each point from @xmath84 can be in at most two pairs from @xmath72 .
+ any pair , @xmath85 that lies on @xmath83 will contribute as many triples of points , @xmath75 , to @xmath25 as there are points coincident to both the @xmath18-line of @xmath50 and the @xmath19-line of @xmath51 .
as these families of shared lines are parallel for point pairs along @xmath83 , each point , @xmath42 , can be on at most one @xmath18-line , regardless of a possible overlap with a @xmath19-line .
this means that each point @xmath63 can be in at most one triple of the form @xmath86 with a pair of points , @xmath50 and @xmath51 , from @xmath83 .
the total number of triples contributed by pairs of points in @xmath83 is therefore no more than @xmath3 .
as there are no more than @xmath3 points , there can be no more than @xmath3 distinct radial lines to consider .
since each radial line can contribute no more than @xmath3 triples to @xmath25 , the maximum contribution to @xmath25 by pairs in @xmath72 is no more than @xmath87 .
first , we define @xmath18-lines and @xmath19-lines as in the proof of theorem [ general ] .
again , we let @xmath68 and @xmath69 denote the set of all points incident to the @xmath18-line or @xmath19-line for the given point @xmath42 . now , consider the @xmath18-lines and @xmath19-lines for each @xmath88 .
referring back to the definition of @xmath25 , we see that a triple of points , @xmath75 , will be in @xmath25 precisely when @xmath50 lies on the @xmath18-line of @xmath42 and @xmath51 lies on the @xmath19-line of @xmath42 .
so the quantity we aim to estimate is latexmath:[\[\label{bigest } it follows that we want a bound on the number of times that a point , @xmath88 , is incident to an @xmath18-line or @xmath19-line .
the following result will be proved in section [ lemmaproofs ] .
[ stlemma ] in the setting above , the number of point - line incidences , @xmath90 , is @xmath91 recalling the definitions of @xmath92 and @xmath93 we write @xmath90 as : @xmath94 we now dyadically decompose @xmath21 into two families of disjoint sets defined by the number of incidences they contribute .
@xmath95 where @xmath96 is defined similarly .
now , if a point has roughly @xmath97 points from @xmath21 on its @xmath18-line , and @xmath98 points from @xmath21 on its @xmath19-line , then it will be in the intersection of @xmath99 and @xmath96 .
let us define these intersections as : @xmath100 the intersection of any @xmath18-line or @xmath19-line with @xmath40 ^ 2 $ ] can be no longer than @xmath101 .
all of our points are contained in @xmath40 ^ 2 $ ] , so it will suffice to estimate the maximum number of points of @xmath21 on any line segment of length @xmath102 .
fix such a segment , @xmath103 .
recall the separation condition , @xmath104 so each point on @xmath103 must have a vacant length of segment at least @xmath30 long in either direction .
so if @xmath103 had points packed on it maximally , there would be no more than @xmath105 we can see that for every point @xmath88 , @xmath106 is empty for @xmath107 or @xmath108 bigger than @xmath109 .
this also tells us that for all relevant indices @xmath107 and @xmath108 in the sums to follow , we have : @xmath110 by combining the above : @xmath111 so , for any pair of indices , @xmath107 and @xmath108 , we have the following bound latexmath:[\[\label{parta } we now dyadically decompose the sum in as we did with the sum estimating @xmath90 .
@xmath113 let @xmath114 and @xmath115 be a pair of indices that give the largest summand in .
now we have : @xmath116 burying the constants from the logarithms in , we get : @xmath117 , , and , we conclude @xmath118 as desired .
to see @xmath119 , notice that any line that yields the prescribed dot products for a given point is perpendicular to that point s radial line . for the @xmath18-line of @xmath50 and @xmath19-line of @xmath51 to be coincidental , they must both be perpendicular to the same radial line , and thus , be generated from two points upon the same radial line . for @xmath120
, assuming the @xmath18-line of @xmath50 is coincidental to the @xmath19-line of @xmath51 , we get the following two equations of lines in the plane : @xmath121 @xmath122 we know that @xmath50 and @xmath51 are on the same radial line , so there must exist a @xmath123 such that , @xmath124 we set the equations equal to one another and get : @xmath125 by substituting for the coordinates of @xmath50 , @xmath126 after simplifying , we see that @xmath127 there is a small technical obstruction to a direct application of the celebrated szemerdi - trotter point - line incidence theorem , which is that two types of line may be coincident .
this turns out to not be a problem in our case . by appealing to the definitions of @xmath18-line and @xmath19-line
, we can see that no two points can determine the same @xmath18-line or @xmath19-line .
however , as we have seen above , it is possible for the @xmath18-line of a point to overlap the @xmath19-line of a point ( a different point , unless @xmath128 ) .
we prove lemma [ stlemma ] using techniques of sz ' ekely , from @xcite .
draw a graph ( possibly a multigraph ) with the @xmath3 points as vertices , and the segments of the @xmath18-lines and @xmath19-lines connecting adjacent points as edges .
observe that no pair of points can have more than two edges connecting them , and this can only happen if the @xmath18-line of a point is coincident to the @xmath19-line of a point . if this does happen , we can replace the line segments by two curves , whose endpoints are the adjacent points in question , drawn in such a way so as not to cross any more or fewer edges than the initial line segment crossed . notice that the number of edges contributed by each line is equal to the number of point - line incidences on that line minus one ( for example , if there are seven points on a given line , we would connect consecutive points with six edges . ) .
so the number of edges is equal to the number of incidences minus the number of lines that contribute incidences .
there are exactly @xmath129 lines ( @xmath3 of each type ) that may contribute incidences .
let @xmath90 denote the number of point - line incidences , and @xmath130 denote the number of edges in the graph .
we have that @xmath131 we let @xmath132 denote the _ crossing number of @xmath133 _ , which is the maximum number times that edges of @xmath133 must cross one another at some point which is not a vertex , for any redrawing of @xmath133 .
we appeal to the crossing number lemma in @xcite . for our setup
, we have that @xmath137 , as there can be no more than two edges between any given pair of points .
now , either @xmath138 or @xmath135 . in the first case , we recall to see that @xmath139 in the case that @xmath135 , we appeal to lemma [ szek ] and get @xmath140 notice that the crossing number of the graph can be no more than the number of times that the @xmath18-lines and @xmath19-lines crossed one another . since there are @xmath129 total lines , and each line could potentially cross almost all of the others , the total number of crossing lines is bounded above by @xmath141 . as this would correspond to a drawing of the graph , @xmath133
, we are guaranteed that the crossing number of @xmath133 is no more than @xmath141 . comparing upper and lower bounds on @xmath132 yields
@xmath142 which gives us that @xmath143 in either case , the claimed estimate holds .
a. iosevich and s. senger , _ sharpness of falconer s estimate in continuous and arithmetic settings , geometric incidence theorems and distribution of lattice points in convex domains _
, arxiv:1006.1397 , ( 2010 ) . | given a large finite point set , @xmath0 , we obtain upper bounds on the number of triples of points that determine a given pair of dot products .
that is , for any pair of positive real numbers , @xmath1 , we bound the size of the set @xmath2 [ multiblock footnote omitted ] |
we would like to thank dmitri diakonov , victor petrov and georges ripka for helpful discussions .
the project has been partially supported by the vw stiftung , dfg and cosy ( jlich ) . | we study the pseudoscalar , vector and axial current correlation functions in su(2)-njl model with scalar and vector couplings .
the correlation functions are evaluated in leading order in number of colors @xmath0 . as it is expected in the pseudoscalar channel pions
appear as goldstone bosons , and after fixing the cutoff to reproduce the physical pion decay constant , we obtain well - known current - algebra results . for the vector and axial channels we use essentially that at spacelike momenta the correlation functions can be related to the experimentally known spectral density via dispersion relations .
we show that the latter imposes strong bounds on the strength of the vector coupling in the model .
we find that the commonly used on - shell treatment of the vector and axial mesons ( identified as poles at large timelike momenta ) fails to reproduce the behavior of the corresponding correlation functions at small spacelike momenta extracted from the physical spectral density .
the parameters of the njl model fixed by the correlation functions at small spacelike momenta differ noticeably from those of the on - shell treatment .
# 1#1 # 1 # 1 # 1#2#1 # 2 #
1#1 - 6pt/ = 0.4 cm rub - tpii-19/94 + december , 94 the correlation functions of different colorless hadronic currents built of quarks and gluons offer a general tool to study the structure of the qcd vacuum and the hadron spectrum @xcite .
one can obtain information about the physical intermediate states from the pole structure of the correlation functions in the timelike region as well as from its behavior at spacelike momenta .
for an exact theory both ways should lead to identical results .
this , however , is not necessarily true for the effective models , which in the present lack of non - perturbative solution to qcd attract an increasing interest . among them ,
the nambu jona - lasinio model , proposed @xcite in early sixties in analogy with the bcs theory of superconductivity , plays an essential role .
this model incorporates the basic symmetries of qcd and in particular , the chiral symmetry .
it is able to manifest the spontaneous chiral symmetry breaking and , as a natural consequence , the appearance of the pseudoscalar mesons as goldstone particles . as such it offers a simple but quite successful working scheme to study the role of the dynamical chiral symmetry breaking mechanism in hadron physics ( for review see refs .
@xcite ) . however , the model suffers from two important drawbacks .
first , it is not renormalizable and to make the theory finite one needs a finite cut off fixed in a consisitent way ( usually to reproduce the pion decay constant ) and the limit of large @xmath0 ( number of colors ) .
the typical value of the cut off is about 1 gev .
the second problem is the lack of confinement .
obviously , these drawbacks make the so far used on - shell treatment of the ( axial ) vector mesons questionable @xcite . in the present work
we will show that the proper way to overcome this problem is to consider the corresponding current correlation function at small spacelike momenta . in this case
we do not face the problem of the strong coupling to the @xmath1 continuum and we apply the model at small momentum region where it is designed for . to this end
we make essentially use of the fact that the correlator at small spacelike momenta can be related via dispersion relations to well - established rich phenomenology without detailed knowledge of the dynamics at large timelike momenta .
we will show that such a scheme allows to describe the vector modes in the meson sector of the njl model in a consistent way .
the simplest su(2 ) nambu jona - lasinio ( njl ) model is defined by lagrangean @xcite : @xmath2 \nonumber\\ & -&{g_v\over 2 } \ , [ ( \bar \psi\gamma_\mu \vec\tau \psi ) ^2 + ( \bar \psi \vec\tau\gamma_\mu \gamma_5 \psi ) ^2\ , ] \,,\label{njllag}\end{aligned}\ ] ] which contains chirally invariant local four - fermion interactions with scalar @xmath3 and vector @xmath4 couplings.the current quark mass @xmath5 stands for both @xmath6 and @xmath7 quarks ( assumed to be degenerated in mass ) . applying the well - known bosonization procedure @xcite we arrive at the model generating functional expressed as a functional integral over auxiliary boson fields @xmath8 : @xmath9\bigr)+ i_m[s,{\vec p},{\vec v_\mu},{\vec a_\mu}]}\ , , \label{genf}\ ] ] where the quarks are already integrated out . to simplify our notations we adopt for @xmath10 , @xmath11 and @xmath12 that @xmath13 etc . the euclidean dirac operator @xmath14 is given by @xmath15 and the trace over colors is explicitly done .
the symmetry breaking term is included in the boson part : @xmath16 in the large @xmath0 limit the integral in eq.([genf ] ) is given by its saddle point value .
for the translational invariant case ( vacuum ) we have a non - trivial stationary meson configuration given by @xmath17 which means that the chiral symmetry is spontaneously broken .
the quarks acquire a constituent mass @xmath18 which is related to the scalar coupling constant @xmath3 via the stationary condition ( saddle point ) : @xmath19}\bigr)+m_0=m_0-g_s < \bar \psi \psi > \ , . \label{gap}\ ] ] here @xmath20 is the chiral quark condensate .
as it was mentioned the model is not renormalizable and in order to make the fermion determinant finite it must be regularized .
thus , the model has four parameters , namely two coupling constants @xmath3 and @xmath4 , the current mass @xmath5 and the cut off @xmath21 .
usually they are fixed reproducing the physical pion decay constant and the physical masses of pion and rho meson .
it leaves one parameter free which is commonly chosen to be the constituent mass @xmath18 because with eq.([gap ] ) one can eliminate @xmath3 in favor of @xmath18 . in principle , one can use the empirical value for the quark condensate @xmath20 to constrain @xmath18 . however
, being quadratically divergent @xmath20 is very sensitive to the details of the regularization scheme . in this paper
we suggest an alternative way of fixing the model parameters @xmath3 , @xmath4 , @xmath5 and @xmath21 . to this end
we study the correlation functions of the pseudoscalar , vector and axial - vector currents : @xmath22 @xmath23 @xmath24 in the njl model .
the correlation functions are defined by the vacuum - to - vacuum matrix elements of the currents and can be expressed in terms of invariant functions as follows : @xmath25 @xmath26 @xmath27 where as usual @xmath28 is defined as @xmath29 .
in contrast to the vector current in the real world the axial - vector current is not conserved and the correlation function in the axial channel has an additional longitudinal part @xmath30 which vanishes in the chiral limit @xmath31 . in the njl model the matrix element of rhs of eqs.([pcrdeff]-[acrdeff ] ) can be expressed as a path integral using the model generating functional ( [ genf ] ) . in the leading order in @xmath0 only two diagrams ( shown in fig .
[ figr1 ] ) contribute to the matrix elements .
the first one is a simple one - quark loop whereas the second includes a boson line . for the pseudoscalar function
we get the following result : @xmath32 where @xmath33 function @xmath34 corresponds to the quark - loop diagram with two pseudoscalar - isovector vertices ( @xmath35 ) . as already mentioned , the model is not renormalizable and in order to make the quark loop finite we need a regularization . in our calculations
we use the proper - time as well as the pauli - villars regularization scheme . both schemes
preserve the symmetries of the model .
here we present only the proper - time regularized expression for @xmath34 @xcite : @xmath36}\ , , \label{bpq}\ ] ] and for the condensate : @xmath37 details as well as regularized expressions for the quark condensate and the function @xmath34 in the case of pauli - villars regularization can be found in ref .
@xcite . in the chiral limit the pseudoscalar correlation function ( [ pq ] )
develops the expected goldstone pole at @xmath38 : @xmath39 where the pion decay constant @xmath40 is given by the residue of the pole : @xmath41 the non - zero current mass @xmath5 shifts the pion pole in eq.([pq0 ] ) to @xmath42 and one recovers the gell - mann - oakes - renner relation : @xmath43 in the vector channel the correlation function @xmath44 is given by @xmath45 where , similarly to @xmath34 , the function @xmath46 corresponds to the quark - loop diagram with two vector vertices ( @xmath47 ) ( lhs of fig .
[ figr1 ] ) .
in particular , in the proper - time scheme it has the following form : @xmath48}\ , .
\label{bvq}\ ] ] for the transverse invariant function of the axial current correlator in the njl model we have : @xmath49 whereas the longitudinal one includes only a pion contribution : @xmath50 the correlation functions ( [ pvq2]),([ptaq0 ] ) possess ( at some model parameter values ) poles in the timelike region @xmath51 which is in fact the pole of the propagator of the auxiliary ( axial ) vector field .
this is used in the so - called on - shell treatment @xcite where one fixes the vector coupling constant @xmath4 adjusting the position of the pole to the physical rho mass .
the other model parameters , namely the cutoff @xmath21 and the current mass @xmath5 are fixed reproducing the physical pion mass and the physical pion decay constant from eqs .
( [ fpi]),([gmor ] ) , whereas the constituent quark mass is treated as a free parameter as usual . to illustrate the on - shell scheme we present in fig.[figr2 ]
the corresponding parameter values ( index `` pole '' ) as a function of @xmath18 for the case of the the pauli - villars regularization .
since the results with the proper - time regularization are almost identical with those of the proper - time ones we do not discuss them separately . as can be seen the behavior of the coupling constants and
the cutoff as function of the constituent mass is not smooth the value @xmath52 is a singular point where all curves show a kink . at @xmath53
the vector current correlation function has a real pole at @xmath54 which is an indication for a bound state in the spectrum .
however , because of lack of confinement a second peak in the @xmath1 continuum appears , which is not taken into account in the definition of the physical @xmath55 field .
obviously , it is hard to relate the second peak to the excited @xmath56-resonance . at smaller values of @xmath57 in the vector channel
a broad resonance appears in the @xmath1 continuum because the corresponding pole moves to the complex plane of @xmath58 .
the position of the peak depends on the vector coupling constant @xmath4 . using this we determine @xmath4 fixing the position of the peak in the vector spectral function @xmath59 at the physical @xmath55 mass .
however , as can be seen from fig.[figr2 ] it is not always possible to find a solution for @xmath4 : for @xmath18 between 230 and 260 mev in fact there is no such a solution .
it should be also noted that at @xmath53 both @xmath4 and @xmath3 , as well as the cutoff @xmath21 show a stronger dependence on @xmath18 as @xmath4 is dominant ( @xmath60 . this changes for smaller values of @xmath18 where @xmath3 and @xmath21 stay almost constant whereas @xmath4 decreases . in the axial channel for the values of the constituent mass @xmath18 considered ( 200 - 450 mev ) the @xmath61 meson appears as a very broad peak being centered around 1 gev in the @xmath1 continuum . in fig.[figr2 ] we also show the parameter values ( index `` grad '' ) obtained by means of a gradient or heat - kernel expansion of the effective action @xcite .
this procedure is frequently used to fix the vector coupling constant @xmath4 in njl model and in fact it is an approximation to the on - shell description of vector modes .
as can be seen from fig.[figr2 ] this procedure provides a quite crude approximation to the on - shell values because of the large masses of the vector mesons . from the above discussion one
concludes that the on - shell treatment is not able to provide a consistent description of the vector meson modes in njl model . as a next step
we suggest to make use of the fact that at spacelike momenta @xmath62 the vector and axial - vector current correlation functions can be related to the experimentally known spectral density via dispersion relations . in the vector channel
the invariant function @xmath44 satisfies a dispersion relation with one subtraction : @xmath63 where the physical spectral density @xmath64 is experimentally accessible : @xmath65 for the rhs we use a simple ( @xmath55 + continuum ) parameterization @xcite of the experimental data @xmath66 with the parameter values : @xmath67 eqs.([disprel ] ) and ( [ expspv ] ) provide an independent way to fix the vector coupling constant @xmath4 confronting the model with the experiment by equating the phenomenological dispersion integral and the model correlation function at low positive @xmath58 : @xmath68 the resulting @xmath4 as well as @xmath21 and @xmath3 as a function of @xmath18 , obtained by adjusting ( @xmath69-fit ) the lhs of eq .
( [ phvq2 ] ) to the rhs with the physical spectral density ( [ expspv ] ) at low positive momenta ( we consider @xmath70 gev@xmath71 ) , are shown ( index `` corr '' ) in fig.[figr2 ] a ) and b ) .
apparently the correlation function at spacelike momenta and the on - shell treatment yield rather different results .
in contrast to the curves resulting from the on - shell scheme those of the correlation function have the advantage to behave smoothly as a function of @xmath18 .
this , and the very fact that the njl model is designed for low - energy structures involving small momenta , makes the above correlation method preferable . with increasing @xmath18 larger values for @xmath4 are needed to fit the phenomenological side whereas the @xmath3 is slightly decreasing . in the axial channel ,
since the longitudinal part of the axial correlation function is almost independent on the @xmath4 and @xmath18 , we will concentrate our further discussion on its transverse part @xmath72 .
it obeys a dispersion relation with two subtractions : @xmath73 similar to the vector channel for @xmath74 we take a simplified ( @xmath61 + continuum ) parameterization @xcite @xmath75 with parameters @xcite : @xmath76 confronting the model @xmath61-correlation function ( [ ptaq0 ] ) and the phenomenological expressions ( [ pat ] ) , ( [ expspa ] ) one has another way to fix the vector coupling constant @xmath4 .
the obtained vector coupling constant ( @xmath77 ) as a function of @xmath18 is shown in fig.[figr2 ] a ) in comparison with the results from fitting the @xmath55 phenomenology as well as from the on - shell @xmath55 treatment and the derivative expansion .
in contrast to the vector channel the model axial function @xmath78 depends much stronger on the constituent mass and because of that the model is able to fit simultaneously both @xmath55- and @xmath61-channels only within a narrow window of values for the constituent quark mass @xmath18 around 240 mev .
it is interesting to note that in this window the on - shell values are also very close to those obtained from the correlation functions .
the particular value of @xmath18 depends on the parameter values ( [ vparam]),([aparam ] ) used for the ( resonance + continuum ) parameterizations .
in fact the estimate of the constituent mass @xmath79 mev in the random instanton model @xcite of qcd vacuum , applied successfully also to the meson correlation functions @xcite , is not far from our numbers .
the resulting vector coupling constant @xmath4 is of the same order as the scalar one @xmath3 . in this point
we disagree with ref .
@xcite in which constraining the njl model via qcd sum rules it is concluded that @xmath4 should be almost zero or at least @xmath80 .
it is due to the fact that in contrast to us in ref .
@xcite the njl model is treated in a quite simplified approximate way ignoring the full @xmath58-dependence of the correlation functions as well as the relations between the model parameters fixed to reproduce the physical pion properties . in our calculations
so far we used simple parameterizations ( [ expspv]),([expspa ] ) for the experimental spectral densities in which the resonance widths are neglected . in order to check this approximation we repeated the calculations using a much more elaborated ( finite - width resonance+continuum ) parameterization @xcite of the experimental data .
the obtained results follow qualitatively the picture of fig .
[ figr2 ] .
the only difference is that the narrow window , where in both vector and axial channels the experimental low - energy behavior of the correlation functions can be reproduced simultaneously , is shifted to higher values of the constituent mass @xmath81 mev .
we also related the model with @xmath4 , fixed in the present scheme , to the chiral effective lagrangean of gasser and leutwyler @xcite .
following ref .
@xcite we calculated the corresponding low - energy coefficients @xmath82 in the notation of ref .
@xcite and low - energy pionic characteristics .
as can be seen from table [ tabl1 ] our theoretical predictions are in good agreement with the empirical values .
to summarize , in this letter we study the vector meson modes in the njl model in terms of current correlation functions .
we find that the on - shell treatment of the vector meson fails to reproduce the low - energy behavior of the correlation functions fixed by the experimentally known spectral density via dispersion relations .
we offer a different scheme to fix the njl - model parameters based on the behavior of the correlation functions at small spacelike momenta . treating the vector coupling constant as a free parameter
the model is able to reproduce the phenomenology in both the axial- and vector channels for the constituent mass only in a narrow window around 240 mev .
the vector coupling constant is by no means zero and is of the same order as the scalar coupling constant . in principle , these two schemes , the on - shell treatment ( large timelike momenta ) and the one based on the behavior of the correlation function at small spacelike momenta , should provide identical results within an `` exact '' theory , which however is not the case for an effective model like njl .
in the latter case the correlation function method appears to be preferable .
one also should keep in mind that the present considerations are done in leading order in @xmath0 .
the inclusion of @xmath83 quantum boson ( loop ) corrections could in principle change the present results . |
_ flavourdynamics _ , the study off the electroweak lagrangian and its implications , is one of the central areas of research in particle physics . for example , in the united kingdom the _ mission statement _ of the particle physics community , as defined by our research council , is to obtain insights into the following three fundamental questions : * the origin of mass , * the 3 generations of elementary particles and their weak asymmetries , * and * the nature of dark matter .
the second of these topics is flavourdynamics , which is the subject of this lecture course .
i will review the theoretical and experimental work being done in an attempt to determine the parameters of the standard model ( sm ) accurately , to test subtle properties of the sm and to define and search for signatures of _ new physics_. although this primarily involves weak interactions , the major theoretical difficulty in interpreting experimental data on weak hadronic processes is controlling the non - perturbative strong interaction effects necessarily present in these decays .
for this reason any review of flavour physics must contain a discussion also of these qcd effects , and much of the presentation below concerns this subject . in the next few years much of the effort of the experimental community will be concerned with trying to gain an understanding of @xmath0-violation . indeed , one of the sakharov conditions for generating a matter - antimatter asymmetry in the universe in the big - bang cosmology is the requirement for the existence of @xmath0-violation .
the existence of three generations of quarks and leptons , and hence of a complex phase in the mixing matrix , implies the presence of @xmath0-violation at some level .
very little is known , however , about the value of this phase and hence of the magnitude of @xmath0-violation induced by this mechanism , and also about other possible sources .
it is hoped that the intensive studies , which are about to begin will significantly further our understanding and this will be one of the main topics of discussion towards the end of this lecture course .
the program for the four lectures is as follows :
1 . the first lecture will be an introduction to weak decays in the standard model . among the topics which will be covered
will be charged and neutral currents , the cabibbo - kobayashi - maskawa matrix and unitarity triangles , parity and charge conjugation symmetries , effective hamiltonians and operator product expansions and the heavy quark effective theory .
2 . in the second lecture
i will review the determination of the @xmath1 and @xmath2 matrix elements from semileptonic decays of @xmath3-mesons .
lattice computations , which provide the opportunity to evaluate non - perturbative strong interaction effects in weak decays in general , are introduced in this lecture .
the third lecture will contain a review of @xmath4-@xmath5 and @xmath6-@xmath7 mixing .
finally , in the last lecture i review @xmath0-violation in @xmath3-decays and the subject of inclusive @xmath3-decays .
the theoretical framework introduced in these lectures will be applied to several important physical processes .
it will not be possible , however , to discuss the full range of interesting processes which are providing , or will provide , fundamental physical information . a much larger set of physical quantities is considered in detail in the beautiful review by buras and fleischer @xcite , to which i refer the student , and from which i will quote extensively .
i also refer the reader to ref .
@xcite for a broad introduction to the _ dynamics of the standard model _ and to references @xcite and @xcite for modern reviews of _
@xmath3-decays _ and _ @xmath0-violation _ respectively .
i will assume a familiarity with elementary quantum field theory ( as discussed for example by j. petersen and v. zakharov at this school @xcite ) , but will try to avoid technical complexities , focussing instead on the underlying ideas .
( 360,90)(0,15 ) ( 0,100)(30,100)(30,100)(60,100 ) ( 15,107)[b]@xmath8(45,105)[b]@xmath9 ( 100,100)(130,100)(130,100)(160,100 ) ( 115,107)[b]@xmath8(145,107)[b]@xmath8 ( 200,100)(230,100)(230,100)(260,100 ) ( 215,107)[b]@xmath8(245,107)[b]@xmath8 ( 300,100)(330,100)(330,100)(360,100 ) ( 315,107)[b]@xmath8(345,107)[b]@xmath8 ( 30,100)(30,60)2.55(32,55)[t]@xmath10 ( 30,30)[t](a ) ( 130,100)(130,60)2.55(132,55)[t]@xmath11 ( 130,30)[t](b ) ( 230,100)(230,60)2.55(230,53)[t]@xmath12 ( 230,30)[t](c ) ( 330,100)(330,60)27(330,55)[t]@xmath13 ( 330,30)[t](d ) the interactions between the quarks and gauge bosons in the standard model are illustrated in fig .
[ fig : vertices ] . in these lectures we will be particularly interested in the weak interactions .
the vertex for the charged current interaction in which quark flavour @xmath8 changes to @xmath9 is depicted in fig [ fig : vertices](a ) and has the feynman rule @xmath14 where @xmath15 is the coupling constant of the @xmath16 gauge group and @xmath17 is the @xmath18 element of the cabibbo - kobayashi - maskawa ( ckm ) matrix ( @xmath19 ) @xcite .
( [ eq : ccvertex ] ) illustrates the @xmath20 ( vector@xmath21axial - vector ) structure of the charged - current interactions . at low energies , so that the momentum in the @xmath22-boson is much smaller than its mass @xmath23 , the four - fermion interaction mediated by the @xmath22-boson can be approximated by the local fermi @xmath24-decay interaction with coupling @xmath25 , where @xmath26 ( see fig .
[ fig : fermi ] ) .
( 170,40)(0,60 ) ( 10,100)(70,100)(10,60)(70,60 ) ( 40,100)(40,60)2.55 ( 30,80)[r]@xmath22(85,80)(100,80)(100,80)(102,80 ) ( 147,80)(132,93)(147,80)(162,93 ) ( 147,80)(132,67)(147,80)(162,67 )
the charged - current interactions are of the form @xmath27 where ( assuming the standard model with 3 generations ) the ckm - matrix , @xmath28 is a unitary 3@xmath29 3 one , relating the weak and mass eigenstates .
the 1996 particle data book @xcite gives the following values for the magnitudes of each of the elements : @xmath30 the subscript @xmath31 in eq .
( [ eq : cc ] ) represents _ left handed _ , ( @xmath32 ) .
if we have @xmath33 quark flavours , ( @xmath34 is the number of generations ) then @xmath28 is a @xmath35 unitary matrix .
it therefore has @xmath36 real parameters , however @xmath37 of these can be absorbed into unphysical phases of the quark fields rather than @xmath38 , is that if the fields are multiplied by the same phase factor then @xmath39 is unchanged .
thus there is one fewer phase , which can be absorbed . ] , leaving @xmath40 physical parameters to be determined . in the two flavour case
there is just one parameters , which is conventionally chosen to be the cabibbo angle : @xmath41 with three flavours there are 4 real parameters .
three of these can be interpreted as angles of rotation in three dimensions ( e.g. the three euler angles ) and the fourth is a phase .
the general parametrisation recommended by the particle data group @xcite is @xmath42 where @xmath43 and @xmath44 represent the cosines and sines respectively of the three angles @xmath45 , @xmath46 .
@xmath47 is the phase parameter .
the parametrisation in eq .
( [ eq : ckmgenpar ] ) is general , but awkward to use . for most practical purposes it is sufficient to exploit the empirical fact that the elements get smaller as one moves away from the diagonal of the matrix ( see eq .
( [ eq : ckmvalues ] ) ) , and to use a simpler , but approximate parametrisation .
it has become conventional to use the wolfenstein parametrisation @xcite : @xmath48 @xmath49 is approximately the cabibbo angle , and the terms which are neglected in the wolfenstein parametrisation are of @xmath50 .
much of the current effort of the particle physics community is devoted to determining the four parameters @xmath51 and @xmath52 , with ever increasing precision , and this will continue for some time to come .
for this reason i will spend a significant fraction of these lectures on the theoretical issues related to the determination of these parameters . from the unitarity of the ckm - matrix we have a set of relations between the entries .
a particularly useful one is : @xmath53 in terms of the wolfenstein parameters , the components on the left - hand side of eq .
( [ eq : ut ] ) are given by : @xmath54 + o(\lambda^7)\nonumber\\ v_{cd}v_{cb}^ * & = & -a\lambda^3 + o(\lambda^7)\label{eq : utelements}\\ v_{td}v_{tb}^ * & = & a\lambda^3 [ 1 - ( \bar\rho + i \bar\eta ) ] + o(\lambda^7)\nonumber\ , \end{aligned}\ ] ] where @xmath55 and @xmath56 . the unitarity relation in eq .
( [ eq : ut ] ) can be represented schematically by the famous
unitarity triangle " of fig .
[ fig : ut ] ( obtained after scaling out a factor of @xmath57 ) .
( 250,125)(0,-20 ) ( 0,0)(200,0)(0,0)(140,80)(140,80)(200,0 ) ( 150,90)[b]@xmath58(138,72)[t]@xmath59 ( 10,-10)[t]@xmath60(22,2)[bl]@xmath12 ( 210,-10)[t]@xmath61(188,2)[br]@xmath24 ( 23,43)[l]@xmath62 ( 242,43)[r]@xmath63 ) the notation in fig .
[ fig : ut ] is standard , and in the following lectures i will describe current and future attempts to determine the position of the vertex @xmath64 , and hence the values of the angles @xmath65 and @xmath12 . by using many different processes to overdetermine the position of @xmath64 ,
one will be able to test the consistency of the sm , and check for evidence of physics beyond this model .
although much of the remainder of these lectures will be devoted to the determination of the entries in the ckm - matrix , since the cabibbo sector is the best determined , i will not consider it beyond this subsection .
i briefly review the status of the four entries in the @xmath66 top - left submatrix of @xmath28 @xcite .
( 320,120)(0,20 ) ( 10,100)(50,100)(50,100)(90,110 ) ( 50,100)(90,80)26 ( 30,105)[b]@xmath67(70,110)[b]@xmath68 ( 95,80)[l]@xmath69(50,70)[t](a ) ( 250,100)(20,30)(0 ) ( 220,100)(200,100)(280,100)(300,100 ) ( 220,100)50.5(280,100)50.5 ( 250,120)(270,130)24 ( 270,130)(278,138)(270,130)(278,122 ) ( 280,130)[l]leptons ( 195,100)[r]@xmath70(305,100)[l]@xmath71 ( 235,108)[b]@xmath72(265,108)[b]@xmath68 ( 250,70)[t](b ) ( 320,90)(0,50 ) ( 20,80)(50,80 ) ( 15,80)[r]@xmath73 ( 50,80)(80,110)(60,100)[r]@xmath67 ( 80,110)(110,110)(95,115)[b]@xmath74 ( 52,77)(92,77)(52,83)(92,83 ) ( 80,110)(50,130)25 ( 50,80)50.5(70,89)10(70,94)10 ( 50,65)[t](c ) ( 250,100)(20,30)(0 ) ( 220,100)(200,100)(280,100)(300,100 ) ( 220,100)50.5(280,100)50.5 ( 250,120)(270,130)24 ( 270,130)(278,138)(270,130)(278,122 ) ( 280,130)[l]leptons ( 195,100)[r]@xmath75(305,100)[l]@xmath70 ( 235,108)[b]@xmath74(265,108)[b]@xmath72 ( 250,65)[t](d ) *
@xmath76 : @xmath77 this is the best determined of the elements and is obtained by studying superallowed @xmath24-decays in nuclei ( see fig .
[ fig : vcabibbo](a ) ) . since the publication of the 1996 particle data book , a result with a smaller error has been presented ,
@xmath78 @xcite .
* @xmath79 : @xmath80 this matrix element is obtained from semileptonic decays of @xmath70-mesons ( @xmath81 , @xmath82 , see fig . [
fig : vcabibbo](b ) ) or of hyperons . in the @xmath70-meson decays , parity symmetry implies that only the vector component of the @xmath20 current contributes to the decays .
since the momentum transfer to the @xmath71-meson is small , and the vector current for degenerate quarks is conserved , the uncertainties due to strong interaction effects are small and can be estimated using chiral perturbation theory . *
@xmath83 : @xmath84 the @xmath85 element is obtained from charm production in deep inelastic neutrino ( or antineutrino ) nucleon scattering , see fig .
[ fig : vcabibbo](c ) . *
@xmath86 : @xmath87 this matrix element can be determined from semileptonic decays of charmed mesons ( see fig . [
fig : vcabibbo](d ) ) . in this case
the large difference in the masses of the quarks make it difficult to estimate the strong interaction effects accurately ( see sec .
[ subsec : vcb ] ) , and this is the reason for the relatively large error . the errors in eqs .
( [ eq : vud])-([eq : vcs ] ) are those in the measurements of these matrix elements themselves .
comparing these uncertainties with those in eq .
( [ eq : ckmvalues ] ) , where the constraints of unitarity have been imposed ( assuming just three generations ) , we see how much these constraints reduce the errors . in the standard model
, unitarity implies that there are no flavour changing neutral current ( fcnc ) reactions at tree level , i.e. that there are no vertices of the type : ( 210,40)(0,-10 ) ( 0,20)(30,20)(30,20)(60,20 ) ( 150,20)(180,20)(180,20)(210,20 ) ( 30,20)(30,-10)2.54 ( 180,20)(180,-10)2.54 ( 15,25)[b]@xmath88(45,25)[b]@xmath72 ( 165,25)[b]@xmath68(195,25)[b]@xmath74(230,-10)[c ] .
quantum loops can , however , generate fcnc reactions , through _ box _ diagrams ( see fig . [
fig : box ] ) or _ penguin _ diagrams ( see fig .
[ fig : penguin ] ) , and we will discuss some of the physical processes induced by these loop - effects in the following lectures .
( 250,70)(0,45 ) ( 10,100)(40,100)(40,100)(80,100)28 ( 80,100)(110,100 ) ( 40,60)(10,60)(40,60)(80,60)28 ( 110,60)(80,60 ) ( 40,100)(40,60)(80,60)(80,100 ) ( 25,105)[b]@xmath67(95,105)[b]@xmath88 ( 25,55)[t]@xmath88(95,55)[t]@xmath67 ( 35,80)[r]@xmath89(85,80)[l]@xmath89 ( 160,100)(190,100)(190,100)(230,100 ) ( 230,100)(260,100 ) ( 190,60)(160,60)(230,60)(190,60 ) ( 260,60)(230,60 ) ( 190,100)(190,60)28(230,60)(230,100)28 ( 175,105)[b]@xmath67(245,105)[b]@xmath88 ( 175,55)[t]@xmath88(245,55)[t]@xmath67 ( 210,105)[b]@xmath89(210,55)[t]@xmath89 ( 430,130)(10,15 ) ( 10,100)(40,100)(40,100)(70,100)(70,100)(100,100 ) ( 100,100)(130,100 ) ( 70,100)(30,180,0)320(70,100)(70,60)28 ( 70,105)[b]@xmath89(20,95)[t]@xmath67(120,95)[t]@xmath72 ( 70,140)[b]@xmath90(70,55)[t]@xmath91 ( 70,30)[t](a ) ( 160,120)(190,120)(190,120)(250,120 ) ( 250,120)(280,120 ) ( 190,120)(220,100)27(250,120)(220,100)27 ( 220,100)(220,60)28 ( 220,125)[b]@xmath89(170,115)[t]@xmath67(270,115)[t]@xmath72 ( 202,105)[r]@xmath90(238,105)[l]@xmath90 ( 220,55)[t]@xmath91(220,30)[t](b ) ( 310,100)(340,100)(340,100)(370,100)(370,100)(400,100 ) ( 400,100)(430,100 ) ( 370,100)(30,180,0)320(370,100)(370,60)28 ( 370,105)[b]@xmath89(320,95)[t]@xmath67(420,95)[t]@xmath72 ( 370,140)[b]@xmath90(370,55)[t]@xmath92 ( 370,30)[t](c ) as a consequence of the `` gim''-mechanism @xcite all the fcnc vertices vanish in the limit of degenerate quark masses .
this is a consequence of the unitarity of the ckm - matrix .
for example , consider the vertex in the diagram of fig .
[ fig : penguin](a ) , in the hypothetical situation in which there is a `` horizontal '' symmetry such that @xmath93 . in this case
the contribution from each of the positively charged quarks in the loop is simply proportional to the corresponding ckm - matrix elements , so that the total contribution is @xmath94 which vanishes by unitarity of the ckm - matrix .
a similar argument holds for all the fcnc vertices .
of course , in reality the masses of the quarks are not equal and the fcnc vertices are not zero . depending on the process being studied , however , the gim - mechanism may lead to a suppression of the amplitude .
symmetries play a fundamental rle in physics in general and in particle physics in particular . in this subsection
i briefly introduce the violation of @xmath0-symmetry , which will be a focus of many of the studies at @xmath3-physics facilities in the next few years .
[ [ parity ] ] * parity : * + + + + + + + + + the parity transformation reverses the sign of spatial coordinates , @xmath97 .
the vector and axial - vector fields transform as:@xmath98 and the vector and axial - vector currents transform similarly .
left - handed components of fermions @xmath99 transform into right - handed ones @xmath100 , and vice - versa .
since weak interactions only involve the left - handed components , parity is not a good symmetry of the weak force , in distinction to the strong and electromagnetic forces ( qcd and qed are invariant under parity transformations ) . [
[ charge - conjugation ] ] * charge conjugation : * + + + + + + + + + + + + + + + + + + + + + for free fields ( @xmath101 say ) we can perform a fourier decomposition , with coefficients which contain annihilation and creation operators : @xmath102 where @xmath103 and @xmath104 represent the annihilation and creation operators for the particle and antiparticle .
the charge conjugation transformation is the interchange of @xmath103 and @xmath88 . in an interacting field theory
this is not in general the same as interchanging physical particles and antiparticles for which we need the @xmath105 combined transformation ( @xmath106 is time reversal transformation )
. nevertheless @xmath96-tranformations are also interesting in field theories . under @xmath96
the currents transform as follows : @xmath107 where @xmath108 represents a spinor field of type ( flavour or lepton species ) @xmath8 .
[ [ mathbfcp ] ] @xmath109 + + + + + + + + + + + + + + + + + + + + + + + + + + under the combined @xmath0-transformation , the currents transforms as : @xmath110 where the fields on the left ( right ) hand side are evaluated at @xmath111 ( @xmath112 ) .
consider now a charged current interaction : @xmath113 where @xmath22 represents the field of the @xmath22-boson and @xmath114 and @xmath115 are up and down type quarks of flavours @xmath8 and @xmath9 respectively . under a @xmath0 transformation , the interaction term in eq .
( [ eq : ccint ] ) transforms to : @xmath116 where the parity transformation on the coordinates is implied . comparing eqs .
( [ eq : ccint ] ) and ( [ eq : cpcc ] ) , we see that @xmath0-conservation requires the ckm - matrix @xmath17 to be real ( or more strictly that any phases must be absorbable into the definition of the quark fields ) .
if the only source of @xmath0-violation in nature is the phase in the ckm - matrix , then at least three generations of quarks are required ( see subsec .
[ subsec : ckm ] ) .
a central element of the research of the forthcoming generation of experiments in @xmath3-physics will be to determine whether the phase in the ckm - matrix is the only ( or the main ) source of @xmath0-violation .
@xmath0-violation appears to be small compared to the strength of the weak interaction , so that @xmath0 is a fairly good symmetry .
we recall that the presence of cp - violation is one of the sakharov conditions for the creation of the baryon - antibaryon asymmetry of the universe .
as already mentioned , the main difficulty in making predictions for weak decays of hadrons is in controlling the non - perturbative strong - interaction effects
. we will encounter this problem several times in subsequent lectures , but , as an introduction , i now discuss it briefly in a particularly simple situation , that of the leptonic decays of pseudoscalar mesons in general , and of the @xmath3-meson in particular .
( 180,50)(0,15 ) ( 10,41)(50,41 ) ( 80,41)(15,30)(0 ) ( 79,56)(81,56)(81,26)(79,26 ) ( 50,41)70.8 ( 110,41)(150,41)37 ( 150,41)(167,51)(167,31)(150,41 ) ( 20,35)[tl]@xmath117 ( 80,62)[b]@xmath88(80,20)[t]@xmath118 ( 172,53)[l]@xmath119(172,30)[l]@xmath120 ( 132,48)[b]@xmath22 the leptonic decay of a pseudoscalar meson ( e.g. the @xmath3-meson ) is illustrated in the diagram of fig .
[ fig : fb ] .
all the qcd effects corresponding to the strong interactions between the quarks in fig .
[ fig : fb ] are contained in the matrix element of the @xmath20 hadronic current : @xmath121 lorentz and parity symmetries imply that the matrix element of the vector component of the @xmath20 current vanishes ( @xmath122 ) .
this component is an axial - vector , and there is no axial - vector which can be constructed from the single vector we have at our disposal ( @xmath73 ) and the invariant tensors ( @xmath123 and @xmath124 ) .
these symmetries also imply that the matrix element of the axial component of the @xmath20 current is a lorentz - vector , and hence that it is proportional to @xmath125 : @xmath126 the constant of proportionality , @xmath127 , is called the _ decay constant _ of the @xmath3-meson .
thus all the qcd - effects for leptonic decays of @xmath3-mesons are contained in a single number , @xmath127 .
an identical discussion holds for other pseudoscalar mesons ( @xmath128 ) .
the normalisation used throughout these lectures , corresponds to @xmath129 mev ( @xmath127 is unknown at present ) . for leptonic decays of vector mesons , space - time symmetries also imply that the strong interaction effects are contained in a single decay constant .
thus , the rle of quantitative non - perturbative methods , such as lattice qcd ( discussed in sec .
[ subsec : lattice ] below ) or qcd sum - rules , is to provide the framework for the evaluation of matrix elements such as that in eq .
( [ eq : bldecay ] ) .
even though this is a course in flavour physics , we can not avoid the fact that the quarks are interacting strongly and hence that qcd effects must be considered .
indeed , as already mentioned above , our rather primitive control of non - perturbative qcd is the principal source of uncertainty in interpeting experimental data of weak decays and in determining the fundamental parameters of the standard model . in paolo nason s lecture course @xcite
, we have seen how the property of _ asymptotic freedom _ , which states that the interactions of quarks and gluons become weaker at short distances , enables us to use perturbation theory to make predictions for a wide variety of short - distance ( or light - cone ) dominated processes . for separations
@xmath130 ( @xmath131 fm say ) or corresponding momenta @xmath132 ( @xmath133 gev say ) , we can use perturbation theory .
the natural scale of strong interaction physics is of @xmath134 , however , and so in general , and for most of the processes discussed in this course , non - perturbative techniques must be used .
( 350,100)(0,20 ) ( 30,100)(90,100)(30,60)(90,60 ) ( 60,100)(60,60)2.55 ( 25,100)[r]@xmath72(25,60)[r]@xmath135 ( 95,100)[l]@xmath68(95,60)[l]@xmath118 ( 70,80)[l]@xmath22(60,40)[t](a ) ( 240,100)(340,100)(240,60)(340,60 ) ( 310,100)(310,60)2.55 ( 270,100)(270,60)1.58 ( 235,100)[r]@xmath72(235,60)[r]@xmath135 ( 345,100)[l]@xmath68(345,60)[l]@xmath118 ( 320,80)[l]@xmath22(290,40)[t](b ) ( 262,80)[r]@xmath136 as an example , consider a decay of a @xmath70-meson into two pions , for which a tree - level diagram of the quark subprocess is depicted in fig .
[ fig : kdecay](a ) .
the amplitude is proportional to @xmath137 i.e. it is proportional to the matrix element of the operator @xmath138 between @xmath139 and @xmath140 states .
a one - loop ( i.e. one - gluon exchange ) correction to this diagram is shown in fig .
[ fig : kdecay](b ) .
this clearly generates a second operator , @xmath141 s are the eight matrices , representing the generators of the @xmath142 colour group in the fundamental representation .
the coupling of quarks and gluons is proportional to the elements of these matrices . ] ,
which , using fierz identities can be written as a linear combination of @xmath143 and @xmath144 where @xmath145 this discuusion can be generalized to higher orders , as i now explain .
using the operator product expansion ( ope ) one writes the amplitude for a weak decay process from an initial state @xmath146 to a final state @xmath147 in the form : @xmath148 @xmath149 is the renormalization scale at which the composite operators @xmath150 are defined and @xmath151 is the appropriate product of ckm - matrix elements .
the expansion ( [ eq : ope ] ) is very convenient .
the non - perturbative qcd effects are contained in the matrix elements of the operators @xmath150 , which are independent of the large momentum scale , in this case of @xmath23 .
the wilson coefficient functions @xmath152 are independent of the states @xmath146 and @xmath147 and can be calculated in perturbation theory . since physical amplitudes manifestly
do not depend on @xmath149 , the @xmath149-dependence in the operators @xmath153 is cancelled by that in the coefficient functions @xmath152 .
the _ effective hamiltonian _ for weak decays can then be written in the form : @xmath154 in order to gain a little further intuition into the origin of eq.([eq : ope ] ) consider the one - loop graph of fig .
[ fig : kdecay](b ) .
counting at large loop momenta @xmath155 readily shows that the graph is ultra - violet convergent ( in the feynman gauge , say ) : @xmath156 where the first two factors in the integrand ( @xmath157 ) correspond to the two quark propagators , the third ( @xmath158 ) to the gluon propagator and the final factor to the propagator of the @xmath22-boson .
it can be deduced from eq.([eq : powercounting ] ) that the contribution from this graph contains a term proportional to @xmath159 , where p is some low - momentum ( infra - red ) scale .
( 150,60)(0,50 ) ( 20,100)(90,80)(90,80)(130,100 ) ( 20,60)(90,80)(90,80)(130,60 ) ( 90,80)100.7 ( 45,92.86)(45,67.14)1.57 ( 15,100)[r]@xmath72(15,60)[r]@xmath135 ( 135,100)[l]@xmath68(135,60)[l]@xmath118 ( 38,80)[r]@xmath92 ( 92,65)[t]@xmath143 consider now the one - loop graph of fig .
[ fig : operator ] , depicting a correction to a matrix element of the operator @xmath143 defined in eq .
( [ eq : o1def ] ) .
now the @xmath22-propagator is absent , and the power counting in the large momentum region gives : @xmath160 i.e. a logarithmically divergent contribution .
after renormalization at a scale @xmath149 there will be a contribution to this graph proportional to @xmath161 . in the low momentum ( non - perturbative ) region ( i.e. for @xmath162 ) the contributions from the graphs in figs .
[ fig : kdecay](b ) and [ fig : operator ] are the same .
thus we can split the factor of @xmath163 from the graph of fig .
[ fig : kdecay](b ) into two parts ; the first , @xmath161 is contained in the matrix element of @xmath143 , and the second , @xmath164 , into the coeficient function @xmath165 . since the infra - red behaviour of the graphs in figs .
[ fig : kdecay](b ) and [ fig : operator ] are the same , the coefficient functions are independent of the treatment of the low - momentum region . moreover , in order to calculate the coefficient functions , we can choose any convenient external states , and in practice one chooses quark or gluon states .
we refer to the evaluation of the coefficient functions as the process of _
matching_. in @xmath70-decays
it is natural , although not necessary , to choose @xmath149 to be as small as one can without invalidating perturbation theory . in this way
one might hope to avoid _ large logarithms _ ( @xmath166 ) ) which would spoil any insights we might have about the operator matrix elements from non - relativistic quark models or other approaches .
of course this implies the presence of large logarithms of the type @xmath167 in the coefficient functions , but these can be summed using the renormalization group , leading to factors of the type : @xmath168^{\gamma_0/2\beta_0}\ , \label{eq : leadinglogs}\ ] ] where @xmath169 is the one - loop contribution to the _ anomalous dimension _ of the operator ( proportional to the coefficient of @xmath170 in the evaluation of the graph of fig .
[ fig : kdecay](b ) ) and @xmath171 is the first term in the @xmath24-function , ( @xmath172 ) ) , the expression in eq .
( [ eq : leadinglogs ] ) is generalized into a matrix equation , representing the mixing of the operators ] .
( [ eq : leadinglogs ] ) represents the sum of the _ leading logarithms _ , i.e. the sum of the terms @xmath167 . for almost all the important processes ,
the first ( or even higher ) corrections have also been evaluated .
we end this subsection by making some further points about the use of the ope in weak decays and other hard processes : * we shall see below that for some important physical quantities ( e.g. @xmath173 ) , there may be as many as ten operators , whose matrix elements have to be estimated . * one may try to evaluate the matrix elements using some non - perturbative computational technique , such as lattice simulations or qcd sum - rules ( see below ) , or to determine them from experimental data .
in the latter case , if there are more measurements possible than unknown parameters ( i.e. matrix elements ) , then one is able to make predictions . * in weak decays the large scale , @xmath23 , is of course fixed . for other processes , most notably for deep - inelastic lepton - hadron scattering ,
the ope is useful in computing the behaviour of the amplitudes with the large scale ( e.g. with the momentum transfer ) .
most of the important physical properties of the hydrogen atom are independent of the mass of the nucleus .
analogous features hold in the physics of heavy hadrons , where by _ heavy _ i mean that @xmath174 and @xmath175 is the mass of the heavy quark . during the last few years the construction and use of the _ heavy quark effective theory _ ( hqet ) , has proved invaluable in the study of heavy quark physics . in this subsection
i briefly introduce the main features of the hqet , which will be used in the following lectures .
consider the propagator of a ( free ) heavy quark : ( 0,10)(100,50 ) ( 10,50)(70,50)(40,45)[t]@xmath73 ( 80,50)[l]@xmath176 . if the momentum of the quark @xmath73 is not far from its mass shell , @xmath177 where @xmath178 and @xmath179 is the ( relativistic ) four velocity of the hadron containing the heavy quark ( @xmath180 ) , then ( 0,10)(100,50 ) ( 10,50)(70,50)(40,45)[t]@xmath73 ( 80,50)[l]@xmath181 .
@xmath182 is a projection operator , projecting out the _ large _ components of the spinors .
this propagator can be obtained from the gauge - invariant action @xmath183 where @xmath184 is the spinor field of the heavy quark ) . ] .
@xmath185 is independent of @xmath175 , which implies the existence of symmetries relating physical quantities corresponding to different heavy quarks ( in practice the @xmath88 and @xmath74 quarks ) .
the light degrees of freedom are also not sensitive to the spin of the heavy quark , which leads to a spin - symmetry relating physical properties of heavy hadrons of different spins .
consider , for example , the correlation function : @xmath186 where @xmath187 ( @xmath188 ) is an interpolating operator which can create a heavy hadron @xmath189 , which we take to be a pseudoscalar or vector meson .
the hadron is produced at rest , with four velocity @xmath190 .
we can use the interpolating operator @xmath191 for the pseudoscalar meson and @xmath192 ( @xmath193 ) for the vector meson .
this means that the correlation function will be identical in the two cases except for the factor @xmath194 when @xmath189 is a pseudoscalar meson , and @xmath195 when it is a vector meson .
since the correlation functions behave with time as @xmath196 , this implies that the pseudoscalar and vector mesons are degenerate , up to relative corrections of @xmath197 : @xmath198 where @xmath199 and @xmath200 are the masses of the pseudoscalar and vector mesons respectively , or equivalently @xmath201 this relation is reasonably well satisfied experimentally : @xmath202 which is encouraging .
however , i also mention in passing that the light mesons , to which the hqet certainly does not apply , also satisfy these relations numerically : @xmath203 the usefulness of the hqet does not lie in the fact that the _ residual _ momenta @xmath155 are always much smaller than @xmath204 ; indeed this is not true , in qcd one does have hard gluons with arbitrarily large momenta .
however the effects of hard gluons can be evaluated in perturbation theory .
the non - perturbative strong interaction effects are the same in qcd and in the hqet ( up to corrections of @xmath205 , which for the moment we neglect ) , and so the hqet relations , such as that between the decay constants of pseudoscalar and vector mesons , which can be deduced from the above proportionality of correlation functions , are violated only by perturbatively calculable corrections .
the principal application of the hqet is in @xmath88-physics , and to a lesser extent to charm physics . in both cases
the corrections of @xmath205 are significant , and hence one would like to compute these corrections . in practice , this gets progressively more difficult , and it is a much debated point as to whether even the first corrections ( i.e. those of @xmath205 relative to the leading terms ) have been reliably calculated for any interesting quantity .
in this lecture i will discuss exclusive semileptonic decays of @xmath3-mesons , in which the @xmath88-quark decays into a @xmath74- or @xmath68-quark , and from which one can determine the @xmath1 and @xmath2 elements of the ckm - matrix .
the decays are represented in fig .
[ fig : sl ] .
it is convenient to use space - time symmetries to express the matrix elements in terms of invariant form factors ( using the helicity basis for these as defined below ) .
when the final state is a pseudoscalar meson @xmath95 , parity implies that only the vector component of the @xmath20 weak current contributes to the decay , and there are two independent form factors , @xmath206 and @xmath207 , defined by@xmath208 + f^0(q^2)\,\frac{m_b^2 - m_p^2}{q^2}\,q^\mu\ , \label{eq : ffpdef}\ ] ] where @xmath209 is the momentum transfer , @xmath210 .
when the final - state hadron is a vector meson @xmath211 , there are four independent form factors , @xmath211 , @xmath212 , @xmath213 and @xmath64 : @xmath214 where @xmath215 is the polarization vector of the final - state meson , and @xmath216 . below
we shall also discuss the form factor @xmath217 , which is given in terms of those defined above by @xmath218 , with @xmath219 ( 180,70)(20,10 ) ( 10,41)(43,41 ) ( 15,35)[tl]@xmath3 ( 100,41)(20,50)(0 ) ( 157,41)(190,41 ) ( 180,35)[tl]@xmath220 ( 100,61)3 ( 50,41)70.8(150,41)70.8 ( 75,48)[b]@xmath88(117,48)[b]@xmath221 ( 100,16)[t]@xmath222 ( 100,71)[b]@xmath20 ( 101,21)(99,21 ) ( 70,57)(72,58)(128,57.5)(130,56.5 ) semileptonic @xmath224 and , more recently , @xmath225 decays are used to determine the @xmath1 element of the ckm matrix .
heavy quark symmetry is rather powerful in controlling the theoretical description of these heavy - to - heavy quark transitions ( see e.g. the review article by neubert for details and references @xcite ) . in the heavy quark limit , all six form factors in eqs .
( [ eq : ffpdef])-([eq : ffvadef ] ) are related and there is just one universal form factor @xmath226 , known as the isgur wise ( iw ) function , which contains all the non - perturbative qcd effects . specifically : @xmath227^{-1 } a_1(q^2 ) = \frac{m_b+m_d } { 2\sqrt{m_bm_d}}\,\xi(\omega)\ , \label{eq : iw}\ ] ] where @xmath228 is the velocity transfer variable ( @xmath229 are the four velocities of the corresponding mesons ) . here
the label @xmath75 represents the @xmath75- or @xmath230-meson as appropriate ( pseudoscalar and vector mesons are degenerate in this leading approximation ) .
vector current conservation implies that the iw - function is normalized at zero recoil , i.e. that @xmath231 .
this property is particularly important in the extraction of @xmath1 .
to get some insights into the origin of the heavy quark relations above , let us consider @xmath223 decays .
we can rewrite eq .
( [ eq : ffpdef ] ) in terms of the form factors @xmath206 and @xmath232 ( instead of using the helicity basis as in eq .
( [ eq : ffpdef ] ) ) : @xmath233 where @xmath234 . defining the four - velocities @xmath235 and @xmath236 by @xmath237 and @xmath238 , eq .
( [ eq : ffpdef2 ] ) can be rewritten as : @xmath239(v+v^\prime)^\mu \nonumber\\ & & \hspace{0.5in}+\frac{1}{2 } \left [ ( m_b - m_d)f^+(q^2 ) - ( m_b+m_d)f^-(q^2)\right](v - v^\prime)^\mu \ .
\label{eq : ffpdef3}\end{aligned}\ ] ] now in the hqet : @xmath240 where @xmath241 , and the projection operators ( e.g. @xmath242 ) have been absorbed into the heavy quark fields .
@xmath95 represents a heavy pseudoscalar meson , and the factors @xmath243 are chosen in order to introduce a mass - independent normalization of states ( so that the matrix elements are independent of any masses ) .
there is no term proportional to @xmath244 on the right hand side of eq .
( [ eq : ffphqet ] ) , since @xmath245 , and so we obtain the relations : @xmath246 to understand the normalization condition @xmath247 , consider the forward matrix element : @xmath248 where the last relation follows from the conservation of the current @xmath249 which implies that @xmath250 . in the hqet
this matrix element is equal to @xmath251 , and so @xmath247 .
the spin symmetry relates the form - factors of @xmath224 decays to @xmath226 . the relations in eq .
( [ eq : iw ] ) are valid up to perturbative and power corrections .
the precision with which @xmath1 can be extracted is limited by the theoretical uncertainties in the evaluation of these corrections .
nevertheless we are in the fortunate situation that it is uncertainties in _ corrections _
( which are therefore small ) which control the precision .
allowing for these corrections to the results in the heavy quark limit , one writes the decay distribution for @xmath224 decays as : @xmath252 |v_{cb}|^2\ , { \cal f}^2(\omega)\ , \label{eq : distr}\ ] ] where @xmath253 is the iw - function combined with perturbative and power corrections .
it is convenient theoretically to consider this distribution near @xmath254 , in order to exploit the normalization condition @xmath255 . in this case
there are no @xmath256 corrections ( where @xmath257 or @xmath74 ) by virtue of luke s theorem @xcite , so that the expansion of @xmath258 begins like : @xmath259 where @xmath260 represents the perturbative corrections .
the one - loop contribution to @xmath260 has been known for some time now , whereas the two - loop contribution was evaluated last year , with the result @xcite : @xmath261 where i have taken the value of the two loop contribution as an estimate of the error .
the power corrections are much more difficult to estimate reliably .
neubert has recently combined the results of refs .
@xcite to estimate that the @xmath262 terms in the parentheses in eq .
( [ eq : fexpansion ] ) are about @xmath263 and that @xmath264 .
bigi , shifman and uraltsev @xcite , consider the uncertainties to be bigger and obtain @xmath265 . combining the latter , more cautious , theoretical value of @xmath266 , with the experimental result @xcite @xmath267 , readily gives @xmath268 .
here i am only discussing exclusive decays .
buras and fleischer , perform an analysis of both exclusive and inclusive semileptonic decays and quote @xmath269 as their best value @xcite .
@xmath270 gives the scale of the unitarity triangle : @xmath271 in this subsection we consider the heavy - to - light semileptonic decays @xmath273 and @xmath274 which are now being used experimentally to determine the @xmath2 matrix element @xcite .
heavy quark symmetry is less predictive for heavy - to - light decays than for heavy - to - heavy ones . in particular
, there is no normalization condition at zero recoil corresponding to the condition @xmath247 , which is so useful in the extraction of @xmath1 ( see subsection [ subsec : vcb ] ) .
the lack of such a condition puts a premium on the results from nonperturbative calculational techniques , such as lattice qcd or light - cone sum rules .
heavy quark symmetry does , however , give useful scaling laws for the behaviour of the form factors with the mass of the heavy quark ( @xmath175 ) at fixed @xmath275 : @xmath276 these scaling relations are particularly useful in lattice simulations , where the masses of the quarks are varied .
moreover , the heavy quark spin symmetry relates the @xmath277 matrix elements @xcite ( where @xmath211 is a light vector particle ) of the weak current and magnetic moment operators , thereby relating the amplitudes for the two physical processes @xmath278 and @xmath279 , up to @xmath142 flavour symmetry breaking effects .
these relations also provide important checks on theoretical , and in particular on lattice , calculations .
recent compilations of results for @xmath280 , include 0.06 - 0.11 obtained from exclusive decays only ( or @xmath281 from inclusive decays ) @xcite and @xmath282 from both inclusive and exclusive decays @xcite .
in terms of the parameters of the unitarity triangle ( see fig . [
fig : ut ] ) : @xmath283 thus a measurement of @xmath280 implies that the locus of possible positions of the vertex @xmath64 is a circle centred on @xmath96 ( see fig . [
fig : vub2 ] ) . in practice
, of course , because of the theoretical and experimental uncertainties , the circle is replaced by a band .
( 160,90)(0,0 ) ( 10,10)(150,10)(10,10)(60,60)(60,60)(150,10 ) ( 60,65)[lb]@xmath284 ( 140,2)[lt]@xmath285 ( 0,2)[t]@xmath286 ( 10,10)(71,10,110 ) we have seen that the main difficulty in making predictions for weak hadronic decays is our inability to control the long - distance strong interaction effects present in these decays .
the principal technique for the evaluation of these non - perturbative effects is lattice qcd , and i now digress from the main theme of this course to outline briefly the principles underlying lattice calculations of non - perturbative qcd effects in weak decays and to discuss the main sources of uncertainty present in these computations .
in subsection [ subsec : ope ] we have seen that , by using the operator product expansion , it is generally possible to separate the short- and long - distance contributions to weak decay amplitudes into wilson coefficient functions and operator matrix elements respectively . thus , in order to evaluate the non - perturbative qcd effects , it is necessary to compute the matrix elements of local composite operators .
this is achieved in lattice simulations , by the direct computation of correlation functions of multi - local operators composed of quark and gluon fields ( in euclidean space ) : @xmath287[d\psi][d\bar\psi]\,e^{-s}\,o(x_1 , x_2,\ldots , x_n)\ , \label{eq : vev}\ ] ] where @xmath288 is the partition function @xmath289[d\psi][d\bar\psi]\,e^{-s}\ , \label{eq : zdef}\ ] ] @xmath290 is the action and the integrals are over quark and gluon fields at each space - time point . in eq .
( [ eq : vev ] ) @xmath291 is a multi - local operator ; the choice of @xmath292 governs the physics which can be studied .
we now consider the two most frequently encountered cases , for which @xmath293 or @xmath294 . as a first example , let @xmath295 be the bilocal operator @xmath296 where @xmath188 is an interpolating operator for the hadron @xmath189 whose properties we wish to study . in lattice computations
we evaluate the two point correlation function @xmath297 inserting a complete set of states @xmath298 we have : @xmath299 in eq .
( [ eq : c2exp ] ) , the ellipsis represents the contributions from heavier states than @xmath189 , which we assume to be the lightest hadron which can be created by the operator @xmath187 . using the translation operator to move the argument of @xmath188 to zero
, we find in eq .
( [ eq : c2exp2 ] ) is implicitly included as part of the definition of the summation over states in eqs .
( [ eq : c2exp0 ] ) and ( [ eq : c2exp ] ) . ] : @xmath300 where @xmath301 . at large positive times
@xmath302 the contribution from each heavier hadron , @xmath303 with mass @xmath304 say , is suppressed by an exponential factor , @xmath305 , so that the contribution from the lightest state @xmath189 is isolated .
this is illustrated in the following diagram : ( 120,20)(0,10 ) ( 20,20)(100,20)(60,25)[b]@xmath189 ( 20,20)10(100,20)10 ( 20,14)[t]@xmath306(100,14)[t]@xmath302(120,14)[t ] . in lattice simulations
the correlation function @xmath307 is computed numerically , by discretizing space - time ( hence the word _ lattice _ ) , evaluating the functional integral in eq .
( [ eq : vev ] ) by monte - carlo integration . by fitting the results to the expression in eq .
( [ eq : c2exp2 ] ) both the mass @xmath308 and the matrix element @xmath309 can be determined . ] . as an example
consider the case in which @xmath189 is the @xmath3-meson and @xmath188 is the axial current @xmath310 ( with the flavour quantum numbers of the @xmath3-meson ) . in this case one
obtains the value of the leptonic decay constant @xmath127 , defined in eq .
( [ eq : fbdef ] ) .
it will also be useful to consider three - point correlation functions : @xmath311 where , @xmath312 and @xmath313 are the interpolating operators for hadrons @xmath314 and @xmath315 respectively , @xmath292 is a local operator , and we have assumed that @xmath316 . inserting complete sets of states between the operators in eq .
( [ eq : c3def ] ) we obtain @xmath317 where @xmath318 , @xmath319 and the ellipsis represents the contributions from heavier states .
the exponential factors , @xmath320 and @xmath321 , ensure that for large time separations , @xmath322 and @xmath323 , the contributions from the lightest states dominate .
the three - point correlation function is illustrated in the diagram ( 160,20)(0,10 ) ( 20,20)(80,20)(50,25)[b]@xmath314 ( 80,20)(140,20)(110,25)[b]@xmath315 ( 20,20)10(140,20)10 ( 80,20)30.5(80,14)[t]@xmath322(80,27)[b]@xmath292 ( 20,14)[t]@xmath306(140,14)[t]@xmath302(160,14)[t ] .
all the elements on the right - hand side of eq.([eq : c3states ] ) can be determined from two - point correlation functions , with the exception of the matrix element @xmath324 .
thus by computing two- and three - point correlation functions the matrix element @xmath324 can be determined .
the computation of three - point correlation functions is useful , for example , in studying semileptonic and radiative weak decays of hadrons , e.g. if @xmath314 is a @xmath3-meson , @xmath315 a @xmath75 meson and @xmath292 the vector current @xmath325 , then from this correlation function we obtain the form factors relevant for semileptonic @xmath223 decays .
i end this brief summary of lattice computations of hadronic matrix elements with a word about the determination of the lattice spacing @xmath103 .
it is conventional to introduce the parameter @xmath326 , where @xmath327 is the bare coupling constant in the theory with the lattice regularization .
it is @xmath24 @xmath328or equivalently @xmath327@xmath329 which is the input parameter in the simulation , and the corresponding lattice spacing is then determined by requiring that some physical quantity ( which is computed in lattice units ) is equal to the experimental value .
for example , one may compute @xmath330 , where @xmath331 is the mass of the @xmath332-meson , and determine the lattice spacing @xmath103 by dividing the result by @xmath333 mev .
( 132,160)(0,-10 ) ( 66,66)400.8 ( 0,0)(12,0)12(0,12)12 ( 6,145)[b]@xmath103 ( 11,139)(12,139)(1,139)(0,139)(1,139)(11,139 ) ( 66,-10)[t]@xmath31(1,-13)(61,-13)(71,-13)(131,-13 ) ( 1,-13)(0,-13)(131,-13)(132,-13 ) although lattice computations provide the opportunity , in principle , to evaluate the non - perturbative qcd effects in weak decays of heavy quarks from first principles and with no model assumptions or free parameters , in practice the precision of the results is limited by the available computing resources .
for these computations to make sense it is necessary for the lattice to be sufficiently large to accommodate the particle(s ) being studied ( @xmath334fm say , where @xmath31 is the spatial length of the lattice ) , and for the spacing between neighbouring points ( @xmath103 ) to be sufficiently small so that the results are not sensitive to the granularity of the lattice ( @xmath335 ) , see fig .
[ fig : lattice ] .
the number of lattice points in a simulation is limited by the available computing resources ; current simulations are performed with about 1620 points in each spatial direction ( up to about 64 points if the effects of quark loops are neglected , i.e. in the so called `` quenched '' approximation ) . thus it is possible to work on lattices which have a spatial extent of about 2 fm and a lattice spacing of 0.1 fm , perhaps satisying the above requirements .
i now outline the main sources of uncertainty in these computations : * _ statistical errors : _ the functional integrals in eq .
( [ eq : vev ] ) are evaluated by monte - carlo techniques .
this leads to sampling errors , which decrease as the number of field configurations included in the estimate of the integrals is increased . *
_ discretization errors : _ these are artefacts due to the finiteness of the lattice spacing .
much effort is being devoted to reducing these errors either by performing simulations at several values of the lattice spacing and extrapolating the results to the continuum limit ( @xmath336 ) , or by `` improving '' the lattice formulation of qcd so that the error in a simulation at a fixed value of @xmath103 is formally smaller @xcite@xmath337@xcite .
in particular , it has recently been shown to be possible to formulate lattice qcd in such a way that the discretization errors vanish quadratically with the lattice spacing @xcite , even for non - zero quark masses @xcite .
this is in distinction with the traditional wilson formulation @xcite , in which the errors vanish only linearly . in some of the simulations performed in recent years
improved actions and operators have been used . in some of these these studies
the improvement is implemented at tree - level , so that the artefacts formally vanish more quickly @xmath328like @xmath338@xmath329 than for the wilson action .
this tree - level improved action is denoted as the sw ( after sheikholeslami - wohlert who first proposed it @xcite ) or `` clover '' action . * _ extrapolations to physical quark masses : _ it is generally not possible to use physical quark masses in simulations . for the light ( @xmath68- and @xmath67- )
quarks the masses must be chosen such that the corresponding @xmath71-mesons are sufficiently heavy to be insensitive to the finite volume of the lattice .
in addition , as the masses of the quarks are reduced the computer time necessary to maintain the required level of precision increases rapidly . for the heavy quarks @xmath339 ( i.e. for @xmath74 , and particularly for @xmath88 ) the masses must be sufficiently small so that the discretization errors , which are of @xmath340 or @xmath341 , are small . the results obtained in the simulations
must then be extrapolated to those corresponding to physical quark masses . *
_ finite volume effects : _ we require that the results we obtain be independent of the size of the lattice . thus , in principle , the spatial size of the lattice @xmath31 should be @xmath342fm ( in practice @xmath3433fm ) , and as mentioned above , we can not use very light @xmath68- and @xmath67-quarks ( in order to avoid very light pions whose correlation lengths , i.e. the inverses of their masses , would be of @xmath344 or greater ) . * _ contamination from excited states : _ these are the uncertainties due to the effects of the excited states , represented by the ellipsis in eq .
( [ eq : c2exp2 ] ) . in most simulations , this is not a serious source of error ( it is , however , more significant in computations with static quarks ) . indeed , by evaluating correlation functions @xmath345 for a variety of interpolating operators
@xmath346 , it is possible to obtain the masses and matrix elements of the excited hadrons ( for a recent example see @xcite ) . * _ lattice - continuum matching : _ the operators used in lattice simulations are bare operators defined with the lattice spacing as the ultra - violet cut - off . from the matrix elements of the bare lattice composite operators , we need to obtain those of the corresponding renormalized operators defined in some standard continuum renormalization scheme , such as the @xmath347 scheme .
the relation between the matrix elements of lattice and continuum composite operators involves only short - distance physics , and hence can be obtained in perturbation theory .
experience has taught us , however , that the coefficients in lattice perturbation theory can be large , leading to significant uncertainties ( frequently of @xmath348 or more ) . for this reason ,
non - perturbative techniques to evaluate the renormalization constants which relate the bare lattice operators to the renormalized ones have been developed , using chiral ward identities where possible @xcite or by imposing an explicit renormalization condition @xcite ( see also refs .
@xcite ) , thus effectively removing this source of uncertainty in many important cases . + to * _ `` quenching '' : _ in most current simulations the matrix elements are evaluated in the `` quenched '' approximation , in which the effects of virtual quark loops are neglected . for each gluon configuration @xmath349 ,
the functional integral over the quark fields in eq .
( [ eq : vev ] ) can be performed formally , giving the determinant of the dirac operator in the gluon background field corresponding to this configuration .
the numerical evaluation of this determinant is possible , but is computationally very expensive , and for this reason the determinant is frequently set equal to its average value , which is equivalent to neglecting virtual quark loops .
gradually , however , unquenched calculations are beginning to be performed , e.g. in fig .
[ fig : sesamtxl ] we show the lattice spacing obtained by the sesam and t@xmath350l collaborations from four physical quantities in both quenched and unquenched simulations @xcite . in the quenched case
there is a spread of results of about @xmath351 , whereas in the unquenched case the spread is smaller ( although the errors are still sizeable for some of the quantities used ) . in the next 35 years
it should be possible to compute most of the physical quantities discussed below without imposing this approximation .
having discussed the basics of lattice computations , i now briefly present some results from recent simulations studying @xmath352 semileptonic decays .
i start by making the simple observation that from lattice simulations we can only obtain the form factors for part of the physical phase space for these decays . in order to control discretization errors we require that the three - momenta of the @xmath3 , @xmath71 and @xmath332 mesons be small in lattice units .
this implies that we determine the form factors at large values of momentum transfer @xmath353 .
experiments can already reconstruct exclusive semileptonic @xmath354 decays ( see , for example , the review in @xcite ) and as techniques improve and new facilities begin operation , we can expect to be able to compare the lattice form factor calculations directly with experimental data at large @xmath355 .
a proposal in this direction was made by ukqcd @xcite for @xmath278 decays .
to get some idea of the precision that might be reached , they parametrize the differential decay rate distribution near @xmath356 by : @xmath357 where @xmath103 and @xmath88 are parameters , and the phase - space factor @xmath49 is given by @xmath358 .
the constant @xmath103 plays the role of the iw function evaluated at @xmath359 for heavy - to - heavy transitions , but in this case there is no symmetry to determine its value at leading order in the heavy quark effective theory .
ukqcd obtain @xcite @xmath360 the fits are less sensitive to @xmath88 , so it is less well - determined . the result for @xmath103
incorporates a systematic error dominated by the uncertainty ascribed to discretization errors and would lead to an extraction of @xmath361 with less than 10% statistical error and about 12% systematic error from the theoretical input .
the prediction for the @xmath362 distribution based on these numbers is presented in fig .
[ fig : vub ] . with sufficient experimental data
an accurate lattice result at a single value of @xmath355 would be sufficient to fix @xmath361 .
0.7to to in principle , a similar analysis could be applied to the decay @xmath363 .
however , ukqcd find that the difficulty of performing the chiral extrapolation to a realistically light pion from the unphysical pions used in the simulations makes the results less certain .
the @xmath274 decay also has a smaller fraction of events at high @xmath355 , so it will be more difficult experimentally to extract sufficient data in this region for a detailed comparison
. we would also like to know the full @xmath355 dependence of the form factors , which involves a large extrapolation in @xmath355 from the high values where lattice calculations produce results , down to @xmath364 .
in particular the important radiative decay @xmath279 ( which is dominated by penguin diagrams in the standard model ) occurs at @xmath364 , so that existing lattice simulations can not make a direct calculation of the necessary form factors .
much effort is being devoted to extrapolating the lattice results down to small values of @xmath355 exploiting the hqet , light - cone sum - rules @xcite and axiomatic properties of quantum field theory .
these studies are beyond the scope of these lectures and i refer the students to the review articles @xcite for a discussion and references to the original articles .
in this lecture i will discuss @xmath70-@xmath367 and @xmath3-@xmath368 mixing in turn , and consider the implications of the experimental measurements of these processes for our understanding of the unitarity triangle . the process of @xmath4-@xmath5 mixing appears in the standard model at one - loop level through the box - diagrams of fig .
[ fig : box ] ( with the @xmath88-quark replaced by the @xmath72-quark ) , together with possible long - distance contributions . the @xmath0-eigenstates ( @xmath371 and @xmath372 ) are linear combinations of the two strong - interaction eigenstates . ] : @xmath373 and @xmath374 because of the complex phase in the ckm - matrix , the physical states ( the mass eigenstates ) differ from @xmath375 and @xmath376 by a small admixture of the other state : @xmath377 ( the parameter @xmath378 depends on the phase convention chosen for @xmath379 and @xmath380 ) . for exclusive decays of @xmath70-mesons ( which have angular momentum zero ) into two or three pion states , the two pion states are @xmath0-even and the three - pion states are @xmath0-odd .
this implies that the dominant decays are : @xmath381 which is the reason why @xmath382 is much longer lived than @xmath383 ( @xmath31 and @xmath290 stand for _ long _ and _ short _ respectively ) . @xmath382 and @xmath383
are not @xmath0-eigenstates , however , and the decay @xmath384 may occur as a result of the small component of @xmath375 in @xmath385 , and similarly the decay @xmath386 may occur because of the small component of @xmath376 in @xmath387 .
such @xmath0-violating decays which occur due to the fact that the mass eigenstates are not @xmath0-eigenstates are called _ indirect @xmath0-violating decays_. as mentioned above , the parameter @xmath378 depends on the convention used to define the phases of the states .
a measure of the strength of indirect @xmath0-violation is given by the physical parameter @xmath369 defined by the ratio : @xmath388 where @xmath64 represents the amplitude for the corresponding process and @xmath389 implies that the two pions are in an isospin 0 state . the numerical value in eq .
( [ eq : epskdef ] ) is the empirical result .
directly @xmath0-violating decays are those in which a @xmath0-even ( -odd ) state decays into a @xmath0-odd ( -even ) one : @xmath390 ( 300,124)(-27,-10 ) ( 194,120)(194,80)(194,80)(194,78 ) ( 194,73)[t]@xmath391(186,90)[rb]direct @xmath392 ( 237,120)(237,100)(237,100)(257,100)(257,100)(259,100 ) ( 264,100)[l]@xmath391(235,90)[tl]indirect @xmath393 consider the possible quark subprocesses contributing to @xmath394 decays shown in fig .
[ fig : direct ] . the diagrams in fig .
[ fig : direct](b ) and ( c ) are purely real , whereas that of fig .
[ fig : direct](a ) is complex .
the two pions in the final state can have total isospin @xmath389 or 2 ; @xmath395 is not possible by bose symmetry .
the final states in the diagrams of fig .
[ fig : direct](a ) and ( b ) clearly have @xmath389 , whereas that in the diagram of fig .
[ fig : direct](c ) can have either @xmath389 or @xmath396 .
thus direct @xmath0-violation in kaon decays manifests itself as a non - zero relative phase between the @xmath389 and @xmath396 amplitudes .
( 420,120)(0,-30 ) ( 10,10)(110,10)(10,60)(110,60)(60,10)(60,60)210 ( 60,60)(30,180,0)220 ( 5,60)[r]@xmath72(5,10)[r]@xmath135 ( 115,60)[l]@xmath67(115,10)[l]@xmath135 ( 60,-10)[t]@xmath389 , complex ( 60,-30)[t](a ) ( 160,10)(260,10)(160,60)(260,60)(210,10)(210,60)210 ( 155,60)[r]@xmath72(155,10)[r]@xmath135 ( 265,60)[l]@xmath68(265,10)[l]@xmath118 ( 210,-10)[t]@xmath389 , real ( 210,-30)[t](b ) ( 310,10)(410,10)(310,60)(410,60 ) ( 360,60)(400,35)29 ( 305,60)[r]@xmath72(305,10)[r]@xmath135 ( 400,35)(420,47)(400,35)(420,23 ) ( 415,60)[l]@xmath68(425,47)[l]@xmath118 ( 425,23)[l]@xmath67 ( 360,-10)[t]@xmath389 or @xmath397 , real ( 360,-30)[t](c)(415,10)[l]@xmath135 of course the gluonic corrections to the diagrams in fig .
[ fig : direct ] generate _ strong phases _ , which we denote by @xmath398 , where the suffix @xmath306 or @xmath397 represents the isospin of the final state .
these phases are independent of the form of the weak hamiltonian . using the clebsch - gordan coefficients for the combination of two isospin 1 pions we write : @xmath399 the parameter @xmath400 , which is used as a measure of direct @xmath0-violation in kaon decays ,
is defined by : @xmath401 where is manifestly zero if the phases of the @xmath389 and @xmath396 amplitudes are the same . ]
@xmath402 experimentally the two parameters @xmath369 ( which , following standard conventions i rename from now on as @xmath403 , @xmath404 ) and @xmath400 can be determined by measuring the ratios : @xmath405 direct @xmath0-violation is found to be considerable smaller than indirect violation . by measuring the decays and using @xmath406 the na31 @xcite and e371 @xcite experiments find @xmath407 and @xmath408 respectively for @xmath370 . the particle data group @xcite summarises these results as : @xmath409 we will have to wait for the results of the current generation of experiments , which will have a sensitivity of @xmath410 , to confirm or deny whether the value is indeed different from zero . in this subsection
i discuss the theoretical determination of @xmath403 , following the discussion and notation in the review article by buras and fleischer @xcite .
we need to know the matrix element : @xmath411 where we write the effective hamiltonian in the form : @xmath412\nonumber\\ & & \mbox{\ \ \ } \times \left[\alpha_s^{(3)}(\mu)\right]^{-\frac{2}{9}}\,\left[1 + \frac{\alpha_s^{(3)}(\mu)}{4\pi}j_3\right]\,o^{\delta s=2}(\mu)\ + \ \textrm{h.c.}\ , \label{eq : heffexpr}\end{aligned}\ ] ] the renormalization scale @xmath149 is chosen to be much smaller than @xmath413 , @xmath414 and @xmath415 ( the dependence on @xmath416 is eliminated by unitarity , @xmath417 ) .
the non - perturbative qcd effects for @xmath4-@xmath5 mixing are all contained in the matrix elements of the single local composite operator @xmath418 the remaining terms in eq .
( [ eq : heffexpr ] ) can be calculated in perturbation theory .
the @xmath419 s are the expressions for the box - diagrams without qcd corrections : @xmath420 + \cdots\ .
\label{eq : soxx}\end{aligned}\ ] ] @xmath421 and @xmath422 are mass - dependent short - distance qcd factors , at nlo ( next - to leading order ) @xmath423 where the errors reflect the uncertainty in @xmath424 and the quark masses .
the remaining factors in eq .
( [ eq : heffexpr ] ) are short - distance qcd effects , whose @xmath149-dependence cancels that in the operator @xmath425 ( in the ndr - renormalization scheme @xmath426=1.895 ) .
as mentioned above , all the non - perturbative qcd corrections are contained in the matrix elements of @xmath425 .
it is conventional to introduce the @xmath427 parameter by the definition : @xmath428 the motivation for introducing the parameter @xmath427 in this way comes from the vacuum saturation approximation ( in which @xmath429 ) , but there is no loss of generality in this definition .
@xmath430 depends on the renormalization scale , and it is convenient to introduce an ( almost ) renormalization group invariant @xmath427 by : @xmath431^{-\frac{2}{9}}\,\left[1 + \frac{\alpha_s^{(3)}(\mu)}{4\pi}j_3\right]\ .\ ] ] the weakest link in the theoretical calculation of @xmath403 is the evaluation of @xmath427 .
many lattice computations have been performed , and a recent compilation of results gave @xmath432 @xcite . a calculation based on the @xmath433 approximation gave @xmath434 @xcite .
buras and fleischer perform their analysis with @xmath435 @xmath403 is given in terms of the quantities introduced above by @xmath436 -\textrm{re}\,\lambda_t\,\eta_2s_0(x_t)\right\}\,e^{i\frac{\pi}{4}}\ , \label{eq : epsexpr}\ ] ] where @xmath437 and the mass difference @xmath438 .
we obtain information about the unitarity triangle by comparing the theoretical prediction with the experimentally measured value of @xmath403 . in terms of @xmath439 and @xmath440 the prediction is the hyperbola @xmath441\,a^2\,b_k\,=\,0.226\ , \label{eq : bkhyperbola}\ ] ] where @xmath442 .
this is illustrated in fig .
[ fig : eps ] .
( 210,100)(-50,0 ) ( 10,10)(150,10)(10,10)(60,60)(60,60)(150,10 ) ( 65,65)[rb]@xmath284
( 140,2)[lt]@xmath285 ( 0,2)[t]@xmath286 ( -50,30,60,60,110,110 ) in this subsection i briefly review the elements in the theoretical predictions for @xmath370 and related quantities .
i start by rewriting eq .
( [ eq : epsprimedef ] ) in the form : @xmath444 where @xmath445 even after about 40 years , we still do not understand theoretically the @xmath443 rule , i.e. the empirical observation that @xmath446 .
the rate for transitions in which the isospin changes by 1/2 ( @xmath443 transitions ) is enhanced by a factor of about 450 over those in which it changes by 3/2 ( @xmath447 transitions ) .
there is a relatively small enhancement of the @xmath443 amplitudes , by a factor of 2 - 3 , due to perturbative effects ( i.e. in the calculation of the wilson coefficient function ) .
the remaining enhancement is assumed to be due to long - distance qcd effects , perhaps because the matrix - elements of penguin operators are large , but this has still to be convincingly demonstrated .
in the recent analyses of @xmath370 , one combines the experimental values for the real parts of the amplitudes ( re@xmath448gev , re@xmath449gev , @xmath450 ) , with theoretical predictions for the imaginary parts .
the effective hamiltonian for these @xmath451 transitions takes the form : @xmath452 where @xmath453 .
there are 10 independent operators in the @xmath451 effective hamiltonian , and the major uncertainty is in the values of their matrix elements and in the mass of the strange quark @xmath454 .
the most complete analyses have been performed by buras et al @xcite who find : @xmath455 where @xmath456 and @xmath457 refer to two different ways of treating the many uncertainties ( more and less conservative respectively ) , and by the rome group @xcite who find @xmath458 the various contributions to @xmath370 have different signs , and there are significant cancellations .
for this reason even the expected sign can not be predicted with full confidence .
i end this subsection with a brief summary of the main points : * an understanding of the @xmath443 rule would be an important milestone in controlling non - perturbative qcd effects .
this is a realistic , but non - trivial , challenge for the lattice community .
* a measurement of a non - zero value for @xmath370 would be a very important qualitative step in particle physics ; confirming the existence of direct @xmath0-violation . * the uncertainties in the values of the matrix elements ,
make it difficult to make a precise prediction for @xmath370 .
we expect it to be at the several @xmath459 level , but accidental cancellations between contributions with opposite signs may make it smaller than this . in the coming years , experiments at cern ,
fnal and at da@xmath460ne will be sensitive to a value of about @xmath461 .
in this subsection we consider @xmath3-@xmath368 mixing , from which we get information about the @xmath462 and @xmath463 elements of the ckm - matrix .
i start with some general formalism for the mixing of a neutral psudoscalar meson @xmath464 with its antiparticle @xmath465 .
i write the wave - function in two - component form : @xmath466 the time dependence of @xmath467 is given by the schr " odinger equation : @xmath468 where @xmath469 is the @xmath66 mass - matrix " , whose elements are equal to @xmath470 . using the optical theorem ,
the absorbtive part of this matrix comes from real intermediate states : @xmath471 for @xmath6-@xmath7 mixing , there are no contributions to @xmath472 from intermediate states containing the top quark .
the mass - matrix takes the form @xmath473 where the diagonal terms are equal by cpt - invariance , and @xmath64 , @xmath474 and @xmath355 are complex .
the eigenvalues of this matrix are @xmath475 , so that the mass difference between the two physical eigenstates is @xmath476 for the @xmath3-system , the box diagrams have contributions from @xmath477-quarks in the loops ( see fig .
[ fig : box ] ) . by unitarity
these contribute only to the real part of the mass - matrix .
the mass behaviour of the box diagrams in fig .
[ fig : box ] , and the large value of @xmath478 imply that these contributions dominate and that @xmath479 for @xmath3-@xmath368 mixing ( this is not the case for @xmath70-@xmath367 mixing ) .
thus @xmath480 , and can be calculated from the box diagrams .
the evaluation of the box diagrams follows similarly to that in the kaon system .
the non - perturbative qcd effects are contained in the matrix element of the @xmath481 operator : @xmath482 and it is also convenient and conventional to introduce the @xmath483 parameter analogously to the definition of @xmath427 in eq .
( [ eq : bkdef ] ) .
the theoretical expression for the mass difference is : @xmath484 ) . throughout this discussion
i have been implicitly assuming that we are considering neutral @xmath485 mesons , but an analogous discussion holds for the @xmath486 system . in the kaon system ,
all the non - perturbative qcd effects were contained in the parameter @xmath427 , since the leptonic decay constant @xmath487 is a measured quantity .
for the @xmath3-mesons this is not the case , and one is obliged to use model estimates or lattice or sum - rule calclulations for both the decay constants @xmath127 and the @xmath483-parameters ( or of the matrix element of @xmath488 itself ) .
for example , in a recent compilation of lattice results we obtained @xcite : @xmath489 and @xmath490 combining all the elements one obtains : @xmath491 ^ 2\ , \left(\frac{\bar m_t(m_t)}{170\gev}\right)^{1.52}\ , \frac{|v_{td}|}{0.0088}\ \frac{\eta_b}{0.55 } \label{eq : dmd}\\ \delta m_{b_s } & = & 15.1\,\textrm{ps}^{-1}\,\left [ \frac{f_{b_s}\sqrt{b_{b_s}}}{240\mev}\right]^2\ , \left(\frac{\bar m_t(m_t)}{170\gev}\right)^{1.52}\ , \frac{|v_{ts}|}{0.040}\ \frac{\eta_b}{0.55}\ .
\label{eq : dms}\end{aligned}\ ] ] the mass difference @xmath492 is a measure of the oscillation frequency to change from a @xmath6 to a @xmath7 and vice - versa .
imagine that we start with a @xmath6 at time @xmath493 , then the propabilities that we have a @xmath6 or @xmath7 at a later time @xmath477 are @xmath494 and @xmath495 respectively . defining @xmath496 and @xmath497 to be the probabilities ( obtained by integrating over time )
that when the meson decays it is a @xmath3 or @xmath368 respectively , we obtain @xmath498\ \ \textrm{and}\ \
p(\bar b^0 ) = \frac{1}{2}\,\left[\frac{1}{\gamma}-\frac{\gamma}{\gamma^2 + ( \delta m)^2}\right]\ .\ ] ] the ratio of these integrated probabilities is given by @xmath499 where @xmath500 .
the 1996 particle data group review quotes : @xmath501 in agreement with theoretical expectations and @xmath502 the oscillation in the @xmath486 system are too rapid for observation in experiments which have been performed up to date .
what do these measurements imply for the determination of the position of the unitarity triangle ? in principle
a measurement of @xmath503 allows for a determination of the @xmath462 element of the ckm matrix .
the side @xmath504 of the unitarity triangle is given by : @xmath505 so that a measurement of @xmath462 ( assuming knowledge of @xmath1 ) would imply that the locus of possible positions of the vertex @xmath64 lies on a given circle centred on @xmath3 , see fig .
[ fig : vtd ] .
( 210,120)(0,0 ) ( 10,10)(150,10)(10,10)(60,60)(60,60)(150,10 ) ( 65,65)[rb]@xmath284 ( 140,2)[lt]@xmath285 ( 0,2)[t]@xmath286 ( 150,10)(103,0,170 )
in this final lecture i will consider two topics in @xmath3-physics , @xmath0-violation ( and mixing - induced @xmath0-violation , in particular ) and the application of the heavy quark expansion to inclusive decays ( and to beauty - lifetimes and the semileptonic branching ratio , in particular ) .
i will also make some brief remarks about how difficult it is to make predictions for exclusive nonleptonic decays .
measurements of @xmath0-asymmetries in neutral @xmath3-meson decays into @xmath0-eigenstates will allow us to address 3 fundamental questions : 1 .
is the phase in the ckm matrix the only source of @xmath0-violation ?
2 . what are the precise values of the ckm parameters ? we shall see that the principal systematic errors , particularly those due to hadronic uncertainties , are reduced significantly in some cases .
3 . is there new physics in the quark - sector ? a particularly important class of decays of neutral @xmath3-meson systems , are _ mixing induced _ @xmath0-violating decays ( these are not possible in decays of charged @xmath3-mesons ) .
i now briefly review this important topic , for which we need to return to the formalism introduced in the discussion of @xmath3-@xmath368 mixing in subsec .
[ subsec : bbar ] above .
the two neutral mass - eigenstates are given by @xmath506 starting with a @xmath6 meson at time @xmath493 , its subsequent evolution is governed by the schr " odinger equation ( [ eq : se ] ) : @xmath507 where @xmath508\,\exp[-imt]\ , \cos\left(\frac{\delta m\,t}{2}\right)\ , \label{eq : gp}\\ g_-(t ) & = & \exp\left[-\frac{\gamma t}{2}\right]\,\exp[-imt]\ , i\sin\left(\frac{\delta m\,t}{2}\right)\ , \label{eq : gm } \end{aligned}\ ] ] and @xmath509 meson at @xmath493 , the time evolution is @xmath510 . ] .
let @xmath511 be a @xmath0-eigenstate and @xmath512 be the amplitudes @xmath513 defining @xmath514 we have @xmath515\ \ \ \textrm{and}\ \ \ \langle f_{cp}\,|\,\mathcal{h}\,|\,\bar b^0_{\mathrm{phys}}\rangle = a\,\frac{p}{q}\,[g_-(t)\,+\,\lambda\,g_+(t)]\ .
\label{eq : btof}\ ] ] the time - dependent rates for initially pure @xmath6 or @xmath7 states to decay into the @xmath0-eigenstate @xmath511 at time @xmath477 are given by : @xmath516\label{eq : btofcp}\\ \gamma(\bar b^0_{\mathrm{phys}}(t)\to f_{cp } ) & = & |a|^2\,e^{-\gamma t } \left[\frac{1-|\lambda|^2}{2}\,+\,\frac{1-|\lambda|^2}{2}\cos(\delta m\,t ) + \textrm{im}\,\lambda\,\sin(\delta m\,t)\right]\ .\label{eq : bbartofcp}\end{aligned}\ ] ] the time - dependent asymmetry is defined as : @xmath517 if , as is the case in some important applications , @xmath518 ( which is the case if @xmath519 ) and @xmath520 ( examples of this will be presented below ) , then @xmath521 and the first term on the right - hand side of eq.([eq : aexpr ] ) vanishes . the form of the amplitudes @xmath64 and @xmath522 is : @xmath523 where the sum is over all the contributions to the process , the @xmath524 are real , @xmath525 are the strong phases and the @xmath526 are the phases from the ckm matrix . in the most favourable situation ,
all the contributions have a single ckm phase ( @xmath527 say ) and @xmath528 . under the assumption we are making that @xmath529 , @xmath530 , and @xmath531 .
thus @xmath532 from the box diagrams of fig .
[ fig : box ] we obtain : @xmath533 to illustrate the above discussion let us consider three processes in which the @xmath88-quark decays through the subprocess @xmath534 .
the corresponding tree - level diagram is @xmath535{$b$}\text(85,45)[l]{$u_i$ } \text(110,25)[l]{$d_j$}\text(110,5)[l]{$\bar u_i$ } \end{picture } } \textrm{for which}\ \ \ \ \frac{\bar a}{a}=\frac{v_{ib}v_{ij}^*}{v_{ib}^*v_{ij } } \ .\label{eq : abarovera}\ ] ] [ [ mathbfb_dto - jpsik_s ] ] @xmath536 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in this case @xmath537 the first factor in eq . (
[ eq : li ] ) is the factor @xmath538 in eq .
( [ eq : qoverpb ] ) , the third factor is the analogous one for the final state kaon , and the second factor is @xmath539 as in eq .
( [ eq : abarovera ] ) with @xmath540 and @xmath541 .
we recall from the discussion of the first lecture that the angle @xmath542 there is also a small penguin contribution to this process , due to the subprocess : ( 140,60)(0,20 ) ( 10,50)(130,50)(5,50)[r]@xmath88(135,50)[l]@xmath72 ( 70,50)(30,180,0)320 ( 65,56)[b]@xmath477(100,75)[b]@xmath22 ( 70,50)(120,20)210 ( 120,20)(130,28)(120,20)(130,12 ) ( 95,27)[t]@xmath92(140,12)[c ] , which is proportional to @xmath543 , whose phase is , to a very good approximation , equal to that of @xmath544 .
thus hadronic uncertainties are negligible in the determination of the angle @xmath24 from this process , and for this reason we can consider it a _ gold - plated _ one .
this is an ( almost ) ideal situation .
[ [ mathbfb_dtopipi- ] ] @xmath545 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + for this process , in the notation of eq . (
[ eq : abarovera ] ) , we have @xmath546 and @xmath547 , so that @xmath548 and im@xmath549 , where from the first lecture we recall that @xmath550 in this case there is also a small penguin contribution , which is proportional to @xmath551 , which has a different phase from the tree diagram .
this implies that the hadronic uncertainites are greater than for the process @xmath552 , although they are still reasonably small ( and can be reduced further using isospin analysis ) .
[ [ mathbfb_storho0-k_s ] ] @xmath553 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + for this process in the notation of eq . (
[ eq : abarovera ] ) , we have @xmath546 and @xmath547 again , so that @xmath554 using the unitarity relations and the fact that @xmath555 we find that im@xmath556 . although this process is frequently proposed as a potential way of measuring the angle @xmath12 , the penguin contributions are relatively large and buras and fleischer stress that it is a _ wrong way
_ @xcite .
it has been suggested that it might be possible to use combinations of @xmath557 amplitudes to determine the angle @xmath12 @xcite .
this is illustrated in fig .
[ fig : triangles ] , where @xmath558 , but clearly a measurement of all the amplitudes is a huge experimental challenge .
( 300,150)(0,0 ) ( 30,10)(250,10 ) ( 30,10)(270,120)(250,10)(270,120 ) ( 30,10)(100,80)(250,10)(100,80 ) ( 140,5)[t]@xmath559 ( 55,45)[r]@xmath560 ( 265,65)[l]@xmath561 ( 280,130)[br]@xmath562 ( 212,128)(232,108)(232,108)(234,106 ) ( 70,85)[lb]@xmath563 ( 130,84)(122,76)(122,76)(120,74 ) ( 248,27)[r]@xmath564(250,10)(32,79,156 ) the exciting experimental and theoretical program of research into mixing induced @xmath0-asymetries will continue at the ( near ) future @xmath3-factories and dedicated @xmath3-experiments .
observation of @xmath0-violation in decays of charged @xmath3-mesons would demonstrate the existence of direct @xmath0-violation , but the hadronic uncertainties make it unlikely that they can be used to determine the parameters of the unitarity triangle with the required precision .
since we need an interference between different contributions to the amplitude , we write the amplitudes in the form : @xmath565 where i have exhibited two contributions to each amplitude , @xmath566 are ckm - matrix elements and the @xmath567 are strong interaction phases .
for example in the process @xmath568 the interference is between curent - current and penguin contributions , and in the decay @xmath569 it is between penguin diagrams with internal @xmath68 and @xmath74 quarks .
the @xmath0-violating asymmetry is now defined by @xmath570 from which we see that in order to have a non - zero asymmetry we require both im@xmath571 and @xmath572 to be non - zero .
the strong interaction phases are very difficult to quantify , so it is hard to use measurements of these @xmath0-asymetries to determine the ckm matrix elements . in this subsection
i discuss inclusive nonleptonic decays of beauty hadrons in general , and two very interesting phenomenological problems in particular the lifetimes of the hadrons and their semileptonic branching ratios .
the discussion will use the formalism of bigi ( see ref .
@xcite and references therein ) , developed and used by them and many other groups , in which inclusive quantities are expanded in inverse powers of the mass of the heavy quark , e.g. @xmath573 where @xmath574 is the full or partial width of a beauty hadron @xmath575 , @xmath576 are coefficients which can be computed in perturbation theory ( the coefficient functions obtained when matching qcd onto the hqet ) and @xmath577 are the matrix elements of the kinetic energy and chromomagnetic operators respectively : @xmath578 and @xmath184 is the field of the heavy ( static ) quark .
an important feature of the general expression in eq .
( [ eq : widthexpansion ] ) is the absence of terms of @xmath579 , which is a consequence of the absence of any operators of dimension 4 which can appear in the corresponding ope @xcite .
i will not discuss the derivation of the expansion in eq .
( [ eq : widthexpansion ] ) in detail .
one starts , similarly to the derivation of the ope for moments of deep - inelastic structure functions , by using the optical theorem which states that the width is given by the imaginary part of the forward elastic amplitude . in this case
this amplitude has two insertions of the weak hamiltonian . since the @xmath88-quark is heavy and decays into lighter states , the separation of these two insertions is small , and eq .
( [ eq : widthexpansion ] ) is the corresponding short distance expansion ( for the structure functions it is a light - cone expansion ) .
we now apply this expansion to a study of beauty lifetimes and the semileptonic brabching ratio . using the expression in eq .
( [ eq : widthexpansion ] ) for the widths , one readily finds the following results for the ratios of lifetimes : @xmath580 where @xmath581 and @xmath582 . in order to obtain the result in eq.([eq : ratiolambdab ] ) , one needs to know the difference of the kinetic energies of the @xmath88-quark in the baryon and meson . to leading order in the heavy quark expansion we have : @xmath583 from equation ( [ eq : kediffs ] ) , and using the recent measurement of @xmath584 from cdf @xcite
, one finds that the right hand side is very small ( less than about 0.01 gev@xmath585 ) .
the matrix elements of the chromomagnetic operator are obtained from the mass difference of the @xmath586- and @xmath3-mesons ( see eq.([eq : l12def ] ) ) and from the fact that the two valence quarks in the @xmath587 are in a spin - zero state ( which implies that @xmath588 ) .
the theoretical predictions in eqs .
( [ eq : ratiomesons ] ) and ( [ eq : ratiolambdab ] ) can be compared with the experimental measurements @xmath589 the discrepancy between the theoretical and experimental results for the ratio @xmath590 in eqs .
( [ eq : ratiolambdab ] ) and ( [ eq : ratiolambdabexp ] ) is notable .
it raises the question of whether the @xmath591 contributions are surprisingly large , or whether there is a more fundamental problem .
i postpone consideration of the latter possibility and start with a discussion of the @xmath591 terms .
( 300,50)(-30,15 ) ( 10,41)(40,41)(100,41)(40,41)(100,41)(130,41 ) ( 70,41)(15,30)(0 ) ( 40,41)1.50(100,41)1.50 ( 69,56)(71,56)(69,26)(71,26 ) ( 20,36)[tl]@xmath88(115,36)[tr]@xmath88 ( 67,63)[bl]@xmath74(67,19)[tl]@xmath67(80,44)[br]@xmath118 ( 140,41)(158,41)(158,41)(160,41 ) ( 180,41)(230,41)(205,41)20 ( 205,33)[t]@xmath592 ( 300,50)(260,40 ) ( 300,41)(330,41)(330,41)(300,26 ) ( 330,41)(390,41 ) ( 390,41)(420,41)(420,26)(390,41 ) ( 361,56)(359,56 ) ( 330,41)1.50(390,41)1.50 ( 430,41)(448,41)(448,41)(450,41 ) ( 475,51)(495,41)(495,41)(515,51 ) ( 495,41)(475,31)(515,31)(495,41 ) ( 495,41)20 ( 310,46)[bl]@xmath88(312,28)[tl]@xmath135 ( 361,63)[b]@xmath118(361,34)[t]@xmath74 ( 410,46)[br]@xmath88(407,28)[tr]@xmath135 ( 500,23)[t]@xmath593 at first sight it seems strange to consider the @xmath594 corrections to be a potential source of large corrections , when the @xmath595 terms are only about 2% . however , it is only at this order that the `` spectator '' quark contributes , and so these contributions lead directly to differences in lifetimes for hadrons with different light quark constituents ( consider for example the lower diagram in fig .
[ fig : diagrams ] , for which , using the short - distance expansion , one obtains operators of dimension 6 ) .
moreover , the coefficient functions of these operators are relatively large , which may be attributed to the fact that the lower diagram in fig . [
fig : diagrams ] is a one - loop graph , whereas the corresponding diagrams for the leading contributions are two - loop graphs ( see , for example , the upper diagram of fig .
[ fig : diagrams ] ) .
the corresponding phase - space enhancement factor is 16@xmath596 or so .
we will therefore only consider the contributions from the corresponding four - quark operators , neglecting other @xmath591 corrections which do not have the phase space enhancement @xcite . for each light - quark flavour @xmath209 , there are four of these : @xmath597 where @xmath598 are the generators of colour @xmath142 and @xmath209 represents the fields of the light quarks . thus we need to evaluate the matrix elements of these four operators . for mesons , following ref .
@xcite , i introduce the parametrization @xmath599 where @xmath149 is the renormalization scale .
we have chosen to use @xmath600 as the renormalization scale .
bigi et al .
( see ref .
@xcite and references therein ) prefer to use a typical hadronic scale , and estimate the matrix elements using a factorization hypothesis at this low scale .
operators renormalized at different scales can be related using renormalization group equations in the hqet ( sometimes called hybrid renormalization @xcite ) .
for example , if we assume that factorisation holds at a low scale @xmath149 such that @xmath601 , then , using the ( leading order ) renormalization group equations , one finds @xmath602 and @xmath603 s and @xmath604 s had been defined at this scale ( instead of @xmath600 ) they would have been 1 and 0 respectively . ] . in the limit of a large number of colours @xmath605 , @xmath606 whereas @xmath607 . for the @xmath587
, heavy quark symmetry implies that @xmath608 so that there are only two independent parameters .
it is convenient to replace the operator @xmath609 , by @xmath610 defined by @xmath611 where @xmath612 are colour labels , and to express physical quantities in terms of the two parameters @xmath613 and @xmath614 defined by @xmath615 we do not know the values of these parameters . in quark models
@xmath616 , and @xmath617@xmath618 .
using experimental values of the hyperfine splittings and quark models , it has been suggested that @xmath614 may be larger @xcite , e.g. @xmath619 the lifetime ratios can now be written in terms of the six parameters @xmath620 and @xmath614 ( as well as @xmath127 ) : @xmath621 where the effective weak lagrangian has been renormalized at @xmath622 .
the central question is whether it is possible , with reasonable " values of the parameters , to obtain agreement with the experimental numbers in eq .
( [ eq : ratiolambdabexp ] ) . at this stage in our knowledge ,
the answer depends somewhat on what is meant by _
reasonable_. for example , neubert , guided by the arguments outlined above , has considered these ratios by varying the parameters in the following ranges @xcite : @xmath623\,;\ \ \epsilon_i\in\left[-\frac{1}{3},\frac{1}{3}\right]\ , ; \ \
r\in[0.25 , 2.5]\ , ; \ \ \left(\frac{f_b}{200\mev}\right)^2\in[0.8,1.2]\ .
\label{eq : ranges}\ ] ] he concludes that , within these ranges , it is just possible to obtain agreement at the two standard deviation level for large values of @xmath614 ( @xmath624 ) and negative values of @xmath625 .
lattice studies of the corresponding matrix elements are underway ; a recent qcd sum - rule calculation has found a small value of @xmath614 , @xmath626@xmath627 @xcite .
if the lattice calculations confirm that the parameter @xmath614 is small , or find that the other parameters are not in the appropriate ranges , then we have a breakdown of our understanding .
if no explanation can be found within the standard formulation , then we will be forced to take seriously the possible breakdown of local duality .
this is beginning to be studied in toy field theories @xcite .
to this was the name given by blok @xcite to the observation that the experimental value of the semileptonic branching ratio @xmath628 appeared to be lower than expected theoretically . in eq .
( [ eq : bsldef ] ) the sum is over the three species of lepton , and @xmath629 and @xmath630 are the widths of the hadronic and rare decays respectively .
bigi concluded that a branching ratio of less than 12.5% can not be accommodated by theory @xcite . since then bagan have completed the calculation of the @xmath631 corrections , and in particular of the @xmath632 component of the decay ( including the effects of the mass of the charm quark ) @xcite ; these have the effect of decreasing @xmath633 . with m. neubert
, we used this input to reevaluate the branching ratio and charm counting ( @xmath634 , the average number of charmed particles per @xmath3-decay ) @xcite finding , e.g. @xmath635 @xmath149 is the renormalization scale and the dependence on this scale is a reflection of our ignorance of higher order perturbative corrections .
the experimental situation is somewhat confused , see fig .
[ fig : bsl ] . in his compilation at the ichep conference last year @xcite ,
richman found that the semileptonic branching ratio obtained from @xmath3-mesons from the @xmath636 is : @xmath637 whereas that from lep is : @xmath638 the label @xmath88 for the lep measurement indicates that the decays from beauty hadrons other than the @xmath3-meson are included . using the measured fractions of the different hadrons and their lifetimes , and assuming that the semileptonic widths of all the beauty hadrons are the same , one finds : @xmath639 amplifying the discrepancy .
it is very difficult to understand such a discrepancy theoretically , since the theoretical calculation only involves @xmath640 ( and not @xmath629 for which the uncertainties are much larger ) . in view of the experimental discrepancy
, i consider the problem of the lifetime ratio @xmath590 , described in subsection [ subsubsec : lifetimes ] above , to be the more significant theoretical one .
a large amount of experimental data is becoming available , particularly from the cleo collaboration ( see ref .
@xcite and references therein ) on two - body ( exclusive ) nonleptonic decays of @xmath3-mesons .
this is an exciting new field of investigation , which will undoubtedly teach us much about subtle aspects of the standard model .
unfortunately , at our present level of understanding we are not able to compute the amplitudes from first principles , and are forced to make assumptions about the non - perturbative qcd effects . frequently these assumptions concern factorization , i.e. whether matrix elements of operators in the effective hamiltonian which are products of two @xmath20 currents can be written in terms of products of matrix elements of the currents , e.g. @xmath641 where the first matrix element on the right hand side could be obtained from semileptonic decays ( see sec .
[ subsec : vcb ] ) and the second from the known value of @xmath642 . in some cases
, the factorization hypothesis can be motivated by the concept of _ colour transparency _ , that a light colour singlet meson , produced with large energy at the weak decay vertex ( and which is hence a small colour - dipole ) , will travel a long way from the interaction region before it grows to a size where it could interact strongly @xcite .
it is possible that in some cases factorization will give a reasonable first approximation and in many others it will not be useful at all . thus current analyses are limited to a semi - quantitative level , and at this stage it is not possible to endorse any of the approaches .
i will not discuss these analyses in these lectures , but to wish to point out the importance of this emerging field , ( see refs .
@xcite and @xcite for more extensive discussions ) .
to in these lectures i have briefly reviewed the formalism required for the study of the weak decays in the standard model of particle physics . from these studies
we are attempting to measure the parameters of the standard model , to understand the origin of @xmath0-violation and to look for signatures of physics beyond the standard model .
i have explained that the major theoretical difficulty in interpreting the wealth of experimental data is our inability to control the non - perturbative strong interaction effects to sufficient precision .
througout these lectures i have illustrated the general discussion with applications to various physical processes .
the current status of the uncertainties in determining the vertex @xmath64 of the unitarity triangle , from the analysis of buras and fleischer is reproduced in fig .
[ fig : rhoeta ] .
the circular arcs are obtained from information on @xmath643 mixing , using a modified analysis to that discussed in section [ subsec : bbar ] above . in order to reduce the hadronic uncertainties
, we can take the ratio of the expressions for @xmath503 and @xmath644 in eqs .
( [ eq : dmd ] ) and ( [ eq : dms ] ) , and , using the upper bounds on @xmath503 ( 0.482 ps@xmath645 ) and @xmath646 ( 0.993 ) , obtain a bound for the quantity in eq .
( [ eq : vtdut ] ) : @xmath647 where @xmath648 is defined in eq .
( [ eq : bresults ] ) , where the lattice result is also given . the curves in fig .
[ fig : rhoeta ] represent these bounds for several choices of @xmath648 and @xmath644 @xmath644 is written as @xmath649 . ] .
i end these lectures by summarising the _ tourist guide _ of buras and fleischer @xcite to different processes , in which stars are awarded to the processes depending on the level of theoretical uncertainty . 1
. these are processes in which there are _ no _ theoretical uncertainties i.e. they are particularly small ( less than 2% say ) . among these
is the mixing - induced @xmath0-violating asymmetry in @xmath650 discussed in section [ subsubsec : cpvneutral ] , from which we expect to determine the angle @xmath24 of the unitarity triangle .
other quantities in this category which i have not had time to discuss include the ratio of ( inclusive ) branching ratios @xmath651 from which one would obtain @xmath652 and the decays @xmath653 ( which would give im@xmath654 ) and @xmath655 ( which would give @xmath656 ) . for the kaon
decays i have explained in section [ subsubsec : sinthetac ] , that the small mass difference between the quarks in the @xmath70 and @xmath71 mesons implies that the hadronic uncertainties are small .
2 . the processes in this category have _ small _ hadronic uncertainties , of the order of 5 - 10% say .
they include the evaluation of @xmath1 discussed in sec .
[ subsec : vcb ] , @xmath657 discussed in section [ subsec : bbar ] ( from which we obtain @xmath652 ) and the mixing - induced @xmath0-asymmetry in @xmath658 ( from which we would obtain the angle @xmath59 ) discussed in section [ subsubsec : cpvneutral ] . there is a large set of other mixing - induced asymetries in this category , yielding , in principle , all three angles .
being perhaps a little optimistic , the determination of @xmath659 in inclusive decays , also belongs to this category .
3 . the processes in this category have _ smallish _ hadronic uncertainties , perhaps in the region of 15% or so . of the quantities discussed in this course , @xmath660 , @xmath649 ( section [ subsec : bbar ] ) and @xmath403 ( section [ subsubsec : epsilon ] ) belong in this category .
4 . as we have seen in these lectures , there are also important quantities for which the theoretical uncertainties are too large for the experimental data to give accurate information about the parameters of the standard model ( and hence these quantities are not ascribed any stars ) , but which nevertheless give interesting qualitative information .
these include @xmath370 ( discussed in section [ subsubsec : deltai12 ] ) and most @xmath0-asymmetries in decays of charged @xmath3 s .
non - leptonic decays of @xmath3-mesons in general belong to this category .
of course it is to be hoped and expected that theoretical progress will continue , and that the interesting physical quantities will gradually acquire more stars .
the theoretical community is working very hard to achieve this .
it is a pleasure to thank the students , lecturers and organisers of this school for providing such a lively and stimulating atmosphere .
it was very thought - provoking and instructive to respond to the many excellent questions from the students without being able to rely on an armoury of theoretical formalism .
i warmly acknowledge my collaborators , particularly , guido martinelli and matthias neubert , with whom i learned many of the topics discussed in these lectures .
i am grateful to giulia de divitiis and luigi del debbio for their comments on the manuscript .
* d#1 * # 1phys .
* # 1 * # 1phys . lett . * b#1 * # 1nucl . phys . *
b#1 * # 1nucl . phys .
* b ( proc . suppl.)#1 * # 1nucl . phys . * a ( proc . suppl.)#1 * # 1z .
phys * c#1 * # 1nucl .
* a#1 * # 1commun . math .
* # 1 * # 1phys . rep .
* # 1 * 99 a.j .
buras and r. fleischer , hep - ph/9704376 , to appear in _ heavy flavours _ ( 2nd edition ) , eds .
buras and m. lindner , ( world scientific , singapore , 1998 ) j.f .
donoghue , e. golowich and b.r .
holstein , _ dynamics of the standard model _ ( cambridge university press , cambridge , uk , 1992 ) _ @xmath3-decays _
2nd edition , ed .
s.stone ( world scientific , singapore , 1994 ) _ @xmath0-violation _ , ed . c. jarlskog , ( world scientific , singapore , 1988 ) j. petersen , v. zakharov , these proceedings n. cabibbo , ( 1963 ) 531 ; + m. kobayashi and t. maskawa , prog .
* 49 * ( 1973 ) 652 .
barnett , ( 1996 ) 1 l. wolfenstein , ( 1983 ) 1945 e. hagberg et al .
, nucl - ex/9609002 s.l .
glashow , j. iliopoulos and l. maiani , ( 1970 ) 1285 p. nason , these proceedings m. neubert , hep - ph/9702375 , to appear in _ heavy flavours _
( 2nd edition ) , eds .
buras and m. lindner , ( world scientific , singapore , 1998 ) m. luke , ( 1990 ) 447 a. czarnecki , ( 1996 ) 4124 ; + a. czarnecki and k. melnikov ( 1997 ) 3630 a. falk and m. neubert , ( 1993 ) 2965 and 2982 . t. mannel .
( 1994 ) 428 .
shifman , n.g .
uraltsev and a.i .
vainshtein , ( 1995 ) 2217 .
i. bigi , m. shifman and n. uraltsev , hep - ph/9703290 l. di ciaccio , to appear in the proceedings of the 7th international symposium on heavy flavor physics , santa barbara , california , july 1997 l.k .
gibbons , proc .
patterson , proc .
n. isgur and m. b. wise , ( 1990 ) 2388 p.a .
griffin , m. masip and m. mcguigan , ( 1994 ) 5751 h. albrecht et al . , ( 1987 ) 245 ; ( 1991 ) 297 ; r. fulton et al .
, ( 1990 ) 16 ; j. bartelt et al . , (
1993 ) 411 k. symanzik , in mathematical problems in theoretical physics , eds .
r. schrader , lecture notes in physics vol .
153 ( springer , new york , 1982 ) ; ( 1983 ) 187 ; ( 1983 ) 205 m. lscher and p. weisz , ( 1985 ) 59 ; erratum ( 1985 ) 433 b. sheikholeslami and r. wohlert , ( 1985 ) 572 g. heatlie , ( 1991 ) 266 m. lscher , ( 1996 ) 365 ; m. lscher and p. weisz , ( 1996 ) 429 ; m. lscher , ( 1997 ) 323 g. martinelli et al . , (
1997 ) 141 k.g .
wilson , ( 1974 ) 2445 ; k.g .
wilson , in new phenomena in subnuclear physics , ed .
a. zichichi ( plenum press , new york , 1975 ) p. 69
p. lacock et al .
, ( 1995 ) 6403 m. bochicchio et al . , ( 1985 ) 331 ; l.h . karsten and j. smit , ( 1981 ) 103 ; l. maiani and g. martinelli , ( 1986 ) 265 ; g. martinelli , s. petrarca , c. t. sachrajda and a. vladikas , ( 1993 ) 241 ; erratum ( 1993 ) 660 g. martinelli , c. pittori , c.t .
sachrajda , m. testa and a. vladikas , ( 1995 ) 81 m. lscher , s. sint , r. sommer and h. wittig , ( 1997 ) 344 sesam and t@xmath350l collaborations , h. hoeber , proc .
, hep - lat/9709137 collaboration , j.m .
flynn , ( 1996 ) 327 a. ali , v. braun and h. simma , ( 1994 ) 437 ; p. ball , proc .
31st rencontres de moriond : electroweak interactions and unified theories , les arcs , france , march 1996 , p. 365
braun , proc .
europhysics conf . on high energy physics , brussels , belgium ,
july 1995 , p. 436 ; p. ball and v.m .
braun , ( 1997 ) 5561 j. flynn and c.t .
sachrajda , hep - lat/9710057 , to appear in _ heavy flavours _
( 2nd edition ) , eds .
buras and m. lindner , ( world scientific , singapore , 1998 ) c.t .
sachrajda , hep - ph/9711386 , to be published in the proceedings of the 18th international symposium on lepton - photon interactions ( lp97 ) , hamburg , germany , 28 july - 1 aug 1997 g.d .
barr et al . , ( 1993 ) 233 l.k
. gibbons et al . , ( 1993 ) 1203 j. flynn , proc .
bardeen , a.j .
buras and j.m .
grard , ( 1988 ) 343 ; j.m .
grard , acta phys . pol .
* b21 * ( 1990 ) 257 ; j. bijnens aand j. prades , ( 1995 ) 523 a.j .
buras , m. jamin and m.e .
lautenbacher , ( 1996 ) 749 m. ciuchini et al . , ( 1995 ) 239 ; m. ciuchini , ( 1997 ) 149 m. gronau and d. wyler , ( 1991 ) 172 i.i .
bigi , b. blok , m.a .
shifman , n.g .
uraltsev and a.i .
vainshtein , in _ @xmath3-decays _ ( second edition ) , edited by s. stone , ( world scientific , singapore 1994 ) , p. 132
i. bigi , n. uraltsev and a. vainshtein , ( 1992 ) 430 ; erratum : ( 1993 ) 477 ; b. blok and m. shifman , ( 1993 ) 441 ; ( 1993 ) 459 ; particle data group , r.m .
barnett et al . , 1997 off - year partial update for the 1998 edition available on the pdg www pages ( url : http://pdg.lbl.gov/ ) m. neubert and c. t. sachrajda , ( 1997 ) 339 m.a .
shifman and m.b .
voloshin , sov .
j. nucl .
* 41 * ( 1985 ) 120 ; jetp * 64 * ( 1986 ) 698 ; sov .
j. nucl .
* 45 * ( 1987 ) 292 ; h.d .
politzer and m.b .
wise , ( 1988 ) 681 ; ( 1988 ) 504 j. rosner , ( 1996 ) 267 m. neubert , hep - ph/9707217 , to appear in proc .
2nd int . conf . on b physics and cp violation , honolulu , hawaii , march 1997 p. colangelo and f. defazio , ( 1996 ) 371 b. blok , hep - ph/9707218 ; b. blok , m. shifman and d.x .
zhang , hep - ph/9709333 p. colangelo , c. a. dominquez and g. nardulli , hep - ph/9705390 i.i .
bigi , b. blok , m. shifman and a. vainshtein , ( 1994 ) 408 e. bagan , p. ball , v. braun and p. gosdzinsky , ( 1994 ) 3 , ( 1995 ) 362 [ erratum ( 1996 ) 363 ] ; e. bagan , p. ball , b. fiol and p. gosdzinsky , ( 1995 ) 546 j.d .
richman , proc .
m. feindt , to appear in the proceedings of the 1997 europhysics conference on high energy physics , jerusalem , august 1997 p. drell , hep - ex/9711011 , to be published in the proceedings of the 18th international symposium on lepton - photon interactions ( lp97 ) , hamburg , germany , 28 july - 1 aug 1997 j.d .
bjorken , in _ new developments in high - energy physics _ , edited by e.g. floratos and a. vergazelakis , ( 1989 ) 325 m.j . dugan and b.
grinstein , ( 1991 ) 583 m. neubert and b. stech , hep - ph/9705292 ; to appear in heavy flavours ( 2nd ed . ) , ed .
buras and m. linder ( world scientific , singapore ) | in these four lectures i review the theory and phenomenology of weak decays of quarks , and their rle in the determination of the parameters of the standard model of particle physics , in testing subtle features of the theory and in searching for signatures of _ new _ physics .
attempts to understand @xmath0-violation in current and future experiments is discussed .
shep 98/01 + hep - ph/9801343 * flavour physics * + c.t .
sachrajda + department of physics and astronomy , university of southampton + southampton so17 1bj , uk * abstract * _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ in these four lectures
i review the theory and phenomenology of weak decays of quarks , and their rle in the determination of the parameters of the standard model of particle physics , in testing subtle features of the theory and in searching for signatures of _ new _ physics .
attempts to understand @xmath0-violation in current and future experiments is discussed . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ to appear in the proceedings of the 1997 european school of high - energy physics , menstrup , denmark , 15 may - 7 june 1997 .
january 1998 |
experimental observations ( see , for example , @xcite and references therein ) via imaging and other means have established the occurrence of bifurcations in the stick - slip dynamics of the peeling of a stretchable adhesive tape pulled with a constant velocity off a circular spool which is free to rotate around its center .
the older theoretical studies @xcite focused on the equations of motion and the more recent ones @xcite recognized the differential - algebraic equation ( dae ) structure of the formulations , and proposed mathematical approaches such as the inertial regularization @xcite to simulate the dynamics .
however , research issues such as ( i ) the differential index ( see later in this work ) of the dae model , ( ii ) relationship of the differential index in the stick - slip regime to the bifurcation , and ( iii ) satisfaction of the adhesion - shear constitutive relationship ( the algebraic constraint ) in relation to the stick - slip regime of the dynamics follow naturally from the dae structure of the peeling dynamics . in this context
it may be recalled that ordinary differential equation ( ode ) analysis has its limitations in dealing with daes since daes are not odes @xcite . in this work
, we address the above questions and show ( i ) that the peeling dynamics dae is a variable differential index system , with differential index ranging from two to possibly very high values in the stick - slip regime compared to being one otherwise , ( ii ) that the higher differential index implies an exponentially faster time scale for the peeling velocity relative to the peel front angle between the tape and the spool , and/or the tensile displacement of the tape , and ( iii ) that the above two properties lead to the nonlinear bifurcations .
finally , ( iv ) we develop a homogenized ode representation of the peeling dynamics dae model along with a characteristic time scale for the homogenization .
the homogenized ode scheme is illustrated with physically consistent numerical simulations .
following the modeling approach in @xcite , the dynamics of the peeling of an adhesive tape off a circular spool is given as @xmath1 [ peel ] where @xmath0 , the contact or the peel front angle , is the angle between the radius passing through the point where the tape peels off the spool and the radius coinciding with the horizontal , @xmath2 is the tensile displacement or stretch of the tape which has an young s modulus @xmath3 and cross - sectional area @xmath4 so that @xmath5 , @xmath6 is the speed at which the tape is peeled at the point of the contact with the spool , or simply the peeling velocity and @xmath7 is the constant speed at which the free end of the tape is pulled along its longitudinal axis .
the peeled tape from the point of contact to the point where it is pulled , is assumed to have negligible mass compared to the circular spool which has only in - plane rotational degree of freedom about its center of mass .
the radius of the spool is @xmath8 ; @xmath9 is the constant distance between the center of mass of the spool and the point where the longitudinal axis of the peeled tape meets the horizontal .
the moment of inertia of the spool about its center of mass is @xmath10 .
the angular velocity of the spool is denoted by @xmath11 .
the function @xmath12 is the constitutive model for the adhesion depending on the constant pulling velocity @xmath7 and the peeling velocity @xmath6 .
usually @xmath12 is modeled such that it increases as the tape sticks until it reaches a maximum and decreases as the tape peels off by slipping @xcite,@xcite .
the parameters @xmath13 are constants with respect to time and the other variables . the time varying quantities are @xmath14 and @xmath6 .
the angular velocity @xmath11 , @xmath15 , the peel front or contact angle @xmath0 and @xmath16 are positive counter - clockwise .
the model is shown in figure [ model ] .
the model ( [ peel ] ) can be formally obtained from the lagrangian of the system . with reference to figure [ model ] and in line with the lagrangian approach in @xcite
the following work - energy relationships can be written in the generalized coordinate system @xmath17 .
@xmath18 [ lag ] from which the euler - lagrange equations @xmath19 [ lagderive ] yield @xmath20 [ lageqn ] respectively . in ( [ lag ] ) , one can alternatively view the potential energy part of the lagrangian as the elastic energy stored in the tape , @xmath21 , to which the net work done on the system , @xmath22 , is added .
it may be noted that when the approximations @xmath23 and that @xmath0 is a small angle hold , ( [ lageqn ] ) reduces to the same model as used in @xcite .
the equations in ( [ peel ] ) coincide with those in @xcite when @xmath0 is considered positive clockwise and @xmath24 holds .
the equations of motion ( [ av][tb ] ) , the material stretch rate equation ( [ str ] ) and the shear - adhesion constitutive model of the peeling ( [ sac ] ) constitute a dae system in which @xmath6 is the algebraic variable .
the variables @xmath25 and @xmath2 are the differential variables .
a measure of difficulty of solving a dae system is its differential index ( for definition and detailed treatment see @xcite ) . often referred to simply as index , the differential index is effectively the number of times differentiations of the equations in the dae system that should be done for obtaining a canonical first order ode system in all the unknown variables of the dae .
theorem 5.4.1 in @xcite connects the difficulty of a numerical solution of a dae to its differential index as the condition number of the jacobian of the implicit integration method and shows it to increase exponentially with the differential index of the dae system .
we characterize the peeling model ( [ peel ] ) with respect to its structure to show that the stick slip is essentially a problem of the dae ( [ peel ] ) switching to an arbitrarily large differential index from differential index 1 and that there is a resultant change in the time scale of evolution of the algebraic variable @xmath6 , the peeling velocity , compared to the differential variables of ( [ peel ] ) .
we define the vectors @xmath26 we introduce the subscript @xmath27 for indexing a time varying quantity , e.g. , @xmath28 and @xmath29 .
the derivatives @xmath30 are jacobian matrices whereas @xmath31 is a gradient vector .
the differential index of ( [ peel ] ) is analyzed in the following .
let @xmath32 exist and be non - zero at a time point @xmath27 with a @xmath33 satisfying ( [ peel ] ) .
then ( [ peel ] ) has a unique local solution @xmath34,~ \delta t \ge 0 $ ] as a function of time.[sol ] we observe that using the fundamental theorem of integral calculus , the jacobian of the dae ( [ peel ] ) with respect to @xmath35 can be written as @xmath36 of which @xmath37 is the schur complement , @xmath38 being the identity matrix .
the jacobian @xmath39 is invertible since @xmath37 is non - zero . by the implicit function theorem , a unique solution @xmath40,~ \delta t \ge 0 $ ] giving @xmath41 as a function of time
can be found over a neighborhood of @xmath42 if the jacobian @xmath39 is invertible .
here we remark that we primarily intend to study the local behavior of the stick - slip dynamics and the existence of a local solution suffices for the purpose .
subject to certain conditions , one can use the gronwal or bihari inequality , the leray - schauder principle and schauder fixed point theorem ( similar to the peano existence for an ode ; cf .
chapter 3 of @xcite ) to obtain a proof of peano existence and osgood uniqueness of the global solution of ( [ peel ] ) for a given consistent initial condition .
however a treatment of the global solution of ( [ peel ] ) is outside the scope of the present work .
the local differential index of ( [ peel ] ) is @xmath43 where @xmath44 are as defined in lemma [ sol ] .
[ pi ] consider the jacobian of ( [ peel ] ) with @xmath45 as the @xmath46 operator .
then , by the implicit function theorem , we can rewrite @xmath39 as @xmath47 which must be invertible at a @xmath33 satisfying ( [ peel ] ) for a unique local solution of ( [ peel ] ) to exist .
being the integration operator , @xmath48 is @xmath49 , @xmath50 .
it is obvious that the 2-norm of the local jacobian @xmath39 at any time point @xmath27 is @xmath51 .
the 2-norm of the inverse of @xmath39 is the same order as that of absolute value of @xmath52 which is the inverse of the schur complement . using the neumann series yields @xmath53 the order of @xmath54 depends on the 2-norms of the coefficients of @xmath55 in the neumann expansion and
let this be @xmath56 where @xmath57 is a natural number . since @xmath50 , @xmath58 is @xmath59 or @xmath60 .
hence the 2-norm condition number of @xmath39 is @xmath60 . thus scaling the right hand side of ( [ peel ] ) with an operator of @xmath61 will well - condition @xmath39 , which by lemma [ sol ] is sufficient for obtaining a unique local solution of ( [ peel ] ) .
this indicates @xmath62 differentiations of the equations in ( [ peel ] ) .
hence @xmath63 , i.e. , @xmath43 differentiations are needed to obtain a canonical ode for @xmath64 and @xmath63 is thus the differential index of ( [ peel ] ) .
the slip of the peeled tape occurs when the sticking resistance that has reached a maximum is overcome .
the maximum adhesive force is reached when @xmath65 occurs in the constitutive relationship ( [ sac ] ) . in the following lemma ,
we show that in the neighborhood of the maximum sticking resistance or adhesion and the subsequent slip on the yielding of the adhesive , ( [ peel ] ) tends to have to an arbitrarily high order of singularity as characterized by its high differential index .
this is in contrast to the regime when the adhesion is not in the neighborhood of a maximum , or , @xmath66 or greater implies that ( [ peel ] ) has differential index one , since @xmath64 can be determined uniquely as a function of @xmath67 from ( [ sac ] ) by the implicit function theorem . in the peeling model ( [ peel ] ) , let @xmath68 be a function which has at least a maximum over the range of values that the solution @xmath64 of ( [ peel ] ) takes and @xmath69 monotonically and smoothly goes to zero and has a negative value for values of peeling velocity which are greater than @xmath64 at which @xmath70 . then , along its solution in the neighborhood of @xmath42 , ( [ peel ] ) has differential index two if @xmath68 is a maximum at @xmath64 .
further , in a neighborhood of @xmath64 at which @xmath68 is a maximum , ( [ peel ] ) can have an arbitrarily large differential index .
when @xmath71 at the maximum , the peeling velocity solution @xmath72,~\delta t > 0 $ ] can not be determined from ( [ sac ] ) , i.e. , one differentiation is not sufficient to obtain a notional first order ode in @xmath6 .
hence the differential index of ( [ peel ] ) is higher than one if @xmath71 . by the implicit function theorem and from ( [ av][str ] )
, we obtain @xmath73 where @xmath74
. then , by lemma [ sol ] , @xmath6 can be determined explicitly in terms of @xmath75 and @xmath27 in the neighborhood of @xmath42 if @xmath76 is non - zero .
if @xmath65 at @xmath64 , and a @xmath77 is chosen , then @xmath78 which is @xmath79 and hence , by lemma [ pi ] , ( [ peel ] ) has differential index two .
let @xmath80 and @xmath81 be small positive numbers . if @xmath12is a maximum at @xmath82 , then in a neighborhood @xmath83 , @xmath84 holds since @xmath85 decreases monotonically from positive value to zero as @xmath6 decreases to @xmath64 .
for the values of @xmath86 , @xmath87 is a small negative number by the property of the function @xmath12 .
then , @xmath88 at some time point @xmath89 is such that @xmath90 and @xmath91 .
the differential index of ( [ peel ] ) is consequently @xmath92 by lemma [ pi ] .
since @xmath93 goes to zero smoothly and monotonically and since @xmath74 , @xmath93 can cancel the first @xmath94 arbitrary large number of terms of the neumann series expansion .
the local jacobian @xmath39 becomes rank deficient as the differential index tends to be arbitrarily high and @xmath95 . from lemma [ sol ] and [ pi ] ,
then ( [ peel ] ) will no longer have a unique local solution over @xmath96 $ ] for some suitable @xmath97 . from the above lemmas and the rank deficiency of the jacobian @xmath39 at the stick - slip , we conclude the following .
the peeling model ( [ peel ] ) is a dae that has variable differential index which is at least one and can be arbitrarily high making the local solutions of ( [ peel ] ) non - unique at time points leading up to and following the maximum adhesion .
[ ithm ] physically , the high differential index near the maximum adhesion and the slip affects the process modeled by ( [ peel ] ) over short time sub - intervals over which @xmath98 , @xmath99 being the local differential index of ( [ peel ] ) .
thus , the two time scales emerge with respect to the stick - slip dynamics of the peeling of an adhesive tape : one during when differential index is one and the other when the differential index rapidly increases to an arbitrarily high value .
the second time scale is of interest with respect to a study of nonlinear bifurcation , the behavior of the peeling velocity @xmath6 at slip and the homogenization of @xmath6 over the same time scale .
in this section we investigate the time scale of the stick - slip process .
let @xmath77 .
then @xmath100 where @xmath37 is as defined in lemma [ sol ] .
[ dvdxlem ] taking total differentials on ( [ sac ] ) we get @xmath101 where @xmath102 from lemma [ hi ] .
let @xmath103 be @xmath104 , and @xmath105 $ ] .
then , @xmath106 where @xmath107 } \sqrt { |\alpha(s)-\alpha(t)|^2 + |u(s)-u(t)|^2}$ ] with @xmath108 at any time point @xmath109 . [ ts ]
we observe that @xmath110 is of @xmath111 . from ( [ dvdx ] ) in lemma [ dvdxlem ]
, we obtain by pre - multiplying both sides with @xmath112 , @xmath113 where @xmath114 is the pseudo - inverse and @xmath115 is a rank one matrix , @xmath116 being the unit vector : @xmath117 using the structure of @xmath116 , and by taking norms we obtain from the right hand side of ( [ lsqx ] ) @xmath118 in which @xmath105 $ ] is a constant independent of @xmath55 .
then , taking norms on the both sides of ( [ lsqx ] ) and by applying the cauchy - schwarz inequality , we get @xmath119 where @xmath120 are constants independent of @xmath55 . from the above lemmas [ dvdxlem ] and [ ts ] together with lemma [ pi ] , and since @xmath121 , we conclude the following result . in ( [ peel ] ) the peeling velocity @xmath64 changes
exponentially faster than the time scale of change of either the peel front angle or the tape s tensile displacement or a linear combination of both , the exponent of the time scale being one less than the local differential index of ( [ peel ] ) .
[ expv ] as a corollary to theorem [ expv ] and from lemma [ hi ] it is obvious that in the neighborhood of the time points at which @xmath122 , i.e. , at the stick - slips , the magnitude of the change in the peeling velocities can be arbitrarily high .
thus the peeling velocity undergoes a ( relatively ) stiff change when the maximum adhesion is approached or just following the slip .
consequently , we have two distinct regimes in the dynamics of peeling of an adhesive tape : * a regime during which the variables @xmath123 change with respect to time in the same scale as the peeling velocity @xmath6 .
this happens when the local differential index of ( [ peel ] ) is one , i.e. , @xmath124 . we call this the _ slow scale_. * another regime when @xmath125 , i.e. , when the local differential index of ( [ peel ] ) is greater than one .
this is the stick or adhesion regime followed by the slip , during which the peeling velocity changes exponentially faster compared to the change in @xmath0 or @xmath2 or in both .
we shall refer to this as the _
fast scale_.
we remark that as a straightforward consequence of lemma [ ts ] a reformulation of ( [ sac ] ) as a an ode in which a small positive quantity is multiplied with the time rate of change of the peeling velocity may not always satisfy the constitutive relationship ( [ sac ] ) , especially , in the fast scale of the stick - slip regime when the rate of change of the peeling velocity is as large as the inverse of the small multiplier . in this context
, we consider the ode formulation in @xcite , which is arrived at by considering an additional kinetic energy term in the lagrangian due to the stretch rate of the extremely small mass of the tape and not by simply introducing a singular ode .
the constitutive relationship ( [ sac ] ) then is an ode and not an algebraic equation . @xmath126 where @xmath127 is a small mass , involves @xmath128 on the left hand side .
if @xmath94 is such that @xmath129 grows as @xmath130 or faster at the stick - slip points at which the schur complement tends to go to zero , then , @xmath131 does not necessarily go to zero in the high differential index stick - slip regime .
also , @xmath132 and @xmath133 do not grow as fast as @xmath64 at the the stick - slip points with high local differential index .
in such a case the role of the small mass becomes that of an inertial regularization parameter relative to the nearly singular ode model of the peeling dynamics .
if an @xmath134 is chosen , this formulation has the same characteristics as a high differential index constraint . in the slower scale without the stick - slip ( in which the local differential index is one )
@xmath64 does not change exponentially faster in time and the above formulation satisfies ( [ sac ] ) in the limit as @xmath135 .
however , in the present work we have assumed that the stretched tape is mathematically massless , i.e. , any kinetic energy of the stretching tape is negligible .
this is consistent with the standard experimental set - up described in works such @xcite .
thus our approach obtains the dae ( [ peel ] ) and concerns about its behavior and regularization in the high differential index regime .
in this section we see if the dynamics in the fast scale during the intermittent stick - slip affect nonlinear bifurcation in this regime . as @xmath136 , we show that the model ( [ peel ] ) is driven by the kinetics and kinematics ( [ av][str ] ) and not the adhesion - shear constitutive relationship ( [ sac ] ) . as the tape sticks almost the maximum , due to the rotational inertia of the spool peel front angle @xmath0 changes and due to the constant pulling velocity @xmath7 the tensile displacement @xmath2 changes , perturbing the dynamics ( [ av][str ] ) .
let @xmath33 satisfy ( [ peel ] ) at @xmath27 such that @xmath137 and @xmath138 .
a perturbation @xmath139 of the peel front angle @xmath140 or @xmath141 of the tensile displacement @xmath142 or a combination of both regularizes the jacobian @xmath39 .
[ reglem ] by ( [ tseq ] ) in lemma [ ts ] , a sufficient perturbation @xmath139 of the peel front angle @xmath140 or @xmath141 of the tensile displacement @xmath142 or a combination of both can affect a non - zero change @xmath81 , since @xmath143 in ( [ tseq ] ) for @xmath138 .
when @xmath144 is @xmath145 or more , the schur complement @xmath37 of the jacobian @xmath146 of ( [ peel ] ) becomes non - zero and @xmath146 becomes invertible due to this regularization affected by the perturbation .
regularization of the jacobian @xmath146 of ( [ peel ] ) implies that the dae ( [ peel ] ) can have a solution over some time interval containing the time point @xmath27 in the stick - slip regime .
however this solution is not unique since the condition number of @xmath146 now depends on the perturbation @xmath147 from the rotational inertia of the spool or @xmath148 of the tensile diplacement of the tape due to the constant @xmath7 or both .
lemma [ reglem ] shows how the dynamics at stick - slip may continue drawing from the perturbations from rotational inertia of the spool or from the constant pulling velocity or both , rather than the relaxation of the constitutive relationship ( [ sac ] ) .
let @xmath33 at a time point @xmath27 in the stick - slip regime satisfy ( [ peel ] ) such that the local differential index of ( [ peel ] ) is significantly more than unity and possibly arbitrarily large .
let the dae ( [ peel ] ) have a solution @xmath149 , @xmath150 being a time interval containing @xmath27 .
then , following @xcite(cf .
chapter 10 , definition 28.1 ) , we define @xmath33 as a _ local nonlinear bifurcation _ point of ( [ peel ] ) iff @xmath151 hold with @xmath152 being a solution of ( [ peel ] ) over a time interval @xmath153 containing @xmath27 for each @xmath154 such that @xmath155 for all @xmath156 .
let @xmath33 together with @xmath157 satisfy ( [ peel ] ) in the fast scale of the stick slip dynamics .
then ( [ peel ] ) has a local nonlinear bifurcation point at @xmath33 in the stick - slip regime of the peeling of an adhesive tape .
[ bif ] let @xmath140 or @xmath142 or both be perturbed as in the condition of lemma [ reglem ] so that @xmath158 is non - zero in magnitude and that @xmath159 with @xmath160 ( @xmath161 being a suitable vector p - norm ) .
then , by lemma [ reglem ] , @xmath157 , is perturbed successively as @xmath162 . due to the perturbation of the schur complement @xmath37
, we obtain a sequence of invertible jacobian matrices @xmath163 of ( [ peel ] ) . by the implicit function theorem
, each perturbation leads to the existence of a unique solution @xmath164 of ( [ peel ] ) over a time interval @xmath165 containing the time point @xmath27 at which @xmath35 satisfies ( [ peel ] ) . in the limit as @xmath166 , @xmath167 so that @xmath168 tends to @xmath169 and @xmath170 with @xmath171 tending to be arbitrarily small ( due to the schur complement approaching zero ) .
then by the definition stated above , @xmath33 is a local nonlinear bifurcation point of ( [ peel ] ) . while the perturbation due to rotational inertia and constant pulling velocity advances the dynamics , by lemma [ ts ] the time scale in which the peeling velocity changes is exponentially faster than that of @xmath140 and/or @xmath142 .
this indicates that the peeling velocity becomes highly sensitive to small perturbations at the local nonlinear bifurcation point .
thus there is a near jump in the peeling velocity with the shear force peeling the tape remaining almost unchanged .
the shear force is a function of the differential variables @xmath0 and @xmath2 that change exponentially slowly compared to the peeling velocity .
we conjecture that the exponentially fast jump like change in the peeling velocity in the stick - slip regime contributes to the experimentally observed intense release of energy in acoustic or triboluminescence @xcite form due the instantaneous breaking of the molecular bonds in the process of shearing of the adhesive over an almost negligible time interval .
it is obvious that capturing the arbitrarily large changes in the peeling velocity at the local nonlinear bifurcation points of ( [ peel ] ) is difficult because of the possibly arbitrarily high local differential index of ( [ peel ] ) .
most dae solvers can not cope with differential index greater than @xmath172 due to theorem 5.4.1 in @xcite .
a computational method involving finding repeatedly the bifurcation points in time in order to deal with the singular points by stopping and restarting the dae integration algorithm with regularization is expensive especially for those @xmath7 s at which the stick - slip regime dominates . at a pull velocity , @xmath7 ,
for which the peeling dynamics is mostly in the fast scale of stick slip regime , a direct numerical simulation of the dae ( [ peel ] ) may be thus difficult and likely inefficient compared to the computational effort .
this calls for a consistent reformulation of ( [ peel ] ) so that the numerical integration of the reformulated algebraic constraint of ( [ peel ] ) would correctly average out ( weak convergence ) the actual peeling velocity response at the bifurcation points .
we assume that the pulling velocity @xmath7 is such that the peeling dynamics has intermittent stick - slip , i.e. , both the fast and slow scales .
we choose a @xmath173 such that @xmath174 is invertible over a small time interval of length @xmath80 at every @xmath175 , ~ \delta t > 0 $ ]
. then @xmath55 acts as the characteristic homogenization time scale in ( [ dvdx ] ) and we re - write ( [ dvdx ] ) in the following integral form over the time interval @xmath176 $ ] , @xmath177 [ hvv ] where @xmath178 $ ] and @xmath179 acts as a projection of the slower dynamics of differential variables @xmath75 onto the faster stick - slip dynamics of the peeling velocities @xmath6
. the reformulation of ( [ peel ] ) involving the homogenized peeling velocity is then obtained by appending ( [ hv ] ) to ( [ av][str ] ) . by lemma
[ ts ] , the integral equation ( [ hv ] ) can be seen as a scaling of the time of relaxation of the peel front angle and tensile displacement of the tape by @xmath180 to match that of the peeling velocity .
if there is no stick - slip at any @xmath175 $ ] , then @xmath174 will be invertible even with @xmath181 by the lemmas [ pi ] and [ hi ] .
on the other hand , if there is a bifurcation point in @xmath176 $ ] , then ( [ peel ] ) will have a very high differential index at the bifurcation points by the lemmas [ pi ] and [ hi ] and consequently , ( [ hv ] ) can be written only with a @xmath55 significantly greater than zero if @xmath174 must be invertible at each @xmath175 $ ] .
thus the minimum @xmath55 at which the locally high differential index dae ( [ peel ] ) can be integrated by a specific , at least @xmath182-stable ( cf .
@xcite ) implicit numerical method to a prescribed accuracy will indicate the smallest time scale to which the solution of ( [ peel ] ) can be numerically resolved by that particular method .
the multiple time scale expansion of the homogenized peeling velocity @xmath6 is provided by ( [ hvex ] ) .
let @xmath121 and @xmath175 $ ] , @xmath97 being a suitable small time interval , be such that @xmath183 and that @xmath174 is invertibe .
also , let @xmath184 .
then , for @xmath178 $ ] , @xmath185 where @xmath186 are the coefficients in the expansion ( [ msv ] )
. [ msvexp ] follows immediately from ( [ hvv ] ) .
each function @xmath187 is the generalized time derivative of the component of the peeling velocity in the @xmath188th scale with @xmath189 as the test function .
it is in this sense that the multiple time scale expansion indicates homogenization of the peeling velocity . as @xmath181
, the larger eigenvalues including the higher frequencies are captured in the multiple scale expansion ( [ msv ] ) . as @xmath190 , @xmath6
is averaged to the same time scale as the differential variables and the oscillations are damped out since @xmath191 provides significant damping in ( [ hvv ] ) to the effect of the shear force peeling the tape .
we remark here that ( [ hvv ] ) is obtained by taking total differentials on ( [ sac ] ) using the implicit function theorem ; and is not an artifact of derivation from the system lagrangian , i.e. , it does not alter the underlying physics of the system but simply exploits the mathematical structure of the dae ( [ peel ] ) to introduce the characteristic time scale and the resultant homogenized ode ( [ hv ] ) . the homogenization of @xmath6 can be elucidated by considering the following case .
let @xmath33 satisfying ( [ peel ] ) at @xmath27 be such that @xmath192 , and for a @xmath121 , let @xmath193 be @xmath194 and @xmath195 be @xmath145 .
then the local differential index is @xmath196 and ( [ peel ] ) is already in the fast scale of the stick slip dynamics . further , suppose @xmath197 but not @xmath198 .
if @xmath199 $ ] be such a time interval that the above conditions hold at any @xmath200 $ ] , then , in the said time interval , using the re - formulation ( [ hv ] ) of ( [ sac ] ) would lead to the following homogenization as indicated by the multiscale expansion ( [ hvex ] ) .
\nonumber\end{aligned}\ ] ] it may be noted that the homogenization time scaling is exponential in @xmath55 , i.e. , exponent is @xmath172 in this specific example . the time scale for changes in the peeling velocity
are thus squeezed as @xmath202 , scaling the average peeling velocity over the slower time scale of the differential variables to a faster one .
let @xmath203 be successive homogenization time scales such that @xmath174 is invertible at any @xmath175 $ ] . by lemma [ sol ] , this makes it possible to obtain numerical solutions @xmath204 .
obviously @xmath205 where @xmath206 is the least positive real number for which @xmath174 is invertible at any @xmath175 $ ] . as @xmath166 , we can make @xmath207 . corresponding to this
, we obtain @xmath208 where @xmath209 is the homogenized solution of ( [ peel ] ) using ( [ av][str ] ) and ( [ hv ] ) for @xmath210 $ ] .
thus the reformulation ( [ hv ] ) along with ( [ av][str ] ) also captures the local nonlinear bifurcations , if any , at a suitable @xmath206 .
the homogenization time scale @xmath55 effectively perturbs the differential variables in the jacobian matrix @xmath211 and the gradient vector @xmath212 to affect a non - zero schur complement , which in turn , produces a regularized jacobian matrix @xmath146 .
then , by lemma [ reglem ] and theorem [ bif ] , one can conclude that the homogenization is not inconsistent with the local nonlinear bifurcation . the very high differential index of the dae system ( [ peel ] ) at stick slip points accompanied by the possible non - differentiability of @xmath6 due to near jump changes in the exponentially faster time scale makes almost all dae solvers ( such as the runge - kutta solvers in matlab , the gauss - legendre runge - kutta methods and the backward differentiation formula algorithm of the ddaspkcse / software.html ] ) fail to integrate ( [ peel ] ) directly as a dae .
the homogenized approach reduces the stiffness and within an implicit solver the condition number of the jacobian matrix in the non - linear solution phase can be improved by a choice of appropriate @xmath55 .
of course , larger the @xmath55 , the more smoothed the solution of @xmath6 is .
that is , @xmath6 converges only in the weak sense with respect to @xmath55 . hence it is important for the simulation to employ a numerical method that can self regularize its jacobian matrix without losing stability so that the model given by ( [ av][str ] ) along with ( [ hv ] ) can be studied with as small a @xmath55 as possible with a view to capturing the stiff and oscillatory behavior of @xmath6 in the stick - slip regime . at the smallest @xmath55 , the numerical method jacobian approaches numerical rank @xcite deficiency , elucidating local bifurcations ,
if any , in the stick - slip regime . in this work , we use the _ @xmath0-method _
( described in detail in @xcite and not connected with the peel front angle @xmath0 ) for the time integration of ( [ av][str ] ) appended with ( [ hv ] ) , i.e. , the homogenized reformulation of ( [ peel ] ) .
we define @xmath213 the _ @xmath0-method _ numerically integrates ( [ av][str ] , [ hv ] ) over @xmath214 $ ] using a uniform mesh of @xmath215 equal time steps , each of size @xmath216 .
if the data at the @xmath156th discretization point is known , then the numerical solution at @xmath217th ( indicated by the subscript ) is found by the _ @xmath0-method _ as follows .
@xmath218 where @xmath219 is the uniform time step size , and @xmath220 is an algorithmic variable that is merely updated after every time step but does not need to be iteratively solved for in a time step within the nonlinear solver .
the initial condition @xmath221 is calculated as @xmath222 at @xmath223 .
@xmath0-method _ is an implicit @xmath182-stable integrator with the property of producing well - conditioned jacobian matrices @xcite particularly when the algorithmic parameter @xmath224 is closer to zero .
the method is second order accurate in time step size for @xmath225 when ( [ peel ] ) when @xmath226 is at least twice differentiable with respect to time .
when @xmath6 is such that @xmath226 is only lipschitz continuous then the _ @xmath0-method _ is first order accurate in time step size .
however , in the high differential index stick slip regime , @xmath6 may have poor smoothness leading to @xmath226 not being differentiable with respect to time at some points . at these points
the @xmath0 method have an error linear in step size for @xmath225 . the homogenized reformulation ( [ hv ] ) of ( [ sac ] ) averages out the oscillatory and the stiff response in @xmath6 over a stretched time scale and thus increases the smoothness of @xmath6 making @xmath226 smooth almost everywhere over the time interval of simulation . a detailed error analysis of the _ @xmath0-method _ under various continuity conditions may be found in @xcite .
we consider an example with parameter values chosen following @xcite .
the peeling dynamics example is numerically simulated by the _ @xmath0-method _ as the ode system ( [ av][str ] , [ hv ] ) .
@xmath227 is the least @xmath55 for which the @xmath0-method could be effective and @xmath6 shows stiff non - smooth changes and is oscillatory .
a time step of @xmath228 s was the best accuracy that could be obtained without failure in convergence of the newton iterations inside the _ @xmath0-method_. it may be noted that at this pull velocity the local bifurcations happen often and @xmath55 is larger than that for the other two pull velocities used in this work .
also , due to relatively ill - conditioned jacobian at stick - slip , simulation stops early for a smaller @xmath55 . as @xmath55 increases
, the time profile of @xmath6 is smoothed to its average value which is the pull velocity . ] ) response with @xmath229 m / s .
as @xmath55 increases , the time profile of @xmath0 is smoothed to its average value which is zero . at @xmath227 the slower scale compared to @xmath6 is seen . ] plotted against peeling velocity @xmath6 for pull velocity @xmath229 m / s . at
low @xmath55 the bifurcations can be seen as @xmath6 changes sharply corresponding to very little change in @xmath230 .
this also demonstrates the very fast time scale of changes in @xmath6 and its stiffer response in bifurcations compared to @xmath230 which is a function of the slower differential variables . for a larger @xmath55
as @xmath6 averages out to approach @xmath7 , the graph goes to a fixed point at which @xmath231 . ]
m / s . here
smaller time step size could be taken , increasing the accuracy of the simulation .
however , @xmath55 is of the same order as in @xmath229 as the stick - slip caused bifurcations remain significant .
the least @xmath55 has a slightly lower value as the differential variables get faster due to a higher pulling velocity @xmath7 . as @xmath55 increases
, @xmath6 is averaged out and approaches @xmath7 . for a smaller @xmath55 ,
the jacobian is more ill - conditioned and the numerical integration stops early . ]
m / s . at this pull velocity
there are fewer bifurcations in the stick - slip regime .
the time profile of @xmath6 is also smoother and less stiff . under a higher @xmath7 ,
the shear force peeling the tape is higher and the differential variables are faster , almost matching up with the time scale of @xmath6 .
hence a much smaller time step size and much smaller @xmath55 could be used . as @xmath55 increases
, @xmath6 is averaged out and approaches @xmath7 . ] as a solution of the homogenized ode for various @xmath7 s . for a higher @xmath7 the problem
could be solved with a smaller minimum @xmath55 for which the integrator jacobian does not become non - invertible .
also a higher @xmath7 could admit a smaller time step size allowing more accuracy in the solution .
this is because at a lower @xmath7 the fast scale of the stick - slip regime dominates . for @xmath229
m / s the simulation stops early as a time step encounters relatively ill - conditioned jacobian during a sticking phase . also ,
for this pull velocity the high frequency smaller amplitude oscillations happen during the sticking phase due to the high stiffness and high differential index of the model . ]
the parameters are in si units : @xmath232 and for computational purposes , the approximations @xmath233 , @xmath234 and @xmath235 hold .
the initial data for all the simulations are : @xmath236 and the adhesion model @xcite is taken as @xmath237 .
the user - selectable parameter @xmath224 in the _ @xmath0-method _ is set to zero for keeping the integrator jacobian matrix ( cf .
@xcite for theoretical details ) as well - conditioned as possible .
the values of the constant pulling velocity , @xmath7 , are chosen following the difficulty encountered in simulation of the peeling dynamics in @xcite .
since the peeling velocity @xmath6 is characterized by the local bifurcations and is the algebraic variable in the stick - slip dynamics , we study in figures [ vode1 ] , [ vode2 ] and [ vode3 ] the numerically computed time profiles of @xmath6 for various values of the pull velocity , @xmath7 .
figure [ vode1 ] corresponds to @xmath229 m / s , figure [ vode2 ] to @xmath238 m / s and figure [ vode3 ] to @xmath239 m / s . as the pulling velocity @xmath7 increases , the shear force overcoming the adhesion increases and the local bifurcations occur less frequently .
for @xmath239 m / s , bifurcations are less accentuated and the time profile of @xmath6 is smoother and less stiff as the spool rotates relatively unhindered by the adhesion . for @xmath229 and @xmath240
m / s , the stick slip regime dominates .
a smaller @xmath55 for @xmath241 m / s would mean capturing the very stiff near - jump changes in @xmath6 during the stick - slip regime .
the actual computation is limited by the smallest real number the machine can represent and the numerical method can render the peeling velocity in the stick - slip regime only to the extent the integrator jacobian remains numerically full rank ( cf .
@xcite for numerical rank of a matrix ) .
however , for @xmath239 m / s the smoother and the less stiff response of @xmath6 allows one to use a smaller @xmath55 .
thus the least @xmath55 for which the numerical integration can proceed without encountering a numerically rank deficient integrator jacobian is more at @xmath242 m / s than that at @xmath239 m / s .
this illustrates that the parameter @xmath55 is essentially the time scale to which we are able to resolve and observe the stick - slip dynamics numerically .
lesser the bifurcations , computationally it is easier to attain a finer resolution .
figure [ fvode1 ] shows the bifurcations with respect to @xmath243 for the pull velocity @xmath229 m / s which has the most dominant stick - slip regime .
the faster changes in @xmath6 compared to @xmath243 may be noted in the same figure .
for all three values of @xmath7 considered , the increase in @xmath55 leads to coarser time resolution of @xmath6 and the largest @xmath55 resolves it only to a smoothed time profile which goes to the average value @xmath231 .
figure [ avode1 ] shows the time profile of @xmath0 for @xmath229 m / s . as @xmath55 increases
, @xmath0 reaches its average value which is zero and for the least @xmath55 it shows smaller amplitude faster oscillations along the slower and smoother trajectory . at @xmath227
it can be seen that @xmath6 is stiffer and more oscillatory than @xmath0 .
figure [ vcomp ] compares the relative time profiles of @xmath6 at various values of @xmath7 and shows the difficulty in the numerical simulation when trying to capture the stick - slip regime behavior of the peeling velocity for the lower pull velocities which produce more frequent and pronounce stick - slips .
we have shown that the bifurcations in the peeling dynamics of an adhesive tape are a consequence of the rank deficiency of the jacobian and the high local differential index of the model which are structural properties of the dae model of the peeling dynamics .
the bifurcations characterize the peeling velocity and also makes changes in the peeling velocity exponentially faster than the peel front angle and/or the tensile displacement of the tape .
the homogenized ode model presented in this work captures the characteristic time scale of the stick - slip dynamics by introducing the parameter @xmath55 .
this is important since a dae can not be studied as an ode because of its inherent singularity and any ode approximation , such as the homogenized ode presented in this work , must have consistent convergence properties that smooth out the singularity inherent in the dae . at the smallest @xmath55 the homogenized ode approximation approaches the dae behavior since its jacobian approaches rank deficiency .
the numerical simulations corroborate the smoothing property of the homogenized ode approach and at smaller values of @xmath55 elucidate the local bifurcations .
the authors are grateful to prof .
g. ananthakrishna of the materials research centre of the indian institute of science , bangalore , india for sharing his insights and for introducing them to the problem .
cortet , m .- j .
dalbe , c. guerra , c. cohen , m. ciccotti , s. santucci , and l. vanel , `` intermittent stick - slip dynamics during the peeling of an adhesive tape from a roller , '' _ physical review e _ , vol .
87 , no . 2 , p. 022601
, 2013 .
cortet , m. ciccotti , and l. vanel , `` imaging the stick slip peeling of an adhesive tape under a constant load , '' _ journal of statistical mechanics : theory and experiment _ , vol .
2007 , no . 03 , p. p03005 , 2007 .
m. ciccotti , b. giorgini , and m. barquins , `` stick - slip in the peeling of an adhesive tape : evolution of theoretical model , '' _ international journal of adhesion and adhesives _ , vol .
18 , no . 1 ,
pp . 3540 , 1998 .
k. e. brenan , s. l. campbell , and l. r. petzold , _ numerical solution of initial - value problems in differential - algebraic equations_.1em plus 0.5em minus 0.4emsociety for industrial and applied mathematics , 1996 . c. g. camara , j. v. escobar , j. r. hird , and s. j. putterman , `` correlation between nanosecond x - ray flashes and stick slip friction in peeling tape , '' _ nature _ , vol .
7216 , pp . 10891092 , 2008 .
e. hairer and g. wanner , _ solving ordinary differential equations ii .
stiff and differential - algebraic problems .
_ , second revised ed.1em plus 0.5em minus 0.4emspringer series in computational mathematics , 1996 , vol . | the stick - slip dynamics is considered from the nonlinear differential - algebraic equation ( dae ) point of view and the peeling dynamics is shown to be a switching differential index dae model . in the stick - slip regime with bifurcations , the differential index can be arbitrarily high .
the time scale of the peeling velocity , the algebraic variable , in this regime is shown to be exponentially faster compared to the angular velocity of the spool and/or the stretch rate of the tape .
a homogenization scheme for the peeling velocity which is characterized by the bifurcations is discussed and is illustrated with numerical examples computed with the _ @xmath0-method_. |
the quantum walk ( qw ) as a generalization of the random walk ( rw ) is important in quantum algorithm research @xcite because of the exponentially speed up the hitting time in glued tree graphs @xcite .
qws have been demonstrated using nuclear magnetic resonance @xcite , trapped ions @xcite , single photons in bulk @xcite and fibre optics @xcite and the scattering of light in waveguide arrays @xcite .
with decoherence the spreading of the qw is transited from a ballistic to diffusive scaling @xcite and with quantumness losing the qw is transited to rw @xcite .
an open quantum system loses its quantumness when information of the quantum state leaks into the environment around it .
the unidirectional flow of information in which the decoherence and noise act consistently characterizes a markovian process @xcite .
however , there are some systems such as soft- and condensed - matter systems strongly coupling to the environment and the coupling leads to a different regime where information also flows back into the system from the surroundings , which characterizes a non - markovian process @xcite . during a markovian process , the distinguishability tends to monotonically decrease for pair of states , here in this paper , the quantum walker+coin states @xcite .
memory effects caused by the information flowing back to the system during a non - markovian process can temporarily increase it for the walker+coin states .
in this paper , we use quantum correlation between quantum walker and his coin as a distinguishability to present the behavior of a qw on a line with a non - markovian dephasing coin .
the quantum correlation between quantum walker and his coin decays quickly but is interrupted by revivals .
correspondingly , in the short time regime , the position variance of a non - markovian decoherent qw deviates from both those of the ideal qw and rw . in the long time
limit , the quantum correlation tends to be zero and the position variance shows a diffusive spread .
this paper is organized as follows : in the next section , we introduce an idea qw on a line . in sec .
iii , the non - markovian dephasing on the coin state two - level system is shown .
we study the effects from the non - markovian dephasing on coin .
compared to both idea qws and rws , non - markovian decoherent qws show different behavior in the short time regime . in sec .
iv , the quantum correlation between quantum walker and his non - markovian dephasing coin is studied via the quantum discord ( qd ) and measurement - induced disturbance ( mid ) .
it is observed that both of qd and mid show oscillating behavior in the short time regime and tend to zero in the long time limit . in sec .
v , we analyze the behavior of non - markovian decoherent qws in the long time limit and observe that because of disappearance of the quantum correlation between quantum walker and his coin , a non - markovian decoherent qw in the long time regime shows a diffusive behavior .
finally we summarize the paper briefly .
for an ideal qw on a line @xcite , the hilbert space is @xmath0 with the walker hilbert space @xmath1 spanned by the position vectors @xmath2 and @xmath3 the coin space spanned by two orthogonal vectors which we denote @xmath4 .
each step by the walker is effected by two subsequent unitary operators : the coin - flip operator @xmath5 for @xmath6 the hadamard matrix and the conditional - translation operator @xmath7 where @xmath8 and @xmath9 .
the resultant step operator is @xmath10 with @xmath11 the identity operator on @xmath1 .
the choice of initial state @xmath12 is important in studies of qws because the interference features sensitively depend on the choice of state .
this sensitivity is persistent because the dynamics are unitary hence do not die out . on the other hand the general properties of qws do not depend on the choice of initial state so the choice of initial state is not crucial provided that the focus is on such characterization .
as we are interested in general properties , the initial state is not important so we choose the initial state with the walker at the origin of a line and holding a coin in an equal superposition of the @xmath13 and @xmath14 states : @xmath15 after @xmath16 steps , the final state of the walker+coin system is @xmath17 we now follow the evolution by performing a fourier transform of the evolution operator to the `` momentum '' @xmath18 space . the eigenvectors @xmath19 of @xmath20 and @xmath21 in eq .
( [ eq : s ] ) have the eigenrelations @xmath22 for @xmath18 a continuous real quantity .
the inverse transformation is @xmath23 the walker is initialized at the origin of a line so the walker s initial state is @xmath24 in the @xmath25 basis for the walker , the evolution operator becomes @xmath26 with @xmath27 the coin state and @xmath28 the general density operator for the initial state of the system in the @xmath18 basis can be expressed as @xmath29 the final state after @xmath16 steps is @xmath30 in terms of the superoperator @xmath31 , @xmath32 the walker s position is on a line labeled @xmath33 with the initial position localized at @xmath34 .
measurement of the walker s position corresponds to the projection - valued measure @xmath35 on the walker reduced state ( after tracing out the coin ) .
the probability @xmath36 that the walker will be found at the position @xmath33 is @xmath37 which is obtained by tracing over the coin of the walker+coin state and then measuring the walker s position .
we can characterize @xmath36 by the moments of this position distribution @xmath38 .
the mean @xmath39 and variance @xmath40 can be used as the measure of qws and show the signature of qws compared to rws . instead of variance
we can also use its square root , namely dispersion @xmath41 for a rw , @xmath42 , which is characteristic of diffusive motion , whereas , for a qw , a quadratic enhancement is achieved : @xmath43 @xcite .
as a function of time @xmath16 with @xmath44 and different @xmath45 : @xmath46 in red , @xmath47 in green , @xmath48 in blue presenting non - markovian dephasing and @xmath49 in black presenting markovian dephasing.,width=340 ] we now generalize to allow for decoherence .
open quantum system loses its quantumness when information about the state leaks into its surroundings .
unidirectional flow of information characterizes a markovian process @xcite . whereas , for a non - markovian process @xcite , information also flows back into the system . recently , though it is not well understood yet , non - markovian processes with memory have become of central importance in the study of open system for both theory @xcite and experiments @xcite .
suppose that before each unitary flip of the coin , a completely positive map is performed on the coin .
we consider a pure dephasing quantum process for which the density matrix @xmath50 of the open system evolves according to the master equation @xmath51 + \frac{\gamma(t)}{2}\left[\sigma_z\rho_\text{c}(t)\sigma_z,\rho_\text{c}(t)\right ] .
\label{eq : mastereq}\ ] ] here @xmath52 represents the time - dependent energy shift and @xmath53 the time - dependent rate of the decay channel described by the pauli operator @xmath54 .
the dephasing process influences the coherence between the walker and his coin and the evolution can be described by the time - dependent function @xmath55\right\}\ ] ] which is connected to the energy shift and the decay rate of the master equation ( [ eq : mastereq ] ) by the relations @xmath56 , \gamma(t)=-\text{re}\left[\frac{\dot{\kappa}(t)}{\kappa(t)}\right].\ ] ] if the relaxation rates are positive function @xmath57 , the generator in eq .
( [ eq : mastereq ] ) is in lindblad form for each fixed @xmath58 .
such a process with @xmath59 may be called time - dependent markovian . the generalized master equation involving a certain memory kernel
can describe a non - markovian process , in which the rate @xmath53 must take on negative values for some interval of time . in the markovian regime , information of the coin state leaks into its surroundings and @xmath60 is a monotonically decreasing function of time . in the non - markovian environment , in contrast , information also flows back into the system of coin state and a revival of the distinguishability can be observed in the time evolution . with the definition of non - markovianity
, we see that an increase of @xmath60 leads to a negative rate @xmath53 in the generator of the master equation ( [ eq : mastereq ] ) . as an example
, we consider the case of a lorentzian reservoir spectral density which is on resonance with the coin qubit transition frequency and leads to an exponential two point correlation function @xmath61 , where @xmath62 describes the coupling strength between the coin qubit and environment and @xmath45 denotes the spectral width ( here we consider the coin qubit coupled to a damped jaynes - cummings model ) .
the rate @xmath63 is defined as the solution of the integrodifferential equation , @xmath64 corresponding to an initial condition @xmath65 .
we obtain @xmath66 , \label{eq : kappa}\ ] ] where @xmath67 .
we see that for weak couplings , @xmath68 , @xmath69 decreases monotonically .
the correlation function @xmath70 tends to @xmath71 with @xmath45 increasing . for @xmath72 , @xmath73 and the noise of different time
is not correlated . whereas , in the strong coupling regime @xmath74 , @xmath75 turns to be a complex number and @xmath69 can be rewritten as @xmath76\ ] ] with @xmath77 and starts to oscillate , showing non - markovian behavior @xcite .
the plots of @xmath78 are shown in fig .
[ kappa ] . in the markovian regime , for example
@xmath44 and @xmath49 , the flow of information goes only from the system into the environment , and @xmath78 decreases monotonically . in the non - markovian regime , such as @xmath44 , and @xmath79 respectively , memory effects however can temporarily increase it .
the function @xmath78 decreases faster at @xmath80 but will be interrupted by revivals which are due to memory effects on quantum correlations between the quantum walker and his coin .
the corresponding dynamical map which maps the coin state @xmath81 to the state @xmath82 at time @xmath16 @xmath83 this map can also be written by a set of operators @xmath84 on the coin degree of freedom which satisfy @xmath85 where @xmath86 considering the initial state is @xmath87 let the qw proceed for @xmath16 steps .
then the state evolves to @xmath88 note that for @xmath89 this superoperator preserves the trace .
this implies that @xmath90 for any operator @xmath91 .
steps with a perfect coin in dotted - line , with a non - markovian dephasing coin with parameters @xmath44 and @xmath46 in red and @xmath47 in green , @xmath48 in blue and @xmath49 in black .
here we choose @xmath92 as an initial coin state for calculations .
( a ) comparison of variances between the rw and qw with a non - markovian dephasing coin with parameters @xmath44 and @xmath46 .
( b ) comparison of variances between the qw with a non - markovian dephasing coin with @xmath44 and @xmath46 , @xmath47 , @xmath48 and @xmath49.,title="fig:",width=340 ] steps with a perfect coin in dotted - line , with a non - markovian dephasing coin with parameters @xmath44 and @xmath46 in red and @xmath47 in green , @xmath48 in blue and @xmath49 in black .
here we choose @xmath92 as an initial coin state for calculations .
( a ) comparison of variances between the rw and qw with a non - markovian dephasing coin with parameters @xmath44 and @xmath46 .
( b ) comparison of variances between the qw with a non - markovian dephasing coin with @xmath44 and @xmath46 , @xmath47 , @xmath48 and @xmath49.,title="fig:",width=340 ] now we consider the effect of the non - markovian dephasing on coin . as we know
, without decoherence the variance of position distribution of qw is quadratically dependent on time @xmath16 . with markovian dephasing on coin , in the long time limit the transition from a ballistic to a diffusion spreading behavior
is observed . in this paper , we will show the behavior of a qw with non - markovian dephasing coin .
the variance is used as a signature of the qw compared to rw . in the short time
regime we can calculate the variance of position distribution numerically . in fig .
[ var ] from the numerical results we can see the variance as a function of time starts to oscillate from the first several steps and the period of oscillations depends on the coefficient @xmath75 .
compared to that of ideal qw , in the short time regime ( @xmath93 ) the variance of the qw with non - markovian dephasing coin deviates from a ballistic spreading behavior .
however , in comparison with rw , the oscillating variance neither shows a diffusion behavior , which is due to effect on quantum correlations between quantum walker and his coin from the non - markovian decoherence . from the numerical calculations
the position variance of a rw distribution is the lower bound of that of qw with non - markovian dephasing coin . in fig .
[ position ] , we plot position distribution for a qw with a non - markovian dephasing coin ( @xmath44 and @xmath46 ) at the 38th , 42nd , 46th , and 50th steps . at the 38th step
, the position distribution shows rw behavior , and at the 42nd and 46th steps , there are two small peaks higher than the others nearby in the position distribution .
it shows the memory effect from non - markovian decoherence draws qw behavior back a little .
the small peaks disappear at the 50th step .
the plots of position distribution show the behavior of qw with non - markovian dephasing coin oscillates between the quantum ( ballistic ) and classical ( diffusive ) behaviors . and @xmath46 at the 38th step ( a ) , the 42nd step ( b ) , the 46th step ( c ) and the 50th step ( d).,title="fig:",width=321 ] and @xmath46 at the 38th step ( a ) , the 42nd step ( b ) , the 46th step ( c ) and the 50th step ( d).,title="fig:",width=321 ] and @xmath46 at the 38th step ( a ) , the 42nd step ( b ) , the 46th step ( c ) and the 50th step ( d).,title="fig:",width=321 ] and @xmath46 at the 38th step ( a ) , the 42nd step ( b ) , the 46th step ( c ) and the 50th step ( d).,title="fig:",width=321 ]
the signature of a decoherent discrete - time qw on a line can be demonstrated in various ways such as the position probability distribution becoming increasingly gasussian , with a concomitant fall in the standard deviation .
both of them do not recover the correlations between the quantum walker and his coin , which is known as the reason causing the quantum behavior of qws . in this section
we study the quantum correlation of a decoherent qw by two popular measures , measurement - induced disturbance ( mid ) @xcite and quantum discord ( qd ) @xcite . given the walker+coin state @xmath94 , let the density matrix for the system be diagonalized to @xmath95 for @xmath96 , where @xmath97 is a complete projection - valued measure ( i.e. using von neumann measurements ) for walker w or coin c. summing over joint projections on the state yields the diagonalized state @xmath98 the leftmost @xmath99 is an operator on the density matrix that diagonalizes it in the spectral basis corresponding to @xmath100 projective measurements .
the operator @xmath99 can also be described as a ` local measurement strategy ' .
correlations between the walker @xmath101 and coin state @xmath102 are regarded as classical if there is a unique local measurement strategy @xmath99 leaving @xmath103 unaltered from the original walker+coin state @xmath94 @xcite .
we ascertain whether the walker+coin state is ` quantum ' by determining whether a local measurement strategy exists that leaves the state unchanged .
the degree of quantumness is given by the mid @xcite @xmath104=i\left[\rho(t)\right]-i\left[\pi\rho(t)\right],\ ] ] for @xmath105=s\left[\rho_\text{w}(t)\right]+s\left[\rho_\text{c}(t)\right]-\left[\rho(t)\right]$ ] the quantum mutual information and @xmath106=-\text{tr}\left[\rho(t)\log_2\rho(t)\right]$ ] denotes von neumann entropy . by construction
@xmath103 is classical .
thus , the mid is the difference between the quantum and classical mutual information , which quantify total and classical correlations .
accordingly , eq . ( [ eq : mid ] )
is interpreted as the difference between the total and classical correlations , which are represented by the quantum mutual information and the mutual information .
qd @xcite is defined by @xmath107=i\left[\rho(t)\right]-\sup_{\pi}i\left[\pi\rho(t)\right].\ ] ] the mid has an operational definition , unlike qd , it also tends to overestimate non - classicality because of lack of optimization over local measurements .
applied to qws , we find that mid , while acting as a loose upper bound on qd , still tends to reflect well trends in the behavior of the latter .
we numerically calculate the mid and qd for a qw with perfect , markovian and non - markovian dephasing coins in the short time regime .
the quantum correlation between quantum walker and his coin presented by the mid and qd is oscillating for the first several steps ( @xmath108 ) shown in fig .
[ mid](a ) and ( b ) . without decoherence
there is no difference between mid and qd .
memory effects caused by the non - markovianity of the environment deviate the mid and qd from monotonically behavior as that during the markovian process .
both mid and qd decay steeply and interrupted by revivals .
the period of oscillating is as same as that of corresponding position variance , which proves that quantum correlation between quantum walker and his coin results in the behavior of the qw . either of the mid and
qd tends to zero with time increasing .
whereas for ideal qws , both mid and qd do not decrease with time as shown in @xcite . from the numerical results one can see the quantum correlation in the case of markovian decoherence is the lower bound of that with non - markovian decoherence , which explains the reason that the position variance of a rw distribution provides the lower bound of that of qw with non - markovian dephasing coin . in comparison of mid and qd shown in fig . [ mid](c )
, both of them can show the quantum correlation between quantum walker and his coin changing with time and mid oscillates with bigger amplitude than qd , which makes that the mid might be better signature for observing in this case .
steps with perfect ( in black ) , non - markovian dephasing coin ( in red ) and markovian dephasing coin ( in blue ) , respectively .
here we choose the parameters @xmath44 , @xmath46 and @xmath49 respectively for calculations .
( c ) the comparison of mid in red and qd in green for a qw on a line with the same parameters above.,title="fig:",width=340 ] steps with perfect ( in black ) , non - markovian dephasing coin ( in red ) and markovian dephasing coin ( in blue ) , respectively .
here we choose the parameters @xmath44 , @xmath46 and @xmath49 respectively for calculations .
( c ) the comparison of mid in red and qd in green for a qw on a line with the same parameters above.,title="fig:",width=340 ] steps with perfect ( in black ) , non - markovian dephasing coin ( in red ) and markovian dephasing coin ( in blue ) , respectively .
here we choose the parameters @xmath44 , @xmath46 and @xmath49 respectively for calculations .
( c ) the comparison of mid in red and qd in green for a qw on a line with the same parameters above.,title="fig:",width=340 ]
in the above sections we show the behavior of qw on a line with a non - markovian dephasing coin results in the quantum correlation between quantum walker and his coin , which is presented by the mid and qd . we observe that both mid and qd decrease with time and tend to zero in the long time limit .
that means the quantum correlation will disappear for long enough time . in this section
, we will show in the long time regime , the behavior of a non - markovian decoherent qw . the probability for the walker to reach a point @xmath33 at time @xmath16 is @xmath109
thus we can calculate the moments of this distribution : @xmath110 we can then invert the order of operations and do the @xmath33 sum first .
this sum can exactly be carried out in terms of derivatives of the @xmath111 function : @xmath112 inserting this result back into the expression for @xmath113 we get the first moment @xmath114 we can carry out a similar integration by parts to get the second moment : @xmath115\right\}\right\}.\ ] ] because @xmath116 is a linear transformation we can represent it as a matrix acting on the space of @xmath117 operators .
we choose the representation as @xmath118 the action of @xmath119 on @xmath91 is given by the matrix @xmath120 & \sin2k\text{re}\left[\kappa(t)\right ] & \cos2k\\ 0 & -\cos2k \text{im}\left[\kappa(t)\right ] & -\cos2k \text{re}\left[\kappa(t)\right ] & \sin2k\\ 0 & \text{re}\left[\kappa(t)\right ] & -\text{im}\left[\kappa(t)\right ] & 0 \end{pmatrix}\begin{pmatrix } r_1 \\r_2\\ r_3 \\r_4 \end{pmatrix}\end{aligned}\ ] ] because @xmath121 .
the only nontrivial dynamics results from the @xmath122 submatrix @xmath123 & \sin2k\text{re}\left[\kappa(t)\right ] & \cos2k\\ -\cos2k \text{im}\left[\kappa(t)\right ] & -\cos2k \text{re}\left[\kappa(t)\right ] & \sin2k\\ \text{re}\left[\kappa(t)\right ] & -\text{im}\left[\kappa(t)\right]&0 \end{pmatrix}\end{aligned}\ ] ] we also need to know the effects of left and right multiplying by @xmath54 .
these are given by the two matrices @xmath124 let us take these expressions and apply them to eq .
( [ eq : first ] ) for the first moment . in the integrand
, the initial density matrix for the coin is multiplied @xmath125 times by @xmath116 , then left multiplied by @xmath54 and finally the trace is taken .
given the above expression for @xmath126 , we see that this is the same as multiplying the @xmath127-vector @xmath128 @xmath125 times by @xmath129 and then keeping only the @xmath130 component of the result .
this gives us the new expression for the first moment @xmath131\begin{pmatrix } r_2\\r_3\\r_4 \end{pmatrix}.\end{aligned}\ ] ] the eignevalues of @xmath129 satisfy @xmath132 and then in the long time limit , therefore @xmath133 and the first moment becomes approximately @xmath134\begin{pmatrix } r_2\\r_3\\r_4 \end{pmatrix}\\&=\frac{r_2\text{re}\left[\kappa(t)\right]-r_3\text{im}\left[\kappa(t)\right]+r_4|\kappa(t)|^2}{-1+|\kappa(t)|^2}.\end{aligned}\ ] ] in the long time limit , the second moment can also be calculated as @xmath135\begin{pmatrix } 1\\r_2\\r_3\\r_4 \end{pmatrix}\nonumber\\ & = t-\sum_{j=1}^t\sum_{j'=1}^{j-1}\frac{1}{2\pi}\int_{-\pi}^{\pi}\text{d}k\begin{pmatrix } 0&0&1 \end{pmatrix}m_k^{j - j'}\begin{pmatrix } 0\\0\\2 \end{pmatrix}\nonumber\\ & = t-\frac{1}{2\pi}\int_{-\pi}^{\pi}\text{d}k\begin{pmatrix } 0&0&1 \end{pmatrix}\left[t-\left(1-m_k\right)^{-1}m_k+\left(1-m_k\right)^{-1}m_k^t\right]\left(1-m_k\right)^{-1}m_k\begin{pmatrix } 0\\0\\2 \end{pmatrix}\nonumber\\ & = t\left(1+\frac{2|\kappa(t)|^2}{-1+|\kappa(t)|^2}\right)+\frac{2|\kappa(t)|^4 + 5|\kappa(t)|^2}{\left(-1+|\kappa(t)|^2\right)^2},\end{aligned}\ ] ] which is independent on the initial coin state .
the position variance is obtained as @xmath136-r_3\text{im}\left[\kappa(t)\right]+r_4|\kappa(t)|^2}{-1+|\kappa(t)|^2}.\end{aligned}\ ] ] in the long time regime , @xmath78 ( [ eq : kappa ] ) tends to zero . then the position variance is linear dependent on time @xmath16 .
thus in the case of the non - markovian dephasing coin the variance grows linearly with time and a qw with a non - markovian dephasing coin shows a diffusive spread in the long time limit . and @xmath46 for calculations.,width=340 ] in the long time limit , we can obtain the analytical results of the position variance of a non - markovian decoherent qw . in fig .
[ ana ] , we can see that for the first several steps , the numerical and analytical results agree with each other .
whereas for long time , the variance becomes linear dependent on time @xmath16 showing a diffusive spread , just as in the classical case , though the rate of growth will greater than for the rw .
in this paper , we presented how non - markovian decoherence can influence the evolution of a quantum walker on a line and showed a different behavior compared to the ballistic and diffusive spreads .
we analyze the dynamics and quantum correlation of quantum walker and his coin .
we use the mid and qd as measures for quantum correlations .
the qd would be a valuable measure to use and the mid also suffices to show that pure quantum correlation induced by having non - markovian dephasing on the coin state .
the memory effect caused by the information flowing back to the system during a non - markovian process decreases the quantum correlation quickly which is later interrupted by revivals . in the long time limit , the quantum correlation disappears with @xmath78 tending to zero , which results in a transition from a qw with a ballistic scaling to a rw with a diffusive scaling .
we draw the conclusions as followings : in the short time the behavior of qw with non - markovian dephasing coin oscillates between a ballistic and diffusive spreading behavior , and the quantum correlation between the walker and coin collapses and revivals due to the memory effect . whereas in the long time limit
, qw with non - markovian decoherence does not make much difference from that with markovian decoherence .
thus our analysis is quite valuable in that we characterize this qw on a line with non - markovian dephasing coin carefully and devise appropriate , meaningful qd and mid as measures to study quantum correlation dynamics for this system .
we would like to thank barry c. sanders for useful conversations .
this work has been supported by the national natural science foundation of china under grant nos 10974192 , 11004029 and 11174052 , the natural science foundation of jiangsu province under grant no bk2010422 , the ph.d .
program of the ministry of education of china , the excellent young teachers program of southeast university and the national basic research development program of china ( 973 program ) under grant no 2011cb921203 .
d. aharonov , a. ambainis , j. kempe and u. vazirani , _ in proceedings of the 33rd acm symposium on the theory of computing _
( acm , washington , dc , 2001 ) , p. 50 . v. kendon , math . struct . in comp .
17 * , 1 ( 2007 ) .
a. m. childs , r. cleve , e. deotto , e. farhi , s. gutmann and d. a. spielman , _ in proceedings of the 35th acm symposium on the theory of computing _
( acm , washington , dc , 2003 ) , p. 59 . c. a. ryan , m. laforest , j. c. boilequ and r. laflamme , phys
a * 72 * , 062317 ( 2005 ) ; j. du , h. li , x. xu , m. shi , j. wu , x. zhou and r. han , phys .
rev . a * 67 * , 042316 ( 2003 ) .
f. zhringer , g. kirchmair , r. gerritsma , e. solano , r. blatt and c. f. roos , phys .
* 104 * , 100503 ( 2010 ) . b. do , m. stohler , s. balasubramanian and d. elliott , journal of optical society of america ( b ) * 22 * , 499 ( 2005 ) .
a. schreiber , k. n. cassemiro , v. potoek , a. gbris , p. j. mosley , e. andersson , i. jex and c. silberhorn , phys .
lett . * 104 * , 050502 ( 2010 ) .
m. a. broome et al . , phys .
lett . * 104 * , 153602 ( 2010 ) .
a. schreiber et al . , phys .
lett . * 106 * , 180403 ( 2011 ) . h. b. perets , y. lahini , f. pozzi , m. sorel , r. morandotti and y. silberberg , phys .
lett . * 100 * , 170506 ( 2008 ) .
t. brun , h. a. carteret and a. ambainis , phys .
lett . * 91 * , 130602 ( 2003 ) .
t. brun , h. a. carteret and a. ambainis , phys .
a * 67 * , 052317 ( 2003 ) .
p. xue , b. c. sanders , a. blais and k. lalumire , phys .
rev . a * 78
* , 042334 ( 2008 ) .
p. xue , b. c. sanders and d. leibfried , phys .
rev . lett . * 103 * , 183602 ( 2009 ) .
h. j. carmichael , _ an open systems approach to quantum optics _
, lecture notes in physics , vol . m18 ( springer - verlag , berlin , 1993 ) .
u. weiss , _ quantum dissipative systems _ ( world scientific , singapore , 2001 )
. j. kempe , contemp . phys . * 44 * , 307 ( 2003 )
. n. shenvi , j. kempe and k. b. whaley , phys .
a * 67 * , 052307 ( 2003 ) .
m. m. wolf , j. eisert , t. s. cubitt , j. i. cirac , phys .
* 101 * , 150402 ( 2008 ) .
a. rivas , s. f. huelga and m. b. plenio , phys .
lett . * 105 * , 050403 ( 2010 ) .
e. laine , j. piilo and h. breuer , phys .
a * 81 * , 062115 ( 2010 ) .
h. breuer , e. laine and j. piilo , phys .
lett . * 103 * , 210401 ( 2009 ) . j. m. raimond , m. brune and s. haroche , , 1964 ( 1997 ) .
t. meunier , s. gleyzes , p. maioli , a. auffeves , g. nogues , m. brune , j. m. raimond and s. haroche , , 010401 ( 2005 ) .
j. s. xu , c. f. li , m. gong , x. b. zou , c. h. shi , g. chen and g. c. guo , phys .
lett . * 104 * , 100502 ( 2010 ) .
j. s. xu , c. f. li , c. j. zhang , x. y. xu , y. s. zhang and g. c. guo , , 042328 ( 2010 ) . j. s. tang , c. f. li , y. l. li , x. b. zou , g. c. guo , h. breuer , e. laine and j. piilo , europhysics letters * 97 * , 10002 ( 2012 ) . b. h. liu , l. li , y. f. huang , c. f. li , g. c. guo , e. laine , h. breuer and j. piilo , nat . phys .
* 7 * , 931 ( 2011 ) .
s. luo , phys .
a * 77 * , 022301 ( 2008 ) .
n. li and s. luo , , 032327 ( 2007 ) .
s. luo , , 042303 ( 2008 ) .
h. ollivier and w. h. zurek , phys .
lett . * 88 * , 017901 ( 2001 ) .
l. henderson and v. vedral , j. phys .
a * 34 * , 6899 ( 2001 ) .
r. srikanth , s. banerjee and c. m. chandrashekar , , 062123 ( 2010 ) .
b. r. rao , r. srikanth , c. m. chandrashekar and s. banerjee , phys .
a * 83 * , 064302 ( 2011 ) . | quantum walk acts obviously different from its classical counterpart , but decoherence will lessen and close the gap between them . to understand this process
, it is necessary to investigate the evolution of quantum walk under different situation of decoherence . in this article
, we study a non - markovian decoherent quantum walk on a line . in the short time regime ,
the behavior of the walk deviates from both idea quantum walks and classical random walks .
the position variance as a measure of quantum walk starts oscillating from the first several steps and tends to be linear on time and showing a diffusive spread in the long time limit , which is caused by the non - markovian dephasing affecting on quantum correlations between quantum walker and his coin .
we also study both quantum discord and measurement - induced disturbance as measures of quantum correlations and observe that both of them oscillate in the short time regime and tend to be zero in the long time limit .
therefore quantum walk with non - markovian decoherence tends to diffusive spreading behavior in the long time limit , while in the short time regime it oscillates between a ballistic and diffusive spreading behavior , and the quantum correlation collapses and revivals due to the memory effect . |
[ hatano - sec10 ] first of all , we discuss as to why we have to treat the exponential operator and why we need an approximant in order to treat the exponential operator .
the exponential operator appears in various fields of physics as a formal solution of the differential equation of the form @xmath11 where @xmath12 is a function or a vector and @xmath13 is an operator or a matrix .
typical examples are the schrdinger equation @xmath14 ( we put @xmath15 here and hereafter ) , the hamilton equation @xmath16 ( see eq .
( [ hatano - eq160 ] ) below ) and the diffusion equation with a potential @xmath17 a solution of eq .
( [ hatano - eq100 ] ) is given in the form of the green s function as @xmath18 although it is only a formal solution ; obtaining the green s function @xmath19 is just as difficult as solving the equation ( [ hatano - eq100 ] ) in any other way .
another important incident of the exponential operator is the partition function in equilibrium quantum statistical physics : @xmath20 where @xmath21 is a quantum hamiltonian . the exponential operator , however , is hard to compute in many interesting cases . the most straightforward way of computing the exponential operator @xmath22 is to diagonalize the operator @xmath13 . in quantum many - body problems , however , the basis of the diagonalized representation is often nontrivial , because we are typically interested in the hamiltonian with two terms or more that are mutually non - commutative ; for example , the ising model in a transverse field , @xmath23 and the hubbard model , @xmath24 in the first example ( [ hatano - eq130 ] ) , the quantization axis of the first term is the spin @xmath25 axis , while that of the second term is the spin @xmath1 axis .
the two terms are therefore mutually non - commutative . in the second example ( [ hatano - eq140 ] ) , the first term is diagonalizable in the momentum space , whereas the second term is diagonalizable in the coordinate space . in both examples ,
each term is easily diagonalizable .
since one quantization axis is different from the other , the diagonalization of the sum of the terms becomes suddenly difficult .
the same situation arises in chaotic hamilton dynamics .
consider a classical hamiltonian @xmath26 where @xmath27 is the kinetic term and @xmath28 is the potential term .
the hamilton equation is expressed in the form @xmath29 where the operators @xmath30 and @xmath31 are symbolic ones standing for the operations @xmath32 although each operation of @xmath30 and @xmath31 is simple enough , the
hamiltonian " operator @xmath33 is not easily tractable .
this is because the kinetic part and the potential part , @xmath34 do not commute with each other ; see an example in sect .
[ hatano - subsec20 ] below . to summarize this section
, we frequently encounter the situation where the exponential operator of each term , @xmath35 and @xmath36 , is easily obtained and yet the desired exponential operator @xmath37 is hard to come .
this is the situation where the trotter decomposition ( [ hatano - eq10 ] ) becomes useful .
[ hatano - sec20 ] we discussed in the previous section the importance of the exponential operator and the necessity of a way of treating it .
we here discuss a remarkable advantage of the trotter approximant to the exponential operator .
let us first confirm that the trotter approximant ( [ hatano - eq10 ] ) is indeed a first - order approximant . by expanding the both sides of eq .
( [ hatano - eq10 ] ) , we have @xmath38 where @xmath39 is the identity operator .
the difference between the two comes from the fact that in the approximant ( [ hatano - eq210 ] ) , the operator @xmath2 always comes on the left of the operator @xmath3 .
hence we obtain @xmath40}+{\mathop{\mathrm{o}}}(x^3)}.\ ] ] in the actual application of the approximant , we divide the parameter @xmath1 into @xmath41 slices in the form @xmath42}+{\mathop{\mathrm{o}}}\left(\left(\frac{x}{n}\right)^3\right ) } \right]^n = { \mathrm{e}}^{x(a+b)+\frac{1}{2}\frac{x^2}{n}{\left [ a , b \right]}+{\mathop{\mathrm{o}}}\left(\frac{x^3}{n^2}\right)}.\ ] ] thus the correction term vanishes in the limit @xmath43 .
we refer to the integer @xmath41 as the trotter number .
now we discuss as to why we should be interested in generalizing the trotter approximation .
the trotter approximant ( [ hatano - eq10 ] ) and the generalized one ( [ hatano - eq30 ] ) , in fact , have a remarkable advantage over other approximants such as the frequently used one @xmath44 the approximant of the form ( [ hatano - eq30 ] ) conserves an important symmetry of the system in problems of quantum dynamics and hamilton dynamics . in problems of quantum dynamics , the exponential operator , or the green s function
@xmath45 is a unitary operator ; hence the norm of the wave function does not change , which corresponds to the charge conservation .
we here emphasize that the exponential product @xmath46 is also a unitary operator .
the perturbational approximant ( [ hatano - eq240 ] ) , on the other hand , does not conserve the norm of the wave function ; in fact , the norm typically increases monotonically as the time passes as we demonstrate in sect .
[ hatano - subsec19 ] below . in problems of hamilton dynamics ,
the time evolution of the hamilton system conserves the volume in the phase space @xmath47 , which is called the symplecticity in mathematics .
the exponential product formula , in general , also has the symplecticity .
the time evolution of the hamilton equation ( [ hatano - eq160 ] ) is described by the exponential operator @xmath48 where @xmath21 is the hamiltonian " operator ( [ hatano - eq180 ] ) .
the trotter decomposition approximates the time evolution with the operator @xmath49 with @xmath50 and @xmath51 given by eq .
( [ hatano - eq190 ] ) .
the operator @xmath52 describes the time evolution over the time slice @xmath53 of a hamilton system with only the kinetic energy @xmath54 .
it thereby conserves the phase - space volume , so does the operator @xmath55 .
the whole trotter approximant therefore conserves the phase - space volume .
this holds for any exponential product formula in the form ( [ hatano - eq30 ] ) as well .
hence the exponential product formula , when used in the hamilton dynamics , is sometimes called a symplectic integrator . in equilibrium quantum statistical physics
, the operator @xmath56 does not have a particular symmetry except the symmetries of the hamiltonian itself .
the above advantage of the exponential product formula is hence lost when applied to numerical calculations of the partition function @xmath57 .
in fact , in applying the higher - order decomposition ( [ hatano - eq30 ] ) to the world - line quantum monte carlo simulation , some of the parameters @xmath8 are negative , which causes the negative - sign problem in systems that usually do not have the negative - sign problem @xcite .
the negative - sign problem is the problem that the boltzmann weight of the system to be simulated becomes negative for some configurations .
thanks to a recent development of the world - line quantum monte carlo simulation @xcite , the higher - order decomposition is not necessary anymore in some cases ; the simulation is carried out in the limit @xmath43 from the very beginning and hence the order of the correction term does not matter in such cases .
see appendix [ hatano - app10 ] for a brief review over the recent development .
the fact that the exponential product formula keeps the symmetry of the system is one of its remarkable advantages . in the present and next subsections ,
we demonstrate that this indeed affects numerical accuracy strongly . in the present subsection
, we use a simple example of quantum dynamics , namely the spin precession . consider the simple hamiltonian @xmath58 if we start the dynamics from the up - spin state @xmath59 the spin precesses around the axis of the magnetic field @xmath60 with the period @xmath61 although it is easy to compute the dynamics exactly , we here use the trotter approximant @xmath62 and the perturbational approximant @xmath63
the exact dynamics should conserve the energy expectation @xmath64 .
[ hatano - fig05 ] shows the energy deviation due to the approximations . ) and ( b ) the perturbational approximant ( [ hatano - eq5040 ] ) . in both calculations
, we put @xmath65 and @xmath66 .
the initial state is the one in eq .
( [ hatano - eq5010 ] ) with the energy expectation @xmath67.,title="fig:",scaledwidth=35.0% ] ) and ( b ) the perturbational approximant ( [ hatano - eq5040 ] ) . in both calculations
, we put @xmath65 and @xmath66 .
the initial state is the one in eq .
( [ hatano - eq5010 ] ) with the energy expectation @xmath67.,title="fig:",scaledwidth=35.0% ] the error in the energy of the trotter approximation ( [ hatano - eq5030 ] ) oscillates periodically and never increases beyond the oscillation amplitude . the period of the oscillation in fig .
[ hatano - fig05](a ) is equal to that of the spin precession .
we can understand this as follows : when the spin comes back to the original position after one cycle of the precession , it comes back accurately to the initial state ( [ hatano - eq5010 ] ) because of the unitarity of the trotter approximation , and hence the oscillation .
in contrast , the error in the energy monotonically grows in the case of the perturbational approximant as is shown in fig .
[ hatano - fig05](b ) .
this is because the norm of the wave vector increases by the factor @xmath68 the remarkable difference between fig .
[ hatano - fig05](a ) and fig .
[ hatano - fig05](b ) thus comes from the fact that the trotter approximant is unitary .
we next demonstrate the trotter decomposition ( [ hatano - eq260 ] ) in an interesting example of chaotic dynamics .
we again emphasize that keeping the symplecticity of the hamilton dynamics has an important effect on numerical accuracy .
let us first notice that the operators in eq .
( [ hatano - eq190 ] ) satisfy @xmath69 we therefore have @xmath70 note that applying the two operators in the order @xmath71 is different from applying them in the order @xmath72 ; in the former , the update of @xmath73 in the application of @xmath74 is done under the updated @xmath75 , whereas in the latter , it is done under @xmath75 before the update .
umeno and suzuki @xcite demonstrated the use of symplectic integrators for chaotic dynamics of the system @xmath76 the equipotential contour is given by @xmath77constant ; hence the system is confined in the area surrounded by four hyperbolas as exemplified in fig .
[ hatano - fig10](a ) . ) .
the initial condition is @xmath78 , @xmath79 and @xmath80 with the energy @xmath81 .
the time slice is @xmath66 .
( a ) the movement of the system in the coordinate space @xmath82 for @xmath83 .
the broken curves indicate the hyperbolas @xmath84 .
( b ) the energy fluctuation due to the trotter approximation ( [ hatano - eq260 ] ) .
we plotted a dot every 1,000 steps .
( c ) the energy increase due to the approximant ( [ hatano - eq295 ] ) . ] ) .
the initial condition is @xmath78 , @xmath79 and @xmath80 with the energy @xmath81 .
the time slice is @xmath66 .
( a ) the movement of the system in the coordinate space @xmath82 for @xmath83 .
the broken curves indicate the hyperbolas @xmath84 .
( b ) the energy fluctuation due to the trotter approximation ( [ hatano - eq260 ] ) .
we plotted a dot every 1,000 steps .
( c ) the energy increase due to the approximant ( [ hatano - eq295 ] ) . ] ) .
the initial condition is @xmath78 , @xmath79 and @xmath80 with the energy @xmath81 .
the time slice is @xmath66 .
( a ) the movement of the system in the coordinate space @xmath82 for @xmath83 .
the broken curves indicate the hyperbolas @xmath84 .
( b ) the energy fluctuation due to the trotter approximation ( [ hatano - eq260 ] ) .
we plotted a dot every 1,000 steps .
( c ) the energy increase due to the approximant ( [ hatano - eq295 ] ) . ]
the exact dynamics should conserve the energy .
the trotter approximation of the dynamics , eq .
( [ hatano - eq260 ] ) , gives the energy fluctuation shown in fig .
[ hatano - fig10](b ) . the energy , though deviates from the correct value sometimes , comes back after the deviation .
in fact , the deviation occurs when the system goes into one of the four narrow valleys of the potential ; it is suppressed again and again when the system comes back to the central area .
this is in striking contrast to the update due to the perturbational approximant @xmath85 which yields the monotonic energy increase shown in fig .
[ hatano - fig10](c ) .
the reason of the difference between the approximants , though less apparent than in the case of the previous subsection , must be keeping the symplecticity , or the conservation of the phase - space volume .
we emphasized in the previous section the importance of the exponential product formula . in the present section ,
we describe a way of constructing higher - order exponential product formulas _ recursively _
@xcite .
the easiest improvement of the trotter formula ( [ hatano - eq20 ] ) is the symmetrization : @xmath86 the symmetrized approximant has the property @xmath87 because of which the even - order terms vanish in the exponent of the right - hand side of eq .
( [ hatano - eq300 ] ) . we can thereby promote the approximant ( [ hatano - eq300 ] ) to a second - order approximant .
now we introduce a way of constructing a symmetrized fourth - order approximant from the symmetrized second - order approximant ( [ hatano - eq300 ] ) .
consider a product @xmath88 where @xmath89 is an arbitrary real number for the moment .
the expression ( [ hatano - eq300 ] ) is followed by @xmath90r_3+{\mathop{\mathrm{o}}}(x^5)}.\end{aligned}\ ] ] ( the readers should convince themselves by the taylor expansion that the third - order correction in the exponent of the last line is just the sum of the third - order corrections in the exponents of the second line .
this is not true for higher - order corrections . )
note that we arranged the parameters in the form @xmath91 in eq .
( [ hatano - eq305 ] ) so that ( i ) the first - order term in the exponent of the last line of eq . ( [ hatano - eq320 ] ) should become @xmath92 and ( ii ) the whole product @xmath93 should be symmetrized , or should satisfy @xmath94 . because of the second property , the even - order corrections vanish in the exponent of the last line of eq .
( [ hatano - eq320 ] ) . making the parameter @xmath89 a solution of the equation @xmath95{2 } } = 1.351207191959657\cdots,\ ] ]
we promote the product ( [ hatano - eq305 ] ) to a fourth - order approximant @xcite .
following the same line of thought , we come up with another fourth - order approximant @xcite in the form @xmath96 where the parameter @xmath97 is a solution of the equation @xmath98{4 } } = 0.414490771794375\cdots.\ ] ] we can compare the fourth - order approximants ( [ hatano - eq305 ] ) and ( [ hatano - eq340 ] ) using the following diagram .
suppose that the exponential operator @xmath37 is a time - evolution operator from the time @xmath99 to the time @xmath100 . in the product ( [ hatano - eq305 ] )
, the term @xmath101 on the right approximates the time evolution from @xmath99 to @xmath102 , the term @xmath103 in the middle approximates the time evolution from @xmath104 to @xmath105 , and the term @xmath101 on the left approximates the time evolution from @xmath106 to @xmath107 . let us express this time evolution as in fig .
[ hatano - fig20](a ) . ) and ( b ) the eighth - order approximant ( [ hatano - eq380 ] ) . ] ) and ( b ) the eighth - order approximant ( [ hatano - eq380 ] ) . ]
the product ( [ hatano - eq340 ] ) is similarly represented as in fig .
[ hatano - fig20](b ) .
as is evident , the first product ( [ hatano - eq305 ] ) has a part that goes into the past , " or @xmath108
. this can be problematic in some situations ; in the diffusion from a delta - peak distribution , for example , there exists no
past " of the initial delta peak .
the second product ( [ hatano - eq340 ] ) does not have the problem and hence is recommended for general use .
once we know how to construct the fourth - order approximant from the second - order approximant , the rest is quite straightforward @xcite . following the construction ( [ hatano - eq340 ] ) , we construct the sixth - order approximant in the form @xmath109 with @xmath110{4}}=0.373065827733272\cdots,\ ] ] and further construct the eighth - order approximant in the form @xmath111 with @xmath112{4}}=0.359584649349992\cdots.\ ] ] these approximants are represented by the diagrams in fig .
[ hatano - fig30 ] .
we can continue this recursive procedure , ending up with the exact time evolution , where the diagram ultimately becomes a fractal object .
this is why the series of the approximants is called the fractal decomposition .
it is an interesting thought that the back - and - forth time evolution in a fractal way reproduces the exact time evolution .
before going into another way of constructing higher - order exponential product formulas , let us introduce , as an interlude , an important application of the exponential product formula .
we show how to approximate the time - ordered exponential @xcite .
we have considered until now only the case where the operators @xmath2 and @xmath3 do not depend on @xmath1 , or in other words , only the time evolution of a time - independent hamiltonian .
the fractal decomposition introduced in the previous section needs modification when applied to problems such as the quantum dynamics of a time - dependent hamiltonian ; in quantum annealing @xcite , for example , the transverse field @xmath113 in the hamiltonian ( [ hatano - eq130 ] ) is changed in time .
the time - evolution operator of the quantum hamiltonian @xmath114 is not simply @xmath115 but a time - ordered exponential in the form @xmath116.\ ] ] it is quite well - known that @xmath117 is an approximant of the first order of @xmath118 and @xmath119 is an approximant of the second order .
how do we construct higher - order approximants ?
we here show that a slight modification of the fractal decomposition gives the answer .
the key is to introduce a shift - time operator @xcite defined in @xmath120 note that the operator acts on the function on the left .
the shift - time operator is expressed in the form @xmath121 in the case where @xmath122 is an analytic function , but the definition ( [ hatano - eq430 ] ) does not limit its use to the analytic case . if we have two shift - time operators , the result is @xmath123 with the use of the shift - time operator , the time - ordered exponential ( [ hatano - eq410 ] ) is transformed @xcite as @xmath124 = { \mathrm{e}}^{-{\mathrm{i}}\vardelta t\left({\cal h}(t)+{\cal t}\right)}.\ ] ] we can prove this by using the trotter approximation as follows : @xmath125.\end{aligned}\ ] ] decomposing the hamiltonian into two parts as in eq .
( [ hatano - eq400 ] ) , we have now three parts in the exponent of the time - evolution operator as in @xmath126 = { \mathrm{e}}^{-{\mathrm{i}}\vardelta t\left(a(t)+b(t)+{\cal t}\right)}.\ ] ] we then approximate the exponential in the right - hand side of eq .
( [ hatano - eq460 ] ) .
the first - order approximant is given by @xmath127 and the second - order approximant is given by @xmath128 higher - order approximants are given by the fractal decomposition of the three parts , @xmath2 , @xmath3 , and @xmath129 .
the fractal decomposition of three parts is easily obtained by substituting @xmath130 for eq .
( [ hatano - eq300 ] ) .
the fourth - order approximant is thereby obtained @xcite as @xmath131 with the coefficient @xmath97 given by eq .
( [ hatano - eq350 ] ) .
[ hatano - sec50 ] in the last section before the summary , we discuss the calculus of the correction terms . in the fractal decomposition , we construct higher - order approximants recursively .
is it possible to construct higher - order approximants _ directly _ , not recursively ?
in fact , ruth @xcite found ( not systematically ) a third - order formula @xmath132 which would not be found within the framework of the fractal decomposition .
for the purpose of finding higher - order formulas directly , we need to compute the correction terms in the exponent as @xmath133 this is one of the aims of the quantum analysis developed by one of the present authors ( m.s . )
then we can put the correction terms to zero up to a desired order and solve the set of non - linear simultaneous equations @xmath134 thereby obtaining the parameters @xmath135 .
the main feature of the quantum analysis is to introduce operator differential . in order to motivate the readers ,
suppose that we can write down an identity @xmath136 where @xmath137 is an operator functional .
the derivative with respect to @xmath1 on the right - hand side is well - defined ; for example , @xmath138 for @xmath139 .
now , is it possible to define the differentiation @xmath140 ?
let us discuss as to what should be the definition of the operator differential in order for the identity ( [ hatano - eq530 ] ) to hold .
the definition of the @xmath1 derivative is expressed as @xmath141 the left - hand side of the identity ( [ hatano - eq530 ] ) is given by the definition of the derivative as @xmath142 the identity ( [ hatano - eq530 ] ) suggests that the operator differential @xmath140 must be a hyperoperator that maps the operator @xmath143 to the operator given by eq .
( [ hatano - eq550 ] ) .
thus we arrive at the definition of the operator differential within the framework of the quantum analysis @xcite : if we can express the operator given by @xmath144 in terms of a hyperoperator mapping from an arbitrary operator @xmath145 as in @xmath146 , then we refer to the hyperoperator as an operator differential @xmath140 and denote it in the form @xmath147 we stress here that the operator differential @xmath140 must be expressed in terms of @xmath2 and the commutation relation of @xmath2 , or the inner derivation " @xmath148},\ ] ] but _ not _ in terms of the arbitrary operator @xmath145 . the convergence of eq .
( [ hatano - eq560 ] ) is in the sense of the norm convergence which is uniform with respect to the arbitrary operator @xmath145 .
let us consider the application of the above in a simple example @xmath149 .
the definition ( [ hatano - eq560 ] ) is followed by @xmath150 thus we have @xcite @xmath151 if @xmath152 , we use the result ( [ hatano - eq600 ] ) for eq .
( [ hatano - eq530 ] ) and have @xmath153 } \nonumber\\ & = & 2xb^2 + 4x^2bc+2x^2cb+4x^3c^2 -2x^2(bc - cb)-x^2(cb - bc ) \nonumber\\ & = & 2xb^2 + 3x^2bc+3x^2cb+4x^3c^2,\end{aligned}\ ] ] which is indeed identical to the result of straightforward algebra .
we can not see in this simple example any merit of the use of the quantum analysis .
the readers should wait for more complicated examples given later in sec .
[ hatano - subsec530 ] , where we show that the differential of exponential operators is given in terms of the inner derivation .
the lie algebra is defined by commutation relations , or the inner derivation ; it is hence essential to obtain results in terms of the inner derivation , not in terms of naive expansions such as the right - hand side of eq .
( [ hatano - eq610 ] ) .
we here provide some of the important formulas of the inner derivation ( [ hatano - eq580 ] ) as preparation for the next subsection , where we give the differential of exponential operators .
first , we have linearity : for any c - numbers @xmath154 and @xmath155 , the inner derivation of the operators @xmath2 and @xmath3 satisfies @xmath156 } = a{\left [ a , \quad \right]}+b{\left [ b , \quad \right ] } = a\delta_a+b\delta_b.\ ] ] any powers of the operator @xmath2 are commutable with the inner derivation of any powers of the same operator : @xmath157}=0,\ ] ] because @xmath158}={\left [ a^n , a^mb \right]}=\delta_{a^n}a^mb\ ] ] for an arbitrary operator @xmath3 and any integers @xmath10 and @xmath41 .
we can generalize the identity ( [ hatano - eq630 ] ) to the case of any analytic functions of the operator @xmath2 : @xmath159}=0,\ ] ] where @xmath137 and @xmath160 are defined by the taylor expansion as @xmath161 next , we prove the identity @xcite @xmath162 the proof is as follows : for an arbitrary operator @xmath3 , we have @xmath163}+g(a){\left [ f(a ) , b
\right]}-{\left [ g(a ) , { \left [ f(a ) , b \right ] } \right ] } \nonumber\\ & = & f(a){\left [ g(a ) , b \right]}+{\left [ f(a ) , b \right]}g(a ) = { \left [ f(a)g(a ) , b \right ] } \nonumber\\ & = & \delta_{f(a)g(a)}b.\end{aligned}\ ] ] note that we can rewrite the identity ( [ hatano - eq660 ] ) as @xmath164 because of the identity ( [ hatano - eq650 ] ) . in the special case @xmath165
, we have @xmath166 with the repeated use of the identity ( [ hatano - eq680 ] ) , we then prove the identity @xcite @xmath167 for any integer @xmath41 .
this is proved by means of mathematical induction .
the identity ( [ hatano - eq690 ] ) indeed holds for @xmath168 .
assume now that @xmath169 then the identity ( [ hatano - eq680 ] ) yields @xmath170\left(a-\delta_a\right)+a^{n-1}\delta_a \nonumber\\ & = & a^n-\left(a-\delta_a\right)^n.\end{aligned}\ ] ] an interesting and quite well - known identity is @xmath171 we can prove this by differentiating the left - hand side by @xmath1 .
first , note that @xmath172}{\mathrm{e}}^{-xa}.\ ] ] we thereby have the following in each order of @xmath1 : @xmath173}{\mathrm{e}}^{-xa } \right|_{x=0 } = \delta_a b , \\ \label{hatano - eq735 } \left .
\frac{{\mathrm{d}}^2}{{\mathrm{d}}x^2}{\mathrm{e}}^{xa}b{\mathrm{e}}^{-xa } \right|_{x=0 } & = & \left .
{ \mathrm{e}}^{xa}{\left [ a , { \left [ a , b \right ] } \right]}{\mathrm{e}}^{-xa } \right|_{x=0 } = { \delta_a}^2 b , \\
\nonumber \\ & \cdots , & \nonumber\end{aligned}\ ] ] which proves the identity ( [ hatano - eq720 ] ) . as a corollary ,
we obtain the following identity : @xmath174 the proof is straightforward ; for an arbitrary operator @xmath175 , we have @xmath176 we are now in a position to discuss the differential of exponential operators . we begin with the differential of the power of an operator , @xmath177 , a generalization of the identity ( [ hatano - eq600 ] ) .
the result is @xcite @xmath178 an important comment is in order .
the identity ( [ hatano - eq770 ] ) does not claim that the inverse of @xmath179 is well - defined .
in fact , the inner derivation @xmath179 in the denominator is canceled when we expand the numerator of the second expression .
the denominator is well - defined only in such cases .
we use the identity ( [ hatano - eq690 ] ) in the derivation of the identity ( [ hatano - eq770 ] ) . the definition ( [ hatano - eq560 ] )
is followed by @xmath180 \right\ } { \mathrm{d}}a \nonumber\\ & = & \sum_{j=1}^na^{j-1}\left(a-\delta_a\right)^{n - j } { \mathrm{d}}a = \frac{a^n-\left(a-\delta_a\right)^n}{a-\left(a-\delta_a\right)}{\mathrm{d}}a \nonumber\\ & = & \frac{a^n-\left(a-\delta_a\right)^n}{\delta_a}{\mathrm{d}}a = \frac{\delta_{a^n}}{\delta_a}{\mathrm{d}}a.\end{aligned}\ ] ] note again that the transformation in the fourth line is well - defined only because the expansion of the numerator cancels the denominator
. we can generalize the identity ( [ hatano - eq770 ] ) to any analytic functions defined by the taylor expansion ( [ hatano - eq790 ] ) .
the result is @xmath181 it is interesting to note that the operator differential or the quantum derivative @xcite is expressed by a difference form of hyperoperators .
as a special case , we arrive at the identity @xcite @xmath182 we now use the formula ( [ hatano - eq810 ] ) for the derivation of the baker - campbell - hausdorff formula , or the derivation of higher - order terms of the exponent @xmath183 given in @xmath184 the differential of the left - hand side of eq .
( [ hatano - eq830 ] ) gives @xmath185 owing to eq .
( [ hatano - eq810 ] ) , while the differential of the right - hand side of eq .
( [ hatano - eq830 ] ) gives @xmath186 where we have used the identity ( [ hatano - eq720 ] ) .
equating the both sides , we have @xmath187 the second equality is due to the identity ( [ hatano - eq740 ] ) .
we can expand the right - hand side of eq .
( [ hatano - eq855 ] ) as follows .
note here that @xmath188 thus we transform eq .
( [ hatano - eq855 ] ) as @xmath189 we finally arrive @xcite at @xmath190 it is very important to notice here that all the expansion terms are given by commutation relations .
one of the merits of the quantum analysis is to be able to express the expansion in terms of commutation relations .
let us derive , for example , the term of the third order of @xmath1 , or the second order of @xmath191 of eq .
( [ hatano - eq880 ] ) . up to the second order
, we have @xmath192 and hence @xmath193 summing up the second - order terms with the coefficient @xmath194 , we have @xmath195 which we integrate to obtain @xmath196 } \right ] } + { \left [ { \left [ a , b \right ] } , b \right ] } \right).\ ] ] we now extend the above computation to the exponential product @xmath197 and seek ruth s formula ( [ hatano - eq500 ] ) as a specific solution of the general formula .
we compute the second - order and third - order correction terms of @xmath183 , defined in @xmath198 and put @xmath199 . the same computation as from eq .
( [ hatano - eq840 ] ) through eq .
( [ hatano - eq880 ] ) produces @xmath200 note again that all the terms are given by commutation relations .
for the term @xmath201 , we have up to the second order of @xmath1 , @xmath202p_3a + \cdots \nonumber\\ & = & \left(p_1+p_3+p_5\right)a + \left(p_2+p_4+p_6\right)b \nonumber\\ & & + t\left [ p_1p_2\delta_ab + p_2p_3\delta_ba+ \left(p_1+p_3\right)p_4\delta_ab + \left(p_2+p_4\right)p_5\delta_ba + \left(p_1+p_3+p_5\right)p_6\delta_ab \right ] \nonumber\\ & & + \frac{t^2}{2}\left [ p_1 ^ 2p_2{\delta_a}^2b + p_2 ^ 2p_3{\delta_b}^2a + 2p_1p_2p_3\delta_a\delta_ba + \left(p_1+p_3\right)^2p_4{\delta_a}^2b \right . \nonumber\\ & & \phantom{+x^2 } + 2p_2p_3p_4\delta_b\delta_ab + \left(p_2+p_4\right)^2p_5{\delta_b}^2a + 2\left(p_1p_2+p_1p_4+p_3p_4\right)p_5\delta_a\delta_ba \nonumber\\ & & \phantom{+x^2 } + \left(p_1+p_3+p_5\right)^2p_6{\delta_a}^2b + 2\left(p_2p_3+p_2p_5+p_4p_5\right)p_6\delta_b\delta_ab \bigr].\end{aligned}\ ] ] the zeroth - order term with respect to @xmath191 appears only here and hence we have the conditions @xmath203 using eq .
( [ hatano - eq3040 ] ) and the identity @xmath204 , we can reduce the right - hand side of eq .
( [ hatano - eq3030 ] ) as @xmath205\ ] ] with @xmath206 for @xmath207 , we first have @xmath208 , \ ] ] where we already used the conditions in eq .
( [ hatano - eq3040 ] ) .
applying eq .
( [ hatano - eq3060 ] ) to eq .
( [ hatano - eq3050 ] ) and dropping higher - order terms , we note that the first - order term vanishes and have @xmath209(a+b ) \nonumber\\ & = & t^2 ( 1 - 2q)\left({\delta_a}^2b-{\delta_b}^2a\right ) + \frac{t^2}{2}\left [ { \delta_a}^2b+{\delta_b}^2a-2(1-q){\delta_a}^2b -2q{\delta_b}^2a \right ] \nonumber\\ & = & t^2 \left(\frac{1}{2}-q\right ) \left ( { \delta_a}^2b -{\delta_b}^2a \right).\end{aligned}\ ] ] the second - order term of @xmath191 in the term @xmath210 vanishes just as in eq .
( [ hatano - eq905 ] ) .
thus we arrive at @xmath211+{\mathop{\mathrm{o}}}(x^4 ) .
\ ] ] putting the second - order and third - order terms to zero , we have a set of simultaneous equations of the parameters as @xmath212 we can confirm that ruth s formula ( [ hatano - eq500 ] ) , or @xmath213 is indeed a solution of the above set of simultaneous equations . with six variables for five equations ,
the solution is in fact a continuous line ; ruth s solution ( [ hatano - eq3100 ] ) is just a point on the line . by adjusting the last variable @xmath214
, we have the continuous solution shown in fig .
[ hatano - fig50 ] .
( we can solve the set of equations with five parameters by putting @xmath215 , but the solution is complex . )
we finally present an interesting exercise , motivated by the perturbational composition " @xcite .
suppose that we apply a weak transverse field to an ising spin .
we ask what is the correction term in the exponent of the right - hand side of @xmath216 notice that we expand the exponent with respect to the perturbation parameter @xmath217 , not with respect to @xmath1 as in the preceding sections .
the first - order perturbation term @xmath218 in turn contains higher orders of @xmath1 .
we could explicitly compute the @xmath219 matrices on both sides of eq .
( [ hatano - eq4000 ] ) , expand them with respect to @xmath217 and compare them term by term , but the quantum analysis provides a more elegant way of computation .
we differentiate the both sides of eq .
( [ hatano - eq4000 ] ) with respect to @xmath217 : @xmath220 equating the both sides , we have @xmath221 putting @xmath222 , we have @xmath223 owing to the fact @xmath224 , where we have made the taylor expansion @xmath225 with @xmath226
. we also note that the function ( [ hatano - eq4050 ] ) is even with respect to @xmath1 and hence @xmath227 for odd integers @xmath41 . ) . ] ) . ]
the right - hand side of eq .
( [ hatano - eq4040 ] ) is explicitly calculated as follows . in each order
, we have @xmath228}=2{\mathrm{i}}\sigma_y , \\ \label{hatano - eq4062 } { \delta_{\sigma_z}}^2\sigma_x & = & 2{\mathrm{i}}{\left [ \sigma_z , \sigma_y \right]}=4\sigma_x , \\ \label{hatano-^eq4064 } { \delta_{\sigma_z}}^3\sigma_x & = & 4{\left [ \sigma_z , \sigma_x \right]}=8{\mathrm{i}}\sigma_y , \\ & & \cdots , \nonumber\end{aligned}\ ] ] or in general , @xmath229 we substitute eq .
( [ hatano - eq4070 ] ) for each even - order term of the right - hand side of eq .
( [ hatano - eq4040 ] ) and arrive at @xmath230 ( in the second equality , we used the taylor expansion ( [ hatano - eq4050 ] ) in the reverse direction . ) in summary , we have @xmath231}.\ ] ] the coefficient @xmath232 behaves as shown in fig .
[ hatano - fig40 ] .
we have @xmath233 for small @xmath1 as is expected , but @xmath234 for large @xmath1 , and hence the first - order perturbation term grows as @xmath235 .
in the present article , we have reviewed a continual effort on generalization of the trotter formula to higher - order exponential product formulas . as was emphasized in sect .
[ hatano - sec20 ] , the exponential product formula is a good and useful approximant , particularly because it conserves important symmetries of the system dynamics .
we focused on two algorithms of constructing higher - order exponential product formulas .
the first is the fractal decomposition , where we construct higher - order formulas recursively .
the second is to make use of the quantum analysis , where we compute higher - order correction terms directly . as interludes
, we also have described the decomposition of symplectic integrators , the approximation of time - ordered exponentials , and the perturbational composition .
it is our hope that the readers find the present article a useful and tutorial manual " when they numerically investigate dynamical systems . for more practical applications of the exponential product formulas ,
we refer the readers to the review articles found in refs .
we mention here another kind of the exponential product formula @xcite .
consider the trotter approximant @xmath236}+{\mathop{\mathrm{o}}}(x^3)}.\ ] ] we can cancel out the second - order correction term in the form @xmath237 } } = { \mathrm{e}}^{x(a+b)+{\mathop{\mathrm{o}}}(x^3)}.\ ] ] if , in some problems , the commutation relation @xmath4}$ ] is easily diagonalized , eq .
( [ hatano - eq1010 ] ) may be a useful approximant .
a more complicated one is the fourth - order approximant @xcite @xmath238 } \right ] } } s_a\left(\frac{x}{3}\right)s_b\left(\frac{x}{3}\right)s_a\left(\frac{x}{3}\right ) { \mathrm{e}}^{\frac{x^3}{432}{\left [ b , { \left [ a , b \right ] } \right ] } } = { \mathrm{e}}^{x(a+b)+{\mathop{\mathrm{o}}}(x^5)},\ ] ] where @xmath239 in fact , the diffusion equation is described by @xmath240 and we have @xmath241 } \right]}=\left(\nabla v(\vec{q } ) \right)^2.\ ] ] the above type of the exponential product formula was referred to as the hybrid exponential product formula .
we do not give its details in this article , since commutation relations are not easily diagonalized except for a few specific problems .
in the present appendix , we give a short review of the world - line quantum monte carlo method . the world - line quantum monte carlo method is to transform the partition function ( [ hatano - eq120 ] ) of a quantum system @xmath21 into the partition function of a classical system by means of the path - integral representation and simulate the latter system .
we explain the method using the transverse ising model ( [ hatano - eq130 ] ) , or @xmath242 with @xmath243 the starting point is the trotter decomposition ( [ hatano - eq220 ] ) of the partition function , namely the suzuki - trotter transformation @xcite , of the form : @xmath244 in the second line , we have taken the trace with respect to a complete basis set by using the spin @xmath25 axis as the quantization axis : @xmath245 where the eigenvalue is @xmath246 .
the meaning of the superscript @xmath247 becomes self - evident just below .
we further insert the resolution of unity in between each pair of the exponential operators in the last line of eq .
( [ hatano - eq2010 ] ) , obtaining @xmath248 in the above expression , we used the fact that the operator @xmath2 is diagonal in the representation of @xmath249 and hence made the complete set on the both sides of each operator @xmath250 identical .
in contrast , the operator @xmath251 has off - diagonal elements .
let us calculate the matrix elements in eq .
( [ hatano - eq2020 ] ) .
the matrix element of the operator @xmath250 is easy : @xmath252 this is because the operators @xmath253 are all diagonal in the present representation as in eq .
( [ hatano - eq2015 ] ) . on the other hand , the operator @xmath251 has off - diagonal elements as well as diagonal ones in the following form : @xmath254 with each matrix element given by @xmath255 these matrix elements are expressed in a single equation @xmath256 where the parameters @xmath257 and @xmath258 are defined in @xmath259 or more explicitly defined by @xmath260 the expressions ( [ hatano - eq2030 ] ) and ( [ hatano - eq2060 ] ) give the partition function ( [ hatano - eq2020 ] ) in the form @xcite @xmath261 with the resulting classical hamiltonian @xcite @xmath262 where we dropped a constant term due to @xmath258 .
note that the periodic boundary conditions , @xmath263 , must be required in the second term of eq .
( [ hatano - eq2100 ] ) because the trace operation in eq .
( [ hatano - eq2010 ] ) demands it .
the classical hamiltonian ( [ hatano - eq2100 ] ) is interpreted as follows ( fig .
[ hatano - figa10 ] ) .
suppose that the original quantum system ( [ hatano - eq2000 ] ) is defined on a square lattice .
the first term of eq .
( [ hatano - eq2100 ] ) indicates that the two - dimensional system is replicated into @xmath41 layers with the intra - layer interaction reduced by @xmath41 times . the second term of eq .
( [ hatano - eq2100 ] ) represents the inter - layer interactions .
the coupling is @xmath264 as defined in eq .
( [ hatano - eq2100 ] ) .
thus the quantum system on a square lattice is mapped to an ising model on a cubic lattice . in general , a @xmath265-dimensional quantum system is mapped to a @xmath266-dimensional classical system .
the additional axis is called the trotter direction .
the physical quantities of the quantum system can be estimated by monte carlo simulation of the mapped classical system .
this is the basic idea of the world - line quantum monte carlo method @xcite .
we can use this mapping in order to study the quantum annealing @xcite .
suppose that we look for the ground state of the diagonal part @xmath2 of the system ( [ hatano - eq2000 ] ) .
random exchange interactions @xmath267 may produce many local minima that are only slightly above the ground state in energy but far apart from the ground state in the phase space .
the simulated annealing , a well - known method of ground - state search , is often trapped in a local minima and does not reach the ground state . in quantum annealing , we use the transverse field @xmath113 in order to induce tunneling from local minima to the ground state .
we first apply the off - diagonal part @xmath3 of eq .
( [ hatano - eq2000 ] ) strongly and turn it off gradually , hoping to end up with the ground state of the diagonal part @xmath2 .
this corresponds to a monte carlo simulation of the mapped classical system ( [ hatano - eq2100 ] ) with the intra - layer coupling @xmath257 being infinitesimally weak at the beginning and infinitely strong at the end .
each layer of the system ( [ hatano - eq2100 ] ) is first independent of each other and is gradually frozen into an identical configuration , which we hope is the ground state .
an annoying problem inherent in the algorithm of the quantum monte carlo method is the systematic error due to the finite trotter number @xmath41 .
it used to be that simulations were carried out for various finite values of @xmath41 , quantities were estimated in each simulation , and then the limit @xmath43 was taken in the process of the data analysis , which was called the trotter extrapolation .
a recent development of the quantum monte carlo method dramatically changed the situation .
we here mention the development briefly ; see ref .
@xcite for a tutorial and exhaustive review of the topic . in the most recent quantum monte carlo algorithm ,
it is possible for some systems to take the trotter limit _ before _ we set up the classical system for simulation . taking the trotter limit @xmath43
, we have a continuum trotter axis ( fig .
[ hatano - figa20 ] ) .
( note again that the boundary conditions are required in the trotter direction . )
the interaction is described as follows ( fig .
[ hatano - figa30 ] ) . instead of ising spins on lattice points of a trotter axis
, we have up - spin domains and down - spin domains on the axis . instead of intra - layer interactions between a pair of nearest - neighbor spins , we have parallel - spin areas and anti - parallel - spin areas . in monte carlo simulation , we update the up - spin domains and down - spin domains on the basis of the energy of the parallel - spin areas and anti - parallel - spin areas .
it is thus possible in such situations to carry out a simulation in the trotter limit @xmath43 .
monte carlo estimates of such a simulation are free of the systematic error of the order @xmath268 in eq .
( [ hatano - eq220 ] ) , and hence do not need the higher - order exponential product formula for such systems .
m. suzuki , k. umeno : higher - order decomposition theory of exponential operators and its applications to qmc and nonlinear dynamics . in : _
computer simulation studies in condensed - matter physics vi _ , ed by d.p .
landau , k.k .
mon , h .- b .
schttler ( springer , berlin heidelberg , 1993 ) pp 7486 m. suzuki : convergence of exponential product formula and its applications to hamiltonian systems . in : _ dynamical systems and
chaos _ , vol 2 , ed by y. aizawa , s. saito , k. shiraiwa ( world scientific , singapore , 1994 ) pp 450453 m. suzuki : new scheme of hybrid exponential product formulas with applications to quantum monte carlo simulations . in : _
computer simulation studies in condensed - matter physics viii _
, ed by d.p .
landau , k.k .
mon , h .- b .
schttler ( springer , berlin heidelberg new york , 1995 ) pp 169174 m. suzuki : general theory of exponential product formulas . in : _
computational physics as a new frontier in condensed matter research _
, ed by h. takayama , m. tsukada , h. shiba , f. yonezawa , m. imada , y. okabe ( physical society of japan , tokyo , 1995 ) pp 5156 m. suzuki : general theory of exponential product formulas and its applications to quantum fluctuation . in : _
coherent approaches to fluctuations _
, ed by m. suzuki , n. kawashima ( world scientific , singapore , 1996 ) pp 95 - 100 | this article is based on a talk presented at a conference quantum annealing and other optimization methods " held at kolkata , india on march 25 , 2005 .
it will be published in the proceedings quantum annealing and other optimization methods " ( springer , heidelberg ) pp .
3970 . in the present article
, we review the progress in the last two decades of the work on the suzuki - trotter decomposition , or the exponential product formula .
the simplest suzuki - trotter decomposition , or the well - known trotter decomposition @xcite is given by @xmath0 where @xmath1 is a parameter and @xmath2 and @xmath3 are arbitrary operators with some commutation relation @xmath4}\neq0 $ ] . here the product of the exponential operators on the right - hand side is regarded as an approximate decomposition of the exponential operator on the left - hand side with correction terms of the second order of @xmath1 .
mathematicians put eq .
( [ hatano - eq10 ] ) in the form @xmath5 and ask what correction terms appear in the exponent of the right - hand side owing to the product in the left - hand side .
they hence refer to it as an exponential product formula .
( the readers should convince themselves by the taylor expansion that the second - order correction in eq .
( [ hatano - eq10 ] ) is the same as that in eq .
( [ hatano - eq20 ] ) .
the higher - order corrections take different forms . )
we here ask how we can generalize the trotter formula ( [ hatano - eq10 ] ) to decompositions with higher - order correction terms .
we concentrate on the form @xmath6 or equivalently @xmath7 we adjust the set of the parameters @xmath8 so that the correction term may be of the order of @xmath9 .
we refer to the right - hand side of eq .
( [ hatano - eq30 ] ) as an @xmath10th - order approximant in the sense that it is correct up to the @xmath10th order of @xmath1 .
( see appendix [ hatano - app05 ] for another type of the exponential product formula . )
one of the present authors ( m.s . )
has studied on the higher - order approximant continually @xcite .
the present article mostly reviews his work on the subject .
we first show the importance of the exponential operator in sect .
[ hatano - sec10 ] and the effectiveness of the exponential product formula in sect .
[ hatano - sec20 ] .
we demonstrate the effectiveness in examples of the time - evolution operator in quantum dynamics and the symplectic integrator in hamilton dynamics .
section [ hatano - sec30 ] explains a recursive way of constructing higher - order approximants , namely the fractal decomposition .
we present in sect .
[ hatano - sec40 ] an application of the fractal decomposition to the time - ordered exponential .
we finally review in sect .
[ hatano - sec50 ] the quantum analysis , an efficient way of computing correction terms of general orders algebraically .
we can use the quantum analysis for the purpose of finding approximants of an arbitrarily high order by solving a set of simultaneous equations where the higher - order correction terms are put to zero .
we demonstrate the prescription in three examples .
we mention in appendix [ hatano - app05 ] a type of the exponential product formula different from the form ( [ hatano - eq30 ] ) ; it contains exponentials of commutation relations .
we give in appendix [ hatano - app10 ] a short review on the world - line quantum monte carlo method with the use of the trotter approximation ( [ hatano - eq10 ] ) . |
studies of lorentz violation ( lv ) and cpt violation ( cptv ) arguably are best performed in the context of the general effective - field - theory ( eft ) approach called the standard - model extension ( sme).@xcite the sme lagrangian contains all lv and cptv operators that one can construct from the conventional standard - model fields , coupled to fixed - valued lorentz tensors which parametrize the symmetry breaking and whose values presumably arise from a more fundamental theory .
one of the great merits of the sme is that it provides a concrete and general way for experiments to constrain lv and cptv.@xcite in this respect , clock - comparison , maser , and magnetometer experiments have been particularly successful . as a result
very stringent bounds have been placed on a number of effective neutron and proton coefficients.@xcite however , the connection between these effective coefficients and the quark and gluon coefficients from which they must arise is obscured by the fact that qcd is nonperturbative at the relevant ( low ) energies . in the usual lorentz - symmetric case ,
the low - energy effects of qcd are addressed successfully by the eft of qcd , called chiral perturbation theory ( @xmath0pt ) ( for reviews see , e.g. , ref . ) .
it thus seems natural to extend this approach to the qcd sector of the sme .
the conceptual basis of this extension has been established in our recent paper@xcite in which we treated three cptv and lv quark and gluon operators of mass - dimension five , given by @xmath1 + h_{\mu\nu\rho}\,{\rm tr}\left(g^{\mu\lambda}d^\nu\tilde{g}^{\rho}_{\;\;\;\lambda}\right)\ . \label{lagr1gevq}\ ] ] here , the lv coefficient @xmath2 is fully symmetric on its lorentz indices and @xmath3 , with @xmath4 .
these operators were first written down in ref . and are relevant below the electroweak scale @xmath5 gev , but above the qcd chiral - symmetry breaking scale @xmath6 gev , where @xmath7 mev is the pion decay constant .
the natural expected size of the lv coefficients is @xmath8 , where @xmath9 is the scale at which the lv occurs and is commonly identified with the planck scale @xmath10 gev . because we are interested in observables for non - strange baryons , we have restricted attention to up and down quarks . in the following we outline the approach described in ref .
to ensure that observables that follow from an eft correspond to the ones that arise from the underlying theory , one has to write the most general lagrangian in terms of the relevant low - energy degrees of freedom that is consistent with all the symmetries of the underlying theory.@xcite in the resulting infinite tower of operators , operators of higher mass dimensionality are suppressed by higher powers of a large mass scale , such that for a given required accuracy the set of terms can be truncated .
the suppressing mass scale determines the range of validity for the eft . in the case of @xmath0pt ,
this mass scale is given by @xmath11 and the relevant degrees of freedom at energies below @xmath11 are the pions , nucleons , and photons . particularly important in formulating the effective lagrangian
is the approximate global @xmath12 chiral symmetry of qcd .
this symmetry is explicitly broken by the small quark masses while inspection of the hadron spectrum reveals that it is also spontaneously broken down to its @xmath13 isospin subgroup .
the pions are subsequently identified as the pseudo - goldstone bosons . in the chiral limit ( vanishing quark masses )
the pions only interact through spacetime derivatives .
this allows for a perturbative expansion of hadronic observables in terms of the small parameter @xmath14 , where @xmath15 is the typical momentum of the process under consideration . in ref .
we constructed the effective lagrangian in the @xmath16 formalism of @xmath0pt .
the form of the effective operators is severely constrained by the transformation properties of the operators in eq . under the chiral @xmath16 group as well as under discrete c , p , and t transformations .
the resulting set of non - redundant leading - order nucleon and pion - nucleon operators is given by @xmath17\right)\sigma^{\nu\rho}\mathcal{d}^\mu n \notag \\ & & + \frac{1}{m_n^2}\tilde{h}_{\mu\nu\rho}\bar{n}\gamma^\mu \gamma^5 \mathcal{d}^\nu \mathcal{d}^\rho n + \frac{i}{m_n f_\pi d } \tilde{d}^-_{\mu\nu\rho}\bar{n}(\boldsymbol{\tau}\times\boldsymbol{\pi})_3\sigma^{\nu\rho}\mathcal{d}^\mu
n \notag \\ & & + \mathrm{h.c.}\ , \label{efflagr}\ ] ] where @xmath18 is the nucleon doublet , @xmath19 is the nucleon mass , @xmath20 is the pion triplet , @xmath21 are the pauli matrices , @xmath22 , and @xmath23 is the chiral- and gauge - covariant nucleon derivative , whose explicit form can be found in ref . .
the lv coefficients are given by @xmath24 , @xmath25 , and @xmath26 , with @xmath27 , @xmath28 , and @xmath29 being low - energy constants whose size can not be determined by symmetry considerations , although chiral symmetry gives @xmath30 .
using naive dimensional analysis,@xcite an order - of - magnitude estimation for the low - energy constants can be obtained : @xmath31 and @xmath32 .
the strongest bounds can be set on kinetic nucleon terms , using magnetometer experiments .
these kinetic terms induce a shift in the energy levels of an atom with total angular momentum @xmath33 and projection number @xmath34 . in ref .
we determined that the dominant part of this shift is given by @xmath35\left\langle [ \sigma_3]_{w , n}\right\rangle\ , \ ] ] where @xmath36 labels the particle species ( neutron or proton ) and @xmath18 runs over all particles of that species that are present in the atom .
the unknown matrix elements @xmath37_{w , n}\right\rangle$ ] are defined in the ` stretched ' state @xmath38 and one has to adopt some nuclear - structure model to determine their value . using the results of a @xmath39he/@xmath40xe comagnetometer experiment,@xcite we were able to put the current best bounds on components of @xmath41 and @xmath42 , on the order of @xmath43 gev@xmath44 .
subleading effects of @xmath41 and @xmath42 on atomic energy levels , as well as their effect on spin - precession , can be found in ref . .
the lv @xmath45 coefficient does not show up in these calculations , because it has no kinetic nucleon terms in eq . .
such terms are forbidden by the symmetries . in ref . , we calculated the contribution of @xmath45 to the electromagnetic form factor of the nucleon , through pion loops .
experimental effects of this are suppressed with a factor of @xmath46 with respect to effects of @xmath41 and @xmath42 , due to small electromagnetic fields .
this leads to a best bound on some components of @xmath45 in the order of @xmath47 gev@xmath44 , which does not probe the expected size of @xmath48 . however , despite the lack of kinetic nucleon terms , the @xmath45 coefficient does contribute to one - pion exchange between nucleons.@xcite in future work it might therefore be possible to improve the bounds on @xmath45 by considering the lv nucleon - nucleon ( nn ) potential
. studies of the nn potential and its applications to deuterons and other light nuclei have been performed successfully in the context of chiral eft with discrete - symmetry violations other than cptv ( see ref . for a review ) .
an analogous treatment in the present case is expected to lead to new bounds from atomic experiments or storage - ring experiments for the deuteron or other light nuclei .
the author acknowledges financial support from the portuguese foundation for science and technology ( fct ) under grant sfrh / bpd/101403/2014 and program poph / fse . | we consider the low - energy effects of a selected set of lorentz- and cpt - violating quark and gluon operators by deriving the corresponding chiral effective lagrangian . using this effective lagrangian , low - energy hadronic observables can be calculated .
we apply this to magnetometer experiments and derive the best bounds on some of the lorentz - violating coefficients .
we point out that progress can be made by studying the nucleon - nucleon potential , and by considering storage - ring experiments for deuterons and other light nuclei . |
final - state interaction ( fsi ) effects play important role in many physical processes , and in particular in various weak decays . these effects may significantly affect determination of fundamental @xmath5-violating parameters since extraction of the latter requires at least some knowledge of fsi .
the role of fsi in b decays was discussed in @xcite .
unfortunately , understanding it constitutes a difficult task for both theory and phenomenology .
our model of coupled - channel final - state interaction is based on a quasielastic approximation and regge pole methods @xcite .
the basic physical idea of regge model is that the high energy behavior of s - channel amplitudes is determined by `` exchanges '' in the crossed channel .
our model considers rescaterings of the type : @xmath6 , where @xmath7 and @xmath8 denote pairs of pseudoscalar mesons : @xmath9 , @xmath1 , @xmath10 , @xmath11 and @xmath12 .
the dominant exchanges in the @xmath13-channel are the pomeron ( @xmath14 ) and the regge trajectories . in that framework
the coupled - channel fsi effects for @xmath15 weak decays into @xmath9 and @xmath1 were discussed in refs .
the calculations @xcite were performed under the assumption of the exchange of the @xmath16 , @xmath17 , @xmath18 , @xmath19 regge trajectories , the trajectories of their @xmath2 partners , and the exactly @xmath2-symmetric pomeron . in this paper
we analyze in some details both the influence of @xmath2 breaking in the pomeron , and the influence of singlet penguin amplitude on the predictions of the quasi - elastic coupled - channel regge approach of ref.@xcite .
if @xmath2 in the pomeron is broken and the singlet penguin is not neglected , the conclusions of refs.@xcite would have to be modified .
we use the following phase conventions for pseudoscalar mesons : @xmath20 for cabibbo - suppressed @xmath21 decays there are nine possible final states composed of two pseudoscalar mesons . in the basis of definite isospin @xmath22 the symmetrized two - boson states @xmath23 are : @xmath24 quark diagram amplitudes ------------------------ the decays of @xmath21 mesons to two pseudoscalar mesons ( @xmath7 ) are described by 7 flavor-@xmath2 invariant amplitudes @xcite , but only 4 of them ( fig . [ 4d ] ) : `` tree '' ( @xmath25 ) , `` color - suppressed '' ( @xmath26 ) , `` penguin '' ( @xmath27 ) and additional penguin involving flavor-@xmath2-singlet ( @xmath28 ) diagrams , are important @xcite .
we assume that @xmath29 , with @xmath30 @xcite , @xmath31 @xcite and latexmath:[$|s|\approx(0.6\pm0.2 ) -flavor amplitudes for the decays of b mesons to a pair of light pseudoscalar mesons .
( @xmath25 ) `` tree '' ; ( @xmath26 ) `` color - suppressed '' ; ( @xmath27 ) `` penguin '' ; ( @xmath28 ) additional penguin involving flavor-@xmath2-singlet .
[ 4d],scaledwidth=40.0% ] short - distance amplitudes : @xmath25 , @xmath26 , @xmath27 and @xmath28 , have weak and strong phases .
we can write : @xmath33 where @xmath34 , @xmath35 , ( @xmath36 ) are weak ( strong ) phases .
it is possible that the short - distance weak amplitudes have large strong phases .
however , since we want to study fsi we neglect these phases ( i.e. we set @xmath37 ) . for the weak phases we assume @xcite @xmath38
furthermore , we neglect the electroweak penguins diagrams @xcite . in terms of quark diagram
amplitudes the weak decays of @xmath15 to the states defined in eq .
[ states ] are given by : @xmath39 we assumed @xmath2 symmetry in weak decays , i.e. equal amplitudes for the production of strange ( @xmath40 ) and nonstrange quark pairs .
the weak amplitude @xmath41 is changed by isospin - conserving strong interaction @xmath42 in the final state @xcite into a fsi - corrected weak amplitude @xmath43 : @xmath44 where subscript @xmath22 denotes isospin .
we describe @xmath42 in the regge pole model as used in @xcite-@xcite . in the energy range
@xmath45 gev@xmath46 the pomeron ( @xmath14 ) contribution to the @xmath13-channel amplitude is phenomenologically well described by the formula @xcite @xmath47 where the residue @xmath48 and slope @xmath49 depend on the scattering process considered .
calculations of the s - channel @xmath50 waves @xmath51 , give , for the pomeron @xcite : @xmath52 from @xcite we obtain @xmath53 and @xmath54 the simple relations between @xmath55 , @xmath56 and @xmath57 , @xmath58 for broken @xmath2 are given by @xmath59 and @xmath60 from eqs .
[ a_p ] [ b_p ] we find : @xmath61 in the @xmath2 symmetric case we have @xmath62 mbgev@xmath46 .
many authors restrict their studies to elastic rescattering only . in regge language
this is described in terms of a pomeron exchange .
but at @xmath63 contributions from other inelastic nonleading regge exchanges are not completely negligible @xcite . )
uncrossed reggeon exchange , ( @xmath64 ) crossed reggeon exchange .
[ cu],scaledwidth=38.0% ] there are two types of contributions from exchange - degenerate reggeons corresponding to two different diagrams ( crossed @xmath64 and uncrossed @xmath65 , see fig .
the contributions of diagrams @xmath65 and @xmath64 differ in their phases @xcite.the calculations of the @xmath66-channel @xmath50 partial waves amplitudes @xmath67 for uncrossed reggeon exchange and @xmath68 for crossed reggeon exchange give @xcite : @xmath69 and @xmath70 where @xmath71 is the regge residue fitted from experiment @xcite,@xcite : @xmath72 and @xmath73 the scale factor @xmath74 is taken as @xmath75 .
inelastic fsi means here the coupled - channel effects of the type : @xmath76 , @xmath77 , @xmath78 , , and @xmath79 , @xmath77 , @xmath78 , etc in the final state .
inclusion of such processes was shown in @xcite to be very important .
there are tree separate non - communicating fsi sectors of different isospin @xmath80 . in the @xmath81 sector
one obtains only the contribution from the crossed diagram of fig .
[ cu ] : @xmath82=0,\ ] ] @xmath83=2.\ ] ] in the @xmath84 sector there are three states , and consequently we have coupled - channel effects described together with quasi - elastic effects by two @xmath85 matrices .
one obtains : @xmath86 = \left [ \begin{array}{ccc } \epsilon^2 & \frac{2}{\sqrt{3}}\epsilon & \sqrt{\frac{2}{3}}\epsilon \\ \frac{2}{\sqrt{3}}\epsilon & \frac{4}{3 } & \frac{2}{3}\sqrt{2 } \\ \sqrt{\frac{2}{3}}\epsilon & \frac{2}{3}\sqrt{2 } & \frac{2}{3 } \end{array } \right]\ ] ] and @xmath87\\ = \left [ \begin{array}{ccc } 0 & -\frac{2}{\sqrt{3}}{|\epsilon| } & 2\sqrt{\frac{2}{3}}|\epsilon| \\ -\frac{2}{\sqrt{3}}{|\epsilon| } & \frac{4}{3 } & \frac{2}{3}\sqrt{2 } \\
2\sqrt{\frac{2}{3}}|\epsilon| & \frac{2}{3}\sqrt{2 } & \frac{2}{3 } \end{array } \right].\ ] ] the states in the rows and columns are ( from top to bottom and from left to right ) : @xmath88 , @xmath89 and @xmath90 . in the @xmath91 sector
there are five states with rows and columns corresponding to the states ( from top to bottom and from left to right ) : @xmath92 , @xmath93 , @xmath94 , @xmath95 and @xmath96 .
one obtains : + @xmath97 @xmath98\ ] ] and @xmath99=$ ] @xmath100.\ ] ] the parameter @xmath101 ( @xmath102 ) describes suppression of propagation of one ( two ) strange quarks in the @xmath13-channel . for the @xmath2 discussion of coupled - channel effects @xmath103 @xcite .
a more realistic assumption used in this paper is : + @xmath104 where @xmath105 and @xmath106 are reggeon s parameters .
let us now connect weak decays and strong interactions in the final state .
we can obtain amplitudes @xmath107 of @xmath21 decay to states @xmath108 from : @xmath109 with @xmath110 are eigenvectors for @xmath111 matrices .
we assume now that the fsi - corrected weak decay amplitudes differ from quark - level expressions eq .
[ diagrams ] by hadronic phase factors only ( @xmath112 ) @xcite,@xcite . using eq .
[ main1 ] one obtains the numbers given in the right - hand side of tables [ tab1 ] , [ tab2 ] and in table [ tab3 ] . for the sake of comparison , in the left - hand side of table [ tab1 ] , [ tab2 ] we added amplitude phases with @xmath2 symmetric fsi ( table [ tab1 ] )
, as well as amplitude phases calculated without fsi effects ( table [ tab2 ] ) .
.comparison of calculated values of amplitude phases for @xmath21 decays with weak phases set to @xmath113 . [
cols="<,^,^,^,^,^,^,^,^,^ " , ] one has to realize , that for l=0 partial wave amplitudes the regge pole methods need not be reliable @xcite and , consequently , the obtained numbers should be considered as rough estimates only . in additional to fsi effects and hadronic phases of bare weak diagrams , @xmath5 violation effects
may strongly depend on electroweak diagrams @xcite , but there is not enough data to determine the corresponding parameters .
in summary , we have discussed the effect of abandoning exact @xmath2 in coupled - channel final - state interactions through reggeon exchange for @xmath114 .
@xmath2 was broken by admitting lower lying trajectories for strange reggeons @xmath115 ( eq .
( [ epsilon ] ) ) and in the pomeron ( eq . ( [ pomeron ] ) )
couplings . as expected in @xcite the singlet penguin diagram may be neglected in intermediate states of @xmath116 decays .
however , it can not be neglected in decays to @xmath11 , @xmath10 , and @xmath12 .
we have shown that strong fsi play an important role in the analysis of @xmath5-violating effects in @xmath117 decays .
the size of @xmath5 asymmetry in @xmath118 decays has been shown to depend strongly on the ratio of penguin to tree amplitude and on the strong phases of short - distance quark diagrams .
99 p. enczykowski , _ acta phys . pol .
_ * b28 * , 1605 ( 1997 ) .
donoghue _ et al .
_ , _ phys .
rev . _ * d33 * , 1516 ( 1986 ) .
l. wolfenstein , _ phys .
rev . _ * d52 * , 537 ( 1995 ) .
grard and j. weyers , _ eur
. phys . j. _ * c7 * , 1 ( 1999 ) .
d. delpine , j .-
grard , j. pestieau , j. weyers , _ phys . lett . _ * b429 * , 106 ( 1998 ) .
grard , j. pestieau , j. weyers , _ phys .
lett . _ * b436 * , 363 ( 1998 ) .
p. enczykowski , _ phys .
* b460 * , 390 ( 1999 ) .
enczykowski , _ acta phys . pol .
_ * b33 * 1833 ( 2002 ) .
m. gronau and j.l .
rosner , _ phys .
rev . _ * d58 * , 113005 ( 1998 ) .
k. anikeev _ at al .
_ , hep - ph/0201071
. z. luo and j.l rosner , _ phys .
rev . _ * d65 * , 054027 ( 2002 ) .
j.l rosner , in _ flavor physics for the millennium _
, p431 ( boulder 2000 ) , [ hep - ph/0011355 ] .
k.m .watson , _ phys . rev . _ * 88 * , 1163 ( 1952 ) .
r. seber , _ phys .
* 13 * , 32 ( 1964 ) . v. barger and r.j.n .
phillips , _ nucl .
phys . _ * b32 * , 93 ( 1971 ) .
b. aubcrt , _ et al._,_phys .
rev . _ * d66 * , 054024 ( 2002 ) .
the babar colaboration , hep - ph/0207065 .
et al . _ , cleo collaboration , _ phys .
* 85 * , 525 ( 2000 ) .
m. gronau and j.l rosner , _ phys .
rev . _ * d65 * , 093012 ( 2002 ) .
xing , _ phys .
* b493 * , 301 ( 2000 ) .
chua , w.s .
hou , k.c .
yang , hep - ph/0210002 . | for weak decays @xmath0 and @xmath1 the effects of @xmath2 breaking in coupled - channel final - state interaction effects are discussed in a regge framework . it is shown that @xmath2 breaking in the inelastic final - state transitions dramatically affects the phases of the isospin @xmath3 amplitudes in the @xmath4 decays .
the effect of the singlet penguin diagram on these phases is studied .
furthermore , on the example of the @xmath0 decays , the dependence of @xmath5 asymmetries on the size of penguin amplitude is analyzed .
pacs numbers : 13.25.hw , 11.30.hv , 11.80.gw , 12.15.hh |
dilute quantum gases are ideal for studying many - body physics , because they provide model systems in which the parameters can be precisely controlled .
external fields can be used to tune the effective isotropic contact interactions between the particles , and the geometry and strength of the confining optical potentials can be controlled by laser beams .
for example , quantum - gas analogues of superconductivity @xcite and the superfluid - to - mott - insulator quantum phase transition @xcite have been observed in the laboratory , and their properties have been shown to agree beautifully with the predictions from theoretical models @xcite .
recently , quantum gases of particles with long - range anisotropic interactions have been created @xcite . for particles with permanent electric dipole moments
, the range of the dipole - dipole interactions can be much larger than typical optical lattice spacings , and interesting new quantum phases and quantum information applications have been proposed @xcite .
a quantum gas of @xmath5k@xmath0rb ground - state molecules is the only such system that presently exists in the laboratory @xcite .
our goal is to generate a dipolar quantum gas of ground - state @xmath0rb@xmath1cs , which , unlike krb , is expected to be collisionally stable because both the exchange reaction 2rbcs @xmath6 rb@xmath7 + cs@xmath7 and trimer formation reactions are endothermic @xcite .
although other approaches are under development @xcite , the only method currently available to produce high phase - space density gases of ground - state molecules is to create weakly bound molecules from ultracold atomic gases by magnetic tuning across a feshbach resonance @xcite , and then to transfer the molecules to the rovibronic ground state by stimulated raman adiabatic passage ( stirap ) @xcite . as a first step ,
we have performed evaporative cooling on rb and cs samples in separate optical traps , combining them at the end to obtain an rb - cs mixture with high phase - space density @xcite .
we have successfully used this mixture to produce ultracold samples of weakly bound rbcs @xcite . in this paper , we present a combined experimental and theoretical study of the interspecies feshbach resonances and weakly bound molecular energy levels of rb - cs and use the results to develop an accurate coupled - channel model of the interaction , based on the derived interaction potentials for the molecular states @xmath8 and @xmath9 .
the work described in this paper involved a close collaboration between experiment and theory . at the start of the work ,
the feshbach resonances and bound states observed experimentally @xcite were unassigned . in initial theoretical work , we developed preliminary coupled - channel models of the bound states and scattering and used these to propose assignments of quantum numbers to observed energy levels and feshbach resonances .
experiments were then carried out to test the assignments and extend the early measurements .
the whole process was repeated several times .
however , to aid understanding , we will describe the experiments in section iii below using quantum numbers based on our final understanding from theory ( section iv ) , even though the quantum numbers were not known at the outset .
two alkali - metal atoms in @xmath10 states interact at short range to form singlet ( @xmath8 ) and triplet ( @xmath11 ) states .
et al . _
@xcite have carried out an extensive spectroscopic study of these states by fourier transform spectroscopy and have developed potential energy curves as shown in fig.[fig - curves ] .
they were able to observe the vibrational ladder up to high - lying levels with outer turning points around 1.5 nm , at which point the coupling between singlet and triplet molecular states is already significant .
they also identified in the observed spectra accidental coincidences of singlet and triplet levels deeper within the potential wells , which fixed their relative energy position very well . in the present work
, we initially constrained the short - range part of the potential to follow these curves and adjusted the long - range parameters to reproduce the feshbach resonances and weakly bound states .
the bound states ( feshbach molecules ) that are of most interest in the present paper have binding energies of at most a few mhz@xmath4 and require a quite different description . for a heteronuclear bialkali molecule ,
there are 4 field - free atomic thresholds , which for @xmath0rb@xmath1cs may be labelled in increasing order of energy by @xmath12 = ( 1,3 ) , ( 2,3 ) , ( 1,4 ) , and ( 2,4 ) , as shown in the inset of fig .
[ fig - curves ] . in a magnetic field , each threshold splits into @xmath13 sublevels labelled @xmath14 .
the feshbach molecules might be described using two different sets of quantum numbers , either @xmath15 or @xmath16 , where @xmath17 is the resultant of @xmath18 and @xmath19 and @xmath20 . in the non - rotating case @xmath17 and @xmath21
are exact quantum numbers if there is no external field , but if there is an external magnetic field , it mixes states with different @xmath17 values , destroying the exactness of @xmath17 as a quantum number ; the character of the feshbach molecules at the magnetic fields considered here is more accurately described by @xmath22 . for high magnetic fields , @xmath18 and @xmath19 are also no longer good quantum numbers . thresholds for @xmath0rb@xmath1cs and the bins " ( vertical boxes ) below each threshold within which each vibrational state must lie for any value of the scattering length .
the horizontal boxes , 1.5 ghz deep , show the energy range within which bound levels can cause resonances at the @xmath23 and @xmath24 thresholds at fields under 500 g. the vibrational `` bin '' for each hyperfine state that contains bound states that can cause resonances at the lowest threshold is colored .
selected levels are shown ( short horizontal lines ) as a function of @xmath25 ( @xmath2 , @xmath3 , @xmath26 , @xmath27 , @xmath28 for @xmath29 , 1 , 2 , 3 , 4 ) for the specific choice of scattering length that gives a least - bound state for @xmath29 with 110(20 ) khz below threshold . for fields below 500 g , only levels @xmath30 and @xmath31 can cause resonances at the @xmath23 threshold and only levels @xmath32 and @xmath33 can cause resonances at the @xmath24 threshold .
] additional quantum numbers are needed for the molecules end - over - end angular momentum @xmath25 and the molecular vibration . for near - dissociation levels
it is convenient to specify the vibrational quantum number with respect to the asymptote of the atom pair , so that the topmost level is @xmath34 , the next is @xmath35 , and so on .
each level lies within a bin " below its associated dissociation threshold , with the boundaries of the bins determined by the long - range forces between the atoms . for rbcs , using the published values of the long - range dispersion coefficients , @xcite , which are the same for the singlet and triplet potentials , we find that the @xmath34 level lies between zero and @xmath36 mhz , and the @xmath35 level lies between @xmath36 and @xmath37 mhz .
subsequent lower bin boundaries lie at 3.7 , 8.6 , 16.7 and 28.8 ghz below threshold for @xmath38 to @xmath39 , respectively . as shown below , the actual levels for @xmath0rbcs lie close to the top of their bins .
feshbach resonances occur at fields where a bound state exists at the same energy as the colliding atoms .
zero - energy feshbach resonances are caused by molecular levels that cross atomic thresholds as a function of magnetic field .
since the level shifts due to the zeeman effect at fields below 500 g are not more than 1.5 ghz , there is only one vibrational level below each field - free threshold that can cause feshbach resonances at the @xmath23 threshold , as shown in fig.[thresholds_and_bins ] ; these are @xmath40 , @xmath41 and @xmath39 for levels associated with @xmath42 , ( 1,4 ) and ( 2,4 ) , respectively .
in addition to this , levels very close to dissociation ( @xmath34 or @xmath43 ) corresponding to the _ same _ zero - field threshold as the incoming wave can also cause low - field resonances . fig.[thresholds_and_bins ] also shows the situation at the @xmath44 threshold , which will be considered in section [ sec : excited ] . in the general case we label weakly bound states with a complete set of quantum numbers @xmath45 , with @xmath46 , etc . designated by @xmath2 , @xmath3 , @xmath26 , etc
@xmath47 is the sum of all angular momenta projected onto the field axis , @xmath48 , and is the only exactly conserved quantum number in an external field .
since , however , @xmath47 is always 4 for the levels studied in this paper ( except in sections [ sec : p - wave ] and [ sec : excited ] ) , we will omit it in the following discussion .
all other angular momenta are approximate quantum numbers , but are sufficient for proper labeling .
we characterize by @xmath49 the partial wave character of the continuum scattering process and speak of incoming @xmath2- and @xmath3-wave resonances for @xmath50 and @xmath51 , respectively .
magnetic feshbach resonances are an important tool for the production of weakly bound molecules and for tuning the scattering length , which determines the elastic and inelastic scattering properties of cold atomic gases @xcite .
in addition , their positions provide important clues to the molecular bound state structure that lies below the scattering threshold . in previous work
@xcite we observed 23 resonances over the range 0 to 300 g , using a mixture of the lowest spin states , @xmath0rb@xmath52 and @xmath1cs@xmath53 .
since this mixture was prepared by evaporating both species simultaneously in the same optical trap , interspecies three - body recombination loss and heating @xcite limited the evaporative cooling efficiency , resulting in comparatively high temperatures of 7 @xmath54k and low particle densities of about @xmath55 @xmath56 for each species . in the current experiment ,
the mixture is created by combining separately cooled atomic clouds @xcite , so it is much colder ( 100 to 200 nk ) and denser ( @xmath57 @xmath56 ) and gives a much better signal - to - noise ratio for the loss features discussed below .
we stop the evaporation procedure before the onset of condensation , because we have previously found the two bose - einstein condensates ( becs ) to be immiscible @xcite .
we hold the mixture at constant magnetic field @xmath58 for 200 ms .
enhanced losses that occur simultaneously for rb and cs are attributed to three - body recombination @xcite at an interspecies feshbach resonance .
we associate the field value @xmath58 at which maximum atom loss occurs with the pole of the resonance .
for example , fig .
[ feshbachraw ] shows the atom loss in the vicinity of the resonance near @xmath59 g. for sufficiently wide resonances , we find that the number of rb atoms exhibits a maximum at fields just above resonance .
rb has a lower trap depth than cs and thus bears most of the heat load through evaporation when the two species are in thermal equilibrium @xcite .
reduced thermalization with cs at zero interspecies scattering length reduces the heat load on the rb part of the sample and thus leads to less loss of rb atoms .
this simple explanation allows us to provide an estimate for the resonance width @xmath60 : it is the difference between the field values for the minima ( for rb and cs ) and the maximum ( for rb ) as indicated in fig .
[ feshbachraw ] .
a detailed comparison with calculated widths requires a thorough analysis , including three - body and evaporation effects , and will be made in a future publication .
example of a feshbach resonance scan showing simultaneous loss for rb and cs near 197.06(5 ) g for a hold time of @xmath61 ms .
the maximum for the number of rb atoms to the right of the resonance is attributed to the zero crossing for the interspecies scattering length .
the difference in the positions of the minima and the maximum as indicated gives an estimate for the resonance width @xmath60.,width=325 ] as part of this work , we have scanned over a wider range ( 0 to 667 g ) than in ref .
@xcite , finding 7 new incoming @xmath2-wave resonances in addition to those reported in ref .
the old and new resonances are collected together in table [ resonances ] .
the resonances observed for temperatures @xmath62 nk are assigned as incoming @xmath2-wave ( @xmath63 ) resonances , while those observed at 7 @xmath54k and not observed at 200 nk are assigned as incoming @xmath3-wave ( @xmath64 ) resonances .
the magnetic field is calibrated near each resonance using rb microwave transitions .
the calibration in our previous work @xcite was based on low - field data and was found to deviate from the current calibration by as much as 0.5 g when extrapolated to 300 g. the positions of the incoming @xmath3-wave resonances have therefore been scaled to the new calibration , using the incoming @xmath2-wave resonances observed in both experiments as a reference .
the incoming @xmath3-wave resonances from 258 to 272 g have been remeasured at @xmath65k with the new calibration . .@xmath0rb@xmath52
+ cs@xmath53 feshbach resonances observed over the range 0667 g for @xmath2-wave scattering and 0300 g for @xmath3-wave scattering .
the magnetic field uncertainties result from a quadrature of resonance position uncertainty due to atom number noise , and an estimated field calibration error of 0.03 g. resonances too narrow to allow a clear width measurement have no width indicated .
[ resonances ] [ cols= " < , < , < " , ] @xmath66 this parameter is set to give continuity between the short - range and mid - range functional forms .
at the conclusion of the two - part least - squares refinement procedure described above , we arrived at the potentials given in tables [ fitparms ] , [ singlet ] and [ triplet ] .
table [ fitparms ] gives the parameters from fitting to the feshbach bound states and resonance positions , whereas tables [ singlet ] and [ triplet ] give the full potentials for the singlet and triplet states , respectively .
although the feshbach bound states and resonance positions do allow all 5 parameters in table [ fitparms ] to be extracted , they are very highly correlated .
the table therefore gives both 95% confidence limits and parameter sensitivities as defined by le roy @xcite .
the 95% confidence limits are correlated properties that describe the uncertainty in an individual parameter , but the parameters need to be specified to within their sensitivities , not their confidence limits , in order to reproduce the results of the calculations . the new version of the electronic potentials for the rbcs ground - state system reproduces the fourier transform spectra as accurately as the original version of ref .
@xcite , with the important improvement that it can also accurately reproduce properties relating to the very top of the electronic potentials , such as feshbach spectra . there are some remaining deviations between the observed and calculated positions of the @xmath35 states as shown in the lower panels of figure [ wiggle ] , but in view of the possible systematic errors in the corresponding measurements described in sections [ sec:13d03 ] and [ sec : magmom ] above , these are not a great cause for concern .
the final results for the resonance positions and weakly bound states are listed in table [ quality - of - fit ] , together with the quality of fit to the experimental data and the quantum label assignments .
the calculated @xmath2-wave scattering length and its match to the resonance positions is shown in figure [ s - wave - low ] , together with an overview of the bound states responsible for the resonances .
the singlet and triplet scattering lengths @xmath67 and @xmath68 obtained from the fitted potentials are included in table [ fitparms ] , together with their fully correlated uncertainties , calculated as described in ref .
the background scattering length derived for the @xmath23 channel is @xmath69 , calculated at @xmath70 g far from resonances .
we have also calculated the binding energy of the least - bound state for @xmath29 at @xmath71 g ( this value was chosen to represent a value far from resonance within the region for which experimental values are available ) and found it to be @xmath72 khz @xmath4 . as described above , some of the resonances observed by pilch _
et al . _
@xcite do not appear for the rb+cs mixture at the lower temperatures studied in the present work and are assigned as resonances in @xmath3-wave scattering .
when @xmath51 , @xmath73 can take values @xmath74 , 0 or + 1 , so @xmath47 can be 3 , 4 , or 5 at the latexmath:[$|1,1\rangle + compares the observed @xmath3-wave resonance positions with the @xmath3-wave scattering lengths for these three values of @xmath47 , calculated using the fitted potentials .
it may be seen that the observed resonances correspond quite well to a subset of the calculated resonances , although it is not altogether clear why pilch _
et al . _
@xcite observed some @xmath3-wave resonances and not others .
in addition to the resonances at the lowest threshold , pilch _ et al . _
@xcite observed two resonances at an excited threshold with the rb atoms in their @xmath76 state , at 162.3 and 179.1 g. at this threshold inelastic scattering is possible , and trap loss can occur through either 2-body or 3-body collisions .
the scattering length is complex , @xmath77 , and the inelastic collision rate is proportional to @xmath78 .
[ ea - scat ] shows the real and imaginary parts of the @xmath2-wave and @xmath3-wave scattering lengths at this threshold , calculated using the fitted potentials , and compares them to the experimental resonance positions .
it may be seen that the two observed resonances are in good agreement with the calculation , with the high - field resonance arising from @xmath2-wave scattering and the low - field resonance from @xmath3-wave scattering . as noted above
, there is one resonance observed in @xmath2-wave scattering , at 273.45 g , that does not appear in coupled - channel calculations on the fitted potential using a basis set with @xmath79 .
however , there are numerous additional resonances that appear when basis sets including more partial waves are used .
in particular , a calculation including @xmath80 functions yields an additional resonance at 275.07 g that arises from the @xmath81 state .
the exact position of this resonance is quite sensitive to variations of the potential within its uncertainty and is plausibly responsible for the otherwise unassigned resonance .
the resonances listed in table [ quality - of - fit ] , at fields up to 553 g , include all those expected from @xmath82 states . however , there are additional resonances that appear at higher field , mostly due to @xmath2 and @xmath26 states of @xmath83 .
some of the corresponding bound states appear in fig .
[ s - wave - low ] .
figure [ s - wave - high ] shows the @xmath2-wave scattering length at fields up to 1000 g ; in particular , the comparatively wide resonance near 790 g ( with width @xmath84 g ) is due to the @xmath85 state
. this wide resonance may be useful for tuning interspecies scattering properties , and for studying few - body properties such as interspecies efimov resonances @xcite .
in particular , since there is a very broad feshbach resonance for cs in state @xmath53 with a pole at @xmath86 g @xcite , rb+cs mixtures may make it possible to study efimov physics near overlapping feshbach resonances .
we have studied and modeled interspecies scattering in an ultracold rb - cs gas mixture with the aim of finding an assignment for the observed interspecies feshbach resonances and in particular to understand the spectrum of weakly bound rbcs molecules .
our results are of great importance for the production of ultracold samples of heteronuclear molecules and for the generation of dipolar quantum gases made of rbcs molecules . with recent work on optical one- and two - photon spectroscopy
@xcite we are now poised to perform stimulated ground - state transfer using the stirap technique .
we expect that a three - dimensional optical lattice will allow us to maximize the molecule creation and state transfer efficiencies , as in recent work on cs@xmath7 @xcite . as detailed in ref .
@xcite , interspecies feshbach tuning will be used to bring a superfluid sample of rb atoms into overlap with a single - atom - per - site mott insulator for cs , in order to optimize the rb - cs pair - creation efficiency . with sufficiently high efficiencies ,
the creation of a dipolar quantum gas of rbcs molecules is within reach .
the innsbruck team acknowledges support by the austrian science fund ( fwf ) and the european science foundation ( esf ) within the euroquam / qudipmol project ( fwf project number i124-n16 ) and support by the fwf through the sfb foqus ( fwf project number f4006-n16 ) .
the durham , jqi and temple university teams acknowledge support from an afosr muri project on ultracold polar molecules .
the durham and jqi teams acknowledge support from the engineering and physical sciences research council .
work at temple university was also supported by nsf grant phy 1005453 .
the durham and hannover teams acknowledge support from the qudipmol project and the hannover team also support by the deutsche forschungsgemeinschaft through the cluster of excellence quest . | we have studied interspecies scattering in an ultracold mixture of @xmath0rb and @xmath1cs atoms , both in their lowest - energy spin states .
the three - body loss signatures of 30 incoming @xmath2- and @xmath3-wave magnetic feshbach resonances over the range 0 to 667 g have been catalogued .
magnetic field modulation spectroscopy was used to observe molecular states bound by up to 2.5 mhz@xmath4 .
we have created rbcs feshbach molecules using two of the resonances .
magnetic moment spectroscopy along the magneto - association pathway from 197 to 182 g gives results consistent with the observed and calculated dependence of the binding energy on magnetic field strength .
we have set up a coupled - channel model of the interaction and have used direct least - squares fitting to refine its parameters to fit the experimental results from the feshbach molecules , in addition to the feshbach resonance positions and the spectroscopic results for deeply bound levels .
the final model gives a good description of all the experimental results and predicts a large resonance near 790 g , which may be useful for tuning the interspecies scattering properties .
quantum numbers and vibrational wavefunctions from the model can also be used to choose optimal initial states of feshbach molecules for their transfer to the rovibronic ground state using stimulated raman adiabatic passage ( stirap ) . |
it has been recognized that feedback from active galactic nuclei ( agn ) plays a significant role in the coevolution between galaxies and supermassive black holes ( smbhs ) in their centers ( for a review see e.g. , * ? ? ?
agn outflows in the form of jets and winds could expel and heat the gas in the interstellar medium , and thus regulating both star formation and mass accretion onto smbhs .
the powerful jets from agns can affect the surrounding structure even at the scale of a galaxy cluster . despite the importance of agn feedback
, several fundamental questions still remain open , such as : what are the physical mechanisms to launch relativistic jets and how the nuclear structure of agns with jets ( radio - loud agns ) differs from those without jets ( radio - quiet agns ) ?
x - ray observations bring us significant insight into the nuclear structure of agns .
there are previous works that compare x - ray spectral properties of radio - loud and radio - quiet agns .
@xcite find rich emission lines in the soft x - ray spectrum of 3c 234 , similar to those of radio - quiet agns .
these lines come from a photoionized plasma , suggesting that the narrow - line region in radio - loud and radio - quiet agns have similar geometrical and physical properties .
warm absorbers similar to those found in radio - quite agns are also detected from several radio - loud agns @xcite .
@xcite discover ultra - fast outflows in some radio - loud agns , and their features are very similar to those previously observed from radio - quiet agns .
however , no consensus has been firmly established yet whether or not there are differences between the two populations in their accretion disk structures and circumnuclear environments . to pursue these issues ,
our team have been working in _ suzaku_follow - up of hard x - ray selected radio - loud agns @xcite detected in the _
swift_/bat survey @xcite .
the torus is a key structure in agns , which is considered to be responsible for supplying mass onto smbhs .
it causes obscuration observed in so - called `` type-2 '' agns , a major agn population in the universe ( e.g. , * ? ? ?
* ) . to understand the origin of the torus , it is critical to identify the essential physical parameters that determine its structure .
previous studies , mainly based on the radio - quiet population , suggest that agn luminosity is one of the most important parameters .
the fact that the fraction of obscured agn is anti - correlated with x - ray luminosity above @xmath15 erg s@xmath1 ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) indicates that the solid angle covered by the torus decreases with @xmath16 .
this correlation can also explain the so - called x - ray baldwin effect @xcite that the equivalent width of the iron k@xmath17 line decreases with x - ray luminosity @xcite .
( j2000 ) & 08 39 50.62 & 07 09 14.09 + decl .
( j2000 ) & 12 14 33.9 & 36 01 21.8 + redshift & 0.198 & 0.111 + _ suzaku _ observation i d & 705007010 & 706008010 + nominal position & hxd & xis + start time ( ut ) & 2010 - 05 - 08t16:32:34 & 2011 - 04 - 02t02:16:10 + end time ( ut ) & 2010 - 05 - 10t11:45:10 & 2011 - 04 - 04t01:32:13 + exposure ( xis ) ( ks ) & 80.6 & 81.1 + exposure ( hxd / pin ) ( ks ) & 65.6 & 57.2 in this paper , we study the spectra of two very luminous radio - loud agns , 3c 206 and pks 070735 , observed with _
suzaku_and _ swift_/bat to reveal their nuclear structure .
the main purpose is to establish if the correlation of the torus opening - angle with luminosity seen in radio - quite agns also holds for radio - loud agns . for quantitative discussion , we utilize the numerical torus model by @xcite to predict the equivalent width of the iron - k line as a function of torus opening angle and column density . our targets , 3c 206 and pks 070735 , are the most luminous unobscured ( type-1 ) and obscured ( type-2 ) radio - loud agns having hard x - ray luminosities of @xmath18 erg s@xmath1 and @xmath2 erg s@xmath1 , respectively , among non - blazar type agns with fluxes above @xmath19 erg@xmath10s@xmath1(14195 kev ) detected in the _
swift_/bat 58-month survey @xcite .
3c 206 ( @xmath20 = 0.1979 ; * ? ? ?
* ) is a radio - loud quasar , classified as seyfert 1.2 from the optical spectrum @xcite .
the source is located in a cluster of galaxies with the abell richness class 1 @xcite .
the observation with the _ einstein _ high resolution imager ( hri ) detected no significant extended x - ray emission within @xmath3 2 arcmins from the agn @xcite .
the host galaxy of 3c 206 has an elliptical morphology .
the black hole mass is estimated to be @xmath21 = 10@xmath22 @xcite from the empirical formula using the continuum luminosity at 5100 and the line widths of h@xmath17 and h@xmath23 .
pks 070735 ( @xmath20 = 0.1108 ; * ? ? ?
* ) is a radio galaxy of fanaroff - riley type ii .
this object has a largely extended radio structure up to @xmath3500 .
the galaxy morphology and black hole mass of pks 0707 - 35 are not known .
we classify this source as an obscured ( type-2 ) object on the basis of the x - ray spectrum , which shows an intrinsic absorption with a hydrogen column density of @xmath24@xmath25 @xmath10(section [ pks ] ) .
this paper is organized as follows .
section [ obs ] summarizes the observations . in section [ ana ] , we present the data analysis and the results including the light curves of _ suzaku_and detailed model fit to the broad - band spectra obtained with _ suzaku_and _
summary and discussion are given in section [ discussion ] .
the cosmological parameters ( @xmath26 , @xmath27 , @xmath28)= ( 70 km s@xmath1 mpc@xmath1 , 0.3 , 0.7 ; @xcite ) are adopted in calculating the luminosities .
the errors on the spectral parameters correspond to the 90% confidence limits for a single parameter .
we observed the two radio - loud agns 3c 206 and pks 070735 with _
suzaku_on 2010 may 8 - 10 and 2011 april 2 - 4 , respectively , for a net exposure of @xmath380 ksec .
the hard x - ray fluxes in the 14195 kev band averaged from 2005 to 2010 @xcite are @xmath29 erg@xmath10s@xmath1 ( 3c 206 ) and @xmath30 erg@xmath10s@xmath1(pks 070735 ) , which are sufficiently bright for detailed broad - band spectral analysis .
the basic information of these targets and observation logs are summarized in table [ targets ] . _ suzaku _ , the fifth japanese x - ray satellite @xcite , carries three currently active x - ray ccd cameras called the x - ray imaging spectrometer ( two front - side illuminated ones ( xis - fis ) , xis-0 , 3 , and one back - side illuminated one ( bi - xis ) , xis-1 ) and the hard x - ray detector ( hxd ) composed of si pin photodiodes and gadolinium silicon oxide ( gso ) scintillation counters .
we utilize the data of the xis - fis , xis - bi , and hxd / pin in the energy bands of 112 kev , 0.58 kev , and 1635 kev ( pks 070735 only ) , respectively , where the source detection is not affected by the background uncertainties and the calibration is reliable .
we do not use the hxd / pin data of 3c 206 because of the unusually high count rate of the non x - ray background ( nxb ) that prevented a firm detection of the source signal .
the data of hxd / gso , covering above 50 kev , are not utilized , because the fluxes of both targets are too faint to be detected .
3c 206 and pks 070735 were focused on the nominal center position of the hxd and the xis , respectively . for spectral analysis
, we also include the _ swift_/bat spectra covering the 14195 kev band averaged over 58-months from 2005 to 2010 @xcite .
we reduce the unfiltered event files processed with the versions of 2.5.16.28 ( 3c 206 ) and 2.5.16.29 ( pks 070735 ) . to reflect the latest energy calibration ,
we apply the _ xispi _ ftool with the caldb files ae_xi\{0,1,3}_makepi_20110621.fits .
these events files are then screened with standard selection criteria described in the _
suzaku_abc guide . to extract light curves and spectra
, we use the ftools package ( heasoft version 6.11 ) .
spectral fitting is performed with the xspec software ( version 12.7.0o ) .
the xis events of each target are extracted from a circular region centered on the source peak with a radius of 3 , and the background is taken from a source free region within the xis field of view .
we generate the rmf ( redistribution matrix file ) and arf ( ancillary response file ) of the xis with the _ xisrmfgen _ and _ xissimarfgen _ ftools @xcite , respectively .
the `` tuned '' nxb event files provided by the hxd team are utilized for the background subtraction of the hxd / pin data .
the simulated cosmic x - ray background ( cxb ) is added to the nxb spectrum based on the formula given by @xcite .
we use the hxd / pin response file ae_hxd_pinxinome9_20100731.rsp .
the spectra of _ swift_/bat and response matrix are available from the heasarc web page of the 58-month survey .
figure [ fig_lc_soft ] ( a ) shows the light curve of 3c 206 in the 0.52 kev band obtained from xis-1 ( upper ) , that in the 210 kev band combined from xis-0 and xis-3 ( middle ) , and the hardness ratio between the two bands ( ratio1 ) .
figure [ fig_lc_soft ] ( b ) represents those of pks 070735 , together with the hxd / pin light curve in the 1635 kev band .
the hardness ratio between the 210 kev and 1635 kev bands ( ratio2 ) is also plotted .
the bin width is 5760 s , the orbital period of _ suzaku _ , chosen to eliminate any systematics caused by the orbital change of the satellite .
the background is subtracted . as noticed from the figures , we find no significant time variability in ratio1 and ratio2 .
hence , we analyze the time averaged spectra over the whole _
suzaku_observations for both objects . in the spectral analysis
, we simultaneously fit the _ suzaku_and _ swift_/bat spectra .
the xis data in the 1.71.9 kev band are excluded because of calibration uncertainties associated with the instrumental si - k edge .
figure [ spectra ] shows the spectra of 3c 206 and pks 070735 folded with the detector response .
the relative flux normalization of the xis - bi to the xis - fi is set to be a free parameter , while that of the hxd / pin to the xis - fis is fixed at 1.16 for pks 070735 ( xis nominal position ) , based on the calibration using the crab nebula @xcite . the galactic absorption , @xmath31 , is always included in the spectral model with * phabs * in the xspec terminology , which is fixed at the value derived from the h i map of @xcite ( @xmath31= @xmath32 @xmath10for 3c 206 and @xmath33 @xmath10for pks 070735 ) . solar abundances by @xcite are assumed throughout our analysis .
we model the emission from the nucleus ( other than the jet component ) with a power law with exponential cutoff ( @xmath34 ) and compton reflection components from cold matter .
unlike most of previous works , in this paper we consider two reflection components , one from the accretion disk and the other from the torus , in order to achieve a physically consistent picture . since the direct emission is subject to time variability between the _ suzaku_and _
swift_/bat observations , we make the relative normalization of the direct component and its disk reflection ( @xmath35 ) free between the two spectra .
since we do not anticipate significant time variability in the reflection component from the torus , which usually has a size of an order of @xmath30.1 pc or larger @xcite , the relative normalization of the torus reflection is tied between the _ suzaku_and _ swift_/bat observations .
in fact , when we apply the formula by @xcite , the torus distances of 3c 206 and pks 070735 are estimated to be @xmath36 pc by assuming a bolometric correction factor of @xmath37 for the disk luminosity .
for both targets , the cutoff energy is fixed at 200 kev , which is suggested from several radio - loud agns ( e.g. , * ? ? ?
* ; * ? ? ?
the averaged cutoff energy for local seyfert galaxies is reported to be @xmath3300 kev by @xcite , which is somewhat higher than that of radio - loud agns .
we confirm that the results are little affected even if @xmath38 300 kev is adopted instead .
to calculate the reflection components , we adopt the * pexmon * model @xcite in xspec , where the fluorescence lines of iron - k@xmath17 , iron - k@xmath23 , and nickel - k@xmath17 are self - consistently included in the reflected continuum computed by the * pexrav * code @xcite .
* pexrav * model assumes a neutral , compton - thick ( @xmath24@xmath39 @xmath10 ) medium with a plane geometry .
the relative strength to the direct component is defined by @xmath40 @xmath41 , where @xmath42 is the solid angle of the reflector viewed from the central irradiating source .
we consider the doppler and relativistic smearing in the disk - reflection component , which is thus modeled as * rdblurpexmon * in the xspec terminology .
since it is difficult to constrain the parameters of * rdblur * from our spectra , we assume the inner and outer radii of the accretion disk to be @xmath43 ( @xmath44 is the gravitational radius ) and @xmath45 , respectively , with a typical emissivity law of @xmath46 . the adopted inner radius can be regarded as a typical value observed from radio galaxies @xcite , although it is yet uncertain .
the inclination angles between the normal axis of the accretion disk and the line of sight are assumed to be @xmath47 for 3c 206 ( type-1 ) and @xmath48 for pks 070735 ( type-2 ) .
even when we adopt an inclination angle of @xmath49 for 3c 206 ( the lowest available value in the * pexrav*/*pexmon * model ) , the result of broad - band fitting is essentially unaffected .
such a low inclination is unlikely in our case , however , because the predicted jet emission largely overpredicts the observed x - ray flux ( see section [ 3c206_ana ] for the details ) . as for the reflection component from the torus
, we ignore any smearing effects , which can not be resolved at the ccd energy resolution .
it is described as a single * pexmon * model , whose inclination angle is fixed at @xmath48 for both targets , as a representable value for the complicated torus geometry . to evaluate possible contamination of x - ray emission from the galaxy cluster around the agn
, we compare the radial profile of 3c 206 obtained from the xis-0 and xis-1 images in the 0.52 kev band with that of mrk 205 , an agn not located in a cluster , observed in 2010 may at the hxd nominal position ( _ suzaku_observation i d : 705062010 ) .
the results are plotted in figure [ radialprofile ] .
there is no evidence for diffuse emission from the cluster of galaxies surrounding 3c 206 , which could be extended over @xmath3 600 kpc ( @xmath3 3 at z = 0.1979 ) .
this is consistent with the result by @xcite using _
einstein_. we thus analyze the spectra of 3c 206 by neglecting any emission from the cluster of galaxies . for the spectral fit of 3c 206 , we utilize the data of xis - fis in the 112 kev band , xis - bi in the 18 kev band , and _ swift_/bat in the 14195 kev band ; because the xis spectra show a soft excess below @xmath501 kev over a power law extrapolated from higher energies , we do not include the xis - bi spectrum below 1 kev in the broad - band spectral analysis to avoid complexity .
possible origins of the soft excess will be discussed later in this subsubsection .
first , we apply a simple model consisting of a direct continuum and its reflection component only from the torus , expressed as * `` phabs ( constant zpowerlwzhighect + pexmon ) '' * in the xspec terminology . here
the `` constant '' factor is introduced to represent the time variability between the _
suzaku_and bat observations .
we obtain the best - fit photon index of @xmath51 and reflection strength from the torus of @xmath52 relative to the time averaged flux of the direct component determined by _
swift_/bat with @xmath53 .
next , we adopt a model expressed as * `` phabs ( constant ( zpowerlwzhighect + rdblurpexmon ) + pexmon ) '' * , adding the reflection component from the accretion disk to the previous model to achieve a physically consistent picture , even though the previous fit without the disk reflection is acceptable in terms of the chi - squared value . since it is difficult to determine the strengths of the two reflection components independently from our data , we fix only that of the disk component .
according to the calculation by @xcite , the equivalent width of the iron - k line is predicted to be ew@xmath54 = 1731 ev for an accretion disk truncated at @xmath55 for a power - law photon index of 1.71.9 .
here we assume an inclination angle of @xmath47 and scale height of the irradiating source of 10 @xmath56 .
we then estimate the corresponding reflection strength of @xmath57 in the * pexmon * model so that the predicted equivalent width ( 1731 ev ) is consistently reproduced .
thus , we fix @xmath58 for the reflection component from the accretion disk .
we find that the _
suzaku_and bat spectra of 3c 206 favor this model with @xmath59 ( @xmath60 ) over the single reflection model , although the difference in the chi - squared value , @xmath61 , is not highly significant .
we obtain the best - fit photon index of @xmath51 and reflection strength from the torus of @xmath62 relative to the time averaged flux of the direct component determined by _ swift_/bat .
emission from the jets may be contained in the x - ray spectra of radio - loud agns . to estimate the jet contribution in the x - ray band , we utilize the radio luminosity in the agn core , which is attributable to the emission from the innermost region of the jet . according to the unified scheme of agns @xcite ,
radio - loud quasars and radio galaxies are intrinsically the same populations of blazars except that the jet axis is not aligned with the line - of - sight .
thus , from the observed radio luminosity , we can estimate the broad band spectral energy distribution ( sed ) of the jet emission in ( non - blazar ) radio - loud agns , by referring to averaged seds of blazars given as a function of radio luminosity ( the so - called blazar sequence ) after correcting for relativistic beaming .
the observed sed from an emitter moving with a relativistic velocity of @xmath63 shifts toward higher frequencies by a beaming factor of @xmath64 ( @xmath65 , where @xmath66 , @xmath67 , and @xmath68 is the angular separation from the line - of - sight ) , and toward higher luminosities by @xmath69 ( section 5.8.1.1 in * ? ? ?
* ) . by assuming the inclination angle of @xmath70 and the jet velocity of @xmath71 for type-1 radio - loud agns , the beaming factor of 3c 206
is estimated to be @xmath72 = 0.52 .
typically , @xmath73 in blazars @xcite .
then , we can convert the sed of typical blazars into that of 3c 206 by multiplying a factor of @xmath74 and @xmath75 in the frequency and luminosity , respectively .
using the blazar seds given in @xcite and the radio luminosity of 3c 206 ( @xmath76 ergs@xmath1at 5 ghz ; * ? ? ?
* ) , we find that the intrinsic sed of the jet emission in 3c 206 correspond to those of the 3 most - luminous blazar classes in figure 10 of @xcite .
we estimate that the observed x - ray luminosity from the jet in 3c 206 should be @xmath77 ergs@xmath1at 1 kev , which is @xmath3120% of the total observed flux , and the spectrum is approximated by a power law with a photon index of 1.21.4 in the x - ray band below 100 kev .
note that when we assume an inclination angle of @xmath78 ( @xmath72 = 1.72 ) , the jet luminosity is estimated to be @xmath79 ergs@xmath1at 1 kev , which is @xmath80 times higher than the total observed flux . to take into account the jet emission ,
we add a power law to the spectral model described above . considering a large uncertainty in the predicted luminosity of the jet component
, however , we fix the photon index at 1.3 but set the normalization to be a free parameter . for simplicity
, we also assume that the flux ratio between the jet ( power law ) and the direct component from the disk ( cutoff power law ) did not change between the _ suzaku_and _ swift_/bat observations .
the final model , which is shown in figure [ fig_3c206_spec ] , is expressed as * `` phabs ( constant ( zpowerlwzhighect + rdblurpexmon + zpowerlw ) + pexmon ) '' * , where the constant factor takes account of flux variation between the two observations , and the first to fourth terms represent ( 1 ) the direct component , ( 2 ) the reflection from the accretion disk , ( 3 ) the jet component and ( 4 ) that from the torus , respectively .
we confirm that this model including the jet emission well reproduces the broad band spectrum in the 1200 kev band ( @xmath81/dof = 818.0/824 ) . besides the requirement from the consistency with the radio core luminosity
, we confirm that the improvement of the fit by adding the jet component is marginally significant with an _
f_-test probability of 99.3% .
it is found that the fraction of the jet flux in the total one is @xmath82 in the 210 kev band and @xmath83 in the 14195 kev band , respectively .
this result is consistent with our rough estimate using the radio luminosity and blazar sequence . as for the non - jet emission
, we obtain the best - fit photon index of @xmath84 and reflection strength from the torus of @xmath85 .
the best - fit parameters are summarized in table [ para1 ] .
although the improvement from the previous fit without the jet component is only marginally significant ( at 99.3% confidence level from the f - test ) , we regard this model as the most realistic description of the broad band data of 3c 206 , considering that the jet contribution in the x - ray spectrum is quite likely .
we check possible systematic errors caused by our assumption on the reflection component from the accretion disk , which is quite uncertain .
if we assume 10 @xmath56 as the inner radius of the accretion disk instead of @xmath86 , we expect a stronger reflection of @xmath87 , which predicts ew@xmath88 100130 ev , based on the calculation by @xcite .
adopting @xmath89 and @xmath90 in the above model ( including the jet component ) yields the best - fit photon index of @xmath91 and reflection strength from the torus of @xmath92 with @xmath81/dof = 822.7/824 .
the fit becomes considerably worse but these parameters are still consistent with those obtained in the case of @xmath93 and @xmath94 within the statistical errors at the 90% confidence level .
cccc ( 1 ) & @xmath31(@xmath95 @xmath10 ) & 0.0568@xmath96 & 0.170@xmath96 + ( 2 ) & norm@xmath97 & @xmath98 & @xmath99 + ( 3 ) & norm@xmath100 & & 1.16@xmath96 + ( 4 ) & norm@xmath101 & @xmath102 & @xmath103 + ( 5 ) & @xmath104 ( photon @xmath10s@xmath1 kev@xmath1 ) & @xmath105 & @xmath106 + ( 6 ) & @xmath107 ( photon @xmath10s@xmath1 kev@xmath1 ) & @xmath108 & + ( 7 ) & @xmath109(@xmath95 @xmath10 ) & & @xmath110 + ( 8) & @xmath111(@xmath95 @xmath10 ) & & @xmath112 + ( 9 ) & @xmath113 & & @xmath114 + ( 10 ) & @xmath115 & @xmath116 & @xmath117 + ( 11 ) & @xmath118 ( kev ) & 200@xmath96 & 200@xmath96 + ( 12 ) & @xmath119 & ( 0.1@xmath96 ) & ( 0.2@xmath96 ) + ( 13 ) & @xmath120 ( @xmath56 ) & 100@xmath96 & 100@xmath96 + ( 14 ) & @xmath121 & @xmath122 ( @xmath123 ) & @xmath5 ( @xmath124 ) + ( 15 ) & @xmath125 ( % ) & & @xmath126 ( @xmath127 ) + ( 16 ) & @xmath128 & 1.3@xmath96 & + ( 17 ) & @xmath129 ) & @xmath130 ( @xmath131 ) & @xmath132 + ( 18 ) & @xmath133 ) & @xmath134 ( @xmath135 ) & @xmath136 + ( 19 ) & @xmath137 ) & @xmath138 ( @xmath139 ) & @xmath140 + & @xmath81/dof & 818.02/824 & 616.31/582 as mentioned above , we detect the soft excess below 1 kev in the xis - bi spectrum when the above model is extrapolated toward lower energies ; when the xis - bi spectrum in the 0.61 kev band is included , the goodness of the fit becomes rather poor with @xmath81/dof = 989.6/926 .
similar features are observed in the x - ray spectra of many seyfert 1 galaxies .
first , we consider possible contribution of optically - thin thermal emission from a gas unrelated to the agn . by adding the * apec * model in the xspec terminology ,
the fit is significantly improved ( @xmath81/dof = 911.1/924 ) with a temperature of @xmath141 kev and a luminosity of @xmath142 ergs@xmath1 . however
, this luminosity is too high as that from star forming regions in the host galaxy of agns ( e.g. , * ? ? ?
as mentioned in the previous section , this could not be the hot gas in the galaxy cluster , either .
thus , we conclude that this model is unrealistic , even though it is statistically acceptable .
a possible origin is a thermal comptonization component of soft seed - photons from the disk ( e.g. , * ? ? ?
* ) , although the physical picture is not yet established .
we find that the * comptt * model @xcite successfully represents the soft excess ( @xmath81/dof = 902.6/923 corresponding to an f - test probability of @xmath143 99.99% ) with the best - fit electron temperature of @xmath144 kev and optical depth of @xmath145 for black body seed photons of 0.01 kev .
the temperature of this comptonization corona is consistent with the result from the seyfert 1 galaxy mcg 63015 obtained by @xcite , while the optical depth is smaller .
we need to be aware of the degeneracy between the temperature and optical depth , however .
the observed flux of this component is @xmath146 erg@xmath10s@xmath1 in the 0.51 kev band .
an alternative explanation of the soft excess is reflection from a mildly ionized disk , which emits strong iron - l lines below 1 kev ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
accordingly , we utilize the * reflionx * model @xcite convolved with * rdblur * @xcite , instead of `` cold '' reflection modeled as * rdblurpexmon*. since such an ionized disk region is expected to be located close to the smbh , we fix @xmath147 , the innermost stable circular orbit ( isco ) of a non - rotating black hole .
this model also reproduces the soft excess of our spectra with the best - fit ionization parameter @xmath148 erg cm@xmath1 s@xmath1 , which predicts prominent iron - l line features below 1 kev ( e.g. , * ? ? ?
we obtain @xmath81/dof = 911.0/923 with an f - test probability @xmath143 99.99% from the model without the soft - excess component .
the reflection strength roughly corresponds to @xmath150 as estimated from its continuum flux above 10 kev . while we can not distinguish which is the more realistic model explaining the soft - excess , our results on the reflection from the torus are not affected . in the spectral analysis of pks
070735 , we use the data of the xis - fis in the 1.012 kev band , xis - bi in the 0.68 kev band , hxd / pin in the 1635 kev band , and _ swift_/bat in the 14100 kev band .
_ swift_/bat did not detect positive signals above 100 kev from this source .
as described below , the jet component is negligible in the x - ray spectrum of pks 070735 .
we find the continuum to be subject to an intrinsic absorption with @xmath24 @xmath151 @xmath10 , on which we base our classification of this source as an obscured agn .
previous studies ( e.g. , * ? ? ?
* ) showed that a scattered ( or photoionized ) component of the direct emission by a gas surrounding the nucleus is almost always observed in the x - ray spectra of absorbed agns , including radio - loud ones ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we thus essentially adopt the same model components as applied to 3c 206 except that ( 1 ) no jet component is included , ( 2 ) intrinsic neutral absorption is applied to the model components , and ( 3 ) a scattered component is added with the same photon index as that of the transmitted one . regarding the third component , we do not find significant improvement of the fit by adding emission lines expected from a photoionized plasma to the power law model due to limited statistics and energy resolution of the data .
we confirm the absence of a significant jet flux in the x - ray spectrum of pks 070735 based on the same discussion as for 3c 206 .
the core radio - flux of pks 070735 is measured to be 25 mjy at 2.3 ghz by @xcite .
assuming the inclination angle of @xmath152 , which is larger than the case of 3c 206 because it is a type-2 agn , and the jet velocity of @xmath153 , we estimate the beaming factor of the jet in pks 070735 to be @xmath64 = 0.11 . by comparison with the seds of blazars @xcite
, we estimate that the fraction of the jet flux is less than @xmath154 0.3% of the total flux at 1 kev corrected for the absorption . as in the spectral analysis of 3c 206
, we first consider the reflection component only from the torus , adopting a model expressed as * `` phabs ( constantzphabszpowerlwzhighect + zphabspexmon + @xmath125zpowerlwzhighect ) '' * in the xspec terminology .
these terms correspond to the direct component , the reflection component from the torus , and the scattered component , respectively .
the constant factor represents the flux variation between the _
suzaku_and bat observations .
we assume that the fluxes of the scattered and torus - reflection components did not vary between the two observations . since it is difficult to constrain the intrinsic absorption separately for the direct and reflection components that are emitted from different regions , we adopt a same @xmath24 value for them .
we find that the model can reproduce the broad band spectra of pks 070735 reasonably well ( @xmath81/dof = 631.0/584 ) with @xmath155 , @xmath24@xmath156 @xmath10 , @xmath157 , and @xmath158 ( the latter two parameters are defined as those relative to the time - averaged flux of the direct component as measured with _
swift_/bat ) .
next , we add a reflection component from the accretion disk . the model is thus expressed as * `` phabs ( constantzphabs ( zpowerlwzhighect + rdblurpexmon ) + zphabspexmon + @xmath125zpowerlwzhighect ) '' * in the xspec terminology , where the second term corresponds to the disk - reflection component .
we estimate the strength of the reflection component from the accretion disk by assuming the inner radius of @xmath55 and the inclination angle of @xmath48 ( instead of @xmath47 as assumed for 3c 206 ) . for a photon index of 1.51.8
, we find @xmath159 0.2 and ew@xmath54 = 1326 ev from @xcite , and hence we fix @xmath160 .
we find that the fit is better ( @xmath81/dof = 629.2/584 ) than the previous one with the same degree of freedom , although the significance of the improvement ( @xmath161 ) is weaker than the case of 3c 206 .
as expected , we obtain a weaker reflection strength from the torus , @xmath162 , while the other parameters are little changed .
since the above fit is not statistically acceptable yet , we then adopt a partial covering model instead of a single absorption , where two different absorbing column densities are considered with fractions @xmath113 and @xmath163 .
this model often better reproduces the x - ray spectra of type-2 agns ( see e.g. , * ? ? ?
. the model , which is shown in figure [ fig_pks0707_spec ] , is thus expressed as * phabs ( constantzphabszpcfabs ( zpowerlwzhighect + rdblurpexmon ) + zphabszpcfabspexmon + @xmath125 zpowerlwzhighect)"*. we find the fit is significantly improved ( @xmath164 ) with an f - test probability of @xmath165 .
the parameters of the partial covering are determined as @xmath109@xmath166 @xmath10(41@xmath167% ) and @xmath111@xmath168 @xmath10(59@xmath169% ) .
we obtain the best - fit photon index of @xmath170 , the reflection strength from the torus of @xmath171 ( to the bat flux ) , and the scattering fraction of @xmath172 ( to the bat flux ) .
the best - fit parameters are summarized in table [ para1 ] .
we regard this model as the best description of the spectrum of pks 070735 .
finally , we also check systematic uncertainties related to the assumption of the reflection component from the accretion disk . if we assume the inner radius of the accretion disk to be 10 @xmath56 , the calculation by @xcite predicts @xmath173 and ew@xmath88 80110 ev for an inclination of 60@xmath174 and a photon index of 1.51.8 .
adopting @xmath175 and @xmath89 , we obtain the following best - fit parameters with @xmath81/dof = 614.6/582 ; @xmath176 , @xmath177 ( to the bat flux ) , @xmath172 ( to the bat flux ) , @xmath109@xmath178 @xmath10(49@xmath179% ) , and @xmath111@xmath180 @xmath10(51@xmath181% ) .
they are all consistent with the case of @xmath182 within the statistical errors at the 90% confidence level .
cccc ( 1 ) & @xmath183 ( kev ) & 6.4@xmath96 & 6.4@xmath96 + ( 2 ) & @xmath184 ( ev ) & 1@xmath96 & 1@xmath96 + ( 3 ) & @xmath185 ( kev ) & 6.4@xmath96 & 6.4@xmath96 + ( 4 ) & @xmath120 ( @xmath56 ) & 100@xmath96 & 100@xmath96 + ( 5 ) & @xmath186 ( photon @xmath10 s@xmath1 ) & ( @xmath187 ) & @xmath188 + ( 6 ) & ew@xmath189 ( ev ) & ( @xmath190 ) & @xmath191 + ( 7 ) & ew@xmath192 ( ev ) & ( @xmath193 ) & @xmath13 + ( 8) & @xmath194 ( photon @xmath10 s@xmath1 ) & ( @xmath195 ) & ( @xmath196 ) + ( 9 ) & ew@xmath197 ( ev ) & ( @xmath198 ) & ( @xmath199 ) + ( 10 ) & ew@xmath200 ( ev ) & ( @xmath193 ) & ( @xmath199 ) + & @xmath81/dof & 312.50/304 & 389.33/366 to confirm the results on the reflection components obtained from the analysis of the broad - band spectra , we also analyze the narrow - band spectra of the xis - fis and xis - bi in the 39 kev and 38 kev bands , respectively , focusing on the fluorescence iron - k@xmath17 line .
figure [ fe_lines ] plots the spectra of the xis - fis .
we model the continuum by the same model from which iron - k emission lines are excluded by replacing the * pexmon * model with * pexrav*. a narrow * gaussian * and a * diskline * component @xcite are added , representing iron - k line emission from the torus and accretion disk , respectively . although iron - k lines from radio - loud agns are usually narrow ( e.g. , * ? ? ?
* ; * ? ? ?
* ) and hence are likely produced by the torus , here we aim to constrain any possible contribution from the accretion disk directly from our data .
both line energies are fixed at 6.4 kev in the rest frame . in the * diskline * model ,
the inner and outer radii are fixed at @xmath120 = @xmath55 and @xmath201 , respectively , with an emissivity law of @xmath46 .
the parameters of the continuum are fixed at those in table [ para1 ] except for the normalization .
the fitting results are summarized in table [ ew ] .
data to continuum model ratio is shown in the lower panel of figure [ fe_lines ] .
we find that the observed equivalent widths of the * diskline * component are @xmath7 53 ev ( 3c 206 ) and @xmath7 68 ev ( pks 070735 ) , and those of the narrow line are @xmath7 30 ev ( 3c 206 ) and @xmath191 ev ( pks 070735 ) , respectively . to discuss the physical origin of these lines ,
we need to correct the observed equivalent widths for the contribution of the jet in 3c 206 .
the jet flux accounts for @xmath202 of the total flux at 6.4 kev in the rest frame .
we thus obtain the equivalent widths of the broad line ew@xmath203 ev and the narrow line ew@xmath204 ev for 3c 206 .
furthermore , for the line originating from the torus , it is necessary to take into account the time variability of the direct component .
the equivalent widths of the narrow iron - k line with respect to the 5 year averaged bat flux are found to be ew@xmath205ev and @xmath13 ev for 3c 206 and pks 070735 , respectively , which can be directly compared with theoretical predictions described in sections [ 3c206_ana ] and [ pks ] .
these results are consistent with those obtained from the broad - band spectra in both targets . when we assume the inner radius of the accretion disk to be 10 @xmath206 , the equivalent widths corrected for the jet contribution and time variability become ew@xmath207 ev and @xmath208 ev and ew@xmath209 ev and @xmath210 ev for 3c 206 and pks 070735 , respectively
. these values are still consistent with those derived for the case of @xmath211 . however , the equivalent width of the diskline component in 3c 206 is smaller than the predicted value of 100130 ev by @xcite ( see section [ 3c206_ana ] ) .
recalling the fact the model with @xmath211 better reproduces the broad - band spectrum , we infer that the accretion disk of 3c 206 is likely to be truncated at a radius much larger than @xmath212 if the disk produces only cold reflection . assuming that the narrow iron - k@xmath17 line entirely originates from the torus
, we can constrain the solid angle and/or hydrogen column density of the torus from the line equivalent width . to quantify the relation between the predicted equivalent width of the iron - k line and torus geometry , we employ the numerical model by @xcite , where the reprocessed emission from the torus irradiated by the central source is calculated via monte carlo simulations .
@xcite consider a three - dimensional torus configuration , characterized by the half - opening angle @xmath6 , the inclination angle of the line - of - sight @xmath213 , the hydrogen column density in the equatorial plane @xmath8 , and the ratio between the inner radius @xmath120 and outer radius @xmath214 , which is fixed at 0.01 ( see figure 2 of * ? ? ?
a cutoff power law with a cutoff energy of 360 kev is assumed for the spectrum of the central source ; this choice little affects the predicted iron - k line intensity when compared with the case of 200 kev . for a photon index of @xmath215 and a column density of @xmath8@xmath216 or @xmath217 @xmath10
, we calculate the iron - k line equivalent widths for various combinations of @xmath213 and @xmath6 , using the spectral database available as a `` table model '' in xspec .
the results applicable for type-1 agns ( i.e. , @xmath218 ) and for type-2 agns ( i.e. , @xmath219 ) are displayed in figures [ fig_type1_oa_ew ] and [ fig_type2_oa_ew ] , respectively .
the inclination is fixed at @xmath220 8@xmath174 , 28@xmath174 or 48@xmath174 for type-1 agns and at @xmath220 88@xmath174 for type-2 agns .
value for type-2 agn little change the result . ] as representative values , we assume @xmath8@xmath216 or @xmath217 @xmath10 in figures [ fig_type1_oa_ew ] , and @xmath8@xmath216 @xmath10 in figures [ fig_type2_oa_ew ] , the same order of the line - of - sight column density of pks 070735 . although the assumed value of @xmath8more sensitively affects the observed equivalent width in type-2 agns than in type-1 agns because of absorption of the continuum , the difference in the attenuation fraction of the direct component at 6.4 kev between the line - of - sight column density of @xmath221 @xmath10(as observed ) and @xmath222 @xmath10is found to be @xmath310% level , which we ignore in the following discussion for pks 070735 .
since the @xcite model is available only for @xmath223 , we extrapolate the result of @xmath224 toward @xmath225 in figure [ fig_type1_oa_ew ] , assuming that the line intensity is proportional to the volume of the torus .
the figures show the trend that the equivalent width decreases with increasing half - opening angle except for the case of @xmath8@xmath12 @xmath10 in figure [ fig_type1_oa_ew ] , which is consistent with figure 13 of @xcite . in figures [ fig_type1_oa_ew ] and [ fig_type2_oa_ew ]
, we also show the range of the `` corrected '' equivalent width of the narrow iron - k@xmath17 of 3c 206 and pks 070735 , respectively , with dashed lines .
we find that the torus of 3c 206 either has a relatively low column density in the equatorial plane @xmath8@xmath9 @xmath10 , or has a large opening angle ( @xmath226 ) if @xmath8@xmath12 @xmath10 .
even when we assume 0.5 times smaller elemental abundances for iron and nickel than the solar values , similar constraints are obtained ; @xmath8@xmath9 @xmath10 , or @xmath227 for @xmath8@xmath12 @xmath10 .
the result indicates that the torus of 3c 206 is far from compton thick or covers only a small solid angle .
the line - of - sight column densities of pks 070735 , @xmath24=
@xmath228 @xmath10 , suggest a small @xmath8 for this object , and we confirm that the iron - k@xmath17 equivalent width is consistent with the prediction from @xmath8@xmath229 @xmath10 as shown in figure [ fig_type2_oa_ew ] .
we have obtained the best quality broad - band x - ray spectra of the two most luminous radio - loud quasars 3c 206 ( type-1 ) and pks 070735 ( type-2 ) with _ suzaku_and _
swift_/bat , which have hard x - ray luminosities of @xmath0 erg s@xmath1 and @xmath2 erg s@xmath1 ( 14195 kev ) , respectively .
the spectra are well described with a physically motivated model consisting of the direct component ( with partial absorptions for pks 070735 ) , its reflection component ( including fluorescence lines ) from the accretion disk , that from the torus , jet emission ( 3c 206 ) , and scattered component ( pks 070735 ) . here
we properly take into account the effect of time variability of the direct component between the _ suzaku_and ( 5 year averaged ) _ swift_/bat observations .
our main findings are that ( 1 ) there is a significant contribution from the jet in the x - ray spectrum only for 3c 206 and that ( 2 ) the reflection strength from the torus is estimated to be quite small in both objects compared with typical seyfert galaxies , which is confirmed by the detailed analysis of the iron - k emission line .
we have shown that the x - ray spectrum of the `` type-1 '' radio - loud quasar 3c 206 is likely to contain significant emission from the jet , while its contribution is negligible for the `` type-2 '' radio quasar pks 070735 . in the unified scheme of agn ,
radio - loud quasars and radio galaxies are intrinsically the same as blazars except for the inclination angle with respect to the jet axis , which is the smallest in blazars and increases from type-1 to type-2 agns .
we show that it is possible to make a rough estimate of the contribution from the jet in the x - ray band using the radio core luminosity , by referring to the `` blazar sequence '' with correction of beaming factor based on reasonable assumptions .
our results imply that non - thermal jet emission may make a significant contribution to the observed x - ray spectrum of type-1 radio - loud agns in general , and this component should be taken into account when spectral modeling is done .
the broad band spectral fit suggests that the fraction of the jet component in the x - ray spectrum of 3c 206 is 26% ( 210 kev ) and 62% ( 14195 kev ) of the total flux , which is consistent with the estimate from the radio luminosity within a factor of 2 .
if this is the case , the very large hard x - ray luminosity of this source is partially due to the jet emission . the same trend that the relative jet power increases with energy
is also reported from 3c 273 @xcite .
the true luminosity arising from the accretion disk is estimated to be @xmath230 erg@xmath10s@xmath1(210 kev ) and @xmath231 erg@xmath10s@xmath1(14195 kev ) , which is very similar to that of pks 070735 .
compton reflection components accompanied by fluorescence lines from matter around the nucleus give us valuable information on the geometry of the surroundings of an agn .
unlike in most of previous works , we separately consider two reflection components , one from the accretion disk and the other from the torus in order to achieve a physically consistent picture .
relativistic blurring is considered in the disk reflection , whose parameters are linked to the reflection strength in a physically consistent manner , based on the calculation by @xcite .
since the torus is located far away from the central smbh and ccd spectroscopy can not resolve the line width , we ignore any blurring for the latter .
it is hard to distinguish the two reflection components from the spectra alone , which would be strongly coupled with each other in the fitting .
thus , we assume the reflection strength from the accretion disk to be @xmath119 = 0.1 ( 3c 206 ) and 0.2 ( pks 070735 ) , and then derive the reflection strengths from the torus as a free parameter .
we obtain @xmath85 ( 3c 206 ) and @xmath5 ( pks 070735 ) from the broad - band spectral fitting .
these results are consistent with the equivalent widths of the iron - k@xmath17 line derived from narrow - band ( 39 kev ) spectral analysis .
since any possible broad iron - k line features look very weak in both objects , it is difficult to directly constrain the innermost radius of the accretion disk from the line profile .
we find that the broad - band fitting of the 3c 206 spectra becomes significantly worse if we assume @xmath232 , which predicts a much larger reflection strength from the disk , compared with the case of @xmath211 .
in addition , the iron - k line analysis in the narrow band yields an upper limit of the equivalent width of a broad iron - k line that is smaller than the prediction by @xcite in the case of @xmath233 .
these results imply that the accretion disk of 3c 206 is likely to be truncated at a radius much larger than @xmath233 as far as the disk is not significantly ionized .
in fact , if we make the disk inner radius a free parameter , by assuming the relation between @xmath120 and @xmath40 based on the equivalent width calculated by @xcite and its corresponding reflection strength in the * pexmon * model @xcite , then we obtain the best - fit inner radius of @xmath234 .
pictures of truncated disks are obtained from other type-1 radio - loud agns , 4c 50.55 ( @xmath235 ; * ? ? ?
* ) , 3c 111 ( @xmath236 ; * ? ? ?
* ) , 3c 390.3 ( @xmath237 ; * ? ? ?
* ) and 4c+74.26 ( @xmath238 ; * ? ? ?
by contrast , the disks of 3c 120 and 3c 382 seem to be extended down to @xmath239 @xcite and @xmath240 @xcite , respectively .
it is not clear if there is any systematic trend of disk truncation in radio - loud agns .
we note that the presence of a highly ionized inner disk extending to @xmath147 , or smaller if the black hole is rapidly spinning , can not be ruled out , however , which might explain the soft excess seen in 3c 206 .
our robust finding is that the reflection component from the torus containing a narrow iron - k line is weak in both 3c 206 and pks 070735 ( @xmath241 ) compared with the typical reflection strength observed in the x - ray spectra of seyfert galaxies ( @xmath242 ; * ? ? ?
utilizing the torus model by @xcite , we have quantified the relation between the half opening angle of a torus ( @xmath6 ) and the equivalent width of an iron - k line . the equivalent width with respect to the time - averaged
, non - jet continuum flux is @xmath7 71 ev for 3c 206 .
this result constrains that the column density in the equatorial plane @xmath8@xmath9 @xmath10 , or @xmath11 if @xmath8@xmath243 @xmath10 is assumed , which corresponds to a very small solid angle of @xmath244 .
also , a small value of @xmath8 , @xmath245 @xmath10 , is inferred for pks 070735 from the line - of - sight absorption column density , which can explain the iron - k line equivalent width .
thus , these luminous radio - loud agns have only poorly developed tori .
there is a possibility that the tori are geometrically thick but highly clumpy with a very small filling factor .
it is interesting to compare this result with those obtained from less luminous radio - loud agns . @xcite
analyze _ suzaku_and _ swift_/bat spectra of the radio galaxy ngc 612 , which has @xmath246 ergs@xmath1 , and obtain the equivalent width of a narrow iron - k line of @xmath247 ev .
they also find that the torus model by @xcite well represents the spectra with the best - fit parameters of @xmath8@xmath248 @xmath10and @xmath249 .
the opening angle of ngc 612 is smaller than the estimate for 3c 206 , @xmath11 , obtained when a similar value of @xmath8@xmath250 @xmath10is assumed .
two narrow line radio galaxies studied by , 3c 403 ( @xmath251 ergs@xmath1 ) and ic 5063 ( @xmath252 ergs@xmath1 ) , show narrow iron - k line with equivalent widths of @xmath253 ev and @xmath254 ev , respectively .
thus , the iron - k equivalent widths obtained from type-2 radio - loud agns with moderate luminosities are all larger than that of most luminous one pks 070735 ( @xmath255 ergs@xmath1 ) .
this can be partially explained by their larger line - of - sight column densities that attenuate the continuum fluxes at 6.4 kev : @xmath256 @xmath10(ngc 612 ) , @xmath257 @xmath10(3c 403 ) , and @xmath258 @xmath10(ic 5063 ) .
the comparison suggests that the torus solid angle and/or column density are smaller in more luminous radio - loud agns .
it is known that the fraction of absorbed agn decreases with x - ray luminosity above @xmath259 erg s@xmath1 from studies using mostly radio - quiet agn samples ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this fact indicates that the solid angle of the torus decreases with luminosity in radio - quiet agns .
a popular explanation is the receding torus model @xcite , where the strong radiation from the central engine affects the structure of the torus ; both the solid angle and column density could be reduced in luminous agns .
our results are consistent with the idea that the same picture holds for radio - loud agns .
this implies that the presence of powerful jets is not related to the structure of the torus , which is responsible for gas feeding into the disk as the boundary condition .
rather , some ( yet not established ) physics taking place in more inner parts of the disk must be involved to explain the difference between radio - loud and radio - quiet agns .
we thank the anonymous referee for his / her many useful comments .
this work was partly supported by the grant - in - aid for jsps fellows for young researchers ( f.t . ) and for scientific research 23540265 ( y.u . ) , and by the grant - in - aid for the global coe program `` the next generation of physics , spun from universality and emergence '' from the ministry of education , culture , sports , science and technology ( mext ) of japan . | we present the results from broad - band x - ray spectral analysis of 3c 206 and pks 070735 with _ suzaku _ and _ swift_/bat , two of the most luminous unobscured and obscured radio - loud active galactic nuclei with hard x - ray luminosities of @xmath0 ergs@xmath1 and @xmath2 ergs@xmath1(14195 kev ) , respectively . based on the radio core luminosity
, we estimate that the x - ray spectrum of 3c 206 contains a significant ( @xmath3 60% in the 14195 kev band ) contribution from the jet , while it is negligible in pks 070735 .
we can successfully model the spectra with the jet component ( for 3c 206 ) , the transmitted emission , and two reflection components from the torus and the accretion disk .
the reflection strengths from the torus are found to be @xmath4 and @xmath5 for 3c 206 and pks 070735 , respectively , which are smaller than those in typical seyfert galaxies . utilizing the torus model by ikeda et al .
( 2009 ) , we quantify the relation between the half opening angle of a torus ( @xmath6 ) and the equivalent width of an iron - k line .
the observed equivalent width of 3c 206 , @xmath7 71 ev , constrain the column density in the equatorial plane to @xmath8@xmath9 @xmath10 , or the half opening angle to @xmath11 if @xmath8@xmath12 @xmath10 is assumed .
that of pks 070735 , @xmath13 ev , is consistent with @xmath8@xmath14 @xmath10 .
our results suggest that the tori in luminous radio - loud agns are only poorly developed .
the trend is similar to that seen in radio - quiet agns , implying that the torus structure is not different between agns with jets and without jets . |
the dynamics of electronic transport in modern semiconductor devices can be described by the semiclassical boltzmann - poisson ( bp ) model @xmath0 @xmath1 @xmath2 is the probability density function ( pdf ) over phase space @xmath3 of a carrier in the @xmath4-th energy band in position @xmath5 , with crystal momentum @xmath6 at time @xmath7 .
the collision operators @xmath8 model @xmath4-th and @xmath9-th carrier recombinations , collisions with phonons or generation effects .
@xmath10 is the electric field , @xmath11 is the @xmath4-th energy band surface , the @xmath4-th charge density @xmath12 is the k - average of @xmath13 , and @xmath14 is the doping profile .
+ the bp model for electron transport on a single conduction energy band for electrons has the form + @xmath15 @xmath16 , \ , \vec{e } = - \nabla_{\vec{x } } v , \label{pois1band}\ ] ] with the quantum mechanical electron group velocity @xmath17 , and the electron density @xmath18 . + the semi - classical boltzmann description of electron transport in semiconductors is , for a truly 3-d device , an equation in six dimensions plus time when the device is not in steady state .
the heavy computational cost is the main reason why the bp system had been traditionally solved numerically by means of direct simulation monte carlo ( dsmc ) methods @xcite .
however , after the pioneer work @xcite , in recent years , deterministic solvers to the bp system were proposed in @xcite .
these methods provide accurate results which , in general , agree well with those obtained from monte carlo ( dsmc ) simulations , often at a fractional computational time .
moreover , these type of solvers can resolve transient details for the electron probability density function @xmath19 , which are difficult to compute with dsmc simulators .
+ the initial methods proposed in @xcite using weighted essentially non - oscillatory ( weno ) finite difference schemes to solve the boltzmann - poisson system , had the advantage that the scheme is relatively simple to code and very stable even on coarse meshes for solutions containing sharp gradient regions . however
, a disadvantage of the weno methods is that it requires smooth meshes to achieve high order accuracy , hence it is not very flexible for adaptive meshes .
+ motivated by the easy _ hp-_adaptivity and the simple communication pattern of the discontinuous galerkin ( dg ) methods for macroscopic ( fluid level ) models @xcite , it was proposed in @xcite to implement a dg solver to the full boltzmann equation , that is capable of capturing transients of the probability density function .
+ in the previous work @xcite , the first dg solver for ( [ be])-([pois ] ) was proposed , and some numerical calculations were shown for one and two - dimensional devices . in @xcite , the dg - ldg scheme for the boltzmann - poisson system was carefully formulated , and extensive numerical studies were performed to validate the calculations .
such scheme models electron transport along the conduction band for 1d diodes and 2d double gate mosfet devices with an analytic kane energy band model . + a dg method for full conduction bands bp models was proposed in @xcite , following the lines of the schemes in @xcite , generalizing the solver that uses the kane non - parabolic band and adapting it to treat the full energy band case .
a preliminary benchmark of numerical results shows that the direct evaluation of the dirac delta function can be avoided , and so an accurate high - order simulation with comparable computational cost to the analytic band cases is possible
. it would be more difficult or even unpractical to produce the full band computation with other transport scheme .
it is worth to notice that a high - order positivity - preserving dg scheme for linear vlasov - boltzmann transport equations , under the action of quadratically confined electrostatic potentials , independent of the electron distribution , has been developed in @xcite .
the authors there show that these dg schemes conserve mass and preserve the positivity of the solution without sacrificing accuracy .
in addition , the standard semi - discrete schemes were studied showing stability and error estimates .
+ the type of dg method discussed in this paper , as was done in @xcite , belongs to a class of finite element methods originally devised to solve hyperbolic conservation laws containing only first order spatial derivatives , e.g. @xcite . using a piecewise polynomial space for both the test and trial functions in the spatial variables , and coupled with explicit and nonlinearly stable high order runge - kutta time discretization ,
the dg method is a conservative scheme that has the advantage of flexibility for arbitrarily unstructured meshes , with a compact stencil , and with the ability to easily accommodate arbitrary _ hp-_adaptivity . for more details about dg scheme for convection dominated problems ,
we refer to the review paper @xcite , later generalized to the local dg ( ldg ) method to solve the convection diffusion equations @xcite and elliptic equations @xcite .
+ regarding boundary conditions ( bc ) , there are several kinds of bc for bp semiconductor models .
they vary according to the considered device and physical situation .
for example , in the case of electron transport along a single conduction band the following bc could arise : + charge neutrality boundary conditions , given by @xcite @xmath20 this bc is imposed in source and drain boundaries where electric currents enter or exit the device in order to obtain there the condition @xmath21 .
+ reflective bc happen in insulating boundaries , usually defined by a neumann boundary @xmath22 , of 2d and 3d devices .
in general , reflective bc can be formulated as the values of the pdf at the inflow boundary being dependent on the outflow boundary values @xmath23 @xmath24 @xmath25 outward unit normal .
the neumann outflow boundary is defined as @xmath26 specular reflection bc over the neumann inflow boundary is imposed by @xmath27 @xmath28 diffusive reflection is a known condition from the kinetic theory for gas dynamics , in which the distribution function at the inflow boundary is proportional to a maxwellian @xcite , @xcite @xmath29 @xmath30 mixed reflection bc models the reflection of the electrons from a rough boundary , giving the reflected wave a linear convex combination of a specular and a diffuse component , as in the following formula : @xmath31 @xmath32 is sometimes called specularity parameter .
it can either be constant or a function of the momentum @xmath33 .
for example , the work by soffer @xcite studies a statistical model for the electrical conduction in metals , and derives a specularity parameter @xmath33 which depends on the momentum .
it is given by @xmath34 where @xmath35 is the rms height of rough interface , and @xmath36 is th angle between the incident electron and the interface surface normal .
+ reflection bc is a widely studied topic in the context of the kinetic theory of gases modelled by boltzmann - equations .
however , there is few work existent related to the study of the effect of reflection bc , such as diffusive , specular , and mixed reflection , in the context of boltzmann models for electron transport in semiconductors . a non - exhaustive list of references where reflection bc are studied for boltzmann equations in the context of kinetic theory of gases would include the works of cercignani @xcite and sone @xcite , where the specular , diffusive , and mixed reflection bc are formulated for the boltzmann eq . for gases .
v. d. borman , s. yu .
krylov , a. v. chayanov @xcite study the nonequilibrium phenomena at a gas - solid interface .
the paper of brull , charrier , mieussens @xcite studies the gas - surface interaction at a nano - scale and the boundary conditions for the associated boltzmann equation .
struchtrup @xcite studies as well the maxwell boundary condition and velocity dependent accommodation coefficients in the context of gases mentioned . in the context of the physics of electron transport ,
soffer @xcite studies a statistical model for the electrical conduction in metals , and derives under certain assumptions , analytical formulas for the specularity parameter @xmath37 associated to this physical phenomena , which we mentioned above in ( [ sofferpformula ] ) . in the case of semiconductors ,
the reference book of markowich , ringhofer , & schmeiser @xcite discusses the mathematical definition of boundaries according to the physics of semiconductors , and according to the kind of bc to be imposed at those boundaries : dirichlet , neumann , inflow and outflow boundaries .
a work of particular importance for our paper is the one by cercignani , gamba , and levermore @xcite .
they study high field approximations to a boltzmann - poisson system and boundary conditions in a semiconductor .
the bp system for electrons in a semiconductor in the case of high fields and small devices is considered .
boundary conditions are proposed at the kinetic level that yield charge neutrality at ohmic contacts , which are dirichlet boundaries , and at insulating neumann boundaries .
both bc , either the one yielding charge neutrality at dirichlet boundaries , or the one rendering zero flux of electrons at the boundary , assume that the pdf is proportional to a ground state associated to an asymptotic expansion of a dimensionless boltzmann - poisson system
. then they study closures of moment equations and bc for both the pdf and for the moment closures .
jngel mentions in his semiconductors book @xcite the different kinds of reflection bc common on the kinetic theory of gases , specular , diffusive , and mixed reflection but no further study of these reflection bc in the context of semiconductors is pursued .
+ we intend to present in this paper a mathematical , numerical , and computational study of the effect of diffusive , specular , and mixed reflection bc in boltzmann - poisson models of electron transport in semiconductors , solved by means of discontinuous galerkin fem solvers .
we study the mathematical formulation of these reflection bc in the context of bp models for semiconductors , and derive equivalent numerical formulations of the diffusive and mixed reflection bc with non - constant @xmath33 , such that an equivalent numerical zero flux condition is satisfied pointwise at the insulating neumann boundaries at the numerical level .
we present preliminary numerical simulations for a 2d bulk silicon diode with an applied bias and reflective bc , comparing the effects of specular , diffusive , and mixed reflection in the simulations .
the kane energy band model is a dispersion relation between the conduction energy band @xmath39 ( measured from a local minimum ) and the norm of the electron wave vector @xmath40 , given by the analytical function ( @xmath41 is a constant parameter , @xmath42 is the electron reduced mass for si , and @xmath43 is planck s constant ) : @xmath44 for our preliminary numerical studies we will use a boltzmann - poisson model as in @xcite , in which the conduction energy band is assumed to be given by a kane model .
we use the following dimensionalized variables , with the related characteristic parameters : @xmath45 a transformed boltzmann transport equation is used as in @xcite as well , where the coordinates used to describe @xmath38 are : @xmath46 , the cosine of the polar angle , the azimuthal angle @xmath47 , and the dimensionless kane energy @xmath48 , which is assumed as the conduction energy band ( @xmath49 is boltzmann s constant , @xmath50 is the lattice temperature , and @xmath51 ) .
so : @xmath52 a new unknown function @xmath53 is used in the transformed boltzmann eq .
@xcite , which is proportional to the jacobian of the transformation and to the density of states ( up to a constant factor ) : @xmath54 where @xmath55 the transformed boltzmann transport equation for @xmath53 used in @xcite is : @xmath56 @xmath57 represent the 2d cartesian components of the electron velocity @xmath58 , in the coordinate system ( @xmath59 , @xmath46 , @xmath60 ) .
+ @xmath61 represent the transport in the phase space of the new momentum coordinates ( @xmath59 , @xmath46 , @xmath60 ) due to the self consistent electric field @xmath62 @xmath63 , \\ & = & \mbox { } - c_k \frac{2 \sqrt{w(1+{\alpha_k}w)}}{1 + 2 { \alpha_k}w } \ , \hat{e}_w \cdot \vec{e}(t , x , y ) \ , , \\
g_4 { ( \mathbf{\cdot } ) } & = & \mbox { } - c_k \frac{\sqrt{1-\mu^2}}{\sqrt{w(1+{\alpha_k}w ) } } \left [ \sqrt{1-\mu^2 } \ , e_x(t , x , y ) - \mu \cos{\varphi}\ , e_y(t , x , y ) \right ] \ , , \\ & = & \mbox { } - c_k \frac{\sqrt{1-\mu^2}}{\sqrt{w(1+{\alpha_k}w ) } } \ , \hat{e}_{\mu } \cdot \vec{e}(t , x , y ) \ , ,
\\ g_5 { ( \mathbf{\cdot } ) } & = & - c_k \frac { - \sin{\varphi}}{\sqrt{w(1+{\alpha_k}w ) } \sqrt{1-\mu^2 } } \ , e_y(t , x , y ) \\ & = & -c_k \frac{1}{\sqrt{w(1+{\alpha_k}w ) } \sqrt{1-\mu^2 } } \ , \hat{e}_{{\varphi } } \cdot \ , \vec{e}(t , x , y ) \ , , \end{aligned}\ ] ] with @xmath64 and @xmath65 the orthonormal vector basis in our momentum coordinate space .
+ the right hand side of ( [ eqphi ] ) is the collision operator ( having applied the dirac delta s due to electron - phonon scattering , which depend on the energy differences between transitions ) @xmath66 \right\ } \\[5pt ] & & \mbox { } - \phi(t , x , y , w,\mu,{\varphi } ) \ , 2 \pi \ , [ c_0 s(w ) + c_+ s(w - { \gamma } ) + c_- s(w + { \gamma } ) ] \ , , \end{aligned}\ ] ] with the dimensionless parameters @xmath67 the electron density is : @xmath68 where @xmath69 hence , the dimensionless poisson equation is @xmath70\ ] ] with @xmath71
the domain of the devices to be considered can be represented by means of a rectangular grid in both position and momentum space .
this rectangular grid , bidimensional in position space and tridimensional in momentum space , is defined in the following way : @xmath72 \times \left [ y_{j - { \frac{1}{2 } } } , \ , y_{j + { \frac{1}{2 } } } \right ] \times } _ { x_{ij } } \underbrace { \left [ w_{k - { \frac{1}{2 } } } , \ , w_{k + { \frac{1}{2 } } } \right ] \times \left [ \mu_{m - { \frac{1}{2 } } } , \ , \mu_{m + { \frac{1}{2 } } } \right ] \times \left [ { \varphi}_{n - { \frac{1}{2 } } } , \ , { \varphi}_{n + { \frac{1}{2 } } } \right ] } _ { k_{kmn}}\ ] ] where @xmath73 , @xmath74 , @xmath75 , @xmath76 , @xmath77 , @xmath78 @xmath79 the finite dimensional space used to approximate the functions is the space of piecewise continuous polynomials which are piecewise linear in @xmath80 and piecewise constant in @xmath81 @xmath82 @xmath83 is the set of tensor product polynomials , linear over the element + @xmath84 \times \left [ y_{j - { \frac{1}{2 } } } , \ , y_{j + { \frac{1}{2 } } } \right ] $ ] and constant over the element + @xmath85 \times\left [ \mu_{m - { \frac{1}{2 } } } , \ , \mu_{m + { \frac{1}{2 } } } \right ] \times\left [ { \varphi}_{n - { \frac{1}{2 } } } , \ , { \varphi}_{n + { \frac{1}{2 } } } \right]$ ] . + @xmath86 will denote the piecewise polynomial approximation of @xmath53 over elements @xmath87 , @xmath88 , \quad i = ( i , j , k , m , n ) . \\ \ ] ] under this approximation , the density on the cell @xmath89 \times [ y_{j - { \frac{1}{2 } } } , \
, y_{j + { \frac{1}{2 } } } ] $ ] is @xmath90 \delta w_{k } \delta \mu_{m } \delta \varphi_{n } \nonumber\\ & = & \quad \sum_{k=1}^{n_{w } } \sum_{m=1}^{n_{\mu } } \sum_{n=1}^{n_{\varphi } } t_{ijkmn } \delta w_{k } \delta \mu_{m } \delta \varphi_{n } \nonumber\\ & + & \left ( \sum_{k=1}^{n_{w } } \sum_{m=1}^{n_{\mu } } \sum_{n=1}^{n_{\varphi } } x_{ijkmn } \delta w_{k } \delta \mu_{m } \delta \varphi_{n } \right ) \frac{(x - x_{i})}{\delta x_{i}/2 } \nonumber\\ & + & \left ( \sum_{k=1}^{n_{w } } \sum_{m=1}^{n_{\mu } } \sum_{n=1}^{n_{\varphi } } y_{ijkmn } \delta w_{k } \delta \mu_{m } \delta \varphi_{n } \right ) \frac{(y - y_{j})}{\delta y_{j}/2 } \nonumber\end{aligned}\ ] ] the discontinuous galerkin formulation for the boltzmann equation ( [ eqphi ] ) is : find @xmath91 , s.t .
@xmath92 for any test function @xmath93 . in ( [ dgb ] ) ,
the boundary integrals are given by : @xmath94 @xmath95 @xmath96 @xmath97 @xmath98 where the upwind numerical fluxes @xmath99 are defined as @xmath100 the poisson equation ( [ pois ] ) is solved by the ldg method as in @xcite . by means of this scheme
we find a solution @xmath101 , where @xmath102 and @xmath103 , @xmath104 the set of linear polynomials on @xmath105 .
+ the following rk - dg algorithm for bp is a dynamic extension of the gummel iteration map . to evolve from time @xmath108 to time @xmath107 : + 1.- compute the electron density @xmath109 . + 2.- solve poisson eq . for the given @xmath109 by local dg , obtaining the potential @xmath110 and the electric field @xmath111 .
compute then the respective transport terms @xmath112 .
+ 3.- solve by dg the advection and collision part of the boltzmann equation .
a method of lines ( an ode system ) for the time dependent coefficients of @xmath86 ( degrees of freedom ) is obtained .
+ 4.-evolve ode system by runge - kutta from @xmath106 to @xmath107 . + ( if partial time step necessary , repeat step 1 to 3 as needed ) .
we will consider in this work 2d devices in position space , which need a 3d momentum description for kinetic equations modeling semiconductors .
for example , a common device of interest is a 2d double gate mosfet .
a schematic plot of it is given in figure [ mosfet ] .
the shadowed region denotes the oxide - silicon region , whereas the rest is the silicon region .
potential bias are applied at the source , drain , and gates .
the problem is symmetric about the x - axis . +
another possible 2d problem is the case of a bi - dimensional bulk silicon diode , for which the doping is constant all over the physical domain , and which would have just an applied potential ( bias ) between the source @xmath114 and the drain @xmath115 ( no gates ) , with insulating reflecting boundaries at @xmath116 and @xmath117 .
we consider in the following sections the different kinds of boundary conditions for 2d devices and their numerical implementation , either at @xmath5-boundaries or at @xmath118-boundaries .
the bc for poisson eq .
are imposed over the @xmath80-domain .
+ for example , for the case of a 2d double gated mosfet , dirichlet bc would be imposed to the potential @xmath119 , as we have three different applied potentials biases : + @xmath120 volts at the source @xmath114 , @xmath121 volts at the drain @xmath115 , @xmath122 volts at the gates .
homogeneous neumann bc would be imposed for the rest of the boundaries , that is , @xmath123 .
+ for the case of a 2d bulk silicon diode , we impose dirichlet bc for the difference of potential @xmath119 between source and drain : + @xmath120 volts at the source @xmath114 , @xmath121 volts at the drain @xmath115 . for the boundaries @xmath124 we impose homogeneous neumann bc too , that is , @xmath125 .
+ as in @xcite , at the source and drain contacts , we implement the charge neutrality boundary condition ( [ neutralchargebcmath ] ) .
ghost cells for @xmath126 and @xmath127 at the respective boundaries are used , implementing this bc numerically as below : @xmath128 and @xmath129
in the @xmath130-space , we only need to apply a cut - off boundary condition . + at @xmath131 , @xmath86
is made machine zero : @xmath132 no other boundary condition is necessary for @xmath118-boundaries , since analytically we have that : * at @xmath133 , @xmath134 .
* at @xmath135 , @xmath136 ; * at @xmath137 , @xmath138 , so , at those boundaries , the numerical flux always vanishes .
reflection boundary conditions can be expressed in the following form : @xmath139 such that the following pointwise zero flux condition is satisfied at reflecting boundaries : @xmath140 as in the paper @xcite of cercignani , gamba , and levermore , where the given bc at neumann boundaries at the kinetic level is such that the particle flow vanishes at these insulating boundaries .
+ for simplicity we write @xmath141 .
we will study three kinds of reflective boundary conditions : specular , diffusive , and mixed reflection .
the last one is a convex combination of the previous two , but the convexity parameter can be either constant or momentum dependant , @xmath33 .
we go over the math and numerics related to these conditions below .
it is clear that , at the analytical level , the specular reflection bc ( [ eq : defspecreflex ] ) satisfies the zero flux condition pointwise at reflecting boundaries , since @xmath142 specular reflection bc in our transformed boltzmann eq . for the new coordinate system
is mathematically formulated in our problem as @xmath143 to impose numerically specular reflection bc at @xmath124 in the dg method , we follow the procedure of @xcite .
we relate the inflow values of the pdf , associated to the outer ghost cells , to the outflow values of the pdf , which are associated to the interior cells adjacent to the boundary , as given below @xmath144 in the case of the boundary @xmath145 , assuming @xmath146 , @xmath147 , with @xmath148 , if @xmath149 then @xmath150 . the values of @xmath151 at the related inner and outer boundary cells @xmath152 ( @xmath153 ) and @xmath154 ( @xmath155 ) must be equal at the boundary @xmath156 : @xmath157 therefore , from the equality above we find the relation between the coefficients of @xmath86 at inner and outer adjacent boundary cells , given by @xmath158 following an analogous procedure for the boundary @xmath159 , we have that @xmath160 @xmath161 the diffusive reflection bc can be formulated in the following way : @xmath162 where @xmath163 and @xmath164 are the function and parameter such that the zero flux condition is satisfied at each of the points of the neumann boundary : @xmath165 \ , d \vec{k } \ , , \nonumber\\ 0 & = & \int _ { \vec{v } \cdot \eta > 0 } \vec{v } \cdot \eta \ , f|_{\gamma_{n^+ } } \ , d \vec{k } \ , - \ , \sigma(\vec{x},t ) \cdot c \int _ { \vec{v } \cdot \eta < 0 } |\vec{v } \cdot \eta| \ , e^{-\varepsilon(\vec{k})/k_b t_l } \ , d \vec{k } \ , .
\nonumber\end{aligned}\ ] ] it follows then that @xmath166 @xmath167 @xmath168 the diffusive reflection bc , formulated in terms of the unknown function @xmath53 of the transformed boltzmann equation [ eqphi ] , is expressed as @xmath169 @xmath170 @xmath171 we have , over the portion of the boundary considered , that @xmath172 for @xmath116 and @xmath173 for @xmath174 .
therefore : @xmath175 @xmath176 for the case of the dg numerical method , we have to project the bc to be imposed in the space @xmath177 .
our goal is to have at the numerical level an equivalent pointwise zero flux condition at the reflection boundaries .
we formulate then the diffusive bc for the dg method in the following way : @xmath178 where @xmath179 is a function in our piecewise polynomial space for @xmath80 and @xmath180 is a parameter such that the zero flux condition is satisfied numerically @xmath181 in the space @xmath177 of piecewise continuous polynomials which are tensor products of polynomials of degree @xmath32 in @xmath5 and of degree @xmath182 in @xmath118 , the following holds : @xmath183 therefore , for our particular case we have that @xmath184 so for the numerical zero flux condition pointwise we have that @xmath185 we observe then that we can obtain a numerical equivalent of the pointwise zero flux condition if we define @xmath186 in our particular case , in which we have chosen our function space as piecewise linear in @xmath80 and piecewise constant in @xmath81 , the projection of the maxwellian is a piecewise constant approximation representing its average value over each momentum cell , that is @xmath187 therefore , for the particular space we have chosen , we have that @xmath188 we notice that the polynomial approximation @xmath189 is equal to the analytical function @xmath190 operating on the polynomial approximation @xmath191 .
however , the constant @xmath180 needed in order to achieve the zero flux condition numerically is not equal to the value of this parameter for the analytical case .
@xmath180 is in this case an approximation of the analytical value using a piecewise constant approximation of the maxwellian ( its average over cells ) .
+ @xmath192 is a piecewise linear polynomial depending on @xmath80 with time dependent coefficients .
we have that @xmath193 @xmath194 is such that , at the boundary @xmath195 of the cell @xmath196 , is given by @xmath197 we define @xmath198 , so in @xmath199 .
we will have then @xmath200 we summarize the main results of the calculations for @xmath201 and @xmath202 , by showing the ones related to @xmath117 ( the case @xmath116 is analogous ) .
for the boundary @xmath117 , the inner cells associated to outflow have @xmath203 , adjacent to the boundary , whereas the ghost cells related to inflow have the index @xmath204 .
we compute the integral @xmath201 below .
@xmath205 therefore , we have , with @xmath206 , @xmath207 below , that @xmath208 @xmath209 @xmath210 once the coefficients of @xmath201 have been computed , we use them to obtain the polynomial approximation @xmath211 , with @xmath212 , from ( [ sigmahphih+ ] ) @xmath213 \frac { \int_{k } e^{-w } s(w ) \ , dw } { \delta w_k } \quad .
\ ] ] we have at the same time , by definition , that @xmath214 . \nonumber\ ] ] therefore , the coefficients for @xmath215 are @xmath216 @xmath217 @xmath218 keeping in mind that our parameter @xmath180 is given by the formula below @xmath219 the mixed reflection condition is a convex combination of the specular and diffusive reflections : @xmath220 @xmath32 is the specularity parameter , @xmath221 .
@xmath32 can be either constant or @xmath222 , a function of the wave vector momentum . + for @xmath32 constant , it can be shown easily that the previous formulas obtained for the specular and diffusive bc , in particular the previous formulas for @xmath190 @xmath223 , works also in this case to obtain a zero flux condition at the neumann boundaries : @xmath224 \ , d \vec{k } \nonumber\\ & = & \int _ { \vec{v } \cdot \eta > 0 } \vec{v } \cdot \eta \ , f|_{\gamma_{n^+ } } \ , d \vec{k } \ , + \ , p \int _
{ \vec{v } \cdot \eta < 0 } \vec{v } \cdot \eta \ , f(\vec{x } , \vec{k } ' , t)|_{\gamma_{n^+ } } \ , d \vec{k } + \left(1-p\right ) \sigma c \int _ { \vec{v }
\cdot \eta < 0 } \vec{v } \cdot \eta \ , e^{\frac{-\varepsilon(\vec{k})}{k_b t_l } } \ , d \vec{k } \nonumber\\ & = & \sigma(\vec{x},t ) \ , - \ , p \sigma(\vec{x},t ) + \left(1-p\right ) \sigma(\vec{x},t ) \left ( -1 \right ) \nonumber\\ & = & 0\end{aligned}\ ] ] however , for @xmath33 a function of the crystal momentum the same choice of @xmath225 and @xmath223 as in the diffusive case does not necessarily guarantee that the zero flux condition will be satisfied at neumann boundaries .
therefore , a new condition for @xmath180 in order to satisfy this condition must be derived .
we derive it below . + the general mixed reflection bc can be formulated as @xmath226 where @xmath227 and @xmath228 are the function and parameter such that the pointwise zero flux condition is satisfied at the neumann boundaries : @xmath229 \ , d \vec{k } \nonumber\end{aligned}\ ] ] since @xmath230 we conclude then that @xmath231 @xmath232 the general mixed reflection bc then has the specific form : @xmath233 with @xmath234 s.t .
@xmath235 + notice that the product @xmath236 has the form : @xmath237 which for the case of @xmath32 constant , it reduces to the original function @xmath225 and parameter @xmath238 : @xmath239 however , for the non - constant case @xmath33 the new function and parameter @xmath240 , @xmath241 need to be used instead , as the previous @xmath225 , @xmath242 wo nt satisfy the zero flux condition in general for @xmath243 : @xmath244 a more general possible case of mixed reflection bc would have a specularity parameter @xmath245 dependent on position , momentum , and time .
the related reflective bc would then be @xmath246 where @xmath247 is the distribution to which the electrons get into equilibrium when reflecting diffusively on the physical boundary .
@xmath248 and @xmath249 are the functions such that the zero flux condition is satisfied pointwise at insulating boundaries : @xmath250 d \vec{k } \nonumber\\ & = & \int_{\vec{v}\cdot\eta > 0 } \eta \cdot \vec{v } f|_+ d \vec{k } \ , + \ ,
{ \vec{v}\cdot\eta < 0 } \eta \cdot \vec{v } \ , p(\vec{x},\vec{k},t ) f|_+(\vec{x},\vec{k}',t ) d \vec{k } \nonumber\\ & - & \sigma^*(\vec{x},t ) \ , c^*(\vec{x},t ) \ , \int _ { \vec{v}\cdot\eta < 0 } | \eta \cdot \vec{v } | \left(1 - p(\vec{x},\vec{k},t ) \right ) m(\vec{x},\vec{k } ) d \vec{k } \nonumber\end{aligned}\ ] ] therefore we conclude for this reflection case that @xmath251 @xmath252 and then the full bc formula for the @xmath245 reflection case is @xmath253 } { \int _ { \vec{v}\cdot\eta < 0 } | \eta \cdot \vec{v } | \left(1 - p(\vec{x},\vec{k},t ) \right ) m(\vec{x},\vec{k } ) d \vec{k } } \nonumber\end{aligned}\ ] ] the numerical implementation of the general mixed reflection with specularity parameter @xmath33 is done in such a way that a numerical equivalent of the pointwise zero flux condition is achieved .
+ the general mixed reflection boundary condition in our dg numerical scheme is @xmath254 we will be using the notation @xmath255 the specific form of @xmath228 and @xmath256 will be deduced from the numerical analogous of the mixed reflection boundary condition .
we want to satisfy numerically the zero flux condition @xmath257 in the space @xmath177 of piecewise continuous polynomials which are tensor products of polynomials of degree @xmath32 in @xmath5 and of degree @xmath182 in @xmath118 , the following holds : @xmath258 therefore , we have for our particular case that @xmath259\ ] ] using this , our numerical pointwise zero flux condition is @xmath260 d\vec{w } \nonumber\\ & = & \int_{\vec{v}\cdot\eta>0}\vec{v}\cdot\eta\,\phi_h|_+ d\vec{w } \ , - \int_{\vec{v } \cdot \eta < 0 } |\vec{v } \cdot\eta | \ , \pi_h \left\lbrace p(\vec{w})\phi_h|_{+}(\vec{x},\vec{w}',t ) \right\rbrace d \vec{w } \nonumber\\ & + & c ' \sigma'_h(\vec{x},t ) \int_{\vec{v } \cdot \eta < 0 } \vec{v}\cdot\eta \ , \left [ \sum_{k , m , n } \chi_{kmn } \frac { \int_{k_{kmn } } ( 1-p(\vec{w } ) ) e^{-w } s(w ) d\vec{w } } { \int_{k_{kmn } } d\vec{w } } \right ] d\vec{w } \nonumber\\ & = & \int_{\vec{v}\cdot\eta>0}\vec{v}\cdot\eta\,\phi_h|_+ d\vec{w } \ , - \int_{\vec{v } \cdot \eta < 0 } |\vec{v } \cdot\eta | \ , \pi_h \left\lbrace p(\vec{w})\phi_h|_{+}(\vec{x},\vec{w}',t ) \right\rbrace d \vec{w } \nonumber\\ & - & c ' \sigma'_h(\vec{x},t ) \sum_{k , m , n } \chi_{kmn } \int_{\vec{v } \cdot \eta < 0 } | \vec{v}\cdot\eta | \ , d\vec{w } \ , \frac { \int_{k_{kmn } } ( 1-p(\vec{w } ) ) e^{-w } s(w ) d\vec{w } } { \int_{k_{kmn } } d\vec{w } } \nonumber \\ & = & \int_{\vec{v}\cdot\eta>0}\vec{v}\cdot\eta\,\phi_h|_+ d\vec{w } \ , - \int_{\vec{v } \cdot \eta < 0 } |\vec{v } \cdot\eta | \ , \pi_h \left\lbrace p(\vec{w})\phi_h|_{+}(\vec{x},\vec{w}',t ) \right\rbrace d \vec{w } \nonumber\\ & - & \sigma'_h(\vec{x},t ) \ , c ' \sum_{k , m , n , \
, \vec{v } \cdot \eta < 0 } \int_{k_{kmn } } | \vec{v}\cdot\eta | \ , d\vec{w } \
, \frac { \int_{k_{kmn } } ( 1-p(\vec{w } ) ) e^{-w } s(w ) d\vec{w } } { \int_{k_{kmn } } d\vec{w } } \nonumber\end{aligned}\ ] ] we conclude then that we can achieve a numerical equivalent of the pointwise zero flux condition by defining @xmath261 @xmath262 therefore , the inflow bc in our dg numerical method is given by the expression @xmath263 the particular form of the coefficients defining the piecewise polynomial approximation @xmath264 for the general mixed reflection bc is presented below for the boundary @xmath174 , since the calculations for the case of the boundary @xmath116 are analogous .
+ for the boundary @xmath265 , which defines the sign of @xmath266 .
outflow cells have the index @xmath267 .
they are cells inside the domain adjacent to the boundary .
inflow cells have the index @xmath268 .
they are ghost cells adjacent to the boundary .
we have in our case that @xmath269 if @xmath270 ( inflow ) , @xmath271 ( outflow ) , the projection integrand is given by @xmath272 \nonumber\ ] ] the coefficients of @xmath273 are given below .
we have now that @xmath274 @xmath275 , so from the previous two formulas then @xmath276 since on one hand we have @xmath277 \\ & + & \sum_{i , k , m , n}^{n > n_{\varphi}/2 } \chi_{i , n_y+1,k , m , n } \ , \frac{\int_{{kmn } } ( 1- p(\vec{w } ) ) e^{-w } s(w ) d\vec{w } } { \int_{{kmn } } d\vec{w } } \times \nonumber\\ & \times & c ' \left [ { \sigma'}_{i , n_y , k , m , n}^0
+ { \sigma'}_{i , n_y , k , m , n}^x \frac{(x - x_i)}{\delta x_i/2 } + { \sigma'}_{i , n_y , k , m , n}^y ( + 1 ) \right ] \end{aligned}\ ] ] and on the other hand @xmath278 \nonumber\ ] ] we conclude that the coefficients for @xmath211 are then @xmath279
we present results of preliminary numerical simulations for the case of a 2d n bulk silicon diode with an applied bias between the boundaries @xmath280 , and reflection bc at the boundaries @xmath124 .
the required dimensionality in momentum space is 3d @xmath281 .
the specifics of our simulations are : + initial condition : @xmath282 .
final time : 1.0ps + boundary conditions ( bc ) : + @xmath38-space : cut - off - at @xmath283 , @xmath53 is machine zero .
+ only needed bc in @xmath81 : transport normal to the boundary analitically zero at singular points boundaries : + at @xmath133 , @xmath284 . at @xmath285 , @xmath286 = 0 . at @xmath287 ,
@xmath288 = 0 .
+ @xmath5-space : charge neutrality at boundaries @xmath289 . + bias - potential : @xmath290 v , @xmath291 v. + neumann bc for potential at @xmath292 : @xmath293 .
+ reflection bc at @xmath294 : specular , diffusive , & mixed with @xmath295 .
+ we observe an influence of the diffusive and mixed reflection in macroscopic observables , particularly noticeable in the kinetic moments .
for example , the distribution of charges slightly increases with diffusive reflection close to reflecting boundaries , and alters the density profile over domain , due to mass conservation .
momentum & velocity increase with diffusive reflection over the domain , while the energy is decreased by diffusive reflection over the domain .
( @xmath296 ) vs position @xmath80 in @xmath297 plot for specular , diffusive , & mixed @xmath295 reflection.,title="fig : " ] vs. position @xmath80 in @xmath297 plots for specular , diffusive , & mixed @xmath295 reflection.,title="fig : " ] vs. position @xmath80 in @xmath297 for specular , diffusive , and mixed @xmath295 reflection . ] vs. position @xmath80 in @xmath297 for specular , diffusive , and mixed @xmath295 reflection . ] vs. position @xmath80 in @xmath297 for specular , diffusive , and mixed @xmath295 reflection . ] vs. position @xmath80 in @xmath297 for specular , diffusive , and mixed @xmath295 reflection . ] vs. position @xmath80 @xmath297 plot for specular , diffusive & mixed @xmath295 reflection , title="fig : " ] vs. position @xmath80 in @xmath297 for specular , diffusive & mixed @xmath295 reflection ]
we have considered the mathematical and numerical modeling of reflective boundary conditions in 2d devices and their implementation in dg - bp schemes . we have studied the specular , diffusive and mixed reflection bc on physical boundaries of the device for the modeling of surface roughness , comparing the influence of these different reflection cases in the computational prediction of moments close to the boundaries and their associated scale .
+ there is an effect due to diffusive reflection at neumann boundaries over the moments of the probability density function , whose influence is not only restricted to the boundaries but actually to the whole domain , partly due to mass conservation .
+ work in progress is being developed for the computational implementation for mixed reflection with a @xmath298 , and the study of reflective bc in a 2d double gate mosfet device .
future research will be related to reflective bc with an epm - dg - bp full band model , and the inclusion of additional scattering mechanisms . c. cercignani , i. m. gamba , c.d .
levermore , _ high field approximations to a boltzmann - poisson system and boundary conditions in a semiconductor _ ,
1em plus 0.5em minus 0.4emappl .
* 10 * , 4 , 111 - 117 ( 1997 ) .
, _ the monte carlo method for semiconductor device simulation _ , spring - verlag : wien - new york , 1989 .
e. fatemi and f. odeh , _ finite difference solution of boltzmann equation applied to electron transport in semiconductor devices _ , journal of computational physics , 108 ( 1993 ) 209 - 217 .
a. majorana and r. pidatella , _ a finite difference scheme solving the boltzmann poisson system for semiconductor devices _ , journal of computational physics , 174 ( 2001 ) 649 - 668 .
carrillo , i.m .
gamba , a. majorana and c .- w .
shu , _ a weno - solver for 1d non - stationary boltzmann - poisson system for semiconductor devices _ , journal of computational electronics , 1 ( 2002 ) 365 - 375 .
carrillo , i.m .
gamba , a. majorana and c .- w .
shu , _ a direct solver for 2d non - stationary boltzmann - poisson systems for semiconductor devices : a mesfet simulation by weno - boltzmann schemes _ , journal of computational electronics , 2 ( 2003 ) 375 - 380 .
carrillo , i.m .
gamba , a. majorana and c .- w .
shu , _ a weno - solver for the transients of boltzmann - poisson system for semiconductor devices .
performance and comparisons with monte carlo methods _
, journal of computational physics , 184 ( 2003 ) 498 - 525 .
caceres , j.a .
carrillo , i.m .
gamba , a. majorana and c .- w .
shu , _ deterministic kinetic solvers for charged particle transport in semiconductor devices _ , in transport phenomena and kinetic theory applications to gases , semiconductors , photons , and biological systems .
series : modeling and simulation in science , engineering and technology . c. cercignani and e. gabetta ( eds . ) , birkhuser ( 2006 ) 151 - 171 .
carrillo , i.m .
gamba , a. majorana and c .- w .
2d semiconductor device simulations by weno - boltzmann schemes : efficiency , boundary conditions and comparison to monte carlo methods _ , journal of computational physics , 214 ( 2006 ) 55 - 80 . m. galler and a. majorana , _ deterministic and stochastic simulation of electron transport in semiconductors _ , bulletin of the institute of mathematics , academia sinica ( new series ) , 6th mafpd ( kyoto ) special issue vol . 2 ( 2 ) ( 2007 ) 349 - 365 . z. chen , b. cockburn , c. gardner and j. jerome , _ quantum hydrodynamic simulation of hysteresis in the resonant tunneling diode _ , journal of computational physics , 274 ( 1995 ) 274 - 280 .
z. chen , b. cockburn , j. w. jerome and c .- w .
shu , _ mixed - rkdg finite element methods for the 2-d hydrodynamic model for semiconductor device simulation _
, vlsi design , 3 ( 1995 ) 145 - 158 . y .-
x . liu and c .- w .
shu , _ local discontinuous galerkin methods for moment models in device simulations : formulation and one dimensional results _ , journal of computational electronics , 3 ( 2004 ) 263 - 267 .
x . liu and c .- w .
shu , _ local discontinuous galerkin methods for moment models in device simulations : performance assessment and two dimensional results _ , applied numerical mathematics , 57 ( 2007 ) 629 - 645 .
y. cheng , i.m .
gamba , a. majorana and c .- w .
shu , _ discontinuous galerkin solver for the semiconductor boltzmann equation _ , sispad 07 , t. grasser and s. selberherr , editors , springer ( 2007 ) 257 - 260 .
y. cheng , i. gamba , a. majorana and c .- w .
shu , _ discontinuous galerkin solver for boltzmann - poisson transients _ , journal of computational electronics , 7 ( 2008 ) 119 - 123 .
y. cheng , i. m. gamba , a. majorana and c .- w .
shu _ a discontinuous galerkin solver for boltzmann - poisson systems in nano - devices _ , computer methods in applied mechanics and engineering , 198 ( 2009 ) 3130 - 3150 .
y. cheng , i. m. gamba , a. majorana and c.w .
shu _ a discontinuous galerkin solver for full - band boltzmann - poisson models _ , iwce13 ( 13th international workshop on computational electronics ) ( 2009 ) .
y. cheng , i. m. gamba and j. proft , _ positivity - preserving discontinuous galerkin schemes for linear vlasov - boltzmann transport equations _
, mathematics of computation , 81 ( 2012 ) 153 - 190 .
j. chelikowsky , and m. cohen , _ electronic structure of silicon _ , physical review b , 10 ( 1974 ) 5095 - 5107 .
m l. cohen and j.r .
chelikowsky , _ electronic structure and optical properties of semiconductors _ , springer series in solid - state sciences 75 ( 1989 ) . c. jungemann , a.t .
pham , b. meinerzhagen , c. ringhofer and m. bollhfer , _ stable discretization of the boltzmann equation based on spherical harmonics , box integration , and a maximum entropy dissipation principle _ , journal of applied physics , 100 ( 2006 ) 024502 .
a. majorana , _ a comparison between bulk solutions to the boltzmann equation and the spherical harmonic model for silicon devices _ , progress in industrial mathematics at ecmi 2000 - mathematics in industry , 1 ( 2002 ) 169 - 173 . b. cockburn and c .- w .
shu , _ tvb runge - kutta local projection discontinuous galerkin finite element method for conservation laws ii : general framework _ , mathematics of computation , 52 ( 1989 ) 411 - 435 .
b. cockburn , s .- y .
lin and c .- w .
shu , _ tvb runge - kutta local projection discontinuous galerkin finite element method for conservation laws iii : one dimensional systems _ , journal of computational physics , 84 ( 1989 ) 90 - 113 .
b. cockburn , s. hou and c .- w .
shu , _ the runge - kutta local projection discontinuous galerkin finite element method for conservation laws iv : the multidimensional case _ , mathematics of computation , 54 ( 1990 ) 545 - 581 .
b. cockburn and c .- w .
shu , _ the runge - kutta local projection p1-discontinuous galerkin finite element method for scalar conservation laws _ , mathematical modelling and numerical analysis , 25 ( 1991 ) 337 - 361 . b. cockburn and c .- w .
shu , _ the runge - kutta discontinuous galerkin method for conservation laws v : multidimensional systems _ , journal of computational physics , 141 ( 1998 ) 199 - 224 . b. cockburn and c .- w .
shu , _ runge - kutta discontinuous galerkin methods for convection - dominated problems _ , journal of scientific computing , 16 ( 2001 ) 173 - 261 . b. cockburn and c .- w .
shu , _ the local discontinuous galerkin method for time - dependent convection - diffusion systems _ , siam journal on numerical analysis , 35 ( 1998 ) 2440 - 2463 .
d. arnold , f. brezzi , b. cockburn , and l. marini , _ unified analysis of discontinuous galerkin methods for elliptic problems _ , siam journal on numerical analysis , 39 ( 2002 ) 1749 - 1779 . c. cercignani , i.m .
gamba , c.l .
levermore , _ a drift - collision balance asymptotic for a boltzmann - poisson system in bounded domains _ , siam j. appl .
, vol 61 , no . 6 , ( 2001 ) 1932 - 1958 . | we shall discuss the use of discontinuous galerkin ( dg ) finite element methods to solve boltzmann - poisson ( bp ) models of electron transport in semiconductor devices at nano scales .
we consider the mathematical and numerical modeling of reflective boundary conditions in 2d devices and their implementation in dg - bp schemes .
we study the specular , diffusive and mixed reflection bc on physical boundaries of the device for the modeling of surface roughness , comparing the influence of these different reflection cases in the computational prediction of moments close to the boundaries and their associated scale .
we observe an effect due to diffusivity at the reflection boundaries over the kinetic moments of the probability density function , whose influence is not restricted to the boundaries but rather over the position space domain . |
the lattice 4d fermion - gauge scalar ( @xmath0 ) model with u(1 ) gauge symmetry has at quite strong gauge coupling , @xmath1 , a tricritical point frje95a , frfr95a .
the nature of the continuum limit taken at this point is not known .
when this point ( point e in fig .
[ fig:1 ] ) is approached from the phase with spontaneously broken global chiral symmetry and a massive fermion @xmath2 ( nambu phase ) , the mass @xmath3 scales .
the fermion @xmath4 is composed of the fundamental fermion ( @xmath5 ) and scalar ( @xmath6 ) fields and is unconfined .
this raises the hope that a nonperturbatively renormalizable theory with dynamical fermion mass generation might be obtained frje95a .
motivated by this interesting possibility we have started an investigation of the scaling behavior of various observables in the vicinity of the point e. we find that apart from @xmath3 also the mass @xmath7 of a composite scalar boson @xmath8 scales in the vicinity of e. this , as well as some other properties frfr95a make the point distinctly different from the ( nonrenormalizable algo95 ) nambu jona - lasinio model found in the strong coupling limit of the @xmath0 model leshr87a .
this , as well as the experience that tricritical points give rise to the scaling behavior different from that associated with the adjacent critical points lasa84 , contributes to the hope that the point e might correspond to a renormalizable theory .
as reported in frje96c , in 2d an analogous @xmath9 model seems to belong to the universality class of the 2d gross - neveu model and is thus probably renormalizable .
the action is @xmath10 where @xmath11\ ! \nonumber\\ & & + \ , a m_0 \sum_x { \overline{\chi}}_x \chi_x \ ; , \label{sch } \\ s_u \hspace{-2mm}&=&\hspace{-2 mm } \beta \ , \sum_{\rm p } \left [ 1 - { \mathop{\rm re}\nolimits}u_{\rm p } \right ] \ ; , \label{su } \\ s_\phi \hspace{-2mm}&=&\hspace{-2 mm }
- { \kappa}\ , \sum_x \sum_{\mu=1}^4 \left [ \phi_x^\dagger u_{x,\mu } \phi_{x + \mu } \,+\ , { \rm h.c . } \right ] \ ; . \label{sph}\end{aligned}\ ] ] here @xmath12 is the plaquette product of link variables @xmath13 and @xmath14 .
the gauge field link variables @xmath13 are elements of the compact gauge group u(1 ) . the complex scalar field @xmath6 of charge one satisfies the constraint latexmath:[$|\phi infinite ) when the squared bare scalar mass is positive ( negative ) infinite .
the staggered fermion field @xmath5 of charge one leads to the global u(1 ) chiral symmetry of the model in the chiral limit , i.e. when the bare fermion mass @xmath16 vanishes .
the model is meant in the chiral limit , @xmath17 .
however , numerical simulations require nonvanishing @xmath16 and thus also analytic considerations have to include the bare mass .
we therefore show in fig .
[ fig:1 ] schematic phase diagram of the model in the @xmath18 space .
it illustrates that the tricritical point e is indeed a point in which three critical lines meet : * line ne where @xmath3 and goldstone boson mass @xmath19 simultaneously smoothly vanish .
* lines ee@xmath20 and ee@xmath21 with vanishing @xmath7 .
the nambu phase below the ne line at @xmath22 is characterized by @xmath23 and @xmath24 .
the limit cases are * @xmath25 : 4d xy model and free fermion field .
* @xmath26 : nambu - jona - lasinio model with bare mass . *
@xmath27 : u(1 ) higgs model without fermions .
* @xmath28 : compact qed with charged fermion field . * @xmath29 : free fermion field .
as is apparent from fig .
[ fig:1 ] , an approach to the tricritical point requires in principle tuning of three parameters
. obviously also the lattice sizes have to be varied .
but the tricritical points have frequently rather large domains ( angles ) of dominance lasa84 .
therefore it is reasonable to fix @xmath30 , which is our current best estimate of the position of the point e in the @xmath31 direction , and vary only @xmath32 and @xmath33 . even so an investigation of the scaling behavior is a tremendous task .
we can present only some preliminary results obtained for @xmath34 and , rather incomplete , also for @xmath35 , without a proper analytic analysis and in particular without taking into account finite size effects . in fig .
[ fig:2 ] we show the chiral condensate @xmath36 .
an onset of the genuine chiral phase transition is indicated with decreasing @xmath32 .
the fermion mass @xmath3 ( fig .
[ fig:3 ] ) behaves similarly .
the transition appears to be smooth , i.e. of 2@xmath37 order , close to @xmath38 . in fig .
[ fig:4 ] we show the scalar boson mass @xmath7 .
the position of the minimum for each @xmath32 corresponds roughly to the crossection of a continuation to smaller @xmath39 of the nearly horizontal `` wing '' with the @xmath30 plane .
as @xmath32 decreases , the minimum seems to approach the @xmath33 value at which the chiral phase transition at @xmath22 is expected on the basis of figs . 2 and 3 .
thus we are indeed in a close vicinity of the tricritical point . fig .
[ fig:5 ] contains very preliminary results for several @xmath40 ( meson ) masses . for comparison
we show also data for @xmath41
. the goldstone boson mass @xmath19 is small around the transition as expected .
the fact that the @xmath42- and @xmath43- masses cross the @xmath41 line in the vicinity of the phase transition indicates that in the nambu phase these mesons can be interpreted as bound states and that their masses scale similarly to @xmath3 and thus may exist also in the continuum limit .
this is a hint that the spectrum of the underlying continuum theory might be quite complex .
our present results suggest that it is worthwhile to continue the investigation of the tricritical point e. for this purpose the full scaling theory of tricritical points lasa84 and a finite size scaling theory in their vicinity will have to be applied .
some analytic insight into the properties of the model in the relevant coupling region is highly desirable . | we investigate the tricritical point in the lattice fermion gauge scalar model with u(1 ) gauge symmetry . in the vicinity of this point , in the phase with the broken chiral symmetry , we observe the scaling behavior of the chiral condensate and of the masses of composite fermion and composite scalar , indicating the existence of an interesting continuum limit of the model at this point . |
severe acute respiratory syndrome ( sars ) is a new respiratory disease which was first identified in china s southern province of guangdong .
sars is not merely a local endemic disease : it poses a serious risk to the medical community , is a threat to international travelers , is having a substantial negative economic impact in parts of east asia and is spreading world - wide .
the serious danger sars poses to the medical community is illustrated by the numerous cases of transmission to health - care workers . startlingly , the man who awakened the world to the dangers of sars , dr .
carlo urbani , succumbed to the disease .
cases of transmission between aircraft passengers are suspected , and relatively short visits to epidemic regions have resulted in infection .
the most striking feature of sars , however , has proven to be its ability to rapidly spread on a global scale .
one man with sars made @xmath1 flights : from hong kong to mnich to barcelona to frankfurt to london , back to mnich and frankfurt before finally returning to hong kong @xcite .
another individual , a @xmath2-year - old airport worker , appears to have transmitted the disease to @xmath3 people @xcite .
clearly , there is an unfortunate interaction between the incubation period of the virus , the widely distributed severity and infectiousness of sars in different people and the speed and volume of passenger air travel .
the adverse economic impact in parts of east asia far exceeds the disruption of previous outbreaks of avian influenza , earning comparison with the 1998 financial market crisis in that part of the world @xcite .
although the causative agent of sars has been determined @xcite , a detailed understanding of the causative virus pathogenticity and routes of transmission and the dynamics of the epidemic is still at a very early stage .
it is uncertain how the virus is transmitted : by droplet or airborne transmission or person - to - person contact .
the recent development of laboratory tests promises to improve the epidemiological situation somewhat @xcite .
sars is a public health crisis on a scale rarely seen .
the obvious question in such a crisis is , `` can sars be contained ? '' in this study , we report transmission parameters and epidemic dynamics from a model based on classes of people who are susceptible , exposed , infectious , diagnosed , and recovered ( `` seijr '' ) that includes the effect of patient isolation . our model is consistent with the possibility of containment in toronto , ontario .
sars was first identified in november 2002 in the guongdong province of china @xcite . by february @xmath2 ,
@xmath4 officials in hong kong reported their first cases of sars and no later than march @xmath5 of this year the virus reached canada @xcite .
as of april @xmath6 , canada is the only location outside of asia which has seen deaths as a result of sars ( @xmath7 so far ) @xcite . u.s
. health officials are currently investigating @xmath8 cases in @xmath9 states ( apri 17 , 2003 ) @xcite .
an individual exposed to sars may become infectious after an incubation period of @xmath10 days ( or longer ) @xcite with @xmath11 days being most common @xcite .
most infected individuals either recover , typically after @xmath1 to @xmath12 days , or suffer @xmath13 mortality or higher @xcite .
sars appears to be most serious in people over age @xmath14 , especially those who have other medical problems such as heart or liver disease .
its symptoms are similar to pneumonia or other respiratory ailments and include a high fever ( @xmath15c ) , shortness of breath , dry cough , headache , stiff or achy muscles , fatigue and diarrhea @xcite . these symptoms
, however , are not uniform . in the us , for example , the disease seems to be a milder one than in asia @xcite .
the result has been that sars was , and for the moment remains , a diagnosis of exclusion .
presently , there is no treatment for sars @xcite and diagnostic tests are just becoming available @xcite . the mortality rate is reported to be @xmath16% or higher world - wide@xcite .
experts estimate that between @xmath17 and @xmath18 percent of people with sars recover without medical intervention , while the condition of the remaining victims requires medical care @xcite .
as of april @xmath19 , @xmath4 , the world health organization ( who ) reported @xmath20 cases ( a mixture of probable or suspected cases ) in @xmath2 countries .
@xmath21 victims are reported to have died @xcite . although researchers in the erasmus medical center in rotterdam recently demonstrated that a coronavirus ( some of which produce common colds ) is the causative agent of sars , the mode of transmission still remains unknown @xcite .
the current hypothesis is that sars is transmitted mainly by close person - to - person contact which may explain the relatively slow transmission scale .
however , it could also be transmitted through contaminated objects , air or by other unknown ways @xcite .
it is also a mystery how the disease originated , whether in birds , pigs or other animals , nor is it known if the origin is rural or urban @xcite . in this article , a simple model for sars outbreaks
is formulated ( see @xcite ) .
the model is used in conjunction with _ global and local _ sars data to estimate the initial growth rate of the sars epidemic .
these rates are used to estimate sars _ basic reproductive number _ , @xmath22 , the classical epidemiological measure associated with the reproductive power of a disease .
@xmath22 estimates the _ average _ number of secondary cases of infection generated by a _
typical _ infectious individual in a population of susceptibles @xcite and hence , it is used to estimate the initial growth of a sars outbreak .
we estimate ( using data from ontario , hong kong and singapore ) that @xmath22 is about @xmath0 .
this value is not too different from past estimates of @xmath22 for influenza ( see @xcite ) despite the fact that _ superspreaders _ of sars have been identified .
in fact , the parameter values resulting on this @xmath22 , on our population - scaled model , can lead to extremely high levels of infection ) .
we show , via simple extrapolation , that the estimated rate of growth is consistent with the reported date for the first cases of sars in hong kong , however the first cases in toronto may be several weeks earlier than the february 23 date of the first case reported by the canadian health ministries @xcite .
our best `` rough '' estimate for toronto is that the first case occurred sometime around january 29th , and not later than february 28th .
the data for hong kong are fitted by fixing the parameters @xmath23 , @xmath24 and @xmath25 based on estimates of the observed rates for the corresponding processes .
the growth rate @xmath26 is estimated from observed `` model - free '' exponential growth in singapore and hong hong .
the _ average _ diagnostic rate @xmath27 and the measure of heterogeneity between the two susceptible classes @xmath28 and the effectiveness of patient isolation measures ( related to @xmath29 ) are then varied to fit the initial data for hong kong and singapore . to model the data in toronto
, we must postulate that the parameters describing the rate of diagnosis ( @xmath27 ) and isolation ( @xmath29 ) in the canadian outbreak changed radically on march @xmath30 .
two hospitals in toronto were closed about that time : scarborough grace hospital on march @xmath31 and york central hospital on march @xmath32 @xcite .
the remainder of this article is organized as follows : section @xmath16 introduces the basic model and gives a formula for the basic reproductive number @xmath22 ; section @xmath33 describes the results of simulations and connections to data ; and , section @xmath34 collects our final thoughts .
u.s . data is limited and sparsely distributed @xcite while the quality of china s data is hard to evaluate @xcite . on the other hand
, there appears to be enough data for toronto @xcite , singapore and hong kong @xcite to make limited preliminary predictions using a model that includes the effects of _ suspected _ mechanisms for the spread of sars .
limited data and inconclusive epidemiological information place severe restrictions on efforts to model the global spread of the sars etiological agent .
thus , we model _ single _ outbreaks , ignoring demographic processes other than the impact of sars on survival .
the model is applied to data from toronto , hong kong and singapore .
because the outbreak dynamics in singapore and hong kong are different from those in toronto , some of the results may only be indicative of what is happening in those regions of the world ( in particular our parameters @xmath27 and @xmath29 may change )
. the situation must be re - evaluated frequently as sars continues its travels around the world .
here we describe a model that incorporates , in a _ rather crude _ way , some of the important characteristics suggested in the literature ( unequal susceptibility , symptomatic and asymptomatic individuals , mode of transmission , superspreaders , etc . )
the goal is to use the results for single outbreaks as a first step in our efforts to gauge the _ global _ impact of sars .
hence , we focus on three `` closed '' populations ( southern ontario ( toronto ) , singapore and hong kong ) and postulate differences in the degree of susceptibility to sars @xcite
. these differences may be due to variations in contact rates , age - dependent susceptibility or `` unknown '' genetic factors .
this last assumption is handled ( in a rather crude and arbitrary way ) via the introduction of two distinct susceptible classes : @xmath35 , the most susceptible , and @xmath36 , less so .
initially , @xmath37 and @xmath38 where @xmath39 is the proportion of the population size @xmath40 that is initially at higher risk of sars infection .
the parameter @xmath28 is a measure of reduced susceptibility to sars in class @xmath36 @xcite .
@xmath41 ( `` exposed '' ) denotes the class composed of asymptomatic , possibly infectious ( at least some of the time ) individuals .
typically , it takes some time before asymptomatic infected individuals become infectious .
the possibility of limited transmission from class @xmath41 is included , in a rather crude way , via the parameter @xmath42 ( see table @xmath43 ) .
the class @xmath44 denotes infected , symptomatic , infectious , and undiagnosed individuals .
@xmath44-individuals move into the diagnosed class @xmath45 at the rate @xmath27 .
individuals recover at the rates @xmath25 ( @xmath44 class ) and @xmath46 ( @xmath45 class ) . the rate @xmath24 denotes sars disease - induced mortality .
the classes @xmath47 is included to keep track of the cumulative number of diagnosed and recovered , respectively .
furthermore , it is assumed that diagnosed individuals are handled with _
care_. hence , they might not be ( _ effectively _ ) as infectious as those who have not been diagnosed ( if @xmath29 is small ) .
the parameter @xmath29 takes into account their reduced impact on the transmission process ( small @xmath29 represents effective measures taken to isolate diagnosed cases and _ visa versa _ ) .
table @xmath43 includes parameters definitions and the initial values used .
our sars epidemiological model is given by the following nonlinear system of differential equations : @xmath48 which is refered to as `` seijr , '' after the variables used to name the classes .
+ the values of @xmath28 and @xmath42 are not known and are fixed arbitrarily while @xmath29 and @xmath27 are varied and optimized to fit the existing data ( least - squares criterion ) for hong kong , singapore and toronto .
we did not explored the sensitivity of the model to variations in @xmath28 and @xmath42 because they are not known and can not be controlled .
all other parameters were roughly estimated from data @xcite and current literature @xcite .
in particular , the transmission rate @xmath26 is calculated from the dominant root of the third order equation obtained from the linearization around the disease - free equilibrium @xcite .
the parameters @xmath29 and @xmath27 were allowed to vary when fitting the data for each location ( singapore , hong kong and toronto ) .
some restrictions apply , for example , the value of @xmath49 .
we also require that @xmath50 , a statement that members of the diagnosed class @xmath45 recover at the same rate as members of the undiagnosed class @xmath44 .
@xmath51 has been reported to be between @xmath1 and @xmath12 days @xcite . from the second generator approach @xcite , we obtain the following expression for the basic reproductive number : @xmath52\right\ } \left\{\frac{q}{k } + \frac{1}{\alpha + \gamma_1 + \delta } + \frac{\alpha l}{(\alpha + \gamma_1 + \delta)(\gamma_2 + \delta)}\right\}\\ \end{array}\ ] ] which can be easily given an epidemiological interpretation .
the use of parameters estimated from hong kong ( table @xmath43 ) gives a values of @xmath53 ( hong kong ) and @xmath54 ( toronto , assuming exponential growth ) and @xmath55 ( singapore ) .
initial rates of growth for sars outbreaks in different parts of the world ( see figure [ myfig00 ] ) are computed using the data provided by who @xcite and the canadian ministry of health @xcite .
these rates are computed exclusively from the number of cases reported between march @xmath56 and april @xmath57 .
the values obtained are @xmath58 ( world data ) , @xmath59 ( hong kong ) , @xmath60 ( canada ) , @xmath60 ( toronto ) and @xmath61 ( singapore ) . for our numerical simulations , we start with an infectious individual ( not yet diagnosed , @xmath62 ) and _ crude _ estimates for the start of sars outbreaks ( @xmath63 ) are obtained from the formula @xmath64 , which assumes initial exponential growth ( @xmath65 , the estimated `` model - free '' rate of growth from the time series @xmath66 of the cumulative number of sars cases ) .
results for toronto , hong kong , singapore and aggregated world data are shown in table @xmath67 .
the estimated `` world '' start of the outbreak is november @xmath33 , a date consistent with the fact that the first sars case was detected in guangdong , china in november @xcite .
these dates are used as the starting time of the respective outbreaks . + for the case of the province of ontario , canada the total population @xmath40 is approximately @xmath68 million .
we assume that the population at _ major _ risk of sars infection lives in ontario s southern part ( particularly toronto ) , and is approximately @xmath69 of the total population ( @xmath70 in our model ) .
it is worth pointing out that this value of @xmath39 is not critical ( that is , the most sensitive ) in the model .
the `` model - free '' approximately exponential growth rates for the various regions of the world are roughly similar _ except for canada _ from march 31st ( day 61 ) to april 6th ( beginning the day of the jump in the number of reported canadian cases ) , the number of diagnosed cases grew @xmath71 , where @xmath72 is measured in days .
this rate is substantially higher than elsewhere in the world . in the subsequent week
( beginning april 7th , day 68 ) the number of probable or suspected canadian cases rapidly rolls over to a smaller growth rate not too far from the rest of the world .
we conclude , based on the coincidence of the canadian hospital closures , the jump in the reported number of canadian sars cases on march 31st and the rapid rise in recognized cases in the following week , that canadian doctors were rapidly diagnosing pre - existing cases of sars ( in either class @xmath41 or @xmath44 on march 26th ) .
if we make the assumption that the fundamental disease spreading parameters other than @xmath27 and @xmath29 are roughly constant throughout the world prior to march 26th , we can reach two important conclusions .
beginning on march 26th , in toronto : * @xmath27 changed from a number @xmath73 days to @xmath74 days , and * @xmath29 changed from an uncertain and relatively large value @xmath75 to @xmath76 .
if we assume that the fundamental growth rate @xmath26 is essentially constant from one region of the world to another , it is difficult for our model to produce growth rates @xmath65 well above the world average , except as a transient response to differences in diagnostic rate @xmath27 ( due to delays in response or change in policy ) .
similarly , the seijr model requires fairly small values of @xmath29 to achieve a rapid roll - over in the growth rate of recognized cases .
the parametric details of how a `` second '' initial condition for toronto on march 26th is generated do not affect the qualitative aspect of this argument : the canadian data prior to march 31st ( the day of the large jump ) are probably not as meaningful as data after that date , and hence only bound the model from below prior to march 26th .
the essential aspect of this before - and - after hospital closure argument is that there were substantially more undiagnosed people in classes @xmath41 and @xmath44 than in class @xmath45 on march 26th .
this is a reasonable assumption given that the number of cases reported by canadian officials more than double from march 30th to march 31st .
the introduction of behavioral changes starting on march @xmath2 ( @xmath77 days),alters the fate of the disease in a dramatic fashion ( see table @xmath78 ) .
+ fitting the model to the hong kong and singapore data is carried out in a similar fashion with @xmath79 , ( hong kong has about @xmath80 million inhabitants , singapore @xmath81 million ) .
the estimated transmission rate from hong kong data is @xmath82 and for singapore @xmath83 .
both hong kong and singapore s data are fit with the value @xmath84 .
hong kong and singapore s measure of contact between diagnosed sars cases and susceptibles are @xmath85 and @xmath86 , respectively ( see figure 4 ) .
even though there is some heterogeneity in the parameters for hong kong and singapore , they provide an important calibration of our model .
their values for @xmath29 and @xmath27 are roughly consistent with each other , indicating that the difference with toronto is significant within our model , and pointing to the joint importance of rapid diagnosis @xmath87 and good isolation of diagnosed patients @xmath88 in controlling an outbreak .
while there is some indication in the data from hong kong of a possible slowing of the outbreak , we did not attempt to analyze the slowing or assess its significance .
a simple model that can capture the effect of average infectiousness in a heterogeneous population and the effect of isolating diagnosed patients has been introduced to explore the role of patient isolation and diagnostic rate in controlling a sars outbreak . by examining two cases with relatively clean exponential growth curves for the number of recognized cases , we are able to calibrate a seijr model with parameters @xmath89 ( sars diagnostic rate ) and @xmath90 ( isolation effectiveness ) .
we then use our seijr model to examine the non - exponential dynamics of the toronto outbreak .
two features of the toronto data , the steep increase in the number of recognized cases after march 31st and rapid slowing in the growth of new recognized cases , robustly constrained the seijr model by requiring that @xmath91 and @xmath92 days@xmath93 .
the model is also used to look at the impact of drastic control measures ( isolation ) .
the fitting of data shows that the initial rates of sars growth are quite similar in most regions leading to estimates of @xmath22 between @xmath94 and @xmath0 despite the recent identification of _
superspreaders_. model simulations are fitted to sars reported data for the province of ontario , hong kong and singapore .
good fits are obtained for reasonable values of @xmath27 , the rate of identification of sars infections ; `` reasonable '' values of the control parameters @xmath29 ( a measure of isolation ) ; possible values of @xmath28 , a _ crude _ measure of reduced susceptibility ( due to genetic factors , age or reduced contact rates ) ; @xmath42 a _ crude _ measure of the relative degree of infectiousness of asymptomatic individuals ; possible values of @xmath39 a measure of initial levels of population heterogeneity ; and , reasonable values of n the _ effective _ population size .
it is worth noticing that for values of @xmath40 larger than @xmath95 the predictions ( proportion of cases at the end of the outbreak , etc . )
are _ roughly _ the same . the introduction of behavioral changes that follow the identification of the first case ( reduce values of @xmath29 at the time of the identification and moving aggressively to identify cases of sars by increasing @xmath96 ) result in a dramatic reduction in the total number of cases and on mortality in toronto .
given the fact that sars appears to kill between three and seven percent of infected ( diagnosed ) cases ( @xcite ) , it seems quite appropriate to isolate diagnosed people .
although we do not examine the effect of quarantine by varying @xmath42 , it seems intuitive that quarantining those who came into close contact with positively diagnosed individuals will reduce the total number of cases .
+ model results and simple estimates suggest that _ local _ outbreaks may follow similar patterns .
furthermore , the use of relative extreme isolation measures in conjunction with rapid diagnosis has strong impact on the local dynamics ( toronto s situation ) .
however , if sars has shown us anything it is that `` undetected '' and `` unchecked '' local disease dynamics can rapidly become a global issue . the research on this article used the latest data available ( april @xmath97 for canada and april @xmath98 for hong kong and singapore ) .
recent disclosures @xcite reaffirm the importance of carrying out the analysis excluding data from china .
we have redone the analysis including the data collected up to april @xmath99 and , our conclusions , remain the same .
current data seem to support higher values for sars induced mortality rates @xcite . however , our model is _ most sensitive _ to the parameters @xmath29 ( effectiveness of isolation ) and ( @xmath27 ) diagnostic rate .
it is not as sensitive to changes in @xmath24 .
in fact , the consideration of a @xmath100 mortality ( @xmath101 ) rather than @xmath13 reduces the number of cases by about @xmath102 . in toronto ,
we have estimated @xmath103 diagnosed cases with ( @xmath104 and @xmath89 after march 26th ) .
perfect isolation after march 26th , ( @xmath105 ) reduces this number to @xmath106 diagnosed cases .
the assumption of homogenous mixing implies that our model is likely to overestimate the size of the outbreak .
hence , the situation in toronto seems to support the view that this _ outbreak _ is being contained .
obviously , the case of the crude model ( by design ) can not handle high levels of variability ( an stochastic model would be desirable ) .
this possibility is tested ( as it is often done in deterministic models ) by looking at the sensitivity of the model to parameters ( @xmath27 and @xmath29 being the most critical ) .
such sensitivity analyses can also help `` estimate '' the variability in @xmath22 .
we thank penny j. hitchcock , norman l. johnson , krastan b. blagoev , and the t-11 epidemiology discussion group at los alamos national laboratory and hans frauenfelder for enhancing our ability to carry out this research .
we also thank fred brauer , simon levin , james watmough ( who reconfirmed our value of @xmath22 by the method @xcite ) , carlos w. castillo - garsow , and miriam nuno for their recent comments .
this research has been supported through the center for nonlinear studies at los alamos national lab under department of energy contract w-7405-eng-36 and partially supported by nsf , nsa and sloan foundation grants to carlos castillo - chavez . during the final stages of preparation , it came to our attention that prof .
roy anderson is examining similar questions about sars outbreak dynamics .
99 r. m. anderson and r. m. may , _ infectious diseases of humans _ , oxford university press ( 1991 ) . c. castillo - chavez , h. w. hethcote , v. andreasen , s. a. levin , w. m. liu .
_ epidemiological models with age structure , proportionate mixing , and cross - immunity _ , j. math .
27 , 233 - 258 ( 1989 ) .
bradsher , keith .
`` carrier of sars made @xmath1 flights before treatment . '' _ the new york times . _ april 11 , 2003 .
mcneil , donald g. and lawrence k. altman .
`` how one person can fu el an epidemic . '' _ the new york times . _
april 15 , 2003 .
schoen , john w. `` sars business impact spreading . ''
_ msn b news service .
_ april 2 , 2003 .
`` record sars deaths in hong kong . ''
_ msnbc news service .
_ website : http://www.msnbc.com/news/885653.asp .
april 15 , 2003 .
`` sars hits airlines , qantas cuts jobs . ''
_ msnbc news service .
_ website : http://www.msnbc.com/news/897719.asp .
april 9 , 2003 . c. drosten ,
_ , `` identification of a novel coronavirus in patients with severe acute respiratory syndrome , '' _ new england j. med . _ ( 2003 )
. t. g. ksiazek , _ et al .
_ , `` a novel coronavirus associated with severe acute respiratory syndrome , '' _ new england j. med . _ ( 2003 )
. j. l. gerberding , `` faster . . .
but fast enough ? '' _ new england j. med . _
pomfret , john .
`` doctor says health ministry lied about disease . '' _ the washington post .
_ april 9 , 2003 .
`` time line . ''
_ msnbc news service .
_ website : http://www.m snbc.com/news/885653.asp .
record sars deaths in hong kong ; msnbc news service ( april @xmath107 , @xmath4 ) .
stein , rob .
`` scientists identify cause of outbreak . ''
_ the washington post
. _ april 17 , 2003 .
`` iaff issues sars guidelines for first responders . ''
_ fire chief magazine .
_ website : http://www.firechief.com .
april 9 , 2003 .
`` symptoms and treatment of sars . ''
_ c health .
_ website : http://chealth.canoe.ca/ stein , rob .
faq on sars . "
_ the washington post .
_ website : http://www.washingtonpost.com/wp-dyn/articles/a11849-2003apr2.html .
sloan , rennie .
`` will sars wreak havoc here ? . '' _ msnbc n ew service .
_ website : http://www.msnbc.com/news/894251.asp .
april 3 , 2003 .
kalb , claudia .
`` tracking sars . '' _
newsweek_. april 28 , 2003 .
private communication , april 21st , 2003 . international society for infectious diseases .
promed - mail .
archive number 20030325.0744 .
_ sars - worldwide _ ( march 25 , 2003 ) .
cdc ( centers for disease control ) : _ basic information about sars_. website : http://www.cdc.gov/ncidod/sars/factsheet.htm ( april 16th , 2003 ) .
bradsher , keith .
`` sars is here to stay , hong kong and singapore t ell citizens . ''
_ the new york times . _ april 8 , 2003 . f. brauer , c. castillo - chavez , _ mathematical models in population biology and epidemiology _ , springer verlag new york ( 2000 ) .
o. diekmann , j. a. p. heesterbeek , _ mathematical epidemiology of infectious diseases : model building , analysis and interpretation _ , wiley ( 2000 ) .
canadian ministry of health .
summary of severe acute respiratory syndrome ( sars ) cases : canada and international .
http://www.hc-sc.gc.ca/pphb-dgspsp/sars-sras/eu-ae/ international society for infectious diseases .
promed - mail .
archive number 20030416.0930 .
_ sars - worldwide _ ( april 16th , 2003 ) .
coomer , john .
suspected sars cases near 150 in u.s . _ the reporter .
vanderbuilt medical center .
_ april 11 , 2003 .
pomfret , john .
`` underreporting , secrecy fuel sars in b eijing , who says . '' _ the washington post .
_ april 17 , 2003 . world health organization .
cumulative number of reported probable cases of severe acute respiratory syndrome ( sars ) .
http://www.who.int/csr/sarscountry/en/ cdc ( centers for disease control ) : _ isolation and quarantine_. website : http://www.cdc.gov/ncidod/sars/isolationquarantine.htm ( april 3 , 2003 ) . cdc ( centers for disease control ) : _ information for close contacts of sars patients _ .
website : http://www.cdc.gov/ncidod/sars/factsheetcc.htm ( april 16th , 2003 ) .
cdc ( centers for disease control ) : _ sars : frequently asked questions . _
website : http://www.cdc.gov/ncidod/sars/faq.htm .
april 16 , 2003 .
pomfret , john .
`` china seals hospitals to fight sars . ''
_ the washington post .
_ april 25 , 2003 .
`` ministers may review sars status . ''
_ bbc news .
_ http://news.bbc.co.uk/1/hi/health/2979623.stm ( april 27 , 2003 ) .
table 1 . parameter definitions and values that fit the cumulative number of cases in class @xmath45 ( `` diagnosed '' ) for hong kong .
these parameters are used to compute the basic reproductive number @xmath22 .
+ [ cols="^,<,^",options="header " , ] + | in this article we use global and regional data from the sars epidemic in conjunction with a model of susceptible , exposed , infective , diagnosed , and recovered classes of people ( `` seijr '' ) to extract average properties and rate constants for those populations .
the model is fitted to data from the ontario ( toronto ) in canada , hong kong in china and singapore outbreaks and predictions are made based on various assumptions and observations , including the current effect of isolating individuals diagnosed with sars .
the epidemic dynamics for hong kong and singapore appear to be different from the dynamics in toronto , ontario .
toronto shows a very rapid increase in the number of cases between march 31st and april 6th , followed by a _
significant _ slowing in the number of new cases .
we explain this as the result of an increase in the diagnostic rate and in the effectiveness of patient isolation after march 26th .
our best estimates are consistent with sars eventually being contained in toronto , although the time of containment is sensitive to the parameters in our model .
it is shown that despite the empirically modeled heterogeneity in transmission , sars average reproductive number is @xmath0 , a value quite similar to that computed for some strains of influenza @xcite .
although it would not be surprising to see levels of sars infection higher than ten per cent in some regions of the world ( if unchecked ) , lack of data and the observed heterogeneity and sensitivity of parameters prevent us from predicting the long - term impact of sars .
the possibility that 10 or more percent of the world population _ at risk _ could eventually be infected with the virus in conjunction with a mortality rate of three - to - seven percent or more , and indications of _ significant _ improvement in toronto support the stringent measures that have been taken to isolate diagnosed cases . |
quantum cascade lasers ( qcls ) are high - power , coherent light sources emitting in the mid - infrared ( mid - ir ) and terahertz ( thz ) frequency ranges @xcite .
qcls are electronically driven , unipolar devices whose active core consists of tens to hundreds of repetitions of a carefully designed stage .
the qcl active core can be considered a superlattice ( sl ) , in which each stage is a multiple - quantum - well ( mqw ) heterostructure , where confined electronic states with specific energy levels are formed because of quantum confinement .
the concept of achieving lasing in semiconductor sls was first introduced by kazarinov and suris @xcite in 1971 .
the first working qcl was demonstrated by faist _
. _ @xcite two decades later .
qcls are typically iii - v material systems grown on gaas or inp substrates . molecular beam epitaxy ( mbe ) @xcite and metal - organic chemical vapor deposition ( mocvd )
@xcite are the techniques that enable precise growth of thin layers of various iii - v alloys .
it is also possible to incorporate strain into the structure , as long as the total strain in a stage is balanced .
both the precision and the possibility of introducing strain bring great flexibility to the design of the qcl active core , so lasing over a wide range of wave lengths ( from 3 to 190 @xmath0 m ) has been achieved .
the growth techniques produce high - quality interfaces , with atomic - level roughness .
mid - ir qcls ( wave - length range @xmath1 m ) have widespread military and commercial applications .
a practical portable detector requires mid - ir qcls to operate at room - temperature ( rt ) , in continuous - wave ( cw ) mode , and with high ( watt - level ) output power .
furthermore , these qcls must also have high wall - plug efficiency ( wpe , the ratio of emitted optical power to the electrical power pumped in ) and long - term reliability under these high - stress operating conditions . as the stress likely stems from excessive nonuniform heating
while lasing @xcite , improving device reliability and lifetime goes hand - in - hand with improving the wpe . in qcls ,
multiple conduction subbands are formed in the active core by means of quantum confinement .
qcls are unipolar devices , meaning that lasing is achieved through radiative intersubband transitions ( transitions between two conduction subbands ) instead of radiative interband transitions ( transitions between the conduction and valence bands ) in traditional quantum well ( qw ) semiconductor lasers . as a result
, electrons do not combine with holes after the radiative transitions and can be used to emit another photon . in order to reuse electrons ,
the same mqw heterostructure is repeated many times ( 2570 ) in the qcl active core ( the so - called cascading scheme ) .
figure [ fig : bandstructure ] depicts a typical conduction - band diagram of two adjacent stages in a qcl under an electric field .
each stage consists of an injector region and an active region .
the injector region has several thin wells separated by thin barriers ( @xmath2 ) , so a miniband is formed , with multiple subbands that are close in energy and whose associated wavefunctions have high spatial overlap .
typically , the lowest few energy levels in the miniband are referred to as the injector levels .
the injector levels collect the electrons that come from the previous stage and inject them into the active region .
the active region usually consists of 23 wider wells ( @xmath3 ) separated by thin barriers .
consequently , a minigap forms in the active region between the upper lasing level ( 3 ) and the lower lasing level ( 2 ) .
another important energy level in the active region is the ground state ( 1 ) .
there is a thin barrier ( usually the thinnest among all layers ) between the injecting region and the active region , called the injection barrier . by design ,
the injector levels are close in energy and strongly coupled to the upper lasing level because of the thin injection barrier .
the upper and lower lasing levels have large spatial overlap , which allows a radiative transition between the two levels ; the wave length of the emitted light is determined by the energy spacing between these two levels .
the lower lasing level overlaps with the ground state for efficient electron extraction .
electron emission of longitudinal optical ( lo ) phonons is the dominant mechanism for electron extraction , so the energy spacing between the lower lasing level and the ground state is designed to be close to the lo phonon energy to facilitate extraction . with careful design ,
the electron lifetime in the upper lasing level is longer than in the lower lasing level , so population inversion can be achieved . after reaching the ground state ,
electrons tunnel through the injector into the the upper lasing level of the next stage , and the process is repeated .
of course , the lasing mechanism description above is idealized . in reality
, the efficiency of the radiative transition between the upper and lower lasing levels is very low @xcite . in recent years
, considerable focus has been placed on improving the wpe and output power of qcls for rt cw operation .
@xcite showed 8.4 % wpe and 1.3 w output power around 4.6 @xmath0 m in 2008 . shortly thereafter , lyakh _ et al . _
@xcite reported 12.7% wpe and 3 w power at 4.6 @xmath0 m .
watt - level power with 6% wpe at 3.76 @xmath0 m and then lower power at 3.39 @xmath0 m and 3.56 @xmath0 m are reported by bandyopadhyay _
_ @xcite demonstrated 21% wpe and 5.1 w output power around 4.9 @xmath0 m in 2011 .
much higher wpe and/or output power has been achieved at lower temperatures or at pulsed mode @xcite near 4.8 @xmath0 m .
a summary of recent developments can be found in review papers @xcite .
while good output powers and wpes have been achieved , long - term reliability of these devices under rt cw operation remains a critical problem @xcite .
these devices are prone to catastrophic breakdown owing to reasons that are not entirely understood , but are likely related to thermal stress that stems from prolonged high - power operation @xcite .
this kind of thermal stress is worst in short - wave length devices that have high strain and high thermal impedance mismatch between layers @xcite .
in addition to improved device lifetime , we seek better cw temperature performance ( higher characteristic temperatures @xmath4 and @xmath5 , defined below ) @xcite .
the first aspect is a weaker temperature dependence of the threshold current density .
empirically , the threshold current density ( the current density at which the device starts lasing ) has an exponential dependence on the operating temperature @xmath6 : @xmath7 .
higher characteristic temperature @xmath4 is preferred in qcl design , as it means less variation in @xmath8 as the temperature changes .
another key temperature - dependent parameter is the differential quantum efficiency ( also called the slope efficiency of external quantum efficiency ) , defined as the amount of output optical power @xmath9 per unit increase in the pumping current @xmath10 : @xmath11 .
the differential quantum efficiency is directly proportional to the wpe ( @xmath12 , where @xmath13 is the feeding efficiency ) .
therefore , the higher the @xmath5 , the closer @xmath14 is to unity , and the higher the wpe .
recently , deep - well structures with tapered active regions have demonstrated significant improvements in @xmath4 and @xmath5 with respect to the conventional @xmath15 device @xcite , underscoring that the suppression of leakage plays a key role in temperature performance @xcite . still , the microscopic mechanisms and leakage pathways that contribute to these empirical performance parameters remain unclear . under high - power , rt cw operation , both electron and phonon systems in qcls are far away from equilibrium . in such nonequilibrium conditions , both electronic and thermal transport modeling are important for understanding and improving qcl performance .
electron transport in both mid - ir and thz qcls has been successfully simulated via semiclassical ( rate equations @xcite and monte carlo @xcite ) and quantum techniques ( density matrix @xcite , nonequilibrium green s functions ( negf ) @xcite , and lately wigner functions @xcite ) .
inp - based mid - ir qcls have been addressed via semiclassical @xcite and quantum transport approaches ( 8.5-@xmath16 @xcite and 4.6-@xmath16 @xcite devices ) .
there has been a debate whether electron transport in qcls can be described using semiclassical models , in other words , how much of the current in qcls is coherent .
theoretical work by iotti and rossi @xcite show that the steady - state transport in mid - ir qcls is largely incoherent .
monte carlo simulation @xcite has also been used to correctly predict transport near threshold . however , short - wavelength structures @xcite have pronounced coherent features , which can not be addressed semiclassically @xcite .
negf simulations accurately and comprehensively capture quantum transport in these devices , but are computationally demanding .
density - matrix approaches have considerably lower computational overhead than negf , but are still capable of capturing coherent - transport features .
a comprehensive review of electron - transport modeling was recently written by jirauschek and kubis @xcite .
electronic simulations that ignore radiative transitions are applicable for modeling qcls below or near threshold , where the interaction between electrons and the laser electromagnetic field can be ignored .
such simulations are useful for predicting quantities such as threshold current density and @xmath4 . however , in order to accurately model qcls under lasing operations , the effect of the laser field on electronic transport would have to be included . in some cases , the effects of the laser field can be very strong @xcite , especially for high wpe devices , where the field - induced current can be dominant @xcite . when included in simulations ,
the laser field is typically either modeled as an additional scattering mechanism @xcite or as a time - dependent sinusoidal electric field @xcite . in this work
, we ignore the effect of the laser field on electron dynamics .
thermal transport in qcls is often described through the heat diffusion equation , which requires accurate thermal conductivity in each region , a challenging task for the active core that contains many interfaces @xcite .
it is also very important to include nonequilibrium effects , such as the nonuniform heat - generation rate stemming from the nonuniform temperature distribution @xcite and the feedback that the nonequilibrium phonon population has on electron transport @xcite . in this chapter ,
we present a multiphysics ( coupled electronic and thermal transport ) and multiscale ( bridging between a single stage and device level ) simulation framework that enables the description of qcl performance under far - from - equilibrium conditions @xcite .
we present the electronic ( sec .
[ sec : electron ] ) and thermal ( sec .
[ sec : thermal ] ) transport models , then bring them together for electrothermal simulation of a real device structure ( sec .
[ sec : example ] ) .
we strive to cover the basic ideas while pointing readers to the relevant references for derivation and implementation details .
depending on the desired accuracy and computational burden , one can model electronic transport in qcls with varying degrees of complexity .
the goal is to determine the modal gain ( proportional to the population inversion between the upper and lower lasing levels ) under various pumping conditions ( current or voltage ) and lasing conditions ( pulsed or continuous wave ) .
a typical electron transport simulator relies on accurately calculated quasibound electronic states and associated energies in the direction of confinement .
electronic wavefunctions and energies are determined by solving the schrdinger equation or the schrdinger equation combined with poisson s equation in highly doped systems .
section [ sec : kp ] introduces a @xmath17 schrdinger solver coupled with a poisson solver .
more information about other solvers for electronic states can be found in the review paper @xcite and references therein .
the simulations of electronic transport fall into two camps depending on how the electron single - particle density matrix is treated .
the diagonal elements of the density matrix represent the occupation of the corresponding levels and off - diagonal elements represents the `` coherence '' between two levels .
transport is semiclassical or incoherent when the off - diagonal coherences are much smaller than the diagonal terms , and can be approximated as proportional to the diagonal terms times the transition rates between states @xcite . in that case , the explicit calculation of the off - diagonal terms is avoided and only the diagonal elements are tracked , which simplifies the simulation considerably .
when the off - diagonal terms are appreciable , transport is partially coherent and has to be addressed using quantum - transport techniques , discussed below .
semiclassical approaches assume that electronic transport between stages is largely incoherent `` hopping '' transport .
the key quantities are populations of electronic states that are confined in the qcl growth direction , and electrons transfer between them due to scattering events . the scattering rates can be obtained empirically or more rigorously , via fermi s golden rule .
common semiclassical approaches are the rate equations and ensemble monte carlo ( emc ) , the latter solving a boltzmann - like transport equation stochastically . in the rate - equation approach @xcite ,
scattering between relevant states , i.e. , the injector level , the upper and lower lasing levels , and the ground state , is captured through transition rates .
the rates include all relevant ( radiative and nonradiative ) scattering mechanisms , and can be either empirical parameters or calculated @xcite .
the computational requirements of rate - equation models are low , so they are suitable for fast numerical design and optimization of different structures @xcite .
the heterostructure in the qcl active core is a quasi - two - dimensional ( quasi-2d ) system , where electrons are free to move in the @xmath18 plane , while confined cross - plane , in the @xmath19direction ; the confinement results in the formation of quasibound states and discrete energy levels corresponding to the bottoms of 2d energy subbands .
the electron wavefunctions in 3d are plane waves in the @xmath18 plane and confined wavefunctions in @xmath20 .
electronic transport is captured by a boltzmann - like semiclassical transport equation @xcite , which can be solved via the stochastic emc technique assuming instantaneous hops between states in 3d due to scattering @xcite .
the simulation explicitly tracks the energy level and in - plane momentum of each particle in the simulation ensemble ( typically @xmath21 particles ) .
tracking in - plane dynamics makes it more detailed than the rate - equation model .
the transition rates are generally computed directly from the appropriate interaction hamiltonians , and therefore depend on the energy levels as well as the wavefunction overlaps between different electronic states @xcite .
emc allows us to include nonequilibrium effect into transport , which is covered in more detail in sec .
[ sec : emc ] .
density matrix and negf are the two most widely used techniques to describe quantum transport in qcls .
recently , a wigner - function approach was also successfully used to model a superlattice @xcite . in semiclassical approaches ,
the central quantity of interest is the distribution function @xmath22 , the probability of an electron occupying an eigenstate @xmath23 and having an in - plane kinetic energy @xmath24 .
the quantum - mechanical analogue is the single - electron density matrix , @xmath25 , where the diagonal elements @xmath26 are occupations and the off - diagonal elements @xmath25 are the spatial coherences between states @xmath23 and @xmath27 at the in - plane energy @xmath24 .
when employing semiclassical methods , off - diagonal matrix elements are assumed to be much smaller than diagonal elements .
this approximation may fail in some cases , for example , when two eigenstates with a large spatial overlap have similar energies .
this scenario often arises when modeling terahertz qcls @xcite , but can also come up in mid - ir qcls @xcite . in these cases ,
semiclassical models fail .
the density - matrix models that have been employed for qcl modeling can be categorized into two groups .
the first includes hybrid methods , where transport is treated semiclassically within a region of the device ( typically a single stage ) while the effects of tunneling between different regions , separated by barriers , is treated quantum mechanically using a density - matrix formalism with phenomenological dephasing times @xcite .
the second group involves completely quantum - mechanical methods that rely on microscopically derived markovian master equations that guarantee positivity of the density matrix @xcite .
both methods are more computationally expensive than their semiclassical counterparts , because the density matrix contains many more elements than its diagonal semiclassical analogue .
the nonequilibrium green s function technique ( see a good overview in @xcite ) relies on the relationships between single - particle time - ordered green s functions and correlation functions @xcite .
the correlation function @xmath28 , often referred to as the lesser green s function @xcite , is one of the central quantities and can be understood as a two - time generalization of the density matrix , where @xmath29 refers to the magnitude of in - plane wave vector .
the correlation function contains both spatial correlations ( terms with @xmath30 ) as well as temporal correlations between times @xmath31 and @xmath32 ( not included in semiclassical or density - matrix models ) .
typically , the potential profile is assumed to be time independent , in which case the correlation function only depends on the time difference @xmath33 .
fourier transforming over the time difference into the energy domain gives the energy - resolved correlation function @xmath34 , which is the quantity which is usually solved for numerically @xcite .
the main advantages of the negf formalism are that it provides spectral ( energy - resolved ) information and it includes the effects of collisional broadening ( the broadening of energy levels due to scattering ) , which is particularly important when the states are close in energy .
these advantages carry a considerable computational cost , so negf calculations are much more time consuming than density - matrix approaches @xcite . here , we focus on presenting semiclassical modeling of electron transport in qcl structures via emc @xcite .
the solver consists of two parts , a coupled schrdinger poisson solver and a transport kernel .
we solve for the electronic states using the coupled schrdinger
poisson solver and feed the energy levels and the wavefunctions of the relevant electronic states to the transport kernel .
the transport kernel keeps track of the electron momentum , energy , and distribution among subbands .
if the electron density inside the device is high , transport kernel will periodically feed the electron distribution back to the schrdinger
poisson solver and update the electronic states .
this loop is repeated until the electron distribution converges . by doing so
, we solve for both electron transport and the electronic band structure self - consistently .
since the active qcl core consists of repeated stages , the wavefunctions in any stage can be obtained from the wavefunctions in any other stage by translation in space and energy .
this translational symmetry makes it possible to simulate electron transport in only one generic central stage instead of in the whole qcl core @xcite .
typically , electronic states in nonadjacent stages have negligible overlap , which also means that the transition rates between them are negligible . as a result , it is sufficient to limit interstage scattering events to only those between adjacent stages .
figure [ fig : emcsche ] shows a schematic of three adjacent stages under an applied field .
we simulate electron transport in the central stage @xmath35 , while nearest - neighbor interstage ( @xmath36 ) and intrastage ( @xmath37 ) scattering is allowed .
periodic boundary conditions ( pbcs ) are applied in the simulation , i.e. , whenever one electron scatters from the central stage out to the next stage ( process @xmath38 ) , an electron scatters from the previous stage into the central stage ( process @xmath39 ) and vice versa ( process @xmath40 and process @xmath41 ) .
pbcs are justified by the cascading scheme .
we employ the @xmath17 method to solve the schrdinger equation and couple it to a poisson solver @xcite .
the @xmath17 method is an efficient way to solve for the electronic band structure near the band edges , where the transport happens in qcls .
the @xmath17 method considers the contribution from the conduction band ( c ) , light - hole band ( lh ) , and the spin - orbit split - off band ( so ) ( the heavy - hole band ( hh ) decouples from the other three at the band edge ) @xcite .
the contributions from the lh and so are especially important for narrow - gap materials , such as inp .
moreover , in modern qcls , strain - balanced structures have been employed to obtain enhanced performance . in these structures , alternate layers
are compressively or tensilely strained while the whole structure is strain free , with carefully designed thickness of each layer .
the @xmath17 method allows for convenient inclusion of the effects of strain on the band structure .
the implementation details of the @xmath17 solver can be found in @xcite .
the @xmath17 solver can only solve for a finite structure rather than an infinite periodic one . as a result
, we need to simulate a finite number of stages and add artificially high barriers to the two ends to confine all the states .
if a stage is far enough from the boundaries , the calculated band structure in it should be the same as if we were to solve for the whole periodic structure .
tests have confirmed that three stages , which we also use in emc , are enough when solving for the electronic states to ensure that the central - stage states are unaffected by the simulation - domain potential boundaries .
the states from the central stage are then translated in energy and position to the neighboring stages according to the stage length and the applied electric field .
when we need to solve for electron transport and electronic states self - consistently , it is necessary for the solver to be able to automatically pick out the electronic states belonging to the central stage .
one intuitive criterion is to calculate the `` center of mass '' for each state ( the expectation value of the cross - plane coordinate , @xmath42 ) and assign those falling in the central stage to that stage .
however , in our three - stage scheme , this method may pick up the states that are too close to the boundary .
one can either extend the number of stages in the @xmath17 solver to five , so the three stages in the middle are all far from the boundary , or use additional criteria , such as that there be more than 50% possibility of finding an electron in the central stage , based on the probability density distribution , or requiring that the location of the probability - density peak be in the central stage .
additional criteria requiring strong confinement of states have been explored in @xcite .
the emc kernel tracks the hopping transitions of electrons between subbands and stages until convergence , and outputs the transport information for us to calculate the experimentally relevant quantities such as current and modal gain @xcite . in the transport kernel , both electron electron interactions and electron
lo - phonon interactions are considered .
other scattering processes such as intervalley scattering , impurity scattering , interface roughness scattering can be considered under different circumstances @xcite .
photon emission is not considered , either . because emc tracks individual particles , nonequilibrium electron transport can be automatically captured .
( emc tracks individual simulation particles , each of which might represent thousands of real electrons . )
the most important scattering mechanism in qcls is electron lo - phonon scattering , which facilitates the depopulation of the lower lasing level .
as shown in @xcite , phonon confinement has little effect on the electronic transport , therefore , for simplicity , lo phonons are treated as bulklike dispersionless phonons with energy @xmath43 .
the transition rate between an initial state @xmath44 with energy @xmath45 and a final state @xmath46 with @xmath47 can be derived from fermi s golden rule as @xmath48 where @xmath49 is the electronic charge while @xmath50 and @xmath51 are static and high - frequency electronic permittivities of the material , respectively .
the integrals are over the in - plane kinetic energy @xmath24 of the final state and the cross - plane momentum transfer @xmath52 .
@xmath53 is the in - plane momentum transfer .
latexmath:[\[\label{eq : oi } is defined as the overlap integral ( oi ) between the initial and final states , where @xmath52 is the cross - plane momentum transfer .
the integration is over the angle between initial and final in - plane momenta @xmath55 and @xmath56 ( @xmath57 ) , cross - plane momentum component of the final state ( @xmath58 ) , and the kinetic energy of the final state ( @xmath59 ) .
@xmath60 represents the number of lo phonons with momentum @xmath61 .
the expression can be further simplified in the equilibrium case , where @xmath60 follows the bose - einstein distribution @xcite . in order to model nonequilibrium phonon effects
, we numerically integrate the expression using a phonon number histogram according to both @xmath62 and @xmath52 @xcite . according to the uncertainty principle
, position and momentum can not both be determined simultaneously .
since our electrons are all confined in the central stage ( @xmath63 is finite ) , the cross - plane momentum is not exactly conserved during the scattering process ( @xmath64 ) @xcite .
this analysis does not affect the momentum conservation in the @xmath18 plane , because we assume infinite uncertainty in position there .
previously , the cross - plane momentum conservation has been considered through the momentum - conservation approximation ( mca ) @xcite and a broadening of @xmath52 according to the well width @xcite .
the mca forbids a phonon emitted between subbands @xmath65 and @xmath66 to be re - absorbed by another transition between @xmath67 and @xmath68 if @xmath69 or @xmath70 , and thus might underestimate the electron - lo interaction strength @xcite .
the concept of well width is hard to apply in a mqw structure such as the qcl active core @xcite .
we observe that the probability of a phonon with cross - plane momentum @xmath52 being involved in an interaction is proportional to the overlap integral in equation ( [ eq : oi ] ) .
figure [ fig : overlap ] depicts the typical overlap integrals for both intersubband ( @xmath71 and @xmath72 ) and intrasubband ( @xmath73 ) transitions . as a result , in each electron lo - phonon scattering event , we randomly select a @xmath52 following the distribution from the overlap integral ( fig .
[ fig : overlap ] ) . depending on the mechanism ( absorption or emission ) , a phonon with ( @xmath74 ) is removed / added to the histogram according to the 2d density of states ( dos ) and the effective simulation area @xcite . once the phonons with a certain momentum are depleted , transitions involving such phonons become forbidden . from eq . ( [ eq : oi ] ) versus cross - plane phonon wave vector @xmath52 for several transitions ( intersubband @xmath71 and @xmath72 ; intrasubband @xmath73 ) .
figure reproduced from @xcite , y. b. shi and i. knezevic , j. appl .
116 , 123105 ( 2014 ) , with the permission of aip publishing.,title="fig : " ] in order to couple the emc solver to the thermal transport solver , we need to keep a detailed log of heat generation during electron transport . in all the relevant scattering events ,
electron lo - phonon scattering is the only inelastic mechanism and therefore is the only mechanism that contributes to heat generation . as a result ,
the total energy emitted and absorbed in the form of lo phonons is recorded during each step of the emc simulation . the nonequilibrium phonons decay into acoustic longitudinal acoustic ( la ) phonons via a three - phonon anharmonic decay process .
the formulation and the parameters here follow @xcite .
the simulation results of emc including nonequilibrium phonons are shown in section [ sec : example ] .
the dominant path of heat transfer in a qcl structure is depicted in fig .
[ fig : heatflow ] .
the operating electric field of a typical qcl is high , which means that considerable energy is pumped into the electronic system . these energetic , `` hot '' electrons relax their energy largely by emitting lo phonons .
lo phonons have high energies but flat dispersions , so their group velocities are low and they are poor carriers of heat .
an lo phonon decays into two la phonons via a three - phonon process referred to as anharmonic decay .
la phonons have low energy but high group velocity and are the main carriers of heat in semiconductors @xcite . if we neglect the diffusion of optical phonons , the flow of energy in a qcl can be described by the equations @xmath75 where @xmath76 , @xmath77 , and @xmath78 are the lo phonon , acoustic phonon , and electron energy densities , respectively .
@xmath79 is the thermal conductivity in the system and @xmath80 is the acoustic - phonon ( lattice ) temperature .
the term @xmath81 describes heat diffusion , governed by acoustic phonons .
we have also used the fact that the rate of increase in the lo - phonon energy density equals the difference between the rate of its generation by electron lo - phonon scattering and the rate of anharmonic decay into la phonons . in a nonequilibrium steady state ,
both the lo and la energy densities are constant , so @xmath82 as described in the previous section , the right - hand side of equation ( [ eq : heatdiff ] ) is the heat - generation rate @xmath83 and can be obtained by recording electron lo - phonon scattering events in electronic emc @xcite @xmath84 where @xmath85 is the electron density ( @xmath86 is the sheet density and @xmath87 is the length of a single stage ) while @xmath88 and @xmath89 are the number of simulation particles and the simulation time , respectively . @xmath90 and @xmath91 are the energies of the emitted and absorbed lo phonons , respectively . to solve equation ( [ eq : heatdiff ] ) ,
we need information on both the thermal conductivity @xmath79 and the heat - generation rate @xmath83 ; they are discussed in subsections [ sec : iiivsl ] and [ sec : device ] , respectively .
the qcl active core is a sl : it contains many identical stages , each with several thin layers made from different materials and separated by heterointerfaces .
the thermal - conductivity tensor of a sl system reduces to two values : the in - plane thermal conductivity @xmath92 ( in - plane heat flow is assumed isotropic ) and the cross - plane thermal conductivity @xmath93 .
experimental results have shown that , in sls , the thermal conductivity is very anisotropic @xcite ( @xmath94 ) while both @xmath95 are smaller than the weighted average of the constituent bulk materials @xcite .
both effects can be attributed to the interfaces between adjacent layers @xcite . here , we discuss a semiclassical model for describing the thermal - conductivity tensor of iii - v sl structures .
note that the model described here is in principle applicable to sls in other material systems , as long as they have high - quality interface and thermal transport is mostly incoherent @xcite .
in particular , we focus on thermal transport in iii - arsenide - based sls , as they are most commonly used in mid - ir - qcl active cores @xcite . under qcl operation conditions of interest ( @xmath96 k , and typically near rt ) , thermal transport
is dominated by acoustic phonons and is governed by the boltzmann transport equation ( bte ) . to obtain the thermal conductivity , we solve the phonon bte with full phonon dispersion in the relaxation - time approximation @xcite . to capture both the anisotropic thermal transport and the reduced thermal conductivity in sl systems , we need to observe the twofold influence of the interface .
first , it reduces @xmath92 by affecting the acoustic - phonon population close to the interfaces @xcite .
second , it introduces an interface thermal boundary resistance ( itbr ) , which is still very difficult to model @xcite .
common models are the acoustic mismatch model ( amm ) and the diffuse mismatch model ( dmm ) @xcite ; the former assumes a perfectly smooth interface and only considers the acoustic mismatch between the two materials , while the latter assumes complete randomization of momentum after phonons hit the interface . as most iii - v based qcls are grown by mbe or mocvd , both well - controlled techniques allowing consistent atomic - level precision , neither amm nor dmm captures the essence of a iii - v sl interface .
figure [ fig : slsche ] shows a schematic of interface roughness in a lattice - matched sl .
the jagged dashed boundaries depict transition layers of characteristic thickness @xmath97 between the two materials . .
this effective interface roughness leads to phonon - momentum randomization and to interface resistance in cross - plane transport .
figure reproduced from @xcite , s. mei and i. knezevic , j. appl .
118 , 175101 ( 2015 ) , with the permission of aip publishing.,title="fig : " ] we introduce a simple model that calculates a more realistic itbr ( a key part in calculating @xmath93 ) by interpolating between the amm and dmm transmission rates using a specularity parameter @xmath98 .
the model has a single fitting parameter : the effective interface rms roughness @xmath97 . since the growth environment is well controlled ,
using one @xmath97 to describe all the interfaces is justified .
we use @xmath97 to calculate a momentum - dependent specularity parameter @xmath99 where @xmath100 is the magnitude of the phonon wave vector and @xmath57 is the angle between @xmath101 and the normal direction to the interface .
consistent with the twofold impact of interface roughness , @xmath97 affects the thermal conductivity through two channels .
apart from calculating the itbr , an effective interface scattering rate @xmath102 dependent on the same specularity parameter @xmath103 is added to the internal scattering rate to calculate modified @xmath92 ( see detailed derivations in @xcite ) . by adjusting only @xmath97 , typically between 1 - 2
, the calculated thermal conductivity using this model fits a number of different experiments @xcite .
thermal transport inside the active core of a qcl is usually treated phenomenologically : @xmath92 is typically assumed to be 75% of the weighted average of the bulk thermal conductivities of the constituent materials , while @xmath93 is treated as a fitting parameter ( constant for all temperatures ) to best fit the experimentally measured temperature profile @xcite .
we calculated the thermal - conductivity tensor of a qcl active core @xcite and showed that the typical assumption is not accurate and that the degree of anisotropy is temperature dependent ( fig .
[ fig : sltc ] ) .
ga@xmath106as and in@xmath107al@xmath108as .
the solid curve , dashed curve , and dashed - dotted curve show the calculated in - plane , cross - plane , and averaged bulk thermal conductivity , respectively .
@xmath109 in the calculations .
the inset shows the ratio between the calculated in - plane and the averaged bulk thermal conductivities .
figure reproduced from @xcite , s. mei and i. knezevic , j. appl .
118 , 175101 ( 2015 ) , with the permission of aip publishing.,title="fig : " ] the ratio between @xmath92 and the averaged bulk value ( inset to fig .
[ fig : sltc ] ) varies between 45% and 70% over the temperature range of interest .
@xmath93 has a weak dependence on temperature , in keeping with the common assumption in simplified models ; the weak temperature sensitivity means that itbr dominates cross - plane thermal transport .
these results show that it is important to carefully calculate the thermal - conductivity tensor in qcl thermal simulation and we will use this thermal - conductivity model in the device - level simulation .
the active core is not the only region we need to model in a device - level thermal simulation .
figure [ fig : qclsche ] shows a typical schematic ( not to scale ) of a qcl device in thermal simulation with a substrate - side mounting configuration @xcite .
the active core ( in this case , consisting of 36 stages and 1.6 @xmath0 m thick ) with width @xmath110 is embedded between two cladding layers ( 4.5-@xmath0m - thick gaas ) . the waveguide is supported by a substrate ( gaas ) with thickness @xmath111 . an insulation layer ( si@xmath112n@xmath113 ) with thickness @xmath114 is deposited around the waveguide and then etched away from the top to make the contact .
finally , a contact layer ( au ) with thickness @xmath115 and a thin layer of solder ( @xmath116 ) are deposited on top .
there is no heat generation in the regions other than the active core .
further , these layers are typically thick enough to be treated as bulk materials . bulk - substrate ( gaas or inp ) thermal conductivities are readily obtained for iii - v materials from experiment , as well as from relatively simple theoretical models @xcite ( table [ table : bulktc ] ) .
.thermal conductivity as a function of temperature for materials in a qcl structure .
[ cols="<,<",options="header " , ] the length of a qcl device is much greater than its width , therefore we can assume the length is infinite and carry out a 2d thermal simulation .
the schematic of the simulation domain ( not to scale ) is shown in fig .
[ fig : qclsche ] .
the boundary of the simulation region is highlighted in green , and certain boundary conditions ( heat sink at fixed temperature , convective boundary condition , or adiabatic boundary condition ) can be applied ( independently ) to each boundary .
typically , the bottom boundary of the device is connected to a heat sink while other boundaries have the convective boundary condition at the environment temperature ( single - device case ) or the adiabatic boundary condition ( qcl - array case ) .
typical values for the layers thickness are @xmath117 m , @xmath118 m , @xmath119 m , @xmath120 m , @xmath121 m .
we use the finite - element method to solve for the temperature distribution .
the whole device is divided into different regions according to their materials properties .
each stage of the active region is treated as a single unit with the heat - generation rate tabulated in the device table in order to capture the nonuniform behavior among stages .
the active core is very small , but is also the only region with heat generation , small thermal conductivity , and spatial nonuniformity . to capture the behavior of the active region
while saving computational time , we use a nonuniform mesh in the finite - element solver to emphasize the active core region .
figure [ fig : qclmesh ] shows a mesh generated in the simulation .
it is known that among all the stages in the active core , the temperature @xmath122 and the electric field @xmath123
( @xmath65 represents the stage index ) are not constant @xcite , but we have no _ a priori _ knowledge of how they depend on the stage index .
however , we know that the charge current continuity equation must hold , and in the steady state @xmath124 ; this implies that the current density @xmath125 must be uniform , as the current flow is essentially in one dimension , along @xmath20 .
this insight is key to bridging the single - stage and device - level simulations . from sec .
[ sec : emc ] , we can obtain the heat - generation rate @xmath83 inside the active core by running the single - stage emc simulation .
each single - stage emc is carried out at a specific electric field @xmath126 and temperature @xmath6 and outputs both the current density @xmath127 and the heat - generation rate @xmath128 . by sweeping @xmath126 and @xmath6 in range of interest ,
we obtain a table connecting different field and temperature @xmath129 to proper current density and heat - generation rate @xmath130 [ @xmath131 .
however , from the discussion above , the input in the thermal simulation needs to be the constant parameter @xmath125 .
therefore , we `` flip '' the recorded @xmath132 table to a so - called device table @xmath133 , suitable for coupled simulation @xcite .
figure [ fig : thermalflchart ] depicts the flowchart of the device - level electrothermal simulation @xcite . before the simulation
, we obtain the device table [ @xmath133 ] , as discussed above .
we also have to calculate the thermal conductivities ( @xmath92 and @xmath93 ) of the active region as a function of temperature and tabulate them , based on the model described in sec .
[ sec : iiivsl ] .
we also need the bulk thermal conductivity of other materials in the device ( cladding layer , substrate , insulation , contact , and solder ) as a function of temperature .
these material properties are standard and already well characterized . and temperature profile @xmath122 across the whole device . based on the tabulated information from the single - stage simulation and assumed @xmath135 , we get stage - by - stage profile for the electric field @xmath123 and the heat - generation rate @xmath136 profiles
. an accurate temperature - dependent thermal conductivity model , which includes the boundary resistances of layers , and the temperature profile guess are used as input to the heat diffusion equation , which is then iteratively solved ( with updated temperature profile in each step ) until the thermal boundary conditions are satisfied.,title="fig:",width=5 ] each device - level thermal simulation is carried out in a certain environment ( i.e. , for a given set of boundary conditions ) and with a certain current density @xmath125 . at the beginning of the simulation , an initial temperature profile is assigned . with the input from the device table and the thermal conductivity data in each region , we use a finite - element method to iteratively solve the heat diffusion equation until convergence . at the end of the simulation , we obtain a thermal map of the whole device .
further , from the temperature @xmath122 in each stage and the injected current density @xmath125 , we obtain the nonuniform electric field distribution @xmath123 . with the electric field in each stage and given the stage thickness , we can accurately calculate the voltage drop across the device and obtain the current voltage characteristic . by changing the mounting configuration ( @xmath110 , @xmath111 , @xmath114 , @xmath115 , @xmath116 ) or the boundary conditions ,
the temperature profile can be changed .
in this section , we present detailed simulation results of a 9-@xmath0 m gaas / al@xmath137ga@xmath138as mid - ir qcl @xcite based on a conventional three - well active region design .
the chosen structure has 36 repetitions of the single stage ; each stage has 16 layers . starting from the injection barrier ,
the layer thicknesses in one stage ( in @xmath139 ) are * 46*/19/*11*/54/*11*/48/*28*/34/*17*/30/**//**//*26*/30 . here ,
the barriers ( al@xmath137ga@xmath138as ) are in bold while the wells ( gaas ) are in normal font ; the underlined layers are doped to a sheet density of @xmath140 .
the results at 77 k are shown here .
figure [ fig : pageband ] shows the electronic states of the chosen structure under the design operating field of 48 kv / cm , calculated from the coupled @xmath17poisson solver ( see sec . [ sec : kp ] ) .
the active - region states of the central stage are represented in bold red curves ; 1 , 2 , and 3 are the ground state and the lower and upper lasing levels , respectively .
injector states are labeled @xmath141 and @xmath142 .
other blue states together form the miniband .
( when the electron density in the qcl is high , the electronic bands have to be calculated self - consistently with emc . )
-valley subbands in two adjacent stages of the simulated gaas / algaas - based structure .
the bold red curves denote the active region states ( 1 , 2 , and 3 represent the ground state and the lower and upper lasing levels , respectively ) .
the blue curves represent injector states , with @xmath141 and @xmath142 denoting the lowest two .
figure reproduced from @xcite , y. b. shi and i. knezevic , j. appl .
116 , 123105 ( 2014 ) , with the permission of aip publishing.,title="fig : " ] the current density @xmath125 vs field @xmath126 curve , one of the key qcl characteristics at a given temperature , is intuitive to obtain in emc . after calculating the electronic band structure at a certain field @xmath126
, the wavefunctions , energy levels , and effective masses of each subband and each stage are fed into the emc solver . in the emc simulation
, we include all the scattering mechanisms described in sec .
[ sec : emc ] .
since we employ periodic boundary conditions , the current density @xmath125 can be extracted from how many electrons cross the stage boundaries in a certain amount of time in the steady state .
the net flow @xmath144 of electrons is calculated by subtracting the flow between the central stage and the previous stage ( @xmath145 ) from the flow between the central stage and the next stage ( @xmath146 ) in each time step .
the current density is then calculated as @xmath147 where @xmath148 is the time interval during which the flow is recorded .
@xmath149 is the effective in - plane area of the simulated device .
since doping is the main source of electrons , the area is calculated as @xmath150 where @xmath151 is the number of simulated electrons and @xmath86 is the sheet doping density ( in @xmath152 ) in the fabricated device . in the current simulation , @xmath153 and @xmath154 . due to the stochastic nature of emc
, we need to average the current density over multiple time steps . in practice , one can record the net cumulative number of electrons per unit area that leave a stage over time and obtain a linear fit to this quantity in the steady state ; the slope yields the steady - state current density . from each individual simulation
, we extract the current density at a given electric field and temperature . to obtain the @xmath143 curve at that temperature , we sweep the electric field .
to demonstrate the importance of including nonequilibrium phonons effects , we carry out the simulation with thermal phonons alone and with both thermal and excess nonequilibrium phonons .
figure [ fig : jf ] is the @xmath143 curve for the simulated structure with ( filled squares ) and without ( empty squares ) nonequilibrium phonons at 77 k. it can be seen that the current density at a given field considerably increases when nonequilibrium phonons are included and the trend holds up to 60 kv / cm . this difference is prominent at low temperatures ( < 200 k ) and goes away at rt @xcite . )
curve of the simulated device with ( filled squares ) and without ( empty squares ) the nonequilibrium phonon effect at 77 k. the inclusion of nonequilibrium phonons considerably increases the current density at a given field up to 60 kv / cm .
, title="fig : " ] we calculate the modal gain as @xcite @xmath156 where @xmath157 is the permittivity of free space .
some constants are obtained from experiment : waveguide confinement factor @xmath158 , stage length @xmath159 nm , optical - mode refractive index @xmath160 , and full width at half maximum @xmath161 @xcite .
the dipole matrix element between the upper and lower lasing levels ( @xmath162 nm ) and the emission wave length ( @xmath163 m ) are also estimated in experiment @xcite , but we calculate these two terms directly . the dipole matrix element is calculated as @xmath164 the value is slightly different at different fields , as the band structure changes . at 48 kv / cm ,
the calculated matrix element is @xmath165 nm . similarly , the wave length of emitted photon also changes at different fields .
one can calculate the value from the energy difference between the upper and lower lasing levels .
the calculated wave length at 48 kv / cm is 8.964 @xmath0 m .
@xmath166 is the population inversion obtained from emc .
again , due to the randomness of emc , the population inversion needs to be averaged over a period of time after the steady - state has been reached .
figure [ fig : gm ] shows the modal gain of the device with nonequilibrium ( filled squares ) and thermal ( empty squares ) phonons as a function of ( a ) electric field and ( b ) current density at 77 k. horizontal dotted line indicates the total estimated loss in the device , which is used to help find the threshold current density , @xmath8 .
lasing threshold is achieved when the modal gain @xmath155 equals the total loss @xmath167 .
we consider two sources of loss , mirror ( @xmath168 ) and waveguide ( @xmath169 ) , so the total loss is @xmath170 .
the intercepts between the total loss line and the @xmath155 vs @xmath126 [ fig .
[ fig : gm](a ) ] and @xmath155 vs @xmath125 [ fig .
[ fig : gm](b ) ] curves give the threshold field @xmath171 and threshold current density @xmath8 , respectively .
like the current density , the modal gain of the device is also considerably higher when nonequilibrium phonons are considered , which leads to a lower @xmath171 and a lower @xmath8 .
the reason for the increased current density and modal gain with nonequilibrium phonons can be attributed to the enhanced injection selectivity and efficiency @xcite . ) of the simulated device with nonequilibrium ( filled squares ) and thermal ( empty squares ) phonons as a function of ( a ) applied electric field and ( b ) current density @xmath125 at 77 k. horizontal dotted line shows the total estimated loss of the device.,title="fig : " ] the way to obtain the heat - generation rate @xmath83 is similar to how we get the current density @xmath125 .
we record the cumulative net energy emission as a function of time and fit a straight line to the region where the simulation has reached a steady state .
the slope of the line is used in place of @xmath172 .
figure [ fig : hgen ] shows the heat - generation rate as a function of electric field at 77 k. the filled squares and the empty squares depict the situation with and without nonequilibrium phonons , respectively . at 77
k with ( filled squares ) and without ( empty squares ) nonequilibrium phonons.,title="fig : " ] this section serves to illustrate how the described simulation is implemented in practice , and what type of information it provides at the single - stage and device levels .
first , the single - stage coupled simulation has to be performed at different temperatures , as in fig .
[ fig : singlestagejf](a ) .
we note the the calculated @xmath143 curves show a negative - differential - conductance region , which is typical for calculations , but generally not observed in experiment . instead ,
a flat @xmath143 dependence is typically recorded @xcite .
at every temperature and field , we also record the heat - generation rate , as depicted in fig .
[ fig : singlestagejf](b ) .
second , the thermal model for the whole structure is developed .
considering that growth techniques improve over time , structures grown around the same time should have similar properties . since the device studied here was built in 2001 @xcite , we assume the active core should have similar effective rms roughness @xmath97 to other lattice - matched gaas / alas sls built around the same time @xcite . from our previous simulation work on fitting the sl thermal conductivities
@xcite , we choose an effective rms roughness @xmath173 in this calculation .
figure [ fig : pagetc ] shows the calculated thermal conductivities @xmath92 ( solid line ) and @xmath93 ( dashed line ) , along with the calculated bulk thermal conductivity ( dash - dotted line ) for the substrate gaas . ;
solid line ) and cross - plane ( @xmath93 ; dashed line ) thermal conductivities of the active core , along with the bulk thermal conductivity of the gaas substrate ( dash - dotted line ) .
the effective rms roughness @xmath97 is taken to be @xmath174,title="fig : " ] the structure we considered operated in pulsed mode at 77 k. depending on the duty cycle , the temperature distribution in the device can differ considerably .
figure [ fig : tempprofileschematic ] depicts a typical temperature profile across the device , while fig .
[ fig : tempprofiledutycycle ] depicts the profile across the active core alone at duty cycles of 100% ( essentially continuous wave lasing , if the device achieved it ) and 0.01% ( as in experiment @xcite ) .
clearly , cw operation would results in dramatic heating of the active region .
finally , fig .
[ fig : jv ] shows the @xmath175 curve of the entire simulated device at 77 k with a duty cycles of 0.01% , 100% , and as observed in experiment @xcite .
and the duty cycle is 100%.,title="fig:",width=302 ] ) .,title="fig : " ]
we overviewed electronic and thermal transport simulation of qcls , as well as recent efforts in device - level electrothermal modeling of these structures , which is appropriate for transport below threshold , where the effects of the optical field are negligible .
we specifically focused on mid - ir qcls in which electronic transport is largely incoherent and can be captured by the ensemble monte carlo technique .
the future of qcl modeling , especially for near - rt cw operation , will likely include improvements on several fronts : 1 ) further development of computationally efficient yet rigorous quantum - transport techniques for electronic transport , to fully account for coherent transport features that are important in short - wavelength mid - ir devices ; 2 ) a better understanding and better numerical models for describing the role of electron electron interaction , impurities , and interface roughness on device characteristics ; 3 ) holistic modeling approaches in which electrons , phonons , and photons are simultaneously and self - consistently captured within a single simulation .
the goal of qcl simulation should be nothing less than excellent predictive value of device operation across a range of temperatures and biasing conditions , along with unprecedented insight into the fine details of exciting nonequilibrium physics that underscores the operatiuon of these devices .
the authors gratefully acknowledge support by the u.s .
department of energy , basic energy sciences , division of materials sciences and engineering , physical behavior of materials program , award no .
de - sc0008712 .
the work was performed using the resources of the uw - madison center for high throughput computing ( chtc ) .
80ifxundefined [ 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 * * , ( ) link:\doibase 10.1109/5.649646 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.332655 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.3385159 [ * * , ( ) ] link:\doibase 10.1109/jstqe.2012.2237387 [ * * , ( ) ] \doibase http://dx.doi:10.1038/nphoton.2009.262 [ * * , ( ) ] \doibase http://dx.doi.org/10.1038/nphoton.2012.143 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2894569 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.3238263 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.3496489 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4719110 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.3586773 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.3484279 [ * * , ( ) ] link:\doibase 10.1364/oe.23.008462 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2907862 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/j.sse.2010.03.015 [ * * , ( ) ] link:\doibase 10.1049/el.2012.0017 [ * * , ( ) ] @noop * * , ( ) \doibase http://dx.doi.org/10.1063/1.1474613 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1494457 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1882768 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.87.146603 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1644337 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2711153 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4896400 [ * * , ( ) ] link:\doibase 10.1103/physrevb.67.085315 [ * * , ( ) ] link:\doibase 10.1103/physrevb.80.245316 [ * * , ( ) ] , link:\doibase 10.1103/physrevb.79.165322 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.205311 [ * * , ( ) ] http://stacks.iop.org/1367-2630/12/i=3/a=033045 [ * * , ( ) ] \doibase http://10.1007/s10825 - 016 - 0869 - 3 [ ( ) , http://10.1007/s10825 - 016 - 0869 - 3 ] , link:\doibase 10.3390/photonics3020038 [ * * , ( ) ] link:\doibase 10.1103/physrevb.66.245314 [ * * , ( ) ] link:\doibase 10.1002/pssb.201350322 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) \doibase http://dx.doi.org/10.1063/1.4895123 [ * * , ( ) ] http://stacks.iop.org/0034-4885/68/i=11/a=r02 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.3147943 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4863665 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.3284523 [ * * , ( ) , http://dx.doi.org/10.1063/1.3284523 ] link:\doibase 10.1109/jstqe.2013.2239613 [ * * , ( ) ] link:\doibase 10.1109/jqe.2008.922327 [ * * , ( ) ] link:\doibase 10.1007/s00340 - 011 - 4744 - 4 [ * * , ( ) ] link:\doibase 10.1007/s10825 - 012 - 0397 - 8 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4935142 [ * * , ( ) ] @noop ( ) , link:\doibase 10.1103/physrevb.73.245320 [ * * , ( ) ] link:\doibase 10.1109/jqe.2002.1005404 [ * * , ( ) ] link:\doibase 10.1109/jqe.2007.907563 [ * * , ( ) ] _ _ , @noop ph.d . thesis , ( ) _ _ , @noop ph.d .
thesis , ( ) \doibase http://dx.doi.org/10.1063/1.2136420 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/s0370-1573(01)00029-1 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2387485 [ * * , ( ) ] @noop _ _ ( , ) link:\doibase 10.1007/s10825 - 008 - 0175 - 9 [ * * , ( ) ] link:\doibase 10.1103/physrevb.39.7852 [ * * , ( ) ] http://stacks.iop.org/0022-3719/15/i=28/a=021 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2172225 [ * * , ( ) ] link:\doibase 10.1103/physrevb.50.14179 [ * * , ( ) ] link:\doibase 10.1109/jproc.2006.879794 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4832615 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.98526 [ * * , ( ) ] link:\doibase 10.1115/1.2911404 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.114919 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/0921-4526(95)00858-6 [ * * , ( ) ] link:\doibase 10.1103/physrevb.59.8105 [ * * , ( ) ] link:\doibase 10.1115/1.2824212 [ * * , ( ) ] link:\doibase 10.1103/physrevb.57.14958 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1455693 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.165406 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.88.155318 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.61.605 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2222074 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1374520 [ * * ( ) , http://dx.doi.org/10.1063/1.1374520 ] \doibase http://dx.doi.org/10.1063/1.2840136 [ * * , ( ) ] | note : this is a book chapter that will appear in _ handbook of optoelectronic device modeling and simulation _ , taylor & francis books , 2017 . editor : joachim piprek . for table of contents , see http://www.nusod.org/piprek/handbook.html |
[ [ section ] ] a very efficient tool in diophantine geometry is the so - called determinant method which was developed by bombieri and pila in the influential paper @xcite about the number of integral points of bounded height on affine algebraic and transcendental plane curves .
basically , the method consists in using a determinant of a suitable set of monomials evaluated at the integral points , in order to construct a family of auxiliary polynomials vanishing at all integral points on the curve within a small enough box .
building on the estimates in @xcite for algebraic curves , pila proved in @xcite bounds on the number of integral ( resp .
rational ) points of bounded height on affine ( resp .
projective ) algebraic varieties of any dimension , improving on previous results by s. d. cohen using the large sieve method @xcite .
important further improvements going towards optimal bounds conjectured by serre in 13 of @xcite have been made since by heath - brown , browing and salberger @xcite,@xcite,@xcite . in @xcite ,
pila and wilkie proved a general estimate for the number of rational points on the transcendental part of sets definable in an o - minimal structure ; this has been used in a spectacular way by pila to provide an unconditional proof of some cases of the andr - oort conjecture @xcite ( see also @xcite , @xcite and @xcite for surveys on applications in diophantine geometry of the pila - wilkie theorem ) .
lying at the heart of pila and wilkie s approach is the possibility of having uniform - in terms of number of parametrizations and in terms of bounds on the partial derivatives - @xmath2-parametrizations .
these parametrizations are provided by an o - minimal version of gromov s algebraic parametrization lemma @xcite ( see also @xcite ) , itself a refinement of a previous result of yomdin @xcite,@xcite .
such @xmath2-parametrizations enter the determinant method via taylor approximation .
the aim of this paper is to provide a version of the yomdin - gromov lemma in the non - archimedean setting , notably for subanalytic sets over @xmath3 and @xmath4 , and to develop the determinant method in this context in order to obtain non - archimedean analogues of some of the results in diophantine geometry mentionned above . at first sight
one may have doubts about the realizability of such a program , since because of the totally disconnected character of non - archimedean spaces , it seems there is no way for a global taylor formula to make sense in this framework . a first indication that the situation may not be completely hopeless , is provided by the fact that in previous work @xcite ( see also @xcite )
we have been able to prove a version of first - order taylor approximation , piecewise globally , in the definable @xmath0-adic setting . in the present paper , though we extend this first order result to a much wider situation , we have chosen not trying to generalize it to higher order , but instead we show directly the existence of uniform @xmath2-parametrizations that do satisfy taylor approximation , which is enough for our purpose .
the existence of such parametrizations is provided by theorem [ gyt ] which is the main result of section [ sec:2 ] . in section [ sec:3 ]
, we deduce a @xmath0-adic analogue of the theorem of pila and wilkie in @xcite , in the strengthened version given by pila in @xcite in terms of blocks . in section [ sec:4 ] ,
we prove a geometric analogue of results of bombieri - pila @xcite and pila @xcite over @xmath4 where counting number of points is replaced by counting dimensions .
the diophantine applications we give in the @xmath0-adic case ( concerning the density of rational points on the transcendental part of definable sets ) and the motivic case ( concerning the density of rational points on algebraic sets ) are quite different .
one should notice that in the algebraic case , working over @xmath3 instead of @xmath5 would not provide better estimates than those following directly from the archimedean ones in @xcite and @xcite .
[ [ ssec:1.2 ] ] let us spell out in some more detail basic versions of three of our results for subanalytic @xmath0-adic sets . one calls a set @xmath6 semialgebraic if it is definable in the ring language with parameters from @xmath7 . by adding to the ring language symbols for analytic functions , one obtains subanalytic sets
( see section [ sec : gyt ] below , with @xmath8 ) .
the dimension of a nonempty subanalytic set @xmath6 is the largest integer @xmath9 such that there exists a coordinate projection @xmath10 such that @xmath11 has nonempty interior .
we will simply denote by @xmath12 the @xmath0-adic norm of an element @xmath13 and when furthermore @xmath14 , we will denote by @xmath15 the real norm of @xmath16 . for @xmath17 a subset of @xmath18 and @xmath19 a real number , write @xmath20 for the set consisting of points @xmath21 in @xmath22 such that one can write @xmath23 as @xmath24 where @xmath25 and @xmath26 are integers with @xmath27 and @xmath28 . for @xmath17 a subset of @xmath18 ,
write @xmath29 for the subset of @xmath17 consisting of points @xmath16 such that there exists an algebraic curve @xmath30 such that @xmath31 is locally of dimension @xmath32 at @xmath16 . with this notation ,
the following statement is a particular case of theorem [ blocksq ] : let @xmath6 be a subanalytic set of dimension @xmath33 with @xmath34 .
let @xmath35 be given .
then there exist an integer @xmath36 and a semialgebraic set @xmath37 such that @xmath38 lies inside @xmath29 , and such that for each @xmath39 , one has @xmath40
more generally , we also provide estimates for algebraic points of bounded degree on @xmath17 , as follows . for a rational number @xmath41
, we define @xmath42 as @xmath43 when @xmath44 with @xmath45 and @xmath46 integers which are either relatively prime or are such that @xmath47 . for nonrational @xmath41
, @xmath42 is defined to be @xmath48 .
we extend @xmath49 to tuples @xmath50 by putting @xmath51 . for an integer @xmath52 and any @xmath53
we define @xmath54 as @xmath55 where the minimum runs over all nonzero tuples @xmath56 such that @xmath57 if such tuple exists , and as @xmath48 otherwise .
one extends @xmath58 to @xmath16 in @xmath18 by taking the maximum of the @xmath59 for @xmath60 .
for @xmath17 a subset of @xmath18 , @xmath61 an integer and @xmath19 a real number , write @xmath62 for the set consisting of @xmath16 in @xmath17 satisfying that @xmath54 is at most equal to @xmath39 .
the following statement follows from theorem [ algheights ] : let @xmath6 be a subanalytic set of dimension @xmath33 with @xmath34 .
let @xmath35 and an integer @xmath61 be given .
then there exist an integer @xmath63 and a semialgebraic set @xmath64 such that @xmath38 lies inside @xmath29 , and such that for each @xmath39 , one has @xmath65
our proofs of theorems [ algheights ] and [ blocksq ] rely on the existence of reparametrizations provided by theorem [ gyt ] , where we prove the following : let @xmath66 , @xmath9 and @xmath67 be integers and let @xmath68 be a subanalytic set of dimension @xmath33
. then there exists a finite collection of subanalytic functions @xmath69 such that the union of the @xmath70 equals @xmath17 , the @xmath71 have @xmath72-norm bounded by @xmath32 , and the @xmath71 are approximated by their taylor polynomials of degree @xmath73 with remainder of order @xmath45 , globally on @xmath74 . for precise definitions of the @xmath72-norm and approximation by taylor polynomials of certain degree with controlled remainder
, we refer to definition [ tr ] .
note that a key point in this non - archimedean statement is that the approximation holds globally on @xmath74 , which represents a challenging goal because of total disconnectedness . on the opposite , in the real case
, it is enough to consider convex charts , since for such charts global taylor approximation is immediate . in the core of the paper all these results will be stated and proved uniformly for definable families .
[ [ ssec:1.3 ] ] we end this introduction with a quick overview of our results over the base field @xmath5 . in this case , the analogue of theorem [ gyt ] essentially still holds , except one has to replace `` finite '' by `` parametrized by a constructible subset of @xmath75 for some @xmath46 '' . for each positive integer @xmath45
we denote by @xmath76 _ { < r}$ ] the set of complex polynomials of degree @xmath77 . for any subset @xmath78 of @xmath79
we denote by @xmath80 the set @xmath81 _ { < r})^n$ ] and by @xmath82 the dimension of the zariski closure of @xmath80 in @xmath83 _
{ < r})^n \simeq \mathbb{c}^{nr}$ ] .
when @xmath17 is an algebraic subvariety of @xmath84 of dimension @xmath33 , for every positive integer @xmath45 one has the basic estimate @xmath85 ( cf .
lemma [ trivest ] ) .
assume @xmath17 is irreducible of degree @xmath86 .
the main result of section [ sec:4 ] , theorem [ motivicbp ] , states that , for every positive integer @xmath45 , one has @xmath87 which is an improvement on the basic estimate as soon as @xmath17 is not linear .
this result can be seen as an instance of motivic point counting , like in the paper @xcite , where in a somewhat different context a detailed study of the motivic height zeta function leads to asymptotic estimates for dimensions of spaces of sections .
it is a motivic analogue of pila s results of @xcite . as in @xcite
, we reduce to the case of plane curves by lang - weil type arguments .
however , to prove the result in this case , we do not follow the original method of bombieri and pila in @xcite , which seems difficult to adapt in this setting .
we follow instead a strategy introduced by marmon in @xcite , which relies on the yomdin - gromov lemma in place of the elaborate analytic arguments used in @xcite .
although the general results on lipschitz continuity of theorems [ lip ] and [ lipmixed ] are used to prove theorems [ gyt ] and [ gyts ] , as far as our applications to points of bounded height are concerned , they are not needed in full generality .
first , in the @xmath0-adic setting of section [ sec:3 ] , theorems [ lip ] and [ lipmixed ] are not new since they are proved in @xcite and even appear in a slightly sharper form in @xcite . secondly , in section [ sec:4 ] when working over @xmath88 ,
only the one - dimensional case of theorem [ gyt ] , namely with @xmath89 , is used .
this case of theorem [ gyt ] only relies on the one - dimensional case of theorem [ lip ] , which can be proved similarly as the one - dimensional case of @xcite .
such one - dimensional cases require less work and are easier to prove than the general cases .
for example , for the case of general dimension for theorems [ lip ] and [ lipmixed ] one can not use definable sections as in @xcite and we had to find the alternative approach via theorem [ lipcenter ] ; in the one - dimensional case definable sections were not used in @xcite .
another simplification , in the @xmath0-adic case , would be to exploit the fact that the residue rings @xmath90 are finite .
indeed , this finiteness allows one to zoom and scale as in the real case , and this can serve as an alternative to the passage to the algebraic closure of @xmath91 when proving theorems [ gyt ] and [ gyts ] for @xmath8 .
this zoom and scale technique would not work in the case of @xmath92 , unless one is content in theorem [ motivicbp ] with a weakening of the upper bound to @xmath93 , for some constant depending on @xmath86 .
we would like to thank antoine chambert - loir , immanuel halupczok and ehud hrushovski for stimulating discussions related to this work .
in particular , antoine chambert - loir directed us to marmon s paper , immanuel halupczok provided an example that finitely many pieces do not suffice in general for theorem [ lip ] , and ehud hrushovski encouraged us to consider potential diophantine applications . the authors were supported in part by the european research council under the european community s seventh framework programme ( fp7/2007 - 2013 ) with erc grant agreement numbers
246903 nmnag and 615722 motmelsum , and by the labex cempi ( anr-11-labx-0007 - 01 ) .
we would also like to thank the ihes and the fim of the eth in zrich , where part of the research was done .
the paper was finalized while the authors participated in the msri program : model theory , arithmetic geometry and number theory .
in this section and in the next one we prove non - archimedean analogues of the yomdin - gromov parametrization lemma . in the same time , we prove that our parametrizations can be uniformly approximated by their taylor polynomials .
the ability to approximate parametrizations by their taylor polynomials with small error terms is key for counting points of bounded height in the non - archimedean case , as well as in the real case , although in the real case this approximation is classical when the domain is convex and boundeof order @xmath94 ( see theorem [ gyt ] ) .
the yomdin - gromov parametrization lemma , as generalized by pila and wilkie in @xcite , gives for any bounded definable set @xmath17 of dimension @xmath33 in @xmath95 ( in an o - minimal structure on @xmath96 ) and any integer @xmath97 a finite collection of definable maps @xmath98^m\to x$ ] whose ranges together cover @xmath17 and whose @xmath72-norms are bounded by @xmath32 . by convexity of @xmath99^m$ ] , and techniques based on the mean value theorem , one can control the error terms when approximating @xmath100 by a taylor polynomial of degree @xmath73 . both convexity and the mean value theorem
do not carry to our context . for @xmath101 , we do not know , even for @xmath0-adic semialgebraic functions @xmath102 on @xmath103 , whether having small @xmath72-norm allows one to piecewise control the error term , globally on each piece , when one approximates @xmath102 by a taylor polynomial of degree @xmath73 . in the case
where @xmath104 , the desired approximation coincides with lipschitz continuity , and the study of piecewise lipschitz continuity , in a general non - archimedean context , is the content of sections [ sec:5b ] and [ sec : mix ] . from section [ sec : gyt ] on
, we will place ourselves in a more concrete framework of definable sets in complete , discretely valued fields ( possibly with extra restricted analytic functions in the language ) , to treat @xmath72-parametrizations with good taylor approximation when @xmath101 . in @xcite , piecewise lipschitz - continuity for a semialgebraic or subanalytic function @xmath105
is shown to hold whenever @xmath102 is locally lipschitz continuous with a fixed lipschitz constant . moreover ,
the pieces can be taken to be definable . in @xcite ,
the lipschitz constants were further controlled in an optimal way when going from local to global on each piece . in this section ,
we extend the result of @xcite in two ways , namely to many new structures with a non - archimedean geometry , including @xmath5 , and to other languages than the semialgebraic and subanalytic ones , including some weaker languages without multiplication .
when the residue field is not finite , one is led to replace finite definable partitions by definable families with parameters running over the residue field .
our study of lipschitz continuity is subdivided into two cases : equicharacteristic zero and mixed characteristic . both are axiomatically treated . in the mixed characteristic case , residue rings , and not only the residue field , are used .
we first introduce the set - up adapted to the equicharacteristic zero case . the typical example to have in mind
is that of henselian valued field @xmath106 of equicharacteristic zero with ( multiplicatively written ) value group @xmath107 and residue field @xmath108 .
then in section [ sec : mix ] we will consider the mixed characteristic case .
let @xmath109 be the disjoint union of a nontrivial ordered abelian group @xmath110 with a minimal element @xmath111 , where the group operation on @xmath110 is written multiplicatively , and where we put @xmath112 for all @xmath113 . recall that an ordered abelian group is an abelian group with a total order @xmath114 such that @xmath115 implies @xmath116 for all elements @xmath117 of the group .
such a group is automatically torsion free , and hence , the order @xmath114 has no endpoints .
let @xmath106 be an additively written abelian group and suppose we are given a surjective map @xmath118 with the following properties for all @xmath119 * @xmath120 if and only if @xmath121 , * @xmath122 , * @xmath123 * if @xmath124 , then @xmath125 .
an open ball is by definition a subset @xmath126 of the form @xmath127 for some @xmath128 and @xmath129 ; such @xmath130 is unique and is called the radius of the open ball @xmath131 ( not to be confused with the radii of closed balls defined in section [ sec:2 ] ) .
since @xmath110 has no endpoints , each open ball is an infinite set .
consider a set @xmath108 containing a special element @xmath132 and write @xmath133 for @xmath134 .
suppose that we are given a surjective map @xmath135 with @xmath136 and such that for each @xmath137 and @xmath128 , the set @xmath138 is an open ball of radius @xmath130 .
let us more generally introduce the notation @xmath139 note that the family of sets @xmath140 is a disjoint family whose union equals @xmath106 when @xmath141 varies in @xmath108 and @xmath130 in @xmath142 .
clearly @xmath143 equals @xmath111 , and both @xmath144 and @xmath145 are empty for nonzero @xmath130 and nonzero @xmath141 .
we put on @xmath106 the valuation topology , that is , the topology with the collection of open balls as base , and the product topology on cartesian powers of @xmath106 .
note that @xmath106 thus becomes a topological group .
for a tuple @xmath146 , @xmath147 stands for @xmath148 .
next we recall the definition of lipschitz continuity and we define a special variant of continuity , called s - continuity .
[ def : lip ] let a function @xmath149 be given , with @xmath150 . for @xmath151 , the function @xmath152 is called @xmath130-lipschitz ( globally on @xmath17 ) if for all @xmath16 and @xmath153 in @xmath17 , @xmath154 the function @xmath102 is called locally @xmath130-lipschitz if every point of @xmath155 has a neighbourhood on which @xmath102 is @xmath130-lipschitz .
[ defjacprop ] let @xmath156 be a function for some set @xmath157 .
say that @xmath158 is s - continuous if for each open ball @xmath159 the set @xmath160 is either a singleton or an open ball , and , there exists @xmath161 such that @xmath162 on an open @xmath163 is @xmath46-continuous in , say , the variable @xmath164 , by which we mean that @xmath165 is @xmath46-continuous for each choice of @xmath166 then we write @xmath167 for the element @xmath168 witnessing the s - continuity of @xmath165 locally at @xmath164 , namely , @xmath130 satisfies ( [ s - gamma ] ) for the function @xmath169 , where @xmath170 run over some ball @xmath131 containing @xmath164 such that @xmath171 .
note that for an @xmath46-continuous function @xmath172 on an open @xmath78 such that @xmath158 is moreover @xmath32-lipschitz , one has @xmath173 for all @xmath174 .
hence , for such @xmath158 , for @xmath174 with @xmath175 , and for any ball @xmath159 containing @xmath16 , say , of radius @xmath176 , the set @xmath160 is a ball of radius @xmath177 . moreover , for compositions of s - continuous functions one has a certain form of the chain rule which corresponds to the classical chain rule for differentiation , cf .
lemma [ cor : invers ] and its proof .
let @xmath178 be the first order language with the sorts @xmath106 , @xmath108 and @xmath142 , and symbols for addition on @xmath106 , for @xmath135 , @xmath179 , and for the order and the multiplication on @xmath142 .
let @xmath180 be any expansion of @xmath178 .
by @xmath180-definable we mean @xmath181-definable in the language @xmath180 , and likewise for other languages than @xmath180 .
write @xmath182 , @xmath183 , and @xmath184 , with a slight abuse of notation . note that @xmath180 may have more sorts than @xmath178 , since it is an arbitrary expansion .
this language @xmath178 is very basic ( since it does not have multiplication ) , and can be interpreted in many structures .
we give an example of a triple @xmath185 with @xmath178-structure .
let @xmath106 be the laurent series field @xmath186 , seen as a group for addition , put @xmath187 , let @xmath188 be the @xmath189-adic norm with @xmath190 on @xmath106 , @xmath108 the finite field @xmath191 , and let @xmath192 send a nonzero laurent series @xmath193 to the coefficient of its lowest degree nonzero term .
a more natural example of an @xmath180-structure with @xmath180 being @xmath178 together with multiplication on @xmath106 , is for the field @xmath92 with @xmath189-adic norm and @xmath192 defined as for @xmath186 .
[ config ] given integers @xmath194 , @xmath195 , a set @xmath196 and some @xmath197 , say that @xmath39 is in @xmath198-config if there is @xmath199 such that @xmath39 equals the union over @xmath168 of sets @xmath200 for some @xmath201 .
if moreover @xmath202 we speak of an open @xmath198-config , and if @xmath203 we speak of a graph @xmath198-config . if @xmath39 is nonempty and in @xmath198-config , then @xmath141 and the sets @xmath204 such that @xmath140 is nonempty are uniquely determined by @xmath39 and @xmath198 .
say that @xmath205 is in @xmath180-tame config if there exist @xmath206 and @xmath180-definable functions @xmath207 such that the range of @xmath198 contains no open ball , and , for each @xmath208 , the set @xmath209 is in @xmath210-config . by the aforementioned uniqueness in the nonempty case
, one sees that , for an @xmath180-definable set @xmath39 which is in @xmath198-config , the collection of sets @xmath211 can be taken to be an @xmath180-definable family .
the functions @xmath212 and @xmath198 used in a @xmath180-tame config are in general not unique , but still one often calls @xmath198 the center ( of the configuration ) .
[ test ] for any @xmath180-structure @xmath213 which is elementarily equivalent to @xmath214 and for any language @xmath91 which is obtained from @xmath180 by adding some elements of @xmath213 ( of any sort ) as constant symbols , call @xmath215 a test pair for @xmath214 .
[ ktame ] say that @xmath214 is weakly tame if the following conditions hold . 1 .
each @xmath180-definable set @xmath216 with @xmath194 , @xmath195 is in @xmath180-tame config .
2 . for any @xmath180-definable function @xmath217 there exist @xmath218 and an @xmath180-definable function @xmath219 such that , for each @xmath208 , the restriction of @xmath158 to @xmath220 is s - continuous .
say that @xmath214 is tame when each test pair @xmath215 for @xmath214 is weakly tame .
call an @xmath180-theory @xmath221 tame if for each model @xmath222 of @xmath221 , the pair @xmath223 is tame . condition ( 2 ) is a substitute for the so - called jacobian property which holds for henselian valued fields in equicharacteristic zero equipped with the denef - pas language .
we refer to theorem 6.3.7 of @xcite for a closely related jacobian property ; it can be adapted to the denef - pas language using theorem 4.1 of @xcite on elimination of valued field quantifiers
. for henselian valued fields in mixed characteristic equipped with the generalized denef - pas language , see section [ sec : mix ] . in @xcite ,
definition 2.19 and theorem 5.12 , one will find a version of the jacobian property in higher dimensions .
some examples of tame structures are provided in section [ ex - tame ] .
we can now state our first main result on lipschitz continuity , going from local to piecewise global on parts parametrized by variables running over @xmath108 .
[ lip ] suppose that @xmath214 is tame .
let @xmath224 be an @xmath180-definable function which is locally @xmath32-lipschitz .
then there exists an @xmath180-definable function @xmath225 for some @xmath218 such that for each @xmath208 , the restriction of @xmath102 to @xmath220 is @xmath32-lipschitz .
theorem [ lip ] is complemented by theorem [ lipcenter ] about simultaneous partitions of domain and range into parts with @xmath32-lipschitz centers .
this is an improvement of proposition 2.4 of @xcite , where this is done for the domain only , and only in the @xmath0-adic case . for @xmath226
any function between sets and for @xmath227 , write @xmath228 for the set @xmath229 and write @xmath230 or @xmath231 for the function which sends @xmath232 to @xmath233 .
we use similar notation @xmath228 and @xmath230 or @xmath231 when @xmath234 is a cartesian product @xmath235 and @xmath236 for some coordinate projection @xmath237 .
[ lipcenter ] suppose that @xmath214 is tame .
let @xmath238 be an @xmath180-definable function which is locally @xmath32-lipschitz .
then , for a finite partition of @xmath78 into definable parts , the following holds for each part @xmath17 .
there exist @xmath218 , a coordinate projection @xmath239 and @xmath180-definable functions @xmath240 such that , for each @xmath208 , the restrictions of @xmath241 and @xmath242 to @xmath243 are @xmath32-lipschitz and , for each @xmath244 in @xmath245 , the set @xmath246 is in @xmath247-config and the image of @xmath246 under @xmath248 is in @xmath249-config .
note that the projection @xmath0 in theorem [ lipcenter ] _ a priori _ depends on the part @xmath17 .
theorems [ lip ] and [ lipcenter ] are proved by a joint induction on @xmath250 . by the improvement of proposition 2.4 of @xcite given by theorem [ lipcenter ] , we can avoid the usage of definable sections ( skolem functions ) , which were heavily used in @xcite .
this is especially helpful since one does not have definable skolem functions in the general context of tameness .
let us first explain the general strategy of the proofs .
we first prove some general , although easy , results about sets and functions in tame structures , from statement [ fin ] up to [ invers ] .
an analogue of the chain rule for derivation can be used , based on s - continuity .
the lipschitz continuity of @xmath198 in theorem [ lipcenter ] is proved as in @xcite , as well as the case @xmath251 of theorem [ lip ] .
what is new here is that @xmath86 in theorem [ lipcenter ] can be required to be lipschitz continuous as well .
working piecewise in the proof of theorem [ lip ] , one may , after some triangular transformation , assume that the centers of domain and of range are both zero .
in which case the comparison of distances in domain and in range becomes easier .
we prove preliminary statements in view of theorems [ lip ] and [ lipcenter ] .
[ fin ] suppose that @xmath214 is tame .
if @xmath252 is @xmath180-definable for some @xmath253 , then the image of @xmath254 contains no open ball .
let @xmath252 be @xmath180-definable and let @xmath216 be the graph of @xmath254 , with the natural identification .
now take @xmath255 and @xmath256 such that the range of @xmath198 contains no open ball and such that @xmath257 is in @xmath210-config , for any @xmath208 . by definition
[ config ] there exist sets @xmath258 such that @xmath259 equals the union of @xmath260 over @xmath261 . from this description as a cartesian product together with the fact that @xmath39 is the graph of @xmath254
, it follows that @xmath262 is empty whenever @xmath140 contains more than one element .
moreover , whenever @xmath140 is a singleton one has @xmath263 . hence , the range of @xmath254 is contained in the range of @xmath198 , which contains no open ball .
the next proposition has to be compared with the real monotonicity theorem ( see @xcite , ( 1.2 ) chapter 3 ) .
[ inj - cons]suppose that @xmath214 is tame .
let @xmath217 be @xmath180-definable .
then there exist @xmath218 and an @xmath180-definable function @xmath264 such that for each @xmath208 the restriction of @xmath158 to @xmath220 is injective or constant . by tameness , logical compactness , and by going to a test pair
, it suffices to treat the case where @xmath158 is s - continuous and moreover locally injective or locally constant .
indeed , by tameness there exists an @xmath180-definable function @xmath265 such that @xmath158 is s - continuous on each fiber of @xmath266 , and such that each fiber of @xmath266 is either a singleton or open in @xmath106 .
if we can prove the statement for each restriction of @xmath158 to @xmath267 for any @xmath268 , then we are done by logical compactness , since @xmath214 is arbitrary in the proposition and since such a restriction is @xmath269-definable and lives thus in the test pair @xmath270 , also a tame structure .
note that such logical compactness yields finitely many candidate definable functions @xmath71 out of infinitely many ones , but these @xmath71 can be combined to a single one by putting @xmath271 for the minimal @xmath272 such that @xmath158 is either injective or constant on @xmath273 , which is an @xmath180-definable condition .
let us first suppose that @xmath158 is locally injective .
let @xmath274 be the graph of @xmath158 .
by logical compactness and tameness , there exist @xmath218 and definable functions @xmath275 such that for each value @xmath276 , the set @xmath277 is in @xmath278-config . indeed , for each value @xmath153 in @xmath279 and by tameness of the test pair @xmath280 , there exist @xmath281-definable maps @xmath282 from @xmath283 to @xmath284 and @xmath285 such that for each value @xmath286 , the set @xmath287 , considered as a subset of @xmath106 , is in @xmath288-config .
logical compactness yields again finitely many @xmath71 and @xmath289 out of these possibly infinitely many @xmath282 and @xmath290 , and these can again be combined by defining @xmath291 as @xmath292 and @xmath278 as @xmath293 for the minimal @xmath272 such that @xmath294 is in @xmath293-config . set @xmath295 by the local injectivity of @xmath158 and by the definition of being in @xmath278-config
, it follows that @xmath296 is contained in the range of @xmath297 .
fix @xmath298 .
then @xmath299 is the inverse function of the restriction of @xmath158 to @xmath220 , which is thus injective .
let us finally suppose that @xmath158 is locally constant . by tameness , s - continuity , and local constancy of @xmath158 , there exist @xmath253 and @xmath300 such that @xmath158 is constant on each fiber of @xmath254 .
by lemma [ fin ] , @xmath279 contains no open ball . by tameness there exist definable functions @xmath301 such that @xmath302 is in @xmath210-config , meaning that @xmath279 is contained in the range of @xmath198 .
now define @xmath219 as sending @xmath16 to the unique @xmath303 with @xmath304 .
it is clear that @xmath158 is constant on @xmath220 for any @xmath208 . by the following corollary
, tameness appears as a special variant of @xmath305-minimality as defined in @xcite , but tameness has a more geometrical flavor . note that for us the sorts @xmath108 and @xmath142 play a rather different role , while in @xmath305-minimality all sorts other than @xmath106 are treated on the same footing . in particular parameters in @xmath108
play a special role for us , defining the splitting of a space into parts parametrized by the residue field in theorem [ lip ] and elsewhere .
[ b ] suppose that @xmath214 is tame .
let @xmath221 be the theory of the restriction of @xmath214 to the sorts @xmath306 , that is , @xmath221 is the theory of the structure on these sorts having as definable sets the @xmath180-definable sets .
then @xmath221 is @xmath305-minimal , with main sort @xmath106 and where the role of balls is played by open balls .
the proof is immediate from definition 2.1 of @xcite , lemma [ fin ] , and proposition [ inj - cons ] .
in particular , the dimension theory for @xmath305-minimal structures of @xcite applies to tame structures .
[ continuity ] let @xmath224 be @xmath180-definable
. then there exists an @xmath180-definable function @xmath225 for some @xmath218 such that , for each @xmath208 , the restriction of @xmath102 to @xmath220 is continuous .
further , for given @xmath307 with @xmath194 , @xmath195 , one can moreover take @xmath212 such that the fibers @xmath308 are locally independent of @xmath309 for each @xmath208 .
moreover , if @xmath251 , one can ensure that , for any ball @xmath131 contained in @xmath17 , @xmath310 for any @xmath311 which lie in @xmath131 .
finally , if @xmath17 is open and @xmath102 is locally @xmath32-lipschitz in each variable separately , then one can moreover take @xmath212 such that the restriction of @xmath102 to @xmath220 is locally @xmath32-lipschitz for any @xmath208 . the statements for @xmath251 follow from tameness .
indeed , the statement about continuity follows from s - continuity when @xmath251 and the statement about @xmath39 follows from existence of tame configurations .
note that one can preserve continuity of @xmath102 while controlling @xmath308 since it is possible to combine two maps @xmath312 into a single map @xmath313 refining both @xmath314 and @xmath315 , in the sense that each fiber of @xmath212 is included in a fiber of @xmath314 and of @xmath315 . the statements for general @xmath250 follow from tameness , induction on @xmath250 , and logical compactness ( used in the same way as in the proof of proposition [ inj - cons ] ) .
we shall illustrate this by giving now full details for the proof of the statement about the continuity of @xmath102 in the case @xmath316 .
the proof for @xmath317 and and the proof of the statement about @xmath308 are completely similar . by compactness , by the case @xmath251 , and by going to a test pair
, we may suppose that @xmath318 and @xmath319 are continuous for each @xmath320 and @xmath321 in @xmath106 .
likewise , we may suppose that @xmath322 is in @xmath323-config , for some definable function @xmath198 and for each @xmath320 in @xmath106 , and that @xmath324 is in @xmath86-config for some @xmath325 , where @xmath326 is the projection @xmath327 .
similarly , we may further suppose that @xmath198 is continuous on @xmath324 .
again for similar reasons , we may assume that the definable family of sets @xmath328 does locally not depend on @xmath329 , where the @xmath328 are such that , for some @xmath199 and for each @xmath320 in @xmath106 , @xmath330 furthermore , we can also assume that , for each @xmath320 , @xmath331 indeed , after partitioning , the only case left is when @xmath322 or @xmath324 are in graph config for each @xmath320 , which , by continuity of @xmath198 , reduces to the case @xmath251 . for any function @xmath332 and any @xmath227 , let @xmath333 be the set of pairs @xmath334 in @xmath335 such that , for any @xmath305 in @xmath78 , if @xmath336 , then @xmath337 . for each @xmath338 in @xmath106
, consider the sets @xmath339 they form definable families of subsets of @xmath335 with parameters @xmath340 in @xmath341 .
now choose @xmath340 in @xmath17 and @xmath342 in @xmath110 .
choose also @xmath343 such that @xmath334 lies in the intersection of the two sets in ( [ x12c ] ) .
such a @xmath344 exists since @xmath318 and @xmath319 are assumed to be continuous . by continuity of @xmath198 , local independence of @xmath320 for the sets @xmath328 , and by the openness assumption in ( [ xcopen ] ) , we can take @xmath344 so small so that , for any @xmath345 in @xmath17 with @xmath346 , one has that @xmath347 lies in @xmath17 .
now , for any @xmath345 in @xmath17 with @xmath346 , we have @xmath348 which yields the continuity of @xmath102 .
[ invers ] let @xmath349 be @xmath180-definable
. then there exists an @xmath180-definable function @xmath225 for some @xmath218 such that , for each @xmath208 , either @xmath198 is locally @xmath32-lipschitz on @xmath220 , or , the restriction of @xmath198 to @xmath220 is injective and its inverse function is locally @xmath32-lipschitz .
the statement is clear , by proposition [ inj - cons ] and the definitions of tameness and s - continuity . combining lemma [ invers ] with a form of the chain rule for differentiation , we find the following several - variable result .
[ cor : invers ] let @xmath150 be @xmath180-definable and of dimension @xmath350 . then there exist an @xmath180-definable function @xmath351 for some integer @xmath46 , a finite partition of the graph @xmath352 of @xmath212 into @xmath180-definable parts @xmath353 , and for each @xmath272 an injective coordinate projection @xmath354 and @xmath180-definable functions @xmath355 such that the union over @xmath272 of the graphs @xmath356 equals @xmath352 and such that the functions @xmath357 are locally @xmath32-lipschitz for each @xmath208 . by compactness , working piecewise , and going to a test pair , we may assume that @xmath17 is already the graph of a function @xmath358 which is s - continuous in each variable separately .
by induction on @xmath86 we may suppose that @xmath163 is open .
we will treat the case @xmath359 , the general case being similar .
after reordering the variables @xmath360 , working piecewise , we may suppose that @xmath361 is maximal among the @xmath362 on the whole of @xmath163 for @xmath363 , and that @xmath364 on @xmath163 .
moreover , by proposition [ inj - cons ] and compactness , we may suppose for any @xmath365 that @xmath230 is injective on @xmath366 .
now we can reverse the role of @xmath367 and @xmath368 , by reordering the coordinates .
this way @xmath254 is replaced by a function @xmath369 sending @xmath370 to the compositional inverse @xmath371 by s - continuity , with the notation from just below definition [ defjacprop ] , we have , for each @xmath372 and for @xmath373 with @xmath374 that @xmath375 which is at most one by our assumption that @xmath361 is maximal among the @xmath376 .
now we come to the proof of our main results on lipschitz continuity .
[ rem : comp ] as the proof of theorem [ lipcenter ] for @xmath251 will show , the hypothesis that @xmath102 is locally @xmath32-lipschitz is not needed at all when @xmath251 .
hence , theorem [ lipcenter ] for @xmath251 holds even when @xmath102 is not locally @xmath32-lipschitz .
furthermore , when @xmath251 , if @xmath102 is injective and @xmath212 is as given by theorem [ lipcenter ] , the function @xmath102 gives a correspondence between the maximal balls included in @xmath220 and the maximal balls included in @xmath377 .
first suppose that @xmath102 is injective and s - continuous , and that @xmath17 equals an open ball @xmath378 for some @xmath180-definable @xmath379 , @xmath199 and @xmath168 .
write @xmath380 for @xmath381 .
it follows from the case assumptions that @xmath380 is an open ball . by tameness , there exist @xmath382 and @xmath180-definable functions @xmath383 such that @xmath384 is in @xmath385-config for each @xmath386 .
define @xmath387 and define @xmath210 as @xmath388 when @xmath385 lies outside @xmath380 and as @xmath389 when @xmath385 belongs to @xmath380 .
it follows by s - continuity and injectivity of @xmath102 that @xmath220 is in @xmath210-config .
the slightly more general case where @xmath102 is injective and s - continuous and where @xmath17 is in @xmath388-config for some @xmath180-definable @xmath379 is treated similarly , by choosing @xmath254 and defining @xmath212 and @xmath198 as in the above construction .
finally we consider the general case . by tameness and proposition [ inj - cons ]
, there exist @xmath180-definable functions @xmath390 and @xmath391 , such that for each @xmath386 , the restriction of @xmath102 to @xmath392 is s - continuous , and , injective or constant , and such that the set @xmath392 is in @xmath210-config .
now we finish the proof by noting that the above construction , applied to the restrictions of @xmath102 to @xmath392 , works definably and uniformly in @xmath393 .
we proceed by induction on @xmath250 , assuming that theorems [ lip ] and [ lipcenter ] hold for integers up to @xmath394 .
the case @xmath251 of theorem [ lipcenter ] is already proved so we may assume that @xmath395 .
by theorem [ lipcenter ] in the case @xmath251 that we just proved and then by logical compactness , it is enough to consider the case of an @xmath180-definable part @xmath396 such that , for some coordinate projection @xmath239 and some @xmath180-definable functions @xmath397 and @xmath398 , @xmath399 is in @xmath400-config , and @xmath401 is in @xmath402-config for each @xmath403 .
again by theorem [ lipcenter ] in the case @xmath251 and logical compactness ( we keep for simplicity the notation @xmath17 for the part of @xmath78 we have now to work on ) , we may assume that , for some coordinate projection @xmath404 , there are @xmath180-definable functions @xmath405 such that @xmath406 is in @xmath407-config , @xmath408 is in @xmath409-config , and @xmath410 is in @xmath411-config , for each @xmath412 . by corollary [ cor : invers ] and by theorems [ lip ] and [ lipcenter ] for @xmath394 , we may suppose that @xmath17 is open .
the reduction to the case where @xmath198 is @xmath32-lipschitz continuous is done as in @xcite , using s - continuity instead of the norm of the partial derivatives .
let us now describe the strategy of @xcite to make @xmath198 @xmath32-lipschitz continuous , where we refer to @xcite , proof of proposition 2.4 , for the explicit ultrametric calculations .
we shall proceed by decreasing induction on the number of variables on which @xmath198 depends nontrivially , the case when is no such variable being clear . by compactness and tameness
we may assume that @xmath198 is @xmath46-continuous in each variable separately .
after reordering the variables @xmath413 , we may suppose that @xmath414 is maximal among the @xmath415 on the whole of @xmath416 for @xmath417 .
if @xmath418 on the whole of @xmath416 , then we are done by theorem [ lip ] for @xmath394 .
hence , we may further assume that @xmath419 on the whole of @xmath416 .
now we subdivide in two cases ( possibly involving a further finite partitioning ) , where for the complete details we refer to the two cases in the proof of proposition 2.4 of @xcite : either also @xmath420 is in @xmath409-config for each @xmath421 and each @xmath244 with @xmath422 in which case we are done by induction on the number of variables on which @xmath198 depends nontrivially , or , the graph of @xmath198 is included in @xmath17 . in the latter case
, one can finish by taking the inverse function of @xmath423 and by reversing the role of @xmath164 and @xmath424 , using that @xmath406 is in @xmath407-config , and that @xmath423 is s - continuous and injective , as in @xcite . in this case one
concludes similarly as in @xcite , using the chain rule as in the proof of corollary [ cor : invers ] .
thus we may suppose that @xmath198 is @xmath32-lipschitz .
let us now show that we can reduce further to the case where @xmath86 is @xmath32-lipschitz as well , as required by the theorem .
the argument is by decreasing induction on the number of variables on which @xmath86 depends .
let us summarize the relevant current assumptions .
the set @xmath399 is in @xmath400-config , @xmath401 is in @xmath402-config , @xmath410 is in @xmath411-config , and @xmath248 and @xmath425 are s - continuous for each @xmath403 and for each @xmath412 .
furthermore , we may assume that @xmath198 is @xmath32-lipschitz , @xmath17 is open and @xmath248 is injective for each @xmath403 .
moreover , by theorem [ lip ] for @xmath394 and by compactness we may require that @xmath426 is @xmath32-lipschitz for each @xmath164 .
define @xmath380 as the image of @xmath17 under the function @xmath427 sending @xmath16 to @xmath428 .
we may suppose that there are @xmath429 such that , for each @xmath403 , one has @xmath430
@xmath431 for some sets @xmath432 . by proposition [ continuity ] for @xmath251 and by compactness
, we may suppose that for each @xmath412 and each open ball @xmath131 contained in @xmath406 , these sets @xmath433 do not depend on the choice of @xmath434 .
+ * case 1 .
* the function @xmath435 is @xmath32-lipschitz continuous for each @xmath436 , where @xmath436 and @xmath244 are such that @xmath437 .
+ in this case , we may perform the bi-@xmath32-lipschitz transformation @xmath438 , and assume that @xmath198 is identically zero .
this transformation preserves the assumptions summarized above .
if @xmath86 is locally @xmath32-lipschitz continuous in each variable separately , we are done by proposition [ continuity ] and by theorem [ lip ] for @xmath394 .
hence , we may suppose that @xmath86 is not locally @xmath32-lipschitz in at least one variable . by working piecewise
, we may suppose @xmath86 is nowhere locally @xmath32-lipschitz in at least one specific variable .
up to reordering the variables @xmath413 if necessary , we may thus suppose for any @xmath421 that @xmath425 is nowhere locally @xmath32-lipschitz .
suppose that there is @xmath412 and an open ball @xmath131 contained in @xmath406 such that @xmath439 for some @xmath440 .
then this violates the @xmath32-lipschitz continuity of @xmath102 in the variable @xmath424 as follows .
fix @xmath440 satisfying ( [ t1t2 ] ) .
choose @xmath441 in @xmath442 such that the sets @xmath443 and @xmath444 are disjoint balls , with @xmath445 by remark [ rem : comp ] on the correspondence of maximal balls in domain and range of the functions @xmath446 , we can take @xmath130 in @xmath447 , and @xmath448 with @xmath449 such that @xmath450 lies in @xmath353 for @xmath451 . by s - continuity of @xmath425 and
the fact that @xmath425 is nowhere locally @xmath32-lipschitz continuous , and the note below definition [ defjacprop ] , one has @xmath452 since @xmath443 and @xmath444 are disjoint , it follows from their description that one has for any @xmath453 for @xmath451 that @xmath454 combining these inequalities with @xmath455 , one finds @xmath456 which indeed violates the @xmath32-lipschitz continuity of @xmath102 in the variable @xmath424 .
hence , we may suppose that for each @xmath457 and each open ball @xmath131 contained in @xmath406 , the set @xmath458 is independent of the choice of @xmath434 .
but then it follows that @xmath458 is in @xmath411-config , as we wanted to prove . +
* case 2 . *
the graph of @xmath198 is contained in @xmath17 and @xmath459 .
+ in this case we may assume , by theorem [ lip ] for @xmath394 and compactness , that @xmath460 is @xmath32-lipschitz as well .
but then it follows from the case assumption , namely from @xmath459 and from a chain rule , see below definition [ defjacprop ] , that @xmath86 is @xmath32-lipschitz and we are done also for this case .
+ we now explain how one can deduce the general case from case 1 and case 2 .
let us write @xmath461 so that @xmath462 and @xmath463 for the function sending @xmath464 to @xmath465 .
take an @xmath180-definable function @xmath466 for some @xmath46 such that , for each @xmath208 , each @xmath467 the function @xmath468 is @xmath32-lipschitz on @xmath469 for each @xmath436 . in general @xmath246
may not be in @xmath132-config .
however , we may assume that @xmath246 is in @xmath470-config for some @xmath180-definable function @xmath471 .
we may also assume that either the graph of @xmath471 is disjoint from @xmath472 , or is included in @xmath472 .
in the first case , one notes that @xmath246 is in fact in @xmath132-config , and one falls in case 1 . in the remaining case
when the graph of @xmath471 is included in @xmath472 , we may suppose that @xmath473 by the proof of theorem [ lipcenter ] for @xmath251 .
if @xmath471 is @xmath32-lipschitz then we fall in case 2 and we are done .
if @xmath471 is not @xmath32-lipschitz , after performing a permutation of the variables as we did in the beginning of this proof for @xmath198 , and transforming @xmath86 accordingly , we may assume @xmath471 is @xmath32-lipschitz , and one falls again in a case already treated .
we amend on @xcite and @xcite , more precisely on their proofs of the piecewise lipschitz continuity results .
the explanation of the reduction to the cases 1 and 2 as in the proof of theorem [ lipcenter ] is not given in the proof of theorem 2.3 of @xcite , and , only case 1 is treated in @xcite , namely by assuming ( * ) on page 83 of @xcite . either one adds a case 2 and a reduction to cases 1 and 2 , or , one uses the simplified approach of this paper .
if one uses the approach of the present paper , one should adapt @xcite accordingly , and use the monomial approximation result of @xcite to get rid of the constant @xmath474 as created in the proof of theorem [ lipmixed ] below for the analogues of ( [ equi000 ] ) and ( [ equi00 ] ) .
we proceed by induction on @xmath250 , assuming that theorem [ lipcenter ] holds for integers up to @xmath250 . for @xmath475
there is nothing to prove concerning the statement of theorem [ lip ] .
write @xmath476 for the coordinate projection sending @xmath477 to @xmath478 and define @xmath380 as the image of @xmath17 under the function @xmath427 sending @xmath16 to @xmath479 . clearly , by induction on the number of variables on which @xmath102 depends ,
lemma [ invers ] , corollay [ cor : invers ] , theorem [ lipcenter ] , tameness , compactness and by going to a test pair , we may assume that the following basic assumptions hold .
* basic assumptions .
* 1 . @xmath17 is open in @xmath480 .
2 . @xmath481 is s - continuous for each @xmath482 in @xmath416 .
3 . @xmath426 is @xmath32-lipschitz continuous for each @xmath164 .
4 . for each @xmath482 in @xmath416 ,
the set @xmath483 is in @xmath484-config , where @xmath198 is an @xmath180-definable function .
5 . for each @xmath482 in @xmath416 ,
the set @xmath485 is in @xmath486-config , where @xmath86 is an @xmath180-definable function .
@xmath198 and @xmath86 are @xmath32-lipschitz on @xmath416 .
we prove that under these basic assumptions , @xmath102 is globally @xmath32-lipschitz . by replacing @xmath102 by @xmath487
we may suppose that @xmath488 .
we may also assume that latexmath:[\[\label{equi0000 } replace @xmath198 by @xmath132 on the piece where one has @xmath490 .
consider @xmath491 .
if @xmath164 and @xmath492 lie in the same open ball @xmath131 which is included in @xmath483 with @xmath493 , then one derives from the assumptions : @xmath494 which ends the proof in this case .
now suppose that @xmath164 and @xmath492 do not lie in any open ball included in @xmath483 with @xmath493 , and , by symmetry , that @xmath164 and @xmath492 do not lie in any open ball which is included in @xmath495 with @xmath496 .
note that this implies that @xmath497 we also have latexmath:[\[\label{equi00 } by s - continuity as given by ( 1 ) , since @xmath102 is locally @xmath32-lipschitz and @xmath488 .
combining ( [ equi0000 ] ) , ( [ equi000 ] ) , ( [ equi00 ] ) one gets @xmath499 and we are done .
recall that the generalized denef - pas language @xmath500 consists of the sorts @xmath501 for valued field , the @xmath502 for @xmath503 for the residue rings modulo the product ideal of the ideal @xmath504 and the maximal ideal , and @xmath505 for the union of @xmath111 with the multiplicatively written value group @xmath506 , and having as symbols the ring language on @xmath501 , the ring language on the @xmath502 , the language of ordered multiplicative groups @xmath507 on @xmath506 , the norm map from @xmath501 to @xmath505 , and angular component maps @xmath508 for all @xmath503 .
an angular component map @xmath508 is just a multiplicative map @xmath509 , extended by zero on zero , that coincides with the natural projection on the units of the valuation ring to @xmath502 .
the maps @xmath510 are required to form a compatible system , that is , the composition of the projection @xmath511 with @xmath510 has to equal @xmath512 whenever @xmath33 divides @xmath250 .
let @xmath513 be this generalized language of denef - pas @xmath500 , but without multiplication on @xmath514 and on the @xmath502 .
let @xmath180 be any first order language with the same sorts as @xmath515 and such that @xmath180 contains all the symbols of @xmath515 .
let @xmath106 be an @xmath513 structure , where we write @xmath142 for @xmath516 , @xmath110 for @xmath517 and @xmath518 for @xmath519 .
call a set @xmath520 auxiliary if it is a subset of a cartesian product of some copies of @xmath142 and the @xmath518 .
let us use the notation @xmath521 furthermore , let @xmath522 when @xmath523 be shorthand for @xmath111 .
[ config2 ] given a set @xmath524 with @xmath520 auxiliary , say that @xmath39 is in @xmath198-config with depth @xmath250 if there exists @xmath525 such that @xmath39 equals the union over @xmath168 of sets @xmath526 for some @xmath527 . again , if @xmath39 is nonempty and in @xmath198-config , then @xmath250 , @xmath141 and the @xmath204 , where @xmath130 is such that @xmath528 , are uniquely determined by @xmath39 and @xmath198 .
say that @xmath39 is in @xmath180-tame config if there exist @xmath66 , @xmath206 and @xmath180-definable functions @xmath529 such that the range of @xmath198 contains no open ball , and such that @xmath530 is in @xmath210-config with depth @xmath250 for each @xmath531 .
one could as well have used @xmath532 instead of @xmath533 for some independent @xmath46 and @xmath189 , but this would create heavier notation .
suppose that the @xmath513 structure on @xmath106 can be expanded to be a structure for the generalized denef - pas language .
this condition is a simplification that replaces some of the conditions of section [ sec:5b ] , which would otherwise have become more cumbersome in the mixed case .
we maintain this condition throughout this section .
suppose moreover that @xmath106 is an @xmath180-structure .
consider an @xmath180-structure @xmath213 which is elementarily equivalent to @xmath214 . if @xmath91 is a language which is obtained from @xmath180 by adding constants from @xmath213
, then we call @xmath215 a test pair for @xmath214 .
[ mixktame ] say that @xmath214 is weakly mixed tame if the following hold : 1 .
each @xmath180-definable set @xmath524 with @xmath520 an auxiliary set is in @xmath180-tame config ; 2 .
if @xmath217 is @xmath180-definable , then there exists an @xmath180-definable function @xmath534 for some @xmath218 such that , for each @xmath531 , the restriction of @xmath158 to @xmath220 is s - continuous .
say that @xmath214 is mixed tame if each test pair @xmath215 for @xmath214 is weakly tame .
more generally , call an @xmath180-theory @xmath221 mixed tame if for each model @xmath222 of @xmath221 , the pair @xmath223 is mixed tame . by essentially the same proof as that of theorem [ lip ] , we obtain our final result on lipschitz continuity .
[ lipmixed ] suppose that @xmath214 is mixed tame .
let @xmath224 be an @xmath180-definable function which is locally @xmath32-lipschitz .
then there exists an integer @xmath535 and an @xmath180-definable function @xmath536 for some @xmath218 such that for each @xmath531 , the restriction of @xmath102 to @xmath220 is @xmath474-lipschitz .
adapt the proof of theorem [ lip ] , and all its auxiliary results and their proofs , by replacing any occurrence of ( the residue field ) @xmath108 by a residue ring @xmath537 for some @xmath46 , and each occurrence of @xmath192 by @xmath538 for some @xmath46 .
( see for example proposition [ inj - cons - mixed ] for the adaptation of proposition [ inj - cons ] . ) in this process , a constant of the form @xmath474 shows up in the upper bound in equations ( [ equi000 ] ) and ( [ equi00 ] ) , where @xmath539 can be bounded in terms of the depths of the occurring configurations . the natural analogue with @xmath474-lipschitz centers for some integer @xmath535 of theorem [ lipcenter ] in mixed characteristic also holds .
we leave its proof to the reader . instead
, we make explicit the analogue of proposition [ inj - cons ] , as proposition [ inj - cons - mixed ] .
[ inj - cons - mixed ] suppose that @xmath214 is mixed tame .
let @xmath217 be @xmath180-definable .
then there exist an integer @xmath206 and an @xmath180-definable function @xmath536 such that for each @xmath531 the restriction of @xmath158 to @xmath220 is injective or constant .
similar adaptation of the proof of proposition [ inj - cons ] as explained in the proof of theorem [ lipmixed ] .
the following proposition provides examples of ( mixed ) tame structures .
[ pex ] let @xmath500 be the generalized denef - pas language .
suppose that @xmath106 is a valued field of characteristic zero , equipped with angular component maps @xmath510 , and , a separated analytic @xmath540-structure as in definition 4.1.6 of @xcite , where @xmath540 is a weierstrass system as in definition 4.1.5 of @xcite , and write @xmath180 to denote the corresponding expansion of @xmath500 .
then @xmath214 is tame , resp .
mixed tame , if @xmath106 is of equicharacteristic zero , resp . of mixed characteristic .
one readily derives this statement from the version of theorem 6.3.7 of @xcite which is formulated with sorts for quotients @xmath541 with @xmath542 the maximal ideal of the valuation ring of @xmath106 , instead of with the sorts @xmath502 .
note that @xmath106 is automatically henselian because it has a separated analytic @xmath540-structure with @xmath540 a weierstrass system . in some specific cases proposition
[ pex ] follows , alternatively , by results from @xcite , @xcite , @xcite , @xcite , resp .
@xcite ; see also @xcite for a related , one - sorted result in the @xmath0-adic subanalytic case .
all the structures that we will use in section [ sec:2 ] are tame or mixed tame , by examples 4.4(1 ) and 4.4(13 ) of @xcite , section 3.4 of @xcite , and by proposition [ pex ] .
note that by compactness , family versions of theorems [ lip ] , [ lipmixed ] and [ lipcenter ] follow naturally .
uniform versions in all models of a tame theory follow likewise .
[ sec : gyt ] [ [ section-1 ] ] in section [ sec:5 ] , piecewise lipschitz continuity was obtained for a definable function with bounded first partial derivatives . in this section
, we will show that one can parametrize any definable set by a small set of maps with bounded partial derivatives up to any given finite order . in the previous sentence ,
a small set of maps means a set of maps indexed definably by the residue field , or more generally , some residue rings @xmath543 for some @xmath535 .
we will use piecewise lipschitz continuity from the previous section , together with new techniques using a strong kind of analyticity .
we will furthermore define a property @xmath544 for approximation by taylor polynomial of degree @xmath73 with remainder term of degree @xmath45 and the property @xmath544 will be required in our parametrizations ( see theorem [ gyt ] ) . to distinguish from the more abstract setting of section [ sec:5 ] , we will write @xmath91 , instead of @xmath106 , in this section , where @xmath91 will be a valued field . [ [ subsec : gyt ] ]
let @xmath91 be a complete , discretely valued field of characteristic zero such that , for each integer @xmath545 the set @xmath546 of the @xmath250-th powers in @xmath547 has finite index in @xmath547 .
write @xmath548 for the valuation ring of @xmath91 with maximal ideal @xmath549 , and residue field @xmath550 .
let us choose a uniformizer @xmath551 of @xmath548 and let us write @xmath552 for the characteristic of @xmath550 .
write @xmath553 for the ( surjective ) valuation map and write @xmath188 for the multiplicative norm on @xmath91 with normalization @xmath554 for some real number @xmath555 .
when @xmath550 is finite we set @xmath556 . as usual , the norm @xmath147 of a tuple @xmath557 is set to be the maximum of @xmath558 for @xmath60 , and , in the case where @xmath91 is a @xmath0-adic field , @xmath559 is taken to be the inverse of the number of elements in @xmath550 .
further , for each @xmath503 write @xmath560 for @xmath561 and write @xmath562 for the function which sends @xmath132 to @xmath132 and nonzero @xmath563 to @xmath564 . note that , if @xmath550 has characteristic @xmath132 , then one has @xmath565 and @xmath566 for all @xmath503 .
we also write @xmath192 for @xmath567 .
let @xmath568 be the first order language having sorts for @xmath91 , the @xmath560 for @xmath503 , and @xmath569 , and having as symbols the ring language with parameters from @xmath91 on @xmath91 , the ring language on @xmath550 , the presburger language @xmath570 on @xmath569 , the valuation map @xmath553 , and the maps @xmath571 for all @xmath503 .
a function @xmath572 which satisfies
@xmath573 whenever @xmath574 is called a restricted analytic function if there is a power series @xmath575 in @xmath576 , converging on @xmath577 , such that @xmath578 for @xmath579 .
let @xmath580 be the language consisting of @xmath568 and all the restricted analytic functions @xmath581 for all @xmath66 .
an @xmath568-definable subset of @xmath582 is often called a semialgebraic subset of @xmath582 .
an @xmath580-definable subset of @xmath582 is called a subanalytic subset of @xmath582 .
let @xmath180 be either @xmath568 or @xmath580 . from now on in section
[ sec:2 ] , definable sets and functions will be so for the language @xmath180 .
note that the study of definable sets was initiated in the works of macintyre @xcite , denef and van den dries @xcite in the @xmath0-adic case , and was generalized later to this and other settings in e.g. @xcite , @xcite , @xcite , and @xcite . for a nonempty definable set @xmath583
, the dimension of @xmath17 is defined as the largest integer @xmath584 such that , for at least one of the coordinate projections @xmath585 , the set @xmath416 has nonempty interior for the valuation topology on @xmath586 .
the empty set is given dimension @xmath587 .
for an integer @xmath67 , and similarly for @xmath588 , the @xmath72-norm of a @xmath72-function @xmath589 on an open @xmath590 is defined as the supremum over all @xmath16 in @xmath163 , all @xmath60 , and all @xmath591 with @xmath592 , of the values @xmath593 where @xmath594 stands for @xmath595 , and @xmath596 for @xmath597 . for a @xmath72-function @xmath598 and @xmath599 , write @xmath600 ( or @xmath601 ) for the tuple of the taylor polynomials of the @xmath100 at @xmath153 of degree @xmath73 .
we now define a notion of ( global ) approximation by taylor polynomials .
[ tr ] let @xmath17 be a subset of @xmath586 .
let @xmath45 be a positive integer .
we say that a map @xmath602 satisfies @xmath603 ( on @xmath17 ) if @xmath17 is open in @xmath586 , @xmath102 is @xmath72 with @xmath72-norm not larger than @xmath32 , and for every @xmath16 and @xmath153 in @xmath17 one has @xmath604 if a map @xmath102 satisfies @xmath544 on a subset of @xmath605 , then it also satisfies @xmath606 for any @xmath607 with @xmath608 , by the ultrametric inequality . for a @xmath609-function from an open subset of @xmath586 to @xmath577 ,
satisfying @xmath610 is equivalent to being lipschitz continuous with lipschitz constant @xmath32 .
let @xmath583 be a definable set of dimension @xmath33 .
a family @xmath611 of definable functions for @xmath272 running over some set @xmath612 and with @xmath613 , is called a @xmath544-parametrization of @xmath17 if each of the @xmath100 satisfies @xmath544 and @xmath614 a family @xmath17 of sets @xmath615 for @xmath153 running over a definable set @xmath380 is called a definable family if @xmath616 is a definable set . a collection of maps @xmath617 for @xmath153 running over a definable set @xmath380 is called a definable family of maps if the collection @xmath102 of the graphs of the @xmath617 is a definable family of sets .
recall that for @xmath618 a map and @xmath227 , we write @xmath228 for the set @xmath229 and @xmath619 or @xmath620 for the function @xmath621 on @xmath228 . [ gyt ] let @xmath66 , @xmath9 and @xmath622 be integers and let @xmath623 be a definable family of subsets @xmath624 for @xmath153 running over a definable set @xmath380 .
suppose that @xmath615 has dimension @xmath33 for each @xmath625 .
then there exist an integer @xmath539 and a definable family @xmath626 of definable functions @xmath627 such that @xmath628 and for each @xmath153 , @xmath629 forms a @xmath544-parametrization of @xmath615 .
namely , @xmath630 and @xmath631 satisfies @xmath544 on @xmath632 for each @xmath625 and each @xmath633 .
let us first describe the strategy of the proof of theorem [ gyt ] .
we introduce a global notion of analyticity ( global in the sense that the radii of convergence of the power series are large in a certain sense ) , and show a globally analytic cell decomposition theorem . a first step towards theorem [ gyt ]
is to parametrize our set with functions having small @xmath609-norm .
this is done by inverting the roles of some of the coordinates and using the chain rule to bound the @xmath609-norm by @xmath32 .
the analyticity allows us to go further by working with gauss - norms on balls and on boxes ( defined as products of balls ) .
this has two uses : to obtain that composition with well chosen power maps makes the @xmath72-norm less or equal to @xmath32 , and to obtain that @xmath72-norm bounded by @xmath32 implies @xmath544 on each maximal ball included in the domain .
moreover , by using the results of the previous section on @xmath610 , we reduce to the situation where one has globally @xmath610 and locally ( on maximal balls ) @xmath544 . by composing with power maps once more , the previous @xmath610 condition on far away
points implies @xmath544 , which then follows globally on the pieces .
finally , theorem [ gyts ] strengthens theorem [ gyt ] .
consider integers @xmath66 and @xmath634 for @xmath635 .
a nonempty definable set @xmath583 is called a cell around zero with depth @xmath636 if it is of the form @xmath637 for some set @xmath638 and some @xmath639 .
if moreover @xmath640 is a subset of @xmath641 , where @xmath642 and @xmath643 by convention , then @xmath17 is called an open cell around zero .
more generally , for nonempty definable sets @xmath380 and @xmath644 , the set @xmath17 is called a cell around zero over @xmath380 with depth @xmath636 if it is of the form @xmath645 for some set @xmath646 and some definable functions @xmath647 .
if moreover @xmath640 is a subset of @xmath648 , then @xmath17 is called an open cell around zero over @xmath380 .
note that the definability of @xmath640 is not an issue here since @xmath17 is assumed to be definable .
note that some of the sets @xmath615 for some @xmath625 may be empty . by a box
we mean a cartesian product of closed balls , where a closed ball is a subset of @xmath91 of the form @xmath649 for some @xmath45 in @xmath547 and some @xmath650 .
for a box @xmath131 of the form @xmath651 , we define the associated set @xmath652 as @xmath653 where @xmath654 is an algebraic closure of @xmath91 with norm extending the one on @xmath91 .
note that @xmath655 .
we extend the definition of @xmath72-norm , for an integer @xmath67 and also for @xmath588 , of a @xmath72-function @xmath656 on an open @xmath657 in the obvious way .
[ glob ] let @xmath658 be a definable function on an open set @xmath17 .
say that @xmath102 is globally analytic on @xmath17 if , for any box @xmath131 contained in @xmath17 , the restriction of @xmath102 to @xmath131 is given by a tuple of power series , converging on the associated set @xmath652 . by this
we mean that for any @xmath659 , there is a power series @xmath660 such that @xmath661 for all @xmath662 and which converges on @xmath663 .
the notion of globally analytic maps will be most useful when the domain is an open cell around zero .
we can now give in theorem [ gyts ] a strengthened version of theorem [ gyt ] as well as a variant of theorem [ gyts ] in proposition [ gyts ] .
[ gyts ] with data and notation from theorem [ gyt ] , one can take @xmath212 as in theorem [ gyt ] and such that moreover @xmath631 is globally analytic for each @xmath664 and such that @xmath665 is an open cell around zero over @xmath666 , where @xmath667 .
[ gyts ] with data and notation from theorem [ gyt ] , one can take @xmath212 as in theorem [ gyts ] such that moreover for each @xmath664 , and for each box @xmath131 contained in @xmath632 and associated set @xmath652 , the power series corresponding to @xmath631 on @xmath131 satisfies @xmath544 on the whole of @xmath652 .
[ [ section-2 ] ] in order to prove theorems [ gyt ] and [ gyts ] , we give now some preliminary definitions . for a closed ball @xmath649 with @xmath45 in @xmath547 and @xmath650
, the real number @xmath668 is called the radius of the closed ball , while @xmath669 is called the valuative radius .
all balls from now on will be closed balls , as opposed to section [ sec:5 ] where we used open balls .
a ball @xmath131 with @xmath670 for some set @xmath671 is called a maximal ball contained in @xmath17 if @xmath131 is a closed ball which is maximal for the inclusion among all closed balls contained in @xmath17 . by convention ,
@xmath672 stands for @xmath111 , and so do also @xmath673 and @xmath674 , namely the definable set of a true formula without free variables .
we complement the above notion of cells around zero by a notion of cells with a center .
[ cellc ] consider integers @xmath66 and @xmath634 for @xmath635 . for non empty definable sets @xmath380 and @xmath644 , the set @xmath17
is called a cell over @xmath380 with center @xmath675 and depth @xmath676 if it is of the form @xmath677 for some set @xmath646 and some definable functions @xmath647 and @xmath678 , where @xmath679 .
if moreover @xmath640 is a subset of @xmath648 , where @xmath642 , then @xmath17 is called an open cell over @xmath380 ( with center @xmath675 and depth @xmath676 ) . [ cell0 ]
let @xmath17 be a cell over @xmath380 with center , with notation from definition [ cellc ] .
the cell around zero associated to @xmath17 is by definition the cell @xmath472 obtained by forgetting the centers , namely @xmath680 with associated bijection @xmath681 sending @xmath682 to @xmath683 . for a definable map @xmath684 there is the natural corresponding function @xmath685 from @xmath472 to @xmath78 .
[ cella ] suppose that @xmath644 is an open cell over @xmath380 .
if @xmath475 then @xmath17 is a globally analytic cell over @xmath380 . for @xmath545 , if the image of @xmath17 under the coordinate projection @xmath0 from @xmath686 to @xmath687 is a globally analytic cell , and if @xmath688 is globally analytic on @xmath689 for each @xmath625 in the sense of definition [ glob ] , then @xmath17 is called a globally analytic cell over @xmath380 .
[ cda ] given definable sets @xmath380 and @xmath644 and a definable map @xmath690 , there exist @xmath535 and a definable bijection @xmath691 over @xmath17 such that @xmath692 is the disjoint union of a cell with empty interior and an open globally analytic cell @xmath78 over @xmath693 such that @xmath694 is globally analytic on @xmath695 for each @xmath696 and @xmath625 .
we define an expansion @xmath697 of @xmath180 similar to the one of ( 4.1 ) of @xcite , and to the one of definition 6.1.7 of @xcite , by joining division and witnesses for henselian zeros and roots .
[ hf1 ] let @xmath697 be the expansion of @xmath698 obtained by joining to @xmath698 function symbols @xmath699 and @xmath700 for @xmath701 and @xmath702 , where on a henselian valued field @xmath106 of characteristic zero and with value group @xmath703 these functions are : @xmath704 sends @xmath705 to the ( unique ) @xmath33-th root @xmath153 of @xmath16 with @xmath706 and @xmath707 , whenever simultaneously @xmath708 ,
@xmath709 , and @xmath710 is divisible by @xmath33 in @xmath703 , and to @xmath132 otherwise ; @xmath711 sends @xmath712 to the unique @xmath153 satisfying @xmath713 , @xmath706 , and @xmath714 , whenever @xmath141 is a unit , @xmath715 , @xmath716 , and @xmath717 with @xmath718 the derivative of @xmath102 , and to @xmath132 otherwise . [ terms ] given a definable function @xmath719 , there exists a definable bijection @xmath720 over @xmath17 and a tuple of @xmath697-terms @xmath254 such that @xmath721 for all @xmath722 and with @xmath723 . by @xcite ,
theorem 7.5 , the proposition holds in a slightly different setting .
namely in @xcite , an extra value group variable is allowed as input in the function @xmath699 and in parametrizations @xmath724 runs over a cartesian product of @xmath725 with the value group . since in the present case the value group is simply @xmath569 , the proposition as stated follows directly from @xcite , theorem 7.5 .
we proceed by induction on @xmath250 . for @xmath475
, there is nothing to prove .
we will use the classical form of cell decomposition without global analyticity , which follows immediately from tameness and compactness , or alternatively , by theorem 7.4 of @xcite .
suppose now that @xmath503 .
let @xmath697 be the expansion of @xmath180 given by definition [ hf1 ] . by theorem [ terms
] , we may suppose that @xmath102 is given by a tuple of @xmath697-terms @xmath726 .
we may focus on @xmath727 among the @xmath726 .
we proceed now by induction on the complexity of the term @xmath727 .
suppose that @xmath727 equals @xmath728 for some @xmath91-valued terms @xmath729 and a function symbol @xmath254 of @xmath697 . by the classical form of cell decomposition and both ongoing inductions
, we may assume that @xmath17 is already a globally analytic cell over @xmath380 , that the @xmath730 are globally analytic for each @xmath625 , and that , for a chosen @xmath731 , @xmath732 and @xmath733 are constant on each box contained in @xmath734 .
now , by choosing @xmath213 appropriately depending on @xmath254 ( as explained in the proof of lemma 6.3.15 of @xcite for each possibility for @xmath254 ) , the theorem follows .
the following elementary lemma about compositions will often be used without mentioning .
[ prod ] let @xmath735 be integers .
let @xmath736 and @xmath737 be locally analytic functions on some open subsets @xmath738 and @xmath739 .
assume that @xmath102 and @xmath212 satisfy @xmath544 .
then the composition @xmath740 satisfies @xmath544 .
just use that the taylor polynomial of a composition corresponds to the composition of the taylor polynomials ( of respective degrees and up to a certain degree ) , and use a classical ultrametric calculation .
the composition with power maps has been used in the context of real parametrizations by yomdin , gromov , pila and wilkie .
we will use power maps similarly and introduce the following notation for convenience .
[ fnb ] for @xmath741 a definable function , for any integer @xmath535 and any @xmath742 , write @xmath743 for the set of all @xmath744 such that @xmath745 lies in @xmath78 , and write @xmath746 for the function @xmath747 .
let us recall how convergent power series over @xmath548 may also be interpreted in other valued fields , even when they are non - complete or of higher rank . for @xmath9
, put @xmath748 , namely the ring of formal power series in @xmath16 over @xmath548 and converging on @xmath749 .
write @xmath750 for the ring of @xmath380-valued functions on @xmath17 for any sets @xmath751 .
let @xmath752 be a valued field with valuation ring @xmath753 and maximal ideal @xmath754 .
an analytic @xmath755structure on @xmath752 is the data of ring homomorphisms @xmath756 for all @xmath9 , satisfying * [ 1 ) ] @xmath757 , * [ 2 ) ] @xmath758 the @xmath272-th coordinate function on @xmath759 for @xmath760 , and * [ 3 ) ] @xmath761 extends @xmath762 with the natural inclusions @xmath763 and @xmath764 inducing @xmath765 .
we also consider one - sorted variants @xmath766 and @xmath767 of @xmath180 .
[ hfl1 ] for @xmath106 a henselian field , let @xmath768 be the function that associates to @xmath769 the unique zero , @xmath198 , of the polynomial @xmath770 that satisfies @xmath771 , if @xmath772 and @xmath773 , and let @xmath774 output @xmath132 in all other cases . corresponding to the choice of @xmath180 as either @xmath568 or @xmath580 , let @xmath766 be the valued field language @xmath775 with coefficients from @xmath91 , resp . , the valued field language together with function symbols for each element of @xmath776 for all @xmath9 .
let @xmath767 be @xmath766 together with the function symbols @xmath774 for all @xmath66 .
[ qel1h ] let @xmath654 be an algebraically closed valued field with analytic @xmath755structure .
then @xmath654 admits quantifier elimination in ( the one - sorted language ) @xmath767 .
the following lemma gives uniform bounds on gauss - norms and is based on lemma 6.3.9 of @xcite .
recall that the gauss - norm of a power series is the supremum of the norms of the coefficients .
write @xmath777 for the maximal ideal of the valuation ring @xmath778 of @xmath654 .
[ gauss ] let @xmath779 be a definable family of functions for @xmath153 in a definable set @xmath380 .
suppose that for each @xmath625 , @xmath617 is given by a power series with coefficients in @xmath91 which converges on the associated set @xmath780 .
suppose further for each @xmath625 and each @xmath781 that the partial derivatives @xmath782 have norm at most @xmath32 on @xmath652 .
then there is a nonzero integer @xmath213 such that the gauss - norm of @xmath783 is at most one for each @xmath625 .
the proof is somewhat easier in the equicharacteristic zero case , but we will give a uniform treatment .
our lemma is implied by compactness by the following more general and abstract result .
let @xmath752 be an algebraically closed valued field with an analytic structure in the sense of definition 4.1.6 of @xcite and let @xmath784 be given by a power series in @xmath785 of the separated weierstrass system of the analytic structure .
suppose for each @xmath781 that @xmath786 has norm at most @xmath32 on @xmath754 .
then there is a nonzero integer @xmath213 such that the gauss - norm of @xmath787 is at most one .
when @xmath89 this follows from the first part of the proof of lemma 6.3.9 of @xcite ( showing that @xmath198 is not infinitesimal ) .
the statement for general @xmath33 follows from induction on @xmath33 , property ( v ) of definition 4.1.2 of @xcite , and compactness .
the following corollary expresses that , for power series in our setting , bounded @xmath609-norm almost implies @xmath610 .
[ cor : gauss ] let @xmath788 be a definable family of functions for @xmath153 varying in a definable set @xmath380 .
suppose that for each @xmath625 , @xmath789 is a box and that @xmath617 is given by a power series with coefficients in @xmath91 which converges on the associated set @xmath790 .
suppose further that , for each @xmath625 , the function @xmath617 has @xmath609-norm at most @xmath32 on @xmath652 .
then there is a nonzero integer @xmath213 such that , for each @xmath625 , the function @xmath791 satisfies @xmath610 on @xmath790 .
let us first consider the situation for fixed @xmath153 , so we may write @xmath102 instead of @xmath617 , and so on . since @xmath131 is a cartesian product , to prove this case we may furthermore assume @xmath89 . up to a translation
, we may suppose there exists a bijection @xmath792 for some nonzero @xmath793 , and that @xmath794 .
the function @xmath795 is globally analytic and has @xmath609-norm at most @xmath32 on @xmath777 , by the chain rule for differentiation . by lemma [ gauss ] , the gauss - norm of @xmath796 is at most @xmath32 for some @xmath731 .
let us write @xmath797 for the series @xmath798 , and thus @xmath799 for the series @xmath800 . by the bound on the gauss - norm of @xmath796
, one has @xmath801 for @xmath802 let us write @xmath803 and @xmath804 for @xmath805 in @xmath777 , and @xmath806 rewriting @xmath807 by @xmath808 , one thus finds @xmath809 this proves the statement for fixed @xmath153 .
the general case follows from the case @xmath89 and the uniformity in @xmath213 given by lemma [ gauss ] .
we proceed by induction on @xmath33 , the case of @xmath810 being trivial by ( mixed ) tameness , ( see section [ ex - tame ] )
. the proof will combine corollary [ cor : invers ] , theorems [ lip ] , [ lipmixed ] and [ cda ] , and will require going to an algebraic closure @xmath654 of @xmath91 to control the @xmath609-norms on @xmath654 . by using the two constants @xmath132 and @xmath32 in @xmath550 or in @xmath543 to realize disjoint unions , it is clear that we may proceed by working piecewise on @xmath17 .
also , by induction on the dimension of @xmath17 and the dimension theory as in @xcite , we may replace @xmath17 by a definable subset whose complement in @xmath17 has dimension less than @xmath33 .
we may therefore assume that we have a definable bijection @xmath811 where @xmath812 and the @xmath813 are open in @xmath749 for each @xmath625 and @xmath696 , and such that the @xmath814 are @xmath609 for each @xmath153 and @xmath41 .
indeed , such a bijection can be found by ( a basic form of ) theorem [ cda ] . by theorem [ terms ]
, we may suppose that the components of @xmath815 are given by @xmath697-terms @xmath816 , with notation from definition [ hf1 ] .
let @xmath654 be an algebraic closure of @xmath91 with norm extending the one on @xmath91 .
we now explain how to switch between @xmath91 and @xmath654 , to improve @xmath815 .
this passage to the algebraic closure will preserve the necessary information by the term structure given by proposition [ terms ] for @xmath91 and by theorem 3.4.3.(i ) of @xcite for @xmath654 , and by the quantifier elimination result stated as theorem [ qel1h ] .
the field @xmath654 has a natural @xmath697-structure and @xmath767-structure by lemma 3.3.6 and theorem 3.4.1 of @xcite and theorem 4.5.11 of @xcite ; this @xmath697-structure expands the natural @xmath767-structure on @xmath654 .
moreover , every @xmath697-term corresponds in @xmath817 naturally to an @xmath767-term , where the variables from @xmath818 are replaced by ( new ) variables over @xmath654 .
for example , the @xmath697-term @xmath699 is interpreted by the function @xmath819 which sends @xmath820 to the ( unique ) @xmath33-th root @xmath153 of @xmath16 with @xmath821 , whenever simultaneously @xmath822 , @xmath823 , and to @xmath132 otherwise , which is given by an @xmath767-term .
let us now associate to the @xmath697-terms @xmath726 in @xmath654 the corresponding @xmath767-terms @xmath824 and consider them as @xmath654-valued functions on @xmath825 for some @xmath826 .
we will mimic the proof of corollary [ cor : invers ] to make the @xmath609-norm of the @xmath824 small and then go back to @xmath91 . by theorem [ qel1h ] , there is a finite quantifier free @xmath767-definable partition of @xmath825 with pieces @xmath827 such that , up to neglecting lower dimensional parts and possibly permuting coordinates , we may suppose for each piece @xmath827 that @xmath828 is maximal among the @xmath829 for @xmath830 and @xmath781 , and that it is either at most one or larger than @xmath32 on the whole of @xmath827 . in the latter case , we may assume by compactness , propositions [ inj - cons ] and [ inj - cons - mixed ] , and reinterpreting back ( as above ) into @xmath767 if necessary , that the functions @xmath831 are injective for each @xmath832 , with inverse @xmath833 also given by an @xmath767-term by theorem 3.4.3.(i ) of @xcite . replacing the restriction of @xmath834 to @xmath827 with the function @xmath835 and by the chain rule for differentiation ( with a similar calculation as in the proof of corollay [ cor : invers ] ) , it follows that we may suppose that the functions @xmath836 have @xmath609-norm bounded by one on each @xmath837 .
since moreover such terms are almost everywhere locally analytic , the @xmath836 may be assumed to be locally @xmath610 . interpreting this data back in @xmath91
, we obtain an improved @xmath180-definable function @xmath838 . using cell decomposition for @xmath254
as provided by theorem [ cda ] , it follows from theorems [ lip ] , resp .
[ lipmixed ] , corollary [ cor : gauss ] , and by induction on @xmath33 applied to the graphs of the centers to replace @xmath839 by @xmath840 as in definition [ cell0 ] , that one may assume that there is an integer @xmath841 ( with @xmath842 in the equicharacteristic zero case ) such that @xmath843 is as desired , but for @xmath844 instead of @xmath17 , with @xmath845 coming from the use of theorem [ lipmixed ] and corollary [ cor : gauss ] .
we are done if @xmath550 is of characteristic zero , since then @xmath842 .
now suppose that @xmath550 is of positive characteristic .
in that case we shall make use of the fact that , for @xmath16 close enough to @xmath153 in @xmath778 , one has , with @xmath846 , @xmath847 .
note also that it is always possible to increase the depth of the cell @xmath839 over @xmath666 , at the cost of increasing @xmath539 .
we first incease the depth of the cell @xmath839 over @xmath666 by a factor which is a power of @xmath848 , then replace the @xmath849 by composing with @xmath213-th powers for some @xmath213 which is a power of @xmath848 , and restore the condition of having open cells around zero ( which is possible since the group of @xmath213-th powers have finite index in @xmath547 ) . using ( [ powerp ] )
we observe that all conditions are met for the new @xmath254 , for a large enough choice of the powers of @xmath848 , depending only on @xmath845 .
this finishes the proof .
[ gauss0 ] let @xmath850 be a globally analytic function , where @xmath131 is the box @xmath851 for some nonzero @xmath852 .
if for some @xmath853 @xmath854 then , for all @xmath855 , @xmath856 first assume @xmath857 .
the assumptions imply that the @xmath858-norm of @xmath212 on @xmath652 is at most @xmath32 .
since the @xmath858-norm of @xmath212 on @xmath652 coincides with the gauss - norm of @xmath212 , it follows that also the latter is at most one .
hence , the gauss - norm and the @xmath858-norm of @xmath859 on @xmath652 are also at most @xmath32 . in other words , @xmath860 on @xmath652 .
the general case follows by applying the case @xmath857 to the function @xmath861 on @xmath549 . by legendre s formula @xmath862 and thus , for any positive integer @xmath250 divisible by @xmath848 , @xmath863 for all integers @xmath855 .
we now prove a key lemma allowing to go from @xmath609 to @xmath544 .
[ gauss1a ] let a positive integer @xmath45 be given . in the equicharacteristic zero case , let @xmath864 and let @xmath251 .
in the mixed characteristic case , let @xmath250 be a positive integer sufficiently divisible by @xmath848 and let @xmath539 be a positive integer divisible by @xmath865 .
let @xmath131 be the ball @xmath866 for some nonzero @xmath867 .
let @xmath868 be a globally analytic function whose @xmath609-norm is at most @xmath32 on @xmath652 .
let @xmath869 be the map @xmath870 sending @xmath16 to @xmath871 . then , for any ball @xmath872 with @xmath873 , the function @xmath874 satisfies @xmath544 on @xmath875 .
let us write @xmath876 for @xmath874 with domain @xmath234 consisting of @xmath877 such that @xmath878 lies in @xmath652 . by the chain rule and the product rule for differentiation , for any @xmath879 and for @xmath272 with @xmath880 ,
the derivative @xmath881 is a finite sum of terms of the form @xmath882 for @xmath883 and @xmath884 and with @xmath885 moreover , by lemma [ gauss0 ] applied to @xmath886 , one has @xmath887 hence , for @xmath888 with @xmath889 and for @xmath16 in @xmath234 one has @xmath890 and thus we find @xmath891 hence , for each @xmath855 @xmath892 on @xmath875 .
thus , the @xmath72-norm of @xmath876 is at most @xmath32 on @xmath652 by ( [ legendre ] ) .
now choose @xmath893 and choose @xmath894 .
develop @xmath212 around @xmath895 into a series @xmath896
. then @xmath897 by ( [ gni ] ) , which implies that @xmath212 is @xmath544 on @xmath875 as follows .
first note that @xmath898 . using this and the bounds on the @xmath899
, we have @xmath900 where the last inequality follows from ( [ legendre ] ) .
this proves the lemma .
the following is a multi - variable variant of lemma [ gauss1a ] .
[ gauss1 ] let positive integers @xmath45 and @xmath33 be given
. in the equicharacteristic zero case set @xmath251 and @xmath864 . in the mixed characteristic case , let @xmath250 be sufficiently divisible by @xmath848 and @xmath539 be sufficiently divisible by @xmath250 .
let @xmath131 be the box @xmath901 for some nonzero @xmath902 .
let @xmath868 be a globally analytic function .
let @xmath869 be the map @xmath903 sending @xmath904 to @xmath905 .
suppose that the @xmath609-norm of @xmath212 is at most @xmath32 on @xmath652 .
then , for any box @xmath906 such that @xmath873 , the function @xmath874 satisfies @xmath544 on @xmath875 . as for lemma [ gauss1a ]
. we may take @xmath907 with all properties of theorems [ gyt ] , [ gyts ] and proposition [ gyts ] with @xmath104 , since this case has already been proved .
we work with fixed @xmath625 and @xmath696 and omit @xmath153 and @xmath41 from the subscripts , explaining uniformity properties in @xmath153 and @xmath41 along the way .
take @xmath250 and @xmath539 corresponding to proposition [ gauss1 ] and our @xmath45 and @xmath33 .
we may increase the depth of the cell @xmath665 to the depth @xmath908 with @xmath909 . by proposition [ gauss1 ] ,
applied to the restrictions of @xmath212 to any box @xmath131 in its domain we find for any @xmath910 that the function @xmath911 is globally analytic and satisfies @xmath544 on @xmath875 for any box @xmath893 included in @xmath912 .
similarly as at the end of the proof of the case @xmath104 , after rewriting the function @xmath913 , we get a definable function @xmath914 as in the case @xmath104 of theorems [ gyt ] , [ gyts ] and proposition [ gyts ] .
moreover , for each @xmath625 and @xmath915 , the map @xmath916 satisfies @xmath544 on @xmath875 for each box @xmath893 in its domain .
we claim that @xmath917 is as desired in proposition [ gyts ] and theorems [ gyts ] and [ gyt ] .
there is only left to check that @xmath917 satisfies @xmath544 globally .
we still omit @xmath153 and @xmath41 from the notation . to show @xmath544 for @xmath917 we will use @xmath610 for the above @xmath212 .
choose @xmath421 and @xmath244 in the domain @xmath918 of @xmath917 .
let @xmath612 be the set of those indices @xmath272 such that @xmath919 and @xmath920 have the same order , and @xmath921 its complement in @xmath922 .
let @xmath436 be the intermediary tuple @xmath923 such that @xmath924 when @xmath925 and @xmath926 when @xmath927 .
then @xmath244 and @xmath436 lie in the same box contained in @xmath918 , and by @xmath610 for the above @xmath212 , we have @xmath928 the first of these inequalities follows from the ultrametric triangle property , the second from the fact that @xmath917 satisfies @xmath544 on each box in its domain , the third inequality follows from the construction of @xmath917 via @xmath913 , property @xmath610 for @xmath212 and the fact that @xmath929 .
the fourth and final inequality follows from the fact that @xmath930 by construction of the point @xmath436 .
indeed , @xmath931 and similarly @xmath932 .
this finishes the proof of proposition [ gyts ] , theorem [ gyts ] and thus also of theorem [ gyt ] .
[ rem : inj ] in the case where @xmath89 in theorems [ gyt ] , [ gyts ] and by observing their proof , one can further ensure in theorems [ gyt ] , [ gyts ] that the coordinate projection @xmath933 is finite to one on @xmath934 .
[ struct ] let us comment on how the reparametrization results can be generalized to other fields than the fields @xmath91 of this section , and to other languages than @xmath935 or @xmath936 .
the language @xmath180 can be interpreted naturally in many more fields than just in @xmath91 , as explained below definition [ fnb ] , with still a well - understood geometry of the definable sets by section 3.4 of @xcite .
similarly , instead of @xmath935 or @xmath936 , for @xmath180 we can take any analytic language formed by adding function symbols for the elements of a separated weierstrass system as in @xcite to @xmath500 , further enriched with some constant symbols , and interpret it as an analytic structure on a henselian valued field @xmath752 of characteristic zero , as in @xcite .
suppose that @xmath937 is such a more general structure .
if , furthermore , for sufficiently many @xmath535 , @xmath753 is a finite union of sets of the form @xmath938 for @xmath939 where @xmath940 is the set of @xmath539-th powers in @xmath753 , and if @xmath180 has constant symbols for these @xmath724 , then we strongly expect theorems [ gyt ] and [ gyts ] and proposition [ gyts ] to go through on @xmath937 with similar proofs , where the results of @xcite can be used instead of the quoted results from @xcite .
once we have the parametrizations of definable sets provided by theorem [ gyt ] , we may derive a result analogous to lemma 3.1 of @xcite and lemma 2.1 of @xcite similarly as pila does in @xcite .
we now introduce the following notation , that will be used again in subsequent sections : @xmath941 and @xmath942 , @xmath943 .
thus , @xmath944 and @xmath945 .
[ detest]fix @xmath946 .
let @xmath163 be an open subset of @xmath586 contained in a box which is the cartesian product of @xmath33 closed ball of equal radius @xmath947 .
let @xmath320 , , @xmath948 be points in @xmath163 , and @xmath949 , , @xmath950 be @xmath72-functions @xmath951 .
assume 1 .
the integer @xmath45 satisfies @xmath952 2 . the functions @xmath953 satisfy @xmath544 on @xmath163 .
set @xmath954 then @xmath955 with @xmath956 by hypothesis ( 2 ) , one may write @xmath957 with @xmath958 expanding @xmath959 into the sum of @xmath960 monomial terms of type @xmath961 one gets an expansion of @xmath962 into the sum of @xmath963 terms , the last one being @xmath964 . the columns of the matrix @xmath965 being indexed by @xmath966 , we can write each column in @xmath967 , except for the first one which is @xmath968 , as a sum of @xmath963 columns , and then expanding the determinant one may write @xmath967 as the sum of @xmath969 determinants @xmath970 . for each determinant @xmath971
, we factor out from its columns the factors @xmath972 .
this lets us write @xmath971 as a product of factors @xmath972 and of a determinant @xmath973 with columns @xmath974 , called of order @xmath596 , and columns @xmath975 .
note that if @xmath976 , then @xmath973 can not have two identical columns , and thus can not have more than @xmath977 columns of order @xmath108 , @xmath978 .
now @xmath979 is maximized when the number of columns of type @xmath980 in @xmath973 is maximal , that is for @xmath981 columns of order @xmath108 , @xmath982 .
note , by hypothesis ( 1 ) , that the number of these columns is then @xmath983 . in this case
the degree for the monomial factored out from @xmath971 is @xmath984 , and thus of norm @xmath985 , and the number of the remaining columns in @xmath973 , which are of type @xmath975 , is minimal and equals @xmath986 . by hypothesis ( 2 ) and the ultrametric property of the norm , it follows that @xmath987 .
thus , for such a @xmath970 , @xmath988 . finally , again by the ultrametric property of the norm , the statement follows .
[ [ section-3 ] ] we give in this section a @xmath0-adic version of pila - wilkie s theorem 1.10 of @xcite in the form stated by pila in theorem 3.5 of @xcite , that is the so - called block version .
though the arguments involved in the proofs of both versions are the same , the block version has shown to be more useful in applications ( a reason for the effectiveness of the block version is that the image of a block under a semialgebraic map has a controlled number of @xmath132-dimensional blocks ) . for instance
it allows one to bound the number of points of given algebraic degree over @xmath989 and bounded height . in the @xmath0-adic context ,
such a byproduct of the @xmath0-adic block version of theorem 3.5 of @xcite is still possible and given at the end of the section in theorem [ algheights ] .
we shall work in this section with the language @xmath990 .
thus the @xmath180-definable subsets of @xmath18 are exactly the subanalytic sets .
we say an @xmath180-definable subset @xmath17 of @xmath991 is of dimension @xmath108 at a point @xmath16 if for every small open ball @xmath131 containing @xmath16 , @xmath992 is of dimension @xmath108 .
we say @xmath17 is of pure dimension @xmath108 if it is of dimension @xmath108 at each of its points .
let @xmath17 be an @xmath180-definable subset of @xmath991 .
one defines @xmath993 as the union of all semialgebraic subsets of @xmath17 of pure ( strictly ) positive dimension .
note that this description of @xmath993 coincides with the one given in introduction , since by the curve selection lemma , a point @xmath16 of a given semialgebraic set @xmath17 of positive dimension at @xmath16 is always contained in an algebraic curve @xmath994 such that @xmath995 is of positive dimension at @xmath16 .
we denote by @xmath996 the set of points @xmath997 with @xmath998 , for @xmath999 .
let @xmath1000 be an @xmath180-definable subset of @xmath1001 and denote by @xmath380 the projection of @xmath1000 on @xmath1002 .
for @xmath153 in @xmath380 , we denote by @xmath1003 the fiber of @xmath1004 at @xmath153 .
we view @xmath1000 as a family of definable subsets @xmath1003 of @xmath991 parametrized by @xmath380 .
we call @xmath1000 a definable family of definable subsets of @xmath1005 .
we begin this section by a simple but useful remark relating the norm @xmath1006 and the @xmath0-adic norm @xmath1007 .
[ bound ] let @xmath1008 and @xmath19 be such that @xmath1009 . choosing @xmath1010 such that @xmath1011 , since @xmath1012 , one obtains @xmath1013 .
[ basicest ] fix positive integers @xmath1014 .
then for every integer @xmath1015 there exists an integer @xmath1016 and positive constants @xmath1017 and @xmath1018 such that the following holds . for every real number @xmath1019 , for every open subset @xmath163 of @xmath1020 , for every locally analytic mapping @xmath1021 such that the functions @xmath1022 satisfy condition @xmath544 , the subset @xmath1023 is contained in the union of at most @xmath1024 hypersurfaces of degree @xmath1025 .
furthermore , @xmath1026 as @xmath1027 . with the notation of section [ de ] , set @xmath1028 and fix @xmath45 such that @xmath1029 .
note that @xmath45 is unique , so we can denote it by @xmath1030 .
let @xmath1031 , with @xmath163 an open subset of @xmath1020 , such that the functions @xmath953 satisfy the condition @xmath603 .
let @xmath1032 be a ball of radius @xmath1033 .
consider @xmath1034 points ( possibly with repetition ) @xmath1035 ,
, @xmath1036 in @xmath1037 and the @xmath1034 monomials @xmath1038 , @xmath1039 . by lemma [ prod ] the mappings @xmath1040 , for @xmath1039 ,
satisfy @xmath544 as well as the mappings @xmath1022 , @xmath1041 . by lemma [ detest ]
it follows that @xmath1042 with @xmath1043 on the other hand , since @xmath1044 , for all @xmath1045 and all @xmath1046 , after expanding the determinant @xmath1047 one gets a sum of @xmath1048 integers , each of them having a real norm @xmath1049 , with @xmath1050 .
it follows that the integer @xmath1047 has a real norm @xmath1051 , and finally by remark [ bound ] , one has under the condition @xmath1052 , @xmath1053 by putting together ( [ upbis ] ) and ( [ downbis ] ) , it follows that if @xmath1054 then @xmath1055 .
now the end of the proof is quite similar to that of lemma 1 of @xcite . for @xmath1056 as in ( [ boundonrhobis ] ) ,
note that the matrix @xmath1057 has rank @xmath1058 , with @xmath1059 running over @xmath1060 and for any @xmath1061 .
say that , for instance , in this matrix the columns are indexed by @xmath966 .
let @xmath41 be the maximal rank of @xmath1062 over all @xmath1063 and let @xmath1064 be of rank @xmath41 , for some fixed @xmath1065 and some @xmath1066 of cardinality @xmath41 .
since @xmath1067 , we can choose @xmath1068 .
let us denote by @xmath1069 the determinant of the matrix @xmath1070 , where @xmath1071 and @xmath1072 is @xmath213 augmented by the line @xmath1073 .
the polynomial @xmath102 is not zero since the coefficient of @xmath1074 in @xmath102 is the non zero minor @xmath1075 and the degree of @xmath102 is at most @xmath86 .
furthermore for any @xmath1076 we have @xmath1077 , by definition of the maximal rank @xmath41 . since there exists a constant @xmath1078 depending only on @xmath1079 ( and @xmath0 ) , such that @xmath1020 is covered with @xmath1080 balls of radius @xmath1081 such that ( [ boundonrhobis ] ) holds , we get the required result , since , by a straightforward computation done in @xcite , p. 212
( the constant @xmath131 of @xcite being our constant @xmath1082 ) , for fixed @xmath1014 , @xmath1083 as @xmath1027 .
[ ml ] let @xmath1000 an @xmath180-definable family of @xmath180-definable subsets of @xmath1084 parametrized by @xmath1085 .
assume all fibers have dimension @xmath1086 .
let @xmath1087 .
there exists an integer @xmath1088 and a positive real number @xmath1089 such that , for every @xmath153 in @xmath380 and every @xmath19 , the set @xmath1090 is contained in the union of at most @xmath1091 algebraic hypersurfaces of degree at most @xmath86 .
the argument is quite similar to the one in proposition 6.2 of @xcite .
we assume that all fibers of @xmath1000 have dimension @xmath34 .
take @xmath1092 large enough in order to have @xmath1093 and take @xmath1094 , where @xmath1095 and @xmath1096 are given in lemma [ basicest ] . by theorem [ gyt ]
there exist an integer @xmath106 and a definable family @xmath1097 of locally analytic functions @xmath1098 satisfying @xmath544 and such that @xmath1099 by lemma [ basicest ] , since @xmath1100 , for every @xmath1101 , the set @xmath1102 is contained in @xmath1103 algebraic hypersurfaces of degree at most @xmath86 , with @xmath1104 as in lemma [ basicest ] .
we set @xmath1105 .
we now come to the main result of this section , the @xmath0-adic version of theorem 3.5 of @xcite .
we first define our notion of block , essentially in the same way as in @xcite , up to connectedness .
a block @xmath1106 is either a singleton , or , a smooth definable set of pure dimension @xmath1107 which is contained in a smooth semialgebraic set of pure dimension @xmath86 . in particular , for a block @xmath1108 of positive dimension , one has @xmath1109 .
note that the interior @xmath1110 of a definable set @xmath17 of dimension @xmath250 in @xmath18 is a block , since @xmath1110 is the intersection of itself with the semialgebraic set @xmath1111 . on the other hand
the regular part of a definable set is not always a block .
a family of blocks @xmath1112 is a definable set whose fibers @xmath1113 , for @xmath625 , are blocks in @xmath18 . in proposition [ blocks ] , we consider integer points of bounded height .
we deduce from it theorem [ algheights ] , which is about rational points , and we finally prove the rational version of proposition [ blocks ] in theorem [ blocksq ] .
[ blocks ] let @xmath1114 be an @xmath1115-definable family of @xmath1115-definable subsets of @xmath18 parametrized by a definable set @xmath1116 .
let @xmath1117 be given .
there exist @xmath1118 , a constant @xmath1119 and a family of blocks @xmath1120 such that for any @xmath625 , for any @xmath19 @xmath1121 for @xmath1122 of cardinal less than @xmath1123 . in particular , for all @xmath625 , denoting by @xmath1124 the union over @xmath1125 in @xmath1126 of the sets @xmath1127 with @xmath1128 , one has @xmath1129 and , @xmath1130 we follow the strategy of pila s proof of theorem 3.5 in @xcite . since we want to bound the density of integers points of @xmath1000
, one can assume that @xmath1131 . since our result is true for the union of two families when it is true for each family
, one can assume that the dimension @xmath108 of the fibers @xmath1003 of @xmath1000 is constant and that these fibers are of pure dimension .
we proceed by induction on @xmath108 . for @xmath1132 , the family @xmath1000 has a finite number of points for fibers , and this number is bounded with respect to the parameter @xmath153 , by a constant depending only on the set @xmath1000 .
it follows that @xmath1000 itself is a finite union of families of blocks .
assume now that our statement is true for definable families of fiber dimension @xmath1133 and consider @xmath1000 a definable family in @xmath1134 of fiber dimension @xmath1135 .
we first remark that one can easily assume that the fibers @xmath1003 are not of maximal dimension in @xmath1136 . for this
let us denote by @xmath1137 the regular part of a fiber @xmath1003 , that is the set of points of @xmath1003 at which @xmath1003 is smooth .
now if @xmath1138 , the definable subset @xmath1139 of @xmath1134 is a family of blocks , since a fiber is given by the intersection of the semialgebraic set @xmath1140 with the definable set @xmath1137 .
finally since the definable set @xmath1141 has fiber dimension @xmath1142 , we can apply to this set the induction hypothesis , and obtain our statement in this case . from now on
we assume that @xmath1143 , and we fix @xmath35 . by proposition [ ml ] , for any @xmath625 and for any choice of @xmath1144 coordinates in @xmath18 , the projection @xmath1145 onto the corresponding @xmath1146 subspace of @xmath18 is such that @xmath1147 is contained in the union of at most @xmath1148 hypersurfaces of degree at most @xmath1149 , with @xmath1150 , and with some constant @xmath1119 depending only on @xmath1000 and @xmath342 .
let us denote by @xmath1151 , @xmath1152 , the family of algebraic sets of @xmath1153 defined by intersecting the cylinders in @xmath18 over algebraic hypersurfaces of degree at most @xmath86 in @xmath1154 , for every choice of @xmath1144 coordinates in @xmath18 .
the dimension of a given fiber of @xmath1155 depends on the transversality of the @xmath213 cylinders over hypersurfaces in @xmath1146 that give this fiber by intersecting each other , but this dimension is at most @xmath108 , by transversality of the supplementary coordinates in @xmath18 for each choice of @xmath1144 coordinates .
we now consider the family @xmath1156 in @xmath1157 which , for the sake of simplicity , we shall still denote by @xmath1000 .
the set @xmath1090 is contained in the intersection of the cylinders over the hypersurfaces of @xmath1146 containing the points of @xmath1158 , that is , @xmath1090 is contained in at most @xmath1159 fibers of @xmath1160 for some constant @xmath1161 .
now , as in the proof of theorem 3.5 of @xcite , we stratify @xmath1162 in the following way .
let us consider the definable family @xmath1163 where @xmath1164 is the regular subset of @xmath1165 of dimension @xmath108 , that is to say the set of points of @xmath1165 in the neighbourhood of which @xmath1165 is smooth and of dimension @xmath108 .
then @xmath1166 is a definable family with fiber dimension @xmath1167 and we can apply the induction hypothesis : there exists a family of blocks @xmath1168 such that for all @xmath625 , @xmath1169 , with @xmath1170 and @xmath1171 of cardinal less than @xmath1172 . the same kind of data @xmath1173 , @xmath1174 , @xmath1175 , @xmath1176 and @xmath1177 , @xmath1178 , @xmath1179 , @xmath1180 are in the same way provided by the induction hypothesis for the families @xmath1181 @xmath1182 since these two families have fiber dimension @xmath1167 as well as the family @xmath1166 .
now observe that for @xmath1183 a point of @xmath1162 not in the family @xmath1184 , a sufficiently small semialgebraic neighbourhood @xmath1185 of @xmath16 coincides with @xmath1186 .
the family @xmath1187 is therefore a family of blocks with fibers of dimension @xmath1188 , such that , for @xmath625 , a union of not more than @xmath1159 of them contains the whole set @xmath1189 . denoting @xmath1190 and considering that the parameter spaces @xmath1126 , @xmath1191 , @xmath1192 ,
are all contained in a single parameter space , also denoted @xmath1126 for simplicity , one obtains that , for any @xmath625 , the set @xmath1090 is contained in at most @xmath1159 fibers of @xmath1160 over @xmath1126 , each of them being decomposed in at most @xmath1193 family of blocks , providing the existence of the desired @xmath1108 and @xmath1194 . to conclude for the final statement of the proposition ,
observe again that @xmath1129 follows from the definition of @xmath1195 and that blocks of dimension zero are singletons by definition , which implies the final bound of the proposition . before applying proposition [ blocks ] to algebraic points in @xmath18 of bounded algebraic degree over @xmath989 and bounded height ,
let us recall the notion of polynomial height defined in the introduction that will both encode classical height and algebraic degree over @xmath989 . for @xmath1196 , with @xmath45 and @xmath46 relatively prime integers ,
let @xmath1197 and for @xmath1198 , we set @xmath1199 now for @xmath1200 and for @xmath13 , we denote by @xmath1201 the following element of @xmath1202 @xmath1203 @xmath1204 and for @xmath1205 , we finally set @xmath1206 @xmath1207 for @xmath1000 a definable subset of @xmath18 , @xmath1208 and
@xmath19 a real number , we denote by @xmath1209 the set of points @xmath1210 such that @xmath1211 , and by @xmath1212 the set of points @xmath1210 such that @xmath1213 .
[ algheights ] let @xmath1214 be nonnegative integers .
let @xmath1215 be an @xmath1115-definable family of @xmath1115-definable subsets of @xmath18 parametrized by a definable set @xmath1216 .
let @xmath1117 .
there exists @xmath1152 , a constant @xmath1217 and a family of blocks @xmath1218 such that for any @xmath625 , for any @xmath19 @xmath1219 for @xmath1220 of cardinal less than @xmath1221 .
in particular , for any @xmath625 , denoting by @xmath1222 the union over @xmath1223 of the @xmath1224 of dimension @xmath1225 , one has @xmath1226 and @xmath1227 to prove this statement for all @xmath108 it is obviously enough to prove the similar statement for all @xmath108 where rational points are replaced by integer points , that is to say it is enough to work with @xmath1228 instead of @xmath1229 .
for this goal , let us now consider @xmath1230 and @xmath1231 consider the projection @xmath1232 defined by @xmath1233 , let us write @xmath163 for @xmath1234 , and let us denote by @xmath1235 the projection defined by @xmath1236 . by definable choice , and
since @xmath1237 has fibers of size at most @xmath1238 , there exist @xmath1238 semi - algebraic maps @xmath1239 which are sections of @xmath1237 and such that the union of the graphs of the @xmath1240 equals @xmath1241 .
hence , @xmath1242 and thus , by construction , one has for any @xmath625 that @xmath1243_y(\zz , t ) , y ) ) ) _ y.\ ] ] now , given @xmath1117 and applying proposition [ blocks ] to the definable family @xmath1244 with parameter @xmath625 , we obtain a family of blocks @xmath1245 , such that for any @xmath19 and any @xmath625 , @xmath1246_y ( \zz , t ) \subset \bigcup_{\sigma \in { \cal s}_0 } w_{y,\sigma},$ ] for @xmath1247 of cardinal less than @xmath1248 for some @xmath994 .
since the maps @xmath1249 and @xmath1240 are semi - algebraic , by the definition of blocks , and by dimension theory for @xmath180-definable sets , there exist integers @xmath1250 and @xmath1251 and a family of blocks @xmath1252 such that any set of the form @xmath1253 for any @xmath1223 and any @xmath625 , can be written as the union of no more than @xmath213 blocks of the form @xmath1254 for @xmath1255 . combining with ( [ union ] ) and with the information we have about @xmath1256 , the existence of @xmath1194 with the desired properties follows for this @xmath1257 and for any @xmath19 , with @xmath1258 .
one concludes as for the proof of proposition [ blocks ] .
finally note that theorem [ algheights ] implies in particular the following rational version of proposition [ blocks ] , that differs only in its last line from proposition [ blocks ] .
[ blocksq ] let @xmath1215 an @xmath1115-definable family of @xmath1115-definable subsets of @xmath18 parametrized by a definable set @xmath1216 .
let @xmath1117 .
there exist @xmath1118 , a constant @xmath1119 and a family of blocks @xmath1259 such that for any @xmath625 , for any @xmath19 @xmath1260 for @xmath1122 of cardinal less than @xmath1123 . in particular , for all @xmath625 , denoting @xmath1124 the union over @xmath1223 of the @xmath1127 of dimension @xmath1225 , one has @xmath1129 and @xmath1261
[ [ nr ] ] in this section we shall work over the field @xmath1262 . note however that all our results remain valid with identical proofs when @xmath1263 is replaced by any algebraically closed field of characteristic zero . for each positive integer @xmath45
we denote by @xmath76 _ { < r}$ ] the set of complex polynomials of degree @xmath77 .
let @xmath78 be a subset of @xmath79 .
we denote by @xmath80 the set @xmath81 _ { < r})^n$ ] and by @xmath82 the dimension of the zariski closure of @xmath80 in @xmath83 _ { <
r})^n \simeq \mathbb{c}^{nr}$ ] . similarly ,
when @xmath17 is an algebraic subvariety of @xmath84 , we shall write @xmath1264 for @xmath1265 and @xmath1266 for @xmath1267 . up to a @xmath1269-linear coordinate change
, there is a coordinate projection @xmath1270 whose restriction to @xmath17 has finite fibers . the projection @xmath0 induces a map @xmath1271 _ { < r})^n \to ( \mathbb{c } [ t ] _ { < r})^m$ ] . since @xmath1264 is a constructible subset of @xmath83
{ < r})^n \simeq \mathbb{c}^{nr}$ ] , and @xmath1272 has finite fibers on @xmath1264 , the estimate follows . the following result shows the basic bound can be improved , as soon as @xmath17 is not a linear subspace . by the degree of an irreducible affine variety @xmath17 over a field
@xmath108 we mean the number of intersection points when intersecting @xmath1274 with a generic affine space over @xmath1275 of dimension equal to the codimension of @xmath17 , for some algebraic closure @xmath1275 of @xmath108 .
the improved bound then reads as the trivial bound for @xmath1276-dimensional varieties plus @xmath1277 , rounded up . for the plane curve @xmath17 given by @xmath1279 , for every positive integer @xmath45
, one has the equality @xmath1280 thus , taking the product of @xmath17 with an affine space of dimension @xmath1281 , one sees that the upper bound given in theorem [ motivicbp ] is optimal for any value of @xmath33 , @xmath86 and @xmath45 . by simple projection and section arguments
la lang - weil , one reduces , completely analogously as in @xcite , to the case of plane curves ( @xmath1282 and @xmath1283 ) .
for the sake of completeness let us provide some more details .
assume first @xmath1283 and @xmath317 .
linear projections @xmath1284 are written in coordinates as @xmath1285 , @xmath1286 . for the @xmath25 s and @xmath1287 s in a dense open subset @xmath1288 of @xmath1289
, @xmath1290 is surjective and @xmath17 and @xmath1291 have the same degree . by corollary [ trivcor ] ,
@xmath1292 is non empty .
thus take @xmath1290 corresponding to some point in @xmath1293 .
the number of points in the fibers of @xmath1294 is finite and @xmath1295 is contained in @xmath1296 , thus the statement for @xmath17 follows from the one for @xmath142 .
now assume @xmath1297 . by a similar argument ,
after projecting , one may assume @xmath1298 .
in the linear space of hyperplanes @xmath1299 with equations @xmath1300 , @xmath1301 is irreducible of degree @xmath86 outside a closed subset @xmath1302 of positive codimension .
thus , by corollary [ trivcor ] , for some @xmath1303 , @xmath1304 and @xmath1305 , all in @xmath1269 , the corresponding @xmath1299 is not in @xmath1302 .
consider the pencil @xmath1306 of hyperplanes @xmath1300 , @xmath1307 .
since @xmath1308 is not in @xmath1302 , the pencil intersects @xmath1302 in at most @xmath1082 points @xmath1309 such that @xmath1310 lies in @xmath1302 .
if @xmath1311 for some @xmath1309 we are done , so we may assume that , for such a @xmath1309 , @xmath1312 is of dimension @xmath1313 , thus @xmath1314 by the trivial estimate .
for the other @xmath305 s one may apply the induction hypothesis , which gives @xmath1315 . since the linear form @xmath1316 induces a constructible mapping @xmath1317 _ { < r}$ ] , the statement follows by additivity of dimensions .
let @xmath106 be a field . for @xmath1318 ,
we denote by @xmath1319_s$ ] the vector space of homogeneous polynomials of degree @xmath46 .
thus @xmath1319_s$ ] is of dimension @xmath1320 over @xmath106 .
let @xmath1321 be a homogeneous ideal of @xmath1319 $ ] and set @xmath1322_s$ ] .
we set @xmath1323_s /i_s$ ] .
it is the hilbert function of @xmath612 .
let @xmath114 be a monomial ordering on @xmath1319 $ ] in the sense of @xcite def 1 , ch .
denote by @xmath1324 the ideal generated by the leading terms for the ordering @xmath114 of elements of @xmath612 , where the leading term of a homogeneous polynomial @xmath1325 is the term among the nonzero @xmath1326 which is maximal for the ordering . by @xcite prop 9 , ch .
9.3 , @xmath612 and @xmath1324 have the same hilbert function .
let @xmath17 be an irreducible subvariety of dimension @xmath33 and degree @xmath86 of @xmath1330 defined by a homogeneous ideal @xmath612 .
then , for @xmath46 large enough , @xmath1331 is equal to @xmath1332 with @xmath1333 the hilbert polynomial of @xmath17 .
it is a polynomial of degree @xmath33 , leading coefficient @xmath1334 and coefficients bounded in terms of @xmath250 and the degrees of generators of @xmath612 . as explained in @xcite and @xcite
, it follows there exist non - negative real numbers @xmath1335 , @xmath1336 , such that @xmath1337 as @xmath1338 .
note that @xmath1339 [ grevlex]let @xmath17 be a closed equidimensional subscheme of dimension @xmath33 of @xmath1330 .
assume @xmath17 intersects properly the hyperplane @xmath1340 , that is , no irreducible component of @xmath17 is contained in @xmath1340 .
let @xmath114 be the monomial ordering defined as follows : @xmath1341 if @xmath1342 or if @xmath1343 and for some @xmath272 , @xmath1344 and @xmath1345 , for @xmath1346 .
( that is , after reindexing the coordinates , @xmath114 is the reverse graded lexicographic order . ) then @xmath1347 let @xmath1348 and @xmath17 be an irreducible curve in @xmath1349 of degree @xmath86 .
consider the map @xmath1350 and the corresponding embedding @xmath1351 \end{cases}\ ] ] and let @xmath612 denote the homogeneous ideal of the closure of @xmath1352 in @xmath1353 .
let us form the leading term ideal @xmath1324 of @xmath612 for the monomial ordering @xmath114 of lemma [ grevlex ] for @xmath316 .
let @xmath67 be given . fix a positive integer @xmath344 .
set @xmath1354 set @xmath1355 , @xmath1356 for @xmath1357 , and @xmath1358 .
let us write @xmath1359 .
by theorem [ gyt ] , there exists a surjective @xmath1360-definable function @xmath1361 for some integer @xmath206 such that for each @xmath1362 , @xmath1363 satisfies @xmath1364 on @xmath1365 .
fix an integer @xmath1366 , and let @xmath1367 be a closed ball of valuative radius @xmath1059 in @xmath1368 . fix @xmath1369 and , for any choice of points @xmath1370 for @xmath1371 in @xmath1372 ,
consider the determinant @xmath1373 by lemma [ detest ] for @xmath89 and @xmath316 and lemma [ prod ] , we get that @xmath1374 on the other hand , recall that @xmath1375 $ ] belongs to @xmath1376_{<r}$ ] if and only if @xmath1377 , where @xmath1378 stands for the degree in @xmath189 , and hence , @xmath1379 thus , if @xmath1380 , @xmath1381 by putting together ( [ upm ] ) and ( [ downm ] ) , it follows that if @xmath1382 then @xmath1383 .
for such an @xmath1059 , note that the matrix @xmath1384 with @xmath966 running over @xmath1385 and @xmath1370 in @xmath1386 for @xmath1387 , has rank @xmath1058 .
hence , by the arguments in the proof of lemma 1 in @xcite ( which are quite similar to those from the proof of lemma [ basicest ] ) , there exists a nonzero polynomial @xmath1299 in two variables with coefficients in @xmath1388 $ ] and exponents in @xmath1389 which vanishes at all the @xmath1370 , and thus at all points of @xmath1390 .
note that @xmath1299 does not vanish identically on @xmath17 since its exponents lie in @xmath1389 and that its degree is at most @xmath344 .
recall that @xmath97 is given and we want to prove that @xmath1391 .
we will prove this bound by choosing @xmath1392 following @xcite . by properties of hilbert polynomials recalled in section [ hil ] , we have @xmath1393 @xmath1394 and thus @xmath1395 where the @xmath1396 notation is for @xmath344 going to @xmath48 .
thus , @xmath1397 by lemma [ grevlex ] we find @xmath1398 hence , there exist integers @xmath1399 and @xmath1400 , both depending on @xmath45 , such that @xmath1401 now we are ready to bound @xmath1402 , using this choice of @xmath1403 and @xmath1059 . note that @xmath1264 is zariski closed in @xmath1404 _ { < r})^2\simeq \cc^{2r}$ ] being an intersection of zariski closed subsets .
write @xmath1405 for the projection , where @xmath542 is the maximal ideal of @xmath1368 . by remark [ rem : inj ]
, we may suppose that the projection @xmath1406 is finite to one on @xmath380 , and that @xmath380 is a cell over @xmath1407 .
this implies @xmath1408 where @xmath1409 is considered as a constructible subset of @xmath1410 .
the set @xmath1411 of all @xmath1412 in @xmath1413 such that there exists @xmath625 with @xmath1414 and @xmath1415 is a constructible subset of @xmath1416 by proposition [ cons ] . since @xmath212 is surjective , the projection @xmath1417 is also surjective . by construction ,
the projection @xmath1418 has finite fibers ( of size at most @xmath1419 by bzout s theorem ) : indeed , a point in @xmath1420 corresponds to the choice of @xmath1421 together with a ball @xmath1422 as above , and we have shown that the fiber of the projection @xmath1418 over this point is contained in the intersection of @xmath17 with the zero locus of a polynomial @xmath1299 in two variables , with coefficients in @xmath76 $ ] and degree at most @xmath344 , that does not vanish identically on the curve @xmath17 .
it follows that @xmath1423 .
the following proposition is deduced from a result of @xcite on quantifier elimination in an expansion of @xmath180 which includes more auxiliary sorts , namely all the @xmath1424 for integers @xmath1425 ( not to be confused with the @xmath1426 for @xmath545 ) .
[ cons ] let @xmath1427 be an @xmath1360-definable set , let @xmath1425 and @xmath97 be integers and let @xmath1428 be the projection . write @xmath1429 for the intersection of @xmath416 with @xmath1430
. then @xmath1429 , seen as subset of @xmath1431 , is definable in the ring language with coefficients from @xmath1263 .
let @xmath1432 be the language @xmath500 enriched with the auxilliary sorts @xmath1433 for each integer @xmath1425 , ( higher order ) angular component maps @xmath1434 sending nonzero @xmath16 to @xmath1435 and zero to zero , and the bijections from @xmath1433 to @xmath1436 sending @xmath1437 to the tuple @xmath904 . ( the maps @xmath1438 should not be confused with the maps @xmath510 introduced before . ) one has quantifier elimination for all sorts in the language @xmath1432 by @xcite[thm . 4.2 ] .
it follows that @xmath416 is @xmath1432-definable without quantifiers .
moreover , in the variables running over @xmath1439 , finitely many polynomials @xmath1440 over @xmath106 can occur in the formula describing @xmath416 , and we may suppose they occur as arguments of @xmath1441 for some @xmath1059 and of @xmath1442 .
note that , by the geometry of definable sets as summarized by the cell decomposition result , the @xmath1443 take only finitely many values on @xmath1444 .
now the lemma follows by syntactical analysis of quantifier free formulas describing @xmath416 , in relation with the extra condition that @xmath16 lies in @xmath1444 .
[ remdim ] in fact , the inequality ( [ dimpy ] ) holds in a wider generality .
consider an @xmath1360-definable set @xmath1445 and suppose that the projection @xmath1446 is finite to one on @xmath17 , where @xmath1447 .
suppose further that @xmath11 is of dimension @xmath33 .
let @xmath1425 be an integer and write @xmath1448 for the projection which is @xmath1449 on the last @xmath250 coordinates .
then one has @xmath1450 indeed , ( [ dimpx ] ) is easy to show when @xmath17 is a cell , and follows by the cell decomposition theorem [ cda ] in general .
when @xmath1451 with @xmath1447 , one may wonder whether one can bound @xmath1452 in terms of @xmath97 , when @xmath1453 is the transcendental part of a definable subset @xmath17 in @xmath1111 , and with notation from section [ nr ] .
as usual , @xmath1453 is @xmath1454 , where the algebraic part @xmath29 of @xmath17 is defined as the set of those points @xmath722 through which there exists a semi - algebraic ( namely @xmath1360-definable ) @xmath520 of dimension @xmath32 such that @xmath1455 is locally around @xmath16 of dimension @xmath32 .
suppose for contradiction that there is an algebraic curve @xmath1458 with infinite intersection with @xmath1460 .
let us write @xmath1461 for the subset of @xmath1462 of @xmath106-rational points on @xmath994 .
further , let us write @xmath520 for the image of @xmath1463 under the projection @xmath1464 since @xmath1463 is an @xmath1360-definable set of dimension at most @xmath32 , and since @xmath520 is its image under an @xmath1360-definable function , the dimension of @xmath520 is at most equal to @xmath32 .
since @xmath1465 contains an infinite subset of @xmath994 , the dimension of @xmath1465 equals @xmath32 . here
, we have used the natural inclusion @xmath1466 _
{ < r})^n \subset \cc\llb t \rrb^n = \co_k^n.\ ] ] moreover , @xmath1465 is of local dimension @xmath32 at all but finitely many of its points , since it is an @xmath180-definable set .
hence , @xmath1453 is contained in the union of a finite set with @xmath1467 . since @xmath1468 and
since @xmath1469 contains @xmath994 , @xmath1460 can not have infinite intersection with @xmath994 .
finally , let us mention that it seems quite difficult to give sharp bounds on the size of the set @xmath1460 in terms of @xmath45 in general , for @xmath17 of large dimension . under some extra conditions on @xmath17 , like with some non - archimedean analogues of restricted pfaffians instead of the full subanalytic language on @xmath106 , one may hope there exist results for ( low - dimensional ) definable sets , similar to e.g. the results in @xcite for real pfaffian curves .
in this section we shall discuss related results over @xmath1471 .
although our methods are restricted to characteristic zero , due to our limited understanding of the structure of definable sets over henselian fields of positive characteristic , it is possible to use standard methods to deduce from theorem [ motivicbp ] asymptotic bounds for the number of rational points in @xmath1472 $ ] of bounded degree .
this provides in particular a partial answer to a question raised by cilleruelo and shparlinski in @xcite problem 9 , about possible analogues over @xmath1471 of the bombieri - pila bound .
as noticed in remark [ cohenff ] , note however that an analogue of s. d. cohen s bound holds over @xmath1471 .
let @xmath1473 be an algebra essentially of finite type over @xmath1474 , i.e. the localization of a finitely generated @xmath1474-algebra .
we assume @xmath1473 is an integral domain of characteristic zero and we denote by @xmath106 its fraction field .
we consider the category @xmath1475 of ring morphisms @xmath1476 with @xmath158 a field , i.e. the category of field endowed with an @xmath1473-algebra structure .
if @xmath1477 is an @xmath1473-scheme , and @xmath1476 a ring morphism , we denote by @xmath1478 the @xmath158-scheme obtained by base change to @xmath158 .
we consider the affine space @xmath1479}$ ] and @xmath17 a closed subscheme whose ideal is generated by polynomials @xmath1480[x_1 , \dots , x_n]$ ] . for any positive integer @xmath45 ,
we denote by @xmath1481 _ { < r}$ ] the set of polynomials with coefficients in @xmath158 and degree @xmath77 .
we identify @xmath1481 _ { < r}$ ] with @xmath1482 and @xmath1483 _
{ < r})^n$ ] with @xmath1484 .
[ represent ] let @xmath45 be a positive integer . the functor @xmath1485 sending @xmath1476 to @xmath1486 ) \cap ( f[t ] _ { <
r})^n \subset f^{rn}\ ] ] is representable by a closed subscheme @xmath1487 of @xmath1488 .
in particular for any @xmath158 in @xmath1475 we have a natural identification of @xmath1489 with @xmath1490 inside @xmath1491 . after inserting the polynomials @xmath1492 into the system of equations @xmath1493 and developing ,
one gets a system of polynomial equations in the variables @xmath1494 .
the corresponding closed subvariety @xmath1487 of @xmath1488 represents the functor @xmath1264 .
if @xmath108 is a finite field , we denote by @xmath1495 its characteristic and @xmath1496 its cardinality .
we shall use the following lemma , which is a consequence from statements in @xcite based on the lang - weil estimate .
let @xmath1497 be a closed subscheme of @xmath1498 .
let @xmath1499 .
there exists positive integers @xmath1500 , @xmath994 and @xmath213 , such that , for any @xmath1501 in @xmath1475 with @xmath108 finite , if @xmath1502 and @xmath1503 , then for some @xmath1504 and some @xmath1505 , @xmath1506 it follows from proposition 3.3 and proposition 4.9 in @xcite that there exist positive integers @xmath994 and @xmath213 such that , for any @xmath1501 in @xmath1475 with @xmath108 finite , if @xmath1503 , @xmath1507 for some @xmath1505 and @xmath344 the dimension of the zariski closure of @xmath1508 in @xmath1509 .
in particular , @xmath1510 .
since , for some @xmath1500 , @xmath1511 whenever @xmath1512 , the statement follows .
[ ffbp]let @xmath1473 be an algebra essentially of finite type over @xmath1474 and assume @xmath1473 is an integral domain of characteristic zero .
let @xmath106 be the fraction field of @xmath1473 and @xmath1513 an algebraic closure of @xmath106 .
let @xmath17 be a closed subscheme of @xmath1479}$ ] .
assume @xmath1514 is irreducible of dimension @xmath33 and degree @xmath86 .
fix a positive integer @xmath45 .
there exist positive integers @xmath1500 , @xmath994 and @xmath213 , such that , for any @xmath1501 in @xmath1475 with @xmath108 finite , if @xmath1502 and @xmath1515 , then @xmath1516 for some @xmath1517 and some @xmath1505 . from theorem [ motivicbp ] , which stills hold when replacing @xmath1263 by @xmath1513 by remark [ notonlyc ]
, it follows that @xmath1518 . since @xmath1519 ,
the statement is then a direct consequence of lemma [ langweil ] applied to @xmath1520 .
[ cohenff]when @xmath1521 , theorem [ ffbp ] provides a non - trivial improvement on the `` trivial '' bound with @xmath1522 .
note however that , using the function field version of the large sieve inequality due to hsu @xcite instead of the one used in @xcite , one can easily adapt the arguments given in @xcite to get the following function field analogue of s. d. cohen s result in @xcite : if @xmath17 is an irreducible subvariety of @xmath1523 of dimension @xmath33 and degree @xmath1524 , then @xmath1525 , with @xmath1526 ) \cap ( \mathbb{f}_q[t ] _ { < r})^n$ ] .
r. cluckers , i. halupczok , _ approximations and lipschitz continuity in @xmath0-adic semialgebraic and subanalytic geometry _ , selecta math .
* 18 * ( 2012 ) , 825837 .
r. cluckers , l. lipshitz , _ fields with analytic structure _ , j. eur .
* 13 * ( 2011 ) , 11471223 . c. hsu , _ a large sieve inequality for rational function fields _ , j. number theory * 58 * , ( 1996 ) , 267287 .
a. macintyre , _ on definable subsets of @xmath0-adic fields _ , j. symbolic logic * 41 * ( 1976 ) , 605610 .
o. marmon , _ a generalization of the bombieri - pila determinant method _ , zap .
nauchn . sem .
otdel . mat .
inst . steklov .
( pomi ) * 377 * ( 2010 ) , issledovaniya po teorii chisel . *
10 * , 6377 , 242 ; translation in j. math .
( n. y. ) * 171 * ( 2010 ) , 736744 . | we prove an analogue of the yomdin - gromov lemma for @xmath0-adic definable sets and more broadly in a non - archimedean , definable context .
this analogue keeps track of piecewise approximation by taylor polynomials , a nontrivial aspect in the totally disconnected case .
we apply this result to bound the number of rational points of bounded height on the transcendental part of @xmath0-adic subanalytic sets , and to bound the dimension of the set of complex polynomials of bounded degree lying on an algebraic variety defined over @xmath1 , in analogy to results by pila and wilkie , resp . by bombieri and pila . along the way we prove , for definable functions in a general context of non - archimedean geometry , that local lipschitz continuity implies piecewise global lipschitz continuity . |
Subsets and Splits